PHILOSOPHY
Core Questions Philosophy • 1n
A Text with Readings
Elliott Sober University of Wisconsin-Madison
PEARSON Boston Coltunbus Indianapolis New York San Francisco Upper Saddle River A1nsterdam Cape Town Dubai London Madrid Milan Munich Paris Montreal Toronto Delhi Mexico City Sao Paulo Sydney Hong Kong Seoul Singapore Taipei Tokyo
For Aaron
Editorial Director: Craig Ca1npanella Editor in Chief: Ashley Dodge Editorial Product Manager: Carly Czech Editorial Project Manager: Reena Dalal Editorial Assistant: Nicole Suddeth Vice President/Director of Marketing: Brandy
Dawson Executive Marketing Manager: Kate Mitchell Marketing Assistant: Paige Patunas Digital Media Editor: Tho111as Scalzo
Production Manager: Meghan DeMaio Creative Director: Jayne Conte Cover Designer: Mary Siener Cover linage: © Gihnanshin/Shutterstock Editorial Production and Co1nposition Service:
Sudha Balasundara1n/S4Carlisle Publishing Services
Interior Design: Joyce VVcston Design Printer/Binder/Cover Printer: Courier Co1npanies
Credits and ackt10\vledg1nents borro,ved fron1 other sources and reproduced, with pennission, in this textbook appear on the appropriate page within the text.
Copyright© 2013, 2009, 2005, 2001 by Pearson Education, Inc. All rights reserved. Printed in the United States of A1nerica. This publication is protected by Copyright and pennission should be obtained fron1 the publisher prior to any prohibited reproduction, storage in a retrieval systcn1, or transn1ission in any fonn or by any 1neans, electronic, 1nechanical, photocopying, recording, or likewise. To obtain pennission(s) to use 111aterial frorn this \Vork, please sub1nit a \Vritten request to Pearson Education, Inc., Pennissions Depart1nent, One Lake Street, Upper Saddle River, New Jersey 07458 or you n1ay fax your request to 201-236-3290.
Many of the designations by nlanufacturers and seller to distinguish their products are daitned as tradernarks. Where those designations appear in this book, and the publisher \Vas a\vare of a trade111ark clahn, the designations have been printed in initial caps or all caps.
Library of Congress Cataloging-in-Publication Data Sober, Elliott.
Core questions in philosophy: a text with readings/Elliott Sober.-6th ed. p.cm.
Includes bibliographical references (p. ) and index. ISBN-13: 978-0-205-20669-8 ISBN-10: 0-205-20669-7 i. Philosophy-ll)troductions. I. Title. BD2i.S615 2013 100-dc23
10 9 8 7 6 5 4 3 2 1
PEARSON
2012022769
Student Edition ISBN-10: 0-205-20669-7 ISBN-13: 978-0-205-20669-8
Instructor Edition ISBN-10: 0-205-20672-7 ISBN-13: 978-0-205-20672-8
a La Carte Edition ISBN-10: 0-205-19135-5 ISBN-13: 978-0-205-19135-2
Contents
Boxes xv111 Preface xix Acknowledgments xx MySearchLab Connections in This Text xxii
Part One. Introduction i
Chapters
1. What Is Philosophy? 3 Examples 4 Three Theories about What Philosophy Is 6 The Nature of Philosophy Has Changed Historically 7 Philosophical Method 8
MySearchlab Connections 9 CEJ-{Read on MySearchlab The Value of Philosophy, Bertrand Russell
2. Deductive Arguments 11 Argu1nents 11 Good Arguments 12 Deductive Validity Defined 13 "Validity" Is a Technical Term 13 Logical Form 14 Invalidity 14 Testing for Invalidity 15 Circularity, or Begging the Question 18 Truth 18 "True for Me" 20 Wishful Thinking 20 Self-Fulfilling Prophesies 20
MySearchlab Connections 22
3. Inductive and Abductive Arguments 24 Deductive Validity Is a Limitation 25 Nondeductive Inference-A Weaker Guarantee 26 Two Gambling Strategies 26 Universal Laws 26 Detective Work 27
v
vi Contents
''.' ','.' '' '','' '''''' '''' '''' ''''''''
Induction 27 1\vo Factors Influence Inductive Strength 28 Abduction 28 Inferring What Isn't Observed 29 Abduction Differs from Induction 29 Can You Deduce the Explanation front the Observations? 30 Deducing Observational Predictions from a Theory 30 When the Prediction Comes True 30 When the Prediction Turns Out to Be False 31 How True Predictions and False Predictions Are Interpreted 31 The Surprise Principle: When Does Successful Prediction
Provide Strong Evidence? 32 Evidence May Discrilninate bet\veen So1ne Hypotheses
While Failing to Discriminate between Others 34 True Prediction Isn't Enough 35 Modest Favoring 36 The Surprise Principle Summarized 36 The Only Game in Town Fallacy 36
MySearchLab Connections 39
Suggestions for Further Reading 39
Part Two. Philosophy of Religion 41
Chapters
4. Aquinas's First Four Ways 43 IJ!l-[Read on MySearchLab Five Ways to Prove that God Exists,
fro1n Sun1111a Theologiae, Tho1nas Aquinas The Concept of God 44 The First Two Arguments: Motion and Causality 44 Aquinas on the Cause of Motion 46 God Is a Person, Not Just a Cause That Exists Outside of Nature 46 The Birthday Fallacy 46 Why Can't Nature Be Infinitely Old? 47 Why Must Every Event in Nature Have a Cause? 48 The Third Argument: Contingency 48 Necessary and Contingent Beings 48 Possible Worlds 48 Reductio Ad Absurdum 50 Contingency and Eternity 51 Conservation La\vs in Physics 52 The Birthday Fallacy (Again) 52 Necessary Beings other than God 52
Necessary and Contingent Propositions 53 Mathematical Truths 53 Names Differ from the Things Named 53 Nun1bers Aren't Nun1erals 54
Sets 54 Necessity and Certainty Are Different 55 Numbers Are Necessary Beings 56 Aquinas's Fourth Argument: Properties That Come in Degrees 56 Criticizing an Argun1ent versus Sho\ving the Argu1nent's
Conclusion Is False 57
MySearchlab Connections 58
5. The Design Argument 59 Goal-Directed Systems 60 Two Kinds of Design Argument 60 Paley's Watch 61 ~Read on MySearchLab The Design Argurnent, fro1n
Natural Theology, William Paley The Analogy 62 Abductions Arguments Often Postulate Unobserved Entities 62 Hume's Criticisms of the Design Argument 63 ~Read on MySearchLab Dialogues Concerning Natural Religion,
David Hume Is the Design Argument a Weak Argument from Analogy? 63 Is the Design Argument a Weak Induction? 65
MySearchlab Connections 67
6. Evolution and Creationism 68 Creationism 69 Some Creationist Argu1nents 69 Darwin's Two-Part Theory 70 Natural Selection 71 Speciation 73 The Tree ofLife 73 The Principle of the Common Cause 74 Arbitrary Similarities among Organisms 75 Useful Similarities among Organisms 75 Irreducible Complexity 76 Is Creationism Testable? 78 Predictive Equivalence 78 Prediction versus Accon1n1odation 79 Does Evolutionary Theory Make Novel Predictions? Bo Concluding Remarks 81
MySearchlab Connections 82
Contents vii
viii Contents
. '' ..... ' ... '.'.''.'' ' .. '' ... '.,'' ' ... ''.'''.'.','''.' .... ' '''' .... ' ... ''' 7. Can Science Explain Everything? 83
Scientific Ignorance 84 The Only Game in Town Fallacy 85 The Two Questions 85 What Is a Scientific Explanation? 85 A Thesis about Explanation 86 Why Is There Something Rather than Nothing? 87 Can Physics Explain the Origin of the Universe? 87 Leibniz: God Chooses Which Possible World to Actualize 87 Clarke: God Explains Why the Actual World Consists
of One Total History Rather than Another 88 The Only Game in Town Fallacy, Again 88 Causality 89 The Principle of Sufficient Reason 90
MySearchLab Connections 91
8. The Ontological Argument 92 A Posteriori and A Priori 92 Definitions and Existence 93 Anselm's Argument 93 CE!-[ Read on MySearchLab Debate, Guanilo and Anselm Gaunilo's Criticism 95 Anselm's Reply 97 Dispensing with Perfection 98 Conclusion 98
MySearchlab Connections 100
9. Is the Existence of God Testable? 101 Logical Positivis1n 101 The Testability Theory of Meaning 102 Analyticity 102 Falsifiability 102 Auxiliary Assumptions Needed 104 Auxiliary Assumptions Must Be Independently Established 106 "God Exists" Is Meaningful 106 CE!-[ Read on MySearchlab Of Miracles, from An Enquiry Concerning
Human Understanding, David Hume MySearchlab Connections 108
10. Pascal and Irrationality 109 Prudential and Evidential Reasons for Belief 109 CE!-[ Read on MySearchLab The Wager, from Pensees, Blaise Pascal When Does It Make Sense to Gamble? no Pascal's Argument 111
Contents ix
'" .. '.' ...... ,.'', '''. ''''.' '' .. "'' '' ... ' '',. ', .. ' '.' '.''.''.' ...... '''., First Criticis1n of Pascal's Argu1nent 112 Second Criticism of Pascal's Argument 113 The Role of Reason 113 Freud's Psychological Explanation of Theism 114
A New Prudential Argument 115 Prag1natism 115
i'.E-CRead on MySearchlab The Will to Believe, William James CEJ-[Read on MySearchlab 111e Ethics of Belief, W. K. Clifford MySearchlab Connections 118
11. The Argument from Evil 119
First Version of the Argument 119 Two Kinds of Evil 120
Possible Reactions to the Argument 120 Theodicy and Defense 121 Soul-Building Evils 121 Second Version of the Argument 123 Free Will 123
Examples and a Third Version of the Argument 123 A Criticism of the Argument 124 Testability, Again 125 Another Kind of Argument-The Evidential Argument from Evil 125
MySearchlab Connections i27 i'.E-CRead on MySearchlab Theodicy, Gottfried Leibniz
Suggestions for Further Reading 127
Readings 129
Five Ways to Prove That God Exists, Thomas Aquinas 129 The Design Argument, William Paley 129 Critique of the Design Argument, David Hume 129 The Ontological Argument, Anselm 130 The Meaninglessness of Religious Discourse, A.). Ayer 130 Belief in God-What Do You Have to Lose?, Blaise Pascal 134 The Will to Believe, William James 134
Part Three. Theory of Knowledge 135
Chapters
12. What Is Knowledge? i37 Epistemology 137 Three Kinds of Knowledge 138
x Contents
Two Requirements for Knowledge: Belief and Truth 140 Plato: True Belief Isn't Sufficient for Knowledge 140 Justification 141 The JTB Theory 141 Three Counterexamples to the )TB Theory 142 What the Counterexa1nples Have in Con11non 143
An Argument for Skepticism 143 Problems for Further Thought 145
MySearchlab Connections 145
13. Descartes' Foundationalism 147 Foundationalis1n 147 t'.1!1-[Read on MySearchLab Meditations 1-5 of Meditatio11s 011
First Philosophy, Rene Descartes Euclid's Parallel Postulate 148 Descartes' Method of Doubt 149 The Method Applied to a Posteriori Beliefs 150 Dubitability Is a Logical, Not a Psychological, Property 150 The Method Applied to Beliefs Based on Rational Calculation 150 I Am Thinking, Therefore I Exist 151 Thesis of the Incorrigibility of the Mental 152 Do First-Person Psychological Beliefs
Provide a Sufficient Foundation? 153 An Additional Foundational Belief: God Exists and Is No Deceiver 154 How to Prove that God Exists 154 The Clarity and Distinctness Criterion 156 The Cartesian Circle 157 Conclusion 157 MySearchlab Connections 159
14. The Reliability Theory of Knowledge 161 Descartes: Knowledge Is Internally Certifiable 161 What Makes a Thermometer Reliable? 162 Relevance to the Problem of Knowledge 164 Three Concepts oflmpossibility 164 To Have Knowledge, You Don't Have to Be Able to
Construct a Philosophical Argument Refuting the Skeptic 165 A Consequence of the Reliability Theory 167 Thesis of the Relativity of Knowledge 168 What Does the Relativity Thesis Say about Skepticism? 169
MySearchlab Connections 171
15. Justified Belief and Hume's Problem oflnduction 172 Knowledge versus Justified Belief 172
Contents xi
' ... '' '''''''' '''' '''',' ''' '''''' ', .. ' , '''.' '',' . ''.'.' ' ... '.'. , '''''. Skepticism about Justified Belief 173 t:E:J-[Read on MySearchlab Section IV of An Enquiry Concerning
Hu111an Understanding, David Htune Hume's Skeptical Thesis about Induction 174 Hume's Argument that Induction Can't Be Rationally Justified 175 Why Caiit PUN Be Justified? 176 Sununary of Hume's Argu1nent r76
MySearchlab Connections 177
16. Can Hume's Skepticism Be Refuted? 178 What, Exactly, Does the Principle of the Uniformity of Nature Say? 178 A New Concept: Degrees of Reliability 179 What Is a Rule oflnference? 180 Does the Past Reliability of Induction Provide an Answer? 180 Hume's Argument Reformulated 181 Strawson: It Is Analytic that Induction Is Rational 181 Black: Induction Can Be Inductively Justified 182
MySearchlab Connections 184
17. Beyond Foundationalism 185 Hume's Proble1n and Descartes' Proble1n 185
Whether X Is Evidence for Y Depends on Background Assumptions Z 187 Another Relativity Thesis 188 Foundationalism Leads to Skepticism 188 A Nonfoundationalist Approach to Justification 189 Standards of Justification Often Depend on the Audience 189 Two Metaphors-Building a Building and Repairing a Raft 190
MySearchlab Connections 191
18. Locke on the Existence of External Objects 192 t:E:J-[Read on MySearchlab Chapter u of Essay Concerning
Human Understanding, john Locke Locke's First Argument-"Those That Want the Organs of Any Sense" 193 Locke's Second Argument-"Ideas Which Force Themselves upon Me" 194 Locke's Third Argument-"Pleasure or Paiii' 194 Locke's Fourth Argument-"Our Senses Assist One
Another's Testimony" 195
MySearchlab Connections 196
Suggestions for Further Reading 196
Readings 198
The Theaetetus-Knowledge Is Something More than True Belief, Plato 198 Meditations 011 First Philosophy, Rene Descartes 198
xii Contents
Induction Cannot Be Rationally Justified, David Hume 199 Essay Concerning Human Understanding, john Locke 199
Part Four. Philosophy of Mind 201
Chapters
19. Dualism and the Mind/Body Problem 203 What ls the Mind/Body Problem? 204 Descartes' Dualis1n 204 The Mind/Brain Identity Theory 204 CEJ--[Read on MySearchLab Meditation VI, from Meditations
on First Philosophy, Rene Descartes Immortality of the Soul 205 Leibniz's Law 205 Descartes' First Argument for Dualism-The Indubitable
Existence Argun1ent 206 An Analogy 207 Propositional Attitudes and Aboutness 207 Descartes' Second Argument for Dualism-The Divisibility Argument 210 Causality between the Physical and the Nonphysical 210 CEJ-[Read on MySearchlab Correspondence with Princess Elizabeth,
Rene Descartes
MySearchlab Connections 212
20. Logical Behaviorism 214 The Attack on "the Ghost in the Machine" 215 Logical Behaviorism Says Mentalism Is False
Because It Leads to Skepticism 215 Do We Know about the Mental States of Others
by Analogy with Our Own Case? 216 Abduction 216 Logical Behaviorism's Positive Thesis-Its Analysis of
Mentalistic Vocabulary 217 The Dispositional Analysis of Desire ls Incomplete 218 A Dispositional Analysis Does Not Refute Mentalism 218
MySearchlab Connections 220
21. Methodological Behaviorism 221 The Negative Thesis: Psychology Should Avoid Belief/Desire
Explanations 222 Methodological Behaviorism's Positive Thesis 223 First Objection to Behaviorism's Positive Thesis: Novel Behaviors 224
''. ' •• '' 0'','''' ''' '''' '' '' ''''''''
Second Objection to Behavioristn's Positive 'I'hesis: It Assun1es that Environmental Detern1inisn1 Is True 226
The Two Objections Summarized 227
MySearchlab Connections 228
22. The Mind/Brain Identity Theory 229 The Identity Theory ls an A Posteriori Claim 229 Materialistn 230 Progress in Science 230 Dualism Resembles Vitalism 231 A Correlation Experitnent 231 The Principle of Parsimony 232
MySearchlab Connections 235
23. Functionalism 236 Functionalism's Negative Thesis: What's Wrong
with the Identity Theory? 237 Multiple Realizability 237 Could a Computer Have Psychological Characteristics? 238 Multiple Realizability within the Class of Living Things 239 Functionalism's Positive Thesis 240 Sensations 241
MySearchlab Connections 243
24. Freedom, Determinism, and Causality 244 The Problem of Freedom 245 Examples of Unfree Acts 246 Are All Behaviors Like Those Produced by
Braitnvashing and Con1pulsions? 247 A Clash of Plausible Conceptions 247 What Is Causality? 248 Detern1inisn1 248 Indeterminisn1 249 Does Indeter1ninism Make Us Free? 250 Causality Is the Issue, Not Detern1inism 251 What Does Determinism Say about the Causation of Behavior? 252 Determinism Differs fro1n Fatalisin 252
MySearchlab Connections 254
25. A Menu of Positions on Free Will 255 "Compatibility" Defined 255 Incompatibilism and Compatibilism 256 Libertarianisn1 257 r::El-{Read on MySearchLab The System of Nature, Baron d'Holbach
Contents xiii
xiv Contents
'''.' > ''.' •• ' •• ' •• ' ••••••••• '.' ••••• ' ••• '.' ••• '.'.'.' ••• ' •• ,- • ' ' •• '. ' •• ''.
Two Soft Determinist Theories 259 Hume 259 t:EJ-[Read on MySearchLab Of Liberty and Necessity, David Hume First Objection to Hume's Theory: Compulsive Behavior 260 Second Objection to Hume's Theory: Locke's Locked Room 261 Does Coercion Rob Us of Free Will? 261 A Second Compatibilist Proposal: The Relevance of
Second-Order Desires 262 t:EJ-[Read on MySearchLab On Free Choice, Thomas Aquinas MySearchLab Connections 263
26. Compatibilism 265 The Weather Vane Analogy 265 Function and Malfunction 266 What Does It Mean to Ascribe a Function to Something? 267 The Function of the Desire-Generating Device 268 Reply to the Distant Causation Argument 269 What Does Responsibility Mean? 269 Moral Responsibility 270 Reply to the Could-Not-Have-Done-Otherwise Argument 271 Are Coerced Actions Unfree? 272 An Objection to the Weather Vane Theory: Freely Chosen,
Rational Self-Sacrifice 273
MySearchLab Connections 274
27. Psychological Egoism 275 T\vo Truisn1s 276 Goals and Side Effects of an Act 277 A Simple Example 277 Four Preference Structures . 278 People Are Rarely Pure Altruists or Pure Egoists 279 An Experimental Test 281 A Second Experimental Test 282 Conclusion 283
MySearchlab Connections 284 t:EJ-[Read on MySearchLab The Republic, Book II, 357A-367E, Plato
Suggestions for Further Reading 285
Readings 286
Meditation VI, Rene Descartes 286 Other Minds Are Known by Analogy from One~ Own Case, Bertrand Russell 286 Has the Self"Free Will"?, C.A. Campbell 289 Determinism Shows that Free Will Is an Illusion, Baron d'Holbach 300
Of Liberty and Necessity, David Hume 301 What Motivates People to Act Justly?, Plato 301
Part Five. Ethics 303
Chapters
28. Ethics-Normative and Meta 305 Ethics and Religion 305 Metaethics and Normative Ethics 306 Truth and Opinion 306 Alternative Metaethical Positions 307 Subjectivism 307 Realism 307 Conventionalisn1 307 Three Varieties of Conventionalism 308
MySearchLab Connections 310
29. The ls/Ought Gap and the Naturalistic Fallacy 311 Subjectivism: Ethical Statements Are Neither True Nor False 311 Does the Existence of Ethical Disagreement Show
that Subjectivism Is True? 312 The Genetic Fallacy 313 Hume: The Is/Ought Gap 314
Contents
(St): An Argument for Subjectivism with Hume's Thesis as a Premise 314 The Naturalistic Fallacy 316 (S2): An Argument for Subjectivism with Moore's Thesis as a Premise 317
MySearchlab Connections 318
30. Observation and Explanation in Ethics 319 Reasoning about Ethical Issues 319 Testing General Principles by Applying Them to Specific Examples 320 Thought Experiments versus Empirical Experiments 321 Observations Are ((Theory Laded' 321
Observation Does Not Imply Objectivity 322 Insoluble Disagreements 322 Is Subjectivism Preferable to Realism on Grounds of Parsimony? 324 Does Subjectivism Follow? 326 An Explanatory Role for Ethical Principles 326 What Is the Point ofEthics? 327 l::EJ-[Read on MySearchLab Treatise of Human Nature, Part III, 1.1,
David Hume
MySearchlab Connections 328
xv
xvi Contents
'''.'.''' ........... ' ....... '.' ... '' ... '.,' '.' ........ '' '' ... '''.''' '.' 31. Conventionalist Theories 329
What Makes a View Conventionalist? 330 Trivial Semantic Conventionalism 330 Substantive versus Trivial Conventionalisn1 330 Plato's Critique of the Divine Command Theory 331 l'.EJ-{Read on MySearchLab Euthyphm, Plato 1\vo Objections to the Divine Command Theory 332 Ethical Relativism 333 Ethical Relativism Is Normative, Not Descriptive 333 A Further Clarification of Ethical Relativism 334 Ethical Relativistn Is a Version of Conventionalisn1 334 If Imperialism Is Wrong, Does That Justify Ethical Relativism? 334 Sartre's Existentialism 336 l'.EJ-{Read on MySearchLab Existentiaiisn1 Is a Hun1anis1n,
Jean-Paul Sartre
MySearchLab Connections 338
32. Utilitarianism 340 l'.EJ-{Read on MySearchLab Utilitarianism, Chapters 1-4, John Stuart Mill Mill's Defense of the Greatest Happiness Principle 341 Reciprocal Illumination 342 What Is Happiness? 343 The Problem of the Experience Machine 343 Mill on "Higher" and "Lower" Pleasures 344 Objection to Hedonistic Utilitarianism 345 Preference Utilitarianism 345 The Apples and Oranges Problem 346 Utilitarianism and Justice: The Case of the Lonesome Stranger 347
Punishment 347 A Reply: Distinguish Rule and Act Utilitarianism 348 Utilitarianism and Tolerance: The Problem of the Fanatical Majority 349 l'.EJ-{Read on MySearchLab On Liberty, Chapters 1-3, John Stuart Mill Utilitarianism and Personal Integrity: The Problem of Dirty Hands 352 Utilitarianism and Personal Loyalties 353 A Psychological Objection to My Criticisms of Utilitarianism 353 l'.EJ-{Read on MySearchlab Principle of Utility, Jeremy Bentham
MySearchLab Connections 355
33. Kant's Moral Theory 357 Hume on Reason's Role 358 Kant Rejects the Idea that Reason Is Purely Instrumental 358 l'.EJ-{Read on MySearchlab Foundations of the Metaphysics of Morals,
Sections I and 2, Immanuel Kant Kant: Moral Rules Are Categorical Imperatives 358
Contents xvii
"''',''','' ''' ''.'',', .. '' .. '''''.''.'.' ..... ,' ''. The Moral Law 359 Kant: The Moral Value of an Act Derives from Its Maxim,
Not from Its Consequences 359 Kant Rejected Consequentialism 360 The Universalizability Criterion 360 Four Examples 361 Evaluation of Kant's Examples 362 A Problem for the Universalizability Criterion 363 Kant: People Are Ends in Themselves 364 The Rabbit and the Hat 365 MySearchLab Connections 366
34. Aristotle on the Good Life 368 How Far Do Obligations Extend? 368 The Theory of the Right and the Theory of the Good 369 Are There General Principles about the Good Life? 369 CE}-{ Read on MySearchLab Nico111achea11 Ethics, Books I, II and X
(Sections 7 and 8), Aristotle What ls a Good X? 370 Human Beings Are Goal-Directed Systems 371 The Capacity to Reason 371 Aristotle: Happiness Is Not a Subjective State 372 Why the Life of Rational Activity Is Best: Two More Reasons 373 The Doctrine of the Mean 373 A Second Criticism of Aristotle's Theory-Defining What
a Good X Is Differs from Saying What Is Good for an X 374 A Third Criticism-Why Single Out Contemplation as the Best Life? 375
MySearchLab Connections 377
Suggestions for Further Reading 378
Readings 379
The Euthyphro-A Critique of the Divine Command Theory, Plato 379 Existe11tialis111, jean-Paul Sartre 379 Defense of Utilitarianism, john Stuart Mill 379 Ethics Founded on Reason, Immanuel Kant 380 Morality and Human Nature, Aristotle 380 Famine, Affluence, and Morality, Peter Singer 380
Boxes
1. Conditionals 18 2. Begging the Question 19 3. Deducing That a Theory is True 32 4. No Surprise/Surprise 34 5. An Investment Swindle 37 6. Hun1ans fron1 Nonhu1nans, Life fron1 Nonlife 76 7. "But How do You Explain God?" 84 8. A Priori/A Posteriori and Necessity/Contingency 95 9. Neptune and Vulcan 105
10. The Expected Value of an Action 110 11. Necessary and Sufficient Conditions 139 12. Is Skepticism Self-Refuting? 144 13. Indubitability 149 14. The KK-Principle 166 15, What's Relative About Einstein's Theory of Relativity? 170 16. Sense and Reference 208 17. Pain Without Pain Behavior? 219 18. Minimal Explanation Versus Deep Explanation 226 19. Ne\vton on Parsin1ony 233 20. The Birthday Fallacy, Again 238 21. Transitivity 245 22. Two Uses of Probability 250 23. The Normative Problem of Freedom 257 24. What is Coercion? 272 25. Apportioning Causal Responsibility 282 26. E1notivisn1 312 27. Values in Science 323 28. Two Views of Democracy 336 29. Liberalism and Conservatism 350 30. Good Versus Green 370 31. "Natural" and "Norn1al" 374
xviii
Preface
THE PHILOSOPHICAL PROBLEMS investigated in this book concern fundamental facts about our place in the universe. Many of us were brought up to believe that God exists, that there is a real difference between right and wrong, that we can freely choose what sort of lives to lead, and that it is possible for us to gain knowledge of the world we inhabit. A major goal of philosophy is to discover whether these opinions can be rationally defended or are just comfortable illusions.
Core Questions in Philosophy emphasizes the idea that philosophy is a subject devoted to evaluating arguments and constructing theories. This is not the same as describing the history of what various philosophers have thought. Although I discuss historical texts, I do so because they are rich sources of ideas pertinent to answering philosophical questions. The point is not to say solemn and respectful words about worthy figures now dead, but to engage them in dialogue-to grapple with the theories they have proposed, to criticize these theories, even to itnprove upon the1n.
Besides proposing answers to philosophical questions, I also try to make clear which questions I have not answered. I hope that the reader will approach what I say the way I have approached the philosophical texts I discuss. This is a book to argue with, to dissect. It isn't my goal to have the reader accept without question the conclusions I reach.
This work is a combination textbook with readings. In Parts 1\vo through Five of the book, the chapters I wrote are followed by related readings (drawn from the writings of Plato, Aristotle, Aquinas, Anselm, Descartes, Locke, Hume, Kant, Mill, and others). The chapters are intended to flow together, so that the main areas covered-philosophy of reli- gion, epistemology, philosophy of mind, and ethics-are connected to each other to make a coherent whole. The chapters often discuss the readings that follow, but the area of over- lap is far from total. Many chapters contain material that isn't touched on in any reading; and the readings raise a wealth of issues that the chapters dmrt address. The chapters are intended to stand on their own as well as to provide points of entry into the readings. The chapters I wrote are intended to be launching pads from which readers can pursue issues on their own. I believe students are best able to think about philosophy if they are first provided with some basic tools and concepts. It is the purpose of the chapters to provide these core ideas.
Each chapter is followed by review questions and problems for further thought. These should help readers to consolidate their understanding of what I have said, and to think creatively about related problems. The chapters often contain material in boxes; these boxes provide, in a nutshell, a restatement of an important idea or a brief discus- sion of a related matter that may interest the reader. A list of the boxes immediately fol- lows the table of contents. Each main part of the book includes suggestions for further reading. There is also a glossary at the end of the book that provides simple definitions of the inain concepts used.
xix
xx Preface
'. ''''' ''.' <'''' " •• ''. ' •• ' ''' •• ' ' •• ,'' '' •••• '.'.'' '''''' '' ' •• ' ' ••• '' '>.' '.
Besides discussing a number of traditional topics, this book also takes up some con- temporary theories and problems, both from philosophy and from other disciplines. Creationism and evolutionary theory are hotly debated now. The issues they raise are con- tinuous with a tradition of argument in philosophy of religion that goes back (at least) to Aquinas, Hume, and Paley. The relation of mind and body is a philosophical problem of long standing, but the ideas of Freud and Skinner get a hearing along with those of Descartes. In ethics, there has long been a debate as to whether ethical truths are discov- ered or created. Plato and Sartre are separated by more than two thousand years, but both speak to this issue. The problem of free will raises the question of whether every event is caused. Here, the contribution of modern physics must be brought into contact with a perennial problem of philosophy. Philosophy isn't the same as biology, psychology, or physics, but the problems of philosophy cannot be isolated from the sciences. One aim of this book is to connect philosophical problems with ideas derived from a wider culture.
In addition to rewriting each chapter for clarity (and in a few instances correcting what I think were mistakes in the previous edition), I added new readings by Plato, Hans Reichenbach, and Peter Singer.
The etymology of the word philosopher is lover of wisdom. This doesn't guarantee that all philosophers are wise, nor even that each individual philosopher is devoted to the attainment of wisdom. Philosophers should strive for wisdom; whether they do so, and 'vhether they attain it, are separate questions. Wisdom involves understanding-seeing how things fit together. When the pieces of a puzzle are fitted together, one attains a sense of wholeness. Current philosophy is embedded in a historical tradition of philosophical discourse. It also is connected to problems in the sciences, the other humanities, and the arts. This book aims to give the reader a sense of these multiple connections.
Acknowledgments
Elliott Sober U11iversity of Wisco11si11-Madison
My debts to colleagues in philosophy at University of Wisconsin, Madison, are enor- mous. A fixed point in my work week has been discussions of the ideas and tech- niques that go into presenting central problems of philosophy to new students. My philosophical outlook, as well as the view I have of teaching, have been shaped by these conversations.
It is a pleasure to thank Michael Byrd, Claudia Card, Fred Dretske, Ellery Eells, Berent Enc;, Malcolm Forster, Martha Gibson, Paula Gottlieb, Daniel Hausman, Andy Levine, Steve Nadler, Terry Penner, Mark Singer, Dennis Stampe, Daniel Wik:ler, and Keith Yandell. They were generous enough to suffer my trespasses onto philosophical ter- rain that belonged more to them than to me. Some read parts of this book and gave me comments; others listened patiently while I tried out what I thought was a new angle.
The first five editions of Core Questions i11 Philosophy elicited a steady stream of cor- respondence and phone calls from teachers of philosophy and their students. These took a variety of forms; there \Vas praise and blan1e, suggestions on ho'iv to do better, and even
Preface xxi
'. '''' '.'.' ''. < ••• '.'.'' •• '''. '.' •• >. '. ' ... '' ' ... ''.'.' .. ' .. ' •... ' '' ... ' •. a few not-so-gentle suggestions that I should turn my attention to other projects. On the whole, though, I was happy with what I heard, though this didn't mean that I felt that I should leave the book unchanged. I thank everyone who took the trouble to let me know what they thought. Most (but not all) will find evidence that I listened to what they said in the way this edition differs from the ones before.
Deserving of special mention are Richard Behling, Keith Butler, John Carpenter, Paul Christopher, William R. Carter, Robert Cummins, Doug Frame, Phil Gasper, Ronald Glass, Richard Hanley, John Hines, Burton Hurdle, Charles Kielkopf, John Koolage, Gregory Mougin, Bradley Monton, Margaret Moore, William Russell Payne, Howard Prospersel, David Ring, Roy Sorensen, Whilhelm S. Wurzer, Stephen Wykstra, and Joel Velasco. The help they provided was extremely valuable. I also must thank the anonymous readers of previous editions of this textbook for their valuable suggestions. In addition, I'm grateful to Hayley Clatterbuck, Stewart Eskew, Casey Hart, Reuben Stern, and Naftali Weinberger for helping me prepare this new edition.
Writing an introduction to philosophy is a challenge. The challenge is to reconstruct what a problem or idea would sound like to someone who hasn't studied the subject be- fore. The project requires that one return to the beginning-to the fundamentals of the subject. I hope what I found by beginning again will be useful to those who are beginning for the first time.
Get Connected with MySearchlab with eText
Provided with this text, MySearchLab with eText provides engaging experiences that per- sonalize, stimulate, and rneasure student learning. Pearson's MyLabs deliver proven results from a trusted partner in helping students succeed.
Features available with this text include:
A Complete eText-Just like the printed text, you can highlight and add notes to the eText online or download it to your iPad.
• Assessment-Topic-specific assess1nent, flashcards, and chapter exan1s offer and report directly to your grade book.
• Additional Support Readings-Key readings are available exclusively online, includ- ing short answer questions tying readings to text content.
• Writing and Research Assistance-A 'vide range of '\vriting, gra1nmar and research tools and access to a variety of academic journals, census data, Associated Press newsfeeds, and discipline-specific readings help you hone your writing and research skills. MySearchLab with eText can be packaged with this text at no additional cost- just order the ISBN on the back cover. Instructors can also request access to preview MySearchLab by contacting your local Pearson sales representative or visiting '\V'\V\v,1nysearchlab.co1n.
xxll
MySearchlab Connections in This Text
In addition to the outstanding research and writing tools and a complete eText in MySearchLab, this site contains a wealth of resources for philosophy students such as videos, audio files, study tools, and n1ore.
Below is a list of readings found exclusively in MySearchLab connected to chapters in this text.
Readings Available through MySearchlab
Aristotle, Nicomachean Ethics, Books I, II, and X (Sections 7 and 8) (34) G11ani/o and Ansel111, "Debate" (8) Thomas Aq11i11as, "Five Ways to Prove that God Exists", from S11111ma Theologiae (4) Thomas Aq11inas, "On Free Choice" (25) Jeremy Bentham, "Principle of Utility" (32) W. K. Clifford, "The Ethics ofBelief" (10) Rene Descartes, ((Correspondence \vith Princess Elizabeth'' (19) Rene Descartes, Meditation VI from Meditations on First Philosophy (19) Rene Descartes, Meditations 1-5 of Meditations on First Philosophy (13) Baron d'Holbach, "The System of Nature" (25) David Hume, Dialogues Concerning Natural Religion, Part 2 (5) David H11111e, "Of Liberty and Necessity" (25) David Hume, "Of Miracles': from An Enquiry Concerning Human Understanding (9) David Hume, Section IV of An Enquiry Concerning Human Understanding (15) David Hu111e, Treatise of Human Nature; Part Ill, 1.1. (30) William James, "The Will to Believe" (10) I111111a11ue/ Kant, Foundations of the Metaphysics of Morals, Sections 1 and 2 (33) Gottfried Leibniz, Theodicy (n) John Locke, Chapter 11 of Essay Concerning Human Understanding (18) William Paley, "The Design Argument", from Natural Theology (5) Blaise Pascal, "The Wager': from Pensees (10) Plato, "The Republic" Book II, 357A-367E (27) Plato, "Euthyphro'' (31) Bertrand Russell, "The Value of Philosophy" (1) Jean-Pa11l Sartre, "Existentialism ls a Humanism" (31) John Stuart Mill, "Utilitarianism;' Chapters 1-4 (32) John St11art Mill, "On Liberty;' Chapters 1-3 (32)
part one •••••••••••••••• ·-· ·-5·_.-~ •• ~. ~-. ~-~
Introduction
chapter l ..............................
What Is Philosophy?
Chapter Outline
Examples
Three Theories about What Philosophy Is
The Nature of Philosophy Has Changed Historically
Philosophical Method
Summary
When asked "do you have a philosophy?" most people say "yes;' but what do they mean? They usually have in mind a set of beliefs that they admit are difficult to prove are true, but that nonetheless are important to the way they think of themselves and the world they inhabit. Sometimes people describe their philosophies by saying what they think makes an action right or wrong. The statement "it's part of my philosophy that people should help each other" might be an example. A person's philosophy might include the fundamental ethical principles he or she believes. But people often have more than ethics in mind when they talk about their philosophies. A religious person might say that it is part of his or her philosophy that God exists; an atheist might say that it is part of his or her philosophy that there is no God and that there is no life after death. These propositions are important to the people who believe them. They describe what exists; philosophers would say that they are part of metaphysics, not ethics. Metaphysics is the part of philosophy that attempts to describe, in very general tenns, \Vhat there is.
If every day people think of their philosophies as the important beliefs they have that are difficult to prove, how does this idea of philosophy relate to how philosophers under- stand their own subject? Sometimes a term is used in ordinary talk in a way that differs dramatically from the way it is used by specialists. People sometimes say that tomatoes are vegetables, but a botanist will tell you that tomatoes are fruits. Every day people say they
3
4 Part One: Introduction
''' '''' , ' ',' '''''' ''''' '' '''' , & ',''
are concerned about "ecology;' but biologists understand "ecology" in a very different way. Perhaps philosophers use the term "philosophy" in a way that departs fundamentally from what ordinary people mean when they say that they have a philosophy.
To gain a better purchase on what philosophy is, I'm going to discuss the question of what is distinctive about philosophy from two angles. First, I'll sketch some of the main philosophical problems that I'll examine in this book. That is, I'll describe some examples of philosophy. But giving examples doesn't really answer the question of what philosophy is. If you asked, "What is a 1nanunal?" and I sho'ived you a human being, a hippo, and a cat, these exan1ples n1ight give you a hint about 'ivhat a 1nanunal is. Ho'ivever, citing exa1nples isn't the sa1ne as saying 'ivhat it is to be a 111anunal. That is 'ivhy there 'ivill be a second stage to my discussion of what philosophy is. After giving some examples of philosophical problems, I'll present some theories about what philosophy is. I believe these theories have n1erit, though I adn1it none is entirely adequate.
Examples
The first philosophical problem we'll consider in this book is whether God exists. Some philosophers have constructed arguments that attempt to establish that God exists, others have tried to show there is no God, and still others think that the question can't really be answered. I'll evaluate some of the more influential arguments and try to see whether they work.
The second problem we will consider concerns knowledge. It is pretty clear that belief and knowledge are different. Long ago some people thought that the earth is flat. They believed this, but they didn't know it, since it isn't true. Of course, they thought they knew it, but that's different. It is also pretty clear that true belief isn't the same as knowledge. If you believe something for no reason at all, but happen to be right by accident, you have true belief but not knowledge. For example, think of a gullible gambler at a racetrack who believes for no good reason that the first horse in every race will win. Occasionally this person will be right-he will have a true belief. But it isn't plausible to say that he knew, on those races about which he turned out to be right, which horse would win. So having knowledge involves something more than having a true belief.
The philosophical problem about knowledge will split into two parts. First, there are the questions: What is knowledge? What makes knowledge different from true belief? Second, there is the question: Do human beings ever know anything? One philosophical position we 'ivill consider ans'ivers this last question in the negative. Sure, vve have beliefs. And granted, son1e of our beliefs turn out to be true. Kno\vledge, ho,vever, \Ve never have. We don't even know those things that we take to be most obvious. This position is called philosophical skep- ticisn1. We \Vill consider argu1nents for skepticisn1 and argun1ents that attempt to refute it.
The third philosophical subject that will be addressed in this book consists of a collection of topics from the philosophy of mind. The first of these is the so-called mind/ body problem. You have a mind; you also have a brain. What is the relationship between these items? One possible answer is that they are identical. Although "mind" and "brain'' are different words, they name the same thing, just like the names "Superman" and "Clark Kent:' An alternative position in this area is called dualism; it says that the mind and the brain are different things. We will consider other theories that have been advanced about the mind/body problem as well.
Chapter 1: \Y/hat Is Philosophy? 5
''' '''',' ''''''''' '.''' ''.' '' .. '.''' ... ''. '. Another topic from the philosophy of mind that we'll address concerns human freedom.
Each of us has the personalities 've have because 've inherited a set of genes fron1 our parents and then gre'v up in a sequence of environ1nents. Genes plus environn1ents n1ake us the sorts of people 've are. We didn't choose the genes 've have, nor did '"e choose the environn1ents we experienced in early life. These were thrust upon us from the outside. Each of us per- fonns certain actions and abstains from perfor111ing others. This pattern of ,vhat 've do and don't do results fron1 the personalities 've have. Can 've be said to perfor111 actions freely? Is it really in our control to perform some actions and abstain from others? Perhaps the fact that our actions are the results of factors outside our control (our genes and our early environ- ment) shows that it is a mistake to say that we freely choose what we do. Of course, we talk in everyday life about people doing things "of their own free will:' We also think of ourselves as facing real choices, as exercising control over 'vhat "Te do. Ho,vever, the philosophical prob- lem of freedom asks whether this common way of thinking is really defensible. Maybe free- don1 is just an illusion. Perhaps 've tell ourselves a fairy tale about our o'vn freedom because we can't face the fact that we are1it free. The philosophical problem will be to see whether we can be free agents if our personalities are the results of factors outside our control.
The last problem area we will address is ethics. In everyday life, we frequently think that some actions are right and others are wrong. The philosophical problem about this familiar attitude divides into two parts. First, we'll consider whether there really are such things as ethical facts. Maybe talk about ethics, like talk about freedom, is just an elabo- rate illusion. Consider a parallel question about science. In every science, there are ques- tions that are controversial. For example, physicists have different opinions about how the solar system began. But most of us think that there is something else to physics besides opinions. There are facts about what the world is really like.
Clashes of opinion occur in what I'll call the subjective realm. Here we find one human mind disagreeing with another. But facts about physics exist in the objective realm. Those facts exist independently of anybody's thinking about them. They are out there, and science aims to discover 'vhat they are. In science, there are both subjective opinions and objective facts-people have beliefs, but there also exists, independently of what anyone believes, a set of facts concerning the way the world really is. The question about ethics is whether both these realms (subjective and objective) exist in ethics, or only one of them does. We know that people have different ethical opinions. The question is whether, in addition to those opinions, there are ethical facts. In other words, does ethics parallel the description I've just given of science, or is there a fundamental difference here? The accompanying two-by-two table illustrates this question. Ethical subjectivism is the philosophical thesis that there are no ethical facts, only ethical opinions. According to this position, the claim that ((n1urder is ahvays 'vrong" and the clahn that "n1urder is sometimes permissible'' are both misguided-there are no facts about the ethics of murder for us to have opinions about. We'll consider argu111ents supporting and criticizing this position.
Subjective Realm Objective Realm
Science Scientific opinions Scientific facts
Ethics Ethical opinions Ethical facts?
6 Part One: Introduction
....... '' .... '' .... '. ' .. ' ......... ' .. '''.' ''' '' ' .... ''. '.'. '.' '' .. '' ... ' ' ' The second question that arises in ethics is this: If there are ethical facts, what are they?
Here \Ve assume a positive ans,ver to the first question and then press for 1nore details. One theory we'll consider is utilitarianism, which says that the action you should perform in a given situation is the one that will produce the greatest happiness for the greatest number of individuals. This may sound like common sense, but I'll argue that there are some serious problems with this ethical theory.
Three Theories about What Philosophy Is
I've just described a menu of four central philosophical problems: God, knowledge, mind, and ethics. What makes them all philosophical problems? Instead of giving examples, can we say something more general and complete about what distinguishes philosophy from other areas of inquiry? I'll offer three theories about what is characteristic of at least some philosophical problems.
Several of the problems just described involve fundamental questions of justification. There are many things that we believe without hesitation or reflection. These beliefs that are second nature to us are sometimes called "conunon sense;' Common sense says that the sense experiences we have (via sight, hearing, touch, taste, and smell) provide each of us with knowledge of the world we inhabit. Common sense also says that people often act ((of their ovvn free \vill:' and conunon sense holds that some actions are right \vhile others are wrong. Philosophy examines the fundamental assumptions we make about ourselves and the world we inhabit and tries to determine whether those assumptions are rationally defensible.
Another characteristic of many philosophical questions is that they are very general; often they're more general than the questions investigated in specific sciences. Physicists have asked whether there are electrons; biologists have investigated whether genes exist; geologists have sought to find out whether the continents rest on movable plates. However, none of these sciences really bother with the question of why we should think that physical objects exist. The various sciences sin1ply assunie that there are things outside the 1nind; they then focus on more specific questions about what those things are like. In contrast, it is a characteristically philosophical question to ask why you should believe that there is anything at all outside your mind. The idea that your mind is the only thing that ex- ists is called solipsism. Philosophers have addressed the question of whether solipsism is true. This is a far more general question than the question of whether electrons, genes, or continental plates exist.
The third view of what philosophy is says that philosophy is the enterprise of clarifying concepts. Consider some characteristic philosophical questions: What is knowledge? What is freedom? What is justice? Each of these concepts applies to some things but not to others. What do the things falling under a concept have in common, and how do they differ from the things to which the concept does not apply?
We must be careful here, since many questions that aren't especially philosophical sound like the exan1ples just given. Consider so1ne characteristic scientific questions: What is photosynthesis? What is acidity? What is an electron? How does the first batch
Chapter 1: What Is Philosophy? 7
''''.''.' ''' ''' '.'' '. ''' ''.'' • '' •• ''' •• '.' <.' ''. ''''''.'''' ''''' ''''."''
of questions differ from these? One difference between these questions concerns the 'vays in ,vhich reason and observation help ans,ver then1. You probably are a'vare that philosophy courses don't include laboratory sections. Philosophers usually don't per- form experiments as part of their inquiries. Yet, in many sciences (though not in all), laboratory observation is central. This doesn't mean that observation plays no role in philosophy. Many of the philosophical arguments we will consider begin by making an observation. For example, in Chapter 5, I'll consider an argurnent for the existence of God that begins with the following assertion: Organisms are complicated things that are remarkably well adapted to the environments they inhabit. The thing to notice here is that this fact is something we know by observation. So philosophers, as well as scientists, do rely on observations.
Nonetheless, there is son1ething distinctive about ho'v observations figure in a philosophical inquiry. Usually the observations that are used in a philosophical the- ory are familiar and obvious to everyone. A philosopher will try to show by reasoning that those observations lead to some rather surprising conclusions. That is, although philosophy involves both observation and reasoning, it is the latter that in some sense does more of the work. As you will see in what follows, philosophical disputes often involve disagreements about reasoning; rarely are such disputes decidable by making an observation.
Each of these ways of understanding what philosophy is should be taken with a grain of salt (or perhaps with two). I think there is something to be said for each, even though each is somewhat simplified and distorting.
The Nature of Philosophy Has Changed Historically
One thing that makes it difficult to define "what philosophy is" is that the subject has been around at least since the ancient Greeks and has changed a great deal. There are many problems that are just as central to philosophy now as they were to the ancient Greeks, but there are other problems that have broken away from philosophy and now are thought of as purely scientific.
For example, ancient Greek philosophers discussed what the basic constituents of physical things are. Thales (who lived around 580 n.c.E.) thought that everything is made of 'vater; many other theories 'vere discussed as 'vell. No1v such questions are thought to be part of physics, not philosophy. Similarly, until the end of the nineteenth century, universities put philosophy and psychology together in the same academic de- partment. It is only recently that the two subjects have been thought of as separate. Scientists in the seventeenth century-for exan1ple, Isaac Ne,vton-used the ter1n "natural philosophy" to refer to 'vhat \Ve no'v think of as science. The ter1n "scientist" was invented in the nineteenth century by the British philosopher William Whewell. The idea that philosophy and science are separate subjects may seem clear to us now, but the separation we now find natural was not so obvious in the past. Many of the problems that we now regard as philosophical are problems that have not broken away from philosophy and found their way into the sciences. Perhaps there are problems now
8 Part One: Introduction
. ' .. ' ........ '' '' .. '.'',' .. '.''.' ''.''' . '.'' ' ..... '. ' ..... ' ' ..... '.'.'' taken to be philosophical that future generations won't regard as such. The shifting his- torical nature of what counts as philosophy makes it difficult to say anything very pre- cise about what that subject is.
Philosophical Method
Having tried to say something about what philosophy is, I now want to say something about what philosophy is not (at least not in this book). You may have the impression that doing philosophy involves lying under a tree, staring up at the sky, and making deep and mysterious pronouncements off the top of your head that sound very important but that are hard to make sense of when you try to think about them clearly. I'll call this the mysti- cal guru model of philosophy. Your experience reading this book won't correspond to this itnpression.
There is, however, another experience you've probably had that comes closer. If you took a high school geoinetry course, you'll re1ne1nber proving theore1ns fro1n axio1ns. If your geometry course was like the one I had, the axioms were given to you with very little explanation of why you should believe them. Maybe they looked pretty obvious to you, and so you didn't wonder very much about their plausibility. Anyhow, the main task was to use the axioms to prove theorems. You started with the axio1ns as assun1ptions and then showed that if they are true, other statements must be true as well.
Philosophers tend to talk about "arguments" rather than "proofs:' The goal is to try to reach answers to important philosophical questions by reasoning correctly from assumptions that are plausible. For example, in Chapter 4, I'll examine some attempts to prove that God exists. The idea here is to start with assumptions that practically anybody would grant are true and then show that these assumptions lead to the conclusion that there is a God. This resembles what you may have done in geometry: Starting with simple and supposedly obvious assumptions, you were able to establish something less obvious and more complex-for example, that the sum of the angles of a triangle equals two right angles (180°).
Sometimes the philosophical questions we'll consider will strike you as diffi- cult, deep, even n1ysterious. I \Von't shy a\vay from such qt1estions. I'll try, ho\vever, to address them with clarity and precision. The goal is to take hard questions and deal with them clearly, which, I emphasize, should never involve trying to pull the \Vool over someone's eyes by making deep-sounding pronouncements that mean who-knows-what.
Summary
I began this chapter by describing how every day people use the term "philosophy:' In fact, their usage is not so distant from what philosophers mean by the term. Philosophy does address the most fundamental beliefs we have about ourselves and the world we in- habit. Precisely because these assu1nptions are so central to the 'vay "\Ve think and act, it is difficult to step back for a moment from these assumptions and examine them critically.
Chapter 1: \Y/hat Is Philosophy?
The French have an expression: "the most difficult thing for a fish to see is water:' Some assu1nptions are so natural and see1ningly obvious that it is hard to see that \ve are mak- ing assumptions at all. Philosophy is the effort to help us identify these assumptions and evaluate them. Each of us does have a philosophy. What divides some people from others is their willingness to ask probing questions about what they believe and why. This is what philosophy as a discipline tries to add to the philosophies that each of us carries with us through our lives.
Review Questions
i. What is the difference between objective and subjective?
2. If you \Vant to say \Vhat philosophy is, why isn't it enough to list some examples of philosophical problems?
A Problem for Further Thought
Which of the ideas presented here about what philosophy is also apply to mathematics? Which do not?
Supplementary Reading
BERTRAND RUSSELL
The Value of Philosophy
Bertrand Russell (1872-1970), a British philosopher and logician, was one of the greatest and most infiuential philosophers of the twentieth century. In this short passage, Russell defends the practice of philosophy against the criticism that it is frivolous or has resulted in few definite answers, arguing that philosophy's value lies in the questions it asks rather than in the answers it yields.
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
CEf-{ READ AND REVIEW
READ
l. Bertrand Russell, "The Value of Philosophy" According to Russell, what does the study of philosophy contribute to a person's life? How does philosophy, and not science or other pursuits, provide that benefit?
CEf-{ Read "The Value of Philosophy" on www.mysearchlab .com
9
10 Part One: Introduction
........................ ''' ' ... ''.' ... '' ''.'' WATCH AND LISTEN
©{WATCH {<•>{LISTEN l. Listen "Edward Craig on What Is Philoso-
phy?"-Philosophy Bites podcast Edward Craig, editor of the Routledge Encyclopedia of Philosophy, argues that philosophy has been too narrowly defined in the past and that there is no sharp divi· sion between "philosophical" thought and "nonphilosophicol" thought. Why does Craig think that scientific and religious statements are also philosophical? Does this make the definition of "philosophy" too vogue?
2. Listen "John Armstrong on What Can You Do With Philosophy?"-Philosophy Bites podcast , john Armstrong claims that philosophy
provides its students with a wide variety of intellectual resources. Jn what kinds of fields does Armstrong think that these resources are useful? Do you see any instances of philosophical reasoning in your daily life?
3, Listen "Joking, and Learning, About Philos- ophy"-Weekend Edition interview
In this clip, Thomas Cathcart and Daniel Klein, authors of Plato and a Platypus Wolk into a Bar, tell some of their favorite philo· sophicaljokes. According to the authors, what do jokes have in common with philosophy?
•••• '''' '''' ''' 0''' ''.'''.'
RESEARCH AND EXPLORE
'1,T(EXPLORE l. Research philosophical methodology 2. Research nottirol philosophy
chapter 2 ..............................
Deductive Arguments
Chapter Outline
Argun1ents
Good Argun1ents
Deductive Validity Defined
''Validity'' Is a Technical Tern1
Logical Form
Invalidity
Testing for Invalidity
Circularity, or Begging the Question
Truth
"True for Me,,
Wishful Thinking
Self-Fulfilling Prophesies
Philosophy involves constructing and evaluating arguments. In this respect, philosophy is no different frotn any other rational activity-1nathematicians do this, as do economists, physicists, and people in everyday life. The distinctive thing about philosophy isn't that philosophers construct and evaluate arguments; what is distinctive is the kinds of ques- tions those arguments aim to answer. In the previous chapte1; I talked about what makes a question philosophical. The goal in this chapter is to develop some techniques that can be used to tell whether an argument is good or bad.
Arguments
An argument divides into two parts: the premises and the conclusion. The premises and the conclusion are state1nents; each is expressed by a declarative sentence. Each is either true or false. When people argue that a given statement is true, they try to provide reasons for thinking this. The reasons are the premises of their argun1ent; premises are assump- tions. The statement to be established is the argument's conclusion.
In high school geometry, you talked about axioms and theorems. Axioms are assump- tions (premises); the theorem (the conclusion) is what is supposed to follow from those
11
12 Part One: Introduction
' •••••••••••••••••••• t •• '' •••• ' •• ' •••••••• ' •••••••• ' ••• ' ••• '. ''. ''' '''.','
assumptions. In geometry you may have spent little or no time asking whether the axioms are true. Not so for the philosophical arguments I discuss in this book. We'll want to see whether the premises are plausible. We'll also want to see whether the premises, if they were true, would provide a reason for thinking the conclusion is true as well. I'll pose these t\vo questions again and again.
Good Arguments
I now want to talk about different kinds of "good arguments:' What does "good" mean? A good argument is rationally persuasive; it gives you a substantial reason to think the conclusion is true. Advertisers and politicians sometitnes use argu1nents that trick people into believing what they say. These arguments sometimes persuade people, but they don't always provide good reasons.
A good argument should have true premises; if the premises are false, how could they give you good reasons to believe the conclusion? But more is required than this. In the following argument, the premise is true, but it doesn't provide you a good reason to think that the conclusion is true:
Grass is green
Roses are red
What is wrong here is that the premises are irrelevant to the conclusion. A good argument should contain true premises, but it should also cite premises that are related in the right way to the conclusion. The truth of the premises should give you a reason to think that the conclusion is true. The three types of"good argument" that I'll now describe differ in what relationship their premises and conclusions have to each other.
Good arguments can be divided into two categories, and one of those categories can be divided into two more:
Good Arguments --- ------Not Deductively Valid
Deductively Valid
~ ----Abductively Inductively Strong Strong I'll treat the three categories (deductively valid, inductively strong, and abductively strong) as mutually exclusive. If an argument belongs to one category, it can't belong to any of the others. At the end of Chapter 3, I'll modify this classification slightly.
You may have heard some of this terminology before. Deduction is what you do in a mathematical proof. Induction involves sampling from a population to decide what its characteristics are. "Abduction'' may be a less familiar term. It has nothing to do with
Chapter 2: Deductive Arguments 13
.. '''.''.''.' .. ' .... ' ..... ''. '.'' ... '.' .. '.' ' ..... ' ... " ... ' .. ' .. ' '.'. '.'' kidnapping. The word was invented by the great nineteenth-century American philoso- pher Charles Sanders Peirce. Philosophers sometimes use the longer label "inference to the best explanation" to describe what Peirce meant by abduction.
I'll consider deduction in this chapter, induction and abduction in the next. The goal in each case is to describe so1ne of the considerations that are relevant to deciding \Vhether an argument is good or bad.
Deductive Validity Defined
The first type of good argument consists of ones that are "deductively valid:' Here are two examples of this type of argument:
All fish swim.
All sharks are fish.
All sharks swim.
All particles have mass.
All electrons are particles.
All electrons have mass.
In these argu1nents, the pren1ises are the statements above the horizontal line; the con- clusion is the statement below. These arguments say that the premises are true and that, therefore, the conclusion also is true.
Here is what deductive validity means:
A deductively valid argument is an argument that has the following property: IF its premises \Vere true, its conclusion \VOtlld have to be true.
I've capitalized the word IF. I'd print it in bright colors if I could because it is important not to forget this two-letter word. A valid argument need not have true premises. What is required is that the conclusion would have to be true IF the premises were true. Take a minute to look at these two arguments. Convince yourself that they are deductively valid.
"Validity" Is a Technical Term
What philosophers and logicians mean by "valid" doesn't have much in common with what we mean by "valid" in ordinary English. In everyday life, we say that a statement is "valid" if it is plausible or true. The technical use of the term that I just explained differs fro1n ordinary usage in t\vo \Vays. First, \Ve never say that a statetnent or an idea is valid or invalid. Validity is a property of arguments and of arguments only. Second, an argument can be valid even if the statements it contains are wildly implausible. A valid argument can have false pre1nises and a false conclusion.
Here is an example:
All plants have minds.
All ladders are plants.
All ladders have minds.
14 Part One: Introduction
' .. ''' '' '''.''.'. ' ... ''''' .. '. '' ... '' .. '' .... ''.'.,.' ''.' ... ' ' .. '.' '.' ' .. Logical Form
What makes an argument deductively valid? The three example arguments described so far have different subject matters. The first is about fish, the second is about particles, and the third is about plants. Although they are about different things, they have the same structure. The structural property that they have in common is called their "logical form:' Think of each argument as the result of substituting terms into the following skeleton:
All Bs are Cs
All As are Bs
All As are Cs
This is the logical form of the three arguments given. You can think of A, B, and C as blanks into which terms may be substituted. Take a minute to see how the arguments just stated can be obtained from the above skeleton by substitution-by "filling in the blanks:'
An argument is valid or invalid solely because of the logical form it has. The subject matter of the argument is irrelevant. Since the three example arguments have the same logical form, they are all valid or all invalid. They have the same logical form, so they are in the same boat. As already mentioned, they are valid. Indeed, each and every one of the millions of arguments you can construct by substitution into the above skeleton is valid as well.
Invalidity
The definition of validity tells you what a deductively invalid argument will be like. If there is even the smallest possibility that the conclusion could be false when the premises are trne, then the argument is deductively invalid.
The ladder argument is valid, although all the statements it contains are false. Is the reverse situation possible? Can an argument be invalid, even though all the statements it contains are true? The ans\ver is yes. Here's an example:
Emeralds are green.
Lemons are yellow.
The premise is true, and so is the conclusion. So why isn't the argument deductively valid? The definition of validity says that the premises in a valid argument must provide an absolute guarantee that the conclusion is true. But the fact that emeralds are green doestit guarantee that lemons must be yellow. The color of lemons isn't entailed by the fact that emeralds are green. Validity concerns the relationship of premises to conclusion, not the question of whether the premises and the conclusion each happen to be true.
Sometimes it isn't so obvious that an argument is invalid. The above example is pretty blatant-the premise has nothing to do with the conclusion. But what do you think of the following argument? Is it valid or not?
Chapter 2: Deductive Arguments 15
...... '' ....... ' ' ... , .. '' .. '' .. ' ......... ' .. '' ........ '.'.'''' '' '';' .. ', '' If Jones stands in the heavy rain without an umbrella, then Jones will get wet.
Jones is i;vet.
Jones was standing in the heavy rain without an umbrella.
Imagine that all three of the statements in this argument are true. Imagine that Jones is now standing before you soaking wet and that Jones just came in from the rain.
Even if all the statements in this argument are true, this argument is still invalid. It is just like the argument about emeralds and lemons. Although the premises and the conclusion happen to be true, the premises don't guarantee that the conclusion must be true.
How can we see this more clearly? I said before that all arguments that have the same logical form are in the same boat. This means that if the argument about Jones is invalid, so is each and every argument that has the same form. Let's begin by isolating the argument's logical form. Here it is:
If P, then Q
Q
p
What do P and Q stand for in this argument skeleton? You can substitute any statement (declarative sentence) you please for these letters to obtain an argument with this logical form. Notice that the letters in this skeleton differ in their function from the letters in the previous skeleton. There, A, B, and Care blanks into which terms denoting kinds of things ("fish:' "electrons;' etc.) can be substituted. Anyhow, we now have the logical form of the argument about Jones. If it is invalid, so are all arguments that have the same logical form. This means that if there is even one argument that has this logical form in which the prem- ises are true and the conclusion false, then the argument form is invalid. This will mean that the initial argument about Jones is invalid as well.
Here is an argument that has the same logical form as the argument about Jones that settles the question:
If Sam lives in Wisconsin, then Sam lives in the United States.
Sam lives in the United States.
Sam lives in Wisconsin.
The premises of this argument are true, but the conclusion, I assure you, is false. The Sain I'm talking about lives in Georgia.
Testing for Invalidity
So here's a strategy to use if you want to know whether an argument is invalid: First, ignore the argument's subject matter and isolate the logical form (the "skeleton") of the argument. Second, see if you can invent an argument that has this logical form in which the premises are true and
16 Part One: Introduction
. '.''.' ...... ' .. '.'' '. ''''''. '.'. '' ''. '' ''' '.'' ''.' '''. ,, '.'.'.'' ''.'.' '' '. the conclusion is false. If you can find even one rotten apple of this type, you are finished. If there is even one argument with this property, then every argument of that form is invalid.
When an argument has true pren1ises and a false conclusion, it is quite obvious that the truth of the premises doesn't guarantee that the conclusion must be true. The premises can't be guaranteeing this, as the conclusion is false. But this tells you something general. It tells you that each and every argument of this form will be such that the premises don't guarantee the truth of the conclusion.
So far, I've presented some examples of arguments. I've explained that a valid argument needn't have true statements in it and that an argument composed solely of true state- ments needn't be valid. This should make you wonder whether there is any connection at all between the question of whether an argument is valid and the question of whether the premises and conclusion are true.
There is a connection. It is illustrated by the following table. If an argument is valid, it can exhibit three of the four following combinations in which the premises are either all true or not all true and the conclusion is either true or false:
Pnnnises
All true Not all true
True Possible Possible
Conclusion
False Impossible Possible
This table indicates that a valid argument can't have true premises and a false conclusion. However, the fact that an argument is valid leaves open which of the other three cells the argun1ent occupies.
What can be said of an invalid argument? If an argument is invalid, are any of the four combinations impossible? I leave this to you to figure out by consulting the definition of validity.
When you find an invalid argument, you may want to ask if the argument can be repaired. Is there anything that can be done to an invalid argument to turn it into an ar- gument that is valid? There is. By adding premises, you can turn a deductively invalid argument into a valid one. Consider the following argument:
Smith lives in the United States.
S1nith lives in Wisconsin.
This is invalid, but it can be made valid by adding a premise:
Smith lives in the United States.
Everyone \vho lives in the United States lives in Wisconsin.
Sn1ith lives in Wisconsin.
Chapter 2: Deductive Argurnents 17
• '.'' '''' •• ' •• ' •• '.'' ''' ''' ' •• '' '.'''''' '', ''' •• ''' '''' '''.''' ''.''.' 0' •• '
Notice that the conclusion now follows from the premises. The trouble is that the second prernise is false.
In the preceding pair of arguments, fixing the defect of invalidity just substitutes one problem for another; instead of having to criticize an argument for being invalid, you now have to criticize an argument for having a premise that isn't true. The following argument pair is different. Here you can repair the defect of invalidity and obtain a perfectly fine argument. Notice first that the following argument is1it deductively valid:
Smith lives in Wisconsin.
Smith lives in the United States.
The argument can be repaired, however, by adding a premise:
S1nith lives in Wisconsin.
Everyone \vho lives in Wisconsin lives in the United States.
Smith lives in the United States.
This argu1nent is valid and has true premises as \Yell. You can see fro1n these hvo pairs of arguments that invalidity is easy to fix. Just add premises. What is harder is to add premises that not only make the argument valid, but that are true as well.
This idea will come up repeatedly when I discuss various philosophical arguments. I will sometimes claim an argument is invalid. When this happens, you should ask your- self whether the argument can be repaired. Often the price of making the argument valid (by adding a premise) is that you have to supply a new premise that you think is false. In making this addition, you are trading one defect (invalidity) for another (false premises). If you carit repair the argument so that it is both valid and has all true premises, then you should consider the possibility that there is something fundamentally flawed about the whole line of argument. On the other hand, sometimes an invalid argument can be replaced by a valid one merely by supplying a true premise that maybe you neglected to 1nention because it is so obvious. In this case, the defect in the original argument isn't fundamental.
So far I've emphasized two questions that we will want to ask about arguments:
I. ls the argument deductively valid? 2. Are all the premises true?
If the answer to both questions is yes, the conclusion of the argument must be true. Arguments are tools. We use them to do things. When the goal is rational persuasion,
a good argument will provide a good reason to think that the conclusion is true. If an argu- ment is deductively valid and has true premises, is that sufficient to make the argument good? To see why validity and true premises arerit enough, consider the following argument:
Lemons are yellow.
Lemons are yello\v.
Here the conclusion 1nerely repeats \Vhat the premise asserts. This argtunent is valid and the premise is true. But there is something defective about this argument. What is it?
18 Part One: Introduction
'''' .... '' .. '' ..... ' ... ' .. ' ... ' ..... '.'.'' ..... ' ... ' .... ' ''' .. '. '.,.' .. ' '. CONDITIONALS
If/then statements are called conditionals. Conditional statements have other state- ments as components. For example, the statement "If pigs fty, then grass is green" is a statement of the form "If P, then Q," where P and Q are themselves statements.
In the statement "If P, then Q," Pis called the antecedent and Q is called the consequent. A conditional doesn't say that its antecedent is true; the statement "If Joe drinks arsenic, then Joe will die" does not say that Joe drinks arsenic. And a condi- tional doesn't say that its consequent is true; "If there is a nuclear war, then Washing- ton will be attacked" doesn't say that Washington will be attacked.
Conditionals can be rewritten without changing what they say. Consider the state- ment "If you live in Wisconsin, then you live in the United States." This is equivalent in meaning to "If you don't live in the United States, then you don't live in Wisconsin." The conditional "If P, then Q" is equivalent to "If not-Q, then not-P," no matter what P and Q happen to be. Here's a piece of terminology: The statement "If not-Q, then not-P" is the contrapositive of the conditional "If P, then Q." A conditional and its con- trapositive are equivalent.
Consider the following two conditionals: "If P, then Q" and "If Q, then P." Are they equivalent? That is, do they mean the same thing? The answer is no. "If you live in Wisconsin, then you live in the United States" is true, but "If you live in the United States, then you live in Wisconsin" is false. These two if/then statements can't mean the same thing, because one is true while the other is false. "If Q, then P" is termed the converse of the conditional "If P, then Q." A conditional and its converse are not equivalent.
Circularity, or Begging the Question
The previous argument is circular; it begs the question. Suppose you didn't already have an opinion as to whether lemons are yellow. The above argument wouldn't help you resolve your uncertainty. The argument would be useless in this regard.
Good arguments are tools that help answer questions about whether their conclusions are true. A good argument should give you a reason to accept the conclusion if you don't already believe the conclusion is true. So besides checking to see if an argument is deduc- tively valid and has true premises, you should also see if the argument begs the question.
You'll notice in what I've just said that I am using the expression "begging the question'' to name a defect in an argument. Unfortunately, the phrase is often used now to simply mean that a question is being asked. You'll hear this usage on the evening news- "the re- cession has not improved. This begs the question of 'vhether governn1ent action can tnake things better:' As with the term "validity;' the term "begging the questimi' is used by phi- losophers with a special, technical meaning, one that doesn't coincide with ordinary usage.
Truth
One other idea needs clearing up before I leave the topic of deductive validity. You'll notice that the definition of validity makes use of the concept of truth. What is truth?
Chapter 2: Deductive Arguments
There are deep philosophical questions here, most of which I'll skirt. My goal is to describe the concept of truth I use in this book. It is beyond the scope of this book to de- fend this choice or to fully develop its implications. To begin with, whether a statement is true is an entirely different question from whether you or anybody happens to believe it. Whether someone believes the statement "The Rocky Mountains are in North America'' is a psychological question. If beings with minds had never populated the earth, no one would have thought about the location of this mountain range. But this doesn't affect the question of whether the statement is true. There can be truths that no one believes. Sym- metrically, there can be propositions that everyone thinks are true, but that aren't. There can be beliefs that aren)t true.
BEGGING THE QUESTION
To understand what makes an argument question-begging, it is useful to examine some examples.
Suppose you were trying to convince someone that God exists. The argument you give for thinking that this is true is that the Bible says that there is a God. Would this argument convince someone who didn't already believe that there is a God? Probably not. Anyone who doubts that there is a God probably doesn't think that everything the Bible says is true.
Here's a second example. Someone is very suspicious about the reliability of consumer magazines. You try to convince him that Consumer Reports is reliable by pointing out that Consumer Reports ranks itself very highly in an article evaluating the reliability of consumer magazines. Probably your argument will fail to convince.
In these two examples, identify the premises and conclusion in each argument. Then describe what it is about the argument that makes it question-begging.
When I say that a certain sentence has the property of being true, what am I saying? For example, when I say that the English sentence "The Rocky Mountains are in North America'' is true, am I attributing some mysterious property to the sentence? Not really. All I'm saying is that the world is the way the sentence says it is. When I say that the sen- tence is true) all I'm saying is that the Rocky Mountains are in North An1erica. So in a way, the concept of truth is often "redundant:' Sometimes when I use the concept of truth, I could say the same thing without using that concept.
In high school English, your teacher might have told you to avoid redundancy. If you hand in an essay containing the sentence "Oscar is an unmarried bachelor;) the essay might co1ne back 'vith ((unmarried" crossed out and the inarginal conunent ((avoid redun- dancy:) The "'Ord ((unmarried)' is redundant because «Oscar is an un1narried bachelor1' means exactly the sa1ne thing as «Oscar is a bachelor.)) Adding the 'vord ((unmarried" is to spill useless ink. The Redundancy Theory of Truth claims that the word "true'' is redun- dant in just this sense. "It is true that the Rockies are in North America" says exactly what the sentence "The Rockies are in North America'' asserts. This helps show why truth isn't
19
20 Part One: Introduction
'.' .. '.' '' '"''.'.'' '.,,' .. '',' '' .. ' ... ''' '.'.',.'.' '' ... ' .. '.''.' ... '.''' a mysterious property. If you believe a statement P, you also believe that P is true. So, if you have any beliefs about .the world at all, you should be quite comfortable applying the concept of truth to those beliefs.
"True for Me"
You'll see from these remarks that the expression "It is trne for me" can be dangerously lnisleading. Son1etimes saying that a statement is true ((for you" just ineans that you believe it. If that is what you want to say, just use the word "belief" and leave trnth out of it. How- eve1; there is a 1nore controversial idea that inight be involved here. Sometimes people use the expression "true for n1e)) to express the idea that each of us makes our U\Vll reality and that the beliefs we have constitute that reality. I'll assume this is a mistake. My concept of truth assumes a fundamental division between the way things really are and the way they may seem to be to this or that individual. This is what I meant in Chapter 1 by distinguish- ing the objective realm and the subjective realm.
Wishful Thinking
Closely related to this distinction between objective and subjective is a piece of advice: We should avoid wishful thinking. Most of the things we believe areiit made true by our believing them. That the Rockies are in North America is a fact that is independent of our thought and language. We don't bring this geographic fact into being by thinking or talk- ing in the way we do.
Self-Fulfilling Prophesies
In saying this, I'm not denying that the thoughts we have often affect the world outside the mind. If I think to myself, "I ca1it hit a baseball;' this may have the effect that I do badly in the batter's box; here my believing something has the effect that the belief is made true. This is the idea of a "self-fulfilling prophesy:' Notice how this causal chain works:
Thought I believe that I \Von't hit the baseball.
Action
I swing too high.
Truth I don't hit the baseball.
My believing a proposition causes an action, which has the effect of making the proposi- tion true.
I have no problem with the idea that various statements may be caused to be true by individuals thinking thoughts to themselves. What I deny is that the mere act of thinking, unconnected 'vith action or some other causal path\vay, can n1ake statements true in the world outside the mind. I'm rejecting the idea that the world is arranged so that it sponta- neously confor1ns to the ideas '"e happen to entertain.
Chapter 2: Deductive Arguments 21
'. '''' .. '.'' '' .... '' ... '.,' '.'.'' '. ,, '. ''.' '' .. '. ''''' ........ ' '.' .. ' .... . Later in this book, I'll investigate whether there are any exceptions to this principle
that says that we should avoid wishful thinking. Maybe there are some statements that become true just because we think they are. Here are some philosophical claims we'll consider:
Mathematical statements and definitions are made true by our regarding them as such; for exa1nple, "2 + 3:::::: 5" is true just because v;e choose to define our ter1ninol- ogy ("2;' "+;'etc.) in the way we do (Chapter 4).
Some statements about the contents of my own mind (for example, "I am in pain") are made true just by my believing they are true (Chapter i3).
• Ethical statements are true just because God, society, or some individual agent thinks they are (Chapter 31).
I'm mentioning these philosophical claims here without tipping my hand as to whether I think any of them is plausible. If any of them were correct, they would be exceptions to the pattern I've just described. For the moment, though, I'm merely noting that belief and truth are generally very separate questions.
Review Questions
l. When is a statement or idea valid? (a trick question)
2. Define what it means to say that an argument is deductively valid.
J. Invent an example of a valid argument that has false premises and a true conclusion. Invent an example of an invalid argument that has true premises and a true conclusion.
4. Can a statement be a premise in one argument and a conclusion in another? If you think so, give an example.
5. \Vhich of the follov;ing argument forms is valid? Which is invalid? For each of the invalid ones, construct an example of an argument \"lith that form in which the premises are true and the con- clusion false:
If P, then Q If P, then Q p Q
(a) (b)
Q p
If P, then Q If P, then Q
Not-P Not·Q (c) (d)
Not-Q Not-P
For the argument forms you think are fallacious, invent names for these fallacies by using the vocabulary about conditionals presented in the box on page i8.
6. A sign on a store says, "No shoes, no service." Does this mean that if you v1ear shoes, then you will be served?
Part One: Introduction
.. ' .. ''.' .. ''.'.' ........... ' ....... '.' ............. ' .... '. ' ......... '' .. . J. What does it mean to say that an argument is "circular," that it "begs the question"? Construct
an example of an argument of this type different from the ones presented in this chapter.
8. What does it mean to say that truth is objective, not subjective?
Problems for Further Thought
1. The Redundancy Theory of Truth may seem plausible as an account ofvJhat the following sen· tence means:
It is true that the Rockies are in North America.
Does it \Vork as well as an explanation of what the following sentence means?
Some statements that are true have not been formulated yet.
2. Consider the follo\ving argument:
I release an othenvise unsupported apple from my hand a few feet from the earth's surface.
The apple falls to earth.
Is this argument deductively valid? What is the logical form of this argument?
J. (Here is a problem that was drawn to my attention by Richard Behling.) In this chapter, I said that each argument has a single logical form. This is the skeleton into which terms can be substituted to obtain the argument. \VJhat I said is an oversimplification. A given argument can be obtained from many logical forms. For example, consider the following argument:
Fred Jives in California. If Fred lives in California, then Fred lives in the United States.
Fred lives in the United States.
This argument can be obtained from both of the following skeletons by substitution:
X R ifX, then Y S
(a) (b) y T
Argument form (a) is valid, but (b) is invalid. The argument about Fred is valid.
Here's the problem: Use the concept of logical form to define when an argument is valid, and \vhen it is invalid, without falling into the trap of thinking that each argument has exactly one logical form.
4. In this chapter, I claimed that there are "objective truths." Do you agree? Construct an argument in which you try to demonstrate that such things exist, or that they do not.
MySearchlab Connections
Watch. Listen, Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
Chapter 2: Deductive Arguments 2.3
•• < •• '< ' ••• '.'' ''' ''.'.' '' '''.'.''''' ''' '''.'' ''. ''' '',., •• '' ' ••• '' •• '. ''
WATCH AND LISTEN
©{WATCH (<'"-[LISTEN
i. Listen "Truth: A How-to Guide"-Talk of the Nation
Jn this interview with NPR, Professor Simon Blackburn discusses the debate be- tween objectivists and subjectivists about truth throughout history ond today. What does Blackburn think is wrong with the subjectivist position? Do you think that there is absolute truth or that each person determines what is "true for him or her"?
RESEARCH AND EXPLORE
:!(<-(EXPLORE
l. Research deductive validity 2. Research objective truth
chapter 3 ..............................
24
Inductive and Abductive Arguments
Chapter Outline
Deductive Validity Is a Limitation
Nondeductive Inference-A Weaker Guarantee
Two Gambling Strategies
Universal La\vs
Detective Work
Induction
Two Factors Influence Inductive Strength
Abduction
Inferring What Isn't Observed
Abduction Differs from Induction
Can You Deduce the Explanation front the Observations?
Deducing Observational Predictions from a Theory
When the Prediction Comes True
When the Prediction 1\irns Out to Be False
How True Predictions and False Predictions Are Interpreted
The Surprise Principle: When Does Successful Prediction Provide Strong Evidence!
Evidence May Discritninate between Some Hypotheses While Failing to Discriminate between Others
True Prediction Isn't Enough
Modest Favoring
The Surprise Principle Summarized
The Only Game in Town Fallacy
In Chapter 2, I explained the idea of deductive validity. In a deductively valid argument, the premises provide an absolute guarantee that the conclusion is true: If the premises are true, there is no way in the world that the conclusion can be false.
Chapter 3: Inductive and Abductive Arguments 25
'' .. ' ....... ". '. ' ............... ' ......... ' .... ' ....... ' ........... ' ..... ' Deductive Validity Is a Limitation
This feature of deductive arguments may sound like a virtue. It is a good thing when an argument provides this sort of strong guarantee. This virtue, ho,vever, can also represent a kind oflimitation. Granted, a deductively valid argument that has true premises can't have a false conclusion; but it is also a property of such arguments that the conclusion can't say anything that wasn't already contained in the premises.
To see what this means, consider what you could validly deduce from the result of an opinion survey. Suppose you were interested in finding out what percentage of registered voters in a county are Democrats. You do1it feel like contacting each of them and asking, so you open the phone book and make, let's say, one thousand telephone calls.
Suppose the result of your survey is that 60 percent of the people called say they are Democrats. What you want to know is the percentage of Democrats in the whole county. Could you construct a deductively valid argument here? Can you deduce that (approxi- mately) 60 percent of the voters in the county are Democrats from a premise that describes the result of your survey? The answer is no, for two reasons. The fact that 60 percent of the people called said they are Democrats doesn't deductively guarantee that any of them are. And even if 60 percent of the people called are Democrats, you can't validly deduce from this that (approximately) 60 percent of the voters in the county are Democrats. That is, neither of the following arguments is deductively valid:
60 percent of the people called said they are Democrats.
60 percent of the people called are Democrats.
60 percent of the people called are Democrats.
Approximately 60 percent of the voters in the county are Democrats.
Why can't you validly deduce these conclusions? The reason is that in a deductively valid argument, it is impossible for the conclusion to be false if the premises are true. But it is possible that everyone you called in your survey lied. In addition, it is possible that the percentage of Democrats in the whole county is only 25 percent, even if nobody lied in your phone survey. In saying this, I'm not saying that the people you called actually lied, and I'm not saying that the real percentage in the whole county is only 25 percent. I'm just saying that these are possible, given the result of your phone calls. The result of your telephone survey doesn't absolutely rule out these possibilities; this means that you caiit deduce the percentage of Democrats in the whole county from what the one thousand people said on the phone.
So the absolute guarantee that a deductively valid argument provides has this limitation: Insisting that an argument be deductively valid prohibits you from reaching conclusions that go beyond the information given in the premises.
It would make sense to insist that an argument be deductively valid if you wanted to avoid even the smallest risk of having a false conclusion 'vith true premises. Ho"vever, '"e are often willing to gamble. For example, we might think that the result of the phone survey does
26 Part One: Introduction
'. '''.' ........ ' ........ ' ... ' .. ' ... ' .... ''"'' ... ' ... '' .. ''', '' .. '' ''' '''' provide information about the composition of the county. We might think that the survey provides a pretty good reason for concluding that about 60 percent of the county voters are Democrats. Hov/ever, the "good reasoll' isn't a deductively valid one.
Nondeductive Inference-A Weaker Guarantee
We have here a fundamental characteristic of nondeductive inference. Suppose 've conclude that about 60 percent of the voters in the county are Democrats, based on the premise that 60 percent of the people called said they were Democrats. In this case, the premise doesn't provide an absolute guarantee that the conclusion is true. Ho,vever, there is a lesser kind of guarantee that this premise may provide. If the argument is a strong one, the premise makes the conclusion probable; it provides a good reason for thinking the conclusion is true; it makes the conclusion plausible. Instead of an absolute guarantee, we have here a 'veaker guarantee. You are running a risk of being \vrong about your conclusion, even if your premise is true. But this risk might be a reasonable one to take. The conclusion might be a good bet, given that the premise is true.
Two Gambling Strategies
The language in the previous paragraph suggests that you can think about the difference betv/een deductive and nondeductive argu1nents in tern1s of ideas about gambling. Consider two sorts of gamblers. The first I'll call the extreme conservative. This individual refuses to wager unless winning is a sure thing. The second individual I'll call the thoughtful risk taker. This individual at times enters into risky gambles hoping to win. Each strategy has its virtue and its limitation. The virtue of the conservative strategy is that you'll never lose a gamble. Its limitation is that there are gambles you will decline to take that you could have won. The limitation of thoughtful risk taking is that you can lose money. Its virtue is that it can lead you to win wagers by taking risks.
Limiting yourself to deductively valid arguments is a conservative strategy. You avoid the risk of reaching false conclusions from true premises. The limitation is that you decline to say anything that goes beyond the evidence. Nondeductive arguments are riskier. The gain is that you can reach true conclusions that go beyond what the premises say; the risk is that you may reach a false conclusion fro1n true pre1nises.
In science as 'vell as in everyday life, \Ve make nondeductive inferences all the tin1e. We are often prepared to take risks. Each of us has beliefs about the future. These, however, are1it deduced from the observations we made in the present and past.
Universal Laws
Science is a risky business in another '\vay. Scientists often try to reach conclusions about universal laws. An example is Isaac Newton's (1642-1727) universal law of gravitation, which you may have studied in high school. This law says that the gravitational attrac- tion between two bodies is proportional to the products of their masses and is inversely
Chapter 3: Inductive and Abductive Arguments
proportional to the square of the distance between them. This law describes how much gravitational attraction there is behveen any h\'O objects, no matter 'vhere those objects are located and no 1natter vvhen those objects exist. Ne,vton's Ja,y is universal in scope- it describes what is true at any time and place. This isn't an isolated example. In lots of sciences, there are universal staten1ents that scientists think are 'vell supported by evidence.
Could Newton have deduced his law from the observations he made and the experi- 1nents he conducted? No. His la'v is universal in scope. His observations, ho,vever, 'vere conducted in a rather narrowly limited range of places and times. Newton didn't go back- ward in time to check if his law held true 3 million years ago. Nor did he send a space- ship to a distant galaxy to do the required measurements. When scientists conclude that a universal la,v is true or probably true, based on pre1nises that describe the observations they have made, they aren't making deductively valid arguments.
Science is an ainbitious enterprise. Science ventures beyond 'vhat is strictly observed in the here and no'iv, just as the conclusion in a nondeductive argutnent ventures beyond the infor1nation strictly contained in the pre1nises.
Detective Work
I said before that nondeductive argun1ents are constantly used both in science and in everyday life. Newton was my scientific example. Let me describe the calculations of Sherlock Holmes as my everyday one.
Holmes was constantly telling Watson that he figures out detective problems by "deduction:' Although Holmes was a very good detective, I doubt that he solved his puzzles by strictly deductive methods. Holmes didn't observe the crimes he was later called upon to investigate. What he observed were clues. For example, suppose Holmes is trying to solve a murder. He wonders whether Moriarty is the murderer. The clues Holmes gathers include a gun, a cigar butt, and a fresh footprint, all found at the scene of the crime. Suppose the gun has an ''M'' carved in the handle, the cigar is Moriarty's favorite brand, and the footprint is the size that would be produced by Moriarty's ample foot. Can Holmes deduce from these clues that Moriarty is the murderer? No. Although the information may make that conclusion plausible or probable, it doesn't absolutely rule out the possibility that someone else did the dirty deed.
I've been emphasizing that in a strong nondeductive inference, the premises n1ake the conclusion plausible or probable; they don't absolutely guarantee that the conclusion must be true. I now want to talk about the difference between two sorts of nondeductive inference-inductive and abductive.
Induction
Inductive inference involves taking a description of some sample and extending that descrip- tion to items outside the sample. The voter survey discussed before provides an example:
60 percent of the county voters called are Democrats.
Approximately 60 percent of the county voters are Democrats.
28 Part One: Introduction
Notice that in this exa1nple the vocabulary present in the argun1ent's conclusion is already used in the premise. Although the conclusion goes beyond what the premise asserts (\vhich is \Vhat Jnakes the argument nondeductive), no ne\v concepts are introduced in the conclusion.
Two Factors Influence Inductive Strength
In the case of deduction, I said that an inference is either deductively valid or it isn't. Validity is a yes/no affair. It is like pregnancy. Inductive strength, however, isn't a yes/no tnatter; inductive argun1ents are either stronger or \veaker. Inductive strength is a 1natter of degree.
T\vo factors affect ho\v strong or \veak an inductive argument is. The first is sample size. If you called one thousand individuals in your phone survey, that would make your conclusion stronger than if you had called only one hundred. The second factor is the representativeness or unbiasedness of the sample. If you called one thousand indi- viduals drawn at random from a list of voters, that would make the resulting inference stronger than if you had called one thousand members of labor unions. The percentage of Democrats in labor unions may be higher than that found in the population as a \vhole. If so, you are biasing your sample by dra,ving it exclusively fro1n a union inen1- bership list.
By making a telephone survey, you are failing to contact people who dotit have phones. Is this a problem? That depends. If the percentage of Democrats with phones is approxi- mately equal to the percentage of Democrats among registered voters, no bias is intro- duced. On the other hand, if people with phones are disproportionately Democrats or disproportionately Republicans, yam phone survey will have introduced a bias.
How do you avoid having a biased sample? Sometimes this is done by "randomiza- tion:' If you have a list of all the county voters, dra\ving names "at rando1n" means that each name has the same chance of being selected. However, this process of selecting at rando1n can fail to ensure an unbiased san1ple. For exan1ple, suppose you dra'v na1nes at random, but all the people you contact happen to be women. If women are disproportion- ately Democrats or disproportionately Republicans, your sample is biased. I don't say that random draws from the voter list will probably result in this sort of bias. My point is just that randomizing doestit absolutely guarantee that your sample is unbiased. I wotit say more here about how you can avoid having a biased sample. This fine point aside, the basic idea is this: Inductive argu1nents are stronger or \Veaker according to (1) the sa1nple size and (2) the unbiasedness of the sample.
Abduction
I now move to abduction-inference to the best explanation. I'll begin with an example of an abductive inference that was important in the history of science. After saying what is distinctive about this form of inference, I'll describe two principles that are relevant to deciding \Vhether an abductive inference is strong or 'veak.
Chapter 3: Inductive and Abductive Arguments
'. ' .. ' ... ' ... ' ' ' ... ''.'.'.' ........ " ............ '' .......... '.' ' .... '. '.'. Inferring What Isn't Observed
Gregor Mendel (1822-1884) was an experimental biologist who worked in a monastery in Mora.,ia. He is credited with having discovered genes, the particles in living things that allow parents to transn1it characteristics to offspring in reproduction.
The first thing to note about Mendel's discovery is that he never actually saw a gene (or an "element;' as Mendel called them). Although more powerful microscopes made this possible later, Mendel never saw even one of them. Rather, Mendel reasoned that the observations he made could be explained if genes existed and had the characteristics he specified.
Mendel ran breeding experiments in the monastery's garden. He crossed tall pea plants with short ones and noted the proportion of tall and short plants among the offspring. Similarly, he crossed plants that had wrinkled peas with plants that had smooth peas, and he noted the mix of wrinkled and smooth plants among the progeny. He then crossed some of those offspring with each other, and he saw the proportion of various characteris- tics found in the next generation.
Mendel observed that when plants of certain sorts are crossed, their offspring exhibit characteristics in very definite proportions. Mendel asked himself a question that never figured in my discussion of induction. He asked why the crosses produced offspring with characteristics distributed into such proportions.
This why-question led Mendel to invent a story. He said, suppose each plant contains particles (genes) that control the observed characteristics of tall and short, wrinkled and smooth, in certain specific ways. He conjectured that each parent con- tributes half its genes to the offspring and that this process occurs in accordance with definite rules. The whole invented story had this property: If the story were true, that would explain why the breeding experiments had the results that Mendel observed them to have.
It should be quite clear that Mendel's theory of the gene went beyond the observations then available to him. He never saw a gene, but his theory postulates the existence of such things. I noted before that it is a general feature of nondeductive inference, whether induc- tive or abductive, that the conclusion goes beyond the premises. We see here, ho,vever, a respect in which abduction differs from induction.
Abduction Differs from Induction
If Mendel had made an inductive inference, he simply would have claimed that the observed results of the experiments he ran in his garden would also occur elsewhere. His experiment \Vas made in Europe at the end of the nineteenth century. An inductive extension of the description of his experiment might conclude that the same results would occur in twentieth-century North America as well. Had Mendel limited himself to this suggestion, no one 'vould re1ne1nber hhn now as the father of genetics. His itnportant inference was abductive, not inductive. He didn't simply claim that the experiment could be replicated. Rather, he formulated a theoretical explanation of why the results
30 Part One: Introduction
..... ' ................................. ' .................... ' ........... '. occurred. Mendel's inference dre'v a conclusion concerning something he did not see (genes), based on premises that described what he did see (the results of the experimen- tal crossings).
Can You Deduce the Explanation from the Observations?
Let's attend to Mendel's inference more carefully. The following is not a deductively valid argument:
Experimental crosses in the pea plants were observed to exhibit such-and-such results.
There are genes, and they obey laws L.
Remember: You can't validly deduce a theory from a set of observations. Why can't you do this? Basically, the last argument attempts to infer a theory about the
cause from the observation of its effects. There are, however, lots of possible causes that might have been responsible for the observed effects. The argument is deductively invalid for the same reason the following argument is also invalid:
A pistol with an "M" on the handle, an El Supremo cigar butt, and a size 12 footprint were found next to the murder victim's body.
Moriarty is the murderer.
Although Moriarty may be the most plausible suspect, the clues, in themselves, don't ab- solutely guarantee that he must have done the deed. Mendel and Holmes were making an inference about what is probably true, given the observations. They weren't inferring what is absolutely guaranteed to be true by the observations.
Deducing Observational Predictions from a Theory
If a set of observations doesn't deductively imply a theory, then perhaps the reverse is true: Maybe a theory deductively implies some observations. This corresponds more closely to what Mendel did. He saw that his theory of the gene implies that certain experimental re- sults ought to occur. He then saw that those predictions came true. He concluded that the truth of the predictions was evidence that the theory is true.
When the Prediction Comes True
So a better representation of Mendel's inference might go like this. The theory entailed a prediction. The prediction came true. Hence, the theory is probably true. What we now need to see is that the following form of argument is not deductively valid:
Chapter 3: Inductive and Abductive Arguments 31
....... ' .......................... ' ......................... '.' .......... . If there are genes and they obey laws L, then experimental crosses in the pea plants should exhibit such-and-such results.
Experimental crosses in the pea plants were observed to exhibit such-and-such results.
There are genes and they obey laws L.
This is deductively invalid for the same reason that the following argument is too:
If Jones lives in Wisconsin, then Jones lives in the United States.
Jones lives in the United States.
Jones lives in Wisconsin.
Note that these two arguments have the same logical form. Scientists often test their theories by seeing whether the predictions made by the theories
come true. There is nothing wrong with doing this. The point, however, is that the truth of the theory doesn't follow deductively from the truth of the prediction. Scientists are reason- ing 11011deductively when they decide that a theory is plausible because its predictions have come true. Successful prediction isn't absolutely conclusive proof that the theory is true.
When the Prediction Turns Out to Be False
On the other hand, if the predictions entailed by Mendel's theory had come out false, that would have allowed him to deduce that the theory is mistaken. That is, the following argument is deductively valid:
If there are genes and they obey laws L, then experimental crosses in the pea plants should exhibit such-and-such results.
Experimental crosses in the pea plants did not exhibit such-and-such results.
It is false that there are genes and they obey laws L.
In other words, a failed prediction is conclusive proof that the theo1y implying the prediction is false.
How True Predictions and False Predictions Are Interpreted
Let's generalize these points. Let T be a theory and Pa prediction the theory makes. If the prediction comes out true, we can't deduce that the theory is true. If, however, the predic- tion comes out false, we can deduce the theory is false:
Invalid Valid
If T, then P lfT, then P p NotP
T NotT
32 Part One: Introduction
........... '' ....... ' '' .. '.'.''.' .. ''.' ... ''' '.'''' '''' '' .. ''.','. ''''.'' DEDUCINC: THAT A THEORY IS TRUE
Recall from Chapter 2 that a deductively invalid argument can be turned into a valid one by adding premises. I will now exploit this fact to show how the truth of a theory can be deduced from the fact that it makes a successful prediction, if certain further assumptions are made.
Suppose we wish to design an experiment that tests two theories. Here the prob· lem isn't one of evaluating a single theory, but of seeing which of two theories is more plausible. To test one theory (T,) against another (T,), we want to find a prediction over which they disagree. Suppose T, predicts that P will be true, while T, predicts that P will be false. If we assume that one or the other theory is true, we can find out whether P comes true and then deduce which theory is true. For example, if P turns out to be true, we can reason as follows:
T1 or T2 If T,, then P.
If T,, then not·P.
p
T,
Notice that this argument is deductively valid. Also note that if P had turned out to be false, we could have deduced that T, is correct.
The difference between these two arguments-one deductively valid, the other not- suggests there is an itnportant difference behveen the \\'ay scientists argue that theories are true and the way they argue that theories are false. It is possible to reject a theory just on the basis of the false predictions it makes, using a deductively valid argument; but it isn't possible to accept a theory just on the basis of the true predictions it makes, using a deduc- tively valid argument. I will discuss this difference again in Chapter 8.
So far I've explained how a deductively valid argument can lead a scientist to reject a suggested explanation. But ho\v do scientists ever interpret observations as providing strong evidence in favor of the explanations they consider? This must involve a nondeduc- tive inference. But 'ivhat are the rules that govern such inferences? I no\v present t\VO ideas that are relevant to evaluating abductive arguments. I call these the Surprise Principle and the Only Game in Town Fallacy.
The Surprise Principle: When Does Successful Prediction Provide Strong Evidence?
I've argued that you emit validly deduce that a theory is true just from the fact that some prediction it makes comes true. But maybe only a small modification of this idea is needed. Perhaps all we need to say is that a theory is made highly probable or plausible when a pre- diction it nlakes co1nes true. I no\v \Vant to explain \vhy this refor1nulation is also 1nistaken.
Chapter 3: Inductive and Abductive Argun1ents 33
'''.' '.'''.''' ''' '''''' '''' '',. ''',' '''. ''.''. ''''.' '''.'.''''' ''. ''' ','
An unconscious patient is brought into the en1ergency roon1 of a hospital. What is wrong? What would explain the fact that the patient is unconscious? The doctor on duty considers the hypothesis that the patient is having a heart attack. How should the doctor test whether this hypothesis is true? Well, the hypothesis predicts that the patient will have a heart (after all, if someone is having a heart attack, he or she must have a heart). The doctor verifies that this prediction is correct-the patient, indeed, does have a heart. Has the doctor thereby obtained strong evidence that the patient is having a heart attack? Clearly not. This is an example in which you don't obtain serious support for a hypothesis just by showing that one of its predictions is correct.
You go to the gym, and someone tells you he is an Olympic weight lifter. You reason that ifhe is an Olympic weight lifter, he should be able to pick up the hat you are wearing. You offer him your hat; he lifts it without difficulty. Have you thereby obtained strong evidence that he is an Olympic weight lifter? Clearly not. Once again, the hypothesis under test isn't strongly supported by the fact that one of its predictions turns out to be correct.
What has gone wrong in these two cases? In the first example, the presence of a heart isn't strong evidence that the patient is having a heart attack. The reason is that you would expect the individual to have a heart even if he weren't having a heart attack. In the second example, the man's lifting the hat isn't strong evidence that he is an Olympic weight lifter. The reason is that you would expect him to be able to do this even ifhe weren't an Olympic lifter.
What should you look for if you want to test the hypothesis that the patient has had a heart attack? What you want to find is a symptom that you would expect to find if the patient were having a heart attack, but would expect not to be present if the patient were not having a heart attack. Suppose an erratic electrocardiogram (EKG) almost always occurs when there is a heart attack but rarely occurs when there is no heart attack. This means that the presence of an erratic EKG would be strong evidence that the patient is suffering a heart attack.
What sort of test should you use if you want to see if the man in the gym really is an Olympic weight lifter? Suppose Olympic weight lifters (of this man's weight) can almost always lift four hundred pounds, but people who aren't Olympic lifters can rarely do this. This means that his managing to lift four hundred pounds would be strong evidence that he is an Olympic weight lifter.
Think of the unconscious patient and the weight lifter as posing discrimination prob- lems. The problem is to find evidence that strongly discriminates between two hypotheses. In the first example, the competing hypotheses are as follows:
H,: The patient is having a heart attack.
H,: The patient isn't having a heart attack.
An erratic EKG strongly favors H, over H,; the mere fact that the patient has a heart does not.
The same holds true for the second example. The problem is to find evidence that discriminates between the following two hypotheses:
H,: This man is an Olympic weight lifter.
H,: This man isn't an Olympic weight lifter.
The fact that the man can lift four hundred pounds strongly favors H, over H,; the fact that he can lift a hat does not.
34 Part One: Introduction
The Surprise Principle describes what it takes for an observation to strongly favor one hypothesis over another:
The Surprise Principle: An observation 0 strongly supports hypothesis H, over hypothesis H, if both the following conditions are satisfied, but not otherwise: (1) if H, were true, 0 is to be expected; and (2) if H, were true, 0 would have been unexpected.
Let's apply this principle to the first example. Consider the observation that the patient has an erratic EKG. If the patient were having a heart attack (H.), we would expect him to have an erratic EKG. And ifthe patient weren't having a heart attack (H,), we would ex- pect him not to have an erratic EKG. This explains why the erratic EKG is strong evidence of a heart attack-the EKG strongly favors H, over H,. Now consider the observation that the patient has a heart. If the patient were having a heart attack, we would expect him to have a heart. But 've \Vould expect hin1 to have a heart even if he \Veren't having a heart attack. This explains why the presence of a heart isn't strong evidence that the patient is having a heart attack-the observation doesn't strongly favor H, over H,. We were looking for an explanation of why the erratic EKG provides telling evidence, whereas the presence of a heart does not. The key is in condition 2 of the Surprise Principle.
NO SURPRISE/SURPRISE
The Surprise Principle involves two requirements. It would be more descriptive, though more verbose, to call the idea the No Surprise/Surprise Principle.
The Surprise Principle describes when an observation 0 strongly favors one hy- pothesis (H,) over another (H,). There are two requirements:
(1) If H, were true, you would expect 0 to be true.
(2) If H, were true, you would expect 0 to be false.
That is, (1) if H, were true, 0 would be unsurprising; (2) if H, were true, 0 would be surprising.
The question to focus on is not whether the hypotheses (H, or H,) would be surpris- ing. The Surprise Principle has nothing to do with this. To apply the Surprise Principle, you must get clearly in mind what the hypotheses are and what the observation is.
Take a few minutes to apply the Surprise Principle to the example of the weight lifter. Make sure you see how the principle explains why being able to lift a hat isn't strong evi- dence, whereas being able to lift four hundred pounds is.
Evidence May Discriminate between Some Hypotheses While Failing to Discriminate between Others
In the examples of the heart attack and the weight lifte1; H, says only that H, is false. In other abductions, ho'ivever, H21nay say n1ore than this. Suppose you see someone crossing
Chapter 3: Inductive and Abductive Argu1nents
campus carrying several philosophy books. You wonder whether the person is a philoso- phy major. Here are two hypotheses you might consider:
H,: 111e person is a philosophy major.
H2 : The person is an engineering inajor.
According to the Surprise Principle, the observation you have 1nade favors H1 over H 2 • But now consider the following third hypothesis:
H3: The person isn't a student, but is in the business of buying and selling philosophy books.
Although your observation discrin1inates bet,veen H1 and H2 , it does not discrin1inate bet,veen H1 and Hy
This brings out an important fact about how the Surprise Principle applies to abduc- tive inferences. If you want to know whether an observation strongly supports a hypothesis, ask yourself what the alternative hypotheses are. For an observation to strongly support a hypothesis is for it to strongly favor that hypothesis over the others with which it competes.
True Prediction lsn 't Enough
The point of the examples about the unconscious patient and the weight lifter was to show 'vhy successful prediction doesn't automatically provide strong evidence. If so1neone is having a heart attack, that predicts that he will have a heart; but the presence of a heart isn't strong evidence the person is having a heart attack. If someone is an Olympic weight lifter, that predicts that he will be able to lift a hat; but lifting a hat isn't strong evidence the person is an Olympic weight lifter.
There is a scene in the Monty Python movie The Life of Brian that illustrates this idea. The setting is a marketplace. Around the perimeter of the market are assorted prophets and sooth- sayers. The camera pans from one to the other. We see that in each case, the prophet holds a crowd of people in rapt attention. The first prophet predicts that tomorrow a purple monster will rise out of the desert and devour three villages. The crowd is amazed at these predictions and no doubt will conclude that the prophet has special powers to foresee the future if the predictions come true. After shovving us a fevv prophets of this sort, the can1era conies to an individual who very calmly makes the following predictions: Tomorrow, many people will get up early. Others will sleep longer. Some people will decide to have breakfast, while others will postpone eating until later in the day. And so on. The joke is that the crowd that is looking at this guy is just as awestruck as the crowds that were in front of the more outlandish prophets.
There is a lesson here. How would you test the hypothesis that someone has special powers to foresee the future? If he predicts events that people without special powers can easily predict, it isn't very iinpressive that his predictions con1e true. If, ho\vever, he pre- dicts events that nor1nal people aren't able to foresee, and then these predictions co1ne true, 've are n1ore impressed. The Surprise Principle explains \vhy the success of "safe)) predictions provides less compelling evidence than the success of"risky" predictions.
Here's a related example of the Surprise Principle in action. Many people thought that the astrologer Jeanne Dixon had special powers to predict the future. After all, she predicted the assassination of President John F. Kennedy and several other events that
35
Part One: Introduction
no one could have guessed were going to happen. If Jeanne Dixon played it safe and only predicted events that everybody knew were going to happen, we wouldn't be impressed. But aren't we being sensible in reasoning that she probably did have special powers, since this and other daring predictions came true?
Although this reasoning may seem to conform to the Surprise Principle, it doesn't. The thing people sometimes forget is that Jeanne Dixon made thousands of predictions and most of them turned out false. It isn't surprising at all that some handful of these should have come true. Although Jeanne Dixon predicted the Kennedy assassination, which surprised (practically) everyone when it happened, it isn't at all surprising that someone with no special powers should be lucky a few times every thousand tries. The Surprise Principle, properly understood, tells us why we shouldn't take Jeanne Dixon's few successes as strong evidence that she had special powers.
Modest Favoring
The Surprise Principle shows when an observation strongly favors one hypothesis over another. Ho,vever, so1netin1es our observations are not so telling and unequivocal. Some- times the observations ( 0) favor one hypothesis (H,) over another (H,), but only modestly. This will be true when H, confers on 0 a higher probability than H, does, but the differ- ence is modest. What the Surprise Principle and this idea about O's modestly favoring H, over H2 have in co1n1non is this: both make use of the idea that an observation favors one hypothesis (H,) over another (H,). This will be true when the probability of 0, according to H,, exceeds the probability of 0, according to H,.
The Surprise Principle Summarized
In summary, the Surprise Principle gives advice on what a hypothesis must do if it is to be strongly supported by the predictions it makes. First, the hypothesis shouldn't make false predictions. Second, among the true predictions the hypothesis makes, there should be predictions we would expect not to come true if the hypothesis were false.
When we ask whether an observation 0 strongly supports some hypothesis H,, the Surprise Principle requires that we specify what the alternative hypotheses are against which H, competes. It may turn out that 0 strongly favors H, over H,, but that 0 doesn't strongly favor H, over H,. The example of the person carrying philosophy books across campus illustrates this point.
The Only Game in Town Fallacy
I now turn to a second principle for evaluating abductive inferences. Suppose you and I are sitting in a cabin in the woods. We hear a strange rumbling sound in the attic. You ask, "! wonder what that could be?" I reply, "That is the sound of gremlins bowling in the attic:' You, being a sensible person, reply, "I really don't think there are gremlins in the attic:' I then challenge you to produce a more plausible explanation of the noises. You reply, "Gosh, I really don't have any idea why those noises occurred. I just think your story is implausible:' To this humble admission on your part, I make the following
Chapter 3: Inductive and Abductive Argu1nents 37
''''.'' '.' '''. ''.' ''' .. ' ... '. ' ..... ''.' ''.'.''.' .. ' ....... . rejoinder: "Look, rny story, if true, "\Vould explain ,vhy \Ve just heard those strange noises. If you don't \Vant to accept 1ny explanation, you 1nust produce a n1ore plausible explanation of your o\vn. If you can't, you have to accept rny explanation of the noises:'
AN INVESTMENT SWINDLE
Suppose you received a letter every month for a year from an investment firm. In each letter, a prediction is made as to whether "the stock of the month" will increase or decline in value during the next thirty days. You keep track of what happens to the stocks described each month. Each prediction comes true. Would you conclude from this that the investment firm has a method for reliably predicting stock market events?
Some years ago an "investment firm" sent out such letters, but the mailing was a swindle. The firm began with a list of10,ooo investors. In the first month, 5,000 inves- tors received letters predicting that stock A would go up; the other 5,000 received letters saying that stock A would go down. The firm then waited to see which prediction came true. During the second month, the firm sent letters to the 5,000 people who had received a true prediction during the first month. In the second month, 2,500 investors received letters predicting that stock B would go up; the other 2,500 received letters saying that B would go down. The process was repeated, so that by the end of ten months, a small number of investors had received ten letters, each containing a successful prediction.
The company then wrote to those people, asking each to invest a large sum of money. Most did so. The company then absconded with the funds. (This story is from Daniel Dennett's Brainstorms: Philosophical Essays on Mind and Psychology, Cambridge, MA, MIT Press, i978.)
The investors who were swindled thought they were making a reasonable abduc- tive inference on the basis of the company's track record. Describe the premise, the conclusion, and the reasoning that led the investors to think the conclusion was highly plausible. Were the investors making a strong inference?
What I just did was commit an abductive fallacy (mistake in reasoning), which I'll call the Only Game in Town Fallacy. The fact that you can't think of a more plausible explanation of the noises doesn't oblige you to accept the story I constructed. There is an alternative, which is simply to admit that the noises are something you don't know how to explain.
Abduction is sometimes described loosely as follows: If a theory explains some observation, and if no rival account is available that can do a better job of explaining it, then you should accept the theory. Although this description of abduction is roughly correct, it makes the mistake of sanctioning the Only Game in Town Fallacy. The fact that no rival account is better than the explanation I construct doesn't sho'v 1ny explanation is even minimally plausible. My gremlin theory is pretty silly, although maybe there is nothing now available that is clearly superior to it.
I \\'on't at this point try to fine-tune the idea of abductive inference any further, even though there is a lot more to be said about it. We now have before us the basic idea of inference to the best explanation. Wive seen that it is an important part of the scientific method. And I've described two principles that help guide us in evaluating whether an abductive inference is strong or \Veak.
Part One: Introduction
................................... ''. ' .. ' .. ' .......... ' .......... ' ..... '. Review Questions
l. What is the difference between deductive validity and inductive strength?
2. What is the difference between induction and abduction?
J. What factors affect how strong an inductive argument is?
4. Suppose a given observation discriminates between t\vo hypotheses, but a second observation fails to do this. Construct an example, different from the ones presented in this chapter, illustrat- ing the point. Show how the Surprise Principle applies to your example.
5. An observation can succeed in discriminating between hypotheses H1 and H2 but fail to discrimi- nate between H1 and H3. Construct an example that illustrates this point that is different from the ones presented in this chapter. Show ho\v the Surprise Principle applies to your example.
6. What is the Only Game in Town Fallacy? What does it mean to call it a "fallacy''?
Problems for Further Thought
l. Suppose you wanted to find out \Vhat percentage of the adults in your county are vegetarians. You obtain a list of unmarried adults in the county and contact them to do a survey. Is this sample a biased one? Why or why not?
2. Although induction and abduction were described in this chapter as separate kinds of inference, they have a good deal in common. The Surprise Principle was introduced as applying to abduc- tion, but it applies to induction as well. Suppose an urn is filled with one thousand balls, each of them either red or green. You reach into the urn, sampling at random, and bring out one hundred balls. Your sample contains fifty-one red balls and forty-nine green ones. Here are some hypotheses to consider:
H,: All the balls in the urn are green. H,: 75 percent of the balls are green. H3: 50 percent of the balls are green. H4: 25 percent of the balls are green.
Suppose you think that H, and H2 are the only possibilities. Does the observation strongly sup- port one over the other? If so, which? How does the Surprise Principle apply to this question? Suppose, instead, that you think that H, and H3 are the only possibilities. Does the sample strongly support one over the other? If so, which? Suppose, finally, that you think that H, and H4 are the only possibilities. Which is best supported? Why?
3. Why think that any of the beliefs you have about the world outside your own mind are true? For exam· ple, why are you no\'J entitled to think there is a printed page in front of you? Presumably you believe this on the basis of sense experience (sight, touch, etc.). Construct an abductive argument \Vhose conclusion is that there is a printed page in front of you. Make sure your inference obeys the Surprise Principle. Does this abductive argument prove there ls a printed page in front of you? Explain.
4. At the beginning of Chapter 2, I presented deductive validity, inductive strength, and abductive strength as mutually exclusive categories. This means that if an argument belongs to one category, it doesn't belong to any of the others. This is generally correct, but not always. Here is an inductive argument that isn't deductively valid:
I've observed one thousand emeralds and all have been green.
All emeralds are green.
Chapter 3: Inductive and Abductive Arguments 39
' .. '.''' .. '.' .. ' .... '.' .. '' .. ' .... ' ...... ' ............. ' ...... ' .... ' ..... . However, by adding a premise, I can produce a deductively valid argument:
I've observed one thousand emeralds and all have been green. If there are over five hundred emeralds in the universe, they \Vill all have the same color.
All emeralds are green.
Both of these arguments are inductive in the sense that both involve dra\ving a sample from a population and reaching a conclusion about that population. If so, some inductively strong argu- ments are also deductively valid.
The same point holds for abductively strong arguments. Usually they aren't deductively valid, but sometimes they are. Construct a strong abductive argument that obeys the Surprise Principle. Show how it can be made deductively valid by adding or modifying a premise.
5. There is a difference between not expecting 0 and expecting not-0. A person who never consid- ers vvhether 0 is true does not expect 0, but it would be wrong to say that she expects not-0. Of course, if someone expects not-0, it will also be true that she does not expect 0. Thus, "S expects not-0" implies "5 does not expect O," but the reverse is not true. \Y/ith this logical point in mind, explain why condition 2 of the Surprise Principle is formulated by saying "H2 leads you to expect not-0," rather than saying that "H2 does not lead you to expect O."
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchLab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
RESEARCH AND EXPLORE
'HEXPLORE
1. Inductive reasoning 2. Abductive reasoning 3- Surprise Principle
Suggestions for Further Reading
On the Nature of Philosophy
The essays in Charles Bontempo and S. jack Odell, eds. The Owl ofMinervo: Philosophers on Philosophy. New York: McGraw-Hill, i975.
Bertrand Russell. "The Value of Philosophy." In The Problems of Philosophy. Oxford: Oxford University Press, i950.
On Different Forn1s of Argu1nent
Monroe Beardsley. Thinking Straight. Englewood Cliffs, NJ: Prentice Hall, i975. Irving Copi. Informal Logic. New York: Macmillan, i986. Robert Fogel in. Understanding Arguments. New York: Harcourt, Brace, Jovanovich, i978. Ronald Giere. Understanding Scientific Reasoning. New York: Holt, Rinehart, and Winston, i984. Merrilee Salmon. Logic and Critical Thinking. San Diego, CA: Harcourt, Brace1 Jovanovich, i984.
part two .·· • • • • • • • • • • • • • • • • • .. ,• • '• •• ~ !
Philosophy of Religion
chapter 4 .......................... ~ ...
Aquinas's First Four Ways
Chapter Outline
The Concept of God
The First Two Arguments: Motion and Causality
Aquinas on the Cause of Motion
God Is a Person, Not Just a Cause That Exists Outside of Nature
The Birthday Fallacy
Why Can't Nature Be Infinitely Old?
Why Must Every Event in Nature Have a Cause?
The Third Argument: Contingency
Necessary and Contingent Beings
Possible Worlds
Reductio Ad Absurdum
Contingency and Eternity
Conservation Lalvs in Physics
The Birthday Fallacy (Again)
Necessary Beings other than God
Necessary and Contingent Propositions
Mathematical Truths
Names Differ from the Things Named
Nun1bers Aren,t Nun1erals
Sets
Necessity and Certainty Are Different
Numbers Are Necessary Beings
Aquinas,s Fourth Argu1nent: Properties That Come in Degrees
Criticizing an Argu1nent versus Showing the Argun1cnt's Conclusion Is False
Thomas Aquinas (ca. 1224-1274) was an enormously accomplished theologian and philosopher. In his masterwork, the Summa Theologiae, he presents five proofs that God exists (Aquinas called them "the five ways"). I'll discuss the first four now and the fifth in the next chapter.
Each of Aquinas's arguments begins with a simple observation that is supposed to be obvious to everyone. For exa1nple. the first argun1ent begins \vith the observation that
CEJ--[ Read u Five \Vays to Prove That God Exists", Pt. 1, Question 2, Art. 3 of Summa Theofogiae on www.mysearchlab .com
43
44 Part Two: Philosophy of Religion
physical objects are in motion. Each argument then proceeds through various other pre1nises to the conclusion that there is a God. Aquinas intends each of his proofs to be deductively valid.
In Chapter 3, I stressed that most of the hypotheses that scientists are interested in testing can't be deduced from observations. For example, Mendel couldn't deduce the existence and characteristics of genes fro1n the observations he n1ade on his pea plants. The sa111e is true in Aquinas,s argu1nents, as he fully realizes: You can't deduce the existence and characteristics of God just fron1 the sin1ple observations v;ith 'vhich his argu1nents begin. The existence of motion doesn't, all by itself, deductively imply the existence of God. Aquinas's arguments always include additional premises. It is these further premises that are supposed to link the starting observation with the conclusion that God exists.
The Concept of God
Before describing Aquinas's argun1ents, I need to say something about 'ivhat he n1eans by "God" and how I'll use that term. Aquinas took God to be a person-one who is all- powerful (omnipotent), all-knowing (omniscient), and entirely good (omnibenevolent)- all-PKG, for short. This conception of God is a familiar one in the traditions of)udaism, Christianity, and Islam (though there is room to debate, in these traditions, whether God is properly described in this way and other religions have other views about what charac- teristics God has). I'll assume provisionally that God, if such a being exists, has the three characteristics just mentioned. If we don't start with some preliminary picture of what God is, 've 'iVon't kno'iv 'ivhat 'iVC are talking about 'ivhen 'iVC ask 'iVhether God exists. Ho,vever, it is important to bear in mind that this conception of God is not the only one that is pos- sible. Indeed, in Chapter u, I'll consider an argument that suggests that God, if there is such a being, can't be all-PKG. The definition of God as an all-PKG being is a useful place to begin discussion, but it is only a point of departure.
Another caveat I should 1nention is that my versions of Aquinas's arguments 'vo1lt be ac- curate in all respects. It is often a subtle historical question what this or that philosopher had in mind in a given text. In this case and later in this book, when I discuss the ideas of other philosophers, I will often examine somewhat simplified versions of the arguments they con- structed. This may lead you to ask: Why is it worthwhile studying simplified versions of a great philosopher's arguments? Admittedly, there is a loss, but there is also a gain. The main point to be made here, at the beginning of an introductory text, is that it is useful to evaluate these simpler arguments before more subtle arguments are addressed. At any rate, there is ample philosophical material to think about in the arguments I'll describe, even if these arguments don't capture the thoughts of various great thinkers with total accuracy and completeness.
The First Two Arguments: Motion and Causality
Aquinas's first argu1nent for the existence of God is the argurnent fron1 motion. Here it is, formulated as a deductive argument:
(I) (1) In the natural world, there are objects that are in motion.
(2) In the natural world, objects that are in motion are always caused to move by objects other than themselves.
Chapter 4: Aquinas's First Four \Y/ays
'' '''' '., '''' "'. '.',.'' .. '.'' '.'. '' "'"'' '' .. ' .. ' ' ' ' '.' .. ,''' .. ''. (3) In the natural world, causes must precede their effects.
(4) In the natural world, there are no infinite cause/effect chains.
(5) Hence there is an entity outside of the natural world (a supernatural be- ing), which causes the motion of the first moving object that exists in the natural world.
( 6) Hence, God exists.
Aquinas's second argu1nent generalizes the ideas found in the first. Whereas the first argun1ent is about 1notion in particular> the second argument is about causality in general:
(II) (1) The natural world includes events.
(2) In the natural \Vorld, every event has a cause, and no event causes itself.
(3) In the natural world, causes must precede their effects.
(4) In the natural world, there are no infinite cause/effect chains.
(5) Hence there is an entity outside of nature (a supernatural being), which causes the first event that occurs in the natural \Vorld.
( 6) Hence, God exists.
In both these arguments, I've drawn two horizontal lines to indicate that (5) is supposed to follow from premises (1)-(4) and that (6) is supposed to follow from (5).
What do premises (2) and (3) mean in these two arguments? Lefs begin with an object (01) that is in inotion no\v. Since it is in 1notion, there 1nust be an earlier object-call it 0 2-that sent 0 1 into motion. If 0 2 \Vas itself in motion, 0 2 1nust trace back to a previous n1over, 03' and so on. In the second argument> the subject is causality, not motion, but the idea is basically the same.
Premise (4) says that there can't be a cause/effect chain that extends infinitely far back into the past. The idea is that cause/effect chains (or mover/movee chains) leading from the present back into the past have a finite number of links. The arrow in the following diagram represents the relationship of causality:
Past Present Future
Or~···~ 03 ~ 02 ~ 0 1 ~
Although Aquinas's argument focuses on chains that extend from the present back into the past, his principle (4) also has implications about chains that extend from the present forward into the future. These also must have a finite number oflinks. Cause/effect chains must be finite in both directions.
I'll begin with a criticism aimed just at argument I; after that, I'll lump the two argu- ments together and formulate some objections that apply to both.
45
Part Tw·o: Philosophy of Religion
'' ''.' ''' ' ...... ' ... '. '' .. '. '''. ' ... '' '' .. '.'' ''' '' ''' '''' '' '' ''' '' ''''. Aquinas on the Cause of Motion
Aquinas thinks (premises 1-2) that if an object is in motion, it must be caused to move by something outside itself. Aquinas got this idea from Aristotle's physics. Aristotle {384-322 B.c.) held that if an object continues to move, its motion must be sustained by a force that keeps it in motion. If you remove the force, the object stops moving.
This idea didn't survive into modern physics. You may remember from your high school physics course that Newton, in the seventeenth century, held that an object remains in constant uniform motion unless acted on by a force. Recall that one of Newton's laws of motion is F = ma. This means that if an object of mass 111 is acted on by a force of value F, then it will accelerate to degree a. This Newtonian law says that an object that isn't acted on by a force \Von't accelerate, \vhich means it \Vill ren1ain at rest or in unifor111 1notion. Ne,vton's la\vs say that an object can re1nain in nlotion forever \vithout there being any force that sustains its n1otion.
Newton's laws do not exclude the following possibility: The universe contains exactly one physical object, which always moves in uniform motion without any forces ever acting on it. Of course, Newton's laws don't say that the universe we live in is like this. Our uni- verse obviously contains more than one material object. My point, though, is that Newton's theory of motion was different from Aristotle's and so was different from Aquinas's. Aristotle and Aquinas thought that motion requires an outside force; Newton and more modern physical theories hold that it is acceleration, not simply change in position, that requires a force.
It is not difficult to rescue Aquinas•s first argu1nent fro1n this Ne,vtonian objection. Just replace his talk of motion with the concept of acceleration. If objects accelerate, there must be a force that causes them to do so. Then Aquinas' line of reasoning will lead us to the conclusion that there must exist a supernatural entity that causes the first accelerating object in nature to accelerate.
God Is a Person, Not just a Cause That Exists Outside of Nature
I now turn to some problems that the two arguments share. First, it is important to see that proposition (5)-that there is an entity outside of nature that causes the first moving object in nature to n1ove, or that causes the first event in nature to occur-does not guarantee the existence of God, where God is understood to mean a person with something like the three properties of omnipotence, omniscience, and omnibenevolence (an all-PKG being). As Aquinas himself realized, conclusion ( 6) does not follow from proposition (5), in either argutnent.
The Birthday Fallacy
Another problem arises when we ask whether the argument shows that there is precisely one first cause, or instead shows only that there is at least one. Suppose we grant that each causal chain in nature has a first member. According to Aquinas, each of these first
Chapter 4: Aquinas's First Four \Y/ays 47
" .. '' '''. '''.''.''' '' ' .. ' ... '' ''.'' '.' "' ' ... ' "'' ''' '' ...... ', ''.'' '.'''' '. 1nembers 1nust be caused by son1e event outside of nature. Ho,vever, it does not follo\v that there is exactly one such event outside of nature that set all causal chains in the natu- ral world in motion. Here it is important to see the difference between the following two propositions; the first is different from and does not deductively imply the second:
Every event in the natural \Vorld traces back to an event that occurs outside nature.
There is a single event outside of the natural world to which each event in nature traces back.
The difference here parallels the logical difference between the following two proposi- tions, the first of which is true and the second false:
Every person has a birthday-a day on which he or she was born.
There is a single day that is everybody's birthday.
I want to give a name to the mistaken idea that the second proposition follows from the first. I'll call this mistake the Birthday Fallacy.
So one problem with arguments I and II is that they don't show that there is exactly one first cause or unmoved mover; at best, they show that there is at least one. To think otherwise is to commit the Birthday Fallacy.
Why Can't Nature Be Infinitely Old?
Another objection to Aquinas's first two arguments is his claim that cause/effect chains cannot extend infinitely far into the past. Why is this impossible? If the natural world were infinitely old, each event could be caused by an earlier event. Every event that occurs in nature could have a cause that also existed in nature, so there 'vould be no reason to infer that something outside of nature must exist as the cause of \Vhat occurs inside.
Aquinas thinks he has an answer to this question. He doesn't simply assume that causal chains extending into the past must be finite in length; he has an argument that he thinks sho\vs why this is so. Here is his argu1nent, \vhich I'll reconstruct in ter1ns of an example of a present event-you are reading this page no\v:
You are reading this page no\V.
A causal chain that extends from this present event infinitely into the past, by definition, lacks a first member.
If a causal chain lacked a first member, then all subsequent events in the chain could not occur.
Hence, the causal chain leading up to your now reading this page must be finitely old.
The third premise is \Vhere this argu1nent goes 'vrong. Even if \Ve assu1ne that no event in nature can happen without its having a cause, it does not follow that there has to be a first natural event.
Part Two: Philosophy of Religion
Many traditional theists will agree that the world could have an infinite future-that it could go on forever. Ho\\•ever, if an infinite future is possible, \vhy is an infinite past ruled out? If there doesn't have to be a last event in the history of the natural world, why must there be a first? VVhy accept Aquinas's clahn that causal chains can't extend infinitely into the past? Here Aquinas gets son1e help fron1 n1odern physics, \Vhich vie"rs the universe as finitely old. Although it is conceivable that the universe is infinitely old, apparently there are scientific reasons to think that this isn't so.
Why Must Every Event in Nature Have a Cause?
This brings rne to n1y last objection to Aquinas's first hvo arguments for the existence of God. Even if the natural world is only finitely old, why must there be an explanation of the first event that occurs in nature? 'fhat is, ,vhy inust every event that occurs in nature have a cause?
Do scientists assu111e that every event has a cause? Well, it is true that scientists often try to discover the causes of events that they observe. Ho\vever, this activity of searching does not require that one actually believe that every event has a cause. Perhaps there are exceptions to this generalization. Scientists try to find causal ex- planations for the events they observe for a very simple reason: If you don't look, you won't discover the cause if there is one. Better to look and fail to find than never to look at all.
The Third Argument: Contingency
Aquinas's third argun1ent for the existence of God, like the first hvo, begins \Vith an obser- vation that everybody would agree is true. The observation is put in language that may be unfamiliar, but once explained, it see1ns clear enough. 'fhe observation is that contingent things exist. What does it mean to say that a thing is contingent? The opposite of contin- gency is necessity. VVhat inakes an object contingent or necessary?
Necessary and Contingent Beings
You, I, the Washington Monu1nent, the htunan race, the earth, the solar systen1, and the Milky Way are all examples of contingent things. Although all of us exist, we needn't have existed. The world could have failed to include any of us. If the particular sperm and egg that produced you hadn't encountered each other, you \Votddn't have coine into existence. Your parents would have had other children or no children at all. Likewise for the human race; although the world obviously includes human beings, it needn't have. Contingent things depend for their existence on so1nething or other happening. In contrast, a neces- sary being is something that must exist no matter what. It doesn't depend for its existence on an)1hing.
Chapter 4: Aquinas's First Four \Y/ays
Possible Worlds
I 'vant to represent the concepts of necessity and contingency by introducing a ne'v idea. Consider the totality of things that have existed, now exist, or will exist at any place in the universe. This totality comprises a giant object that I'll call "the actual world:' Imagine a census of the objects in the actual world. We might display this census on a very long time line} in \Vhich the durations of various objects are represented:
/-w-/
/-.\~;
Past
/-)~!
No\v
/-;1.;-/
Future
Notice that the life spans, or durations, of objects wand x overlap. Objects wand x existed in the past and no'v are no more. Object y came into existence in the past, exists no\v, and will continue to exist until some future date, when it will cease to be. Object z has1it yet come into being, but it will. Of course, this is a dreadfully incomplete inventory of every- thing that was, is, or will be. Never mind. You get the idea: The actual world consists of everything that has existed, exists no\v, or \Vill exist, anY'vhere in the universe.
We know the world needtit have had precisely the census it does. Some of the objects in this census might have failed to exist. Other objects, which don't actually exist, might have done so. Let's call each possible way the world might have' been a "possible world:' TI1ere are many of these.
We may now say that an object is contingent if it exists in some but not all possible worlds. And an object is necessary if it exists in all possible worlds. Suppose the following are the censuses found in several possible worlds. As before, the horizontal line represents time within a possible world:
11JX y z Actual world ab )' z Possible world I ac d z Possible world 2
Past Present Future
Notice that w exists in the actual world, but not in the two alternative possible worlds represented. On the other hand, y exists in the actual world and in the first possible world, but not in the second. Object z is different-it exists in all the possible worlds depicted. Of course, there are inore than three possible \Vorlds1 as there are inore than hvo alternative ways the actual world might have been (the actual world, of course, is one of the possible worlds). Would z continue to appear if we listed not just a few possible worlds, but all of them? That is the question of whether z is a necessary being.
I cited some examples of contingent things-you, the Washington Monument, the earth, and a few others. These examples might strike you as typical of everything that ex- ists. That is, you might find plausible the following philosophical conjecture: Everything that exists is contingent. Familiar objects clearly have this property. But is everything con- tingent? Aquinas argues in his third proof of the existence of God that not everything is contingent. There is at least one necessary being-narnely, God.
49
50 Part Two: Philosophy of Religion
''. '''.''' ' ••• < •• ' •• ' ' '. '.' '' •••••• ' •• '''' '.'.'.' ' •• < •• ' •• ' ••• ' •• '.' ••• '
Here is how his proof goes:
(III) (1) Contingent things now exist.
(2) Each contingent thing has a time at which it fails to exist (contingent things are not omnipresent).
(3) So if everything were contingent, there would be a time at which nothing exists (call this an "empty time").
(4) That empty time would have been in the past.
(5) If the world were empty at one time, it would be empty forever after (a conservation principle).
( 6) So if everything were contingent, nothing would exist now.
(7) But clearly, the world is now not empty (see premise 1).
(8) So there exists a being who is not contingent.
Hence, God exists.
Aquinas's argument has two stages. First, there is his defense of proposition ( 6), which is contained in premises (1)-(5). Second, there is his use of proposition (6) to establish the existence of God. Before consideriµg why Aquinas thinks ( 6) is correct, I want to focus on his use of (6) and (7) to infer proposition (8).
Reductio Ad Absurdum
Aquinas's proofof (8) from ( 6) and (7) has a distinctive logical form. His proof is a reductio ad absurdum argument (a "reductio:· for short).
When I talked about deductive validity in Chapter 2, I emphasized that the word "valid" in logic and philosophy doesn't mean what it means in everyday life. The same holds for the idea of reducing something to absurdity. In ordinary speech, this means something like "making a tnockery of an idea.,, In logic, ho'ivever, reductio argu1nents are perfectly good arguments-they are deductively valid.
Here's ho'iv a reductio argument \Vorks. You \Vant to prove a proposition P. To do this, you argue that if P were false, some proposition A would have to be true. But you construct the argument so that A is obviously false (it's an "absurdity"). From this, you may validly conclude that Pis true. Reductio arguments, in other words, have the following valid logi- cal form:
If Pis false, then A is true.
A is false.
Pis true.
Chapter 4: Aquinas's First Four \Y/ays
Aquinas tries to establish the existence of a necessary being by reductio. He argues that if only contingent beings existed, the world would now be empty. But it is obviously false that the 'vorld is no'iv en1pty. Hence, not everything is contingent~there 1nust exist at least one necessary being.
Let's now examine Aquinas's defense of proposition ( 6). Why should we think that the world would now be empty if there were no necessary beings? Aquinas's reasons for think- ing this are contained in premises (1)-(5).
Contingency and Eternity
I'll begin by registering an objection to premise (2). Contrary to Aquinas, I submit that a contingent thing can be eternal; the fact that an object is contingent doesn't mean that there must be a time at which it fails to exist. I'll grant that familiar contingent objects aren't eternal. You are an exa1nple. You might have failed to exist; in addition, there 'ivas a time before you were born and there will be a time after you die. But this is just an ex- ample. We want to know whether all contingent beings must fail to be eternal.
Notice first that the considerations affecting contingency differ from those affecting the question of eternity. How can you tell whether an object is contingent in a diagram like the one given before showing the actual world and two possible worlds? You look down the list of possible worlds, checking to see if the object in question is present in each one. In that diagram, we might say that contingency is represented "vertically:' Eternality is differ- ent. To see whether an object present in a given possible world is eternal, you look across the representation of that possible world, checking to see if it exists at all times in that world. So eternality is represented "horizontally:'
Could an object exist in only some possible worlds and still exist at all times in the actual world? That is, could an object be both contingent and eternal? Aquinas says no. Here is a consideration that suggests he may be wrong.
The idea that material objects are made of very small indivisible particles has had a long history, going back at least to the ancient Greeks. Of course, we no longer think of at- oms as indivisible particles-we have long since learned to talk about subatomic particles. However, suppose for the moment that the word "ato1n'' names such indivisible particles. These particles are the basic building blocks of all material things.
One idea that has been put forward in atomistic theories is that atoms (fundamental particles) can't be created or destroyed. Large objects made of atoms can be created and destroyed by assembling 'and disassembling collections of atoms. But the basic particles themselves can't be destroyed, because you can't break them into pieces. Again, I'm not saying that this is true of the objects we now regard as the smallest material particles. I mention it only to describe a possible view of atoms.
Ato1ns, on the vie'v I'tn describing, are eternal. If a given atom exists no'iv, it has ahvays existed, and it will always exist. Let's call one of the atoms that populate the actual world by the name "Charlie:' Charlie is eternal. Does it follow that Charlie is a necessary being- that the world couldn't have failed to include him? I would say not. The world could have contained more ato1ns than it does, or fe,ver. Indeed, it doesn't seem impossible that the world might have been entirely empty of matter. So Charlie, like all the atoms that happen
52 Part Two: Philosophy of Religion
to actually exist, is a contingent being. If this is right, then Aquinas's premise (2) is mis- taken. Charlie, I've claimed, is a contingent being, but he is eternal. So contingent things needn't have a time at which they fail to exist.
The argument I've just given against premise (2) depends on a particular theory about the nature of atoms. If atoms really are indestructible, I can claim to have refuted Aquinas's premise. But are they, in fact, indestructible? Again, remember I'm discussing the smallest units of n1atter here. This is ho\v I'tn using the \Vord "aton1:'
Conservation Laws in Physics
Current physics doesn't support the view that atoms are indestructible. Before Einstein, phys- ics upheld a principle that said that the quantity of mass in the universe is constant. Einstein replaced this classical conservation la\v \Vith another one. Mass isn't conserved; rather, a quantity that Einstein called "mass-energy" is conserved. This principle says that mass can be destroyed; it may be converted into energy. Modern physics, therefore, doesn't seem to allow me to tell my story about Charlie, the indestructible atom. So matter can be destroyed, accord- ing to current theory. But this doesn't mean that every contingently existing particle WILL be destroyed, sooner or later. It is the latter claim that Aquinas advances in premise (2). I don't know of any good reason to think this is so, although I'll leave the question open.
My conclusion thus far is that pre1nise (2) is inistaken; a contingent thing doesn't have to have a tin1e at \Vhich it fails to exist. Let's 1nove on.
The Birthday Fallacy (Again)
The transition from premise (2) to statement (3) involves a fallacy. Even if we grant that every contingent entity has a ti1ne at \vhich it fails to exist, \Ve can't conclude that there is an en1pty ti1ne. The fallacy here is the sa1ne one discussed concerning Aquinas's first hvo arguments. Recall the point about birthdays; "Everybody has a birthday" does not deduc- tively imply that there is a day that is everyone's birthday.
To make this point graphic, consider the following possible world, in which no object is eternal, and yet there is no empty time:
<--- /-1-/ /-3-/ /-5-/ ----> /-2-/ /-4-/ /-6-/
Past Present Future
The arrows at the beginning and end indicate that the pattern should be repeated indefi- nitely into the past and into the future. So the supposition that everything is contingent doesn't imply an empty time. But there is an additional unjustified step. Even if there does have to be an e1npty titne, \vhy inust it have been in the past? I see no reason to assu1ne this.
Necessary Beings other than God
I've noted some problems in Aquinas's attempted reductio proof that not everything is contingent. Suppose, however, that Aquinas could show that there is at least one necessary being. Could he conclude from this that God exists? I want to argue that this, too, doesn't
Chapter 4: Aquinas's First Four \Vays
follo,v. In discussing Aquinas's first t'\VO argu1nents, I claitned that the existence of God doesn't follo'v fro1n the existence of a first cause. I no'\\' '\Vant to suggest that the existence of God doesn't follow from the existence of a necessary being.
To explain why this is so, I need to develop an idea from the philosophy of mathemat- ics. Until no\v I've described necessity and contingency as properties of entities. You and I are contingent beings, for exa1nple. No'\v, ho'\\'ever, I '\Vant to discuss the idea that contin- gency and necessity are properties of propositions. Propositions, or state1nents, are true or false. They also have the properties of necessity and contingency.
Necessary and Contingent Propositions
Necessity and contingency as properties of propositions are definable in a \vay that paral- lels the way this distinction applies to entities like you and me. A proposition is necessarily true if it is true in all possible worlds. A proposition is necessarily false if it is false in all possible worlds. Contingent propositions are true in some possible worlds and false in others. A true statement is contingent if it is true in the actual '\vorld, though false in so1ne possible world.
It is easy to cite examples of contingent propositions. Consider the fact that the United States is a country in North America. This is true, though only contingently so. Consider the following contingent falsehood: the Rocky Mountains are no higher than six thousand feet above sea level. This is false, but there are possible worlds in which it is true.
Mathematical Truths
Are there any necessary truths? Many philosophers have held that mathematical truths are necessarily true. Consider the fact that 7 + 3 = 10. It iStrt just that this happens to be true in the actual world. There is no possible world in which it is false. The sum of 7 and 3 could not fail to equal 10.
To see why this is at least a plausible view of mathematical propositions, you've got to be clear on what the proposition says. The proposition doesn't say that if you put seven rabbits together with three others, there will be ten bunnies forever after. Rabbits repro- duce, but this doesn't contradict any fact of arithmetic. Take another example: If you pour three cups of sugar into seven cups of water, you will not obtain ten cups of liquid. But, again, this does1rt contradict the fact that 7 + 3 = 10. Arithmetic is unaffected by this result; it just so happens that liquids and solids don't always combine additively.
Names Differ from the Things Named
Here is a somewhat more subtle point to notice. The proposition that 7 + 3 = 10 doesn't say anything about the language we use in expressing that fact to ourselves. It is a contin- gent fact that human beings use the numeral "i' to refer to the number 7. We might have called that number by another name-'4/e 1night have used the nun1eral "2» to denote the number 7. Similarly, it is a contingent fact that we use the symbol"=" to represent the rela- tion of equality. So it is a contingent fact that the sentence "7 + 3 = 10" expresses what it does, and it is a contingent fact that the sentence expresses a true proposition rather than a
53
54 Part Two: Philosophy of Religion
false one. But this doesn't show that 7 + 3 might not have equaled 10. The fact could1it have been otherwise, even though we might have used our terminology differently.
The distinction I'1n dra,ving here is very obvious in so1ne contexts, though it is easy to lose sight of it in others. Few things are more different than a thing and its name. You are a person, but your name isn't. Your name contains a certain nun1ber of letters, but you don't. Your na1ne is a linguistic entity, but you aren't.
We mark this difference by using quotation marks. If we want to talk about a thing, we use its na1ne. If we 'vant to talk about the name of a thing, '"e put the na1ne in quotation marks. So if I want to talk about a big mountain in western North America, I might say that the Rocky Mountains are tall. But if! want to talk about the mountain range's name, I'd say that the phrase "the Rocky Mountains" contains seventeen letters.
Numbers Aren't Numerals
This is obvious \vhen you think about it. When it co1nes to mathen1atics, ho,vever, \Ve tend to confuse these ideas. Numerals are names of numbers. But the number and its name are different, just as the phrase "the Rocky Mountains" differs from the mountains themselves. I grant that it is a contingent matter that the sentence "7+3 = 10" expresses a truth. I deny, however, that 7 + 3 might have failed to equal 10. The proposition is necessarily true, even though it is a contingent fact that the sentence expresses the proposition it does. Just as your name is part of a language, though you aren't, so the sentence "7 + 3 = 10" is part of a language, though the proposition it expresses is not.
Sets
So what does the proposition that 7 + 3 = 10 actually mean? It doesn't say anything about rabbit reproduction or about the way sugar and water combine when poured together. Nor does it say anything about the language we use. You can think of this statement as describ- ing a basic property of mathematical objects called sets.
First, I need a bit of terminology. The union of two sets is the set that includes all the members of the first and all the members of the second, and no others. A set is any collec- tion of objects. So the set {Groucho Marx, Napoleon, the Eiffel Tower} is a three-member set, and the set {Napoleon, the Rocky Mountains, the French Revolution} is also a three- member set. The items in a set needn't have anything special in common with each other. There are sets of similar things, but there are also sets of dissimilar things. The union of the two sets just mentioned is this set: {Groucho Marx, Napoleon, the Eiffel Tower, the Rocky Mountains, the French Revolution}.
The proposition that 7 + 3 = 10 can be understood to say the following: Consider any two sets where the first contains exactly seven objects and the second contains exactly three. If these sets have no common members, then the union of these two sets will have precisely ten members. Notice that the arithmetic proposition doesn't say that the Rocky Mountains exist, or that there are at least seven Jnaterial things in the universe. The arith- metic fact does1it rule out the possibility that the world contains no material objects at all. It simply describes a basic property of the operation of set-theoretic union.
Chapter 4: Aquinas's First Four \Vays 55
'''.'.' '''. "'.' '''.'''' '''.' ' .. '' '. '. ''.''.' '' .. '' .. ' '.'' '' .. ''''' '.'.''' Necessity and Certainty Are Different
I need to add a final clarification of the thesis that mathematical truths are necessary. In saying that it is a necessary truth that 7 + 3 = 10, I'm not saying that I know that the propo- sition is true with absolute certainty. Nor am I saying that I'll never change my mind on the question of whether 7 + 3 = 10.
This is a rather subtle point, because 've so1neti1nes express our lack of certainty about a proposition by saying that maybe the proposition is1it so. If I think there will be rain tomorrow, but entertain some small doubt that this will be so, I may express this by say- ing, "Possibly it won't rain tomorrow:' But my certainty or uncertainty is a fact about me. It is a fact in the subjective realm. I want to distinguish this question from the question of whether a given proposition has the property of necessity or contingency. The latter ques- tion does1it have anything special to do with me. Necessity and contingency are objective.
Here is a nonmathematical example that may make this clearer. Could there be per- petual motion machines? A machine of this sort wouldn't require any energy input to keep running, but it would provide a constant output of energy. Scientists for many hundreds of years tried to design such a machine. They always failed. Finally, in the nineteenth cen- tury, physicists working in the area called thermodynamics proved that perpetual motion n1achines are hnpossible.
Consider the proposition that there are or will be perpetual motion machines. This is, we now believe, a falsehood. Is it a necessary falsehood? I would say yes. It isn't just that no one will bother to build one; the point is that it is impossible to build one. This is what science tells us.
Now I'll ask an entirely separate question: Am I absolutely certain that no such ma- chine will ever exist? I guess I'm not absolutely certain. After all, science has been wrong before, and so maybe it now is wrong when it says that such machines are impossible. Opinion has changed through history on the question of whether such machines could be built. Before the nineteenth century, many serious scientists thought such machines are possible. Later, opinion changed. Maybe it will change again.
So the certainty that a single scientist or a community of scientists may have about the issue may change. But there is something that doesn't change. Either it is possible for such things to exist or it is1it. This should convince you that certainty and necessity are different. People may change their degree of certainty about a proposition; they may even think at one time that the proposition is true, but later on think that it is false. However, the propo- sition itself doesn't change from true to false. Nor does a proposition cease to be necessary just because people stop believing it. So, to say that arithmetic truths are necessary isn't to say that people are certain about arithmetic. Nor is it to say that people have never changed their minds about arithmetic propositions. Again, these are questions about our attitudes toward the propositions. But whether a proposition is necessary or not has nothing to do with our attitudes. In this respect, necessity is like truth-both are objective issues, inde- pendent of what people happen to believe.
To sum up, I've described a prominent view in the philosophy of mathematics. It holds that arithmetic truths are necessary. I haveiit provided that view with a complete defense, but I hope you can see what it asserts. I also hope you find it at least somewhat plausible, even if it is not entirely convincing.
Part Two: Philosophy of Religion
What has this material about mathematics to do with the conclusion of Aquinas's third proof? Recall that Aquinas reasons that if there is a necessary being, then God must exist. Part of my point in talking about mathematical necessity here is that I want to describe a necessary being that no one would think is God.
Numbers Are Necessary Beings
Consider the fact that arith1netic includes various existence staternents. Besides asserting that 7 + 3 = 10, arithn1etic also asserts that there is a pritne nun1ber in1n1ediately after 10. Furthermore, if arithmetic truths are necessary, then it is a necessary truth that there exists a prime number immediately after 10. To put it bluntly, the philosophy of mathematics I've described holds that the number n is a necessary being. It exists in all possible worlds. This follows from the thesis that arithmetic truths are necessary and from the fact that arithme- tic contains existence clailns.
The conclusion that there is a God doesn't follow from the assertion that not every- thing is contingent. A philosopher of mathematics might claim that the number 11 is a nec- essary being, but this wouldn't entail that God exists. This is my last criticism of Aquinas's third argument.
Aquinas's Fourth Argument: Properties That Come in Degrees
I 'von't spend much tilne on Aquinas,s fourth proof of the existence of God. It is rooted in an Aristotelian view of causality that seems radically implausible now. Here's the argument:
(IV) (1) Objects have properties to greater or lesser extents.
(2) If an object has a property to a lesser extent, then there exists some other object that has that property to the maximum possible degree (call this a maximum exemplar of the property).
(3) So, there is an entity that has all properties to the maximum possible degree.
Hence, God exists.
The argument begins with an observation that is obviously true: Some things are more po\verful and so1ne less; so1ne are more intelligent and others less, and so on. The second premise, however, seems entirely implausible. The fact that Charlie Chaplin is less than maximally funny doesn't mean that there must exist a maximally funny comedian. And the fact that we are somewhat intelligent, though not perfectly so, doesn't seem to require that there exists a perfectly intelligent being.
The Aristotelian idea that Aquinas is using here is roughly as follows. Aristotle thought that fire is the maximally hot substance. When other objects are hot to some lesser degree, this is because fire is mixed in them to some extent. The property of heat can occur to a less than maxitnal degree in hu1nan beings, for exa1nple, only because there exists this sub- stance, fire, \Vhich is hot to the 1naxin1um possible degree. Fire is the 1naximun1 exemplar of heat, from which lesser degrees of heat derive.
There are other problems with the argument. If each property has a maximum exem- plar, it doesn't follow that there is an entity that is a maximum exemplar of all properties.
Chapter 4: Aquinas's First Four \Vays
Recall the Birthday Fallacy: The fact that everyone has a birthday doesn't imply that there is a single day on \Vhich everyone \Vas born. So even if there is a maximum exemplar of in- telligence, a n1axin1u1n exemplar of po\ver, and a 1naxitnu111 exen1plar of 1noral goodness, it doesn't follow that there is a single entity who is all-knowing, all-powerful, and all-good.
Finally, there is the problem of contradiction. If intelligence has its maximum exemplar, then stupidity \Vould have to have its n1axin1u1n exe1nplar as 'veil. By Aquinas's argument, this leads us to say that there is a single being who is both maximally intelligent and maxi- mally stupid. But this is impossible. That concludes what I want to say about Aquinas's fourth argun1ent for the existence of God.
Criticizing an Argument versus Showing the Argument's Conclusion Is False
None of the four arguments I've discussed here is successful. Does this 1nean there is no God? It means no such thing. There may be other arguments for the existence of God that are convincing. That the four argtunents discussed here don't \Vork doesn't 1nean that no argument will work. Not one word has been said here that shows that atheism is true. All \Ve have seen is that so1ne argun1ents for theistn are fla,ved. In the next chapter, I'll con- sider Aquinas's fifth argument for the existence of God. Maybe it will fare better.
Review Questions
1. \Y/hat objections are there to the first cause argument?
2. \Y/hat is the Birthday Fallacy? How does it figure in the discussion of Aquinas's arguments?
J. Explain what it means for an object to be necessary or contingent. What is a "possible world"?
4. How are necessity and eternality related? How does this bear on Aquinas's third argument?
5. \Y/hat is a reductio argument? Give an example.
6. What is the difference between necessity and certainty? What is meant by saying that necessity is "objective"?
7. What would it mean for something to be a first cause \Vithout being God? What would it mean for something to necessarily exist \Vithout being God?
Problems for Further Thought
1. I formulated Aquinas's proofs by having him talk about objects that exist in "nature" (in "the natural world"). \Y/hat does "nature" include? Does it include just the things we can see or hear or touch or taste or smell? After all, the selection from Aquinas in the Readings translates Aquinas as talking about the "sensible" world. How would interpreting his argument in this way affect its plausibility? What other interpretations of "nature" make sense in this context?
2. In discussing Aquinas's third proof, I talked about Charlie the atom as an example of a thing that is both eternal and contingent. Could something exist that is both necessary and noneternal? It would exist at some time in each possible world, though it would not exist at all times in the actual \Vorld. Can you give an example of such a thing?
). I criticized Aquinas's third argument by discussing numbers, which I claimed exist necessarily. Can the argument be reformulated so that this objection no longer applies?
4. I criticized Aquinas's fourth argument by discussing "maximum stupidity." Can Aquinas reply to this objection by claiming that stupidity is just the absence of intelligence?
57
58 Part Two: Philosophy of Religion
MySearchLab Connections
Watch. Listen. Explore. Read. Visit MySearchLab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
CEJ-( READ AND REVIEW
READ
l. Thomas Aquinas, "Five Ways to Prove That God Exists," Pt. 1, Question 2, Art. 3 of Summa Theofogiae
While reading the selection, refer back to the arguments given in the chapter and compare them to the way in which they are presented in the text. Does Aquinas formulate any of his argutnents in a significantly different form than I have reconstructed them in the chapter?
RESEARCH AND EXPLORE
:j(l{EXPLORE
i. Thomas Aquinas 2. Arguments for the existence of God
chapter 5 ...............................
The Design Argument
Chapter Outline
Goal-Directed Systems
Two Kinds of Design Argument
Paley's Watch
The Analogy
Abductive Arguments Often Postulate Unobserved Entities
Hun1e)s Criticis111s of the Design Argument
Is the Design Argument a Weak Argument from Analogy?
Is the Design Argument a Weak Induction?
There are three main traditional arguments for the existence of God-the cosmological argmnent, the design argument, and the ontological argument. Aquinas's first, second, and third ways, surveyed in the previous chapter, are cosmological arguments. Cosmological arguments take different forms; each cites a general feature of the whole universe as evidence that there is a God. The second type of traditional argument, the design argument, is the one we'll consider in the present chapter. The ontological argument will occupy our attention in Chapter 8.
Aquinas's fifth argument for the existence of God is an instance of what has come to be called the Argument from Design. The design argument has a variety of forms, some of which I'll describe. To start things off, here is a formulation that is close to the one Aquinas uses:
(1) Among objects that act for an end, some have minds whereas others do not.
(2) An object that acts for an end, but does not itself have a mind, must have been designed by a being that has a mind.
(3) Hence, there exists a being with a mind who designed all mindless objects that act for an end.
Hence, God exists. 59
60 Part Two: Philosophy of Religion
Note as a preliminary point that the transition from (2) to (3) commits the Birthday Fallacy described in Chapter 4. If each mindless object that acts for an end has a de- signer, it doesn't follow that there is a single designer of all the mindless objects that act for an end.
Goal-Directed Systems
What does Aquinas mean by ('act for an end"? This phrase corresponds to the tnodern idea of a goal-directed system. Human beings act for an end because they have desires; these desires represent the ends or purposes or goals to \Vhich behaviors are directed. Hu1nan beings are capable of goal-directed behavior because they have minds. Conside1; however, a different example: a guided missile. It is a goal-directed system. Its goal or function is to reach and destroy its target. If the target veers off to the side, the missile can adjust its behavior so that it will achieve its purpose. Guided missiles are goal-directed systems, but they do11t have minds. How is this possible? The answer is consistent with what Aquinas says in premise (2). Guided missiles are artifacts. They are devices built by creatures with minds-namely, human beings. This is how missiles obtained the machinery that allows them to engage in goal-directed behavior.
Are there other examples of goal-directed systems besides human beings and artifacts? Nonhuman organisms provide a third category. Even bacteria, which evidently d011t have beliefs and desires, seek out nutrients and avoid poisonous chen1icals. Because of this, it seems plausible to describe them as having the goal of surviving and reproducing. They are able to modify their behavior to achieve these ends.
Does the list stop \Vith human beings, artifacts, and nonhu1nan organis111s? Aquinas followed Aristotle in thinking that even inanimate objects such as rocks and comets have goals. This idea went out of fashion with the Scientific Revolution in the seventeenth century. It UO\Y see1ns itnplausible to describe a rock as being hard "in order to resist de- struction:' It also seems strange to say that rocks fall toward the center of the Earth when they are released "in order to attain the location that it is in their nature to seek:' But this is how Aristotle thought about rocks, and Aquinas followed him here. Both thought that everything, whether living or not, should be understood teleologically-that is, as a goal- directed system. (This will be discussed further in Chapter 34.) I won't take issue with this general teleological picture, except to note that it is far more encompassing than the one provided by modern science. However, this point does not affect the design argu- ment as I have formulated it. What is required is just that some mindless objects are goal directed.
Two Kinds of Design Argument
It will be useful to distinguish two kinds of design argument. Aquinas would have been willing to endorse them both. David Hume (1711-1776), who examined various design arguments in his Dialogues Concerning Natural Religion (1779), discusses both sorts. I'll call these two sorts of arguments global and local design arguments.
A global design argument cites some general features of the whole universe and argues that this feature should be explained by the hypothesis that it is the product of intelligent
Chapter 5: The Design Argurnent
''' '' '',,''' '''' ', '''.' ''''. '',,' '''' ''.".' ''.'.
design. An example would be the argument that proposes to explain why the laws of nature are simple. Newton himself argued that the simplicity of natural laws is evidence that there exists an intelligent and perfect God who was their author.
A local design argu1nent focuses on a 1nore specific feature that one or more object in the universe has and claims that the hypothesis that God exists is the best or the only plausible explanation of that fact. The example I'll consider here concerns features of the organis1ns 've observe on Earth. They are goal-directed syste1ns; they are con1plex systen1s equipped with the ability to modify their behavior so that they can survive and reproduce. In this chapter, I'll focus on the local argument, which claims that special features of living things (including ourselves) are said to require explanation. In Chapter 7, I'll return to design arguments that are global.
Paley's Watch
During the eighteenth century and the early nineteenth century in Great Britain, design ar- guments were the rage. Numerous books were published arguing that the existence of God was required to explain this or that feature of the world we observe. One of the most influ- ential works of this sort was produced by William Paley (1743-1805). Whereas Aquinas for- mulated his version of the design argument as a deductively valid argument, I'm going to interpret Paley's argu1nent as an abductive argument, an inference to the best explanation.
Paley's striking formulation of the design argument goes like this: Suppose you are walking on a beach (actually, Paley talks about walking on a "heath") and find a watch lying on the sand. Opening it up, you see it is a complex and intricate piece of machinery. You also see that the parts of the watch work together to allow the hands to measure out equal intervals of tin1e 'vith considerable precision. What could explain the existence and characteristics of this object?
One hypothesis to consider I will call the Random Hypothesis. By the random action of the waves on the sand, a watch was accidentally produced. Even if you think that this explanation is possible, I bet you agree that it isn't very plausible. The idea that waves beat- ing on sand could produce a useful object of such intricacy doesn't make a lot of sense. It is about as plausible as suggesting that a monkey randomly pounding on a typewriter will write out the complete works of Shakespeare. A far more plausible explanation is the Design Hypothesis. The intricacy and usefulness of the watch suggest that it is the product of intelligence. This hypothesis says that the watch exists because there was a watchmaker who produced it.
Why do we think the Design Hypothesis is more plausible than the Random Hy- pothesis? If there were a designer at work, then it wouldn't be surprising that the watch is complex and well suited to the task of measuring temporal intervals. If, however, the only process at 'vork 'vere \vaves pounding on sand, then it 'vould be enormously surprising that the watch has these characteristics. The observed features of the watch are possible according to each hypothesis. But they are highly probable according to one and vastly improbable according to the other. In preferring the Design Hypothesis, we prefer the hypothesis that strains our credulity less.
I hope you see that Paley's argument uses the Surprise Principle described in Chapter 3. You've made some observations (call them 0) and are considering whether 0 strongly
CEJ-[ Read "The Design
Argument,~ Chapter i of Natural Theology on \'NA'l.mysearchlab .com
62 Part Two: Philosophy of Religion
favors one hypothesis (H,) over another (H,). The Surprise Principle says that 0 strongly favors H, over H, if H, says that 0 is very probable while H, says that 0 is quite improbable. 0 would be unsurprising if H, were true, but 0 would be very surprising if H, \vere true.
Not only do we infer the existence of a watchmaker from the watch we found, \Ve can also infer so1nething about the \Vatchmaker's characteristics. We can say that the designer must have been fairly intelligent to produce an object of such intricacy. Chimps are somewhat intelligent, but the idea that a chimp could have made the watch is dubious. Rather, what we naturally infer is that the watchmaker must have had an intelligence at least on the order of human intelligence, given the features of the watch that we observe.
The Analogy
So far in this argument, Paley is simply describing what common sense would say about the watch on the beach. Paley then suggests an analogy. Look around the living world. Notice that it is filled with organisms that are extremely intricate and well adapted to living in the environ1nents.they inhabit. In fact, organisms are far n1ore cornplicated than \vatches. And as well suited as a \Vatch is to the task of ineasuring tin1e, organisms are even better suited to the tasks of surviving and reproducing.
How can we explain the fact that organisms are so amazingly intricate and well adapted? One possibility is the Random Hypothesis-that by a process akin to waves pounding on sand, orchids, crocodiles, and people came into existence. The other alternative is the Design Hypothesis-that an organism maker made the impressive pieces of machinery we call living things. Which explanation is more plausible? If the Random Hypothesis says that the existence of the watch is very improbable, then the Random Hypothesis must also say that the existence of these intricate and adapted organisms is very improbable. So if inferring the existence of a watchmaker is plausible in the first case, then inferring the existence of a designer of all life is plausible in the second.
Finally, we may ask how intelligent this maker of organisms must be, given the in- tricacy and fineness of adaptation that organisms exhibit. From what watches are like, we can infer that watchmakers must be pretty smart. By the same reasoning, we infer that the maker of organisms must be very, very intelligent-far more intelligent than human beings are. Paley's design argument concludes that the intricacy and adaptedness of organisms are best explained by postulating the existence of an extremely intelligent designer.
Abductive Arguments Often Postulate Unobserved Entities
There is a point that pertains to all of Aquinas's arguments that should be emphasized here. The design argument claims that there is something we observe-the complexity and adaptedness of living things-that is best explained by the hypothesis that there is a God. The conclusion of the argument concerns the existence of something we have not directly observed. Although there may be defects in this argument, the fact that it reaches a con- clusion about a being we have not observed isn't one of them. Recall from Chapter 3 that
Chapter 5: The Design Argurnent
........ '' ... ' ......... >.'. ' ...... '. '.' .. ' .... '.' ........................ . abductive arguments frequently have this characteristic. It would cripple science to limit theorizing to a description of what scientists have actually observed. So my view of Paley's argun1ent is that it is an abductive argutnent:
Organisms are intricate and well suited to the tasks of survival and reproduction.
Hence, organisms 'vere created by an intelligent designer.
I've drawn a double line here to indicate that the argument does not aim at being deductively valid.
To show this is a strong abductive argument, Paley argues that it is analogous to a second inference to the best explanation:
The watch is intricate and well suited to the task of measuring time.
Hence, the watch was created by an intelligent designer.
Paley claims that if you grant that the watch argument is convincing, you should grant that the organistn argument is convincing as 'vell.
Hume's Criticisms of the Design Argument
Hu1ne's Dialogues on Natural Religion contain several criticis1ns of the design argument. Sometimes he is talking about global design arguments-ones that argue that the entire universe must be the product of intelligent design. At other times, Hume addresses local arguments-ones that focus on the adaptedness and intricacy of organisms.
I've claimed that design arguments are abductive. Hume paints a very different picture. He represents the arguments as being inductive arguments or arguments from analogy. This may not look like a very important difference, as all of these formulations involve a nonde- ductive inference. But you will see in what follows that two of Hume's criticisms of design arguments aren't very convincing if we think of the design argument as an inference to the best explanation.
Is the Design Argument a Weak Argument from Analogy?
In this section, I'll discuss a criticism of the argument from design that Hume develops in Part II of his Dialogues (see the paragraph beginning "What I chiefly scruple .. :').
To see what Hume has in mind when he talks about arguments from analogy, consider the following example of an analogy argument:
Human beings circulate their blood.
Dogs circulate their blood.
I've drawn a double line between the premise and the conclusion, again to indicate that the argument isn't supposed to be deductively valid. In this argument, let's call human beings the analogs and dogs the targets. I say that dogs are the targets here because they are the
r:EJ-[ Read Part 2 of Dialog!,es
Concerning Natural Religion on ww·w .mysearchlab.com
Part Two: Philosophy of Religion
••••• <'.' ' ••• ' •• ' ••••• ' •• '.' ', '' '.' ' •• ''''' '''' '''' ''''''''''' ''' ''
iten1s about 'ivhich the argun1ent aims to reach a conclusion. Htune suggests, 'iVith so1ne plausibility, that such argun1ents are stronger or 'iveaker depending on ho'iv similar the analogs are to the targets. To see 'ivhat he ineans here, co1npare the above argu1nent 'ivith the following one:
Human beings circulate their blood.
Plants circulate their blood.
This argument is pretty weak because human beings and plants aren't very similar. We can formulate Hume's point by saying that an analogy argument has the following
logical form:
Object A has property P.
Object A and object Tare similar to degree n. n[~~====~
T has property P.
A is the analog and Tis the target. The number n measures the degree of similarity be- t'iveen A and T. It goes fro1n a 1ninin1tun value of o (1neaning that A and T are1it siinilar at all) to a maximum value of I (meaning that they share 100 percent of their characteristics). This number also represents a probability-that is why "n" is next to the double line sepa- rating pre1nises from conclusion. A high value of n means that A and T are very shnilar and that the premises make the conclusion very probable. This expresses Hume's idea that the more similar A and Tare, the more probable it is that the target object T has the prop- erty that is found in the analog A.
Hun1e uses this idea about analogy argtunents to criticize the design argtnnent. He thinks the design argument has the following form:
Watches are products of intelligent design. n[~====~=
The universe is a product of intelligent design.
This is a very 'iveak argument, Hun1e says, as the analog is really not very shnilar to the target. Watches rese1nble the universe as a 'iVhole in some \\Tays, but fail to do so in a great many others. So n has a low value.
Here Hume is criticizing 'ivhat I've called a global design argu1nent-an argument that focuses on some large-scale features of the entire universe. Hu1ne's point, ho'ivever, also applies to local design arguments-to arguments that focus on organisms and their characteristics:
Watches are products of intelligent design.
n[~======
Organisms are products of intelligent design.
Hun1e's criticism is that organis1ns are really not very sin1ilar to 'ivatches. Watches are made of metal, but organisms aren't. Kangaroos hop around, but 'ivatches don't. Organisms re- produce and obtain nutrition fro1n their environ1nent, but 'ivatches don't. And so on. As
Chapter 5: The Design Argument 65
' •• ','' '' '"'.' ' ••• '' '' <' '' '' '.' ', ••• ''.
analog and target are so dissin1ilar, the analogy argument is a very '\Veak one; n is lo\v here as well.
Hume's idea is that the strength or weakness of an analogy argument depends on the overall similarity of target and analog. You look at all the known characteristics of target and analog and try to say how similar they are overall. I grant that if you did this, you \Vould conclude that '\Vatches and kangaroos aren't very sitnilar. My view, ho,vever, is that this doesn't undermine the design argument at all when that argument is taken to be ab- ductive. It is entirely irrelevant whether watches and kangaroos both have fur, or whether both hop around, or whether both reproduce. The design argument focuses on a single pair of features of each of these and asks how it should be explained. A watch's intricacy, as well as its being well suited to the task of measuring time, requires that we think of it as the product of intelligent design. Paley's claim is that an organism's intricacy, as well as its be- ing well suited to the tasks of survival and reproduction, ought to be explained in the same way. It doesn't matter that the one is made of metal while the other isn't. Overall similarity is irrelevant.
The fundamental idea of Paley's argument is that the Surprise Principle tells us that the Design Hypothesis is better supported than the Random Hypothesis, given the ob- servations \Ve have 1nade about living things. This argun1ent stands on its o\vn. To use the Surprise Principle in this case, it doesn't matter \vhether organis1ns are similar to \vatches or to anything else. I conclude that Hume is 1nistaken to criticize the design argtunent as a weak argument from analogy.
Is the Design Argument a Weak Induction?
A second criticism that Hume levels at the design argument rests on his assuming that the argument must be inductive if it is to make sense. (Here I have in mind the paragraph in Part II of the Dialogues that begins "And can you blame me ... ";see especially the passage that begins "When two species of objects ... ")
Recall from Chapter 2 that inductive arguments involve observing a sample and extrapolating from it to some claim about one or more objects not in the sample. For example, suppose I call a large number of voters registered in a county and find that most of them are Democrats. This seems to license the inference that the next voter I call will probably be a Democrat. Hume observes, again with some plausibility, that the strength of an inductive inference is influenced by sample size. In particular, if my sample had in- cluded only five individuals, I would be on rather shaky ground if I used this as my basis for predicting what the next voter called would be like. My inference would be on even shakier ground if I ventured a guess about the next telephone call having never sampled even a single voter. A sa1nple size of zero is just plain silly; an inductive argument can't be weaker than that.
Hume claims that if we are to have a reason for thinking that the universe we inhabit is the product of intelligent design, we must base this conclusion on induction. What would this involve? We \Vould have to examine a large number of other universes and see that most or all of them were the result of intelligent design. If our sample size were suffi- ciently large, that would justify a conclusion about the universe that we inhabit. But how big is our sample size? How many universes have we observed being made by an intelligent
66 Part Two: Philosophy of Religion
... '.'.' ..... '. '" ..... ' .. ' ... , ... ' .. '' ...... '.' ................ ' ... ' .. ' .. . designer? The ans\ver is zero. The only universe vve have ever experienced is the one \Ve inhabit. We have not seen our universe being tnade by an intelligent designer, nor have \Ve seen an intelligent designer make the organisms that exist in our universe. So no inductive argument can be constructed here.
My vie\v is that this is true, but irrelevant. Small satnple size does \Veaken an inductive argu1nent. Ho\vever, the design argument isn't an inductive argun1ent. Hu1ne assumed that the only sorts of inferences worth taking seriously are inductive and deductive. I think this is a mistake. There is abduction as well. Mendel didn't have to observe that lots of differ- ent organisms have genes before he could conclude that his pea plants have genes. Mendel never sa\v a single gene, but that didn't prevent him from inferring their existence. His inference "ras abductive, not inductive.
I've reviewed two of Hume's criticisms of the design argument. They don't work. Of course, this doesn't Jnean that the argument has no fla\vs, only that \Ve have yet to uncover one. The design argument that Paley formulated considers two competing hypotheses- the hypothesis of intelligent design and the hypothesis of random physical processes. In the mid-nineteenth century, a new hypothesis was formulated that we now need to con- sider as a third alternative-this is Darwin's theory of evolution by natural selection. In the next chapter, I'll describe what this hypothesis asserts and discuss how it compares with the hypothesis of intelligent design.
Review Questions
1. What does it mean to say that the design argument is an abductive argument?
2. \Vhat is the difference between a global design argument and a local design argument?
3. How does Paley's argument about the watch use the Surprise Principle?
4. Hume formulated a principle that states ho\v the strength of an analogy argument may be mea· sured. What is it?
5. What t\vo criticisms did Hume make of the design argument? Are these good criticisms if the argument is understood to be abductive in character?
Problems for Further Thought
1. It might be suggested that one difference bet\veen Paley's argument about the watch and his argument about organisms is that we have seen watchmakers, but have never directly observed God. Does this point of difference undermine the force of Paley's design argument?
2. I mentioned in passing that modern science no longer takes seriously the idea that all things are goal-directed systems. Consider the following pair of propositions. Can you think of a reason that the first of them might be true, whereas the second might be rejected?
The function of the heart is to pump blood. The function of rain is to provide farm crops \Vith water. What does it mean to attribute a "function" to something?
J. ln addition to the two criticisms that Hume makes of the design argument that are described in this chapter, Hume presents a third. He says that even if the design argument succeeds in show· ing that a designer made the universe {or the organisms in it), it does not succeed in establishing what characteristics that designer has. For this reason, the argument does not show that God exists. ls Hume's claim correct? How seriously does this undermine the design argument?
Chapter 5: The Design Argurnent
.. ' ..... '.' ...... ' .. ' ..... ' ....... ''' .. ' ..... ''. ' ..... ' .. ' ... , ...... '' '''. MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts,
research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
mJ-[ READ AND REVIEW
READ
1. William Paley, "The Design Argument," Chapter 1 of Natural Theology · In this selection from his Natural Theology (1836 edition), William Paley develops an argument
for the existence of God that traces back at least to the last of Aquinas's five ways. His discussion of how one interprets a watch found on a heath is one of the most famous analogies proposed in the history of philosophy. The term "natural theology" means that the author attempts to establish the existence and nature of God by the same methods used in the natural sciences- observation and reasoning. How does Paley's argument resemble arguments used in science?
2. David Hume, Part 2 of Dialogues Concerning Natural Religion In this selection from his Dialogues Concerning Natural Religion (1779), Hume describes a conversation between Cleanthes, Demea, and Philo. Hume did not publish this work during his lifetime, probably because he was apprehensive about what the public reaction would be. How does Hume try to undermine the analogy between evidence of human design and evidence of supernatural design that Paley makes?
RESEARCH AND EXPLORE
l)I<{ EXPLORE
1. Intelligent design
2. Inference to the best explanation 3. William Paley
chapter 6
68
Evolution and Creationism
Chapter Outline
Creationisn1
So1ne Creationist Argun1ents
Darwin's Two-Part Theory
Natural Selection
Speciation
The Tree of Life
The Principle of the Common Cause
Arbitrary Similarities a1nong Organis111s
Useful Shnilarities among Organis1ns
Irreducible Complexity
Is Creationism Testable?
Predictive Equivalence
Prediction versus Acco1n1nodation
Does Evolutionary Theory Make Novel Predictions?
Concluding Remarks
Aquinas and Paley maintained that the intricacy and adaptedness of organisms can be explained only by viewing them as the product of intelligent design, but they were not able to consider an alternative theory that Charles Darwin (1809-1882) put on the table in 1859 when he published his book The Origin of Species. Modern-day creationists do consider what their predecessors could not; they reject Darwin's theory and maintain that the old design argument is still correct. My goal in this chapter is to give a sample of the kinds of argun1ents that one needs to consider in thinking about the evolution versus creation debate. As I promised at the end of Chapter 3, I'll here introduce a new principle that is important to abductive inference, which will supplement the Surprise Principle and the Only Game in Town Fallacy. Unfortunately, 1 won't have time to provide a full treatment of the philosophical issues, and there are lots of biological details that are important here that I won't be able to discuss.
Chapter 6: Evolution and Creationism
'''''' '' '''''' ' ''','' '' '.'' ''' '' ''.''''.' '' .. ' ..... ' .. ' .... ' .... ' ..... Creationism
Creationists (son1etin1es calling the1nselves «scientific creationists" or «intelligent design theorists") are present-day defenders of the design argument. Although they agree among themselves that intelligent design is needed to explain some features of the living world, they disagree with each other about various points of detail. Some hold that the earth is young (around 10,000 years old), \Vhereas others concede that it is ancient-about 4-5 billion years old, according to current geology. Some creationists maintain that each species (or basic "kind" of organism) was separately created by an intelligent designer, whereas others concede that biologists are right when they assert, as Darwin did, that all life on Earth traces back to a comn1011 ancestor.
To clarify \vhat creationis1n asserts, let's consider three possible relationships that might obtain among God ( G), mindless evolutionary processes (E), and the complex adap- tations that organisms are observed to have ( 0):
(Theistic evolutionism)
(Atheistic evolutionisn1)
(Creationism)
G-;E-;Q
E-; 0
G-->E---f->O ~
Theistic evolutionisn1 says that God set inindless evolutionary processes in 1notion; these processes, once under\vay, suffice to explain the co1nplex adaptations \Ve observe organisms to have. Atheistic evolutionism denies that there is a God, but otherwise agrees with theistic evolutionisn1 that inindless evolutionary processes are responsible for complex adaptations. Creationistn disagrees \vith both theistic evolutionisn1 and atheistic evolutionis1n. It n1ain- tains that mindless evolutionary processes are incapable of giving rise to co1nplex adaptations and that God directly intervenes in nature to bring these about. Creationism does not deny that evolution is responsible for so1ne of the features \Ve observe in nature; creationists con- cede that quantitative changes in a feature found in a species might be due to natural selection (an example of this sort of change will be described below). However, the emergence of genu- inely novel, con1plex adaptive features is, for creationists, another story entirely.
You can see from these three options that belief in evolutionary theory is not the san1e as atheis1n. In rny opinion, current evolutionary theory is neutral on the question of whether there is a God. Evolutionary theory can be supplemented with a claim, either pro or con, concerning \vhether God exists. It is also consistent \Vith agnosticism, \Vhich is the view that we don't know whether there is a God. Evolutionary theory, however, is not consistent with creationism. It holds that mindless evolutionary processes (including the process of natural selection) are responsible for the complex adaptations we observe; creationisn1 denies this.
Some Creationist Arguments
Son1e of the 1nost frequently repeated creationist arguments contain mistakes and confu- sions. For example, some creationists have argued that evolutionary theory is on shaky ground because hypotheses about the distant past can't be proven with absolute certainty.
Part Two: Philosophy of Religion
They are right that evolutionary theory isn't absolutely certain, but then nothing in sci- ence is absolutely certain; recall the remarks in Chapter 3 about gambling. What one le- gitimately strives for in science is po,verful evidence sho,ving that one explanation is far more plausible than its competitors. Biologists now think that the hypothesis of evolution is about as certain as any hypothesis about the prehistoric past could be. Naturally, no scientist 'vas on the scene some 3.8 billion years ago. Ho,vever, it is nonetheless possible to have strong evidence about rnatters that one can't directly observe, as I hope rny previous discussion of abduction has made clear.
Another example of an error that some creationists make is in their discussion of the Second Law of Thermodynamics. They claim that this law shows that it is impossible for order to arise frorn disorder by natural processes. Natural processes can lead an automo- bile to disintegrate into a junk heap, but creationists think the Second Law of Thermo- dynamics says that no natural process can cause a pile of junk to assemble itself into a functioning car. Here creationists are arguing that physics is inconsistent with the claim that life evolved from nonlife. What the Second Law actually says is that a closed system will (with high probability) move from states of greater order to states oflesser order. But if the system is1it closed, the law says nothing about what will happen. So if the earth were a closed systern, its overall level of disorder would have to increase. But, of course, the earth is no such thing-energy from the sun is a constant input. If the universe as a whole is a closed system, then thermodynamics does entail that disorder will increase overall. But this overall trend doesn't prohibit pockets of order from arising and being maintained. The Second Law of Thermodynamics offers no basis for thinking that life couldn't have evolved from nonlife.
A full treatment of the evolution versus creationism debate would require me to describe the positive explanations that creationists have advanced. If you want to com- pare evolutionary theory and creationism, you ca1it just focus on the criticisms that creationists make of evolutionary ideas. You also should look carefully at what the al- ternative is. Doing this produces lots of difficulties for creationism. The reason is that creationists have either been woefully silent on the details of the explanation they want to defend, or they have produced detailed stories that can't withstand scientific scru- tiny. For example, young earth creationists, as I rnentioned, nlaintain that the earth is only a few thousand years old. This claim conflicts with a variety of very solid scientific findings from geology and physics. It isn't just evolutionary theory that you have to reject if you buy into this version of creationism, but a good deal of the rest of science as well.
As I have also indicated, there are nlany different versions of creationism. Creationism is not a single theory, but a cluster of similar theories. In the present chapter, I will focus on only a few of the options. First let's look at the ABCs of Darwinism.
Darwin's Two-Part Theory
Many of the main ideas that Darwin developed in The Origin of Species are still regarded by scientists as correct, but others have been refined or expanded. Still others have been junked entirely. Although evolutionary theory has developed considerably since Darwin's time, I'll take his basic ideas as a point of departure. Dar,vin's theory contains t'vo rnain
Chapter 6: Evolution and Creationis1n 71
'' •••• '' ' •• ' <'.' •• ' '.'' ''.' ' •••• '4 •• ' ••• ''.' ••• '.' ' •• ' ••• <'''''.''.'.' ... '
elements. First, there is the idea that all present-day living things are genealogically re- lated. The organis1ns \Ve see didn,t come into existence independently by separate cre- ation. Rather, organisms are related to each other by a family tree. You and I are related. If we go back far enough in time, we'll find a human being who is an ancestor of both of us. The same is true of us and a chimp, though, of course, we must go back even further in time to reach a co1n1non ancestor. And so it is for any t\VO present-day organisms. Life evolved from nonlife, and then descent with modification gave rise to the diversity we no\v observe.
Notice that this first hypothesis of Darwin's says nothing about why new characteris- tics arose in the course of evolution. If all life is related, why aren't all living things iden- tical? The second part of Darwin's theory is the idea of natural selection. This hypothesis tries to explain why new characteristics appear and become common and why some old characteristics disappear. It is very important to keep these t\VO elements in Dar\vin,s theory separate. The idea that all present-day living things are genealogically related isn't at all controversial in modern science. The idea that natural selection is the prin- cipal cause of evolutionary change is somewhat controversial, although it is still by far the majority view among biologists. One reason it is important to keep these ideas sepa- rate is that some creationists have tried to score points by confusing them. Creationists sometimes suggest that the whole idea of evolution is something even biologists regard with great doubt and suspicion. However, the idea that all life is related isn't at all contro- versial. What is controversial, at least to some degree, are ideas about natural selection.
Natural Selection
Here's a simple example of how natural selection works. Imagine a population of zebras that all have the same top speed. They can't run faster than 38 mph. Now imagine that a novelty appears in the population. A mutation occurs-a change in the genes found in some zebra-that allows that newfangled zebra to run faster (at 42 mph), say. Suppose running faster is advantageous, because a fast zebra is less likely to be caught and eaten by a predator than a slow one is. Running fast enhances the organisin's fitness-its ability to survive and reproduce. If running speed is passed on from parent to offspring, what will happen? What will occur (probably) is that the fast zebra will have more offspring than the average slow zebra. As a result, the percentage of fast zebras in the population increases. In the next generation, fast zebras enjoy the same advantage, and so the characteristic of being fast will again increase in frequency. After a number of generations, we expect all the zebras to have this new characteristic. Initially, all the zebras ran at 38 mph. After the selec- tion process runs its course, all run at 42 mph. So the process comes in hvo stages. First, a novel inutation occurs, creating the variation upon \Vhich natural selection operates. Then, natural selection goes to work changing the composition of the population:
St mt
100% run
at38mph
~ Then A novel inutant runs at 42 mph; the rest run at 38 1nph
~Finish
100% run at42 mph
72 Part T\'10: Philosophy of Religion
We nlay su1nmarize ho\v this process "\Vorks by saying that natural selection occurs in a population of organisms \Vhen there is inherited variation in fitness. Let's analyze ,vhat this means. The organis1ns must vary; if all the organisrns are the san1e, then there \Vill be no variants to select an1ong. What is 1nore, the variations n1ust be passed do,vn fron1 parents to offspring. This is the requirement of inheritance. Last, it must be true that the varying characteristics in a population affect an organis1n's fitness-its chance of surviving and reproducing. If these three conditions are met, the population will evolve. By this, I mean that the frequency of characteristics will change.
The idea of natural selection is really quite simple. What Darwin did was to show how this simple idea has many implications and applications. Merely stating this simple idea wouldn't have convinced anyone that natural selection is the right explanation of life's diversity. The power of the idea comes from the numerous detailed applications.
Notice that the introduction of novel characteristics into a population is a precondi- tion for natural selection to occur. Dar'ivin didn't have a very accurate picture ofho\v novel traits arise. He theorized about this, but didn't come up with anything of lasting impor- tance. Rather, it was later in the nineteenth century that Mendel started to fill in this detail. Genetic 1nutations, \Ve no\v understand, are the source of the variation on \Vhich natural selection depends. One central idea that Darwin had about mutation, which twentieth- century genetics has vindicated, is that n1utations do not occur because they \vould be useful to the organis1n; this is \vhat biologists mean 'vhen they say that 1nutations occur ((at rando1n:'
Creationists son1etitnes say that the process of evolution by natural selection is like a tornado blowing through a junkyard. The latter process cannot sweep together the scraps of 1netal lying around on the ground and asse1nble them into a functioning auto- mobile. From this, creationists conclude that the former process is likewise incapable of creating novel adaptations. This analogy is fundamentally misleading. The tornado is a totally random process, like the spinning of a huge roulette wheel. However, evolution by natural selection has two parts; n1utations appear randon1ly, but then it is not a random 1natter 'vhich mutations increase in frequency and \\1hich decline. Selection is nonrando1n. A better analogy than the tornado in the junkyard is one that Darwin proposed in his 1873 book 11ie Variation of Animals and Plants Under Domestication:
Let an architect be compelled to build an edifice \Vith uncut stones, fallen fro1n a precipice. The shape of each fragment may be called accidental; yet the shape of each has been determined by the force of gravity, the nature of the rock, and the slope of the precipice-events and circumstances all of "'tvhich depend on natural laws; but there is no relation behveen these la\vs and the purpose for \vhich each frag1nent is used by the builder. In the san1e n1anner, the variations of each crea- ture are determined by fixed and in11nutable la\vs; but these bear no relation to the living structure 'vhich is slo,vly built up through the po,ver of natural selection. (Darwin 1876, p. 236)
Notice that the little story I've told about zebra running speed describes a rather modest change that occurs within an existing species. Yet Darwin's 1859 book was called 11ie
Chapter 6: Evolution and Creationism
Origin of Species. I-Io\v does change \Vithin a species help explain the coining into existence of ne\v species?
Speciation
Darwin's hypothesis was that small changes in a population (like the one I just described) add up. Given enough little changes, the organisms will become very different. Modern evolutionists usually tell a story like the following one. Think of a single population of zebras. ImagiI]e that a small number of zebras are separated from the rest of the popula- tion for son1e reason; rnaybe they \Vander off or a river changes course and splits the old population in t\vo. If the resulting populations live in different environments, selection will lead them to become increasingly different. Characteristics that are advantageous in one population may not be advantageous in the other. After a long time, the populations will have diverged. They will have become so different from each other that individuals from the one can't breed with individuals from the other. Because of this, they will be two species, not hvo populations belonging to the san1e species.
Pretty much everybody in Darwin's day, including those who thought that God created each species separately, would have agreed that the little story about zebras evolving a greater speed could be true. The real resistance to Darwin's theory focused on his thesis that the mechanism responsible for small-scale changes within species also gives rise to large-scale changes, na1nely. to the origin of 11e1v species. This \Vas a daring hypothesis, but it is no\v the inainstream vie\v in evolutionary theory. Even so, biologists continue to debate the importance that natural selection has had in the evolutionary process. Modern evolutionary theory describes other possible causes of evolutionary change. Which traits \Vere due to natural selection and \Vhich \Vere due to other evolutionary processes? There are a number of still unans\vered questions in evolutionary biology about natural selection. Even biologists who hold that natural se- lection is the 1najor cause of evolution are sometitnes puzzled about how it applies in particular cases. For exa1nple, it is still rather unclear \vhy sexual reproduction evolved. Some creatures reproduce sexually, others asexually. Why is this? Although there are open questions pertaining to natural selection, I \Vant to en1phasize that it isn't at all controversial that hu1nan beings share co1n1non ancestors \Vith chimps. Don't con- fuse the idea of con1n1on ancestry \vith the idea of natural selection; these are separate elen1ents in Dar\vin's theory.
The Tree of Life
I now turn to this uncontroversial idea. Why do biologists think it is so clear that living things are related to each other-that there is a family tree oflife on Earth just as there is a family tree of your family? Two kinds of evidence have seemed persuasive. I won't give the details here; rather, I want to describe the kinds of arguments biologists deploy. As a philosopher, I'm more interested that you grasp the logic of the arguments; for the biologi- cal details, you should consult a biology book.
73
74 Part Two: Philosophy of Religion
•••••••••• ' ••• ' ••••• ' ••• ' •• ' ••• ' •••••••••• ' ••• ' '.'' •••• ' > ••• '. ''' '.''.'.'.
To illustrate how one line of argument works, consider this simple problem. Suppose I assign a philosophy class the job of writing an essay on the meaning of life. As I read through the papers, I notice that two students have handed in papers that are word- for-word identical. How should I explain this striking similarity? One possibility, of course, is that the students worked independently and by coincidence arrived at exactly the same result. The independent origin of the two papers isn't impossible. But I would regard this hypothesis as extremely implausible. Far more convincing is the idea that one student cop- ied from the other or that each of them copied from a common source-an article down- loaded from the Internet, perhaps. This hypothesis is a more plausible explanation of the observed similarity of the two papers.
The Principle of the Common Cause
The plagiarism example illustrates an idea that the philosopher Hans Reichenbach (in The Direction of Time, University of California Press, i956) called the Principle of the Com- mon Cause. Let's analyze the example more carefully to understand the rationale of the principle.
Why, in the case just described, is it more plausible that the students copied from a common source than that they wrote their papers independently? Consider how prob- able the matching of the two papers is, according to each of the two hypotheses. If the two students copied from a common source, then it is rather probable that the papers should closely resemble each other. If, however, the students worked independently, then it is enormously improbable that the two papers should be so similar. Here we have an ap- plication of the Surprise Principle described in Chapter 3: if one hypothesis says that the observations are very probable whereas the other hypothesis says that the observations are very improbable, then the observations strongly favor the first hypothesis over the second. The Principle of the Co1n1non Cause makes sense because it is a consequence of the Surprise Principle.
The example just described involves hypotheses that describe mental activity-when students plagiarize they use their minds, and the same is true when they write papers in- dependently. However, it is important to see that the Principle of the Common Cause also makes excellent sense when the hypotheses considered do not describe mental processes. Here's an example: I have a barometer at my house. I notice that '\vhen it says "lo,v:' there is usually a storm the next day, and when it says "high;' there is usually no storm the next day. The barometer reading on one day and the weather on the next are correlated. It may be that this correlation is just a coincidence; perhaps the two events are entirely indepen- dent. However, a far more plausible hypothesis is that the reading on one day and the weather on the next trace back to a common cause-namely, the weather at the time the reading is taken:
Barometer reading today Weather tomorro'\v
\ I Today's weather
Chapter 6: Evolution and Creationism 75
""' '' '' '.' '.''' '' '' '''' ''' '''. ''' '.'' ''.' ''.','.' '''.''' '''
The common cause hypothesis is more plausible because it leads you to expect the correlation of the two observed effects. The separate cause hypothesis is less plausible be- cause it says that the observed correlation is a very improbable coincidence. Notice that the hypotheses in this example do not describe the mental activities of agents.
Arbitrary Similarities among Organisms
I'll now apply this principle to the evolutionary idea of common ancestry. One reason biologists think all life is related is that all organisms (with some minor exceptions) have the same genetic code. To understand what this means, think of the genes in your body as a set of instructions for constructing nlore complex biological items-an1ino acids and then proteins. The total sequence of genes in your body and the sequence in a frog's are different. The striking fact, however, is that the gene that codes for a given amino acid in a frog codes for that very san1e an1ino acid in people. As far as 've kno"' no,v, there is no reason why the genes that code for a given amino acid had to code for that acid rather than so1ne other. The code is arbitrary; there is no functional reason \vhy it has to be the \vay it is. (Don't be misled by my talk of codes here. This word may suggest intelligent design, but this isn't 'vhat biologists mean. Genes cause amino acids to form; for present purposes, this is a perfectly satisfactory way to understand what it means for genes to "code for" this or that amino acid.)
How are we to explain the detailed similarity among the genetic codes that different species use? If the species arose independently of each other, we would expect them to use different genetic codes. But if those species all trace back to a common ancestor, it is to be expected that they will share the same genetic code. The Principle of the Common Cause underlies the belief that evolutionary biologists have that all living things on Earth have co1nmon ancestors.
Useful Similarities among Organisms
The reason a shared genetic code is evidence of common ancestry is that the code is arbi- trary. There are lots of possible codes that would work. If there were only a single functional code, the fact that different species use this one code would not be evidence of common ancestry. Consider, for example, the fact that sharks and dolphins both have a streamlined body shape. Both are shaped like torpedoes. Is this strong evidence that they have a com- 1non ancestor? I \vould say not. There is an obvious functional reason 'vhy large predators that spend their lives swimming through water should be shaped like this. If there is life in other galaxies that includes large aquatic predators, we would probably expect those organisms to have this torpedo shape. Even iflife on Earth and life on other galaxies are not descended fro1n common ancestors, there are son1e sin1ilarities 've \Votild still expect to find. I conclude that the streamlined shape of dolphins and sharks isn't strong evidence that they evolved from a common ancestor. The Surprise Principle explains why some similarities, but not others, are good evidence for the hypothesis that there is a tree of life uniting all organisms on Earth.
The genetic code is just one example of a similarity that can't be explained by its useful- ness to the organism. There are lots of others. Consider the fact that human beings have
Part Two: Philosophy of Religion
tail bones and that human fetuses have gill slits. Neither of these features is useful. Biologists interpret these features as evidence that 've share comn1on ancestors 'vith nonhuman or- ganis1ns. This and lots of other evidence point to the conclusion that 've share conunon ancestors with monkeys, with other mammals, with fish, and with all other living things. In drawing this conclusion, biologists are using the Surprise Principle and the Principle of the Comrnon Cause.
Irreducible Complexity
Although creationists have usually rejected the Darwinian hypothesis of common ancestry, not all have done so. What creationists universally reject is the thesis that natu- ral selection is the correct explanation of the cornplex adaptations '\Ve observe in nature. Modern-day creationists are usually willing to grant that selection can explain small mod- ifications in existing species, as in 1ny earlier exan1ple about zebra running speed. But how can the gradual accumulation of modifications explain a feature like the vertebrate eye? This is the key objection that Michael Behe develops in his 1996 book Darwin's Black Box when he introduces the concept of irreducible complexity. Behe defines an irreduc- ibly complex system as one in which the whole system has a function, the system is made of many parts, and the system would not be able to perform its function if any of the parts 'vere ren1oved. Bebe's idea is that the Dar,vinian process of natural selection involves adding one small part to another, with each modification improving the fitness of the organism. But what good is 1 percent of an eye? Creationists, Behe included, think that the answer is obvious-no good at all-and that this shows that evolutionary theory can't explain complex adaptations.
HUMANS FROM NONHUMANS, LIFE FROM NONLIFE
When people hear about the idea of evolution, there are two parts of the theory that sometimes strike them as puzzling. First, there is the idea that human beings are descended from apelike ancestors. Second, there is the idea that life evolved from nonliving materials.
Scientists believe the first of these statements because there are so many striking similarities between apes and human beings. This isn't to deny that there are dif- ferences. However, the similarities (for example, the fact that both have tail bones) would be expected if humans and apes have a common ancestor, but would be quite surprising if they came into existence independently.
There is a big difference between having evidence that humans are descended from apelike ancestors and having an explanation of precisely why this happened. The evidence for a common ancestor is pretty overwhelming, but the details of why evo- lution proceeded in just the way it did are less certain. Students of human evolution continue to investigate why our species evolved as it did. In contrast, the claim that we did evolve isn't a matter of contemporary scientific debate.
What about the second idea-that life arose from nonlife? Why not maintain that God created the first living thing and then let evolution by natural selection pro- duce the diversity we now observe? Notice that this is a very different idea from what
Chapter 6: Evolution and Creationisn1 77
• '''' '''' '' '' '.' ''' ' •• ' '''. ' •••• '. '. '''.''.'.'.' ••• >.' ••• ' '.' ••.• '' •• ' •• '
creationists maintain. They hold that each species (or "basic kind" of organism) is the result of separate creation by God. They deny that present-day species are united by common descent from earlier life forms.
One main sort of evidence for thinking that life evolved from non life on Earth about 4 billion years ago comes from laboratory experiments. Scientists have created labora- tory conditions that resemble the ones they believe were present shortly after the earth came into existence about 4.5 billion years ago. They find that the nonliving ingredi- ents present then can enter into chemical reactions, the products of which are simple
1 organic materials. For example, it is possible to run electricity (lightning) through a "soup" of inorganic molecules and produce amino acids. Why is this significant? Amino acids are an essential stage in the process whereby genes construct an organ- ism. Similar experiments have generated a variety of promising results. This subject in biology-prebiotic evolution-is very much open and incomplete. No one has yet been able to get inorganic materials to produce DNA, but the promising successes to date suggest that further work will shed more light on how life arose from non life.
Laboratory experiments don1t aim to create a multicellular organism from inor- ganic materials. No one wants to make a chicken out of carbon, ammonia, and water. Evolution by natural selection proceeds by the accumulation of very small changes. So the transition from nonlife to life must involve the creation of a rather simple self. replicating molecule. Chickens came much later.
A self-replicating molecule is a molecule that makes copies of itself. A molecule of this sort is able to reproduce. With accurate replication, the offspring of a molecule
, will resemble its parent. Once a simple self-replicating molecule is in place, evolu- tion by natural selection can begin. It may sound strange to describe a simple self. replicating molecule as being "alive." Such a molecule will do few of the things that a chicken does. But from the biological point of view, reproduction and heredity (that is, similarity between parents and offspring) are fundamental properties.
Biologists sometimes respond to this creationist claim by arguing that gradual modifi- cation can explain structures like the eye. A piece oflight-sensitive skin allows the organ- ism to tell the difference between light and dark, and this is advantageous. Then, if this skin is shaped into a cup, the organism can tell not only whether it is light or dark, but also from what direction the light is coming, and this provides a further advantage. Maybe it is plausible that the eye could have evolved one small step at a time. However, there is another, more fundamental, problem with Bebe's argument. What we call "the parts" of a system may or may not correspond to the historical sequence of accumulating details. Consider the horse and its four legs. A horse with zero, one, or two legs cannot walk or run; suppose the san1e is true of a horse with three. In contrast, a horse \vith four legs can walk and run, and it thereby gains a fitness advantage. So far so good-the four-legged arrangement satisfies the definition of irreducible complexity. The mistake comes from thinking that horses (or their ancestors) had to evolve their four legs one leg at a time. In fact, it's a mistake to think that a separate set of genes controls the development of each leg; rather, there is a single set that controls the development of appendages. A division of a system into parts that entails that the system is irreducibly complex may or may not
Part T\vo: Philosophy of Religion
'.' .. ' '''''''' .. ''''''.' '' ''. '. ''' '' '''' ''.,'' ''.' ., ' ' ''.',.'''' ''' ''' correspond to the historical sequence of events through which the lineage passed. This point is obvious '\'ith respect to the horse's four legs, but needs to be borne in Jnind ,vhen other, less familiar, organic features are considered. What 've call the "parts" of the eye n1ay not correspond to the sequence of events that occurred in the eye's evolution.
Is Creationism Testable?
So far I have outlined what Darwin's theory of evolution amounts to, the kind of evidence that biologists take seriously for the claim of common ancestry, and some objections that creationists Jnake to this theory. 1'he question I 'vant to consider no'v concerns the theory that creationists themselves present. What is it? What predictions does it make? What evi- dence is there for the theory they present?
Here 've need to consider so1ne of the different versions that creationis1n might take. To get started, let's consider:
H,: A superintelligent designer fashioned all the complex adaptations we observe or- ganisms to have so that organisms would be perfectly adapted to their environment.
This hypothesis is disconfirmed by what we observe. Organisms often have highly imper- fect adaptations. The eye that human beings use has a blind spot, though the eye of the octopus does not. And many spiders have eyes with built-in sunglasses, though human beings do not. Our eye is imperfect and so are lots of features that we and other organisms possess. Can we repair this defect in H,? One way to do so is to make the hypothesis of intelligent design more modest in what it says:
H,: An intelligent designer fashioned all the complex adaptations we observe organisms to have.
The problem with H, is not that it makes false predictions, but that it makes none at all. Hi is consistent 'vith 'vhat vve observe, no matter 'vhat our observations turn out to be. Features that are useful are consistent 'vith H2, as are features that are neutral and features that are harmful. Scientists expect scientific theories to be testable, which means that theo- ries should make predictions that can be checked against observations. H,, it appears, is not testable. This defect can also be remedied. Consider, for example, a third version of creationism:
H3: Organisms did not evolve. Rather, God created each species separately and endowed them with the very characteristics they would have had if they had evolved by natural selection.
H3 is a wild card; it makes the same predictions that evolutionary theory makes. If so, what reason can there be to choose behveen these hvo theories?
Predictive Equivalence
Evolutionary theory and H3 are predictively equivalent. If evolutionary theory pre- dicts that life will have a particular feature, so does H,. Although imperfect adap- tations disconfirm H,, they are perfectly consistent with H,. Does this mean that
Chapter 6: Evolution and Creationisn1 79
'.' ''.' '' ''' '' '< '''.' •• '.'''. '' •••• ' •• ''.'.' ••• '' '''' ''''' '' ''.' ''. '''' '.
evolutionary theory is not well supported? I would ~ay not. Consider the following pair of hypotheses:
/,: You are now looking at a printed page.
/,: You are now looking at a salami.
You have excellent evidence that/, is true and that/, is false./, predicts that you are having particular sensory experiences; if ]1 is true, you should be having certain visual, tactile, and gustatory sensations (please take a bite of this page). /, makes different predictions about these 1natters. The sensory experiences you are no"' having strongly favor J1 over fr Now, however, let's introduce a wild card. What evidence do you have that/, as opposed to J3 is true:
/ 3: There is no printed page in front of you, but someone is now systematically misleading you into thinking that there is a printed page in front of you.
f, and /3 are predictively equivalent. The experiences you are having now tell you that/, is more plausible than/,, but they don't strongly favor/, over /3• In the section of this book that focuses on Descartes's Meditations (Chapter 13), the problem of choosing between/, and /3 will be examined in detail. For now, what I want you to see is this: when you ask whether some hypothesis is strongly supported by the evidence, you must ask yourself what the alternatives are against which the hypothesis is to be compared. If you compare f, with/,, you'll conclude that/, is extremely well supported. However, the problem takes on a different character if you compare /, with /3• The point applies to the competition between evolutionary theory and creationism. When you compare evolutionary theory '\Vith creationism, everything depends on '\vhich version of creationism you consider. Son1e versions nlake false predictions, son1e make no predictions, and son1e "piggy~back" on evolutionary theory, relying on that theory's ability to make predictions and then using those predictions.
Prediction versus Accommodation
A fourth version of creationism raises interesting questions. Suppose '\Ve examine the human eye in detail and observe that it has a set of features that we'll call F. We can then construct the follo,ving version of creationism:
H4: An intelligent designer made the human eye and gave it the set of features F.
Does H4 make predictions? Well, it entails that we should observe that the eye has features F. In fact, no theory can do a better job than H
4 does of fitting what we observe, as H4
hits the nail precisely on the head. The problem is that H4 merely accommodates the observations; it does not provide novel predictions. It is easy-too easy-to construct hypotheses such as H,. Whenever you observe that 0 is true, you just construct the hypothesis that "an intelligent designer wanted 0 to be true and had the means to bring this about:' If such hypotheses were satisfactory, there would be no need to do real science; we could shut down all of the costly research now underway and just invoke this tidy formula.
So Part Two: Philosophy of Religion
. '''''''.'.'''''' ', '''' '' ''' '', ''''''''','.'' '''''''' ''' '''''''.'' ''' ''' This suggests that we should supplement the two rules for abduction presented in
Chapter 3. In addition to the Surprise Principle and the Only Game in Town Fallacy, we should require that a good theory 1nakes predictions that \Vere not used in the construction of that theory. This is the requirement that theories shouldn't merely accommodate what we observe after the fact; in addition, they should make some predictions that are novel.
Does Evolutionary Theory Make Novel Predictions?
What is sauce for the goose is sauce for the gander. If \\'e de1nand that creationis1n 1nake novel predictions, we should demand that evolutionary theory do the same. By "prediction;' I do not restrict n1yself to saying \Vhat \vill happen in the future. A prediction can involve observations that have not yet been made that reflect events that took place in the past. For example, the part of evolutionary theory that says that organis1ns possess co1n1non ancestors predicts that we should find intermediate fossil forms. Biology says that whales and cows have a common ancestor, so there should be fossils that have characteristics that are "in between:' And if birds and dinosaurs have a con1n1on ancestor, there should be fossils that are intennediate here as \\'ell. Dar,vin \Vorried about "the incon1pleteness of the fossil record:) but there have been numerous fossil finds that fulfill the predictions of the hypothesis of common ancestry.
What about the other part of Darwhis theory, the hypothesis of natural selection? What does it predict? Here it is itnportant to realize that there are 1nany detailed theories in evo- lutionary theory that predict which traits will be favored by natural selection in which cir- cumstances. For exa1nple) in hu1nan beings) slightly 1nore boys than girls are born. In other species, there is a more extre1ne 1nale bias, or a fetnale bias, in the sex ratio at birth. Evolution- ary theory provides theories that predict ,vhen a species should evolve one sex ratio and 'vhen it should evolve another. Another exa1nple concerns the nature of infectious disease. Suppose you are infected with a disease that is spread through the air; if you get sick, you will take to your bed and you will spread the disease less frequently. Compare this with a disease like di- arrhea that is spread through feces. Even if you take to your bed, your feces will be taken from your roo1n and sent into the environ1nent, just as if you \Vere \veil. No'v think about these hvo patterns fron1 the point of vie\v of the organism that causes the disease. In the case of an air- borne disease, selection favors infections that are n1ild over ones that are severe; in the case of a "'aterborne disease, selection favors just the opposite pattern. The vector of the disease-the mechanism that spreads the disease from infected to uninfected individuals-allows biolo- gists to predict what traits the disease should evolve. These evolutionary ideas have recently been important in biological thinking about AIDS-HIV; see Paul Ewald's The Evolution of Infectious Disease (Oxford University Press, 1993) for discussion.
When I say that evolutionary theory makes predictions, I do not mean that it is now in a position to predict the whole detailed future of the planet's biosphere, or that every detail of the past can be predicted. There are many open questions in evolutionary biology, as there are in any science. Ho\vever, the theory has 1nade impressive strides since 1859. The same is not true of creationism. Paley compared the random hypothesis with the hypoth- esis of intelligent design and argued that the latter was better supported; Behe compares evolutionary theory to the hypothesis of intelligent design and draws the same conclusion. Creationists, then and now, criticize the theories that they dislike. But they have not devel- oped a positive theory of their own.
Chapter 6: Evolution and Creationism
Concluding Remarks
Creationisn1 comes in lnany forms. Son1e of then1 make very definite predictions about what we observe. The version that says that God made organisms so that they are per· fectly adapted to their environments makes predictions that do not accord with what we observe. Young earth creationism, which says that the earth is ten thousand years old, also rnakes predictions that conflict i;vith \Vhat scientists observe. A third version of creation- ism says that God made organisms to look exactly as they would if they had evolved by the lnindless process of natural selection; this niakes the same predictions that evolutionary theory makes, and so our observations do not allo\v us to discriminate bet\veen evolu- tionary theory and this "mimicking" or "piggy-backing" version of creationism. Finally, the bare, minimalistic version of creationism that says that God had some (unspecified) impact on the traits of living things is, I suggest, untestable. We have not found a version of creationism that Jnakes definite predictions about i;vhat i;ve observe and that is better supported by the observations than evolutionary theory is. What is wanted here is not just a version of creationism that can accotnmodate the observations after i;ve have made then1, but one that tells us what we will observe before we make those observations. Is there a version of creationism that can do this?
Review Questions
l. \Vhat are the two main elements of Darwin's theory?
2. Describe what the Principle of the Common Cause says. How is this principle re!ated to the Surprise Principle? How is it used by biologists to decide whether different species have a common ancestor?
J. The geneticist Fran~ois Jacob said (in "Evolution and Tinkering," Science Vol. i96, 1977, pp. n61-1166) that "natural selection does not work as an engineer \Vorks. It works like a tinkerer-a tinkerer who does not know exactly \vhat he is going to produce but uses whatever he finds around him." \Vhat does Jacob mean here? How is this point relevant to evaluating vvhether the hypothesis of evolution or the hypothesis of intelligent design is a more plausible explanation of the characteristics of living things?
4. \Vhat does it mean to say that two theories are predictively equivalent? Can the design hypothesis be formulated so that the existence of imperfect adaptations isn't evidence against it?
Problems for Further Thought
1. Louis Pasteur (1822-1895) developed scientific evidence against the hypothesis of "spontaneous generation." For example, he argued that maggots developing on rotten meat aren't the result oflife springing spontaneously from nonliving materials; the maggots were hatched from eggs laid there by their parents. Does Pasteur's discovery mean life couldn't have evolved from nonliving materials?
2. Suppose you are a crew member on the starship Enterprise, bound for a new planet. You know there is intelligent life there; the question you want to ans\ver is whether these life forms have ever had any contact with Earth. Which sorts of observations would be relevant and which irrelevant to this question? Defend your interpretation. How does the Principle of the Common Cause apply to this problem?
3- Compare what Paley says about the watch v./1th what Behe says about irreducible complexity.
81
82 Part Two: Philosophy of Religion
'' ''''.'', . ''.' ''' '.'' ' .• ''.' ''' ''.' ''' ''.'. <'''''.' .. ''.' ''. ''' '' '. ,, '
MySearchLab Connections
Watch. Listen. Explore. Read. Visit MySearchLab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
RESEARCH AND EXPLORE
:¥,~EXPLORE
1. Creationism 2. Charles Darwin
3. Irreducible complexity
References
Charles Darwin (1809-1882), "The Ongin of Species", 1859.
Can Science Explain Everything?
Chapter Outline
Scientific Ignorance
The Only Game in Town Fallacy
The Two Questions
What Is a Scientific Explanation?
A Thesis about Explanation
Why Is There Something Rather than Nothing?
Can Physics Explain the Origin of the Universe?
Leibniz: God Chooses Which Possible World to Actualize
Clarke: God Explains Why the Actual World Consists of One Total History Rather than Another
The Only Game in Town Fallacy, Again
Causality
The Principle of Sufficient Reason
In the previous chapter, we failed to identify a version of the design hypothesis that not only makes predictions, but also makes novel predictions, about the features we should observe in living things. It is a further question whether there are versions of creationism that do better than the versions I described. If there are not, then creationism should be discarded as an explanation of the characteristics that organisms possess. But this result would not show that the design hypothesis has no role to play at all. Perhaps there are other features of the universe, distinct from the ones discussed in biology, that we should explain by postulating the existence of an intelligent designer.
Here is another way to pose this problem: A naturalistic explanation seeks to explain features of the world by describing the processes in nature that produced them. A super- naturalistic explanation, on the other hand, attempts to explain features of the world by describing the supernatural processes (the processes that occur outside of nature) that
chapter 7
Part Two: Philosophy of Religion
'.'' ''''' ''','' '''' ''' ''' < ' ''' ','.' ''' ''''. ' •• ''''
produced them. Is there reason to think that every feature of the natural world has a natu- ralistic explanation? If so, the hypothesis of a supernatural God isn't needed to explain anything.
The question I want to explore in this chapter is this: should we expect that science will sooner or later explain everything that's true about the world we inhabit? That is, will science sooner or later be able to explain everything and do this without needing to postu- late the existence of a supernatural God?
Scientific Ignorance
It is easy to find facts about the world that science can't explain now. Every scientific dis- cipline has its open questions. These are the things that keep present-day scientists busy. Scientists don't spend their time repeating to each other things everybody already knows; rather, they devote their energies to trying to puzzle out ans\vers to heretofore unans\vered questions.
Because there are plenty of facts about the world that science can't explain now, it is easy to construct a particular kind of abductive argument for the existence of God. I mentioned in Chapter 6 that evolutionary theory is currently unable to explain why some species reproduce sexually while others do not. Consider the following theological explanation for this puzzling fact: there is a God, and God decided to make organisms reproduce in precisely the way they do. Whereas evolutionists caiit (now) explain pat- terns of reproduction, a theologian can do this just by saying "it was God's will:' I doubt that any scientist would argue that patterns of reproduction will never be explained scientifically. 'fa be sure, evolutionists cannot no1v explain certain facts about sex. But science isn't over yet; it is reasonable to suspect that this example of scientific ignorance is only temporary.
"BUT HOW DO YOU EXPLAIN GOD?"
Sometimes people object to the suggestion that the existence of God is the explana- tion of a fact by asking, "But how do you explain the existence of God?" The implica- tion behind this question is that it isn't plausible to say that God explains something unless you are able to explain why God exists.
This criticism of abductive arguments for the existence of God isn't a good one. Mendel was right to think that the existence of genes explains something. However, Mendel had no idea how to explain the fact that genes exist. Here are two more ex- amples that exhibit the same pattern. For hundreds of years before Newton's time, sailors had good evidence that the tides are correlated with the phases of the Moon, even though they had no very plausible explanation for why this should be so. As noted in the Chapter 6, evolutionists have excellent evidence that human beings are descended from apelike ancestors, even though the explanation of this evolutionary event is still somewhat unclear. My conclusion is that if abductive arguments for the existence of God are defective, this isn't because no one knows how to explain why God exists.
Chapter 7: Can Science Explain Everything?
''' ''',,''' ''.,'' '' ''.'' ',"'''
The Only Game in Town Fallacy
What should we say now? Apparently, we now have two choices: We can accept the theo- logical explanation sketched above, or \Ve can adn1it that \Ve have no explanation at all for \vhy son1e species, but not others, reproduce sexually. I think it is clear that \Ve aren't obliged to accept the theological explanation. To think we are required to accept the theo- logical explanation just because it is the only one on the table is to commit the Only Game in Town Fallacy.
Recall the anecdote from Chapter 3. If you and I are sitting in a cabin and hear noises in the attic, it is easy to formulate an explanation of those noises. I suggest the noises are due to gremlins bowling. This hypothesis has the property that if it were true, it would explain why we heard the noises. But this fact about the gremlin hypothesis doesn't mean that the hypothesis is plausible.
Instead of accepting a theological explanation of why organisms reproduce as they do, we should consider the option of simply admitting that we at present don't understand why this is so. If my prediction about the future of evolutionary theory is correct, we will eventually have a satisfactory scientific explanation. We just have to be patient.
I no\v \vant to consider a very different kind of scientific ignorance. Suppose there are facts about the world that science in principle ca1it explain. It isn't a temporary gap, but a permanent one, that I no\v \Vant to consider. If there are such facts, then the choice \Vo1lt be behveen a theological explanation and a naturalistic explanation that science \\1ill even- tually produce; the choice will be between a theological explanation and no explanation at all-not no\v and not in the future.
The Two Questions
There are two questions I want to consider: (1) Are there any facts about the world that science is inherently incapable of explaining? (2) If there are, can we plausibly argue that the best explanation of why those facts are true is that God exists? To answer the first of these questions, I'll sketch a picture of how scientific explanation works. I then will argue that this vie\V of scientific explanation implies that there are certain facts about the \Vorld that science ca1it ever explain. It is1it just that scientists have1it worked out the explana- tion yet; the point is that the nature of scientific explanation precludes their ever doing so.
What Is a Scientific Explanation?
There are t\VO kinds of things science aitns to explain. It aitns to explain particular events, and it aims to explain generalizations. An example of the first type of problem would be the astronon1er's question, ''Why did the solar system come into existence?" An exa1nple of the second type would be the chemist's question, "Why does it take two atoms of hydrogen and one of oxygen for hydrogen and oxygen to combine to form a molecule of water?" Don't confuse these hvo explanatory projects. Generalizations are true or false. Events aren't true or false; they happen or fail to happen. An event happens at a certain place and thne. True generalizations, ho,vever, describe \vhat is true of all places and thnes. I \Vant to focus here on \Vhat scientists do \Vhen they explain events.
86 Part Two: Philosophy of Religion
'.''. '''.' '.'''' ... ' .... ''.' ....... ' .. ' .... '. ' ......... ' ... , ' .... ' '.'' '' .. My thesis is that science explains an event by describing its causes. I'll assume further
(although this isn't strictly necessary for what I want to conclude) that a cause must pre- cede its effects. So the causal explanation of an event E, that happens now is provided by citing one or more events in the past and showing how those events produced E .. Sup- pose we show that E, was caused by E,. We then might be interested in explaining why E, occurred. This might lead us to describe a still earlier event, E3, and so on. This project leads us to describe causal chains that trace back into the past. You'll recall that Aquinas (Chapter 4) believed that causal chains extending back into the past can have only a finite number of links. I make no assumption about this.
So far I've described two relationships that can obtain between events: (1) some events precede others in time and (2) some events cause others. There is a third relationship. Some events are parts of other events. Consider the relationship between the storming of the Bastille and the French Revolution. The storming of the Bastille marked the beginning of the French Revolution. Both of these are events; the first is part of the second.
I want to focus on what science does when it tries to explain a spatiotemporal event. By "spatiotemporal;' I simply mean that the event in question happened at some place and at so1ne titne. Of course, lots of events are "spread out:' They don't occur in an instant, but take some length of time to occur. And they take place in a volume of space, not at a single spatial point. For example, the French Revolution lasted about ten years, and it took place in France.
A Thesis about Explanation
My thesis is this: For science to be able to say why a spatiotemporal event E occurred, there must be events in space and time that aren't included in E. I assume that no event explains itself. So if science is to explain E, there must be some event outside of E (but still inside of space and time) that can be cited as the explanation.
It follows from this thesis that answerable questions about why this or that spatiotem- poral event occurred must be about events that are part of what happens in the history of the spatiotemporal universe. Such questions can't be answered if they ask why the whole history of the spatiotemporal universe was as it was. Consider an exhaustive description of the past, present, and future of what happens in space and time:
Past Present Future
E, is explicable because there is something besides E, in the world that can be cited to do the explaining. If E, is part of some bigger event (the way th.e storming of the Bastille is part of the French Revolution), then that more inclusive event will be explicable only if there is something outside of it that we can cite by way of explanation. Such answerable questions about spatiotemporal events I'll call local why-questions. They focus on part of what has happened in the world's history. In contrast to them are global why-questions- questions that ask for an explanation of the totality of what has happened in the whole universe's history. I claim science can't answer global why-questions.
Chapter 7: Can Science Explain Everything?
Why Is There Something Rather than Nothing?
Here's an example of a global why-question: Why is there something rather than nothing? This question doesn't ask why there is something now. That, of course, is a local question, 'vhich can be ans,vered by saying that the universe 'vas none111pty in the past, and then perhaps citing a conservation law (as Aquinas did-see Chapter 4) that says that if the world is nonempty at one time, it will be nonempty forever after. What I have in mind here is the question of 'vhy the universe ever contained anything. That is, 'vhy 'vasn't the universe entirely devoid of material objects throughout its history?
The spatiotemporal universe is the totality of everything that was, is, or will be, any,vhere. There is nothing itnpossible about a universe that contains no nlaterial objects. Obviously the actual world isn't this way. The actual world (the one we inhabit) corre- sponds to one total history. An empty universe corresponds to a different total history. Both of these totalities are possible, but only one is actual. The question of why there is something rather than nothing asks why the first of these possible worlds, rather than the second, is actual. My claim is that science can't ans\ver this question.
Can Physics Explain the Origin of the Universe?
Don't be misled by the fact that physicists talk about explaining the "origin of the universe:' This makes it sound as though they are explaining why the universe is non- empty. My view is that they are doing no such thing. Rather they are addressing a different question. The Big Bang Theory doesn't show why the universe, in my sense of the 'vord, can1e into existence. The universe, according to the \Vay I a111 using the term, includes everything that there ever was, is, or will be. If the Big Bang produced what came after it, then the universe includes two stages: First, there was a Big Bang and then there was what happened next. The Big Bang doesn't explain why the whole universe exists; the Big Bang is part of the universe. The scientific question addressed by physi- cists is local, not global.
So the question "Why this totality, rather than that one?" isn't scientifically answerable because it is global. Can a theological answer be provided? Let's imagine that God isn't a material object; he (or she or it) is outside the spatiotemporal totality I'm calling "the mate- rial universe:' Can 've explain 'vhy the material universe is nonen1pty by saying that God decided to make one totality actual rather than another?
Leibniz: God Chooses Which Possible World to Actualize
The seventeenth-century philosopher and scientist Gottfried Leibniz (1646-1716) thought this form of explanation is plausible. He said that God considered the set of all possible worlds and decided to make one of them actual. He did this by finding which world is best. This is why we live in the best of all possible worlds. If you think the world we inhabit is mor- ally imperfect, you may doubt Leibniz's story. In Chapter 11, I'll consider this issue when I take up the problem of evil. However, the point of importance here is that Leibniz had another kind of perfection in mind in addition to moral perfection. Leibniz thought this is the best of all possible worlds in the sense that this world has the maximum diversity of
88
CE]-{ Read Theodicy on
ww1,v.mysearch!ab.com
Part Two: Philosophy of Religion
phenomena and the maximum simplicity of laws. Scientific laws are simple, but the kinds of events that happen in our world are enormously rich. This joint property of our world is 'vhat n1akes it perfect in this non1noral sense.
I won't discuss the details of Leibniz's theory. My point is that he proposed a theo- logical explanation for why there is something rather than nothing in the world of space and time. This fact about the world, being global, can't be scientifically explained. The question I want to pose is this: If a fact can be explained by the hypothesis that there is a God, but can't be explained scientifically, is this a good enough reason to conclude there is a God?
Clarke: God Explains Why the Actual World Consists of One Total History Rather than Another
Another anticipation of the argument for the existence of God '"e are considering is due to the British philosopher Samuel Clarke (1675-1729). Clarke thought he saw a way to improve an argument that Aquinas proposed. As noted in Chapter 4, Aqui- nas's second argument for the existence of God depends on the universe's being finitely old. Clarke's argument, presented in his 1705 book A Demonstration of the Being and Attributes of God, does not depend on that assumption. He asks us to consider the entire totality of events that occur in the world of space and time. Why did this history occur rather than some other? Clarke says that there are two possibilities: The entire history occurred for no reason at all, or it was brought into existence by a being who exists outside of space and time, namely, God. Clarke rejected the first suggestion as patently absurd; there must be a reason why the world is comprised of one total his- tory rather than another. That reason is God. Clarke's argument is an instance of the cosmological argument.
If science is incapable of explaining why the world of space and time includes one total history rather than another, should we accept the claim that God exists because this hypothesis, if true, would explain this global fact? In similar fashion, should we accept the claim that God exists because it is able to explain why there is something rather than noth- ing (a fact that science is unable to explain)?
The Only Game in Town Fallacy, Again
I think the answer to these questions is 110. Once again, the Only Game in Town Fallacy needs to be considered. There is a second option besides accepting the only explanation available. This is to admit that there are things about the universe we don't understand and never will. Why there is something rather than nothing may be inherently inexplicable; the same may be true of the question of why the world contains one total set of events rather than another. I grant that if(1) there were an all-powerful God who exists outside of space and time and if(z) a God outside of the material world could nonetheless create a material world and if (3) God wanted to make the world nonempty, then this would explain why there is something rather than nothing in the spatiotemporal universe. These suppositions would explain the fact in question, but that doesn't show that this explanatory story is at all plausible.
Chapter 7: Can Science Explain Everything?
Causality
To see \Vhat is problematic about this suggested explanation> consider the causal relation- ship it says exists between God and the totality of the material universe. God is outside of space and tilne, but nonetheless 1nanages to bring the \Vhole 1naterial \Vorld into existence. This causal relationship is extremely hard to understand.
'l'hink of the causal relationships \Ve discuss in science and in everyday life. For example, we say that throwing the rock at the window caused the window to break. Here a causal relationship is said to obtain behveen t\VO events, each of \vhich occurred at a particular place and at a particular titne. This is a characteristic of\\1ell-understood clahns about causality; in this case, causation is a relationship that obtains between events that occur within space and time. This is why it is so puzzling to say that a God outside of space and time causes the world of space and time to come into being.
I think it is misleading to use the statement "God did it" as a substitute for the state- ment "! really can't explain why this or that fact is true:' The first statement advances an explanation, while the second admits to not having one. If the existence of God is to explain something, it must have more going for it than the fact that the observation under consideration is other\vise inexplicable.
I said before that evolutionary theory currently cannot explain why some species reproduce sexually \vhile others do1it. This doesn't 1nean evolutionists can't make vague pronounce- ments such as "the pattern is due to natural selection:· This vague re1nark is available to anyone \\'ho \Vants to make it. No scientist, ho\vever, \VOtdd regard this single sentence as a satisfying explanation. The question-the main question-would be how natural selec- tion managed to produce the results it did. What made sexual reproduction advantageous in so1ne species and asexual reproduction advantageous in others? It is this tnore detailed question that> to date, has not been ans\vered satisfactorily. There is an analogy here \Vith the theological case. The theologian can say, "God did it:' The evolutionist can say, "natu- ral selection did it:' A scientist would regard the latter explanation as unsatisfying because it is too short on details. Why should the theological explanation be taken any more seri- ously? Part of what is missing in the theological story is any indication of how or why God did what he did. This may help explain why the theological explanation of patterns of reproduction should not be embraced. A theological explanation so short on details leaves too nlany questions unans\vered to be convincing. In this circun1stance, perhaps it is more plausible to admit we don't understand than to pretend that "God did it" gives us much by way of understanding.
A similar diagnosis can be offered of why the theological explanation of why there is something rather than nothing is so unsatisfying. To say "God did it" is to leave too much unexplained. Ho\v is it possible for something outside of space and time to cause the entire spatiotemporal universe to come into being? Why should God have preferred to actualize this possible world rather than any of the others? To the degree that we don't understand these matters, we should conclude that the theological explanation is really not much of an explanation at all.
My conclusion is that science may not be able to explain everything. There are facts no\v kt10\vn to be true that science has not been able to account for. And even if science progresses, as I expect it to do, there \vill rernain facts that are intrinsically resistant to
go Part T\'10: Philosophy of Religion
'''.' ''' .. >.'.'. '',.'' '.' 0 •• '''' ''',.'. ''' ''' '.' '' '''' '''''
scientific treat1nent. Not only has science not explained the1n yet, but it never \vill> because it can't. Of course, we should be open to the possibility that explanations from outside of
science may have some plausibility. My point here is that bringing in the existence of God
is often to trade one inystery for another, "'Nith no net gain in our understanding.
The Principle of Sufficient Reason
Clarke's argument makes use of an idea that has long been influential in philosophy. It is called The Principle of Sufficient Reason. This principle asserts that evel')1hing that hap- pens in nature has an explanation; nothing happens for no reason at all. What can be said
in favor of this principle? Scientists seen1 to assume it, or son1ething like it, \Vhen they address a new problem. Consider some body of facts that science is currently not able to explain. For example, what is it about the organization of the human brain that allows us to
have conscious experience? Neuroscientists assun1e that there is an ans'iver to this question. This is why they set themselves the task of finding a scientific explanation of conscious-
ness. But 'vhat assurance do they have that such an explanation exists? Maybe it 1nakes sense for scientists to assun1e that consciousness has an explanation
even if they have no evidence at all that such an explanation exists. After all, if they look
for an explanation, they may find one (if such an explanation exists). But if they don't even
look for an explanation, they will certainly fail to find it, even if there is one to be had. This is a strategic argun1ent for its being sensible to look for an explanation of consciousness. It is not an argun1ent that offers evidence that such an explanation exists. The fact that scien-
tists search for explanations 'vithout a guarantee that they 'vill succeed is not evidence that
the Principle of Sufficient Reason is true.
Review Questions
i. \Vhat is the difference between a global \vhy-question and a local \vhy-question?
2. Can science answer global \vhy-questions? Why or why not?
}. Is it possible to have evidence that something exists \Vithout having an explanation for why the thing exists?
4. Consider the following argument: "If science can't explain some fact that a theological story can explain, then that fact is strong abductive evidence that the theological story is true." How is the Only Game in Town Fallacy relevant to assessing this argument?
5. What does it mean to say that causation is a relationship that obtains bet\veen events in space and time? If this were a correct daim about causation, what implications v,rould it have for the idea that God explains the sorts of global facts discussed in this chapter?
Problems for Further Thought
l. Philosophers sometimes use the term brute fact to describe a fact that is so simple and obvi- ous that it can't be justified or explained by appeal to any other fact. Consider the relationship between x and y described in the statement "x = y." This is the relationship of identity. This is a relationship that everything bears to itself; "the Eiffel Tower~ the Eiffel Tower" is true, but "the Eiffel Tower~ the Statue of Liberty" is false. What explains the fact that everything is self-identical? Is this true because of the way we use the identity sign"="? !sit a brute fact?
Chapter 7: Can Science Explain Everything?
'' ',. '"'.,'.' ........... '. '. 2. In this chapter, 1 claimed that nothing explains itself. In everyday speech, however, we talk about
something's being "self-explanatory." What does this mean? Is there a conflict bet\veen what I've said about explanation and the everyday idea? \Ve also use the term "self-starter" to describe some people. Does this confiict \Vith the idea that nothing causes itself?
). Blaise Pascal, a seventeenth-century philosopher whom I'll discuss in Chapter 10, once said, "The truth of religion lies in its very obscurity." What does this mean, and is it true?
4. I argued in this chapter that science can't explain everything. In particular, it can't explain why there is something rather than nothing. The physicist Lawrence Krauss just published a book called A Universe from Nothing-Why There Is Something Rather than Nothing (Free Press) in which he claims to do what I say is impossible. The philosopher David Albert reviewed the book in the New York Times (Sunday, March 25, 2012). Read Albert's review and think about hovv Krauss's argument (as described by Albert) relates to the argument presented in this chapter. The revie\v can be found at: http://www.nytimes.com/2012/03/25/books/review/a-universe-from-nothing- by-!awrence-m-krauss.html? _r=1&scp=1&sq=vid%2oalbert%2obook%2orevie\v&st=cse.
Supplementary Reading
GOTTFRIED LEIBNIZ
Theodicy
In this selection from Theodicy, Leibniz responds to numerous arguments purporting to show that an all good, all powerful God did create the universe. In his response to Objection VIII, Leibniz argues that there are multiple possible \vorlds that God could have made actual, but because God's will "is always infallible and always tends toward the best," the actual world is the best of all possible \Vorlds.
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
RESEARCH AND EXPLORE
:)\c{EXPLORE
1. Scientific explanation 2. Principle of sufficient reason
[E]-{ Read Theodicy on
wvA'l.mysearchlab.com
chapter 8 ..............................
92
The Ontological Argument
Chapter Outline
A Posteriori and A Priori
Definitions and Existence
Ansehn's Argu1nent
Gaunilo's Criticisn1
Anselm's Reply
Dispensing with Perfection
Conclusion
The various argu1nents I've considered so far for the existence of God include pre1nises we know to be true by experience. By using our senses (of sight, hearing, touch, etc.), we know that objects are in motion and that organisms are intricate and adapted to their environments. What is n1ore, there is no other \Vay to kno'v these things; sensory expe- rience is indispensable. Such propositions are called a posteriori. Notice the prefix post, n1eaning "after.)) An a posteriori truth is a truth that requires experience to be kno,vn (or justified). Accordingly, the arguments for the existence of God considered thus far are called a posteriori arguments.
A Posteriori and A Priori
Are all propositions a posteriori? Many philosophers have thought there are propositions that can be known to be true by reason alone. These propositions are termed a priori. Note the word "prior"; such propositions are knowable prior to, or independently of, ex- perience. Standard examples are mathematical truths and definitions. We can know that 7 + 3 ~ 10 and that bachelors are unmarried merely by thinking about the concepts in- volved. You don't have to do experiments or conduct surveys to find out whether such statements are true. Reason suffices.
Chapter 8: The Ontological Argument 93
'''.''''''''.' '' ''.'' '' ''' '''' ''''' '' ''' '' '' '' ''''' ' .. '' ''' .. ' ' .... ' '' ... Let's be clear on \Vhat it 1neans to clailn that a given proposition is a priori. To say
that \Ve kno\v a priori that bachelors are un1narried doesn't 1nean \Ve \Vere born \vith the concepts of bachelorhood and marriage. For a proposition to be a priori, it isn't essential that the concepts contained in the proposition be innate. Maybe none of us is born \vith an understanding of the concept of a bachelorj nlaybe \Ve need to have various experiences if \Ve are to acquire that concept. Even if the concepts \Vere acquired through experience, however, that wouldn't show the proposition fails to be a priori. To say that a proposition is a priori true n1eans that IF you understand the concepts involved, then reason \vill suffice for you to decide that the proposition is true. Don't forget the IF in this idea. The idea that there is a priori kno\vledge doesn't require that there be innate concepts.
The Ontological Argument for the existence of God is an a priori argument. It aims to establish the truth of theism without reliance on any a posteriori premise. In particular, the idea is that we can prove that God must exist simply by examining the definition of the concept of God.
Definitions and Existence
This should strike you as a very surprising thing to try to do. Consider other definitions. For example, a bachelor is by definition an adult unmarried male. The thing to notice about this definition is that it doesn't imply there are any bachelors. The same point seems to be true of nlost definitions. You can define \Vhat it is to be a unicorn or a golden moun- tain, but the definitions \von't settle whether there are such things. When \\'e consider con- cepts such as bachelo1; unicorn, or golden 11101111tai11, the definitions specify what it would take to be an individual of the kind in question; the definitions don't say whether there are zero, one, or many things of that kind. The Ontological Argument claims the concept of God is different: from the definition of the concept of God, the existence of God is sup- posed to follow.
So the proposition that bachelors are unmarried is a priori, whereas the proposition that bachelors exist is a posteriori. The definition doesn't imply the existence claim. Are there any nontheistic exa1nples of existence claims that can be kt10\vn a priori? Perhaps existence clain1s in 1nathen1atics furnish exan1ples. If mathe1natics is a priori, then math- ernatical existence clain1s are a priori. So it is an a priori truth that there exists a prin1e number between 12 and 14. It is philosophically debatable whether there are any a priori truths at all. I wmit discuss that issue here, however, but will go along with the idea that 1nathe1natical truths are a priori. This rneans that since tnathe1natics contains existence claims, there are some existence staten1ents that are a priori true. It follo\VS that if there is a mistake in the Ontological Argument for the existence of God, it isn't to be found in the fact that the argu1nent tnaintains that so1ne existence claims are a priori true.
Anselm's Argument
The Ontological Argument was formulated by Anselm (1033-1109), who became archbishop of Canterbury. Aquinas, who lived in the thirteenth century, rejected this argument and rec- ommended in its stead the five a posteriori argu1nents that \Ve exa1nined in Chapters 4 and 5.
94 Patt Two: Philosophy of Religion
... '' ...... ' ... ,' ..... ' ..... ' .. ' ....... ' ... ''. '.'.',''' '' .. ' . '' ''.'.' '.'.
t:EJ-[ Read uGuanilo and
Anselm: Debaten on W\'1111.mysearch!ab.com
Philosophers down to the present (including Descartes, whose Meditations on First Philosophy we'll discuss in Part II) have discussed the Ontological Argument and have constructed differ- ent versions of it. Here I'll consider a simplified formulation of the argument:
(0)
(1) God is by definition the greatest being possible.
(2) A being who fails to exist in the actual world (while existing in other possible worlds) is less perfect than a being who exists in all possible worlds.
Hence, God exists, necessarily.
I label the argument "O" for future reference.
Premise (1): Conceivability and Possibility Premise (1) is intended to capture what Anselm meant by saying that God is a being "than which none greater can be conceived:' It isn't just that God is the most perfect being that happens to exist in the actual world. The actual world, after all, might contain only grossly imperfect beings, so that God's being the best of these wouldn't mean he is all-powerful, all-knowing, and perfectly good. The idea is that God is a being who is the best possible being. There is no more perfect being in any possible world. (If you're a bit vague on the idea of a "possible world;' go back to Chapter 4 and review.)
Although I gloss Anselm's first premise by saying that God is the best possible being, I want to note an important difference between possibility and conceivability. What is pos- sible is an objective question; it doesn't depend on what people know or believe. Whether 've can conceive of so1nething, ho\'fever, is a fact about us as kI10,ving subjects. Hence, con- ceivability is a subjective notion. (This should remind you of the point made in Chapter 4 that necessity and certainty are different.)
What can be conceived of may vary from individual to individual. We can conceive of things that dogs can't conceive of; perhaps there are superintelligent beings else- where in the universe who can conceive of things that are beyond our grasp. In con- trast, there is a sense of the term possible, which I am using here, according to which what is possible doesn't vary from one thinker to another.
It is impossible, according to Einstein's special theory of relativity, for an object that has positive mass to go faster than the speed oflight. This simply can't be done, no matter how powerful a rocket ship you build. Yet, I think this is something we can conceive of; we can describe what it would mean to go faster than the speed oflight. Newton thought that travel faster than the speed of light was possible. He not only could conceive of what this would be, he thought it was possible. We now view Newton as mistaken. It isn't possible to do this, and it never has been. For something to be possible, it isn't enough that you can conceive of its happening or that you can't see any reason why it wouldn't happen.
If the human mind is limited in various ways, then maybe the most perfect being we can conceive of is still grossly imperfect in many 'vays. Maybe an ant can't conceive of a creature greater than an anteater, but that doesn't mean that anteaters are the most perfect possible beings. Anselm, though he used the word "conceivable;' really meant possible, I think. God isn't just the best being we frail human beings can conceptualize; God is supposed to be the best possible being.
Chapter 8: The Ontological Argument 95
' .. ' .. ' ''. ''. ' ..... ' ' .... '.'' ' ........... ' .. '.' ............ '' .... ' ....... . Premise (2): Necessary Existence Is a Perfection
If a being exists in all possible worlds, it is a necessary being, not a contingent one. Why does Anselm think God's perfection demands that God be a necessary being? A contin- gent thing depends for its existence on other things. You and I are contingent things; we exist only because certain other contingent beings acted as they did. But a truly perfect being, Anselm thinks, wouldn't depend for its existence on anything contingent. It would exist unconditionally. According to Ansehn's second premise, a necessary being is more perfect than a contingent one. So if God has all the perfections, he must be a necessary being, not a contingent one.
The Ontological Argument looks like it is deductively valid: if God has all the per- fections and if necessary existence is a perfection, then God exists necessarily. But from Anselm's time down to the present, there have been doubters. Many philosophers (both theists and atheists) have thought that the Ontological Argument must involve a mistake.
Gaunilo's Criticism
Anselm describes a criticism formulated by his contemporary, Gaunilo. Gaunilo says the argument must be defective, since if it were not, you could prove the existence of a perfect island by a similar a priori argument. What did Gaunilo have in mind?
Let's start with the concept of an island. The first point is that the concept of islandhood resembles the concept of bachelorhood. Its definition doesn't deductively imply that there are any objects of the kind defined. So any attempt to provide an a priori proof that islands exist must fail. Now let's define a new concept: Call it a P-island. AP-island is an island than which no greater island can be conceived. P-islands are just like ordinary islands, except that they are, by definition, the best possible islands. Gaunilo thoughtthere can be no a priori demonstration that there are P-islands. Perhaps no such thing exists, and if there happen to be P-islands, this isn't something that could be discovered by reason alone.
A PRIORI/A POSTERIORI AND NECESSITY/CONTINGENCY
In Chapter 4, the ideas of necessity and contingency were introduced. A possible world, roughly, is a way the world might be. A necessary truth is a proposition that is true in all possible worlds; a contingent truth is a proposition that is true in the actual world, but not in all possible worlds.
In this chapter, the ideas of a priori and a posteriori propositions were introduced. An a priori proposition can be known (or justified) independent of experience. How is the distinction between necessary and contingent propositions related to the distinction between a priori and a posteriori propositions?
This is a deep, controversial, and fundamental philosophical question. Some philosophers hold that the two distinctions divide up the propositions in exactly the same way. That is, they hold that all necessary propositions are a priori and that all contingent propositions are a posteriori. According to this theory, all the propositions that natural science discovers by observation are contingent.
96 Part Two: Philosophy of Religion
••••••••••••••••••••••••••• ' ••••••••••• ' ' ••• '.' ••• '.' '.'' •••••• '' t "''''.'.
Some of what I've said in previous chapters has assumed that this theory is false. I've suggested that empirical science (that is, science that relies on observation) is able to discover what is really possible. In Chapter 4, I mentioned some examples. I said that thermodynamics has discovered that perpetual motion machines are impossible (that is, it is a necessary truth that there are no perpetual motion machines); I also stated that Einstein's theories of relativity say that no particle (with positive rest mass) can go faster than the speed of light (that is, that objects with positive rest mass necessarily go slower than the speed of light). In other words, I've suggested that some a posteriori propositions are necessary, even though other a posteriori propositions are contingent.
What about a priori propositions? The examples discussed so far of a priori propo· sitions (such as "bachelors are unmarried" and 0 2 + 3 = 511 ) are all cases of necessary truths. So the question remains: Are there any a priori propositions that are contin- gent? The examples I'll give provide a preview of material that will be discussed in Chapter i3. Consider the proposition each of us formulates when we say to ourselves "I exist." You know this proposition to be true without using sight, hearing, taste, touch, or smell. Thought alone allows you to see that it is true. So it is a priori. The same seems to be true for the proposition you formulate by saying to yourself "I am thinking." Yet, it isn't a necessary truth that you exist. You might have failed to do so. If this line of thinking is right, we may conclude that the proposition each of us consid- ers when we say to ourselves "I exist" or "I am thinking" is a priori and contingent.
Fill in the following table with examples of the four types of propositions indicated:
Necessary Contingent
A prioli
A posteriori
Consider, however, the following argument, which mimics the structure of the Onto- logical Argument:
(I)
(1) AP-island is by definition the greatest possible island.
(2) An island that fails to exist in the actual world (while existing in other possible worlds) is less perfect than an island that exists in all possible worlds.
Hence, a P-island exists necessarily.
I label this argument ('I" for future reference. Gaunilo's conclusion wasdt that there are P-islands (as well as P-bagels, P-comedians,
and so on). Rather, he thought this argument about islands must be defective. It seems crazy to think the existence of a perfect island could be established a priori. Gaunilo con- cluded that Anselm's argument must have a mistake in it as well. What sort of mistake did Gaunilo think these two arguments commit? He believed that the two premises of the Island Argument are true. But Gaunilo holds that these premises don't deductively imply
Chapter&: The Ontological Argument 97
'' ... '.' '' ... ' ... ''.'.'. '.' ....... ' .... ' ........ ' .......... ' .. ''.'.'.'' .. ' the conclusion; the existence of a perfect island doesn't follow from these a priori premises. Gaunilo concludes that the Island Argument {I) is invalid. He takes this to show that the Ontological Argument (0) is also invalid.
How Are the Ontological Argument and the Island Argument Related?
Notice that Gaunilds argument was obtained from Anselm's by substituting "P-island" for "God" and "island". for "being:' We saw in Chapter 2 that the validity of an argument depends on its form, not on its subject matter. Gaunilo was suggesting that since his Is- land Argument is invalid, so too is Anselm's argument about God, since they have the same logical form. Gaunilo thought the Island Conclusion does1it follow from the Island Premises. By parity ofreasoning, he claims that Anselms conclusion doesn't follow from Anselm's premises.
Gaunilo is making an argument about these two arguments. In particular, he attempts to show that Anselnis argument is invalid by reductio (see Chapter 4 for this piece ofter- minology). Gaunilo argues as follows:
If the Ontological Argument is deductively valid, then the Island Argument is deductively valid.
The Island Argument is not deductively valid.
The Ontological Argument is not deductively valid.
Anselm's Reply
Anselm scornfully replies to this criticism. He says that his argument applies to God-the greatest possible being. It doesn't apply to a P-island-the greatest possible island. His idea is that islands are not the kind of thing that could have all the perfections. For example, islands can't have the perfection of intelligence. I think this criticism of Gaunilo misses the mark. Islands can have some perfections (shady places, beautiful beaches, nice drinking water, tropical fruit), even if they can't have all of them. AP-island is an island that is the most perfect possible island; this does not require that it be intelligent. As long as existence is one of the perfections that an island can have, Gaunilo's point seems correct.
Even so, Gaunilo's objection is, to n1y 1nind, incomplete. His reductio argument, if it 'vorks, sho,vs that there has to be a mistake some\vhere in Anselm's reasoning, but it doesn't show exactly where the mistake occurs. As noted before, Anselms argument looks valid; if it isn't, we would like to see exactly where it goes wrong.
When Anselm dismisses Gaunilo's Island Argument, maybe he means that the Ontologi- cal Argument is deductively valid, while the Island Argument is not. However, this sugges- tion seems implausible, since deductive validity doesn't depend on an argument's subject matter, only on its form, and the two arguments seem to have the same logical form. Or maybe Anselm is claiming the premises of the Island Argument are implausible, though the premises of the Ontological Argument are quite plausible. If this were correct, the two argu- ments would be disanalogous. But the first premise of the Ontological Argument and the first premise of the Island Argument seem to be on a par. Anselm holds that his own definition of God is correct; it is hard to see how Gaunilo's definition ofa P-island could be mistaken.
gB Part Two: Philosophy of Religion
What about the second premises of the hvo argu1nents? If necessary existence is a per- fection for beings in general, why isn't it a perfection for islands in particular? I feel a little shaky about my ability to make such judgments about "perfection:' but I don't see why the second premise of the Ontological Argument should be accepted and the second premise of the Island Argument rejected. I conclude that the two arguments are fully parallel, contrary to '\Vhat Ansehn says in reply to Gaunilo. Gaunilo, I think, is on to so1nething.
Dispensing with Perfection
Anselm's argument (0) makes use of the idea that God is perfect. Gaunilo's parody of Anselm-Gaunilo's Island Argument (!)-also makes use of that concept. Gaunilo and Anselm shared a worldview according to which everything in the universe occupies a par- ticular position on the Great Chain of Being. At the top is God, next come the angels, then human beings, then animals, and then plants. Anselm and Gaunilo believed it was an entirely objective question how perfect a given thing is.
I want to suggest that we can see a defect in the Ontological Argument more clearly once "\Ve set aside the concept of perfection. Let's not '\vorry about inferring a being's ex- istence from its being perfect. Let's not worry about how perfect a particular thing is or about whether perfection is objective. Let's simply build the property of existence into the definition of a being and see what happens.
Unicorns don't exist. The concept of unicornhood isn't exemplified in the actual world. Consider, hovvever, a nevv concept, 'vhich I call E-unicornhood. An £-unicorn is by def- inition so1nething that is an existing unicorn. Does it fo1lo"f fron1 the definition of an E-unicorn that there are E-unicorns? Not at all. The concept includes the property of exis- tence, but that does1it imply that the concept is actually exemplified. The definition simply describes what a thing must be like if it is to count as an E-unicorn. The definition tells us that if something is an E-unicorn, then that thing will have the property of existing. Now let's define the concept of an E-God. AnE-God is a being who is all-powerful, all-knowing, all-good, and who actually exists. Does it follow from the definition of an E-God that there is an E-God? Not at all. The definition simply describes what a thing must be like if it is to count as an E-God. The definition tells us that if something is an E-God, then that thing will have the property of existing.
Conclusion
When we define God, we are saying what properties a being must have to be God. Parallel remarks apply to defining unicorns or golden mountains. Anselm argued that our concept of God has built into it the idea that God necessarily exists. Just as God is by definition omnipotent, it also is true that God, by definition, necessarily exists. My criticism of the Ontological Argument comes to this: let's grant that God is by definition the greatest pos- sible being. And let us grant that necessary existence is a perfection. What follows from this is that necessary existence is part of the concept of God. Necessary existence is built into the concept of God, just as omnipotence is. But the fact that existence is built into a concept doesn't imply there are things to which the concept applies.
Chapter 8: The Ontological Argu1nent 99
'. '''' ' ••••• '.' '.' ',' ••• '.'. ''. '' •• '.'., '''.' •• '' ''' ' •• ' ' '' 0' ' ''' '".'.''
If we concede that God is defined as an omnipotent being, all this means is that if God exists, then that being must be omnipotent. Likewise, if we concede that God is defined as a necessarily existent being, we are merely saying that if there is a God, then that being necessarily exists. The definitions dorit entail the existence of anything that is omnipotent nor do they entail the existence of anything that necessarily exists. It is for this reason that the Ontological Argument is invalid.
Review Questions
l. \Vhat is the difference between an a posteriori proposition and an a priori proposition? \Vhat is the difference between saying that a proposition is a priori and saying that it is innate? Give an example of an a posteriori existence claim and an example of an a priori existence claim.
2. \Vhat is the difference between saying that a proposition is possible and saying that it is conceivable?
J. Hovi does Gauni!o's point about islands bear on Anselm's argument about God? \Vhat is a reductio argument? Describe hov./ Gaunilo uses this form of argumentation.
4. Suppose God is, by definition, a being that exists in all possible worlds. Does it follow that God exists?
Problems for Further Thought
l. Consult the original text of Anselm's argument and Gaunilo's reply. Give a formulation of the argument that is more true to the text than the one presented in this chapter.
2. I've argued that building the idea of existence into a concept doesn't ensure that anything actu- ally exemplifies the concept. That is the point I made by considering E-unicorns. Novv, however, consider the opposite problem. !s it possible to define a concept so that the definition ensures that nothing can exemplify it? Construct a definition that does this. How is it possible that defini- tions can have negative implications about existence, but not positive implications?
3. Do the following arguments have the same logical form? Explain your ansv1er.
God has every perfection. Existence is a perfection.
God exists.
Sam has every novel by Dickens. David Copperfield is a novel by Dickens.
Sam has David Copperfield.
4. Suppose I define God as a necessarily existing being. Does it follow that God exists? If 1 assume that it is possible that God exists, vvould it then fol!ow that God exists? If so, is this a convincing proof of God's existence?
5. In the selection from Hume's Dialogues Concerning Natural Religion, Philo, Demea, and Clean- thes discuss a priori and a posteriori arguments for the existence of God. What are their views on whether there can be an a priori proof that God exists?
IE-[ Read Part 2 of Dialogue
Concerning Nat!ual Religion on wv1w .mysearchlab.com
100 Part T\'lo: Philosophy of Religion
•• '' ••• '' ••••• ' ••• ' ••••••••••••••• ''. '".' •••• ''.'.' •• < ••• ' ' ••• ' •• '.'. '' •• '
MySearchlab Connections
Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
t:EJ-{ READ AND REVIEW
READ
Guanilo and Anselm, "Debate"
One of Guani/o's criticisms of Anselm's argument is that if it establishes the existence of God, then the existence of a perfect island would be provable a priori. Does this criticism tell us what is wrong with Anselm's argument? If not, what does it show?
RESEARCH AND EXPLORE
~MEXPLORE
1, Ontological argument 2. Anselm
chapter 9 ................. -............ .
Is the Existence of God Testable?
Chapter Outline
Logical Positivism
The Testability Theory of Meaning
Analyticity
Falsifiability
Auxiliary Assumptions Needed
Auxiliary Assumptions Must Be Independently Established
"God Exists" Is Meaningful
St11nn1ary
Up to this point, I've taken the question of whether God exists to be perfectly meaningful. It may be difficult to answer, but I've assumed that there is an answer. Theists, atheists, and agnostics agree on this. I've assumed the sentence "God exists" says something; the ques- tion is '\Vhether '\vhat it says is true. and '\Vhether '\Ve can kno'\v this.
Logical Positivism
Some philosophers-advocates of the position called logical positivism-have an altogether different view of this problem. They hold that the sentence "God exists" is meaningless. It isn't true and it isn't false either. They hold that it is misguided to argue about whether God exists. It isn't that there is a true ans'\ver to the question "Does God exist?" but '\ve'll never discover what that a true answer is. Rather, positivists hold that the question has no factual answer at all. Their idea is that we have been misled by a superficial grammatical sin1ilarity behveen the sentence "God exists" and the sentence "Genes exist:' The latter expresses a scientific statement that can be justified or rejected by appeal to evidence. In contrast, the for1ner sentence expresses no proposition at all. Positivistn maintains that '\Ve shouldn't try to answer the question of whether God exists, but should reject it as mean- ingless. According to positivism, theists, atheists, and agnostics are all mistaken.
101
102 Part Two: Philosophy of Religion
The Testability Theory of Meaning
Positivists advanced a theory that is supposed to justify this philosophical thesis. This is the Testability Theory of Mea11i11g. It holds that for a sentence to be meaningful, it must be either a priori or a posteriori. The Testability Theory of Meaning also includes specific proposals for how these two categories should be understood.
As explained in Chapter 8, a truth is a priori if it can be justified without appeal to sense experience-it can be justified by reason alone. And a truth is a posteriori if sense experience is needed for it to be justifiably believed. This is what it means for a truth to be a priori or a posteriori. Falsehoods can fall into these categories as well. A mathematical truth like "2 + 3 = s" can be justified by reason alone, but the mathematical falsehood "2 + 4 = i' can be disproved by reason alone. Just as the statement "There is a book in front of you no\v" is an a posteriori truth, so the state1nent "There is no book in front of you now" is an a posteriori falsehood.
Some n1eaningful statements are true; others are false. In saying that each meaning- ful sentence is either a priori or a posteriori, positivists are saying that every meaningful sentence is decidable i11 pri11ciple. That is, they maintain that if Sis a meaningful sentence, then it is possible to figure out whether Sis true. If there is no way in principle to figure out whether Sis true, positivists conclude that Sis meaningless.
I novv turn to the positivists' account of \vhat makes a sentence a priori or a posteriori. Once each of these categories is clarified, they think the claim that God exists can be seen to fall into neither category. Positivists conclude the claim is meaningless.
Analyticity
Positivists (and some other philosophers besides) hold that all a priori statements are a11alytic. An analytic truth (as opposed to a synthetic one) is a definition or a deductive consequence of a definition. The truth of "All bachelors are unmarried" follows from the definitions of the terms that occur in it. "Either it is raining or it is not raining" sitnilarly follo,vs from the definitions of ''or1' and <'not.1' Symmetrically, positivism maintains that all a posteriori statements are synthetic-if a sentence depends on sense experience for its justification, its truth does1/t follow from definitions alone.
Where does the sentence "God exists" fit into this classification? Positivists reject the whole strategy behind the Ontological Argument. They hold that "God exists" isn't decid- able from definitions alone. So the sentence isn't a priori. This means that if"God exists" is 1neanlngful, it 1nust be a posteriori. We n1ust no\v examine the account that positivists give of \Yhat 1nakes a sentence a posteriori.
Falsifiability
Positivists say that the sentence "God exists" is 1111testab/e. There is no way to use observa- tions to decide whether it is true. Defenders of the Testability Theory of Meaning have argued that sentences are tested by deducing predictions from them that can be checked by making observations. Mendel tested his genetic theory by deducing from it a predic- tion about the proportions of different characteristics that there should be in the pea
Chapter 9: Is the Existence of God Testable? 103
''.' .. "' .. '.' ....... ' ... ' ' .. ' '.' ....... '.' .. ' ... ' ......... '.' "' .. ' ....... ' plants he grew in his garden. He then was able to check these predictions observationally. Positivists argue that the sentence <(God exists" n1akes no predictions \Vhatsoever, so it is untestable.
Sometimes the idea of observational testability is developed with reference to the con- cept of falsification. If a theory makes a prediction that can be checked observationally, the theory runs the risk of being refuted by observations. This is what the influential phi- losopher of science Karl Popper (not a positivist himself) meant by saying that a scientific theory should be falsifiable. This doesn't mean a scientific theory should be false. Falsifi- able doesn't mean false. What is crucial is that the theory rules out at least some possible observations. It n1ust {(stick its neck out:'
In science, confirmation and disconfirmation are intimately connected. An experi- ment has different possible outcomes. If some outcomes would be evidence in favor of the hypothesis, then other outcomes would be evidence against it. If you propose an experi- ment for testing a hypothesis and you think that every possible result would indicate the hypothesis is true, there is no reason to run the experiment. The only way the experi- ment could offer you real evidence in favor of the hypothesis is if there were some pos- sible results that would undermine the hypothesis. You then get evidence supporting the hypothesis if you run the experiment and find that those possible refuting outcomes don't come to pass. Getting evidence in favor of a hypothesis requires that the hypothesis rule out so1nething. But what does the sentence "God exists,, rule out? Is there anything \Ve can imagine observing that would be evidence against theism? For example, if we observed that there are many moral evils in the world, would that count against the hypothesis that there is a God? In Chapter 11, I'll discuss this question in some detail. For now, I'll merely note that theists have usually not taken the existence of evil to count against theism. They say that if the world had been morally perfect, that would be consistent with the existence of God; if the world happens to contain moral evil, that too can be reconciled with theism. It seems that theists usually can reconcile their belief in God with just about anything they observe.
Consider the doomsday pronouncements that pop up from time to time in different religions. Every so often, some religious group gets the idea that God is about to bring the world to an end. Such predictions have yet to come true. What has been the result of these failed predictions? A few end-of-the-worlders have had their faith shaken. But, after some reflection, most theists have decided that the failed predictions of the end- of-the-worlders don't count against the hypothesis that there is a God. If the world had come to an end, theists would have taken this to support their belief that there is a God; but when the world fails to come to an end, these same people continue to believe that God exists.
I hope these remarks convey the flavor of what positivists are getting at when they claim that the hypothesis that God exists is untestable. I now want to be a little more pre- cise about these ideas. I begin with two definitions:
A hypothesis His falsifiable if and only if H deductively implies at least one observation statement 0.
An observation statement is a statement whose truth or falsehood can be determined by direct observation. Notice that if a hypothesis is falsifiable, it can be refuted
104 Part Two: Philosophy of Religion
by a valid deductive argument if the observational prediction it n1akes turns out to be false:
If H, then 0.
Not-0
Not-H
Auxiliary Assumptions Needed
I agree that the hypothesis that God exists isn't falsifiable. The same is true, however, of most of the hypotheses that science investigates. Hypotheses rarely deductively imply pre- dictions all by themselves. Rather, the usual pattern is that you must conjoin auxiliary assumptions A to a hypothesis H to get H to deductively imply an observational prediction 0. It isn't H that implies 0, but H&A that does the trick.
I'll illustrate this idea with an example from outside science. Consider Sherlock Holmes again. Holmes wishes to test the hypothesis that Moriarty is the murderer. Does this hypothesis, all by itself, tell Holmes what he will observe at the scene of the crime? Not at all. Auxiliary as- sumptions are needed. IfHolines ki10\vs that Moriarty\\'ears a size 12 shoe, smokes El Supren10 cigars, and fires a gun of type X, these auxiliary assumptions may lead him to expect to find cer- tain clues. If he discovers a size 12 footprint, an El Supren10 butt on the carpet next to the victitn, and a bullet from a type X gun, his auxiliary assumptions may lead hin1 to think that Moriarty is probably the murderer. On the other hand, if Holmes were to make other auxiliary assump- tions, then the hypothesis that Moriarty is the murderer would lead Holmes to make quite dif- ferent predictions. If Holmes thought that Moriarty didn't smoke, that he has a size 8 shoe, and that he prefern to use a knife, then the hypothesis that Moriarty is the murderer would not pre- dict the clues Holmes found. The clues found at the scene of the crime don't say anything one way or the other about whethet Moriarty is the murderer until auxiliary assumptions are added.
So most of the hypotheses we test don't, all by themselves, deductively imply observa- tionally checkable statements. They do so only when supplemented with auxiliary assump- tions. Where does this leave the thesis that the hypothesis that God exists isn't testable?
1'he sentence "God exists" is not falsifiable, but that is uninteresting, since most scientific hypotheses aren't falsifiable either. This brings us to the next question: Is the hypothesis that God exists testable when it is supplemented with auxiliary assumptions? The answer is yes. Consider the following candidates:
A,: If there is a God, there will be living things in the world and they will be perfectly adapted to their environments.
A,: If there is a God, there will be no living things. A3: If there is a God, there will be living things in the world, but they will not be
perfectly adapted to their environments.
The hypothesis that there is a God, when conjoined with A,, predicts something we can check by making observations. The same is true if we conjoin the theistic hypothesis with A, or with A,. So there is a twofold parallel between the hypothesis that God exists and a scientific hypothesis (or a detective's hypothesis): The hypothesis isn't falsifiable, but it is testable when conjoined with auxiliary assumptions. Does this mean that the hypothesis that God exists is no less testable than the hypotheses of science?
Chapter 9: Is the Existence of God Testable?
There does seem to be a difference between these cases, although it isn't to be char- acterized in terms of the idea of falsifiability. In the case of Holmes testing whether Moriarty is the murderer, I noted that different sets of auxiliary assumptions entail dif- ferent predictions about what Holmes will find at the scene of the crime. If A4 is true, we get one set of predictions; if A5 is true, we get a very different set.
A4 : Moriarty s1nokes El Supren10 cigars, has a size 12 shoe, and uses a type X gun.
A5: Moriarty does not s1noke, has a size 8 shoe, and uses a knife as his preferred tnurder \Veapon.
If Holmes is to say whether the clues at the scene of the crime count for or against the hypothesis that Moriarty is the murdere1; he must decide which of A
4 and A
5 is true.
NEPTUNE AND VULCAN
Most hypotheses in science don't entail predictions all by themselves; they need to be conjoined with auxiliary assumptions to do so. This means that ifthe prediction turns out to be false, scientists face a choice. They can reject the hypothesis, or they can reject the auxiliary assumptions.
Two episodes in nineteenth·century astronomy illustrate these two options. Working separately, John Adams and Urbain Leverrier used Newtonian physics plus assumptions about the planets then known to exist to predict where the orbit of Uranus should be. These assumptions included the idea that Uranus is the planet farthest from the sun. The prediction about Uranus' orbit obtained by conjoining Newton's theory with this as· sumption came out wrong. Adams and Leverrier therefore had to choose between reject· ing Newtonian physics and rejecting one or more of the auxiliary assumptions. Because Newtonian theory was well supported by other observations, they were reluctant to reject it. They therefore suspected that the auxiliary assumptions were the culprit. They conjec· tured that Uranus isn't the last planet. Two years later, astronomers were able to confirm this conjecture; they observed the planet that we now call Neptune.
This same pattern of reasoning was subsequently applied to the planet Mercury. Newtonian physics plus auxiliary assumptions (including the assumption that there is no planet between Mercury and the sun) predict where Mercury's orbit should be. This prediction turned out to be false. Scientists therefore faced a choice. Either Newton's theory is wrong or at least one auxiliary assumption is mistaken. Following the pattern of reasoning that worked so well for the case of Uranus, some scientists conjectured that there is a planet lying between Mercury and the Sun. It was given the name "Vulcan."
As it turned out, this conjecture was mistaken. There is no such planet. It wasn't the auxiliary assumption (that there is no planet between Mercury and the Sun) that was responsible for the false prediction about Mercury's orbit; it was Newtonian physics itself that led to the error. Only when Newton's theories were replaced by Einstein's theories of relativity could the departure of Mercury's orbit from Newtonian values be explained.
In the reasoning about Uranus and the reasoning about Mercury, three sorts of state· ment play a role. There are theories (T), auxiliary assumptions (A}, and predictions (P). Describe the logical form of the arguments that the scientists formulated in the two cases.
105
106 Part T\vo: Philosophy of Religion
.. '' .. ' ... ' ...... ' .. ' .. ' ... '"' '''.'''' ''' ', '. '' '"'' , ',''' ''"'' '''.' Auxiliary Assumptions Must Be Independently Established
In the scientific case, as 'veil as in the case of detective i;vork, there is an itnportant fact about auxiliary assumptions that goes beyond what I've said so far. Not only are auxiliary assumptions needed if the hypothesis is to make observational predictions, in addition, those auxiliary assumptions must be checkable independent of the hypothesis under test. For example, Hohnes can interpret the clues found at the scene of the critne once he decides 'vhether A4 or A5 is true. Further1nore, he can find out \vhether A4 or A5 is true '\Vithout his having to already know if Moriarty is the murderer.
Can the same be done for auxiliary hypotheses about God's nature, such as A., A,, or A,? This isn't so clear. It is easy enough to invent auxiliary hypotheses that reconcile what we observe with the hypothesis that God is responsible for what we observe. If we find that an organism is perfectly adapted to its environ1nent, \\'e can construct an auxil- iary assumption that effects this reconciliation. If we find that the organism is imperfectly adapted, a different auxiliary assumption can be invented that does the same thing. The problem is: How do we figure out which of these auxiliary assumptions is true without as- suming at the outset that God is responsible for what we observe? This is the problem we ran into in Chapter 5 ,vhen \Ve considered different versions of creationis1n.
In the detective case, Holmes would1it endlessly revise his opinions about auxiliary as- sumptions just to preserve his faith that Moriarty must be the murderer. 'lb do so would be to fall prey to an obsession, not to do good detective work. Holmes would find out which of A 4 and A, is true, and then use these independently confirmed auxiliary assumptions to test the hypothesis that Moriarty is the murderer. The result may be that Holmes is forced to conclude Moriarty is probably innocent of the crime. It is difficult to test the hypothesis that God exists because it is hard to determine which auxiliary assumptions to accept if you don't already believe that God exists.
How is what I'm saying different from what the Testability Theory of Meaning main- tains? To begin with, I'm not using the idea of falsifiability to criticize the hypothesis that God exists. Nor am I claiming that no one will ever figure out how the hypothesis that God exists can be tested. Perhaps someday auxiliary assumptions will be independently discovered that can tell us what the hypothesis that God exists predicts about the observ- able world. My claim is just that this has1it happened yet. In summary, positivists used the Testability Theory of Meaning to try to derive a timeless verdict concerning the sentence "God exists:' In contrast, I've used a different idea about testability (one that emphasizes the role of auxilia1y assumptions) to argue for a verdict concerning how the issue of test- ability stands now.
"God Exists" Is Meaningful
I disagree with positivism on another count. I think the hypothesis that there is a God is meaningful. It is possible to discuss whether there is evidence for or against it. Also, the hypothesis obviously has various logical properties; I noted before that it, in conjunction '\Vith various auxiliary statements, deductively implies various predictions. This is enough to show that the sentence "God exists" isn't meaningless gibberish. A meaningful state- ment need not be decidable in principle.
Chapter 9: Is the Existence of God Testable?
Summary
I'll summarize my assessment of the Testability Theory of Meaning as follows. I think there are difficulties in testing the hypothesis that God exists. This, however, isn't because I en- dorse the criterion of falsifiability; I don't. The bare statement "God exists" does not entail observational predictions, but neither does the bare state1nent «electrons exist:' The typical situation in science is that theoretical staten1ents have observational consequences only 'vhen they are conjoined 'vith auxiliary assu1nptions. It is easy to invent auxiliary assun1p- tions that permit the hypothesis that God exists to make predictions. It is also easy to invent such assumptions so that the result is that the hypothesis makes true predictions. What is not so easy is to show that the auxiliary assumptions that permit the theistic hypothesis to 1nake predictions are the1nselves 'vell confirn1ed. Scientific testing requires that the auxil- iary assun1ptions used be 'vell supported by evidence; inventing them isn't enough.
Review Questions
L What does it mean to say that a hypothesis is falsifiable? Does this mean it is false? Is the hypothesis that organisms contain genes falsifiable?
2. What is an auxiliary assumption? How does testing the hypothesis that God exists depend on a choice of auxiliary assumptions?
3- \Vhat does lt mean to say that Holmes could independently confirm the auxiliary assumptions he needs concerning Moriarty? Independent of what?
4. How does the Surprise Principle explain why the existence of miracles \vould be strong evidence for the existence of God? In what way are auxiliary assumptions needed if one wants to interpret miracles as evidence favoring the hypothesis that God exists?
Problems for Further Thought
1. Consider the follo\ving passage from David Hume's "Essay on Miracles":
That no testimony is sufficient to establish a miracle unless the testimony be of such a kind, that its falsehood would be more miraculous than the fact vvhich it endeavors to establish .... When any one tells me that he saw a dead man restored to life, I immediately consider with myself whether it be more probable that this person should either deceive or be deceived, or that the fact which he related should really have happened. I weigh the one miracle against the other, and according to the superiority \Vhich I discover, I pronounce my decision, and always reject the greater miracle. lf the falsehood of his testimony would be more miraculous than the event which he relates, then, and only then, can he pretend to commend my belief or opinion.
What general point about eyewitness testimony is Hume making here? Does Hume's argument show that miracles never happen?
2. Can positivists say that a sentence is meaningful even though vie will never actually find out \Vhether it is true? Can they say that a sentence is meaningful ifit is impossible for us to discover whether it is true? Does this mean that sentences about the prehistoric past are meaningless?
3. The readings for this chapter include a selection from A. j. Ayer's Language, Truth, and Logic (1936). This was a very infiuential defense of logical positivism. Jn this selection, Ayer says that
108 Part Two: Philosophy of Religion
''. ' .. ' .... ' .... ' ............. ' .......................... '' .... '.' '. '' .. .
CEJ-[ Read "Of Miracles,"
Section 10 of An Enquiry Concerning Human Understanding on www.mysearchlab .com
there is no conflict between science and religion. He also says that positivism agrees v1ith \vhat many religious mystics maintain. Why does Ayer hold these views?
4. The concept of falsifiability was defined in this chapter. Define the parallel concept of verifiability. What problems are there \Vith the claim that statements that are not a priori must be verifiable?
5. Is the Testability Theory of Meaning meaningful according to the standards it specifies? That is, is the theory either a consequence of definitions or testable via sense experience?
Supplementary Reading
DAVID HUME
Of Miracles
In the i8th century climate in which Hume was working, miracles were often taken to provide pow- erful evidence for the truth of certain religions. In this selection from An Enquiry Concerning Human Understanding, Hume develops a skeptical position concerning miracles. Jn reading this passage, think about what evidence, if any, would persuade you that a miracle had occurred.
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
CEJ-[ READ AND REVIEW
READ
1. David Hume, "Of Miracles," Section 10 of An Enquiry Concerning Human Understanding What general point about eyewitness testif".lony is Hume making here? Does Hume's argument show that miracles never happen?
RESEARCH AND EXPLORE
:)!.-(EXPLORE
1. Logical positivism 2. Falsifiability 3. Testability
References
Alfred Jules Ayer, "Religious Discourse is Meaningless," from Language, Truth and Logic (New York: Dover Books, i936), pp. i14-120. Reprinted with permission.
chapter 10 ..............................
Pascal and Irrationality
Chapter Outline
Prudential and Evidential Reasons for Belief
When Does It Make Sense to Gamble?
Pascal's Argun1ent
First Criticis111 of Pascal's Argu111ent
Second Criticis1n of Pascal's Argu111ent
The Role of Reason
Freud's Psychological Explanation of Theism
A New Prudential Argument
Prag1natis111
Blaise Pascal (1623-1662), a French physicist, philosopher, mathematician, and mystic, was interested in the question of whether it could be rational to believe in God even if you think it is enormously improbable that God exists. Suppose all the evidence is against God's existence-you regard the existence of God as a possibility, although you think it is extremely implausible. Can it still make sense to be a theist? Pascal was one of the found- ers of the modern mathematical theory of probability. His argument has come to be called Pascals Wager, because Pascal argues that believing that God exists is a sensible wager, even if there is no evidence that God exists.
Prudential and Evidential Reasons for Belief
Before analyzing Pascal's argun1ent, I \Vant to describe hvo kinds of reasons sorneone tnight have for believing some proposition or for performing some action. Suppose one day we nieet for lunch. I'm carrying a n1achine gun and a briefcase. I open the briefcase and sho\v you that it contains $1,000,000. I show you that the gun is loaded. Then, I make you an offer that I think you can't refuse. You probably do not believe the president of the United States is juggling candy bars at this very moment. You regard this proposition as very improbable,
r'r.1 r Read "The lA:Jl- \Vager," Number 233 in Section 3 of Pensies on www.mysearchlab .com
109
110 Part Two: Philosophy of Religion
'.'.' '.'' '''.' ••• ' ••• ' •• '''"' '''' ''. '','' • '"'.'''.''' ''' '' '<'''' '' 0''. '''
though you grant that it isn't entirely impossible. Here is my offer: If you can get yourself to believe that the president is juggling candy bars right now, I'll give you $1,000,000. If you don't believe this proposition, I'll shoot you with the machine gun. What is it reasonable for you to do, supposing that you think I'm entirely sincere in my threats and promises?
In one sense, it is rational to try to believe, even though you think the proposition in question is very improbable. It would be prudent (that is, in your self-interest) to try to believe, supposing that you don't want to die and would like to have the million dollars. I'll call this kind of reason a prudential reason.
In another sense, however, you don't have any reason to believe that the president of the United States is juggling candy bars right now. You have no evidence at all that the proposition is true. Let's describe this fact by saying that you don't have an evidential rea- son for believing the proposition in question.
Pascal's Wager is intended to provide you with a prudential reason for believing that God exists even if you think there is no evidence that there is a God. Even if you think all the evidence is against the existence of God, Pascal thinks he can give you a prudential reason for theism, so long as you grant that the existence of God is at least possible.
When Does It Make Sense to Gamble?
Before analyzing Pascal's argument, I need to mention one detail about gambling. Here is a fact about reasonable gambles: It can make sense to bet on improbable outcomes if the payoff is big enough. Suppose I offer you the following wager. If you pay me $1, you will be allowed to draw a single card from a standard deck of cards. If the card drawn is the ace of spades, I'll give you $1,000,000. If it is some other card, you receive nothing; you simply forfeit the dollar you paid to play. Question: Does it make sense to buy into this gamble, assuming you can trust me to pay you if you win and that your goal is to maximize the amount of money you possess? The answer is yes. Although the probability of winning is only 1/52, the payoff if you win is so huge and the cost to you if you lose is so small that this is a great gamble to make. Bear this fact in mind as you consider Pascal's argument.
THE EXPECTED VALUE OF AN ACTION
You have to decide between two actions: buying a lottery ticket and not buying one. If you buy the ticket, a card is drawn at random from a standard deck of cards. If the card is the ace of spades, you win $1,000,000. If not, you lose. The ticket costs $i. Should you buy the ticket, if your only goal is to maximize your cash resources?
The expected value (expected utility) of an action is the average payoff you would receive if you performed the action again and again. If you buy the ticket, there is a 51/52 chance that the ace of spades will not appear; in this case, your resources will have dropped by $1 (the cost of the ticket). On the other hand, if you buy the ticket, there is a 1/52 chance that you will win the million; in this case, your resources (taking account of the cost of the ticket) will have increased by $999,999. So the expected value of buying the ticket is
($1)(51/52) + ($999,999)(1/52) ~ $19,230
Chapter lo: Pascal and Irrationality 111
............. ' ......... ' ...... ' ... ' .. '' ..... ' ...... ' .. '' .. ,''' ... '' '' '.' '. What is the expected value of the alternative action-of not buying the ticket? That depends on what you would do with the $1 in question. Suppose that you would simply put it in your pocket if you didn't buy into the lottery. In this case, you can be certain that your resources will neither increase nor decline. So the expected value of not buy- ing a ticket is $0. Notice that buying the ticket has a higher expected value than not buying. So the rule in decision theory tells you to buy the ticket because this is how you maximize your expected utility (or value).
The expected value of buying the ticket is not the amount of money you can expect to get. If you enter the lottery, you'll either end up a dollar behind or $g99,999 ahead. The expected value is what you would average if you bought into the lottery again and again. Most of the time, you'll lose; occasionally, you'll win. The average payoff per ticket over the long run is about $19,230.
The mathematical idea of an expectation also applies to quantities other than value or utility. A probability theorist might say that the expected number of offspring for parents in the United States today is about 2.2. This doesn't mean that a couple can expect to have precisely this number of children. Rather, 2.2 is the expected num- ber because it is the average. The same is true when life insurance companies de· scribe the life expectancy of a person who is now nineteen years old.
Pascal's Argument
Pascal's Wager can be represented as a solution to a problem in decision theory. This is a modern theory that describes how an agent should choose among different available ac- tions on the basis of the utilities of different possible outcomes. Be sure to get clear on how each of these italicized terms figures in Pascal's argument.
You have to decide whether or not to believe there is a God. These are the two pos- sible actions-believing in God or not believing. Separate from what you believe, there is the way the world is-either there is a God or there isn't. Hence, there are four possible outcomes-an outcome being composed of your performing an action when the world is a particular way. The four outcomes are represented in the following two-by-two table:
Actions
Believe that God exists
Do not believe that God exists
Ways the World Might Be God exists God does not exist
+= -10
-= +10
The upper-left entry represents the outcome in which you believe in God and there is a God. What is the result, if this should come to pass? Pascal says that the payoff to you is
112 Part T\vo: Philosophy of Religion
'' '''''' '''' '''.''''''.' '''' '.'' '''' ''''',''''' '''''' '' '''' '''' ''' '' ''''
an infinite reward-you go to heaven. Suppose, on the other hand, that you occupy the lower-left cell. This is the outcome in which you don't believe there is a God, but God in fact exists. Here Pascal says you suffer an infinite punishn1ent~you receive eternal dam- nation. What about the other two entries? If you believe in God, but there is no God, per- haps you will have wasted your time with religious rituals and observances you might have preferred to skip had you only known there is no God listening to your prayers. This isn't a horrible result, though it should be represented as a modest disutility-hence my entry of -10. On the other hand, if you don't believe in God and there is no God (the lower-right cell), you receive a modest benefit. Instead of engaging in religious practices that you may find boring at times, you can get involved in activities that are inore re,varding. Pascal says that the entry in this cell represents a modest benefit; my chosen value for the utility here is +io.
To sununarize the structure of the problem: you 1nust decide ,vhich action to perfor1n. The possible outcomes are the possible pairings of an available action with a different way the world might be. With each possible outcome, you associate a utility (a payoff) that represents how good or bad the outcome is for you.
How should the utilities of the different possible outcomes be assessed to reach a de- cision about which act is preferable? The expected utility of believing that there is a God needs to take account of two possibilities: There is the utility you receive if you believe and there is a God, and the quite different utility you receive if you believe and there is no God. Similarly, the expected utility of not believing must take account of both outcome possi- bilities (represented in the bottom row). Here is Pascal's solution. Even if you think the ex- istence of God is very improbable (say, you assign it a probability of only 1/1,000,000 ), the expected utility of believing is higher than the expected utility of not believing. The reason is that, although the existence of God is very improbable, there is a huge benefit if you are a theist and God exists, but only a small cost if you are a theist and there is no God. It makes sense to bet on the existence of God, even if the existence of God is very improbable, for reasons that parallel the gamble I mentioned before. The gamble makes sense even though it is improbable that there is a God.
First Criticism of Pascal's Argument
Tvvo criticisms have been made of this 'vager) but only one of them is decisive. First, some philosophers have objected to treating belief as an action. We can't decide to believe a proposition; belief isn't an action \Ve control in this \vay. In 1ny story about the president, you may recognize that it is in your interest to believe that the president is now juggling candy bars. Believing this would be good for you, but that doesn't mean you can automati- cally make yourself believe the proposition. The objection to Pascal's argument is that it is a inistake to treat belief as a problen1 in decision theory, since decision theory is about choosing actions that you can perfor1n if you \Vant to.
Pascal recognized this difficulty. After stating his wager, he comments that people who are convinced his reasoning is sound may not be able to suddenly start being theists. He sug- gests to them that they go live among religious people. By doing this, habits of faith will grad- ually take hold. Pascal realized you can't simply decide to believe something. But he did see that you can choose to live your life in such a way that belief will come naturally after a while.
Chapter io: Pascal and Irrationality
Pascal's con1n1ent tneans that the \Vager can be refor111ulated so that the actions considered really are things under our control. Rather than treating belief and nonbelief as things we can choose to do directly, let's recognize that sincere belief can only occur in the right setting, where the right habits have been developed. It is also worth mentioning that Pascal's Wager concerns what it is best for you to believe. The question of whether you can do what is best for you is a separate question.
Second Criticism of Pascal's Argument
The second objection to Pascal's argument is more telling. In Chapter 9, I discussed the difficulty of establishing what God would be like if such a being existed. This is the problem of independently confirming auxiliary hypotheses, as I called them, about God's nature.
Pascal's Wager makes very specific assumptions about what God would be like if there were such a being. The assumption is that God would send believers to heaven and nonbe- lievers to hell. But there are other conceptions to consider. For example, maybe God would reward those who lead a good life and would punish those who are bad, independently of what those individuals' theological opinions happen to be. Or consider another possibil- ity: Perhaps God will send everyone to heaven except those who are convinced by Pascal's Wager.
The main problem with the wager is that it makes assumptions about God's nature that are part of one religious tradition but not part of many others. How can one tell which assu1nptions are true? In the absence of a convincing argun1ent, I see no reason to think Pascal's description of the payoffs is correct.
The Role of Reason
Pascal argued that it can make sense to believe things that are totally unsupported by evi- dence. If you think evidence is the hallmark of rational belief, then you'll interpret Pascal as giving a justification of irrationality. Of course, he didn't abandon the standard of ra- tionality altogether. He tried to provide a rational argument that we should believe in God even if there is no evidence that God exists. Pascal ain1ed to provide a prudential reason, not an evidential reason.
Pascal rejected the idea that evidential reasoning should determine whether or not we believe in God. Other religious thinkers have done the same thing, but for reasons that differ from Pascal's. For example, it has been suggested that God is so radically different from all the other beings that we know about that the methods of reason- ing appropriate for discussing these other beings are no longer appropriate when the question is a theological one. God is said to be an infinite being, one who may stand outside of space and time and yet be able to influence what happens inside of space and time. God is also supposed to be all-powerful, all-knowing, and all-good. In all these respects, God is far different from the familiar objects about which we are accustomed to reasoning.
Let's grant that God is supposed to have these special features. Why do they show that reason isn't an appropriate tool to use in thinking about whether God exists?
113
Patt T\vo: Philosophy of Religion
I noted in Chapter 2 and again in Chapter 9 that deductive validity depends on an argu1nent's for1n, not on its subject n1atter. If this is right, then deductive reasoning should apply to the question of God's existence in the same way that it applies to other questions.
Sometimes, it is suggested that God is so shrouded in mystery that we can't hope to understand God by rational inquiry. Maybe this is right-maybe the issues are so mysteri- ous that you'll never be able to see that God exists if you adhere to rational standards of argument. But why does this mean that those standards should be abandoned? Why not take this to mean that you shouldn't believe that God exists, precisely because no good argu1nent can be produced for this conclusion?
I won't pursue these questions any further. Maybe it can be explained why usual stan- dards of evidence and argument should be abandoned when the question is whether God exists. Ho,vever, I kno\V of no reason to think this.
Freud's Psychological Explanation ofTheism
There is a curious fact here, one that really does require an explanation. If 've \Vere talk- ing about virtually any topic from everyday or scientific life, no one would be very attracted by the advice that we should believe something on the basis of no evidence. Yet, 'vhen the question is 'vhether God exists, people so1nethnes lo\ver their critical standards. 'fhis isn't true of everyone, of course. It "'asn't true of Ansehn, Aquinas, Htune, or Paley, for exa1nple. Still, the idea of s\vitching fron1 reason to faith has been attractive to 1nany.
There is a psychological question I want to pose here, even though I will not try to an- s\ver it. Why are people sometin1es inclined to believe \Vithout evidence \vhen the question is whether God exists, but are much less willing to take this attitude when the question is a different one?
In his essay, "The Future of an Illusion;' Sigmund Freud (1856-1939), the founder of psychoanalysis, gave an answer that is worth pondering. Freud points out that the world is in 1nany \vays a frightening place. It is a relatively recent event in hurnan history that people have co1ne to exercise some inastery over nature, and even this recent inastery is enormously incomplete. Even if you aren't threatened every day by wild animals, light- ning, drought, famine, and flood, you still have to face the inevitability of your own death. In this ultin1ate sense, \Ve are as helpless as our ancient ancestors. With so 1nuch of nature out of our control, beyond our understanding, and threatening our survival, it isn't sur- prising that human beings should invent myths that provide a different and more appeal- ing picture of our place in the cosn1os. Instead of vie\ving the universe as indifferent to us, ilnpersonal, incon1prehensible, and dangerous, "'e construct a comprehensible picture of forces n1ore po,verful than ourselves ,vho take an interest in our fate. Freud's hypothesis was that believing in God persists in human history because of the psychological benefits it provides to believers.
Pascal aimed to provide a prudential argument for believing that God exists. Freud was an atheist, but aimed to provide an explanation for why many people believe that God ex- ists. Pascal sought to justify theism; Freud didn't seek to justify it, but to explain it.
Chapter 10: Pascal and Irrationality
A New Prudential Argument
In spite of these differences, vve can use son1e of Freud's observations to for1nulate a prudential argument for theism that is stronger than the one that Pascal provided. The stumbling block for Pascal was the assumptions he made about what God would be like if he existed. We can avoid these difficult questions about God's nature) ho\\Tever, and focus on so1nething that is 1nore familiar and closer to ho1ne-our o'vn n1inds.
Suppose you are the sort of person who derives comfort and fulfillment from believing in God. Your belief in God provides your life with meaning. It gives you the courage to face adversity. Moreover, let's suppose that without believing in God, you would plunge into despair. You would become depressed, unable to live a productive life.
There are inany people in the 'vorld ,vho aren't like this. First, there are cheerful agnos- tics and atheists-people 'vho tnanage to inake a 1neaningful life for then1selves without needing a theological framework. Second, there are people who remain tormented by de- spair even after they become theists. Jn both these cases, I'm talking about people whose general outlook is rather independent of the theological opinions they happen to have. Nevertheless, let's suppose there are some people for whom belief in God makes all the difference. Perhaps you are a person of this sort. If so, you can use these psychological facts to formulate a prudential argument for theism:
If you believe in God, you will lead a happy and productive life.
If you don't believe in God, you will fall into depression and despair.
The way to maximize expected utility is for you to be a theist.
This argument rese1nbles Pascal's in that it doesn't focus on \vhatever evidence there inay be concerning 'vhether God exists. But unlike Pascal's argu1nent, this prudential argument doesn't make assumptions about what God would do to theists and atheists. This argu1nent focuses on the psychological consequences of theisn1, not on specula- tions about heaven and hell. This argument is more difficult to refute than Pascal's. The major issue it raises is difficult and general. Should we allow prudential reasons to influence what we believe, or should we require that our beliefs be regulated by evi- dence alone?
Pragmatism
Pragmatism is a general philosophical position that has implications for this question. This philosophical idea was developed in the nineteenth and twentieth centuries prin- cipally by Charles Sanders Peirce, William James, and john Dewey. In ordinary speech, "being pragmatic" means being practical. The philosophical position, however, isn't fully captured by the everyday meaning of the word. Pragmatists say that we should believe propositions that are useful to us. In everyday life, this usually means that we should be attentive to evidence. Suppose you are about to cross the street and 'vant to kno'v 'vhether a bus is coming. The most useful thing to do here is to use your senses (sight, hearing, etc.). It would be silly to engage in wishful thinking-to refuse to consult any evidence and
115
116
IE--C Read "The \Y/ill to
Believe," i-5, on wvAv .mysearchlab.com
IE--C Read "The Ethics of
Belief" on ww•v .mysearchlab.com
Part Two: Philosophy of Religion
, ' ''''' ''., .. ','','''.'. ''' .. ' ... '' .. ' simply believe the proposition you'd like to be true. In cases of this sort, the prudent thing to do is to attend to evidence. There are cases, ho"vever, in \Vhich prudential considerations can lead in one direction and evidential considerations in another. My story about the president juggling candy bars provides one example. A prudential argument for theism n1ay provide another.
In The Will to Believe (1897), William James (1842-1910) says that a person is entitled to believe in God for purely prudential reasons (of the psychological sort described earlier) if the belief provides a "vital benefit" and if no decision about theism can be made on the basis of the evidence available. But why shouldn't a pragmatist be more liberal? Even if there were substantial evidence against there being a God, why shouldn't sufficiently large prudential benefits ouhveigh the evidential considerations? James's position doesn't pro~ vide a blanket justification of theism. As mentioned before, people who could lead happy productive lives without being theists can't claim they need to believe in God to avoid falling into despair. Yet, the spirit of pragmatism {if not the letter of )ames's formulation) seems to sanction believing in God if belief would be useful.
'fo do justice to the proble111s pragmatisn1 raises, \Ve would need to consider a variety of cases. If it is a matter of life and death, is it reasonable for someone to be a theist on purely prudential grounds? In less extreme cases, is the comfort provided by theism reason enough to believe in God? This raises complex questions about the attitudes we should take towards our lives. Someone might reject pragmatic arguments for theism by saying that they prefer to look the universe square in the face without the benefit of comforting illusions. This person might prefer a life of honest uncertainty and doubt over a life made comfortable by beliefs that have no evidential basis.
The pragmatist says that we should believe what it is useful to believe. An opponent of pragn1atis1n inight insist that prudential considerations are ahvays irrelevant; '\Ve should believe only those propositions that are well supported by evidence. This is the position that W. K. Clifford defended in his essay "The Ethics ofBelief" (1877). He rejected )ames's position and inaintained that «it is '\vrong ahvays, every\vhere, and for anyone, to believe anything upon insufficient evidence:' Here is one of the arguments that Clifford gives:
If I let nlyself believe anything on insufficient evidence, there may be no great har1n done by the mere belief; it may be true after all, or I may never have occasion to ex- hibit it in out\vard acts. But I cannot help doing this great '\vrong to,vards Man, that I make myself credulous. The danger to society is not merely that it should believe wrong things, though that is great enough, but that it should become credulous and lose the habit of testing things and inquiring into the1n; for then it must sink back into savagery. (p. 79)
Clifford's warning is exaggerated. It just isrit true that society will "sink back into savagery" if you believe a proposition without having any evidence for it. People in benign and pro- gressive societies have been believing various propositions \Vithout having evidence for them for a long time. Clifford is here advancing a consequentialist argument; he is suggest- ing that the action in question is \vrong because it has bad consequences. Another possible argument for Clifford's position would be that it is wrong in itself to believe upon insuf- ficient evidence, regardless of whether this has good or bad consequences. This raises the question of why it is wrong in itself to so believe.
Chapter 10: Pascal and Irrationality 117
Review Questions
l. What is the difference between a prudential reason and an evidential reason for believing something?
2. Can it ever make sense to bet on something that probably v1on't happen?
3. Given the payoffs Pascal assigns to outcomes, what is the expected value of believing in God? Of not believing?
4. Do people believe what they do because they "decide to believe"? If not, does it follow that Pascal's argument is mistaken?
5. Hovi does Pascal's argument depend on a set of auxiliary assumptions (in the sense of Chapter 9) about God's nature?
6. Describe a prudential argument for believing in God that avoids the problem that Pascal's argu- ment confronts.
7. \Vhy did Clifford reject James's pragmatic argument for theism? Evaluate the argument that Clifford presents for his position.
Problems for Further Thought
l. In Chapter7, I explained \Vhat the Principle of Sufficient Reason is and sketched an argument for \Vhy it might be useful to believe the principle even ifthere is no evidence that it is true. Work out the details of this decision·theoretic argument by constructing a tvvo·by·t\VO table for this prob· Jem that resembles the one used for Pascal's Wager. Is the argument successful?
2. In Chapter 7, I argued that there is a kind of question (a global question about why the whole history of the universe is as it is} that science cannot answer. Can science ans\ver the question of whether we should accept james's or Clifford's position on the ethics of belief? If so, which scientific theory provides an answer to this question?
3- Ethicists often distinguish duties from superogatory acts. An action is a duty if it is something you must perform; an action is superogatory ifit would be a good thing to do, but is not a duty. How does this distinction apply to Clifford's argument against James?
4. The readings for this section of the book include an excerpt from Pascal's Pensees in which he states his \Vager. Jn fact, he states several \Vagers, steadily refining the \vager he thinks is most relevant. What are these different \Vagers and how do they differ? Which are more convincing and which are less?
Supplementary Reading
w. K. CLIFFORD The Ethics of Belief
Pragmatist arguments for believing in God, such as Pascal's wager, conclude that we could have reasons for believing that God exists even if we do not have evidence that God exists. ln this passage, Clifford argues that these arguments fail, for "it is wrong always, everywhere, and for anyone to believe anything on insufficient evidence." \Vhy does Clifford think that this is always \Vrong?
O!f--[ Read "The .... Ethics of
BeliefTI on wv11'1 .mysearchlab.com
118 Part Two: Philosophy of Religion
'.' > ••• ' •••• ' •• ' ••••• ' ••••••••••••••••••••••• ' ••••••••••••••••••••••••• ' ••
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchLab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
CEJ--{ READ AND REVIEW
READ
i. Blaise Pascal, "The Wager," Number 233 in Section 3 of Pensees In this passage from his Pensees, a book that was posthumously assembled from his notes, Pas- cal sets forth his nowfamous wager. In fact, he gives the wager different formulations. Try to discern what these different arguments of Pascal's are and how they di/Ju
2. William James, "The Will to Believe," 1-5 In the present essay, James argues that we sometimes face decisions about what to believe that can and should be made on some basis other than the evidence at hand. Why does he think that the decision to believe in God should be made this way?
3. W. K. Clifford, "The Ethics of Belief" Clifford argues that it is wrong to believe something without sufficient evidence because doing so has bad consequences. Do you think it's plausible that believing in God could have the negative consequences that Clifford envisions?
WATCH AND LISTEN
©{WATCH {<•>{LISTEN
1. Listen "Ben Rogers on Pascal's Pensees"- Phllosophy Bites podcast
"The Wager" is part of a larger book by Pascal, the Pensees. In this interview, Ben Rogers discusses the context and themes of the Pensees, focusing primarily on Pascal's view of the human condition. How might Pascal's beliefs about humanity have influ· enced his formulation of the Wager?
References
RESEARCH AND EXPLORE
!Jr.{EXPLORE
1. Research Blaise Pascal 2. Research William James 3. Research pragmatism
William Kingdon Clifford (1845-1879) "The Ethics of Belief', 1877.
chapter 11 ..............................
The Argument from Evil
Chapter Outline
First Version of the Argument
Two Kinds of Evil
Possible Reactions to the Argument
Theodicy and Defense
Soul-Building Evils
Second Version of the Argument
Free Will
Examples and a Third Version of the Argument
A Criticism of the Argument
Testability, Again
Another Kind of Argument-The Evidential Argument from Evil
Most of the arguments considered so far concerning the existence of God have had theistic conclusions. I no'\v turn to an issue many regard as providing an argument for atheism. Sometimes it is thought that the fact that there is evil in the world (or that there is so much of it) proves there is no God. An argument to this effect is called an Argument from Evil. I'll consider three forms this argument might take. I think none of these successfully proves there is no God. Then I'll consider a fourth version of the argument from evil; here, instead of trying to prove that there is no God, the argument attempts to show something more modest-that the existence of evil is evidence against the hypothesis that God exists.
First Version of the Argument
The first form of the Argument from Evil is simple:
(1) If God were to exist, then that being would be all-powerful, all-knowing, and all-good (all-PKG, for short).
119
120 Part Two: Philosophy of Religion
(2) If an all-PKG being existed, then there would be no evil.
(3) There is evil.
Hence, there is no God.
The argument is deductively valid; convince yourself of this by identifying the argument's logical form.
Son1etin1es it is said that the first pre1nise is a definition of \vhat 've tnean by the concept of God. l'll say in a minute why I don't think (1) is a definition. Premise (2) is usually de- fended as follows. If God is all-good, he wants to prevent evil. Ifhe is all-knowing, he knows the difference between right and wrong and knows how to prevent evil from coming into existence. And ifhe is all-powerful, he can prevent evil, should he wish to do so. So God, if he is all-PKG, has both the inclination and the ability to prevent evil from occurring.
Two Kinds of Evil
What does evil mean in premise (3)1 Discussion of the Argument from Evil usually divides existing evils into hvo categories. 'I'here are the evils that are brought into existence by lnunan actions, and there are the evils that exist because of natural events that aren't under human control. The first category includes the suffering that human beings inflict on each other; the second includes the suffering that human beings (and other creatures) experi- ence because of natural disasters, such as earthquakes and plagues.
Possible Reactions to the Argument
As tnentioned before, the argument is valid. Hence, if you \Vish to reject the conclusion, you n1ust reject one or inore of the pre1nises. What options are available?
One option is to deny the existence of evil (premise 3). For example, you could take the view that there really is no difference between right and wrong. This is the idea that the universe is morally neutral and that "re htunan beings itnpose illusory moral catego- ries upon it. This view about morality will be discussed in the fifth section of the book, especially in Chapters 29 and 30. For now I'll ignore this option. I'll assume that some of the murders, tortures, and other brutalities that have peppered human history really are morally bad. I will also assume that the world would have been a better place if some of the suffering caused by natural disasters hadn't occurred. This assumption-that premise (3) is correct-is shared by many, if not all, religious traditions.
Another option is to reject premise (1). You could interpret the existence of evil as showing merely that if there is a God, then God isn't all-PKG. This strategy grants prem- ises (2) and (3), but then concludes that if God exists, he can't be exactly the ''""Y one reli- gious tradition says he is.
The fact that this strategy is available shows, I think, that (1) isn't "true by definition:' If the first pren1ise really did define 'vhat the ter1n "God" n1eans, then the existence of evil couldn't be explained by rejecting this premise. But this does seem to be a real option. What we need to recognize is that premise (1) doesn't have the same status as the follow- ing sentence: ''A bachelor is an unmarried tnan:' Suppose 've found that a Jnan ,vho1n
Chapter 11: The Argun1ent from Evil
we thought is a bachelor in fact is married. Would this lead us to consider the possibility that maybe bachelors could be married? I think not. Premise (1) embodies one theory about what God's nature is. It is one theory among many possible theories, none of which deserves to be viewed as "true by definition:' The ancient Greeks denied that the gods are all-PKG. Manichaeism and Zoroastrianism deny that God is all-powerful. So rejecting premise (1) is an option we need to consider as a way of explaining why evil exists.
Theodicy and Defense
One other option remains: reject premise (2). This is the strategy pursued by theists who think premises (1) and (3) are true. If you think God is all-PKG and that evil is a reality, you need to show how God's being all-PKG doesn't imply that there should be no evil in the world.
There are two types of criticism that might be offered of premise (2). You could try to explain why it is false. This is the project known as "theodicy" in traditional Christian theology. Here you try to explain why an all-PKG God would allow evil to exist. Alter- natively, there is the more modest criticism in which you try to show that we don't have a good reason to think that the premise is true. Contemporary philosophers of religion use the term "defense" to describe this second type of reply to the Argument from Evil. The distinction that separates theodicy from defense can be seen in the discussion of other arguments for the existence of God. For example, in Chapter 4, we considered an objection to the design argument that maintains that if God built organisms, they would be perfectly adapted to their environments. One way to criticize this argument is to sug- gest that there is no reason to think that God would have wanted each organism to be perfectly adapted. A more ambitious criticism would be the attempt to demonstrate what God's motives would be.
In any event, I think there is some plausibility in rejecting premise (2). Notice that premise (2) makes a very ambitious claim. It says that if an all-PKG God existed, there would be zero evil in the world.
Soul-Building Evils
Traditional theodicy has claimed that some evils have the property of being "soul-building:' Soul-building evils are the ones that make us better people. Living through adversity strengthens our character; it makes us better people than we would have been if we had never suffered hardship.
I want to be careful about what I'm conceding here. I grant that some evils are soul- building. But I think it is a gross exaggeration to say this of all the evils that human beings have experienced. Somethnes suffering doesn't make the victin1 stronger; sometimes suffering destroys people. And even when people survive brutal tortures, do we really think that they are always better people for it? Sometimes they survive as mere shells. Even 'vhen torture victims in some sense are made stronger by their ordeals, do 've still 'vant to say that the experience has been worth it? If you think it was worthwhile for them, would you choose to undergo the torture yourself, or choose it for someone you love? There are sotne «strengths,, that aren,t 'vorth having because they cost too much to acquire.
121
122 Part Two: Philosophy of Religion
Sometimes it is suggested that when people are destroyed by the sufferings they endure, the souls of other people are strengthened by watching this happen. I think this is so1neti1nes true) but it is often false. So1nethnes the only effect on "\Vitnesses is that they are sickened by having to watch events they know full well are horrible. Sometimes people "\Vho \Vitness the suffering of others react 1Nith callousness and indifference, and some suffering occurs in private. In these cases, there is no audience \vhose souls are made stronger by watching the spectacle. just as suffering often doesn't make the suf- ferer a better person, it also is true that one person's suffering often fails to itnprove the characters of \Vitnesses. So in granting that so111e suffering is soul-building, I'n1 granting something about a rather s1nall portion of the huge tnountain of agony the hu1nan race has experienced.
Is the fact that some evils are soul-building enough to refute premise (2)? This isn't so clear. If it is desirable that we be strong of heart, why couldn't God have made us this way \Vithout us having to experience evil? The need for soul-building experiences is so1nethnes defended by an analogy concerning the way a benevolent parent treats a child. If you re- ally care about your children, you won't shield them from all adversity. There are some experiences that involve pain and suffering that children need to have if they are to 1nature into morally sensitive and autonomous adults. If your child is being picked on by a bully at school, maybe you will decide that your daughter needs to tough it out for herself. Inter- vening would save her a few scrapes and bruises now, but you might decide as a parent that this wouldn't be in her long-term best interests.
You, as a parent, have the goal of seeing your children develop in a certain \vay. You choose not to prevent some bad things from happening to your child because you recog- nize that there is no other \\•ay for her to inature as you intend. Suppose, ho\vever, that you could have your child become a virtuous adult without her having to be kicked around by the bully. In this case, why would you allow the bully to continue to harass your child? A parent might allow soul-building evils to befall a child when there is no other way to have the child's soul grow strong, but the limitations that might justify a parent's choice aren't supposed to be present where God is concerned. If God is all-PKG, why couldn't he sim- ply make us with strong souls? Defenders of the soul-building idea reply that God wants us to be able to be proud of our own achievements. If we suffer through adversity, we get credit for the strong souls that result. If, however, God simply gave us strong souls gratis, \Ve \vould deserve no such credit. We couldn't be proud of ourselves; at best, \Ve \vould simply be grateful to a benevolent deity. To evaluate this suggestion, we must ask why it is so important that we be able to take credit for having strong characters. Why is it better to experience an evil and then get credit for having endured it? Why is this better than having a strong character because God rnade us that \Vay fro1n the start? I raise this question, but won't attempt to answer it. I'll concede for now that an all-PKG God would allow at least some evils to exist. Some soul-building evils are of this sort. In other \Vords, I'll concede that premise (2) is false.
Does this mean there is no problem of evil-that the existence of evil poses no diffi- culty for the conception of God as an all-PKG being? I think not. The soul-building evils an all-PKG God would allow us to experience are a tiny portion of the evils we need to explain. What needs to be explained isn't why there is some evil rather than none at all. The deeper proble1n is to explain \vhy there is so much evil, rather than one ounce less.
Chapter 11: The Argutnent fron1 Evil
''.' '''' •• ' •• ''' '' •• >. ' •••• '. '.
Second Version of the Argument
The second Argument from Evil is more difficult to refute than the first. It takes the following form:
(1) If God were to exist, then that being would be all-PKG.
(4) If an all-PKG being existed, then the amount of evil would not exceed a soul- building 1nini1num.
(5) The amount of evil does exceed a soul-building minimum.
Hence, there is no God.
By "soul-building tninbnum;' I n1ean the niinimun1 an1ount of evil that 'vould suffice to make us have strength of character. This argun1ent is valid. Hence, if the conclusion is false, at least one of the premises must be false. The project of theodicy is to show that premise (4) is mistaken.
Free Will
Is there a reason that an all-PKG being might have for permitting evils to exist that go be- yond the soul-building minimum? It is sometimes suggested that this reason is to be found in the idea of free will. God made us free; this means that it is up to us whether we do good or do evil. A consequence of this freedom is that there is more evil than would be strictly necessary for soul-building.
There is a question here that 've \Vill exan1ine in detail in the section of this book on philosophy of mind. Why couldn't God have made us free agents who always freely choose to do what is right? Why assume that free people must sometimes perform evil actions? If an evil doer can have free will, why can't someone who behaves better? After all, God is supposed to be a free being who never does evil things. This seems to show that being free and always doing what's right are not incompatible. Even if being free implies that you could do evil things, why does freedom imply that you actually choose to do so?
Ho\vever, let's no\v set this point aside and assun1e, for the sake of argu1nent, that hun1an freedom requires that people sometimes produce evil and that this evil exceeds the soul-building 1ninitnun1. This means that pre1nise (4) in the second argument is n1istaken. But again, I 'vant to emphasize that this doesn't fully solve the problem of evil. The problem is to account for the quantity of evil that exists. Let's grant that some evil must exist for soul-building and some more evil will be a consequence of human freedom. Does this explain the quantity of evil that the world contains?
Examples and a Third Version of the Argument
Let's single out some horrendous evil that has befallen the human race. Think of all the suffering Hitler caused. How could an all-PKG God have allowed this to happen? Is the answer that God gave each of us free will? I think this appeal to the importance of human freedom doesn't begin to address the difficulty. Why didn't God intervene in human affairs
124 Part Two: Philosophy of Religion
........ ' ... ''.' ......... ' ...... ' ...... ' ...... ' ' .. ' ' ... '' '' .. ' ..... ',.''.' on those rare occasions on \Vhich it \vould have made a dratnatic difference? When the sperm and egg came together that produced Hitler, why didn't God prevent that fertiliza- tion event from happening? To do so would still have allowed human beings to be free; in addition, that brief intervention would allow there to be soul-building evils aplenty. Here I'm considering a minor foray into hu1nan affairs, not son1e inajor restructuring ofhu1nan nature. Suppose that if Hitler had never existed, many of the terrible things he did would never have happened. The world would have been a much better place. Most theists would agree with this. So why didn't God influence the course of human history for the better? What you must see here is that appeals to soul-building and to freedom do1it provide an answer. Causing Hitler to never be born would not have robbed his parents of free will.
Another \Vay to make this point is to consider horrendous evils that come into exis- tence for reasons having nothing to do with human action. Consider the bubonic plague. If God had prevented that disease from existing, much suffering and death would have been avoided. To intervene in the lives of the relevant n1icroorganis1ns does1it rob then1 of free will, since they don't have wills (minds) to begin with. Also, the amount of suffering that occurred because of the plague far exceeded what could be explained by the goal of soul- building. About one-third of the European population died agonizing deaths from bubonic plague. Would our souls have been less strong if some smaller percentage had perished?
These considerations lead to a third version of the Argument from Evil:
(1) If God were to exist, then that being would be all-PKG.
( 6) If an all-PKG being existed, then there would be no more evil than the mini- mum required for soul-building and as a consequence of human freedom.
(7) The quantity of evil found in human history exceeds the minimum required for soul-building and as a consequence of human freedom.
Hence, there is no God.
A Criticism of the Argument
How might this third argument be criticized? If there is a God, and if premises (1) and (7) are true, what grounds can there be for rejecting premise (6)? The project oftheodicy needs to produce a reason for thinking that ( 6) is false. The more modest project of defense merely needs to show that we do1it have a good reason to think that ( 6) is true. Although soul-building and free will were just presented in the context of a theodicy, I'll now shift gears and formulate a reply to this last argument that has the character of a defense.
Premise (6), like premises (2) and (4) in the preceding arguments, attempts to specify an upper bound on the amount of evil that an all-PKG God would allow to exist. But why think that we mere humans have the ability to figure this out? After all, God, if he exists, is vastly more intelligent than we are. Therefore, it is entirely possible that his plan for the world contains elements that we cannot understand or even imagine. Perhaps God allowed evils into the world beyond those required for soul-building and as a consequence of human freedom because these additional evils are required to obtain some greater good of which we are unaware. We need to be careful in how we formulate this criticism of the Argument
Chapter 11: The Argument from Evil
', •••• ''.',,'' ''' '.''.'.''' •• ,, ••• '' '>' ' •••• ''. '' '.' ''. '.''.' ''.' •••• ' •• '.
from Evil. I am not asserting that premise ( 6) is false. I am giving a reason for thinking that we don't know that ( 6) is true. All of the versions of the Argument from Evil that I have surveyed attempt to pinpoint exactly how much evil there should be if an all-PKG being existed. To do this requires that we know a lot about what an all-PKG being would do. To the degree that we lack knowledge of what God's plan would be if he existed, we also lack the sort of knowledge that is needed for these arguments for atheism to succeed.
Testability, Again
In Chapters 6 and 9, I discussed whether the existence of God is testable. The reply I just formulated to the Argument from Evil raises the same issue. How much evil should we ex- pect there to be if there were a God? Atheists who advance the Argument from Evil just de- scribed think they can ans\\1er this question. Ho,vever, if 've catit ans,ver this question, \Ve catit test the proposition that God exists by examining how much evil there is in the world.
Another Kind of Argument-The Evidential Argument from Evil
The arguments from evil formulated so far are deductively valid; the goal is to prove that an all-PKG God cannot exist, given this or that fact about the evils that there are.
There is reason to think that no such deductive argument can be made to work. Such an argument would have the following form:
If an all-PKG God exists, then evils of kind X will not exist.
Evils of kind X do exist.
No all-PKG God exists.
This argument is deductively valid (what is its logical form?). The problem with this kind of argument is that it is hard to show that the first premise is true. Atheist might object to this argument by asserting that human beings are in no position to understand why an all-PKG God would allow evils of kind X to exist. The theist's pointis not to assert that the first premise is false but to say that we carit tell whether it is true.
I no\v turn to a more modest for1nulation of the argutnent from evil. It is called the evi- dential argument from evil (see W Rowe's article, "The Problem of Evil and Some Varieties of Atheism;' American Philosophical Quarterly [1979] 16: 335-41). The goal here is merely to show that this or that kind of evil is evidence against the hypothesis that an all-PKG being exists; the argument does not claim that these evils prove that no such being exists. We can use the Surprise Principle to formulate the argument. Let Ebe a proposition that describes in some detail the kinds and quantities of evil that exist:
If an all-PKG God existed, you would expect E not to be true.
If an all-PKG God did not exist, you would expect E to be true.
The Surprise Principle.
Eis strong evidence against there being an all-PKG being.
126 Part Two: Philosophy of Religion
...... '.' ... ' .. ' ............. '' .. "' .......... ' .. ' .... ' ........ ' ... ' ' .. ' .. . The Surprise Principle describes what "strong" evidence means. It can be reformulated to define '\vhat evidence (strong or '\Veak) means, and this allovvs us to construct an even inore modest version of the evidential argument from evil:
The truth of proposition Eis more surprising if an all-PKG God exists than it would be if there were no such being.
The Modest Surprise Principle.
Eis evidence against there being an all-PKG being.
As an exercise, write down what the Modest Surprise Principle says. This modest eviden- tial argument from evil says that E favors the hypothesis that no all-PKG God exists. Is it true that the kinds and quantities of evil that exist are evidence against the existence of an all-PKG God? Ifnot, what could count as evidence against this hypothesis?
Review Questions
1. Consider the proposition that God is all-powerful, all-knowing, and all-good. Should we regard this proposition as a definition of what the word "God" means? If it isn't a definition, does this mean that the proposition is false?
2. It sometimes is suggested that imperfections must exist if human beings are to be able to form the idea of God (a perfect being). Is this a solution to the problem of evil?
3. Does the fact that some evils are soul-building while others are consequences of human freedom solve the problem of evil?
4. What is the difference between theodicy and defense as criticisms of the Argument from Evil? Illustrate this distinction by constructing an argument of each kind that has nothing to do with the existence of God.
5. A criticism was presented of the third Argument from Evil. The argument is valid. And the criti- cism did not claim that any of the premises are false. What, then, is the criticism?
6. What is the evidential argument from evil? Is it more successful than the arguments that attempt to prove that there is no God?
Problems for Further Thought
1. The appeal to free will, which is part of the project oftheodicy, includes the claim that (a) God is omnipotent (all-powerful) and the claim that (b) God cannot give us free will without there also being evil in the world. Are (a) and (b) compatible? How should the concept of "omnipotence" be understood?
2. Are there reasons. beyond the goal of soul-building and the goal of giving us free will that an all-PKG God would have for allowing evil to exist? If so, would those additional reasons solve the problem of evil?
3- Some people interpret the Bible as providing an adequate theodicy. According to their interpretation, Adam and Eve freely chose to sin in the Garden of Eden. When they did so, this effectively un- leashed all of the rest of the evils that we find in the world. This theodicy apparently can explain all of the evil that there is, not just the evils that can be attributed to soul-building and to free will. Is this theodicy an adequate reply to the Argument from Evil?
Chapter 11: The Argument from Evil 127
. ' ...... ' ............. ' ........... ' ...................................... . 4. Here's a question about the free will solution to the problem of evil. If God knows in advance
what we'll do (which he must, ifhe is all-knowing), how can it be true that we have free will I This is a question we'll address in Part lV.
5- In Chapter 7, I discussed Leibniz's (1646-1716) idea that this is the best of all possible worlds. Read this passage from his work Theodicy:
How does Leibniz propose to address the problem of evil I The French Enlightenment philosopher Voltaire (1694-1778) ridiculed Leibniz in his 1759 novel Candide. Did Leibniz deserve Voltaire's ridicule!
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchLab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
READ AND REVIEW
t:EJ-{ READ l. Gottfried Leibniz, "Theodicy"
Why does Leibniz think that our world is the best of all possible worlds? Is Leibniz's response an adequate reply to the problem of evil?
WATCH AND LISTEN
©-[WATCH (<,.-[LISTEN
1. Listen "Marilyn McCord Adams on Evil"- Philosophy Bites podcast
In this podcast, philosopher of religion Mar- ilyn McCord discusses the problems that "horrendous evils" pose for belief in God. She argues that the existence of these evils present an even greater challenge for athe- ists who believe that life has purpose. Why does Adams think that belief in God better equips us to deal with the presence of evil?
Suggestions for Further Reading
RESEARCH AND EXPLORE
~!(~EXPLORE
l. Research problem of evil 2. Research theodicy 3. Research free will
On the Debate between Creationis1n and Evolutionary Theory
Michael Behe. Darwin's Black Box. New York: Free Press, i996. Douglas Futuyama. Science on Trial: The Case for Evolution. New York: Pantheon, 1982. Dwayne Gish. Evolution? The Fossils Soy No! San Diego, CA: Creation-Life Publishers, i979. Phillip Johnson. Defeating Darwinism by Opening Minds. Downers Grove, IL: lntervarsity Press, 1997- Philip Kitcher. Abusing Science: The Case against Creationism. Cambridge, MA: MIT Press, 1982. Henry Morris. Scientific Creationism. San Diego, CA: Creation-Life Publishers, 1974. Robert Pennock. Tower of Babel-The Evidence against the New Creationism. Cambridge, MA:
MIT Press, 1998.
[:EJ-[ Read Theodicy
on \VW\'1.mysearchlab .com
128 Part Two: Philosophy of Religion
.................. ' .... ".'.'. "' '.'. ''' '' .. '' ''.'' On the Testability of Religious Clahns
Essays by Antony Flew, R. M. Hare, and Basil Mitchell. "Theology and Falsification." In New Essays in Philosophical Theology, edited by Antony Flew and Alisdair Macintyre. New York: Macmillan, 1966.
Carl Hempe!. "Empiricist Criteria of Cognitive Significance: Problems and Changes." In Aspects of Scientific Explanation. Ne\v York: Free Press, 1965.
Elliott Sober. "Testability." Proceedings and Addresses of the American Philosophical Association (1999). http://philosophy.wisc.edu/sober. Accessed June 21, 2012.
On the Ontological Argun1ent
John Hick and Arthur McGill, eds. The Mony·Faced Argument: Recent Studies on the Ontological Argument for the Existence of God. London: Macmillan, 1968.
Alvin Plantinga. The Nature of Necessity. Oxford: Oxford University Press, i974.
On Pascal's \Vager
James Cargile. "Pascal's Wager." Philosophy 41 (1966): 229-36. William Lycan and George Schlesinger. "You Bet Your Life: Pascal's \Vager Defended." In Reason
and Responsibility, ed'ited by J. Feinberg and R. Shafer-Landau. Belmont, CA: Wadsworth, 1999. Gregory Mougin and Elliott Sober. "Betting against Pascal's Wager." Nous 28 (1994): 382-95. Blaise Pascal. Pensei. London: Everyman's Library,J. M. Dent&Sons, i958.
On the Argu1nent front Evil
Robert M. Adams. "Must God Create the Best?" Philosophical Review 81 (1972): 317-32. Richard Swinburne. "The Problem of Evil." In Reason and Religion, edited by Stuart Brown. Ithaca,
NY: Cornell University Press, i977. Peter Van lnwagen. "The Magnitude, Duration, and Distribution of Evil-A Theodicy." Philosophical
Topics 16 (1988): 161-87.
On Recent \Vork in Philosophy of Religion
William Mann. The Blackwell Guide to Philosophy of Religion. New York: Blackwell Publishers, 2004.
readings ..............................
Readings
SAINT THOMAS AQUINAS
Five Ways to Prove That God Exists
In this selection from Summa Theologica (Part 1, Question 2, Article 3), Saint Thomas Aquinas provides five proofs of the existence of God. In each, an obvious and uncontroversial observation is the starting premise.
WILLIAM PALEY
The Design Argument
In this selection from his Natural Theology (1836 edition), William Paley elaborates an argument for the existence of God that traces back at least to the last of Aquinas's five ways. His discussion of how one interprets a watch found on a heath is one of the most famous analogies proposed in the history of philosophy. The term "natural theology" means that the author attempts to establish the existence and nature of God by the same methods used in the natural sciences-observation and reasoning.
DAVID HUME
Critique of the Design Argument
Jn this selection from his Dialogues Concerning Natural Religion (1779), Hume describes a conver- sation between Cleanthes, Demea, and Philo. Hume did not publish this \Vork during his lifetime, probably because he was apprehensive about what the public reaction would be.
(E]-{ Read uFive \'!Jays to
Prove That God Exists," Pt. 1, Question 2, Art. 3 of Sununa Theologica on ww\v.mysearchlab .com
(E]-{ Read uThe Design
Argument," Chapter 1 of Natural Theology on wvm.mysearchlab .com
(E]-{ Read Part 2 of Dialogues
Concerning Natural Religion on wwv1 .mysearchlab.com
129
130 Part Tv10: Philosophy of Religion
.. ''' '''.' ''''''''.' '' .. '.' ''' .. '.,' .. ' ' ... '. ' .. '',.' '' '', '' ''.
Cl!J-{ Read uGuanilo
and Anselm: Debate" on Wi'N1.mysearchlab . com
SAINT ANSELM AND GAUNILO
The Ontological Argument
In this selection from the Proslogion,-Saint Anselm attempts to provide an a priori proof of the existence of God. Gaunilo, a contemporary of Anselm's, presents two criticisms of the argument. In one of them, he claims that if Anselm's argument establishes the existence of God, then the exis- tence of a perfect island \Vould be provable a priori. Anselm replies to each of Gaunilo's criticisms .
ALFRED )ULES AYER
The Meaninglessness of Religious Discourse
In this selection from Language, Truth, and Logic (1936), A.). Ayer argues that the positivist theory of meaning sho\vs that the claim that God exists is neither true nor false-it is meaningless.
I t is now generally admitted, at any rate by philosophers, that the existence of a being having the attributes which define the god of any non-animistic religion cannot be demonstratively proved. To see that this is so, we have only to ask ourselves what are the premises from which the existence of such a god could be deduced. If the conclusion that a god exists is to be de1nonstratively certain, then these premises inust be certain; for, as the conclusion of a deductive argun1ent is already contained in the premises, any uncer- tainty there may be about the truth of the premises is necessarily shared by it. But we know that no empirical proposition can ever be anything more than probable. It is only a priori propositions that are logically certain. But we cannot deduce the existence of a god from an a priori proposition. For '\Ve kno\v that the reason a priori propositions are certain is that they are tautologies. And from a set of tautologies nothing but a further tautology can be validly deduced. It follows that there is no possibility of demonstrating the existence ofagod.
What is not so generally recognized is that there can be no way of proving that the existence of a god, such as the God of Christianity, is even probable. Yet this also is eas- ily shown. For if the existence of such a god were probable, then the proposition that he existed would be an empirical hypothesis. And in that case it would be possible to deduce fro1n it, and other e1npirical hypotheses, certain experiential propositions ,vhich \Vere not deducible from those other hypotheses alone. But in fact this is not possible. It is some- tilnes clahned, indeed, that the existence of a certain sort of regularity in nature constitutes sufficient evidence for the existence of a god. But if the sentence "God exists" entails no more than that certain types of phenotnena occur in certain sequences, then to assert the existence of a god will be simply equivalent to asserting that there is the requisite regularity in nature; and no religious n1an '\Votild adtnit that this '\Vas all he intended to assert in as- serting the existence of a god. He would say that in talking about God, he was talking about a transcendent being who might be known through certain empirical manifestations, but certainly could not be defined in terms of those manifestations. But in that case the term "god" is a metaphysical term. And if "god" is a metaphysical term, then it cannot be even probable that a god exists. For to say that "God exists" is to make a metaphysical utterance
Readings 131
''.' '' •• ' ' ••• ''.''.' •• '.' •••• ' '' •••• ''. '''.' '. '.' '' >.'.' '''. '' •• ''.''. ' •• '
'vhich cannot be either true or false. And by the sa1ne criterion, no sentence 'vhich pur- ports to describe the nature of a transcendent god can possess any literal significance.
It is iinportant not to confuse this vie'v of religious assertions 'vith the vie\v that is adopted by atheists, or agnostics.' For it is characteristic of an agnostic to hold that the existence of a god is a possibility in 'vhich there is no good reason either to believe or disbelieve; and it is characteristic of an atheist to hold that it is at least probable that no god exists. And our view that all utterances about the nature of God are nonsensical, so far from being identical with, or even lending any support to, either of these familiar conten- tions, is actually incompatible with them. For ifthe assertion that there is a god is nonsen- sical, then the atheist's assertion that there is no god is equally nonsensical, since it is only a significant proposition that can be significantly contradicted. As for the agnostic, although he refrains from saying either that there is or that there is not a god, he does not deny that the question 'vhether a transcendent god exists is a genuine question. He does not deny that the two sentences "There is a transcendent god" and "There is no transcendent god" express propositions one of which is actually true and the other false. All he says is that ,.,,e have no n1eans of telling ,vhich of thetn is true, and therefore ought not to comn1it ourselves to either. But ,.,,e have seen that the sentences in question do not express proposi- tions at all. And this means that agnosticism also is ruled out.
Thus we offer the theist the same comfort as we gave to the moralist. His assertions cannot possibly be valid, but they cannot be invalid either. As he says nothing at all about the world, he cannot justly be accused of saying anything false, or anything for which he has insufficient grounds. It is only when the theist claims that in asserting the existence of a transcendent god he is expressing a genuine proposition that we are entitled to disagree with him.
It is not to be remarked that in cases where deities are identified with natural objects, assertions concerning them may be allowed to be significant. If, for example, a man tells me that the occurrence of thunder is alone both necessary and sufficient to establish the truth of the proposition that Jehovah is angry, I may conclude that, in his usage of words, the sentence "Jehovah is angry" is equivalent to "It is thundering:' But in sophisticated religions, though they may be to some extent based on men's awe of natural process which they cannot sufficiently understand, the "person'' who is supposed to control the empirical world is not himself located in it; he is held to be superior to the empirical world, and so outside it; and he is endO'\Ved 'vith super-empirical attributes. But the notion of a person whose essential attributes are nonempirical is not an intelligible notion at all. We may have a 'vord 'vhich is used as if it named this "person;' but, unless the sentences in \Vhich it occurs express propositions which are empirically verifiable, it cannot be said to symbolize anything. And this is the case with regard to the word "god;' in the usage in which it is intended to refer to a transcendent object. The n1ere existence of the noun is enough to foster the illusion that there is a real, or at any rate a possible entity corresponding to it. It is only 'vhen 've enquire \Vhat God's attributes are that ,.,,e discover that "God;' in this usage, is not a genuine natne.
1This point 1vas suggested to 1ne by Professor H. H. Price.
132 Part Two: Philosophy of Religion
. '.' .... ' ... ' ..................... ' .... ' ...... ' .... ,' .... ' .. '' '' ... ' '.' .. . It is common to find belief in a transcendent god conjoined with belief in an after-life.
But, in the form which it usually takes, the content of this belief is not a genuine hypoth- esis. To say that men do not ever die, or that the state of death is merely a state of pro- longed insensibility, is indeed to express a significant proposition, though all the available evidence goes to show that it is false. But to say there is something imperceptible inside a man, which is his soul or his real self, and that it goes on living after he is dead, is to make a metaphysical assertion which has no more factual content than the assertion that there is a transcendent god.
It is worth mentioning that, according to the account which we have given of religious assertions, there is no logical ground for antagonism between religion and natural science. As far as the question of truth or falsehood is concerned, there is no opposition between the natural scientist and the theist who believes in a transcendent god. For since the religious utterances of the theist are not genuine propositions at all, they cannot stand in any logical relation to the propositions of science. Such antagonism as there is between religion and science appears to consist in the fact that science takes a\vay one of the 1notives ,vhich make men religious. For it is acknowledged that one of the ultimate sources of religious feeling lies in the inability of men to determine their own destiny; and science tends to destroy the feeling of awe with which men regard an alien world, by making them believe that they can understand and anticipate the course of natural phenomena, and even to some extent con- trol it. The fact that it has recently become fashionable for physicists themselves to be sym- pathetic towards religion is a point in favour of this hypothesis. For this sympathy towards religion marks the physicists' own lack of confidence in the validity of their hypotheses, \Vhich is a reaction on their part from the anti-religious dog1natis1n of nineteenth-century scientists, and a natural outcome of the crisis through which physics has just passed.
It is not within the scope of this enquiry to enter more deeply into the causes of reli- gious feeling, or to discuss the probability of the continuance of religious belief. We are concerned only to ans\ver those questions which arise out of our discussion of the possi- bility of religious knowledge. The point which we wish to establish is that there cannot be any transcendent truths of religion. For the sentences which the theist uses to express such "truths" are not literally significant.
An interesting feature of this conclusion is that it accords with what many theists are accustomed to say themselves. For we are often told that the nature of God is a mystery which transcends the human understanding. But to say that something transcends the human understanding is to say that it is unintelligible. And what is unintelligible cannot significantly be described. Again, we are told that God is not an object of reason but an object of faith. This may be nothing more than an admission that the existence of God must be taken on trust, since it cannot be proved. But it may also be an assertion that God is the object of a purely mystical intuition, and cannot therefore be defined in terms which are intelligible to the reason. And I think there are many theists who would assert this. But if one allows that it is impossible to define God in intelligible terms, then one is allowing that it is impossible for a sentence both to be significant and to be about God. If a mystic admits that the object of his vision is something which cannot be described, then he must also admit that he is bound to talk nonsense when he describes it.
For his part, the mystic may protest that his intuition does reveal truths to him, even though he cannot explain to others what these truths are; and that we who do not possess
Readings 133
. '' '''. '.''' '. '' '.''' ''' ''.' .. '''' ''.'.'.' '' ' ...... '. '.' ............... . this faculty of intuition can have no ground for denying that it is a cognitive faculty. For we can hardly nlaintain a priori that there are no \vays of discovering true propositions except those \vhich \Ve ourselves e1nploy. The ans\ver is that \Ve set no lhnit to the number of \Vays in which one may come to formulate a true proposition. We do not in any way deny that a synthetic truth may be discovered by purely intuitive methods as well as by the rational method of induction. But we do say that every synthetic proposition, however it may have been arrived at, must be subject to the test of actual experience. We do not deny a priori that the mystic is able to discover truths by his own special methods. We wait to hear \vhat are the propositions \vhich embody his discoveries> in order to see \vhether they are verified or confuted by our empirical observations. But the mystic, so far from producing propositions which are empirically verified, is unable to produce any intelligible proposi- tions at all. And therefore we say that his intuition has not revealed to him any facts. It is no use his saying that he has apprehended facts but is unable to express them. For we know that if he really had acquired any information, he would be able to express it. He would be able to indicate in some way or other how the genuineness of his discovery might be empirically determined. The fact that he cannot reveal what he "knows;' or even himself devise an empirical test to validate his "knowledge;' shows that his state of mystical intu- ition is not a genuinely cognitive state. So that in describing his vision> the mystic does not give us any information about the external \vorld; he merely gives us indirect information about the condition of his own 1nind.
These considerations dispose of the argtunent fron1 religious experience, \Vhich many philosophers still regard as a valid argument in favour of the existence of a god. They say that it is logically possible for men to be immediately acquainted with God, as they are immediately acquainted with a sense-content, and that there is no reason why one should be prepared to believe a man when he says that he is seeing a yellow patch, and refuse to believe him when he says that he is seeing God. The answer to this is that if the man who asserts that he is seeing God is merely asserting that he is experiencing a peculiar kind of sense-content> then \Ve do not for a Jnoment deny that his assertion 1nay be true. But, ordi- narily, the man who says that he is seeing God is saying not merely that he is experiencing a religious emotion, but also that there exists a transcendent being who is the object of this emotion; just as the man who says that he sees a yellow patch is ordinarily saying not merely that his visual sense-field contains a yellow sense-content, but also that there exists a yellow object to which the sense-content belongs. And it is not irrational to be prepared to believe a man when he asserts the existence of a yellow object, and to refuse to believe him when he asserts the existence of a transcendent god. For \vhereas the sentence "There exists here a yello\v-coloured nlaterial thing" expresses a genuine synthetic proposition \vhich could be e1npirically verified, the sentence "There exists a transcendent god" has> as \Ve have seen, no literal significance.
We conclude, therefore, that the argument from religious experience is altogether fal- lacious. The fact that people have religious experiences is interesting from the psychologi- cal point of view, but it does not in any way imply that there is such a thing as religious knowledge, any more than our having moral experiences implies that there is such a thing as moral knowledge. The theist, like the moralist, may believe that his experiences are cognitive experiences, but, unless he can formulate his "knowledge" in propositions that are empirically verifiable, we may be sure that he is deceiving himself. It follows that those
134 Part Two: Philosophy of Religion
.................... ' .................................................... .
O!J-[ Read "The \¥/ager,"
Number 233 in Section 3 of Penstes on \'IW\'1.mysearchlab .com
O!J-[ Read "The Will to
Believe," 1-5, on www .mysearchlab.com
philosophers who fill their books with assertions that they intuitively "know" this or that moral or religious "trutli' are merely providing material for the psychoanalyst. For no act of intuition can be said to reveal a truth about any matter of fact unless it issues in ver- ifiable propositions. And all such propositions are to be incorporated in the system of empirical propositions ,vhich constitutes science.
Alfred Jules Ayer, "Religious Discourse Is Meaningless:' from Language, Truth, and Logic (New York: Dover Books, 1936), pp. 114-20. Reprinted with permission.
BLAISE PASCAL
Belief in God-What Do You Have to Lose?
Blaise Pascal (1623-1662) was a follower of Jansenism, a branch of Catholicism that was active in France in the sixteenth to eighteenth centuries. Jansen ism \Vas based on the writings of the sixteenth-century Dutch theologian Cornelius Jansen and was inspired by the writings of St. Augus· tine; it was condemned by the Church as heretical. In a letter he had sewn inside his cloak, Pascal describes having a mystical experience that changed his life; it made him realize that "the God of Abraham, Isaac, and Jacob is not the God of the philosophers." Pascal was convinced that God is far too different from us for us to have any evidence concerning whether he exists. Pascal nonethe- less argued that it is rational to believe in God. In the following passage from his Pensees, a book that was posthumously assembled from his notes, Pascal sets forth his now-famous wager. In fact, he gives the \vager different formulations. Try to discern what these different arguments of Pascal's are and how they differ.
WILLIAM )AMES
The Will to Believe
William James (1842-1910) was an influential psychologist and philosopher who helped found the American school of philosophy known as pragmatism. In the present essay, James argues that we sometimes face decisions about what to believe that can and should be made on some basis other than the evidence at hand. His argument is an updated version of Pascal's Wager (Chapter io).
pa ................
Theory of l<nowledge
chapter 12 ..............................
What Is Knowledge?
Chapter Outline
Epistemology
Three Kinds of Knowledge
Two Requirements for Knowledge: Belief and Truth
Plato: True Belief Isn't Sufficient for Knowledge
Justification
Epistemology
The JTB Theory
Three Counterexamples to the JTB Theory
What the Counterexamples Have in Co1n1non
An Argument for Skepticism
In everyday life, in science, and in philosophy as well, we talk of"knowing" things. We also say that some beliefs are "strongly supported by evidence"; we say that they are "justified" or "well confirmed:' We apply these phrases not only to ourselves, but also to others.
Epistemology is the part of philosophy that tries to understand such concepts. Epis- temologists try to evaluate the common-sense idea that we (often, if not always) have knowledge and that we are (often, if not always) rationally justified in the beliefs we have. Some philosophers have tried to defend these common-sense ideas with philosophical ar- gumentation. Others have developed a philosophical position that involves denying these common-sense ideas. A philosopher who claims that we don't have knowledge, or that our beliefs aren't rationally justified, is defending some version of philosophical skepticism.
In this chapter, I'll begin with some remarks about the problem of knowledge. In the next chapter, I'll examine the views of the seventeenth-century French philosopher Rene Descartes. Descartes tried to show that we really do possess knowledge of the world; he tried to refute the skeptic. After evaluating Descartes' views about knowledge, I'll turn to an alter- native way of thinking about knowledge, the Reliability Theory of Knowledge (Chapter 14).
137
138 Part Three: Theory of Knowledge
After this discussion of the problem of knowledge, I'll examine the idea of rational justification. The eighteenth-century Scottish philosopher David Hume, whose views on the Argument from Design we considered in Chapter 5, argued that the beliefs we have that are based on induction aren't rationally justified. Hu1ne '\Vas a skeptic about induction. We'll consider his argument for this philosophical position and also the attempts some philosophers have made to show that Hume's startling thesis is mistaken. The last chapter in this part of the book discusses a central problem in epistemology-how do we know that things outside our O\Vn tninds exist?
So let's get started with the problem of knowledge. Before we can ask whether we know anything, we have to get clear on what knowledge is. To focus ideas, I want to distinguish three different ways we talk about knowledge. Only one of these will be our concern in what follows.
Three Kinds of Knowledge
Consider the differences that separate the following three statements, each concerning an individual, whom I'll call S (the subject):
(1) S knows how to ride a bicycle.
(2) S knows the president of the United States.
(3) S knows that the Rocky Mountains are in North America.
Right now I'm not interested in saying which of these statements is true. The point is that they involve different kinds of knowledge.
The kind of knowledge described in (3) I'll call propositional k11ow/edge. Notice that the object of the verb in (3) is a proposition-something that is either true or false. There is a proposition-the Rockies are in North America-and (3) asserts that S knows that that proposition is true. Statements (1) and (2) don't have this structure. For example, the object of the verb in (2) is a person, not a proposition. A similar kind of knowledge would be involved if! said that S knows Chicago. Statement (2) says that Sis related to an object-a person, place, or thing-so I'll say that (2) describes an instance of object k11owledge. Is there a connection between object knowledge and propositional knowledge? Maybe to know the president of the United States, you must know some propositions that are about him. But which propositions? To know the president, do you have to know what state he comes from? This doesn't seem essential. The same holds true for each other fact about the man; there doesn't seem to be any particular proposition you've got to know for you to know him.
There is another aspect of the idea of object knowledge, one that is rather curious. Suppose I've read lots of books about the president. I know as many propositions about him as you might wish. Still, it won't be true that I know him, because I've never met him. Kno\ving people see1ns to require some sort of direct acquaintance. But it is hard to say exactly what is needed here. If I once was introduced to the president at some large party, that wouldn't be enough for me to say that I "know" him. It isn't just direct acquaintance, but something more, that is required. I won't try to describe this further. I'll merely con- clude that propositional ki10\vledge, no n1atter ho\v volun1inous, isn't sufficient for object knowledge. Object knowledge requires some sort of direct acquaintance with the object. In contrast propositional knowledge does not require that you have direct acquaintance with the objects that the proposition is about; S can know that the Rockies are in North Ainerica though she has never seen then1.
Chapter 12: What Is Knowledge? 139
................ ' ..................................... ' ............. '.' .. . I turn next to the kind of knowledge described in statement (1). I'll call this know-how
knowledge. What does it mean to know how to do something? I think this idea has little connection with propositional knowledge. My son Aaron knew how to ride a bike when he was five years old. This means that he had certain abilities-he knew how to keep his bal- ance, how to pedal, and so on. If you asked a physicist to describe what Aaron was doing that allowed him to ride the bike, the physicist could write out a set of propositions. There would be facts about gravity, forward momentum, and balancing of forces. But Aaron wasn't a physicist at age five. He didn't know the propositions that the physicist specifies. Aaron obeyed the physical principles that the physicist describes-his behavior conformed
NECESSARY AND SUFFICIENT CONDITIO.NS
Consider the following suggested definition of what a bachelor is:
For any S, Sis a bachelor if and only if
(1) Sis an adult.
(2) Sis male.
(3) Sis unmarried.
I'm not claiming that this definition precisely captures what "bachelor" means in ordi- nary English. Rather, I want to use it as an example of a proposed definition.
The definition is a generalization. It concerns any individual you care to consider. The definition makes two claims. The first is that IF the individual has characteristics (1), (2), and (3), then the individual is a bachelor. In other words, (1), (2), and (3) are together sufficient for being a bachelor. The second claim is that IF the individual is a bachelor, then the individual has all three characteristics. In other words, (1), (2), and (3) are each necessary for being a bachelor.
We can define what a necessary condition is and what a sufficient condition is as follows:
"Xis a necessary condition for Y" means that if Y is true, then Xis true.
"Xis a sufficient condition for Y" means that if Xis true, then Y is true.
What does the expression "if and only if" mean in the proposed definition of bach- elorhood? It means that the conditions listed are both necessary and sufficient. A good definition will specify necessary and sufficient conditions for the concept one wishes to define.
This means that there can be two sorts of defect in a proposed definition. A defini- tion can fail to provide conditions that are sufficient. It can also provide conditions that aren't necessary. And, of course, it can fall down on both counts. Which sorts of defects are present in the following suggested definitions?
Sis a bachelor if and only if Sis male .
. Sis a bachelor if and only if Sis an unmarried human adult male who is tall and lives in Ohio.
The first definition is said to be "too broad"; it admits too much. The second is "too narrow"; it admits too little.
140 Part Three: Theory of Knowledge
''' .. ' . '' '. ''. '' '' '"'. '' .. '.' '''' '.'' ... '' ''' .. ' .. '''' '' '' ''' .. ' '' '''.''' to what they say-but he didn't do this by learning the propositions in question. Aaron had know-how knowledge, but little in the way of propositional knowledge.
Conversely, it is possible for a physicist to have detailed knowledge of the physical principles that describe successful bike riding and yet not know how to ride a bike. The physicist may lack the ability to perform the behaviors, but not because there is some prop- osition he or she fails to grasp. I conclude that propositional knowledge is neither neces- sary nor sufficient for knowing how to perform some task. The two concepts of knowledge are quite separate.
The subject of this and the following two chapters is propositional knowledge, not object knowledge or know-how knowledge. The goal is to understand what propositional knowl- edge is. We want to answer the following question: What are the necessary and sufficient conditions for it to be true that S knows that p, where pis some proposition-for example, the proposition that the Rockies are in North America?
Two Requirements for Knowledge: Belief and Truth
Two ideas that form part of the concept of knowledge should be noted at the outset. First, if S knows that p, then S must believe that p. Second, if S knows that p, then p must be true. Knowledge requires both belief and truth.
I won't try to argue for the first of these requirements, but will just assume it is correct. The second does require so1ne explanation, ho\vever. People so1netin1es say they kno\v things that, in fact, turn out to be false. However, this isn't a case of knowing things that are untrue, but of people thinking they know things that happen to be untrue.
Knowledge has an objective and a subjective side. You should remember this pair of concepts from Chapter i. A fact is objective if its truth doesn't depend on the way anyone's mind is. It is an objective fact that the Rocky Mountains are more than ten thousand feet above sea level. A fact is subjective, on the other hand, if it isn't objective. The most obvi- ous example of a subjective fact is a description of \Vhat's going on in so1neone's 111ind. Whether someone believes that the Rockies are more than ten thousand feet above sea level is a subjective matter, but whether the mountains really are that high is an objective matter. Knowledge requires both an objective and a subjective element. For S to know that p, p must be true and the subject, S, must believe that p is true.
I've just cited two necessary conditions for knowledge: Knowledge requires belief, and knowledge requires truth. Is that it? That is, are these two conditions not just separately necessary but also jointly sufficient? Is true belief enough for knowledge?
Plato: True Belief Isn't Sufficient for Knowledge
In the dialogue called The Theaetetus, the Greek philosopher Plato (ca. 430-345 B.c.), \Vho \Vas Aristotle's teacher, argues that the ans\ver is no. Orators and la\vyers so1neti1nes trick people into believing things; sometimes those things happen to be true. According to Plato, people who have been duped in this way have true beliefs, but they don't have knowledge.
Of course, the fact that orators and lawyers intend to deceive isn't crucial for Plato's point. Think of an individual, Clyde, who believes the story about Groundhog Day. Clyde thinks that if the groundhog sees its shadow, then spring will come late. Clyde puts this
Chapter 12: \Y/hat Is Knowledge?
silly principle to work this year. He receives the news about the behavior of the Official Designated Groundhog, and so he believes that spring will come late. Suppose Clyde turns out to be right about the late spring. If there is no real connection between the groundhog seeing its shadow and the coming of a late spring, then Clyde will have a true belief (that spring will come late), but he won't have knowledge.
So what else is needed, besides true belief, for someone to have knowledge? A natu- ral suggestion is that knowledge requires justification. The problem with Clyde is that he didn't satisfy this further requirement (though perhaps he thought he did). Justification, notice, can't just nlean that the subject thinks he has a reason; other,vise 've'd have to co'n- clude that Clyde knows how long winter will last just because he believes that he has a reason for thinking this.
Justification
What does it 1nean to say that an individual is "justified" in believing a proposition? Son1etin1es 've start believing that a proposition is true because 've consider an argument that describes the evidence available. Mendel believed that genes exist because of the data he collected from his pea plants. Sherlock Holmes believed that Moriarty was the murderer because of the evidence he found at the crime scene. Should we conclude that people are justified in believing a proposition only when their belief in the proposition was caused by their considering an argument? This isn't always plausible. When I believe that I have a headache, I don't construct an argtunent in 'vhich the evidence is laid out in a set of propositions that constitute the argument's premises. Still, it would seem that I am justified in believing that I have a headache. This suggests that there is such a thing as noninferential j11stificatio11. Some of the propositions we believe are apprehended more or less "directly"; they are not inferred frotn other propositions that 've believe.
What, then, does "justification" 1nean \\'hen it is used in accounts of 'vhat kno,vledge is? When we talk about someone's action being morally justified, we mean that the action does not violate any moral duties that the person has. Perhaps "justified belief" should be under- stood in a sitnilar "'ay. We should think of individuals as having certain duties concerning how their beliefs should be formed. A belief is justified if the process by which it was formed does not violate any duties that the person has. To inake sense of this suggestion, 've'd have to say what duties we have that govern how we are supposed to form our beliefs. Are we obliged to base our beliefs on the evidence that is available, and only on the evidence? This, I take it, is what I do when I believe that I have a headache. However, this suggestion requires further exploration; so1ne of the issues 'vere touched on in Chapter 10, on Pascal's Wager.
Even though the concept of "justification" requires further attention, it is plausible to think that this is one of the necessary ingredients that defines what knowledge is. Knowing that a proposition is true requires more than just having a true belief. The third require- ment is that your belief be "justified:'
The JTB Theory
Suppose knowledge requires these three conditions. Is that it? That is, are these conditions not just separately necessary, but jointly sufficient? I'll call the theory of knowledge that asserts this the ]TB Theo1y. It says that knowledge is one and the same thing as justified true belief:
141
Part Three: Theory of Knowledge
()TB) For any individual Sand any proposition p, S knows that p if and only if
(1) S believes that p. (2) pis true.
(3) Sis justified in believing that p.
The )TB Theory states a generalization. It says what knowledge is for any person Sand any proposition p. For example, let S be you and let p ~ "the moon is made of green cheese:' The )TB Theory says this: If you know that the moon is made of green cheese, then state- ments (1)-(3) must be true as well. And if you don't know the moon is made of green cheese, then at least one of statements (1)-(3) must be false. As in the definition of bach- elorhood discussed in the preceding box, the expression "if and only if" says that we are being given necessary and sufficient conditions for the defined concept.
Three Counterexamples to the )TB Theory
In 1963, the U.S. philosopher Edmund Gettier published a pair of counterexamples to the )TB Theory ("Is Justified True Belief Knowledge?" Analysis, 1963, Vol. 23, pp. 121-23). What is a counterexample? It is an exa1nple that goes counter to \Vhat son1e general theory says. A counterexample to a generalization shows that the generalization is false. The )TB The- ory says that all cases of justified true belief are cases of knowledge. Gettier thought his two examples show that an individual can have justified true belief without having knowl- edge. If Gettier is right, then the three conditions given by the )TB Theory aren't sufficient, even if they are necessary.
Here is one of Gettier's examples: Smith works in an office. He knows that someone will soon be promoted. The boss, who is very reliable, tells Smith that )ones is going to get the promotion. Smith has just counted the coins in )ones' pocket, finding there to be ten coins there. Smith therefore has excellent evidence for the following proposition:
(a) )ones will get the promotion and )ones has ten coins in his pocket.
Smith then deduces from this statement the following:
(b) The man who will get the promotion has ten coins in his pocket.
Now suppose that, unknown to Smith, Jones will not get the promotion. Rather, it is Smith himself who will be promoted. And suppose Smith also happens to have ten coins in his pocket. Smith believes (b ), and (b) is true. Gettier also claims that Smith is justified in believing (b), since Smith deduced (b) from (a). Although (a) is false, Smith had excellent reason to think that it is true. Gettier concludes that Smith has a justified true belief in (b ), but Smith does not know that (b) is true.
Gettier's other example exhibits the same pattern. The subject validly deduces a true proposition from a proposition that is very well supported by evidence even though it is, unbeknownst to the subject, false. I now want to describe a kind of counterexample to the )TB Theory in which the subject reasons nondeductively.
The British philosopher, logician, and social critic Bertrand Russell (1872-1970) de- scribed a very reliable clock that stands in a town square. This morning you walk by it and glance up to find out what the time is. As a result, you come to believe that the time
Chapter 12: \Vhat ls Knowledge? 143
.. ' ... ' ... ' '.' ''' ... '.' '.'.'' ' .. '.' .. '.'.'''' ' ... '' '.'' ''';'. '.'. '''''' is 9:55. You are justified in believing this, based on your correct assumption that the clock has been very reliable in the past. But suppose that, unbeknownst to you, the clock stopped exactly twenty-four hours ago. You now have a justified true belief that it is 9:55, but you don't know that this is the correct time.
Let me add a third example to Gettier's Smith/Jones story and to Russell's clock. You buy a ticket in a fair lottery. "Fair" means that one ticket will win and every ticket has the same chance of winning. There are one thousand tickets and you get ticket number 452. You look at this ticket, think for a moment, and then believe the following proposition: Ticket number 452 will not win. Suppose that when the drawing occurs a week later, you are right. Your belief was true. In addition, it was extremely well justified; after all, its probability was extremely close to unity-there was only one chance in a thousand that you would be mistaken. Yet, I think we want to say in this case that you didn't know that the ticket would fail to win. Here is a third case of justified true belief that isn't knowledge. Note that the reasoning here is nondeductive.
What the Counterexamples Have in Common
In all three of these cases, the subject has highly reliable, but not infallible, evidence for the proposition believed. The boss usually is right about who will be promoted; the clock usu- ally is right as to what the time is; and it usually is true that a ticket drawn at random in a fair lottery doesn't win. But usually doesn't mean always. The sources of information that the subjects exploit in these three examples are highly reliable, but not perfectly reliable. All the sources of information are prone to error to at least so1ne degree.
Do these examples really refute the )TB Theory? That depends on how we understand the idea of justification. If highly reliable evidence is enough to justify a belief, then the counterexamples do refute the )TB Theory. But if justification requires perfectly infallible evidence, then these examples don't undermine the )TB Theory.
My view is that evidence can justify belief without the evidence being infallible. I think we can have rational and well-supported beliefs even when we aren't entitled to be abso· lutely certain that what we believe is true. This is the point of strong inductive and abduc- tive arguments (Chapter 3). From this, I conclude that justified true belief is not sufficient for knowledge. The counterexamples have refuted the )TB Theory.
An Argument for Skepticism
What more is required? The lottery example suggests the following idea. In this case, you probably won't be mistaken when you believe that ticket number 452 will lose, but high probability isn't enough. To know that the ticket will lose, it must be impossible for you to be mistaken. You don't have knowledge in this example, because there is a chance (small though it may be) that you will be wrong.
There is some plausibility to this suggestion about what knowledge requires. The prob- lem is that this idea seems to lead immediately to skepticism-to the conclusion that we don't know anything. For it seems that virtually all the beliefs we have are based on evi· dence that isn't infallible. Consider, just briefly, the beliefs we have that depend on the tes- titnony of our senses. We use vision, hearing, touch, and so forth, to gather evidence about
144 Part Three: Theory of Knowledge
the way the world is. Do the resulting beliefs count as knowledge? The problem is that the senses are sometin1es tnisleading.
Right now, you believe you are looking at a printed page. You believe this because of the visual experiences you are having no\\'. Do you kno1v that there is a printed page in front of you? According to the present suggestion, for this to be true, it must be the case that you couldn't possibly be mistaken in believing what you do. But the fact of the matter seems to be that you could be mistaken. You might be hallucinating, or dreaming, or your senses n1ight be rnalfunctioning in some other \vay.
So here's where we are. The )TB Theory is mistaken. )TB doesn't suffice for knowledge. As an alternative to the )TB Theory, we have this suggestion: knowledge requires the im- possibility of error. This suggestion, plausible though it may be as a diagnosis of why you don't have knowledge in the lottery example, allows us to formulate the following argu- 1nent for skepticis1n:
If S knows that p, then it isn't possible that S is mistaken in believing that p.
It is possible that Sis mistaken in believing that p.
S doesrit know that p.
This form of argument can be used to argue that a posteriori knowledge does not exist. Recall from Chapter 8 that a proposition is a posteriori if it can be known only through the testi· 1nony of sense experience. The skeptic clailns that beliefs based on sense experience aren't totally immune from the possibility of error. People make perceptual mistakes-for example, in cases of illusion, hallucination, and drean1ing. Since \Ve can't absolutely rule out the pos- sibility of error, the skeptic concludes that om senses do not provide us with knowledge.
IS SKEPTICISM SELF-REFUTING?
Skeptics claim that people don't know anything. Can skeptics claim to know that what they say is true? If not, does that show that their philosophical thesis is false?
It is a contradiction to say that you know that no one (including yourself) knows anything. But a skeptic can assert that people lack knowledge without claiming to know that this is so. In addition, a skeptic can claim to provide a good argument for skepticism. This isn't contradictory.
The thesis that no one knows anything, if true, can't be known to be true. But that doesn't show that the position is false. If you think that all truths are knowable, you will say that this skeptical thesis can't be true. But why think that all truths are knowable? Why not think, instead, that the universe may contain truths that we are incapable of knowing?
Another way for skeptics to avoid contradicting themselves is to be modest. In- stead of claiming that no one knows anything, they could limit themselves to the claim that no one ever knows anything through the testimony of the senses. If this more limited kind of skepticism could be supported by an a priori argument, the position would not be self-refuting.
Chapter 12: \Y/hat Is Knowledge? 145
'.'' ' ....... ' '.'. ' .. ' '.' '.' ... ' ...... '''' '' ''' '' '.'' ..... '' '' '' '' .. '".''' This argument for skepticism is deductively valid. {Identify its logical form.) The first
premise seems to describe a plausible require1nent for kno'iving the proposition in ques- tion. The second pren1ise also seems plausible; it just says that the beliefs 've have aren't absolutely immune from the possibility of error. If you want to reject skepticism, you must refute one or both of these premises.
Although the above argument has fairly plausible premises, its conclusion is pretty outrageous. I think I know lots of things and I believe this is true of you as well. It is hard to accept the idea that I don't know that I have a hand. If you think that you now know there is a printed page in front of you, you too should balk at this argument for skepticism. The skeptical argu1nent contradicts a funda1nental part of our co1nn1011- sense picture of the \Vay 've are related to the \Vorld around us. Co1n1non sense says that people sometimes have knowledge of the world they inhabit; the skeptical argument says that conunon sense is rnistaken in this respect. If there is a mistake in the argu- ment, 'vhere is it?
Review Questions
1. Explain how object knowledge, kno\v-how knowledge, and propositional knowledge differ.
2. What do the following pieces of terminology mean? (i) Xis a necessary condition for Y; (ii) Xis a sufficient condition for Y; (iii) Xis true if and only if Y is true; (iv) 0 is a counterexample to the statement "All emeralds are green."
J. Why think that true belief isn't sufficient for knowledge? What is the JTB Theory? What is the dif- ference between highly reliable evidence and absolutely infallib!e evidence?
4- Describe the three counterexamples to the JTB Theory (Gettier's, Russell's, and the lottery ex- ample). Do these counterexamples sho\v that JTB isn't necessary or that it isn't sufficient for knowledge?
5. What is skepticism? \Vhat is the argument for skepticism given at the end of this chapter? Is the argument deductively valid?
Problems for Further Thought
l. People sometimes say, concerning a proposition that is difficult to take to heart, "1 knov1 it is true, but I don't believe it." Does this refute the claim that if S knows that p, then S believes that p?
2. I suggested that we lack knov.;]edge in the lottery example because kno\vledge requires the im- possibility of error. Can you think of a different explanation for why \Ve lack kno\vledge in this or in the other two examples that were used against the JTB Theory?
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
Part Three: Theory of Knowledge
'' ''., ... ' ...... ' .... '.'' .. ' ... ' .. '''.' ' .. '.'.''.' .... ' .. '' ....... ' ... ' .. WATCH AND LISTEN RESEARCH AND EXPLORE
©{WATCH (<•-{LISTEN 'l\l{EXPLORE
1. Listen 11Socrates"-BBC: In Our Time 1. Research Theaetetus
Socrates was a tremendously influential 2. Research Gettier cases philosopher, but surprisingly, he did not 3. Research skepticism leave any written works. Here, several philosophers and classicists explain why Socrates did not write anything down and discuss what we do know about Socrates from his students' accounts.
chapter 13 ..............................
Descartes' Foundationalism
Chapter Outline
Foundationalism
Euclid~ Parallel Postulate
Descartes' Method of Doubt
The Method Applied to a Posteriori Beliefs
Dubitability Is a Logical, Not a Psychological, Property
The Method Applied to Beliefs Based on Rational Calculation
I Am Thinking, Therefore I Exist
Thesis of the Incorrigibility of the Mental
Do First-Person Psychological Beliefs Provide a Sufficient Foundation?
An Additional Foundational Belief: God Exists and Is no Deceiver
How to Prove that God Exists
The Clarity and Distinctness Criterion
The Cartesian Circle
Conclusion
Rene Descartes (1596-1650) is sometimes described as the father of modern philosophy. The kind of epistemology he tried to develop is called foundationa/ism. Before launching into the details of Descartes' philosophy, I want to describe the kind of approach to the problem of knowledge that foundationalism provides.
Foundationalism
The word foundationalism should make you think of a building. What keeps a building from falling over? The answer has two parts. First, there is a solid foundation. Second, the rest of the building, which I'll call the superstructure, is attached securely to that solid foundation. Descartes wanted to show that (many if not all of) the beliefs we have about the world are cases of genuine knowledge. To show this, he wanted to divide our beliefs
147
Part Three: Theory of Knowledge
''' ''' ''' • ''' ''''' ','' ''' ''. ''' '''' • ''' '', '' '' ''.' ••• ''.'' •• ' '' '"' >'
CEH: Read Meditations
1-5 of Meditations on First Philosophy on \'NNl.tnysearchlab .com
into two categories. There are the foundational beliefs, which are perfectly solid. Then, there are the superstructural beliefs, which count as knowledge because they rest securely on that solid foundation.
Besides this metaphor from architecture, there is another that should help you understand what Descartes' project is. You probably had a geometry course in high school where you studied Euclid's development of the subject. Recall that Euclid, who lived about 2,200 years ago, divided the propositions of geo1netry into hvo categories. First, there are the axioms of geometry. These are supposed to be simple and totally obvious truths. Second, there are the theoren1s, 'vhich at first glance are so1ne,vhat less obvious. Euclid shows that the theorems are true by showing how they can be deduced from the axioms.
A foundationalist theory of knowledge could also be called a Euclidean theory of knowledge. To show that a given body of beliefs counts as knowledge, we use the following strategy: First, we identify the beliefs that will provide the foundations of knowledge (the axioms). These must be shown to have some special property, such as being totally beyond doubt. In a moment, I'll clarify what Descartes had in mind by this "special property:' Second, we show that the rest of our beliefs count as knowledge because they bear some special relationship to the foundational items. In Euclid's geometry, the special relationship 'vas deductive in1plication.
Euclid, of course, was interested only in the beliefs we have about geometry. Descartes had a wider ambition. He was interested in the totality of what we believe. But whether the problem is to describe the foundations of geometry or the foundations of knowledge as a whole, there are two ideas that must be clarified. We need to identify what the foun- dational items are and we need to describe the relationship that must obtain between foundational and superstructural items that qualifies the latter as knowledge. In doing so, we need to remember that the building metaphor is only a metaphor. What makes the foundation of a house solid is that the walls are thick and hard and are set deeply into solid ground. But foundational beliefs are1it foundational because they are thick and hard and planted deeply in dirt. What, then, does it mean, for a belief to be "foundational"?
Euclid's Parallel Postulate
If your high school geometry course was like mine, you spent most of your time seeing what theorems could be proved from the axioms. You spent little or no time seeing why the axioms should be regarded as true. You were told that they are "obvious;' or that they should just be accepted on faith. Maybe your teacher said that geometry is just a game and the axion1s are the rules. In fact, the question about the axioms is a serious one. One of Euclid's starting assumptions-his so-called fifth postulate-bothered geometers for about two thousand years. This postulate says that if you have a straight line (call it S) and a point not on that line, then there is exactly one straight line that goes through that point and is parallel to S. In other words, the parallel postulate says that if you extend all the lines passing through this point to infinity, exactly one of them will fail to intersect the initial line.
A number of subsequent geometers felt that this assumption of Euclid's was less obvi- ous than the other assumptions he used. As a result, they tried to show that it could be proved from the other axioms. Geometers repeatedly failed to do this, and eventually it was established that the parallel postulate cannot be proved from Euclid's other assumptions.
Chapter 13: Descartes' Foundationalis1n 149
''' '' ''''' ''' ', ''"''
That is, they showed that the parallel postulate is independent of the other assumptions. This means that the denial of the parallel postulate is logically consistent with Euclid's other assumptions. It then transpired that non-Euclidean geometries could be developed, ones that retain Euclid's other assumptions but reject his parallel postulate.
The details of this story don't matter here. The history of geometry does show, how- ever, that it son1eti1nes isn't so clear \Vhether a given state1nent is "obviously true:' Perhaps Euclid thought that the parallel postulate was obvious. If so, that would have been his jus- tification for treating it as so1nething that doesn't need to be proved; it \vould stand on its o\vn as an axio1n/postulate. This doesn't 1nean, ho\vever, that subsequent georneters had to regard Euclid's judgment as beyond question. What strikes one person as obvious may not be so obvious to so1neone else.
Descartes' Method of Doubt
Now let's consider Descartes' approach to the problem of knowledge. Descartes' goal is to refute skepticism. He wants to show that we really do have knowledge of the world we inhabit. His strategy for achieving this goal is foundationalist in character. This means the first item on his agenda is to identify the beliefs that are foundational.
At the beginning of his Meditations 011 First Philosophy (1641), Descartes proposes a method for determining which of his beliefs are foundational. He called this the method of doubt. For each proposition you believe, you see whether it is possible to doubt that propo- sition. If it is possible to do this, you set the belief aside-it isn't foundational. If it isn't possible to do so (that is, if the belief is indubitable), then the belief is foundational. Notice that failing the method of doubt test does1it show that the belief is false. It just means the belief isn't absolutely certain.
Descartes doesn't try to apply this method to beliefs one at a time. Think of all the be- liefs you have-there are millions (at least). You believe that 1 is a number. You believe that 2 is a nu1nber. And so on. It \Vould be thne constuning and also boring to try to consider each belief separately. Rather, Descartes considers kinds of beliefs; he considers whether all the beliefs in this or that category pass or fail the method of doubt test.
INDUBITABILITY
Very few propositions are made true just by your believing them. You can believe that there are unicorns, but that doesn't make such creatures pop into existence.
Consider the proposition "I am thinking." If you believe that this is true, then it must be true. Believing is a kind of thinking. Descartes thinks that the same is true for propositions of the form "I seem to see that p." He also thinks that this is true for propositions that say that you believe that p or want it to be true that q (where p and q are propositions). If you believe them, they must be true.
If a proposition is made true by your believing it, then you can't describe a sit- uation in which you believe the proposition but the proposition is false. Descartes determines whether a proposition is dubitable by seeing whether he can describe a situation of this kind. This means that a proposition is indubitable if the proposition is made true by the act of believing it (or doubting it).
(8-[ Read Meditations
i-5 of Meditations on First Philosophy on wv11.v.mysearchlab .com
150 Part Three: Theory of Knowledge
.. , . ' .. ' .... ' '' .. ' ................ , ................. ' ....... ' ...... '.' ... ' Are there indubitable propositions that aren't made true by your believing
them? Some philosophers have held that simple logical truths, such as "It is rain- ing or it ·1s not raining," are beyond doubt. Is it plausible to maintain that these are made true by your believing them?
The Method Applied to a Posteriori Beliefs
The first category Descartes considers is the set of beliefs that depend for their justifica- tion on sense experience. Many of our beliefs are based on sight, hearing, touch, taste, and smell. These propositions are a posteriori; recall this piece of vocabulary from Chapter 8. Is it possible to doubt these beliefs? Descartes says that the answer is yes. Your present belief that there is a printed page in front of you is based on vision. Vision, however, can be misleading. Psychologists tell us about hallucinations and illusions. Maybe you've had such experiences yourself. If not, remember Macbeth, who was certain that a dagger was hovering in front of him. Besides hallucinations, there is the fact of dreaming. In a dream you may find yourself believing that there is a printed page in front of you. You may find yourself having visual experiences just like the ones you are having now. But in the dream, your belief is mistaken. Descartes takes this to show that the belief you have right now might be mistaken.
So beliefs that rest on the testimony of the senses fail the method of doubt test. Lefs be clear on why Descartes thinks they do so. Descartes shows such beliefs are dubitable by con- strncting a story of a particular kind. You now believe that there is a printed page in front of you on the basis of a set of visual experiences. Let's call the proposition you believe Band the experiences you no\v are havingE. Your belief that Bis true rests on your having experiences E. Descartes holds that B can be doubted because he can describe a situation in which you have E and believe Band yet Bis false. Dreams and hallucinations show how this can happen. Des- cartes shows that a class of beliefs is dubitable by constructing a story of this sort.
Dubitability Is a Logical, Not a Psychological, Property
In saying that a belief is dubitable, Descartes isn't saying that we are able to take seriously the idea that it might be false. The proposition that there is a page in front of you is dubita- ble in Descartes' sense; this doesn't mean that you are now about to take seriously the pos- sibility that no page is present. A proposition is dubitable when a certain sort of story can be constructed; dubitability is thus a logical property that a proposition has. It has nothing to do with whether we can get ourselves to believe that the proposition might be false.
The Method Applied to Beliefs Based on Rational Calculation
Descartes next turns his attention to propositions of mathematics. I believe that 2 + 3 ~ 5. I believe that squares have four sides. Descartes remarks that these are true "whether I'm awake or asleep:' These propositions are a priori (recall this piece of vocabulary from Chapter 8). They don't depend for their justification on sensory experience. Do propositions justified by reason, independent of sense experience, pass the method of doubt test? Descartes thinks
Chapter 13: Descartes' Foundatio11alisn1
... ''.''''' .. '',',' '''' '''.''' '',,''' .. , ''', ... ', '.'' ''.',''.' ',' '. ',, .. they fail. To demonstrate that such beliefs are dubitable, Descartes asks us to imagine that our minds are deceived by an "evil demon:' Imagine that an evil demon causes our faculty of reasoning to find propositions totally obvious that in fact are false. If this were so, we might believe that 2 + 3 = 5 even though the proposition isn't true.
These conclusions-all reached in the first of the Meditations-are entirely negative. A posteriori beliefs about the character of the world outside the mind are dubitable. We see this by considering dreams and hallucinations. A priori propositions also are dubitable. We see this by considering the evil demon. If no belief in these two categories passes the method of doubt test, which beliefs could be the foundations of knowledge? What cat- egory of belief could possibly satisfy this very stringent requirement?
I Am Thinking, Therefore I Exist
In an earlier work, the Discourse on the Method (1637), Descartes identifies a pair of propo- sitions that pass the test:
I noticed that while I was trying to think everything false, it must be that I, who was thinking this, was something. And observing that this truth, I am thinking, therefore I exist [le pense, done je sttis in French; Cogito ergo su1n in Latin], was so solid and secure that the most extravagant suppositions of the skeptics could not overthrow it, I judged that I need not scruple to accept it as the first principle of the philosophy I was seeking.
In the Second Meditation, Descartes focuses on the belief"! am, I exist" as the first proposi- tion that he thinks is beyond doubt.
To understand what Descartes is driving at, you must think about yourself, formulat- ing your thoughts in the first person. When Descartes considers the proposition "! am thinking;' you aren't supposed to consider the proposition "Descartes is thinking:' Rather, you should say to yourself precisely what Descartes says to himself.
Consider my belief that I am thinking. I can't doubt this. It is impossible for me to construct a story in which I believe this proposition, though it is false. For if I believe the proposition, then I am thinking, and so the proposition is true. So the attempt to doubt the proposition fails; the attempt to doubt the proposition proves that the proposition must be true. The proposition "! am thinking'' passes the method of doubt test.
There are two important characteristics of the proposition "! am thinking" that you should note. First, it is important that the proposition is in the first person. When Descartes considers "I am thinking;' he concludes that this proposition passes the method of doubt test. However, if he had considered "Descartes is thinking;' the result would have been different. It is possible for Descartes to doubt that there is so1neone narned "Descartes:' He can invent a story in 'vhich there is no such person as Descartes; the fact that such a story can be constructed shows that the proposition is dubitable. The second feature of the proposition "I am thinking" is that it involves a psychological property. I can't doubt that I am thinking, but I can doubt that I am now in North America. Both beliefs are first- person, but only the former is psychological.
As noted, Descartes maintains that the proposition "I exist" also passes the method of doubt test. Take a few minutes to formulate an argument, similar to the one just described for "! am thinking;' that shows that this is so.
152 Part Three: Theory of Knowledge
So far, \\Te have t\vo propositions that Descartes thinks can serve as foundations for the rest of what we know. I can't doubt that I am thinking and I can't doubt that I exist. This is a meager foundation. Just as it would be hard to erect a big building on a foundation made of two bricks, so it would be difficult to ground the whole of what we know about the world on this paltry foundation of two beliefs.
There is tnore, ho'ivever. Consider propositions solely about the contents of your O\Vll present sensory experiences. Such propositions describe the \vay things see111 to be, not the \vay they in fact are. You 110"\V see1n to see a page in front of you. Descartes thinks that all such first-person descriptions of the way things seem are indubitable. To under- stand Descartes' point, it is essential to recognize the difference bet\veen the follo\\1ing t'iVO propositions:
'fhere is a page in front of 1ne.
I see1n to see a page in front of rne.
It is pretty clear, as noted earlier, that the first of these can be doubted. That's the point about drearns and illusions. Descartes inaintains that the second proposition is different. He holds that it has the property that if you believe that the proposition is true, then you can't be mistaken. If you believe that you seem to see a page in front of you, then you do seem to see a page in front of you. You emit be mistaken in your beliefs about the way things see1n.
So the foundation ofki10\vledge has just been aug1nented. "I an1 thinking" and ('I exist" are indubitable) and so are the many first-person beliefs about the 'ivay things seen1.
Thesis of the Incorrigibility of the Mental
Descartes went even further. He thought that people have infallible access to what they believe and desire. Introspection ("looking in'i\'ard") is a 1nethod that the rnind can use to accurately grasp its O'i\'n contents. This is sometin1es expressed by saying that Descartes believed in the thesis of the incorrigibility of the mental. Although "There is a page in front of me" isn't indubitable, Descartes thought "! believe that there is a page in front of me'' is indubitable. If you believe that you have that belief, then it must be true that you do have that belief. Ditto for desires. If you believe that you \Vant son1e ice crea1n, then it must be
true that you \Vant son1e ice cream. A great deal of work in psychology in the last hundred years shows that this incorri-
gibility thesis isn't plausible. Sigmund Freud, whose views on religion came up in Chapter 10, argued that we often misunderstand what we really believe and what we really want. At titnes, our beliefs and desires 'ivould be very upsetting to us if \Ve 'ivere conscious of them. So, as a defense tnechanisrn, our rninds repress thenl. The result is that \Ve often have mistaken beliefs about what we really think and want. An example of this is Freud's theory of the Oedipus complex. Freud held that little boys want to kill their fathers and marry their mothers. If you asked a little boy whether he wanted to do this, however, he prob- ably would sincerely answer no. You might well ask: What evidence can there be for this theory? If little boys deny having these desires, why think they have them? Freud's answer is that his theory is a plausible explanation of what little boys do. It is their behavior that we must consider, not just what they say they believe and desire. Verbal reports about what we
Chapter 13: Descartes' Foundationalistn 153
','''.'' '''''.'''' '''' ''.'' '.''.' '''' '''' .. '.' '' ' ..... '.'.' '' '.''.'' ' ... '' think and 'vant are evidence as to 'vhat our beliefs and desires really are. Sincere reports, ho,vever, don't settle the nlatter, since behavior 1nay provide other evidence that is relevant.
Although this idea-that we often have a false picture of our own beliefs and desires- is very in1portant in Freud's theories, it isn't unique to the1n. A great deal of other 'vork in psychology also denies Descartes' thesis of incorrigibility. See, for example, the 1977 article by R. Nisbett and T. Wilson called "Telling More Than We Can Know" (Psychological Review 84: 231-59). Descartes wants to include all first-person reports of his own beliefs and desires as foundations. Each, he thought, passes the method of doubt test. Freud and many other psychologists would disagree. I'm on their side.
Where does this leave Descartes' project of identifying the foundations of knowledge? The foundations include I am thinking and I exist and all first-person reports about the way the world seems to be. Another example of an indubitable belief might be "I am in pain:• Descartes' vie'v 'vas that if you believe that you are in pain, then you are in pain.
In Chapter 2, I pointed out that for a very large class of our beliefs, there is a huge differ- ence between believing a proposition and that proposition's actually being true. To believe that the Rockies are over ten thousand feet is one thing; for the mountains to actually be that high is another. It is misguided wishful thinking to hold that believing the proposition guarantees that it is true. This sensible separation of belief and truth, which seems right for a wide class of propositions, is called into question by the special examples that Descartes thought pass the method of doubt test. According to Descartes, "I am in pain" must be true ifI think it is. The same holds for "I am thinking" and "I seem to see a page in front of me:' These are cases in \\1hich wishful thinking seen1s to 'vork.
Do First-Person Psychological Beliefs Provide a Sufficient Foundation?
These propositions, Descartes thought, are foundational. We can't be mistaken in believ- ing them. Some philosophers have disagreed with Descartes' claim. They have maintained that it is possible to be mistaken in holding some of these beliefs. I'm not going to address that matter here. Instead, I want to consider another problem. Recall the architectural met- aphor discussed at the beginning of this chapter. An adequate foundation for a building must have two properties. It must be (1) secure and (2) sufficient to support the super- structure. The foundations of knowledge are subject to the same requirements. According to Descartes, they must be indubitable. In addition, it must be possible to rest everything else we know on them. Don't forget that the method of doubt is intended to identify the foundations; the entire superstructure is supposed to rest on that basis.
Descartes, as I mentioned before, wasn't a skeptic. He thought that we know lots of things about the world outside our own minds. For example, you know that there is a page in front of you now. This belief about the external world isn't indubitable; it didn't pass the method of doubt test. Nevertheless, it is (Descartes would agree) something you know. How can Descartes show that this is true? Recall the Euclidean analogy. We've just identi- fied some axioms of the system of knowledge. Do these suffice to prove some theorems? We must show how you know there is a page in front of you by showing how this proposi- tion is connected with foundational ones. In Euclidean geometry, we justify theorems by deducing them from axioms. How can a superstructural belief be shown to be knowledge?
154 Part Three: Theory of Knowledge
' ..... '.' ............... ' ........ ' .................. ' ................ '.' .. We must show that it is connected with foundational pieces of knowledge in the right way. But what is this special connection supposed to be?
Can deduction do the trick? Euclid deduced theorems from axioms. Can Descartes deduce propositions about the world outside the mind from beliefs that are first-person psychological reports? To show that you know that there is a page in front of you, maybe we should try to deduce that proposition in the following way:
I seem to see a page in front of me.
There is a page in front of me.
The premise of this argument is foundational. The problem is that it does not deductively imply the conclusion. The existence of a page in front of me does not deductively follow from the way things seem to me now. That's the point noted before about dreams and hallucinations. The above argument is not deductively valid.
The premise given is1{t enough. Even if I augmented it with other reports about my present psychological state (for example, "I seem to feel a page;' "! seem to taste a page;' etc.), that still would not be enough. Such premises might be ones about which I'm per- fectly certain, but they don't provide a sufficient foundation for the beliefs I have about the world outside my mind. Such arguments are not deductively valid.
An Additional Foundational Belief: God Exists and Is No Deceiver
Descartes saw quite dearly that the argument just described needs an additional premise, one that will bridge the gap between first-person psychological premises and a conclusion that describes the world outside the mind. Descartes thought the proposition that God exists and is no deceiver provides the additional premise that he needs.
How would this additional premise solve Descartes' problem? Descartes had the follow- ing picture in mind: God created your mind and situated you in the world. What sort of mind did God give you? Obviously, he didn't give you a mind that reaches true conclusions about the world on each and every occasion. Your mind isn't infallible. On the other hand, God would not have furnished you with a mind that leads you to false beliefs about the world no matter how carefully you reason and no matter how much evidence you consult. If God had done this, he would have been a deceiver. Descartes thinks that God created us with minds that have the capacity to attain true beliefs about the world. We are neither infallible nor hope- lessly trapped by falsehood. Rather, we are somewhere in between: We can reach true beliefs if we are careful about how we use the minds that God bestowed upon us. TI1is is the kind of mind that an all-PKG god would have given us. But why think that such a being exists?
How to Prove that God Exists
I've just explained what role the proposition that God exists and is no deceiver is supposed to play in Descartes' theory of knowledge. However, if this proposition is to be a part of the foun- dations, it must be indubitable. It must pass the method of doubt test. Descartes has to show that the proposition that God exists and is no deceiver is not just true, but that it is indubitable.
Chapter 13: Descartes' Foundatlonalism 1 55
• ' •.••••. ' ••• '. '.''.'' '.'.''.' '4.'' ''' '.'' ''.' ••• ' •• '' ••• ' ' ' •.•• ' •••• '. '.
Here Descartes seems to be trying to do the impossible. Isn't it obvious that it is possible to doubt that God exists and is no deceiver? After all, atheists or agnostics don't believe that God exists. Isn't this enough to show that it is possible to doubt that God exists and is no deceiver?
Descartes thinks that atheists and agnostics have not considered the matter carefully enough. They think they can doubt the proposition, but they have not really grasped what this would really involve. (Recall Anselm's comment about the "fool" who "said in his heart what cannot be conceived:') Descartes believes that he has a proof of the existence of God whose premises are indubitable. In addition, the proof is so simple that once you attend carefully to it, you can't doubt that there is a god. This is the proof that Descartes lays out in the Third Meditation. Here it is, in outline:
(1) My idea of God is an idea of a perfect being.
(2) There must be at least as much perfection in the cause as there is in the effect.
Hence, the cause of my idea is a perfect being-namely, God himself.
Once you inspect this proof, the proposition that God exists is like the proposition I am thinking-it is impossible to doubt that it is true.
Descartes thinks he knows premise (1) of this argument by introspection. By looking in- ward at the contents of his own mind, he can discern this fact about himself. Premise (1), Descartes believes, is indubitable. As to premise (2), Descartes thinks that it is an indubitable principle about causality. It can be broken into two components. First, there is the idea that ev- ery event has a cause. Second, there is the idea that the cause must be at least as perfect as the effect. It isn't clear why we should accept these principles about causality. For example, why couldn't so1ne events occur for no reason at all? This is a question \Ve explored in Chapters 4 and 7. Descartes needs to show not only that it is true that every event has a cause, but that this proposition is indubitably true. This is something that he does not succeed in doing.
Premise (2) needs to be clarified. Descartes distinguishes two kinds of perfection (or "reality") that an idea or representation might possess. To find out how objectively perfect a representation is, you must find out how perfect the thing is that the representation rep- resents. 1\vo photographs of a saint must have the same degree of objective perfection; two photographs of a trash can will have the same degree of objective perfection as well. If saints are more perfect than trash cans, the first pair will have more objective perfection than the second.
This characterization of objective perfection requires a little fine-tuning. A picture of a unicorn has a certain degree of objective perfection, even though there are no unicorns. So, to determine how objectively perfect the picture is, you must ask yourself this question: If the picture represented something that actually existed, how perfect would that thing be?
I hope you can see from the definition of objective perfection why Descartes thinks that his concept of God has the maximum amount of objective perfection. This is something that Descartes can say just by examining his concept of God. Indeed, it is something that an atheist might agree to as well. If there were a God, then that being would have all the perfections; this means that the atheist's concept of God is objectively perfect.
The concept of objective perfection allows some representations to be ranked higher than others. My idea of God is at the top. Somewhere below that is my idea of a saint. And
Part Three: Theory of Knowledge
''. ''',,''' '''' . '' ''.','.'.'. '''' ''''. '' '''' '' '''' still further down the list is my idea of a trash can. The second kind of perfection that Descartes discusses is different. This is what Descartes calls an idea's formal perfection. All mental contents have the same degree of formal perfection because they are all made of the san1e stuff. 'fo see \Vhat Descartes has in 1nind here, consider paintings. 'fhese n1ay differ in their degree of objective perfection, but all are n1ade of canvas and paint. In this sense, they have the same degree of formal perfection. Descartes thinks the same holds for all the ideas I may have.
In the argument for the existence of God that I just sketched, Descartes is talking about the objective perfection of his idea of God. Let's reformulate the argument to make this explicit:
(1) My idea of God is objectively perfect.
(2) If an idea is objectively perfect, then the cause of that idea must be a perfect being.
Hence, God exists.
Premise (2) strikes me as implausible. Here is a way to see that it is false. I claim that people could form their concept of God (a perfect being) in the following way. They could look at imperfect things in the world and thereby form the idea of limited intelligence, limited goodness, and limited power. Then, by applying the concept of negation (the concept we express by the word "not") to these ideas, they would obtain the idea of a being who has u11/i1nited intelligence, goodness, and power. This seerns to be an entirely possible causal explanation for why we have the idea of a perfect being. If it is possible, then (2) is false.
In the Third Meditation, Descartes explicitly considers this suggestion. He rejects the idea that we might have acquired the concept of perfection by seeing imperfect things and then applying the concept of negation. You should consider Descartes' argument and decide whether you think it works. Descartes thinks that the only possible explanation for the fact that we have an idea of a perfect being is that a perfect being actually exists and caused us to have this idea. Leaving this exercise to you, I'll conclude that Descartes' causal argument for the existence of God is defective. Descartes wanted to show that "God exists" isn't just true, but indubitably true. This proposition was to be a founda- tional element in the structure of our knowledge. The argument he gave falls short of this ambitious goal.
The Clarity and Distinctness Criterion
If God is no deceiver, how can we tell whether the beliefs we have are true? Descartes thought that if we inspect them carefully, and make sure that they are clear and distinct, we can be certain that they are correct. Descartes maintained that clear and distinct beliefs must be true. If you reason carefully and use your mental faculties in the way that God intended them to be used, you will obtain knowledge of the world you inhabit. Descartes thinks you have an indubitable grasp of the contents of your own mind; he also thinks that you can know indubitably that God exists and is no deceiver. This foundation, he believes, provides a sufficient and secure basis for you to gain knowledge of the world you inhabit.
Chapter 13: Descartes' Foundationalis111 157
''' ''' ,, ' '' '' '' ', '.', ,. ' ' ''.'' '' ''' '' '' ''''' ''' ','' ''' ''''.''' '.'.' '.''
The Cartesian Circle
Philosophers argue over whether Descartes' argument for the existence of God has a defect that goes beyond the fuct that some of his premises are dubious. They have suggested that the argument is circular, given its place in Descartes' larger philosophical program. Recall that Descartes wants to prove that God exists and is no deceiver in order to be able to conclude that clear and distinct ideas must be true.
Descartes' argun1ent for the existence of God, like all argu1nents> involves the use of reason. For us to recognize that the argument establishes the existence of God, \Ve must use the faculty of reasoning \\'e possess. Descartes' argument begins by exan1ining his con- cept of God and determining that this is the idea of a perfect being. Descartes therefore seems to be using the clarity and distinctness criterion in arguing that God exists. Des- cartes vehemently denied that he was guilty of circular reasoning. This problem has come to be known as the Problem of the Cartesian Circle.
Here's ho\v I see this issue: Descartes is using the n1ethod of doubt test to asse1nble foundations for knowledge. His goal is to assemble enough premises by this procedure so that he can show that we have knowledge of the world outside our minds. The existence of a God who is no deceiver is supposed to be one of these indubitable premises. Descartes thought his causal argument for the existence of God is beyond doubt; he thought that if you reason your way through the argument, you must find it irresistible.
Why should we agree that the proposition that every event has a cause is beyond doubt? Descartes' vie\v see1ns to be that \\'hen \Ve consider the proposition carefully, using the full resources of clarification and logic that 've possess, 've \vill be driven to conclude that the proposition must be true. This seems to mean that we are applying the clarity and distinctness criterion. If so, Descartes is reasoning in a circle.
Conclusion
Where does this leave the project of refuting skepticism? Descartes thought that first-person psychological beliefs are indubitable. He recognized that this, by itself, isn't enough to show that we have genuine knowledge of the world outside the mind. Additional premises are needed. The following example argument shows how Descartes wanted to bridge the gap between the indubitable knowledge we have of our own minds and the beliefs we have about the \vorld outside our n1inds:
(1) I now believe that the object in front of me is a page.
(2) My present belief is clear and distinct.
(3) Clear and distinct ideas are true.
There is a page in front of me.
The argun1ent is valid. If the pre1nises are true> the conclusion also inust be true. Descartes' foundationalis1n can be understood as a pair of claims about this argu1nent. First, he claims that the premises of this argument are indubitable. Second, he thinks that the con- clusion of the argun1ent is a proposition we kno1v because it follo\VS from these indubitable premises. I've already discussed premise (1). What of the next two premises? There is a tension bet\veen these t\VO clain1s.
Part Three: Theory of Knowledge
.. ' ..... ' ... ' ....... '.' ''' .......... ' ..... ,' ' ....... ' '' .. ' ' ' ' .. '.' ''''' ''. Suppose we grant Descartes that premise (3) is right-that a clear and distinct idea can't
fail to be true. Ifwe grant this, then it isn't so obvious that my present belief about what is in front of me really is clear and distinct. I may think that it is, but this appearance maybe deceiv- ing. So if (3) is right, (2) can't be beyond doubt. Alternatively, suppose we are able to tell, by in- trospection, whether a belief is clear and distinct. That will be enough to underwrite the truth of (2). In that case, there is no longer any absolute assurance that (3) is right. The problem here concerns 'vhether 've should see "clarity and distinctness)' as a purely subjective characteristic ofa belief. If! can tell just by examining the contents of my own beliefs whether they are "clear and distinct;' I see no reason to say that a clear and distinct belief must be true. On the other hand, if 've treat "clarity and distinctness)' as a characteristic that is necessarily connected 'vith truth, I don't see how I can tell whether a belief is "clear and distinct" just by introspection.
Descartes tried to refute the following skeptical argument, which I described at the end of the previous chapter:
Knowledge requires the impossibility of error.
It is now possible that I am mistaken in believing that there is a page in front of me.
Hence, I do1it know that there is a page in front of me.
Descartes accepted the first premise but denied the second. Descartes concedes that we often make mistakes. The senses sometimes play tricks on us, and so does our faculty of reasoning. But the fact that this sometimes happens doesn't show that it is happening now. If! am reasoning now in a careful and logically rigorous way, then I can't now be mistaken in what I believe, or so Descartes maintained. So Descartes' reply to the skeptic's argument can be put like this:
If God exists and is no deceiver and I now have a clear and distinct belief that there is a page in front of me, then I can't be mistaken in thinking that there is a page in front of me.
God exists and is no deceiver and I now have a clear and distinct belief that there is a page in front of me.
I can't be mistaken in thinking that there is a page in front of me.
The problem is that Descartes' argument for the second premise of this last argument wasn't successful. He wasn't able to prove that we are embedded in the world in the spe- cial way described by the hypothesis that God exists and is no deceiver. Consequently, he wasn't able to refute the skeptic's argument.
Review Questions l. \Vhat is foundationalism in epistemology? What does it have to do with Euclidean geometry?
With building a house?
2. What is the method of doubt test? What does Descartes use this method to do? What fails the test? What passes?
Chapter 13: Descartes' Foundatlonalism 159
. '.' ............................................... ' ........ ' ............ . 3. What does it mean to say that there is a "gap" between our first-person psychological beliefs and
the beliefs we have about the world outside the mind?
4. Descartes thought that proving that God exists would help show why we are able to have knowl- edge of the world around us. Why did he think this? Is the proposition that "God exists and is no deceiver" foundational, according to Descartes?
5. Analyze Descartes' causal argument for the existence of God. Suppose that one of Aquinas's arguments for the existence of God is valid and has true premises. Could Descartes have used it, instead of his causal argument, to prove that God exists?
6. What is the Problem of the Cartesian Circle?
7. In the previous chapter, an argument for skepticism was presented. How would Descartes evalu- ate this argument?
Problems for Further Thought
1. Which of the following propositions pass the method of doubt test? How are they different from each other?
I savJ Joe run.
I seem to remember seeing Joe run.
I remember seeming to see Joe run.
2. Can someone mistakenly believe that he or she seems to see a coffee cup? Can someone mistak- enly believe that he or she is in pain?
3. I suggested that our idea of a perfect being might be obtained by observing imperfect things and then using the concept of negation. Descartes considers this suggestion in the Third Medi- tation and rejects it (consult the paragraph that begins "Nor should I imagine that I do not perceive,,."). What reason does Descartes give for rejecting this explanation? Are his reasons plausible?
4. In his book The Principles of Philosophy (1644), Descartes defines clarity and distinctness as follows:
I term that "clear" which is present and apparent to an attentive mind, in the same way that we see objects clearly when, being present to the regarding eye, they operate upon it with sufficient strength. But the "distinct" is that which is so precise and differ- ent from all other objects that it contains within itself nothing but what is not clear ... Perceptions may be clear without being distinct, but cannot be distinct without also being clear.
Given this definition, can a clear and distinct idea be false?
MySearchlab Connections
Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
160 Part Three: Theory of Knowledge
••••••••••••••• ' •••••• ' •••• '' '' •• '''''. <.' ''.' ''' ''' •• ' "'' '''''.' ''' '''.
c::EJ-{READ AND REVIEW
READ
l. Rene Descartes, Meditations 1-S of Meditations on First Philosophy Why does Descartes pursue his method of doubt? For each of the types of belief that Descartes comes to doubt, think about whether you can doubt these beliefs as well.
WATCH AND LISTEN
©-[WATCH (!o~LISTEN
i. Listen "A. C. Grayling on Descartes' Coglto"-Philosophy Bites podcast
A C. Grayling, a biographer of Descartes, discusses Descartes's famous Cogito argu- ment and the role it ploys in his philosophy. Which skeptical hypotheses does Grayling discuss, and why does Descartes remain certain that he exists despite these skeptical worries?
2. Listen "Barry Stroud on Scepticism11- Philosophy Bites podcast
Descartes's method of doubt leads him to skepticism, the view that he cannot have indubitable knowledge about the world outside of his mind. Here, Barry Stroud discusses whether skeptical worries can be overcome. Why does Stroud think that we might have knowledge of the outside world even if we do not know that we are not dreaming?
References
RESEARCH AND EXPLORE
'Jlc{EXPLORE
1. Research Descartes 2. Research cogito }. Research foundationalism
Rene Descartes (1596-1650). "Discourse on the Method" (1637).
chapter 14 ..............................
The Reliability Theory of Knowledge
Chapter Outline
Descartes: Knowledge Is Internally Certifiable
What Makes a Thermometer Reliable?
Relevance to the Problem of Knowledge
Three Concepts of Impossibility
To Have Knowledge, You Don't Have to Be Able to Construct a Philosophical Argument Refuting the Skeptic
A Consequence of the Reliability Theory
Thesis of the Relativity of Knowledge
What Does the Relativity Thesis Say about Skepticism?
Descartes: Knowledge Is Internally Certifiable
According to Descartes, knowledge is internally certifiable. What does this mean? It means that if I know some proposition p, then there exists an argument that shows that p must be true, whose premises are either a priori true or knowable by introspection. Recall how Descartes would explain why I now know that there is a page in front of me:
I believe that there is a page in front of me.
My belief that there is a page in front of me is clear and distinct.
Clear and distinct ideas are true.
There is a page in front of me.
The first two premises I know by introspection-by gazing within my own mind and examining its contents. How do I know that the third premise is true? I know by
161
Part Three: Theory of Knowledge
••••• <'.' ''' >'.'''.'. '' ••• '' '''' ' •• '.'. ''''"'., •• ' •• '.' '' ••• ''' ••• , ••• ' ••
introspection that I have an idea of a perfect God; I (supposedly) know a priori that an idea of a perfect being must be caused by a perfect being. Putting these two thoughts together, I deduce that God-a perfect being-must exist. The third premise is sup- posed to follow from this.
Notice that the word "knowledge" does not appear in the displayed argument. So what does this argument have to do with the issue of whether I know that there is a page in front of me? For Descartes, I know that the concluding proposition is true because I know the premises are true. And how do I know the premises? I know these by introspection and a priori reasoning.
I \Vant to na1ne the ele1nents in this argu1nent. There is a subjective pren1ise, an objec- tive co11c/usion, and a linking premise. The subjective premise describes what is going on in the subject's mind ("I believe that there is a page in front of me"). The objective conclu- sion makes a claim about the world outside the subject's mind ("There is a page in front of me"). The linking premise (or premises) shows how the subjective premise necessitates the objective conclusion ("If I have a particular belief and it is clear and distinct, then it is true").
This vocabulary can be used to say what is characteristic of Descartes's approach to the problem of knowledge. His idea-that knowledge is internally certifiable-comes to this: If the subject k11ows that the objective co11c/usio11 is true, then the subject must know that the li11king premise is true and must know this independe11tly of sense experie11ce. By introspec- tion and a priori reasoning, I can establish the required connection bet\veen \vhat is inside the mind and what is outside it.
The theory of knowledge I'll discuss in this chapter is very different from Descartes'. It agrees \Vith Descartes that kno\vledge requires the existence of a connection bet\veen what is going on inside the mind and what is going on outside it. Knowledge requires that a particular linking premise be true. But according to the Reliability Theory of Knowledge, the linking premise doesn't have to be knowable by introspection and a priori reason- ing. Jn fact, this theory maintains that the subject doesn't have to know that the linking premise is true at all. It just has to be true. This approach to the problem of knowledge was first proposed by Alvin Goldman (in "A Causal theory of Knowing;' Journal of Philosophy, Vol. 64, 1967), Fred Dretske (in "Conclusive Reasons:• Australasian Journal of Philosophy, Vol. 49, 1971), and David Armstrong (in Belief, Truth, and Knowledge, Cambridge University Press, 1973).
What Makes a Thermometer Reliable?
The Reliability Theory of Knowledge claims that there is an important analogy between knowledge and a reliable measuring device. If you know that there is a page in front of you now, then your belief is related to the world outside your mind in the same way that the reading of a reliable thermometer is related to the temperature.
Thermometers are devices that form representations of temperature. The height of the mercury coluinn is the representation; the ainbient teinperature is the thing repre- sented. Thermometer readings represent temperature, just as your beliefs represent the \Vorld outside your inind. Ther1nometer readings can be accurate or inaccurate, just as beliefs can be true or false. What makes a thermometer reliable? Does this just mean that
Chapter 14: The Reliability Theory of Knowledge
........ '' ....... ' .. '. ' .. ' ... ' ... ''.' ... ' .. ' .. '.' .. ' ... '.'. '''' .. ' ' ' ' .. ''' its readings are accurate? To see \Vhy this isn't enough, consider a thermometer that is used just once. Suppose, on that one occasion, the ther1no1neter said "98°F" and the temperature happened to be 98'R The reading was accurate, but that doesn't mean that the thermom- eter was reliable. For all I've said, the thermometer may be broken-perhaps it would say "98°F" no matter 'vhat the true te1nperature is. If so, it isn't reliable, even though its one and only reading \Vas accurate. A reliable thern101neter is one for \Vhich there is a connec- tion bet\veen readings and temperatures. If the ther1no1neter reads n degrees Fahrenheit, then the temperature must be n degrees Fahrenheit. If a thermometer is reliable, then its reading 1nust be correct; it can't be inistaken in \Vhat it says. A stuck thermo1neter isn't reli- able, even \vhen its readings happen to be correct. It is unreliable because its readings are correct only by accident.
Do reliable thern1ometers exist? I think so. The mercury thertnometers we use to check whether we have fevers are examples. In saying this, I'm not denying two obvious facts. First, a thern1ometer can be reliable in one set of circumstances but not in another. A niercury thermometer \Vottldn't be very useful for measuring temperature if it \Vere wrapped in insulation before being placed in your mouth. The second point is that in saying that a thermometer is reliable, I'm not denying that it would be unreliable if it were broken. Hitting a reliable thermometer with a hammer will usually be enough to make it unreliable. So there are two things that help make a thermometer reliable. First, it has to be used in the right environ1nent (for exatnple, don't \vrap it in insulation if you \Vant to take somebody's temperature). Second, the internal makeup of the device has to be right (for example, the glass tube that holds the mercury can't be broken).
Notice that reliability is an objective feature of the relationship between the thermom- eter and its environment. The question is \vhether the thern1ometer and its environ1nent make the fullowing claim true: If the thermometer says that the temperat11re is n degrees Fahrenheit, then the temperat11re m11st be n degrees Fahrenheit. It is an entirely separate question whether anybody realizes that this thermometer/environment relationship obtains. Whether we notice this fact is a subjective question, but whether the relationship obtains is an objective 1natter.
Let's imagine that you take a thermometer out of a child's mouth, see that it reads 99'F, and then announce: "The thermometer is reliable, and so the baby's temperature is 99'F:' Suppose a contentious philosopher (like me) comes along and tries to refute your claim. I say:
Your thenno1neter isn't reliable. It is unreliable because I can conceive of a circum- stance in which its reading \vould be false. For example, I can conceive of a situation in \vhich the thermotneter is \Vrapped in insulation. In that case, the thennometer reading \vouldn't be correct. I also can conceive of the ther1nometer's being broken. In that case as well, the thermometer reading wouldn't be correct. It follows from the fact that I can conceive of these things that your thermorneter isn't reliable.
Should you be convinced by my argument? I think not. Whether the thermometer is reli- able here and now has nothing to do with what I can imagine. Granted, if the situations I described actually obtained, then the thermometer wouldn't be reliable. But from this it doesn't follow that the thermometer is unreliable in the actual circumstances in which it is used.
Part Three: Theory of Knowledge
. '.' .............. ' ............. ' ........... ' ...... ' ...... ' .... '.' ... ' ... ' Relevance to the Problem of Knowledge
How does this discussion of reliable thermometers bear on the problem of knowledge? The Reliability Theory of Knowledge says that an individual knows a proposition if the individual is related to the proposition the way a reliable thermometer is related to the tem- perature it measures. A reliable thermometer wouldn't say n degrees Fahrenheit unless the temperature were n degrees Fahrenheit. An individual knows that there is a page in front of her precisely when she wouldn't have believed that there is a page in front of her unless there were one there. Another way to express this idea is by using the concept of causality. A thermometer is reliable in a given circumstance if the only thing that could cause the thermometer to read n degrees Fahrenheit is that the temperature really is n degrees Fahrenheit. Similarly, S knows that there is a page in front of her in a given circumstance if the only thing that could cause S to believe this is that there really is a page in front of her.
I hope that the connection between thermometers and knowledge is starting to become clear. Suppose S believes that there is a page in front of her. We want to know whether this belief is an instance of knowledge. The answer should depend on the actual relationship that obtains behveen Sand her environment. Suppose S's sensory syste1n is functioning nor1nally; she isn't hallucinating, for example. In addition, suppose that there are no evil demons around who might choose to provide S with misleading evidence. If so, it may be true that S's sensory system is a reliable indicator of the presence of a page. If you want to answer the question of whether S knows that there is a page present, it will be entirely irrelevant to point out that if Shad taken a hallucinatory drug or if S were plagued by an evil demon, then S's belief wottld be (or might be) false. These hypothetical considerations do nothing to undermine the claim that S's sensory state is a reliable indicator of what is going on in her environment.
Here's how the Reliability Theory of Knowledge characterizes what knowledge is:
S knows that p if and only if (1) S believes that p. (2) p is true. (3) in the circumstances that S occupies, if S believes that p, then p must be true.
The third condition can also be formulated in either of the following two ways:
In the circumstances that S occupies, S wouldn't believe that p unless p were true. • In the circumstances that S occupies, it is impossible that S believes p and p is false.
Notice that the Reliability Theory of Knowledge makes use of the concepts of necessity and impossibility. To understand this theory of knowledge, we need to look more carefully at what these concepts mean.
Three Concepts of Impossibility
I want to discuss three kinds of impossibility. Consider the following three statements:
(1) Joe can't be a married bachelor.
(2) Joe can't go faster than the speed oflight.
(3) Joe can't tie his shoes now.
Chapter 14: The Reliability Theory of Knowledge
'.' '.''' '''' ...... '''' '''.' ... '. '''.' '''' ". ' ...... ''' ................. . The word "can't" in each of these indicates that something is impossible, but different kinds of impossibility are involved.
The first statement is logically necessary. It has to be true, just by virtue of logic and the definitions of the terms involved. If definitions and their deductive consequences are a priori true (Chapter 8), then (1) expresses a necessary truth that is a priori.
In contrast, statement (2) is1it a priori. We kno\\T that it is true because Einstein's theory of relativity-an a posteriori theory-says that it is so. Statement (2) is said to be nomologi- cally necessary (fron111on1os, meaning "}a,v"). It is necessary because of a la'" of nature, in this case, an a posteriori law of physics.
Statement (3) differs from the first two examples. Its truth doesn't follow from logic or definitions. Nor does its truth follow from any law of nature-physical or biological or whatever. Rather, to see that (3) is true, you have to take into account particular facts about Joe. For example, suppose that Joe can't tie his shoes now because he is carrying several bags of groceries. I'll call this third sort of necessity circumstantial necessity. By this I mean that the statement is necessarily true because of facts about the circumstances in which Joe finds himself. There is an additional fact about type (3) necessity that I should note. Whether we judge (3) true or false depends on how we interpret it. That is, (3) is ambiguous. If we inter- pret it to mean that Joe can't tie his shoes because he is otherwise occupied, (3) may be true. Suppose, however, we take (3) to mean that even if Joe put down the bags of groceries, he still couldn't tie his shoes. If that is how we interpret (3), we may judge it to be false. This is an important feature of stateinents that are circun1stantially necessary-they are atnbiguous.
When I say that a thern1on1eter is reliable, I 1nean that it and its environn1ent are related in a special way. I'm saying that its circumstances are such that its readings must be correct. In saying this, I'm using the concept of circumstantial necessity. Similarly, when I say that S knows that there is a page in front of her, I'm saying that she is related to her environment in a special way. I'm saying that she ca1it be mistaken in believing what she does. Here again, I'1n using the concept of circumstantial necessity.
Suppose that a real printed page is the only thing that could get S to believe that a printed page is in front of her. Her senses are functioning normally. There are no evil demons lurking about who provide misleading evidence. If this is so, and if S subsequently believes the proposition in question, then Swill know that there is a printed page in front of her. This is what the reliability theory says. In this circumstance, her belief will be related to the world the way the reading of a reliable thermometer is related to the temperature.
To Have Knowledge, You Don't Have to Be Able to Construct a Philosophical Argument Refuting the Skeptic
It is important to notice that the claim that S knows some proposition p isn't refuted by the fact that Smay not be able to defend her knowledge claim against clever philosophi- cal interrogation. If I asked Show she knows that her Impressions aren't caused by an evil demon, she may draw a blank. If I ask her how she knows that she is1it dreaming, she may admit that she emit construct an argument proving that she isn't dreaming. Such philo- sophical puzzles may even lead S to say, "I guess I don't know that there is a printed page in front of me:' But her comment doesn't show that she lacks knowledge. Rather, it shows only that S doesn't believe that she knows there is a printed page in front of her.
166 Part Three: Theory of Knowledge
...... , ..................................... ' ........... ' ................ . If Sis like a reliable thermometer, she has knowledge. Being like a reliable thermometer,
however, doesn't require that S has the ability to construct fancy philosophical arguments that show she is like a thermometer. Thermometers can't construct philosophical argu- ments, yet they are sometimes reliable. S may be similar. She may be unable to refute the skeptic, but that doesn't show she lacks knowledge. S may even be a skeptic herself. She may believe that she lacks knowledge, but that doesn't entail that she really lacks it. Just as it is possible to mistakenly believe that a thermometer is unreliable, so it is possible to mis- takenly believe that an individual lacks knowledge. In fact, it is possible to have false beliefs about one's own situation; Scan believe that she lacks knowledge and yet be mistaken.
THE KK·PRINCIPLE
According to the Reliability Theory of Knowledge, knowing a proposition doesn't imply that you know that you know it. That is, the reliability theory rejects what is called the KK-principle. The following several paragraphs about this principle will be somewhat tough to grasp-so take a deep breath, read them slowly, and then read them again.
The KK-principle says that if S knows that p, then S knows that she knows that p. But I've just argued that the following is possible: S knows that p, but S doesn't believe that she knows that p. It follows that S might know that p even though she doesn't know that she knows that p. So if I'm right, the KK-principle is false.
If the KK-principle were true, S could prove that she lacked knowledge just by be- coming a philosophical skeptic. This would be very convenient-wouldn't it be nice to be able to prove some thesis just by believing it?
Suppose S believed, maybe for no reason at all, that she didn't know that pis true. The KK-principle would conclude from this that, indeed, S failed to know that pis true. To see why, consider the following argument:
(1) If you don't believe that you know that p, then you don't know that you know that p.
(2) If you don't know that you know that p, then you don't know that p.
If you do not believe that you know that p, then you don't know that p.
The argument can be symbolized as follows:
(1) lf-B[K(p)], then -K[K(p)].
(2) lf-K[K(p)], then -K(p).
lf-B[K(p)], then -K(p).
The argument is deductively valid (what is its logical form?). Notice that premise (1) is correct; it follows from the fact that knowing a proposition requires that you believe the proposition. Premise (2) would be true if the KK-principle were correct. So if the KK-principle were correct, believing that skepticism is true would be enough to ensure that skepticism is true. However, if knowing is like being a reliable thermometer, this has got to be wrong.
Chapter 14: The Reliability Theory of Kno\vledge
. '''.' '' '.' .... '' ' .. ' ... ' .. '.' ...... ' ... '.' '' ... ' ......... '.''.'.'.' ' .... ' The Reliability Theory of Knowledge holds that whether an agent knows some-
thing is settled by the objective relationship that obtains between the agent's belief and the environment. Whether the agent believes that this relationship obtains is irrelevant. The reliability theory rejects the KK-principle.
In summary, the Reliability Theory of Knowledge explains why S knows that there is a page in front of her by describing the following argument:
S believes that there is a page in front of her.
In the circumstances S occupies, she wouldn't believe that there is a page in front of her unless there were a page in front of her.
There is a page in front of S.
This argument doesn't use the word "know:' So what does it have to do with knowledge? According to the reliability theory, S knows that the concluding proposition is true be- cause the premises are true. But S doesn't have to know that the pre1nises are true, nor does S have to produce an argument independent of sense experience for the premises. This shows why the reliability theory rejects Descartes' claim that knowledge is internally certifiable.
One virtue of the reliability theory is that it explains what is wrong with a standard skeptical argument. The skeptic claims that S doesn't know that there is a printed page in front of her on the grounds that it is possible to imagine that Sis deluded in some way. You can imagine a situation in which S believes what she does but she is mistaken. The reliabil- ity theory shows why this act of imagination is irrelevant to the question of whether S has knowledge in the real world situation that she occupies.
A Consequence of the Reliability Theory
To conclude this discussion, I want to consider an implication that the reliability theory has. It involves a fact about circumstantial necessity that I mentioned before. This is the fact that claims of circumstantial necessity are often ambiguous. I'll restate the point and then show how it is relevant to the reliability theory.
Consider the statement ((Joe can't tie his shoes now:' Suppose I make this co1nment to you while we are looking at Joe, who is carrying two heavy bags of groceries. Is the quoted remark true? Here are t\vo 'vays of interpreting it:
1. If Joe tries to tie his shoes while holding the bags of groceries, he will fail.
2. If Joe tries to tie his shoes after first putting down the groceries, he will fail.
If we interpreted "Joe can't tie his shoes now" to mean (i), it will be true; if we interpreted it to mean (2), it will be false. Which reading you choose affects whether you will say that the quoted claim is trne.
To see how this idea applies to the Reliability Theory of Knowledge, let's switch examples-from knowing that there is a printed page in front of you to knowing that there
168 Part Three: Theory of Knowledge
. '.'" .... ' ''.' '' '.''' ''''''' '.'' '' .. ' ''' ''' '' ''.'' '''''''' '' is a barn in the field next to the road on which you are driving. Suppose one day you go for a drive in Dane County, "\Vhich is 'vhere Madison, Wisconsin is located. On your drive, you look at a field and say, "There is a barn in that field'.' Suppose that the proposition you've asserted is true. You believe the proposition. Do you know that it is true? The reliability theory says that '"'e tnust ask "\vhether, in the circu111stances you occupy, there is anything besides a real barn that could have caused you to believe a barn is present. The problem I want to focus on is that it isn't clear how to decide what is included and what is excluded by the expression "in the circun1stances you occupy:'
First, I need to define the idea of a fool~ barn. Fool's gold looks like gold, but isn't gold at all. A fool's barn is something that looks like a barn but isn't one. Suppose there are no fool's barns in Dane County. The only things that look like barns are real barns. This 1neans that if I describe the circu1nstances you occupy by saying that you are in Dane County, then it will be true that the only thing that could have made you think a barn is present is a real barn. Hence, you know that there is a barn in the field.
Now let's broaden our vision. In Hollywood, there are fool's barns. These are the fa- cades used in movie sets. When viewed from one angle, they look just like barns, but they aren't buildings at all. If you were driving around Hollywood and came to believe that a barn is present, your belief would not count as knowledge (even if your belief happened to be true). The reason is that, around Hollywood, there are things besides real barns that can make you believe that a barn is present.
So if your circu1nstances are restricted to Dane County, you 1night ki10\V that a barn is present; if your circu1nstances are restricted to Holly,vood, you 'vouldn't. But no"r con- sider a puzzle. I initially suggested that the circumstances you occupy are limited to the ob- jects in Dane County. Given this array of objects, I argued that a real barn is the only thing (in the circumstances you occupy) that could cause you to think that a barn is present. But 'vhy describe your circu1nstances so narro"1ly? Why not describe your circun1stances as including all the objects in the United States? If! describe your situation in this way, it will be false that the only thing that could cause you to believe a barn is present is a real barn.
Here is the point: When we assert or deny that the only thing in the circumstances that could cause you to believe that a barn is present is a real barn, \Ve are 1naking reference to the environment you occupy. The environment may be thought of as composed of a set of objects. If we describe that set narrowly, it may be true that the only thing in the set that could cause you to believe that a barn is present is a real barn. If we describe the set more broadly, this may no longer be true. So whether you know, on your drive through Dane County, that there is a barn on the hill next to the road depends on how we choose to de- scribe the "circumstances you occupy." There are n1any true descriptions \ve n1ight select. Narrow ones will entail that you have knowledge; broader ones will entail that you do not.
Thesis of the Relativity of Knowledge
The claim I just made about knowledge may be put this way: Knowledge is relative. Whether S knows that p depends on (is relative to) a choice. "Sis in Dane County" and "Sis in the United States" are equally correct descriptions of the circumstances that S occupies. Which one \Ve consider depends on our interests.
Chapter 14: The Reliability Theory of Knowledge
''''.'' ''''''' ''' '',' '''', '''' '''.' '''.' '' ''' '.'''' '' '''
Let's be clear what this relativity thesis asserts. When you say that something is relative, you should always be prepared to say what it is relative to. The thesis that knowledge is relative is the thesis that whether an agent knows a proposition depends on (is relative to) a specification of his or her circumstances. 1'hese circu1nstances can be specified in differ- ent equally correct 'vays, and so there is no unique ans\ver to the question of 'vhether the agent has knowledge.
Let's consider a do,vn-to-earth exa1nple of relativity that does not concern kno,vledge. Mary and Alice are walking down the street, side by side. Is Mary walking to the left of Alice? Well, that depends on the point of view. If you look at them from the front, you get one answer. If you look at them from the back, you get the other. The key thing to notice is that x's being to the left of y is1it a relation between x and y alone. It involves some third iten1, z.
Does S know that p? The suggestion now being considered is that the question is in- complete, that is, whether S knows that p depends on a third item besides the agent S and the proposition p. It depends on a specification of S's environ1nent. There are different true 'vays to describe S's environ1nent. S is in Dane County. But it is also true that S is in the United States. Relative to one specification of the environment-call it E,-S knows that p. But relative to another equally true specification of the environment-call it E,-S doesn't know that p.
What Does the Relativity Thesis Say about Skepticism?
The thesis that knowledge is relative has an interesting implication about skepticism. Skepticism, recall, is the position that people don't have knowledge. Its opposite is the more common-sense idea that people often (if not always) know the propositions they believe. If the relativity thesis is true, then each of these theses is true in one sense but false in another. If the agent's environment is given a very broad specification, then skep- ticisn1 is true. If the environ1nent is given a narro,ver specification, then the con1mon- sense position \Vill be correct. Notice, ho,vever, that there is no conflict bet\veen the following two claims:
(1) S knows that p, relative to E,.
(2) S doestit know that p, relative to E,.
These state1nents don't conflict \Vith each other, any 1nore than the next pair conflicts 'vith each other.
(3) Mary is walking to the left of Sue, when they are viewed from the front.
(4) Mary is not walking to the left of Sue, when they are viewed from the back.
So the dispute bet\\1een skepticisn1 and conunon sense seems to end in a stalemate, not in a victory for either side, if the relativity thesis is correct. Each position is correct in one sense but incorrect in another. This consequence follows from the Reliability The- ory of Knowledge once we acknowledge that claims about circumstantial necessity are a1nbiguous.
Part Three: Theory of Knowledge
••••••• ' •••••• ''.' •••••• ' •••••• ' •••• '' ••• '' •••••• ''.' '. < '. '' •••••• " •• ' •• '.
WHAT'S RELATIVE ABOUT EINSTEIN'S THEORY OF RELATIVITY?
Sometimes people say that Einstein proved that "Everything is relative." This isn't true. Neither is the more modest claim that Einstein proved that everything in physics is relative.
To say that a statement is relative isn't to say that it is subjective or arbitrary. Rather, it is to say that the statement is incomplete in a particular way. Whether Sue is walking to the left of Mary is a relative matter. That means that there exists some third item, beyond the two people mentioned, that must be referred to if the statement is to be true or false. To say just that Sue is to the left of Mary is to fail to express a complete thought.
It takes no enormous insight or creativity to see that whether one person is walk- ing to the left of another is a relative, not an absolute, matter. In contrast, Einstein's idea that simultaneity is relative is anything but obvious. It was a brilliant theoretical conjecture, something that isn't at all suggested by our common-sense talk about space and time. Consider two events that occur at different places. Common sense suggests that either the events are simultaneous with each other or they aren't. Ein- stein's theory says that this isn't true. Whether two events are simultaneous depends on the choice of some third item-something called a rest frame. Relative to one rest frame, the events are simultaneous, but relative to another equally correct rest frame, they aren't. Simultaneity is relative, not absolute.
Einstein didn't say that everything in physics is relative. For example, in the special theory of relativity he defines a quantity called the space-time interval. This quantity measures the amount of separation that there is between events; it takes account of both spatial and temporal distances. The space-time interval between two events isn't relative to the choice of a rest frame. The space-time interval is absolute, not relative.
Review Questions
l. At the beginning of this chapter, I said that Descartes held that knowledge is "internally certifi- able." What does this mean? Does the Reliability Theory of Knowledge agree?
2. What is the difference between logical necessity, nomological necessity, and circumstantial necessity?
3. What does it mean to say that claims about circumstantial necessity are "ambiguous"?
4. \Vhat does it mean to say that a thermometer is "reliable"? What analogy does the Reliability Theory of Knowledge see bet\veen reliable thermometers and knov,rledge?
5. How does the Reliability Theory of Knowledge assess the following skeptical argument: "I can imagine that my senses are now malfunctioning. Hence, I don't know now that there is a printed page in front of me."
6. What does it mean to say that knowledge is "relative"? Relative to what?
7. Does the relativity thesis entail that skepticism is correct?
Chapter 14: The Reliability Theory of Knowledge 171
''.'''' '''". '' '' ' .. '. '.' '' '.' '''. ','.' .. '. '.'' .. ' ' ' ' '·' '.'' ''.' '.' ''' '.' '. Problems for Further Thought
l. For any proposition p, we can construct the sentence "S knows that p." Since "S knovJs that lemons are yello\v" is a sentence, we can construct the sentence "S knows that S knows that lemons are yellov,r." \Ve can repeat this operation as many times as \Ve please. The KK-princip!e says that if S kno\vs that p, then S knows that S knows that p. Formulate similar principles for truth (the TI-principle), belief (the BB-principle), and surprise (the SS-principle). Are any of these principles plausible? Defend your ansv1ers.
2. \Vhether a ring is made of gold depends on the materials that the jeweler chooses when the ring is made. Does this mean that whether a ring is made of gold is a relative matter? How is this example different from the t\vo claims about relativity discussed in this chapter?
3- Do you think that nonhuman organisms have knowledge? For example, do dogs? What is required for dogs to have knowledge, according to the reliability theory? How do these require- ments differ from what would be demanded by the view that knowledge is internally certifiable? If dogs have knowledge, does this count for or against the Reliability Theory of Knowledge?
MySearchlab Connections Watch. Listen. Explore, Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific !~and chapter exam assessments available at www.mysearchlab.com.
RESEARCH AND EXPLORE
oJifc{EXPLORE
i. Reliabilism
chapter 15 ..............................
Justified Belief and Hume's Problem of Induction
Chapter Outline
Knowledge versus Justified Belief
Skepticism about Justified Belief
Hume's Skeptical Thesis about Induction
Htune,s Argument that Induction Can't Be Rationally Justified
Why Can't PUN Be Justified?
Su111111ary ofHume,s Argument
In this chapter, I will present a new problem in epistemology. So far, I've discussed the con- cept of knowledge, examining Descartes' foundationalism (Chapter 13) and the Reliability Theory of Knowledge (Chapter 14). The problem to be considered here concerns the idea of justified belief.
Knowledge versus Justified Belief
Before starting, let's be clear on how this problem differs from the previous one. What is the difference between knowledge and justified belief? Two differences should be clear. First, if S knows that p, then p must be true; whereas S's having a justified belief that p does1tt require that p be true. Knowledge requires truth, but justified belief does not. The second difference is that knowledge requires the impossibility of error; justified belief does not. Recall the lottery example discussed in Chapter 12. Suppose you believe, before the drawing, that ticket number 346 will not win in a fair lottery containing one thousand tickets. Suppose your belief turns out (after the drawing) to be true. You didn't know before the drawing that that ticket wouldn't win, but you nonetheless had strong evidence for thinking the ticket would lose. Your belief was reasonable, or justified. (I'll use these two terms interchangeably.)
Chapter 15: Justified Belief and Hun1e's Problem of Induction 173
'' .. '''.' '''' ''.'' '''". ' ... '' '''. '' '.''' '''.''. ''''.'.'.'' "''' '' .. ''' ''' These comparisons of knowledge and justified belief are summarized in the following
table:
Knowledge
Justified belief
Is truth required?
Yes
No
Is the impossibility of error required?
Yes
No
As the lottery example illustrates, the evidence you have for many of your beliefs falls short of providing absolute certainty. We sometimes use the concept of probability to describe such cases. We say that ticket number 346 probably won't win, to indicate that we aren't absolutely sure that it will not. Similarly, you might also say that your present experiences make the beliefs you have about the world outside your mind highly probable. Your pres- ent experiences don't make it absolutely certain that there is a printed page in front of you, but it does seem plausible to say that your experiences make it very probable that there is a printed page before you.
Skepticism about J ustifled Belief
Until now I've talked about skepticism as if it were a single thesis. Now I want to distin- guish skepticism about knowledge from a second kind of skepticism, which concerns the idea of justified belief.
Although skepticism about knowledge sounds like a shocking thesis at first hearing, once one grasps \vhat it says) it doesllt fundamentally undertnine our picture of ourselves and the world we inhabit. If knowledge requires the impossibility of error, perhaps we should concede that we don't know many things. True, we often talk about the knowledge we have, but perhaps this is just sloppy talk. This concession to skepticism isn't terribly threatening because it allows us to hold onto the idea that our beliefs about the world aren't groundless and arbitrary. We can still say that they are "reasonable" and "well justi- fied;' even if they aren't entitled to be labeled "knowledge:' Skepticism about knowledge doesn't entail skepticism about rational belief. This is why it is possible to abandon the claim that we have knowledge without thereby giving up the idea that our beliefs are rational.
There is a second form of skepticism that is far more disturbing. This form of skepti- cisn1 rejects our claiin to rationality. According to this 1nore radical for111 of skepticism, we aren't rationally justified in believing what we do. It is this type of skepticism that is involved in Hume's views on induction. Hume claimed that the beliefs we have about the future and the beliefs we have concerning generalizations can't be rationally justified. For exan1ple, it isn't just that \Ve can't be certain that the sun \viii rise to1norro,v. And ifs not just that \Ve don't knoiv that the sun \Vill rise ton1orro\v. According to Htune, \Ve have no rational justification at all for this or for any other expectation we have about the future.
CEJ--{ Read Section IV
of An Enquiry Concerning Hutnan Understanding on \'tww.mysearchlab .com
174 Patt Three: Theory of Knowledge
Hume's Skeptical Thesis about Induction
We constantly form expectations about what the future will be like or about which gen- eralizations (statements of the form ''.All As are B") are true based on evidence that isn't deductively conclusive. Our beliefs about the future are based on perception and mem- ory, but you can't deduce what the future will be like from premises that describe just the present and the past. Let's focus on an example in order to get clear on this shnple point. Suppose I have observed many emeralds and have found each of them to be green. I then formulate the prediction "The next emerald I observe will be green", or maybe I formulate the generalization ''.All emeralds are green:' (Just so this can be an example of the kind I want, suppose that it isn't a definitional truth that all emeralds are green.)
Conunon sense says that 've are rational in believing the predictions and generaliza- tions we do if those beliefs are based on lots of evidence. Looking at lots of emeralds and finding that each is green seems to justify my expectation that the next emerald I exam- ine '\Vill be green. No\v it is obvious that \VC can't deduce generalizati.ons or predictions from our past observations. Yet, the following appear to be perfectly sensible nondeductive argun1ents:
I've observed ntunerous en1eralds, and each has been green. (GEN)
Hence, all emeralds are green.
I've observed nu1nerous emeralds, and each has been green. (PRED)
Hence, the next emerald I observe will be green.
I've dra,vn a double line bet\veen pre1nises and conclusions to indicate that these argu- ments aren't deductively valid. In both the generalization (GEN) argument and the pre- diction (PRED) argument, we think that the conclusion reached is rationally justified by the premises. We think that it isn't a mere prejudice to hold, in each case, that the premise provides good evidence for the truth of the conclusion. Hume's thesis is that this convic- tion can't be rationally defended.
It is important to see that Hume's thesis goes far beyond the undisturbing claim that the GEN and PRED arguments just given aren't deductively valid. That much is obvious. Rather, Hume is saying that the premises in those arguments don't rationally justify their conclusions. Hume's view is that there is absolutely no rational justification for the beliefs we have that are predictions or generalizations. He thinks it is merely a habit we have that 've regard such premises as providing good reason to believe such conclusions. This is a habit we can't abandon; it is part of human nature to expect the future to resemble the past. But it is a habit we ca1rt rationally defend. When challenged by the skeptic to rationally justify this pattern in our thinking, we can say only that this is the way human beings in fact operate. We can't produce a good argument to rationally justify this habit of mind.
Chapter 15: Justified Belief and Hurne's Problem of Induction
Hume's Argument that Induction Can't Be Rationally Justified
Why did Hume reach this startling conclusion about induction? (Here I'll reconstruct the argument that Hume gives in Part II of the section from An Enqui1y Concerning Human Understanding called "Skeptical Doubts Concerning the Operations of the Understanding:') Hume thought that the arguments GEN and PRED require an additional premise. As they stand, the premise doesn't support the conclusion. If the observation is to support the generalization or the prediction, then \Ve 1nust assu111e that the future will resemble the past. This assumption Hume calls the Principle of the Uniformity of Nature (PUN).
Hume thought that this principle plays an indispensable role in each and every in- ductive argument \\'e 1nake. The example described above concerns the color of einer- alds. Consider, however, the belief that the sun will rise tomorrow. This predictive belief is based on the premise that the sun has risen on each of the days that we have bothered to make an observation. Why should these past observations support the prediction you n1ake about totnorrow? Hu1ne thought that you must be assun1ing that nature is uniforn1- that the future will resemble the past. Hume says that without this assumption, the past would be no guide to the future.
So each and every inductive argument presupposes PUN; we must assume PUN if the observational premise is to support the prediction or the generalization stated in the argument's conclusion. This means that if the conclusion we reach is rationally defensible, then a good argument must be available for thinking that PUN is true. If PUN can't be defended, then anything you believe that depends on assuming that PUN is trne must likewise be indefensible.
We now can state Hume's skeptical argument:
(1) Every inductive argument requires PUN as a premise.
(2) If the conclusion of an inductive argument is rationally justified by the prem- ises, then those premises must themselves be rationally justifiable.
(3) So, if the conclusion of an inductive argument is justified, there must be a ra- tional justification for PUN.
(4) If PUN is rationally justifiable, then there must be a good inductive argument or a good deductive argument for PUN.
(5) There is no good inductive argument for PUN, since any inductive argument for PUN will be circular.
( 6) There cannot be a good deductive argument for PUN, since PUN is not a priori trne, nor does PUN deductively follow from the observations we have made to date.
(7) So, PUN is not rationally justifiable.
Hence, there is no rational justification for the beliefs we have that take the form of predictions or generalizations.
175
Part Three: Theory of Knowledge
' '''' '"'.'.'' '' ''''''' ''''''' ''' '' '''
In a nutshell, Hu1ne's claitn is that the beliefs "Te have about etnerald color and toinorro\v's sunrise (and lots of other beliefs as well) aretit rationally justifiable. This is because they rest on an asstunption that can't be rationally justified.
Why Can't PUN Be Justified?
Let's look more carefully at steps (4)-(6) in the argument. Consider what PUN asserts; it says that the future will resemble the past-that past uniformities will continue to obtain in the future. ls this something that we could know to be true on the basis of induction? If it were, the inductive argument would look like this:
Nature has been uniform in 1ny past observations.
Nature in general is uniform.
Recall that Hume claims that all inductive arguments require that PUN be assumed as a pretnise, but the above argun1ent is inductive. Notice that if "\Ye insert PUN as a pren1ise, as Hun1e requires, the argurnent becomes circular-it assuines as a pretnise the very proposi- tion it tries to establish as a conclusion.
What about the other sort of justification? Can we give a deductive justification of PUN? Here again, Hume holds that the answer is no. The argument istit deductively valid; the general uniformity principle catit be deduced from the observations I've made in the past.
Hume considers, and rejects, a second possible sort of deductive argu1nent. Could PUN be a definitional truth, deducible from the definitions of the terms it uses? If PUN "\Vere a definitional truth, it \vould have the same sort of a priori justification that ''All bachelors are unmarried" possesses. Hu111e rejects this idea by saying that there is ''no contradiction" in supposing that the universe should suddenly cease to be uniform. It istit a definitional truth that past regularities \Vill continue into the future, Htune says.
Summary of Hume's Argument
It should now be clear how Hume's skeptical argument proceeds. First, there is a claim: Every inductive argu1nent requires the premise that nature is uniform. Second, Hun1e ar- gues that no rational justification can be given for that premise. He defends this assertion by considering three options: (1) an inductive argument in favor of PUN, (2) a deduction of PUN from past observations, and (3) a deduction of PUN from definitions. He asserts that none of these ways of defending PUN is going to work. This means that PUN can't be justi- fied. Hume concludes that the inductive inferences we make aretit rationally justifiable.
Review Questions
l. What is the difference between knowledge and justified belief? Does skepticism about knowledge entail skepticism about rational justification? What do each of these skeptical theses assert?
2. It is obvious that vie can't deduce what the future will be like from our present observations and memories. Does Hume's skeptical thesis say anything more than this?
Chapter i5: Justified Belief and Hurne's Problem of Induction
...... ' .. ' .......... ' '.' ... '' .. ' ... '. ' ... ''.' '.' '''.' .. '. '''.' ' .. '''' ). \Vhat is the Principle of the Uniformity of Nature?
4. \Vhat is Hume's skeptical argument about induction?
Problems for Further Thought
l. Is the claim that nature is uniform falsifiable (Chapter g)? That is, can you describe a possible observation that would count against this claim?
2. Is it possible that all of our beliefs could be rationally justified? Would this be possible if we had a flnite number of beliefs? In answering, explain what you think it means to say that a belief is rationally justified?
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, re- search tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
r:E}{READ AND REVIEW
READ
1. Hume, Section IV of An Enquiry Concerning Human Understanding Hume argues that there can be no rational justification of the inductions that people perform. The reason is that all such inductions depend on the Principle of the Uniformity of Nature, and this principle cannot be given a rational justification. Why can't the Principle of the Uniformity of Nature be given a deductive justification? Why can't it be given an inductive justification?
RESEARCH AND EXPLORE
:)ic{ EXPLORE
1. David Hume 2. Uniformity of nature 3. Problem of induction
177
chapter 16
Can Hume's Skepticism Be Refuted?
Chapter Outline
What, Exactly, Does the Principle of the Uniformity of Nature Say?
A New Concept: Degrees of Reliability
What Is a Rule of Inference?
Does the Past Reliability of Induction Provide an Answer?
Hu1ne,s Argun1ent Refor1nulated
Strawson: It Is Analytic that Induction Is Rational
Black: Induction Can Be Inductively Justified
In the previous chapter, I described Burne's skeptical argu1nent about induction as focus- ing on the Principle of the Uniformity of Nature (PUN). This principle, recall, says that the future will resemble the past. In this chapter, I'll criticize this version of Hume's argument. Then I'll formulate Hume's argument in another way. After that, I'll examine two criticisms that have been inade of Hume's startling argu1nent. Again, you \Vant to be clear on \Vhat Hu1ne's conclusion is. He didn't merely say that our predictions are uncertain; he said that they have no rational justification whatever. These predictions are rationally indefensible, but we can't help making and believing them-they are consequences of a habit that is deeply entrenched in human nature.
What, Exactly, Does the Principle of the Uniformity of Nature Say?
PUN says that the future will resemble the past. Recall that Hume held that every time we inake an induction-inferring that the sun will rise toinorro\v or that the next emerald \Ve observe \Vill be green-\ve are assurning that nature is uniforn1. So PUN is supposed to be an assun1ption that is required by each and every inductive argu111ent. But what, exactly,
Chapter 16: Can Hume's Skepticism Be Refuted? 179
does PUN 1nean? Does it n1ean that nature is uniforn1 in each and every respect? If so, it is pretty clear that we don't think that PUN is true. Summer leaves are green, but we do1ft expect autu1nn to rese1nble su1nmer in this respect. s'o if \Ve understand PUN in this \vay, the principle is not, contrary to Htune, something \Ve are ahvays assuming.
Let's try a second interpretation. Maybe PUN means merely that the future will rese1nble the past in so111e respect. This does seen1 to be so1nething \Ve believe. This is such a modest principle, however, that it is hard to see how it will help very much in the task of inferring what the future will be like. If! want to know whether future emeralds will be green or blue based on my observation that so far they all have been green, this modest version of PUN is pretty useless. It does1{t tell me whether it makes sense to expect emer- alds to remain green or to expect a change.
These two attempts to clarify what PUN says illustrate a general problem. No one yet has been able to clarify the principle so that it has both of the following properties:
(1) PUN is something we believe.
(2) PUN gives definite advice about what we should infer from the observations \Ve have made.
There is a third characteristic that Hun1e's argun1ent about induction attributes to PUN. This is the idea that PUN plays an indispensable role in the way we think about the world:
(3) If we want to make inductive inferences about the world, PUN is something we n1ust believe, no inatter \Vhat else \Ve believe.
This last condition expresses Hume's idea that PUN is an assumption-a presupposition- that underlies the whole project of inductive inference.
If it is hard to formulate PUN so that it satisfies (1) and (2), it will be even harder to clarify the principle so that it satisfies (1)-(3). I have1{t proved that it is impossible to do this. But no one has done so yet. For this reason, I suggest we drop the formulation of Hume's argument that involves PUN. Instead, I'll now describe a different version of Hume's argument that induction can't be rationally justified.
A New Concept: Degrees of Reliability
This ne\v version of Hun1e's argun1ent involves the idea that a method of inference pos- sesses some degree of reliability. Induction is a method of inference. It uses observations to tnake predictions and to say \vhat generalizations are true. We can evaluate a inethod of inference by deter1nining ho\v often the predictions or generalizations it endorses have been true. A method that usually leads to truth is highly reliable; one that rarely does so is very unreliable.
This use of the concept of reliability departs somewhat from the one employed by the Reliability Theory of Knowledge (Chapter 14). There I discussed reliability as an on/off concept; a ther1no1neter is either reliable or it isn't. Moreover, a reliable ther11101neter, in the sense en1ployed there, niust n1ake true clahns about the te1nperature. Notice that the con- cept of degrees of reliability isdt an on/off concept. In addition, a highly reliable method
180 Part Three: Theory of Knowledge
. '.' ' .. ".' '. '' .. '' ....... ' ..... "' .... ' ...... '' ...... ' .......... ' ........ . of inference can sometimes lead to falsehood. The point is that it does so rarely. A method of inference is highly reliable if the predictions it makes are usually true.
Induction is a rule of inference that scientists use to evaluate hypotheses in the light of evidence. A scientist makes a number of observations-call them 0. The goal is to say which hypothesis among a set of competing hypotheses is most plausible in the light of those observations. For example, suppose you've examined numerous e1neralds and all have been green. The problem is to say which of the following hypotheses is better sup- ported in the light of those observations:
(H,) All emeralds are green.
(H,) All emeralds are green until the year 2050; thereafter, they are blue.
Common sense suggests that H, is more plausible than H, in the light of the observations 0. The philosophical question is to explain why this is so. Hume denied that this evaluation of the competing hypotheses can be rationally defended.
What Is a Rule of Inference?
Rules of inference provide licenses. A fishing license permits you to fish. A rule of infer- ence per1nits you to dra\v conclusions. In the previous example, induction per1nits you to conclude that H, is probably true, given 0. That is, induction is a rule of inference that connects pre1nise to conclusion in the follo\ving argument:
I've examined lots of emeralds and all have been green. I[
All emeralds are green.
I've drawn a double line to indicate that the argument is not supposed to be deductively valid. I've written an "!" beside the double line to indicate that the rule of inference being used is induction. The problem of induction is to explain why we are entitled to use this inference rule.
Does the Past Reliability oflnduction Provide an Answer?
Common sense may suggest that \Ve are no\v entitled to use induction because induction has been reliable in the past. Induction has often been used to make predictions, and the predictions endorsed by inductive arguments usually turned out to be correct. This is why we rightly take seriously what induction tells us. Hume rejected this attempt to justify in- duction. Let's examine the attempted justification more carefully. Here is the argument we need to consider:
Induction has been highly reliable in the past.
Induction will be highly reliable now and in the future.
The premise of this argument is something we know by having observed the past track record of the inductive method. The conclusion makes a claim about what will be true no\v and in the future.
Chapter 16: Can Hume's Skepticis1n Be Refuted?
, ' ' ' ' ' ' ' ' ' ' < ' , ' ' ' ' ' ' • ' < ' ' ' ~ ' ' ' ' ' ' ' ' ' ' ' ' , ' ' , ' ' ' ' ' , ' , ' ' ' ' ' ' ' ' ' • , '
Every argument 1nust use a rule of inference to license the transition fron1 its pre1n- ises to its conclusion. What rule of inference is used here? Can the conclusion be validly deduced from the premise? No. That is why I've drawn a double line. This is an inductive argun1ent; induction (I) is the rule of inference that is being used, so an "f' should be inscribed next to the double line.
Hume's point about this argument is that it is circular or question-begging. The philo- sophical question is whether induction can be rationally justified. This argument simply assumes that induction is legitimate. You emit justify using induction now by appeal to the fact that induction has been successful in the past.
Hume's Argument Reformulated
So Hu1ne's skeptical argun1ent about induction can be for1nulated as follo\vs:
(1) To rationally justify induction, you must show that induction will be reliable.
(2) To show that induction will be reliable, you must construct an inductive argu- ment or a deductively valid argument.
(3) You cant show that induction will be reliable by giving an inductive argument; that would be question-begging.
(4) You can't validly deduce that induction will be reliable from premises describ- ing the past reliability of induction (or from definitions).
Hence, induction cannot be rationally justified.
This argument does1it say anything about the PUN. Can it be refuted?
Strawson: It Is Analytic that Induction Is Rational
Hun1e claimed that a rational justification for using induction n1ust sho'v that the inethod will probably be reliable. The British philosopher Peter Strawson (in Introduction to Logi- cal Theor;\ Methuen, 1952) challenged this. Strawson rejects premise (1) of Hume's argu- ment. Strawsoris idea is simple. He thinks that the statement "Induction is rational" is like the statement "Bachelors are unmarried:' Both are a priori truths. Both are deductive consequences of the definitions of the terms that occur in the1n. According to Stra\vson, induction is, by definition, a rational activity.
Strawson doesn't claim that "Induction will be reliable'' is an a priori truth. He agrees with Hume that this isrit true by definition. But he thinks that you dorit need to estab- lish that an inference rule will be reliable if you want to show that using the rule is ratio- nal. Hume thought that rationality requires reliability. Strawson denies this. According to Stravvson, it is entirely rational to use inductive 1nethods to forn1ulate our beliefs about the future, even though we can offer no good reason for expecting that the method will prob- ably lead to true beliefs. If it turns out later that we were entirely mistaken about what we thought the world would be like, no one can accuse us of having been irrational. Even if we turn out to be mistaken in our beliefs, we still behaved perfectly reasonably when we formed them.
181
Part Three: Theory of Knowledge
... '''. '' '" .... ' ...... '.'' ... ' ' .... "'''' ... '. ' .... ' ...... '.''.' .. '.''.' ' .. I find Strawson's argument unconvincing. Methods of inference are methods for
doing things. Consider an analogy-recipes. Suppose I told you that I have a great rec- ipe for making a cake. To determine whether this is true, you would want to see if the method is reliable. The point of the recipe is to produce nice cakes. You evaluate the recipe by seeing if it is a reliable instrument. It would puzzle you if I said that a particular recipe is excellent even though I granted that there is no reason whatever to think that using the recipe will probably result in a nice cake. This, however, is just what Strawson is saying about induction. The point of induction is to reach true beliefs about the world. That is the goal we have when we make inferences. Whether it is reasonable to use in- duction or some other method depends on whether the method is a reliable instrument for attaining the specified goal. If it can't be shown that the instrument will attain the goal, or even that it probably will, how can using the instrument be rationally justified? Strawson's mistake, I think, is to hold that the rationality of a method has nothing to do with its reliability.
Black: Induction Can Be Inductively Justified
I turn now to an attempt at providing an inductive justification of induction. The American philosopher Max Black (in "Inductive Support of Inductive Rules;' Problems of Analysis, Cornell University Press, 1954) argued that once you look carefully at what circularity means, you'll see that the inductive justification of induction isn't circular at all. That is, Black holds that the following argument provides a perfectly good reason for accepting the conclusion:
Induction has been highly reliable until now. I[
Probably, induction will be highly reliable if we use it now and in the future.
Once again, I use the letter "!" to indicate that the rule of inference that licenses passing from the premise to the conclusion is the principle of induction.
What does Black think it means for an argument to be circular? His suggestion is roughly as follows: if an argument is circular, then the conclusion occurs as one of the premises. This isn't Black's exact formulation, but it captures the spirit of what he says. For Black, circularity involves a relationship between the conclusion and the premises. It has nothing to do with what rule of inference is used to get from premise to conclusion. I hope it is clear, if you accept Black's definition of circularity, that the above inductive justifica- tion of induction isn't circular. The point I want to make, however, is that Black's definition is too narrow. I think there is a broader definition of circularity that shows that the above argument is defective. It is this: An argument is circular if it couldn't possibly convince people that the conclusion is true if they didn't believe the conclusion already. A circular argument can't change anybody's mind. If you have doubts about induction, the above argument is not going to lay those doubts to rest.
Besides this, there is another problem confronting Black's suggestion. Consider a bizarre rule of inference that is the mirror i1nage of induction. It is called counterinduction. Induction tells you to expect past regularities to continue into the future. Counterinduction
Chapter 16: Can Hume's Skepticism Be Refuted?
tells you to expect past regularities not to continue. For exa1nple, counterinduction licenses the following inference:
I've examined lots of emeralds and all have been green.
I[
Future emeralds will not all be green.
Of course, induction would lead to a quite different conclusion. The problem of induction is, in part, to explain ,vhy \Ve should use induction rather than counterinduction to for1nu- late our predictions about the future.
What has this to do with Black's proposal? Black claims that the inductive justification of induction isn't circular. Ifhe is right, then neither is the following argument:
Counterinduction has been highly unreliable until now.
CI[
Probably, counterinduction will be highly reliable now and in the futme.
Here \Ve have a counterinductive justification of counterinduction. Counterinduction has had a poor track record in the past. How should we expect it
to do in the future? That depends on what rule of inference we use. If we use induction, \Ve \vill expect counterinduction's dismal past track record to continue into the future. If, however, \Ve use counterinduction, \Ve expect the pattern to reverse. Since the nlethod has been unreliable in the past, counterinduction tells us to expect that counterinduction will be reliable in the future. So we have two arguments to compare: the inductive justi- fication of induction and the counterinductive justification of counterinduction. Which method of inference-induction or counterinduction-should \Ve use now? The problen1 with Black's inductive justification of induction is that it gives us no reason to reject the counterinductive justification of counterinduction. This nleans that Black's favored induc- tive justification of induction, even if it is noncircular in Black's sense of the term, doesn't sho\v \vhy \Ve should use induction rather than counterinduction. Hu1ne's probletn has not been solved.
Is it possible to refute Hume's argument? We have examined two attempts to do so. I think both fail. This does not mean that there is no way to prove Hume wrong. In the next chapter, I'll consider this problem from a different angle.
Review Questions
1. When we try to clarify exactly \vhat the Principle of the Uniformity of Nature says, we run into a problem. What is the problem?
2. Ho\v can Hume's skeptical argument about induction be formulated without mentioning PL!N?
3. What is Strav1son's justification of induction? What objection to it \Vas presented in this chapter?
4. What is Black's inductive justification of induction? What l\vo objections to it were presented in this chapter?
5. \Vhat does it mean to say that a rule of inference is a "license"? What is the difference betv1een the license provided by induction and the license provided by counterinduction?
Part Three: Theory of Knowledge
. '' '''''' '.'' '''' ''.''' ''' ... '.' ''' '.'' '.' Problems for Further Thought
l. The problem of induction involves explaining why the observation of many emeralds, all of them green, supports H1 better than H2 :
H1: A!I emeralds are green. H2 : All emeralds are green until 2050; thereafter, they are blue.
Can the Surprise Principle (Chapter J) be used to show that the observations strongly favor H, over H2 ? Explain your answer.
2. Hume considers, and rejects, hvo kinds of arguments that might justify induction-deductive arguments and inductive arguments. Would considering abductive arguments change the sort of conclusion that Hume reaches about induction?
3- Consider the rules of inference that are used in arguments vie think are deductively valid. How can these rules be justified? Formulate a skeptical problem about deduction that is similar to Hume's problem about induction. Can this problem be solved?
4. Maybe Hume's Principle of the Uniformity of Nature should be formulated as the claim that the lav;s of nature are the same at all times and places. Is this claim true? Does it give definite advice about what we should infer from a given body of observations? ls it an assumption of all induc- tive reasoning?
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchLab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
RESEARCH AND EXPLORE
~)!<{EXPLORE
1. Rules of inference 2. Inductive inference J. "Grue"
chapter 17 ..............................
Beyond Foundationalism
Chapter Outline
Hun1e,s Proble1n and Descartes, Proble111
Whether Xls Evidence for YDepends on Background Assumptions Z
Another Relativity Thesis
Foundationalism Leads to Skepticism
A Nonfoundationalist Approach to Justification
Standards ofJustification Often Depend on the Audience
1\vo Metaphors-Building a Building and Repairing a Raft
In the previous chapter, I briefly considered two attempts to refute Hume's skeptical argu- ment about induction. I argued that neither is successful. However, the failures of these hvo criticis1ns don't sho'v that Hu1ne vvas right, so the question re1nains: Can Hun1e's skep- ticism be refuted?
Hume's Problem and Descartes' Problem
In this chapter, I'm going to argue that if we understand the idea of rational justification in the way Hume did, then he was right that induction can't be rationally justified. Notice that the thesis I'll be arguing for has an IF in it. I won't claim that Hume's skepticism is correct. My conclusion will be more modest: skepticism is correct, IF the task of rational justifica- tion is understood in the 'vay Hume understood it.
To defend this thesis, I want to describe a similarity between a problem that Descartes tried to solve and the problem that Hume addressed. Descartes asked, "How does my pres- ent mental state justify the beliefs I have about the world outside my mind?" Hume asked, "How do the observations I have made of my physical environment justify the beliefs
185
186 Part Three: Theory of Knowledge
''' ''' ''' '' ','' ... '.''' '"'' '', ''. '' ... '' '. ','. ''' ''.'' '.' '''.'' ''.'' '. I have about the future?" The parallelism between these two problems is illustrated in the following diagram:
Hume's problem
Descartes' problem
Ll!Vel
3
2
I
Kind of Belief
Predictions and generalizations Present and past observations
Indubitable beliefs
Examples
"The sun \Vill rise ton1orro\v."
"The sun is no\v rising." "The sun has risen each day that I have n1ade an observation." "I no\\' seen1 to see a sunrise." "I no\v seem to ren1en1ber that the sun has risen each day that I have n1ade an observation." "God exists and is no deceiver."
In this diagram, I've divided beliefs into three categories. Hume was asking how beliefs at level 3 can be justified. Descartes was asking how beliefs in level 2 can be justified.
Both Descartes and Hume were foundationalists in the way they approached questions about knowledge and justification. By this, I mean that each held that IF a belief is ratio- nally justified or known, then it is justified or known solely on the basis of its relationship to beliefs at lower levels. Foundationalism holds that justification flows from bottom to top, not in the other direction. Beliefs are justified because of their relationship to other beliefs that are 1nore certain.
Hume thought that IF beliefs in category 3 were justified, they would have to be justi- fied solely on the basis of items in category 2. Descartes thought thatIFbeliefs in category 2 were justified, they would have to be justified solely on the basis of items in category I. Notice that both of these points are IF statements. They don't assert that the beliefs in question actually are justified; the point is that Descartes and Hume agreed about what would need to be true IF the items in question were justified. This is something on which a skeptic and a nonskeptic might agree.
In spite of this similarity between the way Descartes and Hume formulated their re- spective problems about justification, they ended up defending very different solutions. Ren1e1nber that Hume \Vas a skeptic, vvhereas Descartes \Vas not. Hu1ne claimed that level 3 beliefs cannot be justified, whereas Descartes argued that level 2 beliefs can be justified.
I now want to advance a thesis about Descartes' problem and then extend that thesis so that it applies to Hume's problem as well. Let's grant Descartes that each of us is absolutely certain about his or her own present psychological state. My point is that these level 1 beliefs aren't enough to justify the beliefs I have about the physical environment I inhabit (level 2). I need an additional assumption concerning the relationship between levels i and 2. This connecting principle might take the form of claiming that God exists and is no deceiver, or it might take some other form. Hume thought that it is indubitable that God exists and is no deceiver. I argued before that he \Vas n1istaken in claitning this. But Des- cartes was right that the contents of my own experience are simply not enough, taken all by themselves, to justify the beliefs I have at level 2.
Chapter i7: Beyond Foundationalism
''.' .. ' ' ' ' .. '.''. '.' '.'. '' '''' ' ... ' .. '' '' '.' ..... '. '.' ' ... ' ... '.' ....... '. The parallel thesis about Hume's problem is as follows: if my beliefs about my present
and past environment (items at level 2) are to justify the predictions and generalizations I believe (items at level 3), then I need to assume something about the relationship between levels 2 and 3. The principle of the uniformity of nature ("the future wlll resemble the past") is an example of this type of bridge principle. Or some other bridge principle might be proposed. The point is that my present perceptions and memories are simply not enough, taken all by themselves, to justify the beliefs I have at level 3.
Let's be clear on 'vhat 1'1n suggesting. First) note an obvious point: Level 1 statements don't deductively imply level 2 statements, and level 2 statements dorit deductively imply level 3 statements. My point isrit this obvious fact about deductive relationships. I want to make a more ambitious claim. Not only do lower-level statements fail to deductively imply higher-level statements; I additionally claim that lower-level statements, all by themselves, are not enough to provide a justification for higher-level statements. Lower-level state- ments, taken by themselves, don't even provide good evidence for higher-level statements.
Whether X Is Evidence for Y Depends on Background Assumptions Z
This is a pretty radical thesis, I admit. Do your present experiences provide evidence about the physical environment you occupy? For example, does statement (a) below provide a rational justification for believing statement (b )1
(a) You seem to see a printed page in front of you now.
(b) There is a printed page in front of you now.
I claim that the answer depends on what you assume about the relationship between your present experiences and the 'vorld outside your mind. Given one set of assumptions, state- ment (a) might be excellent evidence that statement (b) is true. But given a different set of assumptions, statement (a) would be good evidence that statement (b) is false. The point is that these assumptions go beyond what is contained in level 1. Assumptions about the relationship of levels 1 and 2 arerit, properly speaking, contained in level 1. I hope my dis- cussion of Descartes makes plausible what I'm saying about the relationship of (a) and (b). If you asstnne that the environment is "normal" and that your senses are functioning prop- erly, then (a) is evidence favoring (b ). If you assume that your senses are misleading you, however, then (a) would be evidence against (b).
What would the parallel claim be about the relationship of statements at levels 2 and 3? For example, does statement (c) provide strong evidence for statement (d)?
(c) I've examined lots of emeralds and all have been green.
( d) All emeralds are green.
We naturally take (c) to be evidence favoring (d). How could this be otherwise? Well, suppose you believed the following statement (the example is due to I.). Good, "The White Shoe Is a Red Herring:' British Journal for the Philosophy of Science, Vol. 17, 1967, p. 322):
Either there are lots of eineralds, of 'vhich 99 percent are green, or there are very fe"' emeralds, and all of them are green.
188 Part Three: Theory of Knowledge
If you believed this, then ( c) would be evidence against ( d). On the other hand, if you believed the following statement, then ( c) would provide evidence in favor of ( d):
If you exa1nine lots of en1eralds and all have been found to be green, then probably all e1neralds are green.
Notice that the two possible assumptions just listed aren't strictly at level 2. Whether (c) is evidence for or against (d) depends on what assumptions you make.
If you make no assumptions about the relationship of levels 2 and 3, then the evidence described in (c) can't be interpreted as either favorable or unfavorable to (d). In a way, the present idea is something that has already come up in this book. I've already pointed out that hypotheses are testable only when background assumptions are added to them (Chapter 9). Here's an elaboration of that idea: When you test a hypothesis Hand obtain some observations 0, it will usually be true that 0 is evidence for or against H only because of the background assumptions (A) you made.
Another Relativity Thesis
We have here another example of a relativity thesis in philosophy. The idea of relativity, recall, came up in Chapter 14 in connection with the problem of knowledge. There it was argued that the Reliability Theory of Knowledge leads to the conclusion that whether S kno\VS some proposition p depends on ho\v one chooses to describe the environn1ent that S inhabits. Different equally true descriptions of S's circumstances lead to opposite conclu- sions about whether S has knowledge. The present point concerns the concept of evidence, not the concept of knowledge. Is statement (c) evidence for statement (d)? No answer can be given until a third ter1n is specified. Relative to one set of assumptions, the ans\ver is 110, but relative to another the ans\ver is yes. We shouldn't ask \vhether one state1nent is evidence for another; rather, \Ve should ask \Vhether one statement is evidence for another relative to a stated set of background assumptions.
Foundationalism Leads to Skepticism
I'm claiming that the evidence relationship involves three things, not just two. Whether 0 is evidence for H depends on background assumptions (A). What impact does this point have on the problems that Hume and Descartes pursued about the concept of rational justification? My suggestion is that the problems they posed lead right to skepticism. If the challenge is to see if level 3 beliefs can be justified solely by level 2 beliefs, we must conclude that this can't be done. Similarly, if the problem is to see if level 2 beliefs can be justified solely by level 1 beliefs, we are forced to reach a negative conclusion in this case as well. Descartes wasn't a skeptic about the rational justifiabilityoflevel 2 beliefs. Hume was a skeptic about the rational justifiability of level 3 beliefs. My claim is that if we adopt a foundationalist understanding of what rational justification involves, then Descartes ,vas wrong and Hume was right. In both Descartes' problem and Hume's problem, foundation- alism leads to skepticism.
Chapter 17: Beyond Foundationalism 189
. '. '' '.''' '''.'' ''.''''. ''.' ... ' .. ' '' .. '. '' '.'. '' .. ''.'.'.' ... '' ... ' ..... . A Nonfoundationalist Approach to Justification
However, I don't think this shows that our beliefs are totally lacking in rational justifica- tion. Rather, I believe that the foundationalist misunderstands what it takes to rationally Justify a belief. We can justify a level 3 belief. We do this by appealing to the other beliefs we have, some of which will themselves be at level 3. The same holds for the beliefs we have at level 2.
Think of how the task of justification works in everyday life and in science. If a scien- tist makes a prediction, or if '"e do this in our everyday lives, ho\v are these predictions to be defended? If we say that the sun will rise tomorrow, we can support this prediction by appealing to other beliefs we have about the way the solar system has worked and will continue to work. In everyday life and science, we allow people to justify their beliefs at level 3 by appealing to other beliefs that are also at level 3. Often, predictions are justified by citing generalizations. In doing this, we aren't following the rules that foundationalism lays down. We aren't justifying a belief at one level strictly in terms of beliefs that are at lower levels.
What does this imply about whether our beliefs about the world are rationally justified? My point is to focus on what it means to ask for a rational justification. If we understand that idea in a foundationalist way, we will be led straight to skepticism-to the conclusion that the beliefs we have aren't rationally justifiable. On the other hand, if we recognize that the idea of rational justification needn't be understood in the way that foundationalism de1nands, skepticism doesn't threaten. In everyday life and in science, we frequently say that our beliefs are justified. When we do this, we aren't using a foun- dationalist understanding of the idea of rational justification. I'm suggesting that we are often perfectly correct when we say that this or that belief is well justified. Such claims are correct, in part, because rational justification is usually not understood in the \Vay the foundationalist demands.
Standards of Justification Often Depend on the Audience
Our standards of rational justification often depend on the audience we have in mind. Suppose you believe some proposition that some other person does not. Your goal is to ra- tionally persuade the other individual that you are right. In doing this, you will feel free to use as a premise in your argument any belief that the two of you share. If you agree about lots of things, there will be many propositions that you can use as premises. If, however, your disagree1nent \\'ith the other person is n1ore pervasive, you 'vill be much n1ore limited in the premises you can use. And if the two of you somehow disagree about everything, it will be impossible for you to construct a rational argument that shows this person that the proposition in question is correct.
In everyday life, it is a familiar occurrence that one person rationally justifies some proposition to another. The satne is true in science and in courts of Ja,v. This can hap- pen because the two parties agree about enough; what they share is sufficient to allow an argument to be constructed that shows whether the proposition under dispute is correct. Matters change, ho\vever, \Vhen 've ask sotneone to justify a proposition to a skeptic.
190 Part Three: Theory of Knowledge
..... ' .. '. ' ....... ' ... ' ' ' ...... '.' .. ' .. '.' .. '.'" ... , '''., ''.'' .... ' ' ' '' ''' A skeptic won't agree with you about many of the things you take for granted. If skeptics differ with you on sufficiently many beliefs, perhaps it will be impossible for you to ratio- nally convince them that the proposition is true.
Foundationalists say that when you try to provide a rational justification of some proposition p, you must construct an argument that would be compelling to a skeptic who doubts all the propositions that are at the same level asp. This is one sort of activity, but there are others. A more familiar proble1n of rational justification is to construct an ar- gument that would convince someone who doesn't already believe p, but who has lots of other beliefs at the same level asp. Even if the foundationalist's problem is insoluble, this doesn't mean that more familiar problems of rational justification are too.
Two Metaphors-Building a Building and Repairing a Raft
In Chapter 13, we examined the foundationalist idea that justifying our beliefs resembles building a building on solid foundations. The foundations must be perfectly solid; other- wise, the entire superstructure will be insecure. This is a good metaphor for foundation- alism. Antifoundationalists think that it is fundamentally misleading. When we take up philosophical problems about rational justification, we already have many beliefs. Our job is not to strip them down to an indubitable core and then build them back up. Rather, what we need to do is to ask if we should change any of the beliefs we already have. Otto Neurath (1882-1945) expressed this view of philosophy in the following analogy: "We are like sailors who must rebuild their ship on the open sea, without being able to dismantle the ship in dry dock or being able to reconstruct it from the best components:' Solid foundations are impossible, but that doesn't mean that our beliefs are completely unjustified.
Review Questions
1. What is meant by saying that we have beliefs at levels i-3? How does a foundationalist under- stand what it means for a belief at a level to be rationally justioed? Are there any beliefs that are located at none of these three levels?
2. A relativity thesis is advanced in this chapter concerning the idea of evidence. What is that thesis? How does it go beyond the "obvious" fact that level 2 beliefs don't deductively imply beliefs at level 3?
). How could the observation of lots of green emeralds be evidence against the claim that all emer- alds are green?
4. Foundationalists understand the problem of rational justification in one way, but people in ordi- nary life usually think of the problem in another. What is the difference?
Problems for Further Thought
In Chapter 7, I argued that scientific explanation is possible only for local why-questions. In the present chapter, I argued that you can rationally justify a proposition to someone only if you agree with that person about something. What similarities are there between these two theses?
Chapter lJ: Beyond Foundationalism
. '' ' .... '.' '.'.' .. '' ' .. '' ... '.' .. ''. '''. '.' '' ''''.' ... ''. '.'.' ....... . MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchLab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
RESEARCH AND EXPLORE
c)li{EXPLORE
1. Foundationalism 2. Skepticism
chapter 18 ..............................
CEJ--[ Read Chapter 11
of Essay Concerning Hu;nan Understanding on www.mysearchlab .com
Locke on the Existence of External Objects
Chapter Outline
Locke's First Argument-"Those That Want the Organs of Any Sense"
Locke's Second Argutnent-((Ideas Which Force Themselves upon Me"
Locke's Third Argu1nent-"Pleasure or Pain"
Locke's Fourth Argun1ent-''Our Senses Assist One Another's Testitnony,,
The last problem I want to discuss in this section of the book on epistemology concerns the beliefs we have about the physical objects in our environment. Consider your pres- ent belief that there is a book in front of you. There are several epistemological questions we could ask about this belief. Does this belief count as knowledge? Are you justified in believing the proposition that there is a book before you? What evidence do you have that this proposition is true? In Chapters 13 and 17, I discussed the first of these questions- the one about knowledge-in connection with Descartes' foundationalism. We discussed Descartes' theistic ans\ver, that God is no deceiver, 'vhich n1eans that if I a1n careful in n1y reasoning, I \Vill not err in trusting the testin1ony of my senses. In the present lecture, I want to take up the third of these questions-the one about evidence. I'll do so by examin- ing some ideas that the English philosopher John Locke (1632-1704) presents in his Essay Concerning Human Understanding.
Philosophers call the book in front of you an "external" object, meaning that it is not in your mind. Your mind can be thought of as "an internal world:' It is populated by be- liefs, desires, and feelings. Books and tables are physical objects in the external world. The experiences you are llO\V having-for example, your present visual and tactile sensations- seem to provide you with evidence that there is a book in front of you. But how can what transpires in the internal world of your 1nind provide evidence concerning \vhat is going on in the external world of physical objects? Locke read and learned from Descartes, but he does not argue for the existence of a God \Vho is no deceiver to ans\ver the question he
Chapter 1&: Locke on the Existence of External Objects 193
... ' .. ' .......... ' ...... ' ... ' ............ ' ...... ' .... ' ........ ' .. ' ...... '' wanted to address. In Book 4, Chapter 11, of the Essay Concerning Human Understanding, Locke tries to sho\v that there are features that 've can discern in our experience that sho'v that those experiences provide evidence for the existence of external objects.
Before we get started on the details of Locke's argument, it is worth considering the kind of argument that Locke is attempting to construct. Suppose you are walking through an art gallery and are looking at several paintings that were all painted by a single artist. You 'vonder 'vhether these paintings are pictures of real-,vorld scenes or are fantasies produced by the artist's imagination. It seems that you can't answer this question just by looking at the paintings themselves. If the paintings are similar to each other, that doesn't seem to help, since an artist can produce similar paintings by looking at a single part of the exter- nal world, or by merely imagining a series of similar scenes. But suppose the paintings all contain unicorns and mer1naids. Isn't this evidence that the paintings are fantasies, not paintings of a real-world scene? Maybe so, but here you are relying on information you have-that there are no unicorns and mer1naids. This information isn't so1nething you learned just by looking at the paintings. If this is the right thing to say about the paintings in the gallery, how can Locke successfully defend his thesis that the content of our experi- ence provides evidence that there are external, physical objects?
Locke points out that "nobody can, in earnest, be so skeptical as to be uncertain of the existence of those things which he sees and feels:' This is a correct psychological point. None of us is capable of seriously doubting that there are physical objects in our environment. One might say that it is part of human nature to think that an external world exists. Howeve1; this psychological point does not address the question of epistemology. Why should you think that the content of your experience provides evidence for the existence of external things? Locke gives the following answer: "This is certain: the confidence that our faculties do not herein deceive us is the greatest assurance '"e are capable of concerning the existence of material be- ings:' This isn't much of an argument. Locke is talking about the following two propositions:
Your sense organs do not deceive you.
There are external objects.
He is saying that you are entitled to be confident that the first is true, and that this assures you that the second proposition is true as well. The problem is that the two propositions aren't sep- arable in the way Locke suggests. Given the experiences you are having now, the first entails the second. Someone who questions the second proposition will also want to question the first; the brief argument of Locke's that I just quoted begs the question (Chapter 2). Locke then provides four inore "concurrent reasons" for thinking that our experience provides evidence for the existence of external objects. Let's see ifhe does a better job of solving this problem.
Locke's First Argument-"Those That Want the Organs of Any Sense"
Locke's first reason has to do with people who are blind or lack some other (functioning) sense organ:
Those that want the organs of any sense, never can have the ideas belonging to that sense produced in their minds. This is too evident to be doubted: and therefore we
194 Part Three: Theory of Knowledge
. ' .... ' ...... ' .. '. '.'' .... ' ... '. '". '. '., .... ' .. ' ... '. ' ... ''. ' .. ' ... ,'.' .. . cannot but be assured that they come in by the organs of that sense, and no other way. 1'he organs themselves, it is plain, do not produce then1: for then the eyes of a 1nan in the dark \vould produce colours, and his nose s1nell roses in the winter: but \Ve see no- body gets the relish of a pineapple, till he goes to the Indies, where it is, and tastes it.
Locke is arguing as follows:
People who are blind do not have visual sensations.
The visual experiences of sighted people are not caused by their having eyes.
The visual experiences of sighted people are caused by external things.
Is the argument valid? If not, maybe a double line should be drawn between premises and conclusion to indicate that the premises are supposed to show that the conclusion is plau- sible or probable, not that it must be true. Also, are the premises true?
Locke's Second Argument-"ldeas Which Force Themselves upon Me"
Locke notes that sensation and memory are experienced differently:
When my eyes are shut, or \Vindo\vs fast, I can at pleasure recall to my mind the ideas of light, or the sun, which fonner sensations had lodged in my n1emory; so I can at pleasure lay by that idea, and take into my view that of the smell of a rose, or taste of sugar. But, if I turn 1ny eyes at noon tovvards the sun, I cannot avoid the ideas \Vhich the light or sun then produces in me. So that there is a inanifest difference bet\veen the ideas laid up in my memory (over which, if they were there only, I should have constantly the same power to dispose of them, and lay them by at pleasure), and those "\vhich force the1nselves upon me, and I cannot avoid having. And therefore it must needs be so1ne exterior cause, and the brisk acting of some objects \Vithout ine, "\Vhose efficacy I cannot resist, that produces those ideas in my mind, \Vhether I ,vill or no.
Locke's argument might be put like this:
I cannot mentally control whether I have the visual experiences I have.
Those visual experiences have causes.
The causes of those visual experiences are external to my nlind.
Again, assess this argument: Do the pre1nises provide a good reason to believe the conclu- sion? Are the pre1nises true?
Locke's Third Argument-"Pleasure or Pain"
Locke's next argument also focuses on a difference between sensation and men1ory-in
this case, the me1nory of a sensation:
Pleasure or pain, "\Vhich acco1npanies actual sensation, accompanies not the returning of those ideas without the external objects. Add to this, that many of those ideas are
Chapter 18: Locke on the Existence of External Objects 195
••• ' •• ,.'. ' ••• ' •• ' •• '.' •••• '.' •• '. ' •••••••• ' ••• 0'.' •• ' ••• ' •••• '. '' •• > •• ''.
produced in us \Vith pain, \vhich after\vards \Ve remen1ber \Vithout the least offence. Thus, the pain of heat or cold, when the idea of it is revived in our minds, gives us no disturbance; \\1hich, \Vhen felt, was very troublesome; and is again, ,vhen actually re- peated: \Vhich is occasioned by the disorder the external object causes in our bodies when applied to them: and we remember the pains of hunger, thirst, or the headache, \vithout any pain at all; \vhich \vould either never disturb us, or else constantly do it, as often as \Ve thought of it, \Vere there nothing 1nore but ideas floating in our niinds, and appearances entertaining our fancies, \Vithout the real existence of things affect- ing us from abroad.
Locke then adds that "the same may be said of pleasure:' Hereis how I see Locke's argument:
When I have a headache, I feel pain.
When I remember the headache, I don't feel pain.
This difference must have a cause.
The difference is that an external object causes the headache, but not the memory of the headache.
Yes, same old drill-evaluate Locke's argument.
Locke's Fourth Argument-"Our Senses Assist One Another's Testimony"
This is my favorite:
Our senses assist one another's testimony of the existence of outward things, and en- able us to predict. Our senses in many cases bear "\Vitness to the truth of each other's report, concerning the existence of sensible things \Vithout us. He that sees a fire, may, if he doubt whether it be anything more than a bare fancy, feel it too; and be convinced, by putting his hand in it.
In Chapter 6, on evolution and creationism, I discussed the Principle of the Common Cause. Locke is using this principle here. When two students submit word-for-word iden- tical essays, it is possible that they wrote their essays independently. However, it is far more plausible to think that they copied their essays from a common source-say, from a Web site on the Internet. Another example discussed in that chapter is the striking resemblance between the words in French, Spanish, and Italian for the numbers. This similarity might be a coincidence, but it is far more plausible to think that the similarities exist because these different languages are descendants of a common ancestor-Latin.
In Locke's example, the visual experience of fire tends to be accompanied by a tactile sensation of \var1nth. Why are these correlated? True, it might be a coincidence. Ho"\vever, it is far more plausible to think that the correlation is due to a common cause-an exter- nal object that causes both the visual impression and the feeling of warmth. The same kind of argument can be developed concerning the different memories you have that are in agreen1ent. Your visual in1pression right no"\v rese1nbles the experience you had a sec- ond ago, and also resembles the experience you had the second before that. You know
196 Part Three: Theory of Knowledge
... '' .. ' ..................... '.' ...... ' .. ' ... ' .. ' ... ' .... ' ' .............. . that this resemblance exists by accessing your memories. The resemblance is plausibly explained by postulating the existence of an external object that persists through time- the book before you-which is the common cause of various experiences and of subse- quent n1emory traces.
These considerations don't prove that there are external objects, but the argument that makes use of the Principle of the Common Cause does seem to establish something more modest-that the content of your experience is evidence for the existence of things that are outside of your experience-an external \Vorld.
Review Questions
1. How is Locke's approach to the question of ho\v we have knowledge of (or evidence for) the existence of physical objects external to the mind different from that of Descartes?
2. Describe Locke's four arguments and evaluate them. Are they deductive, inductive, or abductive?
3. Jn Chapter 6, the Principle of the Common Cause \Vas presented and it \vas said to be a conse· quence of the Surprise Principle, introduced in Chapter 3. Ho\v does the Surprise Principle apply to the fourth of Locke's arguments?
Problem for Further Thought
In discussing Locke's fourth argument, I describe two possible hypotheses. What are they? Are there other possibilities? If so, does considering them affect the conclusion of the argument?
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchLab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
RESEARCH AND EXPLORE
'!!<{EXPLORE
1. John Locke 2. External world skepticism
Suggestions for Further Reading
On Descartes
Bernard Williams. Descartes: The Project of Pure Inquiry. London: Harvester Press, i978. Margaret Wilson. Descartes. London: Routledge and Kegan Paul, 1978.
On the Reliability Theory of Knowledge
David Armstrong. Belief Knowledge, and Truth. London: Cambridge University Press, i973. Fred Dretske. Knowledge and the Flow of Information. Cambridge, MA: MIT Press, i981. Alvin Goldman. "A Causal Theory of Knowing." journal of Philosophy i 967, 64: 357-72. Robert Nozick. Philosophical Explanations. Cambridge, MA: Harvard University Press, i 981.
Chapter 18: Locke on the Existence of External Objects 197
'''' ''','' '.'' .. ' ' '' .... '.',' ... ' .... ''.' .. '.' .. '.' .... ' ................. . OnHtuue
Barry Stroud. Hume. London: Routledge and Kegan Paul, 1977·
On Induction
Nelson Goodman. Fact, Fiction, and Forecast. Indianapolis: Bobbs Merrill, i965. Ian Hacking. An Introduction to Probability and Inductive Logic. London: Cambridge University
Press, 2001.
Wesley Salmon. Foundations of Scientific Inference. Pittsburgh, PA: University of Pittsburgh Press, 1968.
Brian Skyrms. Choice and Chance. Belmont, CA: Wadsworth, 1986. Elliott Sober. Reconstructing the Past~Parsimony, Evolution, and Inference. Cambridge, MA:
MIT Press, 1991, chap. 2.
References
John Locke (1632-1704). "Essay Concerning Human Understanding", (1689).
readings ..... ~ ................. ' ..... .
mJ--[ Read Meditations
i-5 Meditations on First Philosophy on \VW\V.mysearchlab .com
198
Readings
PLATO
The Theaetetus-Knowledge Is Something More than True Belief
In this dialogue, The Theaetetus, Plato (427-347 B.C.E.) describes a conversation bet\veen Socrates, \vho was his teacher, and Theaetetus. Theaetetus discovers through Socrates' questioning, that he, Theaetetus, knovvs much less than he previously thought.
Two main points emerge. First, knowledge is something more than true belief. The suggestion is
made that knowledge requires true belief plus "an account." An account must involve some sort of justification. But \Vhat sort? If the propositions providing the account are not themselves known to be true, then it is hard to see how they can provide knowledge. However, it is circular and unhelpful to say that someone knov1s a proposition p only when the individual believes that p and knows some other propositions that justify p.
RENE DESCARTES
Meditations on First Philosophy
In this vvork, Rene Descartes (1596-1650) tries to discover to \Vhat extent he has genuine knowledge of the world he inhabits. He pursues a foundationalist strategy. First, he attempts to discover which of his beliefs are beyond doubt. These will be the foundations for the rest of what he knows. Second, the task is to sho\V ho\v these foundational items support the rest of his knowledge.
Readings 199
. '''''. '.',' '.'. '' ''''' '''.'. ' .. ' ' ' .. ' .. ''.' ''' .. '' .. ' ' ' ' .. '' ''''.' .. '''' DAVID HUME
Induction Cannot Be Rationally Justified
In this excerpt from his An Enquiry Concerning Human Understanding, Hume (1711-1776) argues that there can be no rational justification of the inductions that people perform. The reason is that all such inductions depend on the Principle of the Uniformity of Nature, and this principle cannot be given a rational justification.
JOHN LOCKE
Essay Concerning Human Understanding
In this passage, Locke argues that we have good reasons for believing that our sensory experiences are caused by, and thus can provide us with knowledge of, external objects. Locke acknowledges that it is possible to remain a skeptic about all sensory experience, like Descartes in the First Medi- tation, but argues that our senses "serve to their purpose well enough." Do you think the skeptic's concerns would be assuaged by Locke's arguments?
r:I!}-[ Read Section IV of An
Enquiry Concerning Human Under- standing on \'NNI .mysearch!ab.com
part four . . . .............. ', .. ·--. '. ~ "' . ~ . . ~ .
Philosophy of Mind
chapter 19 ..............................
Dualism and the Mind/ Body Problem
Chapter Outline
What Is the Mind/Body Problem?
Descartes' Dualis1n
The Mind/Brain Identity Theory
Immortality of the Soul
Leibniz's Lalv
Descartes' First Argu1nent for Dualis1n- The Indubitable Existence Argument
An Analogy
Propositional Attitudes and Aboutness
Descartes' Second Argu1nent for Dualism-The Divisibility Argument
Causality between the Physical and the Nonphysical
In this fourth part of the book, on philosophy of mind, I'll discuss three problems-the mind/body problem (Chapters 19-23), the problem of free will (Chapters 24-26), and the problem of psychological egoism (Chapter 27). The mind/body problem and the problem of free will both address the broad issue of how the mind is related to the physical world of cause and effect. Are minds physical things (for example, are they identical with brains?), or are they nonmaterial? If your beliefs and desires are caused by physical events outside of yourself, how can it be true that you act the way you do of your own free will? The last problem-the problem of psychological egoism-concerns the motives that drive us to act. Are people genuinely moved by the welfare of others, or is all behavior, in reality, selfish?
These three problems concern different stages in the causal chain that leads from genes and environment, to the n1ind, and then to action:
Mind
Genes --> Beliefs Actions
Environ1ncnt --> Desires
(1) (2) (3) 203
204 Part Four: Philosophy of Mind
.. '" .. ' ' ... ''' .. ' '' ''''.'''.' '''' ''. ''''''''
[J!J-{ Read Meditation
Vt from Meditations on First Philosophy on WW\'l.mysearch!ab .com
The mind/body problem concerns the nature of the objects and events that exist at stage 2 in this diagram. What is a mind? What are beliefs and desires? The problem of free will con- cerns the relation of stage I and stage 2. If our beliefu and desires are caused by the genes we possess and the environn1ents 've have inhabited, ho\V can 've possess free \vill? The proble1n of psychological egoism concerns the relationship of stages 2 and 3. If the actions we perform are caused by the desires we have, won't it be true that all action is fundamentally selfish- ain1ed at satisfying the actor's o\vn desires, not at satisfying the needs of others? These are preliminary statements of the three philosophical problems. Each needs to be refined.
What Is the Mind/Body Problem?
The question posed by the mind/body problem is simple: What is the relationship between the mental and the physical? You have a mind, which contains various beliefs, desires, sensations, and en1otions. You also have a brain; this physical thing is a structured piece of tissue containing an intricate \Veb of neurons. Are your 1nind and your brain one and the same thing? Are your beliefs, desires, ernotions, and sensations identical "\Vith physical things found in your brain? Or are your mind and your brain different objects? Perhaps the mind isn't a physical thing at all.
Descartes' Dualism
In the previous section, the discussion of Descartes focused on his episte1nology. Descartes also advanced a solution to the mind/body problem. It is this view, now called Cartesian Dualism, that I'll now discuss. Dualism is the idea that there are two kinds of things in the world. There are physical objects on the one hand, and, on the other, there are mental objects (like minds, pains, and beliefs). According to dualism, brains and the bodies in which they are found are physical things; the mind, which is a nonphysical object, is distinct from both the whole body and is also distinct from all of the body's physical parts. Descartes didn't deny that there are causal interactions behveen the mental and the physical; taking aspirin can cure headaches, and the sound of trumpets can lift your spirits. In the other direction, it seems undeniable that beliefs and desires can cause the parts of your body to move in various ways (speaking, walk- ing, etc.). But granting this two-way interaction did not lead Descartes to abandon dualism.
The Mind/Brain Identity Theory
One alternative to dualism is provided by the Mind/Brain Identity Theory. This theory makes a claim about objects and also about the properties those objects possess. First, it says that your mind and your brain are one and the same object. Second, it claims that the mental properties you have (for example, believing that fire is hot, or wanting some ice crea1n, or being in pain) are physical properties; to be in pain is to have a certain type of physical event occur in your central nervous system.
The identity theory asserts that mental terminology and physical terminology describe the same items in the world. The following analogy may be instructive. For a long time, people used the tern1 "'vater" to denote various items in the 'vorld. This ter1n 'vas used in everyday life. At a certain point in the history of science, it \Vas discovered that 'vater
Chapter 19: Dualistn and the Mind/Body Problen1 205
''',' ''' '' ''' ''. '.'' ''''.''.'.''.' '' '''''"'' ''.','',' '' '.' ... ''.'. '''. is nlade of HzO 1nolecules. The discovery 'vas that 'vater and HzO are one and the same thing. There certainly is a difference bet,veen the hvo pieces of ter1ninology '\vater" and "H,O:' The first word has been used by ordinary people for a very long time. The second term 'vas introduced 1nuch 1nore recently, as part of a scientific theory. Yet it does not fol- low from these facts about the two terms that the terms denote different things. Water is identical 'vith HiO, as che1nists have discovered.
Philosophers who defend the Mind/Brain Identity Theory say that the same point ap- plies to the relationship of nlental ter1ns and neurophysiological terms. Co1nmon sense has for a very long tilne deployed such terms as "belief;' "desire:' "pain:' and "mind:' The identity theory suggests that what happened to water will happen to the mind. Eventually neurophysiology will discover the nature of the mind, just as chemistry discovered the nature of \Vater. Once various scientific theories have been developed, \Ve \Vill understand \Vhat it is to believe that sno\v is \Vhite, \Vhat it is to \Vant so1ne coffee, 'vhat it is to feel pain, what it is to have a mind. In each case, the answer will be given in the vocabulary provided by brain science. We now have only a very partial picture of what these neurophysiologi- cal theories will be like. However, the identity theory predicts that science is headed in the direction of a purely materialistic account of the mind. The mind is a physical thing, even though we now have only an incomplete picture of what its physical nature is.
In addition to dualism and the Mind/Brain Identity Theory, there are other solutions to the mind/body problem that I'll discuss. These are logical behaviorism (Chapter 20) and functionalism (Chapter 23). Rather than describing them right now, I'll turn to the task of analyzing Cartesian Dualism.
Immortality of the Soul
Before I present Descartes' argun1ents for dualism, I should note a connection bet\veen the mind/body problem and an issue in the philosophy of religion. If you believe the doctrine of the immortality of the soul but also hold that the body disintegrates at death, you may be attracted by dualism. The issue of whether the soul is immortal, of course, isn't the same as the question of whether there is a God. After all, there are many religions that deny the immortality of the soul. Conversely, someone might hold that the soul survives the death of the body and yet deny that there is a God. But historically it is worth remembering that what we might call "traditional" Christianity (this includes the Christianity of Descartes' time) espouses the doctrine of immortality. Dualism makes room for this possibility.
If the soul is part of the mind (even if it istit the entirety of the mind) and if the soul lives forever and the body does not, \Ve have an argu1nent for dualisn1. Ho\veve1; this isn't Descartes' argument for dualism. Why not? Perhaps the reason is that anyone who doubts dualism will also probably doubt that any part of the mind survives the death of the body. So you aren't going to convince anyone that dualism is true by beginning with the premise that the soul is immortal.
Leibniz's Law
Note a structural feature of the argument for dualism that I just described. The argument defends dualism by trying to find a property that the mind has but the brain lacks; the property in question is immortality. Quite apart fro1n 'vhether this argutnent is successful,
206 Part Four: Philosophy of Mind
'''''''' ''' '''''''' • '' ''' >'.' '.' ''' ''.'''' '''''' '''' ',''''' ''''.,' ''''''
Cl!J-{ Read Meditation VI
from Meditations on First Philosophy on \'NN1.mysearchlab .com
we should note a perfectly sensible principle that it uses. The idea is that if 111 and b are identical, then they must have all the same properties. This principle is called Leibniz's Law, after the seventeenth-century philosopher/mathematician (who, incidentally, co- invented the calculus with Newton. In Chapter 7, I briefly discussed Leibniz's belief that we live in the best of all possible worlds). Leibniz's Law is sometimes called the principle of the indiscernibility of identicals. It states that if you can find even one property that m has and blacks, then you will have shown that 111 and bare distinct entities. You'll see this principle at work in both of the arguments that Descartes presents for dualism. The following form of argument is deductively valid:
111 has property P.
b does not have property P.
If 111 has a property that blacks, then m 7' b.
Ill " b.
Be sure you see how the argument concerning immortality has this logical form. The argument is valid. This means that if you reject the conclusion, you must reject at least one of the premises. I suggest that the third premise (Leibniz's Law) is true. Therefore, if you reject the conclusion, you 1nust reject one of the first t\\'O premises.
Descartes' First Argument for Dualism-The Indubitable Existence Argument
Now let's look at Descartes' first argument for dualism. In Chapter 13, I discussed how Descartes uses the method of doubt in his epistemology. He also uses the idea of doubt in his discussion of the mind/body problem.
In the Second Meditation, Descartes claims you can't doubt that you have a mind. If you try to doubt that you have a mind, you will find yourself entertaining a thought, and so you must grant that you have a mind after all. Descartes thought that the existence of the body has a quite different status. He thought that it is possible for you to doubt that you have a body. After all, you can entertain the thought that you are a disembodied spirit. Descartes concludes that your mind has a property that your body lacks. You can doubt the existence of the one, but not the existence of the other. Dualism follows, by Leibniz's Law.
Perhaps you are suspicious of what Descartes says about your body. Is it really possible for you to doubt that you have a body? Can you conceive of yourself being a disembodied spirit? You should also consider Descartes' other premise. ls Descartes right that you can't doubt that you have a mind?
I'm not going to pursue these questions here. I'll grant Descartes that he can't doubt that he has a mind and that he can doubt that he has a body. I want to consider whether dualism validly follows from these premises. To see whether this follows, we must be very explicit about what the property is that the mind is said to have and the body is said to lack. Descartes claims that his mind has the property of indubitable existence, and that his body lacks that property. Let's look more closely at this property. For an object X to have this property n1eans that the "o,vnee, of X can't doubt that X exists. I say "ovn1er" since n1y
Chapter 19: Dualism and the Mind/Body Problen1
.. ',' .. ' .. ' '.' '' ' ... ''' ' .... ' ' .. ' ' ' .. '' .. '.'.' .. ''.' .. ' '. '''''','''''''' mind does not have indubitable existence for you. I take it that you have no trouble enter- taining the thought that I don't have a mind. It is the first-person case that matters here-a person can't doubt the existence of his or her own mind.
An Analogy
There is a subtle mistake in Descartes' argument. I'm going to argue that indubitable existence is not a genuine property at all.
I'll illustrate this idea by an example. Lois Lane wants to marry Superman. She doesn't re- alize that Superman and Clark Kent are one and the same person. Clark Kent, you'll recall, is the most incompetent reporter at the Daily Planet. If you ask Lois whether she wants to marry Clark Kent, she will say "No!" Does it follow from this (via Leibniz's Law) that Superman and Clark Kent are two different people? Of course not. The following argument is not valid:
Lois Lane \Vants to tnarry Supern1an.
Lois Lane does not want to marry Clark Kent.
Superman is not identical with Clark Kent.
On the surface, it looks as if this argument is valid, given Leibniz's Law. It describes a property that Superman has and Clark Kent lacks and concludes that they are nonidenti- cal. What you must see is that the argument does not describe any such property. Leibniz's Law, properly understood, does not license the conclusion of nonidentity. The two prem- ises are true, but the conclusion does not follow from them; it is false.
Superman and Clark Kent have exactly the same properties. This isn't contradicted by the fact that Lois Lane wants one of the following propositions to be true and the other to be false:
Lois Lane marries Super1nan.
Lois Lane marries Clark Kent.
The fact that Lois desires one of these propositions to be true and the other to be false does not show that Superman and Clark Kent have different properties. The propositions are different; the people are identical.
Likewise, Descartes says that I am able to doubt one, but not the other, of the following two propositions:
I have a brain.
I have a n1ind.
But from this, it does not follow that my brain has a different property from my mind. The prop- ositions are different, but that doesn't show that your brain and your mind are different objects.
Propositional Attitudes and Aboutness
The crucial distinction we have to note here is this: doubting and desiring are attitudes we have towards propositions; doubting, desiring, and believing are examples of propositional attitudes. Perhaps there are so1ne propositions \Vhose truth cannot be doubted) whereas
208 Part Four: Philosophy of Mind
there are other propositions that we are able to doubt. Surely we desire that some proposi- tions, but not others, should be true. Bnt separate from this issue concerning propositions is the issue of what objects those propositions are about. Here is the lesson I draw from the example about Lois Lane: Even if one proposition is desired whereas another is not, it does not follow that what the first proposition is about differs from what the second proposition is about. The same point holds when we consider other propositional attitudes, like doubt- ing. What follows from these two statements?
Lois Lane 'vants it to be true that Lois Lane 1narries Super1nan.
Lois Lane does not \Vant it to be true that Lois Lane marries Clark Kent.
What follows is that the proposition Lois Lane marries Superman is a different proposition from the proposition Lois Lane marries Clark Kent. What does not follow is that the person the first proposition is about (Superman) differs from the person the second proposition is about (Clark Kent). (Note to the reader: if you think that Superman and Clark Kentare differ- ent individuals, please invent an example in \Vhich a single person has hvo names and sho\V how this example illustrates the relevant point about propositional attitudes and aboutness.)
What follows from these two statements?
I can't doubt that I have a mind.
I can doubt that I have a body.
What follows is that the proposition I have a mind is a different proposition from the proposition I have a body. What does not follow is that the object the first proposition is about (my mind) differs from the object the second proposition is about (my body).
With this diagnosis in hand, let's go back to Descartes' argument for dualism. He says that his mind has the property of indubitable existence, whereas his body does not have that property. it sounds as if Descartes is describing a property that the one thing has but the other thing lacks. But this, I claim, is deceptive. Indubitable existence is not a property of an object; rathet; doubting is something we do or fail to do to propositions. To say that my mind indu- bitably exists is just to say that I cannot doubt a particular proposition. To say that my body does not indubitably exist is just to say that I can doubt a particular proposition. From this difference between propositions, however, nothing follows concerning whether those propo- sitions are about the same or different things. Descartes' first argument for dualism is invalid.
SENSE AND REFERENCE
The philosopher/logician Gottlob Frege (1848-1925) wanted to explain a difference that separates the following two statements:
The Evening Star is the Evening Star.
The Evening Star is the Morning Star.
The first of these is obviously true; it is a logical truth of the form "a~ a." It is a priori; anyone who understands the meanings of the terms in this sentence will be able to see that it is true, there being no need for astronomical observation. The second state- ment is different. It describes a discovery that astronomers made; it is a posteriori.
Chapter 19: Dualism and the Mind/Body Problem 209
.......................................... ' ............. ' ... ' ............ . The phrase "the Evening Star" is used to refer to the first star to appear in the evening. "The Morning Star" is an expression that is used to refer to the last star to disappear in the morning. It was discovered that these are one and the same object-namely, the planet Venus.
Notice that the terms occurring in the two statements refer to exactly the same thing. How, then, can the statements express different thoughts? Frege said that the explanation of this fact is that terms have sense (meaning) as well as refer- ence. Although "the Morning Star" and "the Evening Star" are terms that refer to the same thing (the planet Venus), the two expressions aren't synonymous; they have different meanings. Frege thought that synonymous terms must refer to the same thing, but that terms that refer to the same thing need not be synonymous. The meaning of a term determines its reference, but not conversely. He also be- lieved that the "truth value" of a statement (that is; whether the statement is true or false) is settled just by the reference of the terms it contains.
Consider the statement "The inventor of bifocals is dead." This statement is true. If Frege were right to say that the truth value of a statement is determined by the ref- erence of the terms it contains, then we should be able to remove a term from this statement, replace it with a coreferring term, and have the resulting statement still be true. We can do this: "the inventor of bifocals" and "the first U.S. ambassador to France" are coreferential. Substituting one for the other, we obtain the following sen- tence: "The first U.S. ambassador to France is dead." This statement is true. In this case1 a true statement remains true if one of its terms is replaced by another that is coreferring.
Statements describing propositional attitudes posed a problem for Frege. Even though "Superman" and "Clark Kent" refer to the same person, the first, but not the second, of the following statements is true:
Lois Lane wanted to marry Superman.
Lois Lane wanted to marry Clark Kent.
To account for this fact, Frege suggested that terms in such sentences don't refer to the objects that they normally refer to. Normally, "Superman" and "Clark Kent" refer to a person. But in the above pair of sentences, Frege claimed that a shift occurs; the first refers to the meaning of the term "Superman," whereas the second refers to the meaning of the term "Clark Kent." Since these terms have different mean- ings, it won't be true that the terms in the first sentence refer to precisely the same things that the terms in the second sentence refer to. This makes it possible for the first sentence to be true and the second false. In this way, Frege was able to retain his principle that the truth value of a sentence is determined by the reference of its constituent terms.
When you say, "Lois Lane shook hands with Superman," you are referring to Superman. When you say "Lois Lane wanted to marry Superman," you aren't refer- ring to Superman, according to this proposal of Frege's. If you believe that you are referring to Superman in both statements, then you will reject Frege's account of how statements about propositional attitudes should be understood.
210 Part Four: Philosophy of Mind
Descartes' Second Argument for Dualism- The Divisibility Argument
I no\V turn to Descartes' second argun1cnt for dualism. It is far sitnpler than the one just analyzed. In the Sixth Meditation, Descartes claims that physical things have spatial parts. For example, a surgeon could divide iny brain into pieces. My 1nind, ho\\Tever, does not have spatial parts. If so, dualism follows, by Leibniz's Law. Descartes also says that the body, but not the mind, has extension. By this he means that the body, but not the mind, takes up physical space; it has spatial location. This also leads to dualism, by Leibniz's Law.
I'll treat these two arguments together: if the body has the properties of divisibility and extension, but the n1ind does not, dualistn follo,vs. I think that Descartes' argu1nent here is valid. The question is whether his premises are true. I'll grant Descartes that it sounds odd to say that my mind has spatial parts and that it is located between my ears. It also sounds strange to say that my mind weighs about five pounds and has blood vessels running through it. Ho\v are \Ve to explain the fact that such clain1s sound funny to us? One explanation is that they can't be true. If this were right, dualism would follow, since we then would have cited properties that rny brain has but 1ny inind does not. Ho\vever, there is another possible ex- planation for why it sounds odd to say that my mind has spatial parts, or weighs five pounds, or has blood vessels running through it. The explanation is that these ideas are unfamiliar. The assertions sound jarring because they radically depart from what we happen to believe.
Consider the claim that water is H, 0. Before the advent of the atomic theory, the claim that a liquid is made of numerous tiny particles that are too small to see may have sounded pretty strange to so1ne people. But this, of course, does not n1ean that 'vater could not be inade of molecules of HiO. For this reason, I clairn that Descartes' second argun1ent for dualis111 is in- conclus.ive. If the mind and the brain are really identical, then many surprising facts may follow.
I conclude that Descartes' arguments for dualism do not work. The first argument (in- volving the idea of indubitable existence) is invalid. The second argument (involving the ideas of divisibility and extension) is valid, but it begs the question. There seems to be no reason to accept the premises (that the mind is indivisible and lacks extension) unless you already believe that the conclusion (dualism) is true. These negative verdicts don't show that dualism is false. All I've claimed so far is that these two arguments for dualism are unsuccessful. I no\Y turn to a criticis111 that has been inade of dualis1n.
Causality between the Physical and the Nonphysical
One of the main stumbling blocks for dualism has been the idea, endorsed by Cartesian Dualism, that there can be causal interactions between physical and nonphysical things. Descartes thought that physical events in your body can cause sensations in your inind. These sensations, like all mental events, allegedly lack spatial location. But how can events that are located in space bring about events that lack spatial location? And how are causal relations in the opposite direction possible? In 1f1e Passions of the Soul, Descartes claimed that the pineal gland in the brain is the jumping-off point for this interaction. Nerve im- pulses reach the pineal gland and then manage to affect the mind, even though the mind is not located anywhere at all. Conversely, your mind (which is no place at all) influences your body by making an impact on the pineal gland. This is pretty mysterious, and dualists since Descartes have not rnanaged to make this process any inore co1nprehensible.
Chapter 19: Dualis1n and the Mind/Body Probletn
Causality is sornething vve understand best \vhen \Ve consider t\\'O physical events that are linked by a physical signal. When we say that throwing the switch on the wall caused the ceiling light to go on, we are talking about two physical events that occur at different times and at different places. These two events are connected by the flow of electricity. We not only know that throwing the switch caused the light to go on; we also know how throwing the switch managed to bring this about. If we were unable to detect a physical signal passing from the switch to the light, we would be puzzled about how the first event was able to cause the second. Similarly, if I said that throwing the switch caused an event that isn't located anywhere at all, you would be puzzled how electricity or any other physi- cal signal could reach an event that has no spatial location.
In light of the difficulty of understanding how causality can "cross over" from the mental to the physical and back again, wouldn't it be simpler to account for the causal interaction of the mind and the body by adopting the identity theory? If the mind and the brain are identi- cal, it isn't terribly puzzling ho\v your beliefs and desires can cause you to behave in various ways. This doestrt prove that the Mind/Brain Identity Theory is correct; the point is just that what is a hard question for dualism to answer isn't especially hard for the identity theory.
Review Questions
1. What does dualism assert? What is the Mind/Brain Identity Theory?
2. \Vhat is Leibniz's Lavi? Ho\'J is it used in arguments supporting dualism?
3. "I can't doubt that I have a mind, but I can doubt that I have a body. Hence, my mind isn't identi- cal \Vith my body." Is this argument valid?
4. "My brain is divisible into spatial parts, is located bet\veen my ears, v1eighs about five pounds, and has blood vessels running through it. My mind has none of these properties. Hence, my brain and my mind are nonidentical." Is this argument valid? lfit is valid, must dualism be true?
5. Dualism has been thought to make mysterious hov1 the mind and the body can causally interact. What problem is involved here?
Problems for Further Thought
1. Descartes says that he can conceive ofhimse!fbeing a disembodied spirit (that is, having a mind but not a body). What does conceiving of something mean? Does Descartes' claim entail that it is possible for him to be a disembodied spirit? {See discussion of conceivability and possibility in Chapter 8.)
2. Is a statue identical \vith the stone it is made of? Is an organism identical \Vith the collection of cells in its body? Can Leibniz's Law be used to show that either of these claims of identity is false?
3. In the Sixth Meditation, Descartes argues that he is essentially a thinking thing. An essential property of a thing is a property that the thing must have if it is to exist. Could Descartes be de- prived of thought and still be Descartes? Could Descartes have been born without the capacity of thought and still be Descartes? If Descartes can't doubt that he thinks, is that enough to sho\v that Descartes is essentially a thinking thing?
4. It was suggested in this chapter that we understand causality best v1hen there is a physical signal that passes from cause to effect (the electricity example). Ho\vever, the fact that "absences" are sometimes causes suggests that causality need not involve a physical signal. For example, sup- pose a patient dies because his doctor fails to give him medicine. There is no "physical signal"
211
212
CEJ-{ Read "Corres-
pondence with Princess Elizabeth (selectionsr on W\V\V .mysearch!ab.com
Part Four: Philosophy of Mind
between doctor and patient in this case, but there is causation. Does this point solve the objec-
tion to dualism that concerns the nature of causality?
5. Frank Jackson (in "What Mary Didn't Know," journal of Philosophy 83 [1986]: 291-95] gives the fol- lowing argument against physicalism (which he takes to be the claim that all facts are physical facts):
Mary is confined to a black-and-white room, is educated through black-and-white books and through lectures relayed on black-and-v,ihite television. In this way she learns everything there is to know about the physical nature of the world. She knows all the physical facts about us and our environment, in a \Vide sense of "physical" which includes everything in completed physics, chemistry, and neurophysiology, and all there is to know about the causal and relational facts consequent upon all this, including of course functional roles. lf physicalism is true, she kno\'JS all there is to know. For to suppose otherwise is to suppose that there is more to know than every physical fact, and that is \vhat physicalism denies.
Physical ism is not the noncontroversial thesis that the actual world is largely physical, but the challenging thesis that it is entirely physical. This is why physicalists must hold that complete physical knowledge is complete knowledge simpliciter .... The claim here is not that, if physicalism is true, only what is expressed in explicitly physical language is an item of kno\vledge. It is that, if physicalism is true, then if you know everything expressed or expressible in explicitly physical language, you know everything.
It seems, hov1ever, that Mary does not know all there is to know. For when she is let out of the black-and-white room or given a color television, she will learn \Vhat ·1t is like to see something red, say. This is rightly described as learning-she \Viii not say "ho, hum." Hence, physicalism is false.
Assess Jackson's argument using the ideas concerning aboutness and propositional attitudes presented in this chapter.
Supplementary Reading
RENE DESCARTES
Correspondence with Princess Elizabeth
One of the primary challenges to any dualist theory is accounting for the interaction between material bodies and immaterial souls. For example, how can the decision to pick up a pencil cause your arm to move? In her correspondence with Descartes, Princess Elizabeth of Bohemia raises this worry as an objection to Descartes's dualism.
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
Chapter 19: Dualism and the Mind/Body Problen1 213
.. ''.' ''' ..... ' .. '' .. ' .. ' ... '.'. '''' '' .... '' ''.' .. ' .. '' ''.' .. ' '.' .. ' ''.'' CEJ-{ READ AND REVIEW
READ
1. Rene Descartes, Meditation VI from Meditations on First Philosophy · How do Descartes's arguments in Meditation G relate to the Cogito argument that you learned
about in Chapter 13? Do any of his claims here depend on that earlier argument? 2. Rene Descartes, "Correspondence with Princess Elizabeth {selections)"
What objection does Princess Elizabeth raise against Descartes's dualist theory? Does Descartes adequately address her criticism? If not, can you think of a better response that could be given in defense of Cartesian dualism?
WATCH AND LISTEN RESEARCH AND EXPLORE
©{WATCH (<•>-[LISTEN ~1,~EXPLORE
l. Listen "The Mind/Body Problem"-BBC: 1. Research mind-body problem .In Our Time 2. Research dualism
Philosophers Anthony Gray/in& Julian 3- Research Leibniz's Law Baggini, and Sue James discuss the history of views regarding the mind-body problem, from Plato until today. To what extent do you think that scientific developments have influenced our views about the mind-body problem? What view of the relationship be- tween the mind and body do you find most
plausible?
chapter 20 ..............................
214
Logical Behaviorism
Chapter Outline
The Attack on "the Ghost in the Machine''
Logical Behaviorism Says Mentalism Is False Because It Leads to Skepticism
Do We Know about the Mental States of Others by Analogy with Our Own Case?
Abduction
Logical Behaviorism's Positive Thesis- Its Analysis of Mentalistic Vocabulary
The Dispositional Analysis of Desire Is Incomplete
A Dispositional Analysis Does Not Refute Mentalism
In 1949, Gilbert Ryle, a philosopher at Oxford University, published an influential book called The Concept of Mind. In it he presented a solution to the mind/body problem that has come to be called logical behaviorism. His views have something in common with those expressed by Ludwig Wittgenstein in his posthumously published Philosophical Investigations (1953). Later, an American philosopher deeply influenced by Wittgenstein, Norman Malcolm, elaborated this philosophical position in his book Dreaming (1959).
Logical behaviorism is a thesis about the meaning of the mentalistic terms we use in ordinary speech. Logical behaviorism tries to describe what we mean when we talk of an individual's thoughts, beliefs, intentions, dreams, and sensations. This thesis about mean- ing is quite different from a doctrine called methodological behaviorism, which I'll dis- cuss in the next chapter. Methodological behaviorism is a thesis about science. It doesn't offer a linguistic analysis of ordinary speech; rather, it gives advice about how a productive science of psychology should be developed. Logical behaviorism advances both negative and positive claims. The negative part criticizes other vie\vs of the inind; the positive part offers an account of 'vhat 1nentalistic terminology n1eans.
Chapter 20: logical Behaviorisn1
The Attack on "the Ghost in the Machine"
Ryle thought that the commonsense view of the mind is deeply confused. According to Ryle, con1mon sense is co1n1nitted to the idea that mental states are inner causes of be- havior. To see what he means, suppose we ask why Joe lifted the cup to his lips. A com- monsense answer might be that he did this because he wanted a drink of water and believed that there \Vas \Vater in the cup. This co1nn1on sense explanation says that tnental states are inner states that Joe occupies, ones \Vhich cause his onhvard behavior. According to coin- monsense, we see the behavior (the drinking), but we don't see the beliefs and desires that cause it. The beliefs and desires are "inside''; they emit be directly observed by third parties, though their effects-behavior-can be. I'll use the terrn "1nentalisn1" to na1ne this com- monsense idea that mental states are inner causes of behavior. Ryle thought that this natural picture is deeply confused-mentalism is one-half of what he called the myth of"the Ghost in the Machine:'
Ryle also held that common sense embraces dualism. This is the other half of what Ryle calls the myth of the Ghost in the Machine. Although Ryle lumped mentalism and dualism together, we need to keep them separate. Mentalism and dualism are different. Someone who thinks that the mind and the brain are identical-that mental states are states of the central nervous systen1-\vould reject dualism but \vould agree that 1nental states are inner causes of behavior. So nlentalism does not itnply dualism.
Why did Ryle reject mentalism-that minds exist and are quite different from the behaviors they cause? I'll discuss one of his main reasons. Ryle thought that the view of minds as inner causes leads to what I'll call third-person skepticism.
Logical Behaviorism Says Mentalism Is False Because It Leads to Skepticism
Ryle's argument goes as follows. If mental states were inner, then the mental states of others would be hidden from us. Each of us would be able to observe the behaviors of others, but not the beliefs and desires that others have. Because of this, we wouldrit be able to know what others think or want. At best, the only facts we would know about the mind would come to us via first-person introspection. I can tell by examining my own mind what I think, want, and feel. Ho\vever, if tnental states \Vere inner, I \vould have no \vay of kt10\ving anything about what mental states you occupy. Ryle thought we clearly do have knowledge of the mental characteristics of others. He concluded that mental states can't be inner causes of behavior.
The deductively valid argument that Ryle advances goes like this:
(1) If mental states were inner causes of behavior, we would not have knowledge of the mental states of others.
(2) We do have knowledge of the mental states of others.
Hence, mental states are not inner causes of behavior.
Since the argun1ent is valid, if you \Vant to reject the conclusion, you n1ust reject one or both of the premises. My view is that premise (2) is true. I also think that the conclusion is
215
216 Part Four: Philosophy of Mind
'.'.' '.''' '.'.' >''' '' ''' '''' ',.' •• ''.''''.'.' ''''.' '' ''' ''. '' '''' ''' ''' '<
false. This means that I must reject premise (1). To do this, I must explain how we can have knowledge of the inner mental states of others.
Do We Know about the Mental States of Others by Analogy with Our Own Case?
A traditional explanation of how we are able to have knowledge of the mental states of others is via an argument by analogy. In 1948, Bertrand Russell advanced this argument in his book Human Knowledge: Its Scope and Limits, but the argument has been put forward by many philosophers. It is offered as a solution to the so-called "problem of other minds:' I know by introspection that I have a mind. I also see by observing my own body that I behave in certain \Vays. I note that some of my behaviors tend to be associated \vith son1e of my mental states. For example, when I hurt my finger by stabbing it with a pin, I tend to say "Ouch!" I then look at others and notice that they produce certain behaviors. They say "Ouch!" on some occasions. I reason, by analogy, that others probably have minds and occupy the mental states that I do when I produce the same behaviors.
So the analogy argument goes like this:
In my own case, I notice that when I produce behavior B, I usually am in mental state M.
I observe that another individual ( 0) is now producing behavior B.
[ l So, 0 is nu\v in Jnental state M.
The double line indicates that the argument is1it supposed to be deductively valid. Rather, the analogy argument says that the conclusion is probable if the premises are true. The analogy argu1nent ain1s to sho\v ho\v \Ve can have kno\vledge of the inner mental states of others. If the argument works, it refutes Ryle's claim that mentalism leads to third-person skepticistn. But is the analogy argument persuasive?
The usual criticism of this argument is that it is very weak because the evidence is lim- ited to my own case. It is like arguing that since I own a green chair, probably eve1yo11e has one. I agree \Vith this criticis1n: An induction from one,s O\Vll case to a co11clusion about all htunan beings bases too a1nbitious a conclusion on too slender an evidential basis. 1'he sample size is too small. Although the analogy argument does not refute premise (1) in Ryle's argument, I still think this premise of Ryle's is mistaken. We can see why by thinking about abduction, not induction.
Abduction
Let's view the beliefs and desires we attribute to others as theoretical postulates. We don't directly observe what other individuals think and want; we observe their behavior. We then invent a "theory" whose adequacy is judged by its ability to explain and predict behavior. My suggestion is that there are many cases in which the ascription of mental states to others is well confirmed by the behaviors we observe. (Of course, there also are cases in \Vhich our claims about \Vhat people think and \Vant are under1nined by \vhat \Ve subsequently observe them do, but that isn't in dispute.)
Chapter 20: Logical Behavioristn
'' '' '',' '''' ''''' ''''''''''' '.'''' ''' ...... '.'' ..... ' What's the difference between this solution to the problem of other minds and the
analogy argument? In the analogy argument, I begin with an observation of myself and then seek to extend that description to other individuals. In the abductive argument, I inake no 1nention of introspection. Mendel never observed a gene, but that is no objec- tion to his theory. The fact that genes are inner causes of how tall a pea plant grows, or of 'vhether its peas \vill be 'vrinkled or sinooth, does not n1ean that 've can't kno'v about genes (see Chapter 3). Like,vise, the vie'iv that n1ental states are inner causes of behavior does not lead to third-person skepticisn1. Ryle's criticism of inentalistn is 1nistaken.
There is a simple but important distinction that needs to be drawn here. The following t\VO questions are different; they need not have the sa1ne answers:
(1) Is proposition P about observable things and only observable things?
(2) Is proposition P testable by observation?
The Mendelian case and the case of claims about the mental states of others provide examples in which the answers are (1) no and (2) yes.
This discussion of the problem of other minds should remind you of material dis- cussed in Chapter 5 concerning the role of analogy and induction in the Argument from Design for the existence of God. Hume criticized the Argument from Design for being a weak analogy argu1nent and for being a 'veak induction. Ho'iv is iny reply to Hutne sin1ilar to the reply I just made to Ryle?
Logical Behaviorism's Positive Thesis-Its Analysis of Mentalistic Vocabulary
In addition to criticizing n1entalis1n, Ryle also advances a positive thesis about ho,v state- ments about beliefs and desires should be understood. Logical behaviorists maintain not just that belief is not an inner state; they also make a positive claim about ,vhat it is for an agent to believe something.
Logical behaviorism says that the meanings of mentalistic terms can be specified purely in ter1ns of behavior. For exan1ple, 'vhat does it inean to say that someone 'ivants a drink of water? The following proposal, though false, at least is consistent with the require- ments that logical behaviorism imposes on the problem:
S wants to drink water~,, S drinks.
The subscript "df" means that the two statements are said to be equivalent by defini- tion. This proposal is false; people who want a drink of water don't always get to drink. Notice, however, that the proposal obeys the rules that logical behaviorism lays down. The proposal analyzes the meaning of a mentalistic term ("wants to drink water") in purely behavioral language ("drinks"). Behaviorists usually admit that proposals like this one are too crude. They suggest, instead, that the nleaning of nlentalistic ter1ninol- ogy cannot be given in ter1ns of actual behavior, but can be given in ter1ns of disposi- tions to behave. A 1nore satisfactory suggestion, logical behaviorists maintain, is the following:
S wants to drink son1e 'vater :::::df S is disposed to drink 'vater.
218 Part Four: Philosophy of Mind
I have two criticisms of this proposal. First, it is incomplete. When the proposal is fleshed out, it turns out not to be consistent \Vith the require1nents of logical behaviorism. Second, I'll claim that even if the proposal were true, it would not establish what logical behavior- ists hold-namely, that mentalistic terminology does not describe inner causes of outward behavior.
The Dispositional Analysis of Desire Is Incomplete
I begin with the charge of incompleteness. Wanting to drink water is supposed to be a disposition to drink water. Suppose Joe wants to drink some water and we place a cup of water before him, but he does not drink. Why not? The answer might be that he does not believe the cup contains water. However, this means that the behaviorist's proposed defini- tion must be corrected so that it reads as follows:
S wants to drink water = df Sis disposed to drink those things that S believes are water
Notice that the proposal no'iv is not consistent 'iVith behaviorist requirements. 1'he present proposal analyzes a desire in terms of a disposition to behave and a belief. So we have not analyzed the mentalistic concept in purely behavioral terms.
I doubt that behaviorists can overcon1e this problen1. Our commonsense n1entalistic concepts seem to have the following property: attributing a single mentalistic property to an agent does not, by itself, have implications concerning how the agent will behave. What has such implications are batches of mentalistic properties. If an agent wants, above all else, to drink water and believes that the cup before him contains water, perhaps this implies that he will reach for the cup and drink its contents (but what if he is paralyzed?). How- ever, the desire by itself has no such implication and neither does the belief.
A Dispositional Analysis Does Not Refute Mentalism
My second objection to logical behaviorism's proposal for how mentalistic terms are to be analyzed is this: even if mentalistic concepts could be analyzed purely in terms of behavioral dispositions, that would not disprove the mentalistic thesis that mental states are inner causes of ouhvard behavior. To see 'ivhy, I 'ivant to describe an analogy that David Armstrong (whose views on the Reliability Theory of Knowledge were discussed in Chapter 14) noticed between the supposed dispositional property of wanting a drink and other nonmental properties that clearly are dispositional in character. Dispositional prop- erties are often na1ned in English 'ivith "ible" or "able)' suffixes. For example, to say that a lump of sugar is soluble is to say that it is disposed to dissolve. Solubility is a dispositional property. It is not hard to give a "behavioral analysis" of solubility. To say that something is water-soluble is to say that it is disposed to behave in a certain way when placed in a certain situation:
Xis soluble (in water)= df If X were immersed (in water), then X would dissolve.
Dissolving is a behavior; to itnmerse so1nething is to place it in a particular sort of environ1nent. This definition of solubility confortns to behavioristic requirements, since it makes no mention of the inner state that soluble substances might occupy. I have no
Chapter 20: Logical Behaviorism
" ... ' '.''' ... ''' '' ... ' .... ' .... ' ...... ' .. '. '. ' .... ' .......... ' .......... ' objection to this behaviorist definition of solubility. My point is that the adequacy of this definition does not mean that there is no inner state of lumps of sugar that makes them dissolve 1vhen immersed. In fact, che1nistry provides a scientific description of the inner features of sugar that make sugar dissolve when immersed in water. Solubility is a dispo- sitional property, but this is consistent with the fact that soluble substances have internal, structural properties that make them behave as they do.
How is this relevant to the logical behaviorist's attempt to analyze mental states? Even if wanting to drink water were a behavioral disposition-one that could be described with- out 1nentioning any inner state of the organis1n-this \vould not sho\v that \Vants aren't inner states. On the contrary, there is every reason to think that \vhen something has a dispositional property, there is a physical basis of that disposition. When an organism is disposed to drink water, it is entirely appropriate to ask what it is about the organism's internal makeup that disposes it to act in this way.
This completes my criticism of logical behaviorism's positive thesis. I claim that the meanings of mentalistic terms (like "wants;' "believes") cannot be analyzed in purely behavioral language. In addition, I have argued that even if such an analysis could be pro- vided (by describing mental states as "dispositions to behave''), it would not follow that mental states aren't inner causes of behavior.
PAIN WITHOUT PAIN BEHAVIOR?
Curare, the poison that some South American Indians traditionally used on their darts, paralyzes its subject. In the i93os and i94os, scientists purified curare and studied its effects on the central nervous system. At first, some doctors thought that curare was a painkiller; they noticed that if you give curare to your patient before sur- gery, the patient will not move under the knife. After surgery, the patients complained that they experienced great pain, but for a time, the physicians didn't believe them (many of the patients were children). Eventually a physician volunteered to undergo surgery with curare; he reported that the pain was vivid and excruciating. After that, doctors realized that curare isn't a painkiller; it simply immobilizes the patient.
Daniel Dennett (in "Why You Can't Make a Computer That Feels Pain," Brain· storms, MIT Press, i978) suggests the following puzzle. An amnesiac is a drug that makes you forget. Suppose that you gave someone curare and an amnesiac before surgery. Would the individual feel pain? The patient would not say "ouch" or writhe during surgery. And after surgery, the patient would not say "that was horribly painful." Let's use the term "pain behavior" to name any behavior that usually accompanies pain. In Dennett's puzzle, there is no pain behavior, either during the surgery or after The question is, can there be pain without pain behavior? Dennett's example suggests that the answer is yes. Would you be prepared to undergo surgery with curare and an amnesiac as the only drugs you get? If not, the reason is probably that you think that you would still feel pain.
What does this example show about the claim that mentalistic terms do not describe the inner causes of behavior, but describe only the behavior itself?
220 Part Four: Philosophy of Mind
...... '.' .. ' '' ........ ' ....... ' .. ' .. '' .. '' '' .. ' ... ''.'.''.'.''.' '''.'.''., Review Questions
1. What is the "problem of other minds"? Does the analogy argument from one's ovvn case solve this problem? Ryle thought that the mental states of others would be unknowable if mental states \Vere inner causes of behavior. Why did Ryle think this?
2. What is logical behaviorism? Can a logical behaviorist analysis be given of the statement "Jones believes that there is rat poison in the gravy"?
3- Suppose that a mentalistic concept (like \Vanting to drink \Vater) could be analyzed as a disposi- tion to behave. \Vould that show that \van ts aren't inner causes of behavior?
Problems for Further Thought
1. I've suggested that an abductlve argument can solve the problem of other minds. Construct such an argument, making clear ho\v the Surprise Principle (Chapter 3) applies.
2. What does it mean to say that Xis "nothing but" Y? It can't mean just that X \Vouldn't exist if Y didn't. \Ve wouldn't exist \Vithout oxygen. It doesn't follow that vve are "nothing but" oxygen.
3- In one of the readings for this part of the book, Bertrand Russell attempts to solve the problem of other minds by introducing a postulate that he says "is not required in physics." \Vhat is this postulate/ How does it apply to the problem of other minds? Is the postulate also applicable to examples concerning physical objects that have nothing to do with the mind? Are there excep- tions to this principle? Are there many such exceptions?
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
RESEARCH AND EXPLORE
'l(l{EXPLORE
l. Problem of other minds 2. Logical behavio(1sm
References
Bertrand Russell, "Other Minds Are Known by Analogy from One's Own Case," from Human Knowledge: Its Scope and Limits (London: Unwinn Hyman Ltd., i948), pp. 482-486. Reproduced with permission of Taylor & Francis Books Ltd.
chapter 21 ..............................
Methodological Behaviorism
Chapter Outline
The Negative Thesis: Psychology Should Avoid Belief/Desire Explanations
Methodological Behaviorism's Positive Thesis
First Objection to Behavioristn,s Positive Thesis: Novel Behaviors
Second Objection to Behavioris1n's Positive Thesis: It Assu1nes that Environn1ental Deter1ninis1n Is True
The Two Objections Summarized
Logical behaviorism was the subject of the last chapter; methodological behaviorism is the subject of the present one. Although the names sound similar, the doctrines themselves are very different. Logical behaviorism says that the meanings of conunonsense inentalistic concepts can be analyzed in purely behavioral terms. Methodological behaviorism is not a thesis about "fhat such conunonsense tern1s mean; rather, it is a recomtnendation for ho'" the science of psychology ought to be pursued.
In a curious way, methodological and logical behaviorisms take opposite views on the nature of co1nn1onsense 1nentalistic language. As noted earlier, logical behavioris1n rejects the idea that beliefs and desires are inner states that cause behavior. Method- ological behavioris1n, by contrast, accepts the idea that our commonsense n1entalistic vocabulary refers to inner states. Methodological behaviorism then argues that a scien- tific psychology should avoid talking about beliefs and desires precisely because they are inner states.
In this chapter, I'll divide methodological behaviorism into a negative and a positive thesis. (This is similar to my division of logical behaviorism into a negative and a posi- tive thesis in the previous chapter.) First, I'll analyze the negative thesis; this is the claim that psychology should not attempt to explain behavior in terms of people's beliefs and
221
222 Part Four: Philosophy of Mind
'''' '''.' ''.'''' ''''''',' •• ' '',' •• '' '.' ''. ''' '''' '''''.'''' >'''''
desires. Then, I'll consider methodological behaviorism's positive thesis; this is the idea that people's behavior can be explained solely in terms of their history of conditioning.
The Negative Thesis: Psychology Should Avoid Belief/Desire Explanations
Why should psychology avoid explaining the behavior of agents by attributing beliefs and desires to them? B. F. Skinner, who was a leading proponent of behaviorism, provided several ans\'vers to this question in his books Science and Htanan Behavior (Ne\v York, Free Press, 1953), Beyond Freedo111 and Dignity (New York, Knopf, 1971), and About Behavior- is111 (New York, Random House, 1974). Skinner's first objection to mentalistic explanation is that beliefs and desires aren't observable. They are hidden. All we can directly observe is the behaviors of others, not what goes on in their minds. By now, my view of this kind of argument should be clear. Science legitimately talks about things that aren't or can't be observed. The fact that beliefs and desires can't be observed directly doesn't mean that claims about an individual's beliefs and desires can't be tested by observing behavior. For example, suppose I claim that Jane wants a drink of water and believes that there is water in the cup before her. Suppose I further claim that she has no other desires that would override her desire for the water. These claims about the agent's mind predict that she will reach for the water and drink it. If that prediction fails, I've gained evidence against the description just formulated about Jane's mind. As in the Mendelian case, claims about what is not observed can be tested by examining what they imply about what can be observed.
Another reason that Skinner gives for rejecting mentalistic theories is that they are "too easy:' In the example, I attributed to Jane a combination of beliefs and desires that predict something about her behavior. Suppose she doesn't reach for the water. What will I do? I'll modify my claim about what she thinks and wants. I won't abandon my assump- tion that Jane has beliefs and desires and that these cause her behavior. So, apparently, the assumption of mentalism-that behavior is caused by beliefs and desires of some sort- is1it tested by observing behavior. Indeed, it appears that no matter what Jane does, I can always formulate a belief/desire story that is consistent with what I observe.
I want to clarify this point of Skinner's by distinguishing specific belief/desire hypoth- eses from what I'll call the 111e11talistic thesis, which is the claim that an individual's behav- iors are caused by the beliefs and desires he or she possesses. The specific attributions are testable against the data of behavior; however, the mentalistic thesis does not seem to be. Skinner believes that the claims of science must be testable by observation; hence, he con- cludes, the mentalistic thesis does not deserve to be part of a scientific psychology. Notice that even if the mentalistic thesis is untestable, this would not imply that specific belief! desire attributions are untestable. The requirement that scientists should only put forward theories that are testable has the curious consequence that they are allowed to advance specific belief/desire explanations but are not allowed to say that mentalism is true. But mentalism is logically entailed by each specific belief/desire explanation. Can't we assert \vhat our confirmed theories entail?
In Chapter 9, I discussed the idea that scientific claims should be testable. I argued that the thesis of falsifiability is implausible. I now want to make a related point: There are many pe1fectly respectable scientific state111ents that cannot be refuted by the result of a single
Chapter 21: Methodological Behaviorism 223
'' .. '' .. ''' ..... ' ' ..... '.". '''.'' '' ...... ''''.'.'.' '.'. ' ... ''. ' ...... ' ... . experiment. In the present case, we have an example of this kind of statement. Statements that describe the basic tenets of a research program are not falsifiable. Mentalism is such a thesis; it says that theories explaining behavior can be developed by extending and refin- ing the ordinary concepts of belief and desire. Time will tell whether this is a good idea for psychology or a dead end. The fact that a single theory within this framework fails does not show that the whole framework is bankrupt. One rotten apple need not show that the entire barrel is spoiled.
Curiously enough, precisely the same point applies to Skinner's own research program. Skinner rejects 1nentalism and espouses behaviorism. This is the thesis that behaviors can be explained by describing the stimuli people have received from their environments. Skinner says we don't need to describe the inner states-mental or physical-of individu- als in order to explain their behavior. This behaviorist thesis states the framework of the Skinnerian research program. Suppose I follow its dictates and construct a detailed behav- iorist explanation of why yon are now reading this page. Suppose further that this explana- tion turns out to be mistaken. Does that refute behaviorisn1 as a general research progratn? I would say not (and Skinner would probably agree).
It is appropriate to demand that specific explanations be testable against the data of observations. It is not appropriate to demand that theses about the framework of a research program should be falsifiable. Mentalism can't be faulted on the ground that it isn't testable. Specific mentalistic hypotheses can be tested. Time will tell whether the general thesis of inentalisn1 \Vill turn out to be correct. In this sense, the frame,vork of n1entalism, like the frame\vork of behaviorism, is testable-it is testable "in the long run:'
Methodological Behaviorism's Positive Thesis
I'll no\v consider inethodological behaviorism's positive claitn-that behavior can be ex- plained without describing the inner states of the organism. I believe that there is a simple fact about people that shows that behaviorism can't be successful-that the explanations it demands do not exist. Recall that methodological behaviorism does not deny that we have inner mental states. It claims, rather, that psychology does not need to discuss them in its attempt to explain behavior. So behaviorism grants that the causal chain linking environ- mental stimuli to behavior has the following structure:
Environn1ent ------') Inner state of organisn1 ------') Behavior
Methodological behaviorism maintains that my present behavior can be explained in terms of (1) the past environments I've occupied and the behaviors I produced in them and (2) niy present environment.
Skinner's theory of stimulus conditioning shows how such explanations can be con- structed. In his research, Skinner manipulated the behavior of chickens and other organ- isms by placing them into a totally controlled environment-a Skinner box. For example, a chicken can be conditioned to peck at a key in its cage when the light is on. This is how the conditioning is done. When the chicken is first placed in the box, it will peck occa- sionally. The box is so arranged that if the chicken pecks when the light is on, it receives a pellet of food. If it pecks when the light is off, it obtains no reward. After some time,
224 Part Four: Philosophy of Mind
•••• > •••••••••••••••••••••••••••••••• '.'.'. ''' ' •••••• '' ••• '.' '''' •••• ' •• ''
the chicken's pecking starts to happen pretty much only when the light goes on. In this stimulus/response experilnent, the chicken is conditioned to peck \vhen and only \vhen the light is on. The conditioning \Yorks by re\varding the chicken in so1ne circu1nstances but not in others.
Suppose that this conditioning process has taken place. You then look into the cage and see that the chicken is pecking. How might you explain this behavior? A mentalistic explanation might say that the chicken is now pecking because it wants food and believes, based on its past experience, that pecking produces pellets of food. This is a mentalistic explanation-it is stated in terms of beliefs and desires. Skinner believes that this mental- istic story is entirely unnecessary. You can give a behaviorist explanation of the present behavior as follows: (1) in the past, the chicken was conditioned to peck when the light was on and (2) right now, the light is on. Notice that Skinner's explanation makes no mention of beliefs and desires. We can formulate the behaviorist explanation as a deductive argument:
If the light is on, then the chicken pecks.
The light is on now.
Hence, the chicken now pecks.
You explain the present behavior by deducing it from the two facts listed as premises. Skinner thinks that what is true of the chicken's pecking in the experiment is true of
all the chicken's behaviors, both those produced in the experiment and those that occur in less controlled environments (like the barnyard). In addition, Skinner believes that human beings are like the chicken in relevant respects. Each of our behaviors can be explained by describing our history of stimulus conditioning. This is the thesis on which the behaviorist research program rests. We don't need to talk about what human beings think and want if we are to explain what human beings do.
First Objection to Behaviorism's Positive Thesis: Novel Behaviors
Are we like the chicken in a Skinner box? Critics of behaviorism, like the linguist Noam Chomsky, have claimed that novel behaviors cannot be explained in the way the behavior- ist de1nands. Let's look \Vith soine care at an example of Skinner's to see \vhat this criticism means.
Suppose you are walking down the street one day and a robber jumps out of the shad- O\VS, sticks a gun in your face, and says ((your money or your life:' You fumble nervously and then hand over your wallet. How should your behavior be explained? What makes the behavior novel is that you have never been robbed before. The only time anyone ever waved a gun in your face was when you were a child and the gun was a water gun. In that case, you didn't hand over your wallet; rather, you giggled and ran away. So why didn't you do the same thing this time?
The behaviorist needs to show that the present stimulus situation is similar to past ones in which you behaved similarly. Skinner explains the chicken's present pecking by saying that the present situation is like past situations in which the chicken pecked. The similarity
Chapter 21: Methodological Behaviorism
........................... '' ............ ' .... ' ........... '. ' ... ', .... ' .. . is that the light is on in the Skinner box. The problem is that the present robbery is similar to many past situations you have been in. It is similar to your childhood playing with water guns. It is also similar to the time your mother threatened to spank you if you didn't clean up your room. After all, the robbery involved a gun and a threat. Which of these past expe- riences is the right one to appeal to in explaining your present behavior?
Remember that you have never been robbed before. The only time anyone waved a gun in your face it was a water gun and you giggled and ran away. And when your mom threatened to spank you, you apologized and cleaned up your room; you did not hand her your wallet. What past episodes can a behaviorist say are similar to, and also explain, your current behavior? An obvious answer is that you learned in the past that you should accommodate those who threaten you. Perhaps you were rewarded in the past for doing this and so the behavioral pattern became fixed. Although this sounds like the right sort of explanation to develop, it isn't true to the demands of behaviorism.
In the chicken case, there was a physical property of the earlier stimulus situations that recurs in the present. In the past, the light was on. In the present case, the light is on. It is perfectly permissible for a behaviorist to describe the history of conditioning in terms of such physical properties. What the behaviorist can't say, however; is that you are the sort of individual who behaves in ways you believe are accommodating when you believe that you are being threatened. This description of your stimulus history refers to inner mental states. These are just the items that the behaviorist says should not be included in the explanation.
I doubt that there is a physical similarity between past threats and the present one. In the past, your mom may have said, "If you don't clean up your room, I'll give you a spank- ing:' Your behavioral response was not to hand your wallet over, but to apologize and tidy things up. The only way to describe how this past event is similar to the present robbery is to describe your beliefs and desires.
If our behaviors always obeyed patterns that were describable in terms of physical similarities between past and present stimulus conditions, behaviorism might work. But this often seems to be radically untrue. In lots of cases, what makes the past similar to the present is that we conceptualize past and present in similar ways. Take this mental- istic similarity away and behaviorism delivers radically mistaken predictions. In terms of purely physical similarities, I suppose that your participation in a water gun battle as a child may be more similar to the present robbery than your being scolded by your mother. If you were to behave now in ways that are similar to the ways you behaved in previous physically similar past events, the prediction would be that you would giggle and run away.
It is an important fact about human beings that small physical differences in a stimulus often make enormous differences in the response, and huge physical differences in the stimulus often make no difference in the response. If a robber walks into a bank and hands the teller a note that says "You give me the cash;' the teller produces one response. If the note had said "Me give you the cash;' the response might be very different. The notes aren't that different physically, so why was the response so different? Likewise, receiving a paper with a few marks on it can produce the same result as hearing a verbal command. But the marks on paper ("You give me the cash") and a verbal message ("Empty the cash drawers into this sack") are physically very different.
226 Part Four: Philosophy of Mind
'."'.' ''' .... '' '''.' '.'''.' .... ' ' .. ' .. '' '. '".' '' .... ','''. ''''' ''' .. '' .. '' MINIMAL EXPLANATION VERSUS PEEP EXPLANATION
The behavior of a conditioned chicken in a Skinner box can be explained without de- scribing the chicken's internal state. Why is the chicken now pecking the key? Skinner's answer is that the chicken pecks whenever the light is on, and the light is now on.
Since we have explained the behavior without mentioning the chicken's internal state, Skinner concludes that describing the internal state would be of no explanatory relevance. You don't need to describe what is going on inside, so why bother?
I grant that the purely behavioral account is an adequate explanation. However, I think that there can be a deeper explanation of the chicken's behavior.
Consider an analogy. Why did this white cube just dissolve? A behaviorist rnight an- swer by noting that (1) the cube was made of sugar, (2) it was just immersed in water, and (J) all sugar dissolves when immersed in water. Granted, this is an explanation. But it is possible to do better.
What is it about sugar that makes it dissolve when immersed? Here we request in- formation about the internal structure of the sugar lump. Since sugar is water-soluble but wood is not, we expect that the one has an internal state that the other lacks.
! Once described, this internal state of the lump of sugar is relevant to explaining why it just dissolved.
The chicken pecks when the light is on. Surely we can ask what it is about this chicken that makes it peck when the light is on. A conditioned chicken has a disposi- tion to behave that an unconditioned chicken does not have. It is legitimate to expect that the one has an internal state that the other lacks. Once described, this internal state of the conditioned chicken is relevant to explaining why it is now pecking.
Perhaps it is possible to explain the chicken's behavior in the way the methodolog- ical behaviorist demands. However, this does not mean that other explanations that violate behaviorist principles are irrelevant to understanding why the chicken pecks.
Behaviorism sounds plausible for two reasons. First, it is true that claims about the mental states of individuals should be tested against behavioral data. Second, an individ- ual's behavior is importantly influenced by the environments he or she has experienced. Neither of these truis1ns, ho,vever, sho'ivs 'ivhat 1nethodological behaviorism maintains- that it is possible to explain (and predict) behavior without mentioning mental states.
Second Objection to Behaviorism's Positive Thesis: It Assumes that Environmental Determinism Is True
One final shortcoming of methodological behaviorism should be noted. Behaviorism is a radically environmentalist doctrine. That is, 'ivhen it conies to the nature/nurture problem-the problem of saying how much a behavior is explained by genes and how much by environ1nent-the behaviorist is entirely on the side of the environ1nent. By saying that inner states need not be described in explaining behavior, the behaviorist is excluding genes as well as environmentally acquired mental states. My point is that envi- ronmentalis1n 1nay be true in some cases, but not in others.
Chapter 21: Methodological Behaviorisn1 227
. '' .. ' ..... ' ... ' .. '.' .. ' ' .... ' ... ' ... '.'.'.' ' .. '.'.'' '''.' .. '.''.'. '' '''. Returning to the exa1nple of the chicken, it is interesting to note that there are so1ne be-
haviors that you can't condition a chicken to perform. Speaking English is something that human beings learn to do when they are in the right kind of environment. A chicken in the sa1ne environ1nent 'vill not acquire this skill. Why not? Since the environments are the sa1ne, the ans,ver can't be environmental. Presurnably, there is a genetic difference bet,veen chickens and human beings that explains this behavioral difference.
What is true behveen species inay also be true, in some cases, 'vithin a species. There may be genetic differences between human beings that are relevant to explaining differ- ences in their behaviors. This is novv a controversial area of investigation. My point is that n1ethodological behaviorism assumes in advance of any detailed scientific analysis that en- vironn1ental variables suffice to explain behavior. This should not be assumed as a inatter of doctrine, but should be empirically investigated on a case-by-case basis.
Scientists are now considering the possibility that there may be a significant genetic con1ponent in conditions such as depression and schizophrenia. 1'hat is, in addition to the environ1ne11tal causes of these conditions, there may also be genetic causes. In contrast, no one now believes that there is a genetic component in explaining why some people speak one language while others speak another. In this case, variation in the behavior has an en- tirely environ1nental explanation.
Environ1nental determinis1n is the vie'v that genetic differences do not help explain any behavioral differences. Genetic determinism is the vie'v that environn1ental differences do not help explain any behavioral differences. Both of these views are too sweeping and extreme to be plausible. First, we must consider different behaviors separately. The kind of explanation we offer for schizophrenia may differ from the kind we want to give for speaking English. Second, 've n1ust recognize that there are "1nixed" proposals that deserve a hearing; we 'vant to consider the idea that some behaviors have both environ1nental and genetic causes.
The Two Objections Summarized
In conclusion, there are two kinds of questions that methodological behaviorism can't ad- dress. The first is the kind of question typified by the robbery example. When I ask how your present behavior of handing over the money is explained by your previous experi- ences, it seems essential to describe your beliefs and desires. There is no purely physical si1nilarity behveen the present holdup and your previous environ1nents that provides an explanation. Rather, the siinilarity exists only because you conceptualize the present situ- ation in a way that resembles the way you conceptualized previous (physically different) situations. Second, there is the idea of genetic (nonenvironmental) causes of behavior. If these exist, then methodological behaviorism will be unable to explain in some cases why some organisms behave differently from others. Methodological behaviorism began by re- jecting n1entalism; it claiined that inental states are irrelevant to explaining behavior. But, in fact, its thesis excludes brain science and genetics as 'veil. The real thrust ofbehavioris1n is to claim that the explanation of behavior need not describe the inner state of the organ- ism. Mentalis1n describes 'vhat goes on inside, but so do brain science and genetics.
By focusing on what happens in tightly controlled environments, behaviorism may seem to have established a general pattern for explaining all behavior. It is true that the pecking of the chicken in the Skinner box can be handled in the way Skinner demands.
228 Part Four: Philosophy of Mind
However, it is a mistake to think that successful explanation in this case can be generalized to all other behaviors. Take the chicken out of the Skinner box and perhaps it will produce novel behaviors, just as you did when the robber said "your money or your life:' And apart from the issue of novel behaviors, there is the nature/nurture issue, which methodological behaviorism settles in advance oflooking at the relevant data.
Review Questions
l. What is the difference between methodological behaviorism and logical behaviorism?
2. What is mentalism? Is mental ism testable? How does the distinction bet\veen specific explana· tions and framework assumptions about a research program bear on this question?
3- Why do so-called novel behaviors pose a problem for methodological behaviorism? What does novel mean? Isn't every behavior novel in some respect or other?
4. What does methodological behaviorism say about the relevance of genetics and brain science to the explanation of behavior? Why would innate differences in behavior be a problem for behaviorism?
Problems for Further Thought
1. Some psychotherapists call themselves "behaviorists." They treat patients with phobias (extreme fears) by helping them modify fear behavior (like increased heart rate, perspiration, anxiety) rather than by helping them understand why they have the array of beliefs and desires that produce the reaction. Consider people who are enormously afraid of being in closed areas (claustrophobia). Instead of trying to get a claustrophobic individual to understand why he or she has this fear, a behavioral therapist might concentrate on "desensitizing" the individual. The goal would be to modify the individual's behavior. Treatment might begin with very brief exposures to closed areas; gradually, as the patient gains confidence, longer exposures would be attempted. Here's the question: Do the criticisms l've made of methodological behaviorism imply that behavioral therapy will not be effective? Defend your answer.
2. Saying that lemons are yellow is a kind of event. Can it be defined by its physical characteristics? ls saying the English sentence "lemons are yellow" definable by its physical characteristics? \Vhat is the difference between these two kinds of events?
J. If I strike your knee with a hammer, it will jerk. Is this due to the beliefs and desires you have? If not, why is your going to a movie tonight any different?
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchLab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
RESEARCH AND EXPLORE
!.!(..{EXPLORE
1. Methodological behaviorism 2. B. F. Skinner
chapter 22 ..............................
The Mind/Brain Identity Theory
Chapter Outline
The Identity Theory Is an A Posteriori Claim
Materialisn1
Progress in Science
Dualism Resembles Vitalism
A Correlation Experiment
The Principle of Parsimony
So far I've examined three approaches to the mind/body problem. These were Cartesian dualism (Chapter 19), logical behaviorism (Chapter 20), and methodological behaviorism (Chapter 21). Although dualism and methodological behaviorism agree that mental states are inner causes of outward behavior (a thesis that I have called "mentalism''), each says more than this. Dualism asserts that your mind is a nonphysical object; methodological behaviorism recommends that scientific explanations of behavior ignore the mind (in part because you allegedly can't have knowledge of another person's mind). Logical behavior- ism is the one theory of these three that denies mentalism.
The Identity Theory Is an A Posteriori Claim
Cartesian dualism and logical behaviorism are very different positions. Yet the arguments that each presents have something in common: neither appeals to any scientific finding as evidence for \Vhat it 1naintains. Cartesian dualis1n says that \Ve can recognize by introspec- tion and philosophical reasoning that the mind has a property that the body lacks. Logical behaviorism says that \Ve can see that 1nental states are not inner causes of behavior just by analyzing the meanings of commonsense mentalistic terms. Both theories are defended by arguments that are intended to be a priori.
229
Part Four: Philosophy of Mind
The Mind/Brain Identity Theory isn't just a different theory from those surveyed so far; it is also a different kind of theory. The identity theory argues for its solution to the mind/body problem by describing the progress that science has made so far and by pre- dicting the progress that science will make in the future. The relationship of the mind and the body, says the identity theory, is something that science discovers by observation and experiment. It isn't something that armchair philosophy-a priori reflection and linguistic analysis-can hope to resolve. The identity theory is offered as an a posteriori proposition about the relationship of mind and body.
Materialism
The identity theory asserts that the history of science has been marked by success after success for a doctrine called materialism or physicalism. This is the view that every object is a material (physical) object. But what does that mean? Does it just mean that every object has at least one physical property? There must be more to materialism than this. After all, Descartes and other dualists agree that your thoughts have temporal location. For example, yesterday, you wanted an ice cream cone; this is a physical property that your episode of desiring had. What distinguishes materialism from dualism is this: materialists hold that if two objects are different in any way at all, they must exhibit physical differences. If you love jazz and I do not, then there must be a physical difference between us. If we were physically exactly the same, either we'd both love jazz or neither of us would. The materialist's slogan is 110 difference without a physical difference. Dualists disagree with this principle. They think it is possible that two physically identical lumps of matter might be such that one of them has a mind while the other does not. A jazz lover and a jazz hater might be molecule-for- molecule identical; they would be physically identical but psychologically distinct. Dualists think this is possible because they think that the mind is a not a material object at all.
Progress in Science
Looking back on the history of human knowledge, we find phenomena that weren't well understood scientifically. For example, lightning was once profoundly perplexing. The ancient Greeks thought that lightning was Zeus's thunderbolt. But in the eighteenth and nineteenth centuries, scientists studying electricity established that lightning is an elec- trical discharge. Science was able to take a phenomenon that earlier had seemed to defy physical description and show that it can be understood within the framework of the laws of physics. Consider a statement that describes this scientific discovery:
Lightning is one and the same thing as a kind of electrical discharge.
This statement wasn't established by analyzing the meaning of the word lightning. The statement isn't an a priori truth, knowable by linguistic analysis; rather, it is an a posteriori truth, discovered by scientists using observational evidence. It could be established only by observation and experiment.
In the past century, there was a similar perplexity about the nature oflife. What is the difference between living things and things that are not alive? A doctrine called vita/ism held that living things contain a nonphysical substance that animates them with life-an
Chapter 22: The Mind/Brain Identity Theory
''''''''.'''' ''' '' '''.'' '' '.'' '.' '·''' .. ''' elan vital (a vital fluid). In this century, vitalism has been undermined by the discoveries of molecular biology. We now know that life is a physico-chemical phenomenon. Organisms are made of the same basic elements (carbon, oxygen, etc.) as nonliving things. Life differs fron1 nonlife because of ho\v these basic physical constituents are arranged. There is no extra nonphysical ingredient that magically makes organisms "alive:'
Biology in the last one hundred years has discredited vitalism. The discovery of the mo- lecular basis of life has been a triumph for materialism. By this I don't mean that each and every detail of the physical basis oflife is well understood; if this were true, there would no longer be a need for research in molecular biology and biochemistry. Rather, what I mean is that the details are sufficiently well understood that there is no longer any real doubt that 1netabolis1n, reproduction, digestion, respiration, and so on are all physical processes.
Dualism Resembles Vitalism
Cartesian dualisn1 and vitalis1n are sitnilar. Dualism says that an individual's mind is a non1naterial substance. Thinking beings therefore are said to contain a nonphysical ingre- dient not found in beings that don't have minds. Vitalism says that living things contain a nonphysical substance-an elan vital-that is unique to them. Just as dualism opposes a 1naterialistic solution to the 1nind/body problen1, so vitalism opposes a materialistic an- swer to the question of how biological and physical processes are related to each other.
The Mind/Brain Identity Theory says that what has happened to the problem of life is in the process of happening to the problem of mind. We now have some knowledge of the physical basis-the neurophysiological basis-of the mind. If science progresses the way it has in other areas, this materialistic understanding will broaden and deepen. The identity theory predicts that science will fully vindicate the thesis that an individual's mind and brain are one and the same entity. Each mental characteristic (for example, believing that snow is white, wanting a drink of water, and feeling pain) is identical with some physical characteristic. The identity theory is a version of materialism: Mental objects are physical objects and mental characteristics are physical characteristics.
During the 1950s and 1960s, identity theorists-for example, U. T. Place, in "Is Con- sciousness a Brain Process?" and J. J. C. Smart, in «Sensations and Brain Processes" (both reprinted in V. Chappell [ed.], The Philosophy of Mind, Prentice-Hall, 1962)-gave the following as an example of what they meant. Brain scientists at that time suggested that feeling pain is one and the sa1ne event as having the c-fibers in one's brain fire. 1'his neu- rophysiological hypothesis about the physical basis of pain later turned out to be untrue. This didn't deter identity theorists, however, who are quite happy to leave it to science to work out the details of the general hypotheses they formulate. The identity theorist claims that feeling pain is identical with being in some physical state or other; it is for science to tell us precisely what this physical state is.
A Correlation Experiment
Consider this example in more detail. How might a scientist investigate the connection of pain and c-flber firing? The scientist would want to find out whether people experience pain whenever their c-fibers fire and whether people have their c-fibers fire whenever they
231
Part Four: Philosophy of Mind
' .. ''' .... '' .. '' '' ''.' '' '' ''' ''''' ''' '. '' '.'''' experience pain. Here's a simple experin1ent that 'vould help detennine if these mental and physical events co-occur. A probe inserted into the subject's brain "\vould indicate "\vhen the subject's c-fibers fire. Perhaps the subject's o'vn testilnony, based on introspection, "\vould indicate \Vhen he or she experiences pain. The experirnenter occasionally hits the subject's thumb with a hammer. Ignoring complications, suppose that different human subjects, when placed in this (morally questionable) experiment, generate the same data. Suppose that the data show a perfect correlation between experiences of pain and c-fiber firings- the one kind of event occurs if and only if the other does.
Does this, by itself, show that the identity theory is true? No. Dualism is consistent with there being a perfect correlation of mental events and physical events. Dualism claims that feeling pain and having your c-fibers fire are two distinct kinds of event. The identity theory, on the other hand, says that feeling a pain and having one's c-fibers fire are perfectly correlated because they are one and the same kind of event.
Ho\v are \Ve to choose bet\veen dualism and the identity theory? One consideration, mentioned in Chapter 19, is that dualism makes it difficult to understand how mind and body can causally interact. The identity theory, on the other hand, has no difficulty accommodating this fact. However, the fourth Problem for Further Thought at the end of Chapter 19 suggests that this objection to dualism rests on a questionable assumption about \vhat causality involves. Is there some other basis on \Vhich to choose behveen dual- ism and the identity theory?
The Principle of Parsimony
Identity theorists offer a second reason for rejecting dualism. They claim that the iden- tity theory is more parsimonious (simpler) than dualism. Their idea is that the scientific method says that \Ve should prefer sitnpler theories over 1nore co1nplex ones "\vhen both are consistent \Vith the observations. It isn't just that n1ore parsitnonious ideas are easier to think about or are aesthetically inore pleasing. Rather, their idea is that rnore parsitnoni- ous theories are more likely to be true.
1'he identity theory is 1nore parsitnonious because it claitns (if the experiinent just described were to yield a perfect correlation) that feeling pain and having one's c-fibers fire are one and the same event. A dualist, on the other hand, says that these are two different events. The Principle of Parsimony, sometimes called "Ockham's Razor" after the medieval philosopher William of Ockham, is an abductive principle. It says that we should prefer explanations that n1inimize the nu1nber of entities, processes, and events they postulate. The usual slogan is "don't multiply entities beyond necessity;' though this isn't in Ockham's writings. Since one is a smaller number than two, the Mind/Brain Identity Theory is more parsimonious than dualism. Identity theorists count this as a reason for thinking that the identity theory is true and dualism is false, given that both are consistent with the observed correlation of mental and physical states.
In the vitalism controversy, no biologist took seriously the idea that an elan vital exists once the physical bases of life processes became reasonably clear. Rather, the conclusion was that an organisnis being alive is nothing above and beyond its having certain physical processes going on in its body. To say otherwise would be to admit a further entity-an elan vital-without necessity. This would contradict the Principle of Parsimony. Biologists
Chapter 22: The Mind/Brain Identity Theory 233
. '' '.'". '.'.' '' ' .. ''.' .. ' '.' .. '. ''. '' .. ' .. ".' ... ' .... ' ...... '.' '' ''' .. '. took this principle to heart once they obtained physical explanations oflife processes. They realized that there was no reason to postulate the existence of an e/an vital. For this reason, vitalism was rejected and materialism accepted. If a perfect correlation could be found betvfeen n1ental characteristics and physical ones, \Vhat reason could there be to postu- late the existence of an immaterial mind-one that just happens to exhibit psychological characteristics that are correlated with the physical characteristics of the brain? This seems entirely gratuitous-a violation of the Principle of Parsimony. The identity theory asserts that dualism should be rejected and materialism accepted if science discovers perfect cor- relations between mental and physical characteristics.
Let's be clear on what the Principle of Parsimony recommends. In the problem of life, the principle leads one to deny the existence of an immaterial vital substance. It doesn't say that one should suspend judgment as to whether such a substance exists. Likewise, mind/ brain identity theorists argue that parsimony offers a reason to think that immaterial tninds do not exist. Their recommendation is not to suspend judgn1ent about dualism, but to conclude that dualism is false. In both cases, Ockham's Razor leads one to conclude that X does not exist, not to conclude that one should suspend judgment as to whether X exists.
There is an interesting and difficult philosophical problem here, which I woiit attempt to solve: Why should the greater simplicity or parsimony of a theory be a reason to think that the theory is true? Does using the Principle of Parsimony in this way require that we assume that nature is simple? This seems to be the rationale that Isaac Newton offers for using the principle in scientific reasoning (see the accompanying box). If using the prin- ciple requires that \Ve assun1e that nature is simple, \Vhat evidence do \Ve have that nature really is simple? This question should remind you of the question we explored in connec- tion with Hume's views on induction in Chapters 15-17 concerning the principle of the unifor1nity of nature.
NEWTON ON PARSIMONY
The Principle of Parsimony has been used in many scientific debates. In his Principles of Natural Philosophy (1690), Newton presents four "Rules of Reasoning in Philoso- phy." The first two concern abduction, the second two induction:
1. We are to admit no more causes of natural things than such as are both true and sufficient to explain their appearances. To this purpose, the philosophers say that Nature does nothing in vain, and more is in vain when less will serve; for Nature is pleased with simplicity and affects not the pomp of superfluous causes.
2. Therefore to the some natural effects we must, as far as possible, assign the same causes. As to respiration in a man and in a beast, the descent of stones in Europe and in America, the light of our culinary fire and of the sun, the re- flection of light in the earth and in the planets.
3. The qualities of bodies, which admit neither intensification nor remission of degrees, and which are found to belong to all bodies within the reach of our experiments, are to be esteemed the universal qualities of all bodies whatsoever.
234 Part Four: Philosophy of Mind
For since the qualities of bodies are only known to us by experiments, we are to hold for universal all such as universally agree with experiments, and such as are not liable to diminution can never be quite taken away. We are certainly not to relinquish evidence of experiments for the sake of dreams and vain fie· lions of our own devising; nor are we to recede from the analogy of Nature, which is wont to be simple and always consonant to itself. ...
4. In experimental philosophy we are to look upon propositions inferred by general induction from phenomena as accurately or very nearly true, notwithstanding any contrary hypotheses that may be imagined, till such time as other phenom- ena occur by which they may either be made more accurate or liable to excep· tions. This rule we must follow, that the argument of induction not be evaded by hypotheses.
Rule l is relevant to the debate between the identity theory and dualism. Rule 2 is simi· lar to the Principle of the Common Cause discussed in connection with the Argument from Design (Chapter 6). Rule 3 bears on the relevance of Descartes' method of doubt (Chapter i 3) to the progress of science.
There is an alternative argument for using the Principle of Parsimony that does not rest on the assumption that nature is simple. The history of science shows that using the Principle of Parsimony has repeatedly led to important scientific advances. In Aristotle's physics (which we considered briefly in Chapter 4), the motions of celestial bodies and the motions of objects here on Earth are said to follow different rules. One of the great achievements of Newton's physics is that it provided a single set oflaws that covered both. Ne"vton's theory is in this respect sitnpler than Aristotle's; it unifies diverse observations. The rejection of vitalism in the nineteenth and tvventieth centuries provides another such example. These achievements are important scientific milestones. Ho\vever, \Ve must be careful not to attend only to episodes in which more parsimonious theories turned out to be better than their less parsimonious rivals. There also have been episodes in which the opposite turned out to be true.
Review Questions
l. What is materialism? Is it the vie\v that money is the most important thing?
2. What is vltalism? Is vitalism sho\Vn to be true by the fact that living things contain DNA while nonliving things do not?
). In \Vhat way is vitalism (as a solution to the problem of life) similar to dualism (as a solution to the problem of mind)? What is meant by "the problem of life" and "the problem of mind"?
4. What is the Principle of Parsimony? \Vhy is dualism said to be less parsimonious than the iden- tity theory?
5. Does the Principle of Parsimony agree with the following claim: "Since dualism and the identity theory both predict that mental events \viii be correlated with physical events, there is no way to choose betv·1een the two theories"?
Chapter 22: The Mind/Brain Identity Theory 235
'' .. '. '.' ... ' .. '.'.'.' ... ' ' .. ' .. '.' ... ' .. '.' ...... '.' .. ''.'' ' .. ', .. '.' ''' Problems for Further Thought
1. Suppose \Ve observe a perfect correlation betvveen some mental property {like feeling pain) and some physical property (like having one's c-fibers fire). Apply the Surprise Principle (Chapter 3) to see v1hether this observation strongly favors the identity theory over dualism.
2. In the passage from Principles of Natural Philosophy quoted in this chapter, Newton defends the Principle of Parsimony by saying that "Nature does nothing in vain." Is this idea consistent with what we novv knovv about natural selection (Chapter 6)?
3. Recall the postulate that Bertrand Russell introduced to solve the problem of other minds; you considered this postulate in question 3 on page 220. \Vhat does this postulate say? How is it related to the Principle of Parsimony? Does Russell's postulate solve the problem?
4. The Principle of Parsimony is often thought to be relevant to the question of whether God exists. Formulate and evaluate an argument for atheism that makes use of this principle.
5. In a reading at the end of this chapter, the psychologist U. T. Place defends the identity theory, but does not mention the Principle of Parsimony. He does so by describing a situation in which the correlation between hvo events justifies the conclusion that the hvo are identical. Evaluate Place's argument.
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
RESEARCH AND EXPLORE
~)!<{EXPLORE
l. Mind/Brain Identity Theory 2. Physicafism 3. Principle of Parsimony
References
Isaac Newton (1643-1727). "Principles of Natural Philosophy" (1690). From U. T. Place, "In Consciousness a Brain Process?" British journal for Psychology 47: 44-50.
chapter 23 .............. ' .............. .
Functionalism
Chapter Outline
Functionalisnfs Negative Thesis: What's Wrong with the Identity Theory!
Multiple Realizability
Could a Computer Have Psychological Characteristics?
Multiple Realizability within the Class of Living Things
Functio11alis1n,s Positive Thesis
Sensations
Sununary
The Mind/Brain Identity Theory, our topic in the previous chapter, applies to the mind/ body problem the general thesis that everything is material. Minds aren't made of an im- material substance; rather, people have 1ninds and n1ental characteristics by virtue of the fact that they have brains that possess various sorts of physical strncture.
In the 1960s and 1970s, several philosophers-preeminently, Hilary Putnam (in "The Nature of Mental States" in H. Putnam, Mind, Language, and Reality, Cambridge University Press, 1975) and jerry Fodor (in Psyclwlogical Explanation, Random House, 1968)-developed a criticism of the identity theory. They espoused a point of view that ca1ne to be called "functionalisn1:' Although functionalis1n does not reject n1aterial- ism, it does reject one important part of the identity theory. Functionalism has two parts-one negative, the other positive. The negative part describes what psychologi- cal states are not. This is the functionalist critique of the identity theory. The positive part advances a proposal about what psychological states are. I'll take these two ideas in order.
Chapter 23: Functionalis1n 237
'','' '.'.''. ',,.' ''.''' ''' '. '' ' .... '.' '' '.' Functionalism's Negative Thesis: What's Wrong with the Identity Theory?
To begin 'vith, I need to dra\v a commonsense distinction. Suppose soineone says to you, looking at your clothing, "We own the same shirt:' What might this mean? There are two choices. One is that you and this person own the same kind of shirt. The other is that the very shirt on your back is jointly owned by the two of you.
Let's introduce some tern1inology to n1ark this distinction. Unique physical objects are called tokens. Kinds (or properties) are called types. The unique physical object that you are now wearing is a token of many types. It is a token of the type shirt; it is also a token of the type blue, and a token of the type clothing. To say that a single token is a token of many types is merely to say that a single object has many properties.
Conversely, a given type may have zero, one, or 1nany tokens that fall under it. Unicorn is a type of animal that has no instances; there are no tokens of that type. Golden 111011ntain is a type of geological object, one that happens to have no exemplars. You can see just from this pair of examples that two types (properties) may apply to exactly the same tokens and still be distinct types. Being a unicorn isn't the same property as being a golden mountain, even though the set of unicorns has precisely the same members as the set of golden mountains.
How does the type/token distinction apply to the mind/body problem? The identity theory has two parts. It makes a claim about psychological tokens and also a claim about psychological types. The former category-of psychological tokens-includes the follow- ing: Descartes' mind, the pain I felt in my foot last Thursday, )ones's thinking to herself today that lemons are yellow. Each of these is a token, and the identity theory says that each is identical with some physical token.
In addition, there are psychological types (properties); the identity theory makes a claim about these as well. For example, there is the property of having a mind, the prop- erty of feeling pain, the property of believing that lemons are yellow, and so on. Note that these types each have numerous tokens falling under them. Each type, therefore, describes something that various tokens have in common. The identity theory says that each of these psychological types (properties) is identical with some physical property or other.
Multiple Realizability
Functionalism rejects what the identity theory says about psychological types. Functional- ists hold that psychological properties are not identical with physical properties. Instead, functionalists argue that psychological types are multiply realizable. I'll explain this idea with an example.
Consider a type that isn't psychological-the property of being a 111011setrap. Each token mousetrap is a physical object. But think of all the different ways there are to build a mousetrap. Son1e are n1ade of \Vood and 'vire and are loaded \Vith cheese. Others are made of plastic and catch mice by injecting them with curare. Still others are made of a
238 Part Four: Philosophy of Mind
.. '.'' ........... ' ............... ,. ' ....... ' ... ' .. '.'.,'.'.'.' .. '.'.,' ... ' team of philosophers, who stalk around armed with inverted wastepaper baskets. There are many, perhaps endlessly many, ways to build a mousetrap.
Is there a single physical characteristic that all mousetraps have in common, and which non-mousetraps lack? This seems highly dubious. Each token mousetrap is a physi- cal thing, but the property of being a mousetrap (the type) doesn't seem to be a physical property at all.
How does this analogy bear on the plausibility of the Mind/Brain Identity Theory? There is more than one (physical) way to build a mousetrap. The analogy is that there may be many physical ways to build a creature that has a mind, and more than one way for a being with a mind to have this or that psychological property.
THE BIRTHDAY FALLACY, AGAIN
The Birthday Fallacy, introduced in Chapter 4, is the mistake you make if you confuse the following two statements:
Everyone has a birthday.
There is a single day on which everyone was born.
The first is true; the second false. It's a fallacy to think that the first entails the second. You'd be making the same mistake if you confused the following two statements:
Each mousetrap has physical properties that allow it to catch mice.
There is a single set of physical properties that all mousetraps have in com- mon that allows them to catch mice.
The first is true; the second false. How can the Birthday Fallacy be used to describe functionalism's criticism of the identity theory?
Could a Computer Have Psychological Characteristics?
Consider computers. It isn't very plausible to regard very simple computers as having beliefs and desires. My desk calculator no more "knows" facts about simple arithmetic than 1ny stove ('kno,vs" recipes for making soup. Rather, hutnan beings use these devices to do arithmetic or to make soup. A desk calculator, in this respect, is like paper and pencil. It facilitates our calculations, but doesn't itself engage in mental activity.
Don't be misled by the fact that we sometimes use mentalistic terms to describe sitnple co1nputers. For exan1ple, \Ve talk about computer "memory.'' Ho\vever, I don't think my home computer literally remembers anything. This is just a metaphor. If I write some thoughts on a piece of paper and those scribbles are not erased, it is true that the piece of paper has "retained" the inscriptions I placed there. But the piece of paper does not remember anything. In this respect, I think a simple computer is just like a piece of paper.
Chapter 23: Functionalism 239
' ...... ' ..... ' ' ..... ''. ' ... , '.' ... ' .. ' .. '''. '.' '.'. ' ......... '''.' '' ''' .. . Very simple machines do not have minds or n1ental states. But let's novv use our imagi-
nations and think about what the future may bring. Could there be computers that have mental states? Could computers be built that perceive-see and hear, for example? Could computers be built that remember? Could a computer reason? Could a computer be built that has wants and needs and preferences? Functionalists regard these as very real possi- bilities. Perhaps current computers can't do these things, and perhaps computer scientists at times exaggerate how close they are to making computers with these abilities. But take seriously for a moment the possibility that computers eventually will be built that have one or n1ore of these nlental abilities.
This possibility has serious implications for the identity theory. These computers, let's suppose, won't be made of protein. Perhaps they will be made of silicon chips. There seems to be no reason to expect that these thinking machines must be physically very much like the brains we have. just as there are many physical ways to build a mousetrap, so there seem to be many physical ways to have a mind and to have particular mental states.
Multiple Realizability within the Class of Living Things
We don't need to consider the future of computer science to see the point of the idea of multiple realizability. Organisms other than human beings have psychological states. Some can perceive and remember and feel pain, even if they aren't capable of all the complex thoughts and feelings that human beings can experience. Is it reasonable to suppose that these psychological states in other species must be based on precisely the same physical structures that are present in human beings?
Even within the human species, there may be variation in the physical structures used to encode this or that mental state. Indeed, even within a single human being, there may be many ways that the brain can encode a single piece of information. When you believe that lemons are yellow at two different times, maybe the brain structure that encodes your belief at the one time differs from the brain structure that encodes it at the other time.
This is functionalism's criticism of the identity theory. A given mental type-believing that lemons are yellow, wanting a drink of water, feeling pain-is not identical with any single physical type. The reason the identity theory is false is that psychological types have multiple physical realizations. This point might be formulated slightly sarcastically by say- ing that functionalism accuses the identity theory of being a "chauvinistic doctrine:' (This quip is due to Ned Block, "Troubles with Functionalism:' in W. Savage [ed.], Perception and Cog11itio11, University of Minnesota Press, 1978.) Identity theorists say that a computer or another species must be like us physically if it is to have a mind and mental characteris- tics. Functionalism regards this as implausibly restrictive. This, then, is the negative thesis that functionalism advances. Mental properties are not identical with physical proper- ties. Notice that this rejection of type-identity is consistent with accepting token-identity. Perhaps being a mousetrap isn't a physical property, but each mousetrap is a physical thing. Perhaps believing that lemons are yellow isn't a physical property, but my present state of believing that lemons are yellow might still be identical with a physical state that my brain occupies. Functionalism rejects the identity theory's thesis of type-identity, not its thesis of token-identity.
240 Part Four: Philosophy of Mind
'.' ', ••• '' '.''' '',' ••• ' '', ''' ',''', ''. '' ''. '' '' >. ' •• ' '.' '' '''.' • ''.' ' ••
The identity theory is a version of materialism. So is functionalism. The identity the- ory denies that n1inds are n1ade of an hntnaterial substance; it denies that dise1nbodied spirits exist. Functionalism agrees. The following table shows how these two theories are related to each other, and also to dualism:
Types and Tokens: How Dualism, the Identity Theory, and Functionalism are Related
Dualism Functio11alis1n Identity theory
Are all mental types identical No No Yes
with physical types?
Are all mental tokens identical No Yes Yes
with physical tokens?
Functionalisn1 is an inter1nediate position; it is conceptually "in bet\veen'' dualis1n and the identity theory.
Functionalism's Positive Thesis
So n1uch for the negative thesis. What positive account does functionalis1n provide con- cerning what the mind and its properties are like? The name functionalism suggests what this positive account is. Let's return to the property of being a mousetrap. This property, I've argued, is not identical with any single physical property. What, then, makes some- thing a mousetrap? What do the mousetraps have in common that makes them different from the non-n1ousetraps? A mousetrap is any device that functions to turn a free tnouse into a caught one. Mousetraps are things that play a particular causal role. Any device that produces certain effects in certain circu1nstances counts as a mousetrap, regardless of 'vhat physical materials it is made of or how it is physically constructed. The property of being a mousetrap is a functional property.
What does it mean, then, to claim that psychological properties are "functional" prop- erties? Let's consider a psychological state X that has the following characteristics:
When someone believes that there is water in a cup in front of her and the indi- vidual is in state X, she drinks from the cup.
When someone believes that a well contains water and she is in state X, then the individual will draw water from the well and drink it.
Each of these two statements is a conditional; it describes what an agent will do if the agent has a particular belief and is in state X. Can you guess from these two conditionals what state Xis? A reasonable conjecture is that Xis the state of wanting to drink water. Of course, there is 1nore to this desire than ,vhat these t'vo conditionals describe. Function- alism maintains that these conditionals, and others like them, describe what it is to have a particular desire. Any state that plays the right causal role will be the desire to drink water. The physical composition of the state does not matter. Psychological states are to be understood in terms of their causal relations to behavior and to other mental states. This
Chapter 23: Functionalism
example concerning state X suggests the following functionalist proposal for understand- ing what it is for an agent to have a desire:
S wants proposition P to be true if and only if, if S's beliefs are true, then Swill cause P to become true.
Notice that this proposal describes desire as a state that plays a particular causal role: When added to beliefs, desires produce actions of certain sorts. As plausible as this suggestion might be at first glance, there are problems. First, people dorit always get what they want when their beliefs are true. A pitcher may want to strike out a batter and may have true beliefs about how to achieve this, but still the pitcher may fail. The second objection is that people sometimes cause propositions to come true that they don't want to come true. When I drink from the cup, I cause my moustache to get damp, but this isn't something I want. So there is a defect in the functionalist proposal just stated for understanding what desire is. Perhaps it can be re- paired. Maybe there is something right about the idea that mental states are to be understood in terms of their causal connections 'vith behavior and 'vith other mental states.
Sensations
The kind of difficulty we have been considering is especially prominent when we shift our attention from beliefs and desires to sensations. Consider what a functionalist says about the nature of pain, for example. When people are in pain, they are inclined to say "Ouch!" They are inclined to withdraw their bodies from the stimulus they think is causing the pain. And being in pain diminishes a person's attention span and has other psychological effects as well. Suppose I exhaustively described the causes and the effects of being in pain. These would involve the relationship of being in pain to external stimuli, to behavior, and to other psycho- logical states. (Some of these I just mentioned.) It is sometimes suggested that this function- alist account of pain leaves out the most important fact about pain-the fact that pain hurts. A state may play the functional role of pain and still not be a pain state, or so this objection asserts. If this is right, then functionalism will fail as an account of the nature of pain.
I worit try to answer this question of whether functionalism is adequate. Perhaps there are some psychological phenomena that functionalism cannot adequately characterize. Maybe sensations (like pain) are counterexamples to functionalism. Even if this is true, however, there may be other psychological states for which functionalism is adequate; per- haps belief and desire can be understood along functionalist lines. This is now a contro- versial issue in philosophy.
Summary
I have argued that dualism, logical behaviorism, and the Mind/Brain Identity Theory are each inadequate. I think there is nothing wrong with mentalism-the view that beliefs and desires are inner causes of ouhvard behavior. Behavior provides evidence about 'vhat an agent's mental states are, but behavior does not define what it is to be in a given mental state. To make sense of mentalism, we need not think that an agent's mind is made of some strange immaterial substance (as dualis1n 1naintains). Individuals have 1ninds in virtue of
241
242 Part Four: Philosophy of Mind
'' '' .... '''' .. '.'' .... '.''' .. ''.'.' .. '.'''. ' ... '' ..... ' .... ''. '.'. '.'' .... the physical organization of their bodies. This, however, does not require that all the indi- viduals who possess some psychological characteristic (like feeling pain or believing that lemons are yellow) have some single physical characteristic in common. My conclusion is that functionalism's negative thesis is correct; it is plausible to hold that psychological states are multiply realizable. As for the positive thesis that functionalism advances, it re- 1nains to be seen ,vhether an adequate functionalist account of different mental states can be developed.
Positions in the Mind/Body Problem
If you read fro1n top to botton1 in the following chart, you'll sec how the various positions we have surveyed on the 1nind/body problem are related to each other:
Are tnental states inner?
Yes: Do you need to refer to inner states
to explain behavior?
Yes: Are 1ninds hnn1aterial
substances?
Yes: No:
No: LOGICAL BEHAVIORISM
No: METHODOLOGICAL
BEHAVIORISM
DUALISM Arc psychological properties identical ·with physical properties?
Yes: No: IDENTITY THEORY FUNCTIONALISM
In the rest of this part of the book, I'll explore two other philosophical problems about the mind. If human beings act as they do because of what they think and want, and if what they think and want is caused by things outside of themselves, how can human beings have free will? This problem is addressed in Chapters 24-26. After that, I'll consider the prob- lem of psychological egoism. If human beings act on the basis of what they think and want, doesn't it follow that people always act selfishly-that they try to satisfy their own wants and never care) ulti1nately) about the vvelfare of others?
Review Questions
l. I return home from a furniture store and report to my family, "They're selling our sofa." 1 look at the plate of food in front of my son and say, "That is what I ate for lunch." How does the type/ token distinction help identify an ambiguity that is present in each of these two remarks?
2. What does it mean to say that a psychological property (a type) is multiply realizable?
Chapter 23: Functionalisn1 243
. ' .. '" .. "' ''' ... ' .. '''.' .. ' ' .. ''.''' .. ' .. '' ... ' .. '.'',.' '' ... '.' .. ''. '' ''' 3. Functionalists criticize the Mind/Brain Identity Theory for being "chauvinistic." \Vhat does this
mean?
4. How is it possible to reject the identity theory \Vithout thinking that there are immaterial minds (disembodied spirits)?
5. Functionalists try to characterize \vhat it is to have a psychological property (like feeling pain or believing that Washington is the capital of the United States) by describing that property's "causal role." What does that mean?
Problems for Further Thought
l. The fact that two types apply to exactly the same tokens isn't enough to ensure that they are iden- tical. For example, being a unicorn and being a golden mountain are different types (properties), even though they apply to exactly the same objects (namely, to no objects at all). Can the same point be made with respect to types that are exemplified? Can two types be different, even though they apply to the same (nonempty) set of objects? Describe an example of this sort.
2. In Chapter 6, I discussed the concept of fitness that is used in evolutionary theory. What \'Jould it mean to say that fitness is multiply realizable? Is it plausible to claim that fitness has this characteristic?
3- Here is the inverted spectrum problem: You have a particular characteristic sensation when you look at red things and a quite different sensation when you look at green things. Is it possible that someone has precisely the reverse arrangement? This individual would have a red sensation when he looks at green things and a green sensation when he looks at red things. His behavior, including his use of language, would be precisely the same as yours. For example, he would ap- ply the term "red" to fire engines and the term "green" to grasshoppers. If spectrum inversion is possible, \Vhat consequences does this have for functionalism's positive thesis about the nature of mental states?
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
RESEARCH AND EXPLORE
!~EXPLORE
1. Functionalism 2. Artificial intelligence J. Qualia
chapter 24 ..............................
244
Freedom, Determinism, and Causality
Chapter Outline
The Problem of Freedom
Examples of Unfree Acts
Are All Behaviors Like Those Produced by Brainwashing and Compulsions?
A Clash of Plausible Conceptions
What Is Causality?
Detern1inis1n
Indeterminism
Does Indeterminism Make Us Free?
Causality Is the Issue, Not Determinism
VV.hat Does Deter1ninism Say about the Causation of Behavior?
Deter1ninis1n Differs fron1 Fatalism
Although dualists, identity theorists, and functionalists disagree about a lot, they all grant that the mind and the physical world causally interact. Your beliefs and desires cause your body to move in various ways. In addition, mental states themselves have causal anteced- ents in the physical world. What you now believe and the preferences you now have- indeed, your personality as a whole-can be traced back to experiences you have had. These experiences \Vere themselves caused by iten1s in your physical environment. In addition, n1odern science recognizes that some features of your inind may be influenced by the genetic endowment you received from your parents.
So just as the human mind has effects on the physical environment, so too does the physi- cal world affect our mental states. The following diagram represents these causal relationships:
Mind
Environment -->
Beliefs + Desires --> Behavior
Genes -->
Chapter 24: Freedom, Detern1inisn1, and Causality
The Problem of Freedom
The puzzle about the existence ofhun1an freedo1n can no"' be given a preli1ninary formu- lation. Your beliefs and desires, and hence your behavior, are caused by things outside your control. You didn't freely choose the genes you have or the sequence of environn1ents in which you grew up. If you didn't freely choose them, how can it be said that any of your behaviors is the result of free choices on your part? Ho'v can you be responsible for actions that were caused by events (long ago) over which you exercised no control? It looks like you are no 1nore free than a con1puter; a con1puter behaves as it does because it 'vas pro- grammed to do so, and so, apparently, do you.
Another way to see the problem is to attend to a different feature of the above causal picture. Suppose yom behavior is the result of your beliefs and desires just as a computer's behavior is the result of its programming. There is a feature of the computer's situation that see1ns also to characterize your o'vn. Given the cornputer's program, it can't act differently from the way it does. Suppose a computer is programmed to compute the sum of two numbers. This means that if you input 7 and 5, then the computer can't fail to output 12. The computer can't act other than the way it does.
Are we like computers in this respect? Given the beliefs and desires we presently have, isn't it inevitable that we do precisely what we do? Our beliefs and desires leave open what we will do no more than a computer's program leaves open what it will do. This fact about the computer, and about ourselves if we are like the computer, seems to entail that v,re are not free agents. The reason is that if an action is perforn1ed freely, then it must have been possible for the agent to have done otherwise. If I freely lift a cup of coffee to my lips, then it must have been possible for me to refrain from lifting the cup to my lips. But given the beliefs and desires I had, I couldn't have done any- thing other than lift the cup to my lips. Our actions are the inevitable consequences of the tninds 've have, just as a computer's output is the inevitable consequence of its progra1n.
I've just pointed out two features of the causal relationships depicted in the diagram on the preceding page. Each suggests its own argument for the thesis that no human action is free. I'll call the first argument the Distant Causation Argument; it focuses on the idea that our behaviors are caused by factors (our genes and our early childhood environment) that were beyond our control. I'll call the second argument the Could-Not-Have-Done- Otherwise Argtunent; it focuses on the clahn that it is hnpossible for us to act other than the way our beliefs and desires cause us to act. Both these arguments will be discussed later. Before reading further, 'vrite out each of these argtunents, 'vith premises and conclu- sions stated carefully, and make sure that what you come up with is an argument that is de- ductively valid. Once you do this, it should be clear that if we sometimes act freely, at least one premise in each argument must be false. Notice also that both these arguments reach a conclusion of total generality; each concludes that not one of our actions is performed freely. This goes beyond the more modest idea that some of our actions aren't performed freely.
245
Part Four: Philosophy of Mind
••• , ' '.'.''.' •••• ' ••••••••• ' ' ••• '.' ••• '' •••••••••• ' •••• ' ••• < •• ' ' '.''.' ••••
TRANSITIVITY
Some objects stand in relation to others. Alice is taller than Ben. David loves Edward. The italicized terms denote relationships. The names flanking the italicized term refer to objects that are said to be related to each other.
Some relations are transitive; others are not. If Alice is taller than Ben and Ben is taller than Cathy, then Alice is taller than Cathy. A relation R is said to be transitive precisely when it has this property:
For any three objects a, b, and c, if a bears R to b, and b bears R to c, then a must bear R to c.
Being taller than is a transitive relation. Loving is not a transitive relationship. The fact that David loves Edward and Edward loves Flo does not guarantee that David loves Flo.
Causality is a relation that obtains between events. Holmes's pulling the trigger caused Moriarty to die. Moriarty's death caused alarm to spread in the London under- world. Does it follow that Holmes's pulling the trigger caused alarm to spread in the London underworld? That is, is causality a transitive relation?
Does the Distant Causation Argument, discussed in this chapter, assume that causality is transitive?
Examples ofUnfree Acts
It is controversial whether the Distant Causation Argument and the Could-Not -Have-Done- Otherwise Argument are right in their claim that human beings never act freely. However, it is not especially controversial that some human actions are not performed freely. Here I want to give two examples. The first is a behavior produced by brainwashing. Consider the case of Patty Hearst. Heiress to the newspaper fortune, Hearst was kidnaped and abused mentally and physically by her captors for several months in 1974. She then participated with them in a bank robbery. Later, she was caught and brought to trial. There was never any doubt that Hearst helped rob the bank. The question was whether she did so of her own free will. The defense attorneys tried to establish that Hearst didn't have free will at that time; they argued that her captors had so distorted her mental faculties that she was a mere pawn in their hands. Her actions were an expression of their wants, not of hers, or so they argued. The prosecution tried to show that Hearst was a free agent; they argued that even though she had been abused, she was still a willing participant in the bank robbery. The prosecu- tion won the case, and Patty Hearst went to jail. Quite apart from whether this was the right outcome, I want to draw your attention to an idea on which both the defense and the pros- ecution agreed: People who do what they do because of brainwashing aren't acting of their own free will. Here, then, is one kind of behavior that I think we can safely regard as unfree.
The second category of unfree behavior was described by Freud. Freud describes a man who obsessively washes his hands. He is caught in the grip of a compulsion. Even after normal people would recognize that their hands are clean, the compulsive handwasher keeps scrubbing away. Sometimes the compulsion is so extreme that the person's flesh is
Chapter 24: Freedom, Determinism, and Causality 247
''.' •••• <' ••••• ''.' '.'' '.' •••••• '.' •••••••••••••• '.' •••••••••••• '. ' •• ' ••• '
eaten away, exposing the bone. "They can't help themselves;' we may want to say. Another example of this sort is kleptomania. A kleptomaniac is someone driven by the compulsion to steal. Even when they recognize that stealing is ruining their lives, kleptomaniacs find themselves powerless to change their behavior. Perhaps some people steal of their own free will; this does not seem to be true of kleptomaniacs.
Even philosophers who think that we sometimes act freely usually concede that brain- washing and compulsions rob us of our freedom. The question is whether there are other categories of behavior that are genuinely free. The Distant Causation Argument and the Could-Not-Have-Done-Otherwise Argument answer this question in the negative.
Are All Behaviors Like Those Produced by Brainwashing and Compulsions?
To decide whether we ever act freely, you need to decide whether some of our behaviors differ significantly from the unfree behaviors just described. If brainwashing robs us of freedom, what should we say about the way a normal upbringing shapes us into the kinds of people we are? Ifindoctrination robs people of their freedom, what effect does education have on the possibility of free action? If a victim of brainwashing isn't responsible for what he or she does, how can we say that "normal" individuals are responsible for what they do?
Shifting now to the examples of compulsive behavior, we can ask-if compulsive be- havior is unfree, what should we say about normal cases of rational deliberation in which people act on the basis of the wants they have? Perhaps the only difference between some- one said to have a compulsion and a "normal" person is that they have different wants. It is unusual to want to wash your hands two hundred times a day, but not so unusual to want to wash them once after they get covered with dirt. If this is the only thing that dis- tinguishes compulsive behavior from normal behavior, why say that the one is any less free than the other? To understand what freedom is, we need to answer questions such as these.
A Clash of Plausible Conceptions
The problem of freedom involves an apparent clash between two very fundamental ways we have ofconceiving ofourselves. First, there is the idea that we are part of the causal net- work. Our actions don't spring from nothing; rather, they trace back to the beliefs, desires, and other mental features we possess. These mental characteristics also have their causal antecedents, in our genes and environment. Second, there is the idea that (at least some- times) we perform actions of our own free will. Is there an unresolvable conflict between these two ideas, or can they be reconciled?
The arguments considered above concerned two different stages of the causal pro- cess represented in the figure. The Distant Causation Argument concerns the relationship of your actions to your genes and environment. The Could-Not-Have-Done-Otherwise Argument concerns the relationship of your actions to your beliefs and desires. In spite of this difference, the two arguments have something in common. Both assert that your actions are caused, and both assert that some fact about ho'v your actions are caused sho,vs that all of those actions are unfree. As noted before, if you think that some actions are per- formed freely, you must reject one or the other of these two premises.
Part Four: Philosophy of Mind
' ... ' .. '.'.' '.''''.'' ''' ''' .. '' '' ''' ''''.'' ''.'' '.' . ' ... '.''.'.'.' ''' ''' What Is Causality?
For the rest of this chapter, I want to focus on the second of these key concepts. What does it mean to say that so1ne events "cause)) others? Understanding causality has been a very deep philosophical problem at least since David Hume's work in the eighteenth century. I won't attempt to get to the bottom of this cluster of problems. I will, howeve1; try to clarify what causality is-enough so that we can better understand how causality and freedo1n are related.
What does it mean to say that striking a match caused the match to ignite? This doesn't mean that the striking was, all by itself, enough to get the match to light. After all, besides being struck, the match had to be dry and there had to be oxygen in the air. So the first fact about causality is this: A cause 11eed 11ot be a sufficient co11ditio11 for its effect. The second fact about causality is that there are usually many ways to get a particular effect to occur. Striking the match caused it to light, but that doesn't mean that striking the match was the only way to get the match to catch fire. Using a magnifying glass to focus sunlight on the match head would have done the trick. And putting the match into a heated frying pan also would have worked. So the second fact is this: Causes ofte11 are 11ot 11ecessary for their effects to occw: (See Chapter 12 to review the definitions of necessary and sufficient conditions.)
I no\v vvant to note a difference bet\veen a cause of an event and ,vhat \Ve 1nay term the whole cause of that event. Striking the match is a cause of its lighting. It is, however, only part of the whole set of causally relevant facts. Suppose I somehow were able to list all the causally relevant facts about the match at a given time. At that time, the match is struck in a certain \vay, there is oxygen present, the match is dry, and so on. Given all this infor1na- tion, ,vhat can \Ve conclude about \vhat \vill happen next?
Determinism
Can we say that the match may light? Yes. Can we say something more-namely, that the match must light, given this complete specification of all the earlier causal facts? The idea that a complete description of the causal facts guarantees what will happen next is the the- sis of determi11ism. This thesis says that if you list all the causally relevant facts pertaining to the inatch and its environment at a given time, these facts uniquely detern1ine \vhat will happen next. The match's future is not left open by its present state. Given a complete description of the matclis present state, there is only one option as to what will happen next.
Sometimes, we describe the causally relevant facts about an object so that the descrip- tion does leave open what will happen next. For example, suppose we say that a pair of dice is fair. If this is all we know, we should conclude that the probability that the dice will land double-six on the next roll is 1/36. The fact that the dice are fair and that you're now rolling them does not allow you to say what must happen. All you can say is that the dice may land double-six, but it is also true that they may fail to do so. This does not mean that the pair of dice violates the thesis of determinism. That thesis says only that a complete description of the system determines what will happen next. Perhaps causal determinism is true and my description of the dice is incomplete. This has some plausibility. For example, I didn't de- scribe exactly how the dice are rolled, what the surface is like on which they land, or what the wind conditions are that affect them as they leave your hand. Causal determinism says
Chapter 24: Freedorn, Detenninis1n, and Causality 249
''' ', ''',' '' ',''' '' ''' ''."''' '' ', ''''. ''''' '' '''''' .. ''.'. '. '' ' .. only that if all the causally relevant facts are set out, these will leave open only one possible future for the pair of dice.
Indeterminism
Is deterministn true? Before I ans,ver this question, let's be clear on 'vhat it 'vould be for determinism to be false. If detern1inis1n is false, then the 'vorld is indeterrninistic. This means that a complete description of the causal facts at one time leaves open what will happen next. In such a universe, all you can say is that the complete present state of a system makes some futures more probable than others; even a complete description of the present ,,;rill leave open n1ore than one possible future.
Until the twentieth century, determinism was a plausible and widely accepted thesis about the 'vorld. Ne,vton's lavvs of n1otion, for exa1nple, are detern1inistic in character. You may recall from high school physics that those laws do not make use of the concept of probability. For example, Newto1is law "F = ma" says that if a billiard ball of mass 111 is acted on by a force of magnitude F, the ball will have an acceleration equal to Flm. The law does not say that this will "probably" happen, as if there are other possibilities that could happen instead. Rather, the law says what must happen if the object is acted on in a certain way. This Newtonian picture about how physical objects move provides a sugges- tive model for how all events in nature ultimately will be described. Living things are more complicated than billiard balls; creatures with minds are more complicated than many living things. The idea I want to consider, however, says that all things that are made of matter are ultimately governed by the laws of physics. So, since Newtonian theory is deter- ministic, the idea is that the behaviors of living things and of things with minds must also be deterministic in character. The thought behind this generalized Newtonian thesis is that "detern1inis1n percolates up:' If elementary particles are detern1inistic, so is everything that is made of elementary particles. If a persoris mind is a material object, then his or her beliefs, desires, and subsequent behaviors are governed by detenninistic la,vs. I'm not say- ing that Newton was right about elementary particles, and I'm not saying that the mind is a physical thing made of elementary particles (recall that dualism denies this). Rather, I 'vant you to consider a certain idea: If all matter is deter1ninistic and if a perso1is n1ind is a material thing, then human behavior is physically determined.
One of the if's in what I've just described-the idea that matter is deterministic-was1it much questioned until the twentieth century. It was then that Niels Bohr, Werner Heisenberg, and Erwin Schrodinger developed the Quantum Theory. This theory is now interpreted in different ways by physicists, but the standard interpretation (which is called "the Copenha- gen interpretation;' for Bohr's hometown) is that the Quantum Theory says that the behavior of particles is not deter1ninistic. According to this interpretation of the theory, even a con1- plete description of a physical system leaves open what its future will be like. Some futures will be more probable than others, but the number of possibilities is always greater than one. In short, the present does not determine the future-chance is part of the way the world is.
It is by no means settled that the Copenhagen interpretation is the best interpretation of the Quantum Theory. Nor is it inconceivable that this very well-confirmed scientific theory will one day be replaced by another, which will say that the universe is determin- istic. But 'vhat seen1s clear no'v is that one can't silnply assiune that deter1ninis1n must
Part Four: Philosophy of Mind
''''.' .... '' .. ' ...... ''' .. ' .. '.' ... '' .. ' ' .... ' .. '''.' .. ' .. ' '.'.' ..... ' .. '' be right. Perhaps the universe is deterministic, and then again, perhaps it is not. This is a scientific question to be settled by scientific investigation. One can't decide a priori 'vhether determinism is true.
Suppose we are made of matter. Suppose that our psychological characteristics aren't due to the presence of so1ne itn1naterial substance (as Cartesian Dualisn1 asserts), but are consequences of how the matter we are made of is structured. If this is right, then I sug- gest that our behavior must be like the behaviors of physical particles. If chance influ- ences the behavior of particles, then chance also influences the behavior of people. That is, I'm advancing the thesis that indeterminism percolates up. If physical things d01it obey deterministic Ia,vs, then our beliefs and desires don't determine \vhat our actions \viii be. Rather, those beliefs and desires make some actions vastly more probable than others. Sin1ilarly our genes plus the environments \Ve inhabit do not deter1nine 'vhat our thoughts and wants will be. Again, the relationship is probabilistic, not deterministic.
Most philosophers who have written about the issue of human freedom have not worried about the implications of the Quantum Theory. Usually they have assumed that matter is deterministic and then have considered what this implies about the question of whether we are free. This is entirely understandable for people such as David Hume, who wrote about the issue of freedom long before the Quantum Theory came on the scene; Hume was writing in the heyday of the Newtonian world picture.
TWO USES OF PROBABILITY
We often use the concept of probability to describe our lack of knowledge. When we say that a coin has an equal chance of landing heads or tails on the next toss, we often mean that the coin is evenly balanced (fair) and that we don't know exactly how it will be tossed or what the wind conditions will be.
If determinism is true, then the only reason we need to talk about probabilities is that we lack a complete description of the coin's initial state. Recall that determin- ism says that a complete description of a system at one time uniquely determines what will happen next; the present can't leave open different future possibilities each of which has some chance of coming true. If determinism is true, then we need the concept probability only to describe something subjective-namely, the fact that we sometimes (or always) lack complete information.
On the other hand, if determinism is false, then probability describes an objective fact about the world. Chance is a feature of the way things happen. In this case, prob- ability talk isn't simply a way to represent our ignorance. In short, whether determin- ism is true affects how we should use the concept of probability.
Does Indeterminism Make Us Free?
If the world is indeterministic, how does that fact affect the issue of whether we are free? My suspicion is that the shift from determinism to indeterminism does not make much of a difference. The reason I say this is that if you suspect that determinism rules out
Chapter 24: Freedon1, Deterministn, and Causality
'.' ''' .... ' ',''. ''''''. ''' '.,'.' '''',.'''' .... ' ' ' ''. '' .. ' .. '.' '''' ''.'. freedom, you'll probably be inclined to think that we wouldn't be free if indeterminism 'ivere true instead.
Suppose you think we can't have free will on the grounds that our thoughts and wants are causally determined by factors outside our control. We don't freely choose the genes 'iVe have or the environ1nents \Ve inhabit in early childhood. These are thrust upon us. They shape the kinds of people we are as well as the specific beliefs and desires we have. Given all this, how can our actions be regarded as free? This is the thought expressed in the Distant Causation Argument. Maybe you find the Distant Causation Argument somewhat plausible. The question I want to ask is whether you would change your mind about freedom if chance were introduced into the above story. That is, suppose that your beliefs and desires 'ivere due to genes, environments, and chance. I suggest that if determinism robs you of freedom, chance seems to rob you of freedom as well.
The same point applies to the Could-Not-Have-Done-Otherwise Argument. Suppose you agree with this argument's thesis-that our actions can't be free, since they are inevi- table, given the beliefs and desires we have. Would introducing chance into this story make more room for freedom? I think not. If the fact that your beliefs and desires determine your action makes you unfree, I think you'd still be unfree if your actions were caused by your beliefs, desires, and chance. To make this point graphic, consider a kind of brain surgery. Suppose you are no\v a deterministic system: Your beliefs and desires determine what you will do. I now offer you a brain implant, whereby a tiny roulette wheel is intro- duced into your deliberation processes. If you lack free will when you are a deterministic system, would the operation make you free? This seems implausible. Before the operation, you were a slave to your beliefs and desires. After the surgery, you are a slave to your beliefs, desires, and to the implanted roulette wheel. The shift from being a deterministic system to being an indetern1inistic system has not n1ade a difference as far as the question of freedon1 is concerned.
Causality Is the Issue, Not Determinism
My suggestion is that determinism and indeterminism both pose a problem for human freedon1. The reason is that causality can exist in an indeter1ninistic universe just as much as it can in a deterministic one. If your n1ental states are caused by factors outside your control, how can you be a free agent? Whether these causal factors determine your behav- ior is not essential.
It isn't hard to visualize ho\v so1ne events can cause others in a deter1ninistic universe. If striking the match causes the match to light, determinism tells us that the striking completed a set of causal conditions that determines that the match must light. But how can some events cause others if determinism is false? The following example (due to Fred Dretske and Aaron Snyder, "Causal Irregularity;• Philosophy of Science, Vol. 39, 1972, pp. 69-71) shows, I think, that this is possible. Suppose a roulette wheel is con- nected to a gun, which is pointed at a cat. If I spin the roulette wheel and the ball lands on 00, this will cause the gun to fire, which then will kill the cat. Suppose that the rou- lette 'ivheel is an indeterministic syste1n-spinning the \vheel does not deter1nine 'iVhere the ball will drop. Now suppose I spin the wheel and the ball happens to fall on 00,
Part Four: Philosophy of Mind
. '.'.' '.'.'.' .. '' ...... '.' .... '' thereby killing the cat. It seems to me that my spinning the wheel caused the cat to die, even though the process is not deterministic. The death of the cat traces back to my spin- ning the wheel. For this to be true, it is not required that my spinning the wheel made it inevitable that the cat would die. This is why it is possible to have causation without determinism.
Thus, my suggestion is that the real problem is not to see whether determinism and freedom can be reconciled, but to see whether causality and freedom can be reconciled. However, since the traditional positions about freedo1n all focus on detenninistic causa- tion, I'll do the same.
What Does Determinism Say about the Causation of Behavior?
To fix ideas, let's be clear on how the thesis of determinism would apply to the diagram given earlier. This diagra1n represents the idea that your genes and environ1nent cause your present mental state, and that this, in turn, causes your behavior. What do these causal relations mean if causality is understood deterministically? Let's suppose for a ino1nent that this diagram is complete. That is, suppose it represents all the factors that influence your mental state and your subsequent behavior. What does the thesis of determinism say about this chain? It asserts the folln\ving: Given your genes and environ1nent, you couldn't have had a set of beliefs and desires different from the set that you in fact possess, and given your beliefs and desires, you couldn't have performed an action different from the action that you in fact produced. Determinism, recall, is the thesis that the facts at one titne uniquely detern1ine \Vhat co1nes next.
Determinism Differs from Fatalism
Determinis1n is an entirely different doctrine fro1n the idea tern1ed fatalisn1. You can un- derstand what fatalism says by seeing how it figures in the Greek myth about Oedipus. Oedipus was a victim of fate. The Fates decreed that he would kill his father and marry his mother. In the story, Oedipus does these things in spite of himself. The Fates somehow en- sured that Oedipus would find himself in this sorry state of affairs no matter what choices he made and no matter what he did.
Determinism says that the present determines the future; it does not say that the future is cast in stone, unaffected by \vhat is true of the present. Deter1ninisn1 is entirely con- sistent with the idea that what happens at a later time depends on what happened earlier. Determinism does not rule out the idea that if the past had been different, the present would be different. Determinism does not rule out the idea that I can affect what the future will be like by now acting one way rather than another. Fatalism denies this; it says that the future is independent of what you do in the present.
Newton said that billiard balls are deterministic systems. If you hit the ball in a certain way (call it W,), then the ball will move to a certain position (call it P,) on the table. But Newtonian physics also says that if you hit the ball in some other way (W,), the ball will move to a different position (P,). Notice that in this theory, the present makes a difference to the future. The future is controlled by the present.
Chapter 24: Freedon1, Determinis1n, and Causality 253
.. '. ' ... '. '' ........ ' .... '.' ............. ' ....... ' ..... '.' ............... . What would it be for Oedipus to be a deterministic system? This does not require us
to think that there are such things as the Fates. Oedipus would be a deterministic system if his actions were deterministically controlled by his beliefs and desires. If Oedipus has one set of beliefs and desires (call it S,), then he produces one action (call it A,). But if he had had a different set of beliefs and desires (S,), his action would have been entirely different (A,).
Here is a picture of Oedipus according to which he is a deterministic system, but fatal- is1n is false: Oedipus's beliefs and desires, plus the environment he 'vas in, ensured that he would kill his father and marry his mother. But ifhe had had a different set of beliefs and desires, he would not have done this.
It is easy to confuse determinism and fatalism, even though they are very different ideas. In fact, they are almost opposite in what they say. Fatalism says that our present beliefs and desires n1ake no difference as far as 'vhat actions lve lvill perfor1n and as far as what will happen to us. In a sense, fatalism says that our beliefs and desires are impotent; they don't have the power to make a difference. But determinism, if we model it on the Newtonian idea, says that our beliefs and desires are not impotent; they causally control what we do and thereby powerfully influence what happens to us.
According to fatalism, there is no point in t1ying to do something. If it is fated that you will get an A in a course, you will get an A whether you try to get one or not; and if it is fated that you will get a D, you will, whether you try to avoid this fate or not. On the other hand, to think that your efforts influence what happens to you-to think that trying makes a difference to the grade you will receive-is to reject fatalism.
Greek mythology talked about the Fates. I'll assume that there are no such beings. I'll assume that fatalism is false. This leaves open the separate question of whether determinism is true. As I've noted, inost previous lvriters on the free 'vill problem have assumed that cau- sality requires determinism. I hope you see why I reject this assumption; I think that causal- ity is a fact about the world we inhabit, whether or not determinism turns out to be true.
Review Questions
1. What is the Distant Causation Argument? What is the Could-Not-Have-Done-Otherwise Argument? What do these arguments have in common? How do they differ?
2. Two examples of unfree actions were given. What were they?
). \Xlhat does the thesis of determinism assert? What is indeterminism? How do these theses employ the concept of a "sufficient condition" and the concept of a "complete description"?
4. An unusual sort of brain surgery was described in this chapter. \Xlhat was it and what relevance does it have to the problem of free will?
5. What is fatalism? Is it accurately described by the saying, "que sera, sera" ("whatever will be, will be")? Hov.; does fatalism differ from determinism?
A Problem for Further Thought
Suppose the Fates decided to get Oedipus to kill his father and marry his mother by controlling what Oedipus thinks and wants. If this were so, \Vould \Vhat happens to Oedipus be an example of determinism, fatalism, both or neither?
Part Four: Philosophy of Mind
... ' ............................ ' .......... ' .. ' ....... ' ... ' ..... " .. ' ' .. ' .. MySearchlab Connections
Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
RESEARCH AND EXPLORE
~}(~EXPLORE
1. Freewill 2. Determinism
chapter 25 ..............................
A Menu of Positions on Free Will
Chapter Outline
"Compatibility" Defined
Incompatibilism and Compatibilism
Libertarianisnt
Two Soft Determinist Theories
Hunte
First Objection to l!ume's Theory: Compulsive Behavior
Second Objection to Hume's Theory: Locke's Locked Room
Does Coercion Rob Us of Free Will1
A Second Compatibilist Proposal: The Relevance of Second-Order Desires
In this chapter, I'll outline the standard philosophical positions on the relationship of freedom and determinism. I will then make some critical remarks about some of those positions. This will set the stage for the positive proposal I'll make in the next chapter about how the idea of freedom of the will should be understood.
"Compatibility" Defined
Before proceeding, I need to define a piece of terminology from logic. To say that two propo- sitions are compatible is to say that the truth of one wouldn't rule out the truth of the other. Incompatibility means conflict; if one proposition were true, the other would have to be false. Two propositions can be compatible even though neither is, in fact, true. And two proposi- tions can be incompatible even though both are false. Consider the following three statements:
(1) My shirt is green all over.
( 2) My shirt is red all over.
(3) My shirt is torn. 255
Part Four: Philosophy of Mind
Propositions (1) and (2) are incompatible; (1) and (3) are compatible. Do these facts about incompatibility and compatibility tell you what color my shirt is? Do they tell you whether the shirt is torn? The ans\ver to both questions is no.
lncompatibilism and Compatibilism
Incompatibilism is a thesis about the problem of free will. It doesn't claim that determinism is true. It also doesn't clahn that \Ve are unfree. It merely asserts an if/then state1nent: If determinism is true, then we aren't free.
If two propositions (D and F) are incompatible, there are three possibilities: D is true and Fis false; D is false and Fis true; or D and Fare both false. What is excluded by the claim of incompatibility is that both propositions are true. Of the three possible positions an inco1npatibilist can take, hvo have been pron1inent in philosophical discussion. The first holds that determinism is true and that we aren't free. This position has come to be called hard detern1inis111. The second says that \Ve are free and that our actions are not causally determined. This position is called libertarianism. These two positions agree that freedom and determinism can't both be true, but they disagree about which proposition is true and which is false.
Opposed to incompatibilism is the idea that determinism does not rule out the pos- sibility that we are free. This view, not surprisingly, is called compatibilism. In principle, there are four possible subvarieties of compatibilism. If propositions D and Fare compat- ible, then the options are (1) D and Fare both true; (2) D and Fare both false; (3) Dis true and Fis false; or (4) Dis false and Fis true. Of these four possible positions, only one has been discussed much in the philosophical literature. This view is called soft determinism; it holds that our actions are both free and causally determined. The basic idea here is that freedon1 does not require the absence of deter1ninis1n, but rather requires that our actions be caused in a particular \vay.
So the basic menu of standard positions is as follows:
I. Incompatibilism (if determinism is true, then we lack freedom). A. Hard Detern1inism: Incompatibilis1n and deter1ninism are true, so \Ve lack
freedom. B. Libertarianis1n: Incornpatibilism is true and \Ve are free, so determinis1n is false. c.
II. Compatibilism (if determinism were true, that wouldn't rule out the possibility that \Ve are free). A. Soft Determinism: Con1patibilis1n and detern1inis1n are true, and \Ve are free. B,C,D.
Note that this outline leaves some possible positions unlabeled. Be sure you understand what the unlabeled options are. My own view falls under 11.B; I'm a compatibilist. As explained in the previous chapter, I don't think one can assume that determinism is true. I do, however, think that some of our actions are performed freely. The kind of compatibil- ist position I endorse will be described in the next chapter.
The three labeled positions display three patterns of similarity and difference. Soft de- terminism and libertarianism agree that we are free; hard and soft detern1inisn1 agree that
Chapter 25: A Menu of Positions on Free \Viii 257
........ ' ......................... ' .... ' .................. ' ........... ' .. . determinism is true; and hard determinisn1 and libertarianism agree that incompatibilism is correct. To understand ho\v these three positions are related, \Ve can schetnatize each as endorsing an argument. Let F be the proposition that some of our actions are free, and let D be the proposition that determinism is true. Here is each position's characteristic argument:
Hard determinism:
Libertarianisn1:
Soft determinism:
If D, then not-F
D
NotF
If D, then not-F
F
NotD
F
D
F and Dare compatible
Note that each of these arguments is deductively valid. Our task is to figure out which of the premises are plausible.
Libertarianism
!'11 begin by considering libertarianism. The philosopher C. A. Campbell (1897-1974) is the libertarian on whom I'll focus. Before I describe how Campbell defends libertarian- ism, let me note that "libertarianism" in the problem of free will is an entirely different doctrine from the idea that goes by that name in political philosophy. Libertarians in political philosophy argue that the state should not interfere in buying and selling or in other spheres of life. This is a normative view-a claim about the way things ought to be. On the other hand, libertarianism in the problem of free will is a descriptive claim, not a normative one. It asserts that we are free agents and that determinism is false. Libertari- anism as a view about free will does not say whether it is good news or bad that we have free will.
THE NORMATIVE PROBLEM OF FREEDOM
Libertarianism, soft determinism, and hard determinism take different stances on the question of whether we, in fact, are free. They don't advance claims as to whether it is a good thing or a bad thing to be free. In other words, these theories concern descrip- tive, not normative1 issues.
Part Four: Philosophy of Mind
• , •• ' •• ' •• '' '''.'' •• ' '''' '''.''' ' ••• <'' ' •• ' '''' '' '''. '. ''''' •••• '''. ', •••
There is a separate issue about freedom that arises in political philosophy. It in- volves the question of wh·1ch freedoms people have a right to. Do people have rights to particular liberties, which may not be violated by other individuals or by the state? This is a normative question about what people should and shouldn't do. It isn't the same as the question of whether people have free will.
Freedoms, in this normative sense, can come into confiict. Some defenders of capitalism hold that people should be free to engage in buying and selling without regulation by the state. However, this can result in "boom and bust" cycles that pro- duce widespread suffering. If freedom from want is an entitlement that people have, then certain economic freedoms may have to be regulated or curtailed. This and other normative questions about the freedoms that people are entitled to will be discussed in the section of this book on ethics.
Libertarians usually think they can tell by introspection that at least some of their actions aren't determined by the beliefs, desires, and other psychological characteristics they have; when we freely decide to do something, this decision is not causally determined by earlier events. For example, Campbell, in his book Se/fhood and Godhood (Allen and Unwin, 1957), focuses on the fact that 've so1netimes perfonn actions that are <but of character:' When 've do this, he says, it isn't true that our actions are determined by the characters 've have.
I have two objections to this line of thinking. First, as noted in Chapter 13, there is no reason to place such co1nplete trust in introspection. The beliefs 've for1n about our- selves by "gazing within" can be both incomplete and inaccurate. They can be incomplete because there n1ay be facts about ourselves of \Vhich '\'e aren't conscious-facts that in- trospection fails to detect. In addition, introspection can be inaccurate, since there are psychological mechanisms that systematically distort the way we appear to ourselves. Freud stressed both of these ideas. With respect to the latter category, he argued that some of our beliefs and desires would cause us great pain if we realized that we had them. As a "defense mechanisn1;' introspection provides us \vith a false picture of 'vhat 've really think and want. Although this view about introspection is characteristically Freudian, it is important to realize that many other approaches in psychology also endorse it. There is broad consensus that we shouldn't take the testimony of introspection at face value.
There is an additional difficulty with Campbell's view. He says that when we act out of character, our actions aren't detern1ined by ,vhat our n1inds are like. I disagree. Consider a person ,vho is usually co,vardly but inanages to act courageously on a given occasion. Is it plausible to think that the courageous act can't be explained by what the person's mind was like? I find this highly dubious. I suspect that there are aspects of the person's mind that came into play. Perhaps a con1bination of unusual circumstances led the person to muster his or her courage in a way that earlier had not been possible. So I really don't see that the actions we call "acting out of character" pose a problem for determinism. Campbell assumes that the phrase "acting out of character" means that the action isn't caused by the agent's character. But this is to misunderstand what the phrase really means. When people who usually are cowardly manage to act courageously, we might say "evidently, they had it in them:' This testifies to the fact that "acting out of character" isn't something we think is uncaused.
Chapter 25: A Menu of Positions on Free \Viii
••• '.' •• ' •••• ' •• ' ' •• ''' •••• '., •• '.'. ''' '. '.'.'' '.''' •• 0 ' •• ''' •• < ••• ' •• ' ' ''
Carnpbell accepts incon1patibilis1n, he thinks that introspection sho\vs that \Ve sometin1es produce free acts, and so he concludes that determinism must be false. I've suggested that this introspective argument against determinism is very weak. If you think that free will and detern1inism are inco1npatible, I don't see ho\v the introspective impression \Ve have that \Ve are free can be decisive. The behaviorist psychologist B. F. Skinner, whose views I discussed in Chapter 21, wrote a book called Beyond Freedom and Dignity (Knopf, i971). Like Campbell, Skinner is an incotnpatibilist; but, unlike Ca1npbell, Skinner is a hard determinist. His vie\v is that the introspective picture \Ve have of ourselves as free agents is an illusion. It is a co1nfort- ing idea-a fairy tale we tell ourselves. Although I don't agree with Skinner's hard determin- ism, I think that Skinner is right not to take introspection at face value. Baron d'Holbach, another hard determinist, took the san1e vie\v of introspection in the eighteenth century; an excerpt from his book, The System of Nature, is available on-line.
If I thought that incompatibilism were true, I'd try to find out if human actions are determined. I would do this by seeing what science has to say about determinism and what psychology has to say about the causation of behavior. Campbell accepts incompatibilism, but argues in the opposite direction. He decides via introspection that some of our actions are free, and he concludes that our actions ca1it be causally detern1ined. This, I think, is the wrong order in which to address the questions. Notice that Campbell's position, as I've presented it here, doesn't argue that incompatibilism is trne. Rather, Campbell assumes that incompatibilism is true. He uses this as one of his pre1nises to argue for the conclusion that our actions aren't causally determined.
Two Soft Determinist Theories
I turn now to the standard form that compatibilist theorizing takes-soft determinism. Soft determinist theories try to constrnct a plausible theory of what freedom is, one that shows why freedom is not ruled out by the alleged fact that our actions are determined. I think the two compatibilist theories I'll now consider are defective. However, they pro- vide lessons concerning ho\\' a n1ore successful co1npatibilist theory n1ight be constructed.
The goal ofa compatibilist theory is to show that an act is performed freely ifit is caused in a particular sort of way; freedom does1it require the absence of causality, but the right sort of causality. The goal of this theory is to make room for freedom in a world of causes.
Hume
I'll begin with the compatibilist account of freedom advanced by David Hume. His idea was that an action is performed freely when the agent could have done otherwise, had the agent wanted to. Suppose you accept an employer's offer of a summer job. Hume says you act freely if you could have declined the offer had you wanted to. By the same token, when you hand the robber your wallet after the robber says "your money or your life;' you do so of your own free will, if the following were true: Had you wanted to die rather than stay alive, you could have refused to hand over your wallet. So Hume's theory is that free actions are ones that are under the causal control of the agent's beliefs and desires. When an action is under the agent's control in this way, it will be true that if the agent had had a different set of wants, the agent would have selected and performed a different action.
CEJ-[ Read Chapter 11 of
The Systen1 of Nature on w·ww.mysearchlab .com
CEJ-[ Read "Of Liberty and
Necessity," Section Vlll of An Inquiry Concerning Human Understanding on w-.•m.mysearchlab .com
260 Part Four: Philosophy of Mind
.................... ' ............. ' .. ' ...... ' ... ' ........ ' ' .... ' ...... '' .. Hume's is a compatibilist theory because it says that an action is free if it is causally related to the agent's beliefs and desires in a particular way.
What would an unfree action be like according to Hume's theory? Suppose you want to leave the room but you can't because you are handcuffed to the floor. In this case, you d01it remain in the room freely. You stay in the room whether you want to or not. What you do is not under the control of your beliefs and desires.
Here is another example of unfreedom. Suppose I perform a brain operation on you. I dis- connect your beliefs and desires from the nerves that send impulses to the rest of your body. I then implant a radio transmitter so that your body receives instructions fro1n me. No'v it is my beliefs and desires that dictate what you say and do. Your actions are no longer under the causal control of your own beliefs and desires. In this bizarre arrangement, your body would become a robot-a slave to my will. It would do what! want because I want it. After this op- eration, you do various things-you drink 'vater, deposit money into iny bank account, and so on. But you wouldn't be doing these things of your own free will. Hume's theory explains why your actions are unfree after the operation though many or all of them were free before.
First Objection to Hume's Theory: Compulsive Behavior
I think the main objection to Hume's theory is to be found in cases of compulsive behavior. Think of the kleptomaniac discussed in the previous chapter. A kleptomaniac is a thief whose desire to steal is overpoweringly strong. Kleptomaniacs want to steal even though they may be completely convinced that they will be caught and punished. Even with full knowledge that stealing will hurt them rather than help, they continue to steal. Kleptoma- niacs are caught in the grip of an obsession. They are slaves to a desire that isn't diminished by the realization that acting on the desire does them more harm than good. There are thieves who aren't kleptomaniacs, of course. When they decide whether to steal, their deci- sion is affected by information about the chance of being caught and punished. None of this information makes any difference to a kleptomaniac. The kleptomaniac is stuck; his desire isn't sensitive, in the fan1iliar \Vay, to considerations about self-interest.
I claim that when kleptomaniacs steal, they are not doing so freely. Kleptomaniacs lack freedom of will with respect to whether they steal. Yet, kleptomaniacs satisfy Hume's requirements for what it takes to be free. Kleptomaniacs want, above all, to steal things. When a kleptomaniac steals, he does so because of the desires he has. If he hadn't wanted to steal, he wouldn't have done so. The kleptomaniac's actions are under the control of his beliefs and desires. The problem is that there is something about the desires themselves and the way they work that make the kleptomaniac unfree.
The causal diagram from the previous chapter can be used to illustrate why Hume's account of freedom looks for freedom in the wrong place:
Genes --).
Beliefs + Desires Action
Environment --).
J--Hume----1
Chapter 25: A Menu of Positions on Free \Viii
Hume's theory defines freedom in terms of a relation that obtains between beliefs and desires on the one hand and action on the other. For Hume, free actions are ones that are controlled by the agent's desires. The case of compulsive behavior provides an objection to Hume's account. It is the nature of the kleptomaniac's desires that makes him unfree. This suggests that a compatibilist theory shouldn't ignore the earlier links in the above causal chain.
Second Objection to Hume's Theory: Locke's Locked Room
There is a second, more subtle, problem with Hume's account. In his Essay Concerning Human Understanding (1690), John Locke (1632-1704) describes a man who decides to ren1ain in a roo1n in order to talk 'vith a friend. Unbekno,vn to the lnan, the door of the room has been locked. The man remains in the room voluntarily, although it is false that he could have done otherwise had he chosen to do so. Locke's point is that an action can be voluntary even though the agent was not free to do otherwise. I think the man in the room stays there of his own free will. He freely chooses to stay in the room even though he is not free to exit it. If this is correct, Horne's theory is nlistaken. For you to have performed some action of your o'vn free 'vill, it is not essential that you could have performed some other action if you had wanted to do so.
Here's another example that illustrates the same point. Take any action that an agent performed of his or her own free will. For example, suppose that Joe attended a concert last night of his own free will. Now imagine that if he hadn't decided to go to the concert, he would have been kidnapped and taken there against his will. Joe didn't know about this plot when he was deciding whether he would attend. The point is that Joe attended the concert of his own free will, even though he couldn't have done otherwise.
There is an important difference between the man locked in the room (though he does1it know he is) and the kleptomaniac. The kleptomaniac's thought processes have malfunctioned; it is because of this malfunction that his stealing is not done freely. But the thought processes of the man in Locke's room have not malfunctioned. His will is free, although he isn't free to act in certain ways.
Does Coercion Rob Us of Free Will?
This contrast is important to bear in mind when you think about another important issue for the free will problem-coercion. Consider a robber who coerces you into handing over your cash by saying "your money or your life:' The robber has robbed you of your money. But has he also robbed you of your free will?
The robber presents you with an option-you can keep your money and die, or hand over your wallet and live. One option he has taken away from you is for you to keep your money and also stay alive. You are1it free to do that. But has the robber robbed you of your free will?
This is a hard problem, but let me tentatively venture this answer: In many cases (if not in all), coerced actions don't rob their victims of free will. Of course, they can rob their victims of many things that are precious. And, of course, it is often wrong to place people in situations in which they have the choices (and only the choices) that the robber offers them. My point, however, is that the robbery victinis decision to surrendering his wallet
Part Four: Philosophy of Mind
............. ' ................................ ' ...... '. ' ...... ''.'.' ..... . is very different from the kleptomaniac's "decision" to steal. The robbery victim's mind is functioning well; it is the circumstances in which the victim finds himself that are objec- tionable. In contrast, something has gone \vrong \vith the klepton1aniac1s mind, and it is this mental incapacity that makes him lack free will.
A Second Compatibilist Proposal: The Relevance of Second-Order Desires
I now want to consider a second compatibilist theory, one that is intended to meet the problem posed by compulsive behavior. The theory I'll sketch is based somewhat loosely on ideas developed by Gerald Dworkin (in ''.Acting Freely;' Nous, Vol. 4, 1970, pp. 367-83) and Harry Frankfurt (in "Freedom of the Will and the Concept of a Person;' Journal of Philosophy, Vol. 68, 1971, pp. 5-20 ). The idea is that people in the grip of a compulsion are unfree because they act on the basis of desires they would rather not have, but are power- less to expunge. Suppose you ask a kleptomaniac whether he would like to get rid of his overpowering desire to steal and he replies that he would be glad to have this monkey off his back. This kleptomaniac has two desires: (1) he wants to steal and (2) he wants to stop having the desire to steal. The theory we are considering says that the kleptomaniac lacks free will because the second desire is unable to expunge the first.
The proposal we are considering requires us to distinguish first-order from second- order desires. A second-order desire is a desire about what one's desires should be. A first- order desire is a desire about what should be true in the world outside of one's mind. "I \Vant some ice cream" is first-order; "I \Vish 1ny desire for ice cream \Vere less intense" is second-order.
The compatibilist theory we are considering says that people have free will when their second-order desires are related in a specific way to their first-order desires. According to this theory, it isn't enough that one's first- and second-order desires are in harmony. Suppose an alcoholic is so obsessed with alcohol that his addiction warps his ideas about what sort of person he wants to be. His craving for alcohol makes him want to have this craving; he ca1it even see the harm his alcoholism is doing. The lack of conflict between his first- and second- order desires does not entail that he has free will. The question is whether he would have the power to reshape his first-order desires if they were to conflict with his second-order desires. If he wanted to crave alcohol less, would this diminish the degree to which he wants to drink? Ifhe has free will, the answer must be yes, or so this compatibilist theory maintains.
Although this proposal rightly says that many people who are in the grip of compul- sions lack free will, It is still defective. just as the kleptomaniac and the alcoholic have first-order desires that are compulsions, it is possible to have second-order desires that are compulsiorts. Consider a person who is brainwashed into thinking that sexual desire is always evil. This creates a second-order desire-they want to expunge all of their sexual desires. Suppose their second-order desire is so powerful that it transforms their first- order desires; after a while, they cease to have any sexual desires at all. This does not show that they have free-will. Just as a person can lack free will when their first-order desires are "stuck:' a person earl also lack free \Vill \Vhen their second-order desires are «stuck:'
In this chapter, I have outlined and criticized two compatibilist theories of freedom. In the next chapter, I outline a third compatibilist account.
Chapter 25: A Menu of Positions on Free \Viii
..... ' .... ' ... ' ......... ' .... ' ....................... ' .... '.' .... ' ....... . Review Questions
1. What does it mean to say that X and Y are compatible? \Vhat does incompatibility mean? \Y/hat do X and Y stand for?
2. Define compatibilism, incompatibilism, hard determinism, libertarianism, and soft determinism.
3. What argument have libertarians given for thinking that some actions are free? Assess this argument.
4. Explain Hume's theory of freedom. What makes it a compatibilist theory? Why does compulsive behavior provide a counterexample to this theory? \Y/hy does Locke's example of the man Jocked in the room provide a counterexample to Hume's theory? Hume suggests necessary and suf· ficient conditions for freedom. Which of these objections shows that the conditions aren't neces- sary? Which shows that they are not sufficient?
5. Are the following propositions equivalent in meaning? (I) S did A of her own free will; (II) Swas free to not do A. How does Locke's example of the locked room bear on this question?
6. Describe and evaluate the Dworkin/Frankfurt theory of free will.
Problems for Further Thought
1. In the selection from Campbell's Seljhood and Godhood reprinted at the end of this section of the book, Campbell argues that some actions occur neither by chance nor by necessity. These are examples of creative activity, "in \Vhich { ... ]nothing determines the act save the agent's doing it." According to Campbell, these are the acts that are free. Define what "chance" and "necessity" mean, taking into account the definitions of determinism and indeterminism provided in the previous chapter. Do your definitions allow that some actions may fall into neither category?
2. How does the idea of circumstantial necessity introduced in Chapter 14 apply to the present chapter?
}. In this lecture, I argued that the Dworkin/Frankfurt theory fails to provide a sufficient condition for having free will. This leaves it open whether the theory provides a necessary condition. See if you can construct an example that shows that the condition described in the theory is not neces- sary for having free will.
Supplementary Reading
SAINT THOMAS AQUINAS
On Free Choice
In Chapter 4, you read about Aquinas's arguments for the existence of God. Here, Aquinas defends the claim that humans possess freedom of choice and explores the connection between freedom of the will and sin.
MySearchlab Connections Watch. Listen. Explore. Read. Visit MySearchLab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
[J!}-{ Read uon Free
Choice,~ Articles !-VI on www.mysearchlab .com
Part Four: Philosophy of Mind
........................... ' .... '.' ............. ' ...... ' .. ' ... ''.'.' '.' .. . CEJ-{READ AND REVIEW
READ
1. Baron d'Holbach, "The System of Nature" Chapter 11 · d'Holbach argues that human beings are thoroughly enmeshed in the deterministic processes
that govern nature and that, in consequence, free wifl is an illusion. 2. David Hume, "Of Liberty and Necessity," Section VIII of An Inquiry Concerning Human
Understanding
RESEARCH AND EXPLORE
!Jr.{ EXPLORE
1. Freewill
2. Compatibilsim
3- lncompatibilism
References
C. A. Campbell, "Has the Se/f'Free Will'?" from On Selfhood and Godhood (London Unwin Hyman Ltd., 1957), pp. 158-165, 167-179. Reproduced with permission ofTaylor & Francis Books, Ltd.
chapter 26 ..............................
Compatibilism
Chapter Outline
The Weather Vane Analogy
Function and Malfunction
What Does It Mean to Ascribe a Function to Something?
The Function of the Desire-Generating Device
Reply to the Distant Causation Argu111ent
What Does Responsibility Mean?
Moral Responsibility
Reply to the Could-Not-Have-Done- Otherwise Argument
Are Coerced Actions Unfree1
An Objection to the Weather Vane Theory: Freely Chosen, Rational Self-Sacrifice
In the previous chapter, I presented and criticized two compatibilist theories of freedom. Each seeks to show that freedom does1it require the absence of causality; both say that an action is free when it arises by a certain sort of causal process. In this chapter, I present a third compati- bilist theory, one inspired by an idea suggested by my colleagues Dennis Stampe and Martha Gibson (in their "Of One's Own Free Will;' Philosophy and Phenomenological Research, Vol. 52, 1992, pp. 529-56). After presenting a simplified version of this theory, I'll describe an objec- tion to it. I won't propose a way of meeting this objection, but will leave the problem of free 'vill unsolved. Ho,veve1; I do hope to convince you that co1npatibilism is a plausible vie,v, even though I don't claim that a completely plausible compatibilist theory has yet been articulated.
The Weather Vane Analogy
I'll begin with an analogy. Consider a weather vane. This is a device that farmers put on top of their barns to tell them which direction the wind is blowing. We talk about weather vanes being "stuck:' We also talk about \Veather vanes being "free:' What does this inean?
266 Part Four: Philosophy of Mind
''.' ........ ' .. ' '' ''''. ' .. '. '','' A free \Veather vane is one that responds to certain causal influences. If the \Vind is
blowing north, a free weather vane will point north, The behavior of a free weather vane isn't uncaused. A stuck \Veather vane, on the other hand, doesn't respond to the \Vind's direction. When a \veather vane is stuck pointing due north, there is a causal explanation for why it is pointing in that direction. The explanation, however, has nothing to do with the direction of the \vind. So the difference behveen a free \Veather vane and a stuck one isn't that the position of the for1ner is uncaused, \vhereas the position of the latter has a causal explanation. Both are caused to point in a certain direction by son1ething or other. The difference isn't in whether the behavior is caused, but in what causes it.
I've just described the ordinary meaning of "free" when it is applied to weather vanes. Notice that our commonsense understanding of this idea involves a compatibilist theory of \Vhat it is for a vveather vane to be free. The idea is to use this as an analogy for constructing a compatibilist theory of what it is for the will to be free. Of course, we must remember that weather vanes don't have minds, whereas the issue of freedom of the will has to do with a feature of the mind. Let's press on, however, and see if the analogy can be developed.
Function and Malfunction
Why do we say that the vane is free when it is causally responsive to the wind's direction? Why don't we describe a weather vane whose swivel is rusted solid as free, since in that case the direction it points to isn't influenced by the wind? A stuck weather vane, after all, is free from the wind's influences. I suggest that the answer is to be found in the fact that the weather vane has a function. When the vane performs its function, its behavior is caused by the wind's direction. When it is stuck, it can't perform its function.
How can this idea about function be applied to the human mind? If we think of differ- ent parts of the mind as having functions, then freedom of the will exists when the mind is functioning properly. A malfunction, on the other hand, may result in a "stuck,, \Vill-one that is unfree.
Consider the following diagram, which shows how beliefs and desires are formed and then lead, via deliberation, to intentions and thence to action:
Evidence --> [i@-> Beliefs } I j . ---71 Deliberation I ~ Intention ---7 Action
? ~ DGD ---7 Desires
The box labeled "BGD" is the belief-generating device. It is the part of the mind that pro- duces beliefs. Beliefs are its output. The BGD takes account of evidence (other beliefs and sensory experiences); these are its inputs. 1'he BGD 1nalfunctions \vhen it isn't sensitive to evidence-when it outputs beliefs in a way that ignores the evidence at hand. Perhaps paranoia-the irrational and unshakable conviction that the whole world is conspiring against you-can be described as a malfunction of the BGD. In this situation, the person's beliefs are "stuck:' The BGD has malfunctioned because it isn't responsive to evidence.
The DGD is the desire-generating device. It produces desires. Its inputs are the things that cause us to have the desires \Ve do. I've put a question 1nark \Vhere the inputs to the
Chapter 26: Co1npatibilism
. ''.'' ',_' ''.' '''' ' .. ' ' ' ' .. '''. '' '.''''. '.'. ''''.,.'''' ... ' .... ' ''''' .. ''' DGD go; the inputs to this device are the things the device is sensitive to when it is func- tioning properly. Later, I'll discuss what function the DGD might have.
The BGD and the DGD output beliefs and desires, which are fed into a device that carries out deliberations. This device takes into account the agent's beliefs and desires and decides on the basis of them what the agent should do. This device completes its work when the agent forms the intention to perfor1n some action.
By analogy 'vit_h the weather vane, a person has free 'vill "'Nhen the various con1ponents involved in the process described in the diagram are functioning properly. Proper func- tioning is to be understood to mean that the BGD and the DGD are sensitive to certain sorts of causal influences.
What Does It Mean to Ascribe a Function to Something?
Before we can develop this idea much further, we need to have some understanding of what the 'vord "function" means. In the case of the \Veather vane, \Ve ki10\V \vhat the function of that object is, because it was created by human beings for a certain purpose. If farmers put such objects on their barns for purely ornamental reasons, we might not regard them as n1alfunctioning \vhen their S\vivels rust solid. But farmers \Vant \Veather vanes to respond to the \Vinci's direction; this is their intended function.
If the human mind were created by an Intelligent Designer (God), and if we could find out why that Designer gave people the kinds of minds they have, we could say what functions the mind and its components subserve. Ho,vever, if 've 'vant a naturalistic, rather than a theological, account of \vhat it is for parts of the mind to have functions, \Ve n1ust set this idea about an Intelligent Designer to one side.
Consider an analogous problem. We say that the function of the heart is to pump blood. What makes this true? After all, the heart does lots of things. It takes up space in your chest cavity. It also makes noise. But we don't say that the function of the heart is to take up space or to make noise. The problem is that the heart has many effects, but only some of these effects are what the heart is for-what it has as its function. If the heart were created by an Intelligent Designer, that would solve our problem. The Designer's inten- tions would settle what the function of the heart is. But I'll assume that we want a natu- ralistic, not a theological, account of heart function. We \Vant to n1ake sense of the idea of function without having to rely on the existence of an Intelligent Designer.
My proposal is that we understand the biological function of the heart in terms of the historical explanation of why people and other organisms have hearts. The reason hearts evolved is that they provided a benefit to the organisms possessing them. The benefit was that hearts pumped blood. As a side effect of the evolution of such blood pumps, it turned out that organisms had noise makers in their chests. But the heart didlit evolve because it 1nade noise. Making noise provided no advantage for survival and reproduction.
Here, then, is a naturalistic account of what it is for an object (or a characteristic) to have a function. We doiit need to imagine that the object is present because of Intelligent Design. Natural selection (discussed in Chapter 6) is a perfectly satisfactory mechanism. My proposal is that "the function of a is to do F' means that o is present because o's do R The function of the heart is to pump blood (rather than make noise), since individuals have hearts because hearts pump blood (not because they make noise). (For elaboration
268 Part Four: Philosophy of Mind
......... ' ............. '' ............ ' ..... ' ............ '.' .. ' .. '.' .. ' ... ' of this 'vay of understanding the concept of function, see Larry Wright, "Functions;' Philosophical Revie1v, Vol. 85, 1973, pp. 70-86.)
The Function of the Desire-Generating Device
Why do individuals have desires? More specifically, why do individuals have devices (DGDs) that construct desires? What use could this faculty have had in our evolutionary past? The suggestion I want to consider is that the DGD has the function of representing what would be good for the organism. Consider the case of thirst. When an organism gets dehydrated, the organism would benefit from drinking some water. Organisms with minds seek out water when they are dehydrated because their DGD produces a desire for water. Like any device, the DGD might malfunction; if it fails to produce a desire for water when the organism is dehydrated, this could handicap the organism in its struggle to survive.
Bear in mind that there are other ways besides forming desires to get organisms to perform actions that satisfy their biological needs. Bacteria are capable of behaviors that satisfy their needs, but they don't have minds (or desires) to help them do this. The sug- gestion, then, is that the DGD has the function of representing what would be good for the organism; it isn't part of this suggestion that a DGD is the only way to get an organism to act so as to satisfy its needs.
Recall what this claim about the function of the DGD actually means. It means that we possess DGDs because, historically, such devices represented what was good for the organisms that possessed them; organisms 'vith such devices did better at surviving and reproducing than organisms that lacked them (or had DGDs that were constructed differ- ently). If this is right, we can understand why individuals who are in the grip of a compul- sion aren)t free. Their DGD is stuck; it is inalfunctioning because it no longer can represent what would be good for the agent. Kleptomaniacs want to steal even though stealing isn't good for them; indeed, kleptomaniacs sometimes want to steal even when they know that it isn't good for them to have this want. Likewise, a compulsive handwasher has an over- powering desire to keep washing his hands even though there is ample evidence available that this is a harm, not a help.
A similar explanation can be given for why brainwashing is able to rob people of their freedom. When people are brainwashed, they are given beliefs and desires on which they subsequently act. These beliefs and desires are "implanted" by short-circuiting the normal function of the BGDs and DGDs. The process of brainwashing causes the BGD and the DGD to malfunction. This is why individuals who are brainwashed to perform an action don't perform that action of their own free will.
So my proposal about freedom is this: An agent freely performs an action if the agent's mind is functioning properly. This includes the idea that the agent's DGD is able to per- form the function of representing what is good for the agent. In addition, the idea of well- functioning includes the idea that the agent can rationally deliberate, given the beliefs and desires he or she has.
Although the present proposal and Hume's idea are both compatibilist theories, they are different. Hume thought the crucial consideration was whether the agent's desires
Chapter 26: Con1patibilis1n ~6g
''''.'''.'' '' ' .. '' ''' ''.' .. ''' .. ' .... ''.'''' ', .. '' ' .. '"' '' .. ' ... '' ... ' ... control the agent's behavior. The idea I've just described says that it is the entire psycho- logical process-frorn belief and desire formation through the forn1ation of an intention to act-that settles whether the agent's action is free.
I now want to consider how the present theory of freedom allows us to reply to the arguments described in Chapter 24. Each of these arguments, recall, aimed to show that \ve can't be free because of son1e fact about ho'v our beliefs, desires, and actions are caused.
Reply to the Distant Causation Argument
I first turn to the Distant Causation Argun1ent, ,vhich I represent as follo,vs:
(1) If an agent freely performs an action, then the agent is responsible for the action.
(2) Agents are not responsible for actions that are caused by factors outside their control.
(3) Every action an agent performs is caused by factors (genes and early child- hood environments) over which the agent had no control.
No action is performed freely.
The compatibilist theory of freedom advanced in this chapter (the "Weather Vane Theory") must reject this argument. But how? The argument is deductively valid. (Identify its logical form.) Therefore, if the conclusion is false, one or more of the premises must be false.
What Does Responsibility Mean?
To reply to this argument, I need to examine the concept of responsibility. Consider what it tneans to say that a stor1n 'vas responsible for son1e crop da1nage. Here "responsible foe> sitnply means "causes:>
If we use this understanding of the concept of responsibility to clarify premise (2) of the Distant Causation Argument, we can see that premise (2) is false. A storm can cause crop damage even though there were causes of the crop damage that the storm didn't control. Consider the following causal chain:
X ----> Storm ----> Crop damage
The storm was caused by earlier events X, which existed before the storm occurred. If cau- sation is transitive (see the box on transitivity in Chapter 24), X caused the crop damage. The storm didn't exercise any influence over X, since X happened before the storm even existed. But this does not entail that the storm didn't cause the crop damage.
So the fact that the crop damage had "distant causes" doesn't mean that the storm wasn't responsible for the crop damage. By the same token, I conclude that premise (2) in the Distant Causation Argument is false.
Part Four: Philosophy of Mind
.. ''.' .. ' ............................ ' .... ' ... ',.' .. '.' '. ' ... '.' ' .. '.' ''". Moral Responsibility
It might be replied, in defense of the Distant Causation Argument, that "responsibility" is1it the simple causal concept I've just described. After all, if Patty Hearst was brain- washed by her kidnappers, we might say that she wasn't responsible for robbing the bank. This doesn't mean, however, that she played no causal role in the robbery.
If the storm wasn't responsible for the crop damage, this means that the storm didn't cause the damage. But if Patty Hearst wasn't responsible for what she did, this doesn't mean that she didn't do what she did. What, then, is the difference? We sometimes say that we don't "hold" people responsible for some of the things they do. In contrast, it sounds funny to say that we don't hold the storm responsible for what it did. Holding X responsible, and deciding not to hold X responsible, implies that we regard X as a moral agent.
When we say that we don't hold Patty Hearst responsible for what she did, I think we mean that we dotit blame her for her action. We don't think she deserves moral crit- icism for the robbery if she did it while brainwashed. If an agent is morally responsible for an event (a bank robbery, say), this means (1) that the agent caused the event and (2) that the occurrence of the event reflects on the agent's moral character. Moral responsibility differs from the simpler idea of causal responsibility; it is this latter con- cept that we apply to storms, there being no implication of moral praise and blame in such cases.
Suppose we interpret the Distant Causation Argument to be talking about moral responsibility in the sense just described. Understood in this way, premise (2) is still false. We are morally responsible for many (if not all) of the actions we perform, though the beliefs and desires that cause those actions are themselves results of factors outside our control. The reason is that much of what we do shows what sort of moral character we have.
To say that someone is morally responsible for doing something wrong does1it au- tomatically mean that he or she should be punished. Unfortunately, the law often seems to connect these two issues. If members of the jury think that Patty Hearst shouldn't be punished for robbing the bank, they must find that she wasn't responsible for what she did. There is a question here that needs to be faced head on, however, one that shouldn't be allowed to slip by. Suppose someone does something wrong, and we take this to indi- cate a defect in the persmis moral character. Maybe the rather minor lies that people tell have this feature. But surely we don't want to pass laws that say that anyone who lies will go to jail or pay a fine. The reason might be that this legal apparatus would intrude into private life so much that society would be worse rather than better as a result. We hold people responsible when they do such things, but this doesn't mean we think they should be punished.
The Distant Causation Argument tries to make it look as if agents have nothing to do with the actions they perform-as if it is people's genes and environments that cause their actions, rather than the agents themselves. But this is absurd-we cause things to happen and these actions reflect on our characters, even though there are causal explanations for why we are the way we are.
Chapter 26: Compatibilism
•••• '.'. ' ••. ' ''.' •• ' ' •• ' ••• ' ••• >'. '.' ••••••• ' ••••• '. '.' '' •• ' ••• '' ••• ,.'' '.
Reply to the Could-Not-Have-Done-Otherwise Argument
A second argument was presented in Chapter 24 that attempts to show that we lack free will. This is the Could-Not-Have-Done-Otherwise Argument, which can be represented as follows:
(1) If an agent freely performs some action A, then the agent could have done oth- erwise (that is, the agent could have performed some action other than A).
(2) People can't perform actions other than the ones they do, in fact, perform.
Agents never act freely.
Since I believe that the conclusion of this argument is false, I must reject one or both of the pre1nises.
Actually I think that both are defective. Premise (1) is mistaken for reasons that Locke brought out in his discussion of the man '"ho decides of his ui;vn free '\Vill to re1nain in a roon1 to talk to a friend; unbeknownst to him, the door was locked. This was discussed in the previ- ous chapter. Premise (2) also should be examined carefully. In Chapter 14, I formulated an idea that I called the thesis of the ambiguity of necessity. Statements about necessity and possibility often are a1nbiguous. The exa1nple I discussed there \Vas the statement "Joe can't tie his shoes:' Interpreted one way, the statement is true (Joe is holding two bags of groceries). Interpreted in another way, the statement is false (ifjoe puts down the groceries, he could tie his shoes).
I think premise (2) in the above argument is ambiguous. It could mean either of the following:
(2a) Given the very beliefs and desires the agent had, and given that the agent rationally deliberated, he could not have performed a different action from the one he did.
(2b) Even if the agent had had a different set of beliefs and desires, the agent still could not have produced an action different from the one he did.
Proposition (2a) basically says that rational deliberation is a deterministic process-beliefs and desires determine a unique choice of action, if those beliefs and desires are rationally evaluated. Since I don't think that determinism is true, I doubt that (2a) is correct. Even if it '"ere correct, ho'\vever, I don't s~e 'vhy that '\Vould sho\v that '\Ve are never free. Even if '\Veather vanes 'vere deterministic systems, that \vouldn't show that they are never free. I believe the same holds true of the human mind.
Proposition (2b) denies that the agent's beliefs and desires make a difference; (2b) says that the agent would have done the same thing even if the agent had had different beliefs and desires. I think that this statement is false-it is basically the doctrine of fatalism, which I discussed and set aside in Chapter 24.
In summary, proposition (2) is ambiguous. If it is clarified to mean proposition (2a), then it fails to show that we are unfree. If it is clarified to mean proposition (2b ), then it is false. I conclude that the Could-Not-Have-Done-Otherwise Argument fails to show that freedom is an illusion.
Part Four: Philosophy of Mind
'. ''' '"' .. ' .. '.' ''','' '''. '. ''.'' '''''''' '.'''' '''' ' .. '' '.'.'"' '. ''''''' Are Coerced Actions Unfree?
The Distant Causation Argument and the Could-Not-Have-Done-Otherwise Argument do1it refute compatibilism in general or the Weather Vane Theory in particular. I now want to consider an objection that is aimed specifically at the Weather Vane Theory. It criticizes the theory for what the theory says about coerced action.
It seems natural to say that coerced acts aren't performed freely. However, when the robber coerces you into surrendering your \Vallet, it doesn't seern to be true that your in ind has nialfunctioned. You correctly realize it is in your best interest to hand over your 'val- let, and that is \vhat you do. Action under coercion is often no different from other, less dramatic, actions produced by rational deliberation, at least as far as the inner workings of the 1nind are concerned. So if coerced actions are ahvays unfree, the Weather Vane Theory is in trouble, since the theory says that we act freely when our minds are function- ing properly.
My reply is that coercion doesn't rob an individual of his or her free will. I think that when you hand your wallet to the robber because you want to stay alive, your will is func- tioning perfectly well. Your situation here is different from that of a kleptomaniac or com- pulsive handwasher, whose DGDs have malfunctioned.
Maybe there are some cases of coercion that rob people of their free will. Suppose that the robber's threat frightens you so much that you totally lose your capacity to think clearly. You are so overcome by fear that you hand over your wallet, but not for any good reason. In this case, perhaps the coerced action is unfree. However, I think that many (perhaps most) cases of coercion aren't like this. People in such situations rarely "lose their heads"; often they show a considerable degree of care and self-control (you gingerly hand over your wallet, all the while assuring the robber that you have no intention of resisting). I conclude that a coerced action doesn't have to involve the will's being unfree.
WHAT IS COERCION?
The robber says, "Your money or your life." You would prefer to stay alive without forfeiting your money, but that option isn't available. The choice is clear, and so you hand over the money.
You are in a grocery store and want a loafof bread. The grocer will let you have the loafonly if you pay$}. You would prefer to have the loaf without handing over the money, but that option isn't available. The choice is clear, and so you hand over the money.
The robber has coerced you; the grocer has not. What is the difference? There is a moral difference: What the robber did was wrong, but the grocer did
nothing wrong. However, this doesn't seem to deflne what coercion is. You can coerce people for morally impeccable reasons. Suppose you are using a life preserver for a pillow and I need to throw the preserver to someone who is drowning. If you refuse to hand it over willingly, I may have to coerce you, but in doing so I've done nothing wrong.
There is a difference in magnitude: In the robbery example, you want to stay alive far more than you want to hold onto your wallet. In the grocer case, you would rather get the loaf than keep the$), but the difference between these outcomes is far less momentous.
Chapter 26: Con1patibilism 273
'.''. ' .. '' ... ' ' .. '.' .. ' .... ' ...... . Does this mean that coercion means putting people in situations in which it makes a great deal of difference to them what they do? Suppose I offer you $1,000 if you smile. Have I coerced you into smiling? You would much rather have the money and smile than forego the money and not smile. That doesn't make my offer coercive, however.
Some years ago, there was a program of testing new drugs in prisons. Prisoners would receive time off their sentences if they agreed to run the risk of being guinea pigs. The program was discontinued, partly because the offer was thought to be coer- cive. Do you agree that this offer would be coercive?
Let's return to the original question: What distinguishes the grocer from the rob- ber? What, in short 1 is coercion?
An Objection to the Weather Vane Theory: Freely Chosen, Rational Self-Sacrifice
Although the three objections I've just discussed don't undermine the Weather Vane Theory of Freedom, the one I'll discuss now poses a serious challenge. According to the Weather Vane Theory, the function of the DGD is to produce desires that represent what is good for the agent. I want to argue that this thesis is either too vague or is mistaken, if we consider the case of rational self-sacrifice.
Consider people who calmly and rationally decide to sacrifice their own interests to help others (perhaps their family or friends or nation). A soldier, for example, decides to throw himself on a grenade in order to save the lives of his comrades. ls it possible for a soldier to do this of his own free will?
I would say yes, but what does the Weather Vane Theory say? We have to ask whether the soldier's DGD is functioning properly here. Does the desire that it produces-the desire to sacrifice the agent's life to save the lives of his comrades-really represent what is good for the agent? Interpreted in one 'vay, the ans,ver see1ns to be no. The self-sacrifice 'vould be good for others, but not for the soldier himself. If the DGD has the function of repre- senting what is good for the agent, then it seems to have malfunctioned in this instance.
The other alternative is to interpret the account given of the function of the DGD so that the soldier's DGD is not malfunctioning in this case. We might say that the soldier's DGD is representing what is good for him to do, since he believes a moral principle that says that this sort of self-sacrifice is morally required. But now "representing what is good for the agent" is being interpreted to mean "representing what the agent thinks is good:' The problem with this way of understanding the function of the DGD is that the kleptomaniac who thinks that stealing is the best thing in the world also will be said to have a well-functioning DGD.
This suggests, I think, that the description I've been using of the DGD needs to be modified. I don't think rational self-sacrifice indicates that the individual's will is unfree. According to the Weather Vane Theory, for an action to be unfree, the individual's DGD must have malfunctioned. To be able to say this, however, we need to revise the description of what the function of the DGD is.
This is all I'll say about the problem of freedom of the will. I hope you now understand what a compatibilist theory of freedom would be like. Even though the Weather Vane Theory
274 Part Four: Philosophy of Mind
'.','. "''' '' 0'' '' ' •• ' '.'.'' ''' '' '.'' '''',.'' '. ''. '' '''.'.' '. '''" '.''' '''
isn't satisfactory as it stands, this theory, I think, suggests that there may be some merit in the compatibilist position. There does seem to be a difference between people engaged in normal rational deliberation and people who are caught in the grip of a compulsion. The Weather Vane Theory seeks to characterize what this difference is. It remains an open ques- tion in philosophy how the concept of freedom of the will should be understood.
Review Questions
1. In this chapter, I sketched a commonsense picture of what it means to say that a \veather vane is "free." \Vhat is this commonsense idea? Why say that it is compatibilist in character?
2. A \Veather vane has a function because it was designed by people to do something. \Vhat could it mean to say that the heart functions to pump blood, if the heart isn't the product of Intelligent Design?
3. \Vhat is the difference bet\veen saying that Xis causally responsible for Y (as in "The storm \Vas responsible for the crop damage") and saying that Xis morally responsible for Y?
4. What is mistaken in the Distant Causation Argument and in the Cou!d·Not-Have-Done·Otherv.1ise Argument?
5. Are coerced actions performed freely? What consequence does the answer to this question have for the Weather Vane Theory?
6. \Y/hat problem does the existence of rational self-sacrifice pose for the \Veather Vane Theory?
Problems for Further Thought
1. In Chapter 5 on the Argument from Design, I said that Aristotle's teleological conception of the universe has been rejected by modern science. People, tools, and nonhuman organisms are goal-directed systems, but it seems implausible to think of a mountain or the solar system in this way. ln the present chapter, I presented an account of what it means to ascribe a function to something. Hovv does this account help explain \11hy it is mistaken, for example, to say that the function of rain is to provide farm crops \'lith \Yater?
2. Can the Weather Vane Theory of Freedom be revised so that it avoids the counterexample pre- sented here of freely chosen, rational self-sacrifice? How might this be done?
MySearchLab Connections
Watch. Listen. Explore. Read. Visit MySearchlab for additional readings, videos, podcasts, research tools, and more. Complete your study of this chapter by completing the topic-specific and chapter exam assessments available at www.mysearchlab.com.
RESEARCH AND EXPLORE
:)\>{EXPLORE
1. Mora! responsibility 2. Compatibilism
References
Sir Charles Sherrington, The Integrative Action of the Nervous System (Cambridge i 947), pp. xx-xxi.