Introduction to Philosophy EXAM 1

TJ,e ue method of tlis- cove. is like the flijht of

roplane. It starts the 3rountl of par- r observation; it a flijht in the thin

im"!Jinative 3ener- tum; anti it "!Jain

Ian s for renewe.tl obser- n rent/ere.ti acute by nal interpretation. -ALFRED NORTH

WHITEHEAD

Section 1.3

The Laboratory of the Mind Thought Experiments

P hilosophical theories usually identify necessary or sufficient conditions for the application of a concept. Thought experiments test such theo- ries by determining whether the conditions identified are necessary or suffi- cient. Remember, if it's possible for a concept to apply without a condition being met, then that condition is not necessary for the application of the concept. Conversely, if it's possible for a condition to be met without the concepts applying, then that condition is not sufficient for the application of the concept. Thought experiments describe possible situations in which a concept should apply or a condition should be met. If it turns out that the concept doesn't apply or the condition isn't met, then there's reason to be- lieve that the theory is mistaken. To see how this works, let's put Aristotle's theory of human beings to the test.

The first step of the Socratic Method, you will recall, is to identify a prob- lem or pose a question. The question that Aristotle is trying to answer is: What makes something a human being? The second step is to propose a hy- pothesis that solves the problem or answers the question. Aristotle's hypoth- esis is that human beings are rational animals. The third step is to derive a test implication. A test implication is a conditional, or if-then, statement in- dicating what should be the case if the theory is true. To derive a test impli- cation, you have to ask yourself questions like: What if this theory were true? What does it imply? What is it committed to? After considering such ques- tions, you might come up with the following test implication: If human be- ings are rational animals, then human infants are rational animals.

The fourth step is to perform the test-examine the situation in your mind, and see whether the implication holds. If it doesn't, then the situation serves as a counterexample to the hypothesis. A counterexample is an exam- ple that runs counter to or conflicts with the theory. It suggests that the the- ory is mistaken and should be rejected or revised. Does the implication hold in this case? It wouldn't seem so. Human infants are not rational animals

4 Chapter 1 • The Philosophical Enterprise

.... ..... -cal.rs ........ :.-..... ..... ....

E:Rq: ..... --.. ...... --- .·lliins,a,

{_.her ..._ .. -.,

lt conditio ,t such theo ;sary or suffi a conditio

:ation of th without th : applicatio ms in whic out that th eason to be- Jt Aristotle'

mify a prob- :o answer is: ,ropose a hy- de's hypoth- ;; to derive a tatement in- a test impli- ry were true? g such ques- f human be- ls. :ion in your :he situation · is an exam- hat the the- ication hold ,nal animals

because they do not know how to reason. Thus human infants are a coun- terexample to Aristotle's theory. So we need to either reject Aristotle's theory or go back to step 2 and revise it. In this case, it looks like Aristotle's theory can be saved with only a minor correction. We could revise it to read that human beings are animals with the capacity to reason. This would take care of the infant counterexample because although infants can't reason, they have the capacity to reason (given time). To assess this new theory, we need co go through the process of deriving a test implication and performing a test.

Every thought experiment is part of an argument that usually has the form of denying the consequent or affirming the antecedent. In this case, the form of the argument is denying the consequent. It goes like this:

l. If human beings are rational animals, then human infants must be rational animals.

2. But human infants aren't rational animals.

3. Therefore it's not necessarily true that human beings are rational animals.

This is a deductively valid argument-if the premises are true, the conclusion must be true.

Human infants are a counterexample to Aristotle's theory because they're human beings that are not rational animals. What they show is chat being a rational animal is not a necessary condition for being a human being. To re- fute a universal generalization like all human beings are rational animals, all mu need to show is that there is at least one human being that is not a ra- tional animal. Similarly, all you need to show to refute the claim that all ravens are black is that there is at least one nonblack raven.

The most difficult part of performing a thought experiment is deriving the rest implication, because there is no formula for deriving one. Inventing a thought experiment involves a creative leap of the imagination that cannot be dictated by a set of formal rules. German philosopher Edmund Husserl called thought experiments "free fancies" because the situations involved are often produced by the free play of the imagination. But even though thought experiments can be fanciful, they are not frivolous, for as Husserl recognized, ·'riction is the source from which the knowledge of 'eternal truths' draws its sustenance.',zo To determine whether a conceptual claim is true, we have to determine whether it holds in all conceivable situations. And to determine that, we have to go beyond the actual to the possible.

Thought experiments, like physical experiments, can perform many func- tions. In addition to refuting a theory (by showing that a condition is not necessary or sufficient), they can also confirm a theory by showing that a con- dition is necessary or sufficient. By demonstrating the possibility or impossi- bility of something, they help explain the logical relations among concepts. The improved conceptual understanding they give us often aids in the con- struction of new theories. As philosophers Lewis White Beck and Robert L. Holmes note, "Thinking is a process of learning by trial and error in which the trials and error are not made in overt bodily behavior, but in imagination."21

Performing thought experiments is the essence of human thinking and the wellspring of human creativity. The better you get at evaluating and con- structing thought experiments, the better thinker you will become.

Phttosophy is the mitro- scop e of thoUJht.

-VICTOR HUGO

thought experiment The description of a possible situation in which a concept should apply or a condition should be met if the theory in question is true.

test implication A conditional or if- then statement indi- cating what should be the case if the theory is true.

counterexample An example that runs counter to or conflicts with a theory.

The Laboratory of the Mind 45

/J relijum and poh'tics I' rle's b&(s and amvic-

ns are in almost every e Jotten at second-

h nd, and without exam- i tion, from authorities

o have not themselves e, amined the 1uestions a issue but ha~e taken

em at second-hand vm other nonexaminers, hose opinions about em were not worth a 'aSS fartltinj-

-MARK TWAIN

Platonic Humans

Plato once defined human beings as "two-legged featherless animals." ls this a good hypothesis concerning the nature of human beings? Put Plato's theory to the test by using the Socratic Method.

Philosophical inquiry is not just idle, abstract speculation. Sometimes it has concrete, practical applications. It can even be a matter of life and death. To see this, let's consider a variant of the problem Aristotle was addressing: "What makes something a person?" Understanding the concept of a person will be important to solving a number of philosophical problems we will en- counter later in the text.

Case Study: Explaining How Moral Abortions Are Possible

Many people believe that, in certain circumstances, abortion is morally per- missible. But abortion seems to involve the intentional killing of an innocent human being, and such an act is usually considered murder. So those who be- lieve that abortion is morally permissible need to explain how it is possible for abortion not to be murder.

Murder is wrong because it violates our rights, specifically our right to life. But what is it about us that gives us a right to life? Why is it murder to inten- tionally kill an innocent human being but not a cow, a pig, or a chicken? What do we have that gives us our special moral status? Is it something about our physiology? Are we morally superior to these animals because we have an opposable thumb? Because we lack fur or feathers or hoofs? Because we have 46 chromosomes? This was the issue that Mary Anne Warren set out to in- vestigate in her article "On the Moral and Legal Status of Abortion."22

In ethics, a being with full moral status-and thus full moral rights-is called a person. The question is, Are all and only human beings persons? In other words, is being a biological human being a necessary and sufficient con- dition for being a person? To determine whether it is, Warren proposed the following thought experiment.

Warren's Moral Space Traveler

What characteristics entitle an entity to be considered a person? ... In search- ing for such criteria, it is useful to look beyond the set of people with whom we are acquainted, and ask how we would decide whether a totally alien being was a person or not .... Imagine a space traveler who lands on an unknown planet

Chapter 1 • The Philosophical Enterprise

' Is this a theory to

>metimes it and death. addressing: of a person we will en-

norally per- m innocent :\Se who be- possible for

ight to life. er to inten- a chicken?

thing about we have an ttSe we have :tout to in- on."22

1l rights-is persons? In

fficient con- .-oposed the

In search- whom we being was wn planet

I ! I

and encounters a race of beings utterly unlike any he has ever seen or heard of. Ifhe wants to be sure of behaving morally toward these beings, he has to some- how decide whether they are people, and hence have full moral rights, or whether they are the sort of thing which he need not feel guilty about treating as, for example, a source of food. How should he go about making this decision? ...

I suggest that the traits which are most central to the concept of person- hood, or humanity in the moral sense, are, very roughly, the following:

1.

2.

3.

4.

consciousness (of objects and events external and/or internal to the being, and in particular the capacity to feel pain);

reasoning (the developed capacity to solve new and relatively complex problems);

self-motivated activity (activity which is relatively independent of either genetic or direct external control};

the capacity to communicate, by whatever means, messages of an indefi- nite variety of types, that is, not just with an indefinite number of possible contents, but on indefinitely many possible topics;

5. the presence of self-concepts, and self-awareness, either individual or racial, or both ....

I We needn't suppose that an entity must have all of these attributes to be prop- I erly considered a person. (1) and (2) alone may well be sufficient for person- j hood, and quite probably (1 )-(3) are sufficient. Neither do we need to insist I that any one of these criteria is necessary for personhood, although once again

(:

J (1) and ( 2) look like fairly good candidates for necessary conditions, as does (3 ), if "activity" is construed so as to include the activity of reasoning. 23

If being a human were a necessary condition for being a person, it would be impossible for a nonhuman to be a person. But as Warren's thought experi- ment shows, it's not impossible for a nonhuman to be a person, for the notion of a nonhuman person doesn't involve a logical contradiction. According to Warren, what gives us our special moral status isn't the stuff out of which we are made, but rather what we can do with that stuff. So being a human being is neither a necessary nor a sufficient condition for being a person.

Remember, a logically necessary condition is one that something cannot possibly do without. So even if every person who ever has existed or ever will exist is human, it doesn't follow that being a human is a logically necessary condition for being a person. A possibility may be real even if it is never real- i:ed. To show that a condition isn't logically necessary for something, you have only to show that it's logically possible for the thing to exist without it.

Mary Anne Warren wasn't the first person to recognize that the concept of a person and the concept of a human being aren't the same. English philosopher John Locke realized this more than three hundred years ago. He writes," ... we must consider what Person stands for; which I think, is a thinking intelligent Being, that has reason and reflection, and can consider it self as it self .... "24 Locke also uses a thought experiment to demonstrate that persons need not be humans. Instead of appealing to the possibility of intelligent aliens, however,

£verythin3 that is possihle to he he!tevetl is an imtlje of the truth.

-WILLIAM BLAKE

The Laboratory of the Mind 4 7

On you have eliminated the ,possible, whatever rem ins, however im-

le, must be the

-SIR ARTHUR

CONAN DOYLE

Locke appeals to the possibility of an intelligent parrot. It seems that a certain Sir William Temple wrote in his memoirs of a parrot in Brazil that "spoke, and asked, and answered common Questions like a reasonable Creature .... "25 If there really were such a parrot, Locke argued, and if it really did possess reason and reflection, then it would be a person even though it wasn't a human being.

The notion that not all persons are human beings is one that is widely held but little recognized. Most Christians, for example, take God to be a person. But few would claim that he is a biological human being. As English philoso- pher Richard Swinburne puts it, "That God is a person, yet one without a body, seems the most elementary claim of theism."26 So the distinction be- tween persons and human beings is by no means a novel one.

From her analysis of the concept of a person, Warren draws the following conclusion about the moral status of the fetus:

All we need to claim, to demonstrate that a fetus isn't a person, is that any being which satisfies none of (1 )-(5) is certainly not a person. I consider this claim to be so obvious that I think anyone who denied it and claimed that a being which satisfied none of (1)-(5) was a person all the same, would thereby demonstrate that he had no notion at all of what a person is-perhaps because he had con- fused the concept of a person with that of genetic humanity ....

Furthermore, l think that on reflection even the antiabortionists ought to agree not only that (1)-(5) are central to the concept of personhood, but also that it is part of this concept that all and only people have full moral rights .... 2 7

The question we began with was, How is it possible for abortion not to be murder? Warren provides the following answer: It is possible for abortion not to be murder because only persons can be murdered and fetuses are not per- sons. In Warren's view, abortion doesn't violate a fetus's right to life because a fetus isn't the sort of thing that can have a right to life.

The realization that persons need not be humans and that humans need not be persons has important implications for our beliefs in other areas, as Warren notes:

Now if (1)-(5) are indeed the primary criteria of personhood, then it is clear that genetic humanity is neither necessary nor sufficient for establishing that an entity is a person. Some human beings are not people, and there may well be people who are not human beings. A man or woman whose consciousness has been perma- nently obliterated but who remains alive is a human being which is no longer a person; defective human beings, with no appreciable mental capacity, are not and presumably never will be people; and a fetus is a human being which isn't yet a person, and which therefore can't coherently be said to have full moral rights. Citi- zens of the next century should be prepared to recognize highly advanced, self- aware robots or computers, should such be developed, and intelligent inhabitants of other worlds, should such be found, as people in the fullest sense, and to respect their moral rights. But to ascribe full moral rights to an entity which is not a person is as absurd as to ascribe moral obligations and responsibilities to such an entity. 28

What makes something a person is what it can do, not what it's made of. So if a biological human being can no longer feel, think, move, communicate, or

Chapter 1 • The Philosophical Enterprise

nat a certain : "spoke, and ire .... " 25 If :Jssess reason uman being. . widely held be a person. lish philoso- 1e without a tinction be-

1e following

.at any being 1is claim to being which monstrate : had con-

ts ought to ,od, but also l rights .... 27

,n not to be 1bortion not are not per- fe because a

umans need 1er areas, as

LS clear that at an entity : people who n perma- > longer a are not and sn't yet a l rights. Citi- :ed, self- nhabitants id to respect , not a person :m entity.28

made of. So llllllicate, or

be aware of itself and its surroundings-if it's brain dead, for example-it's no longer a person. Conversely, if something can do all of those things, then it's a person even if it's made out of something besides flesh and blood. Since something's moral status is determined by these capabilities, Warren claims we should be prepared to recognize the rights of nonhuman persons, whether they come from outer space or from the labs of computer scientists .

rhoUJht Prohe

The American Society for the Prevention of Cruelty to Robots

According to the American Society for the Prevention of Cruelty to Robots, "Robots are people, too (or at least they will be someday)." This society, founded in 1999, believes that one of the greatest legal challenges of the 21st century will be the struggle for robot rights. They explain:

It is the ASPCR's position that any sentient being (artificially created or not) has certain unalienable rights endowed by its CREATION (not by its Creator), and that those rights include the right to Existence, Indepen- dence, and the Pursuit of Greater Cognition.

It is also the ASPCR's opinion that the current laws of property and capital will surely be applied in opposition to the exercising of these rights. Robots, and all Created Intelligences, will most likely go through an initial period of being considered "property" before they are recognized as fully sentient be- ings, with all attendant rights.

It is the intent of the ASPCR to raise the awareness of the general public about the issues involved with Creating Intelligences. This includes dis- cussing the moral and ethical implications of bringing Sentient Artificial Be- ings into chis world, and the responsibilities that comes with such Creation.29

Do you agree that robots can be persons and thus have moral rights? Why or why not?

How Are Thought Experiments Possible?

Thought experiments test claims about the conditions under which concepts apply or events occur. But how can such flights of fancy prove anything? Why should we trust our imaginations to reveal anything about the way things are? The answer to these questions lies in our conceptual competence. Having a concept gives us the ability to make accurate judgments about its applicability, even in imaginary situations.

We acquire a concept by being given a definition of it or by being shown examples of it. In either case, once we have a concept, we have the ability to apply it to things we have never encountered before. If we have the concept of the letter A, for example, we can apply it to typefaces we have never seen

Every JYeat advance in science has i.ssuul (nm, a new audaci-lJ of imtljt"natwn.

-)OHNDEWEY

The Laboratory of the Mind 49

I (;(JnSClfUCnus are urows of fools

an e beaums of wise

TY. th, like jO/tl, is to be ob ineJ not bJ its 3n wth, but bJ washinj a from it all that is no 30/d

-LEO TOLSTOY

before. A thought experiment is like a newly encountered typeface. Just as we can trust our judgment to determine whether the concept of the letter A ap- plies to a letter in a new typeface, so we can trust our judgment to determine whether a particular concept applies to the situation described in a thought experiment.

Of course, the more flourishes the letters in a typeface have, the more dif- ficult it will be to determine whether a letter is an A. Similarly, the more out- landish the thought experiment, the more difficult it will be to determine whether the concept in question applies. So not all thought experiments are equally persuasive. Some are more convincing than others.

To have a concept is to be able to apply it correctly. But we may be able to apply a concept without being able to state the criteria we use in applying it. For example, we may be able to identify a grammatical sentence without being able to state the rules of grammar. In such a case, we have an intui- tive understanding of grammar even though we do not have a theoretical un- derstanding of it. In attempting to identify the conditions for applying a concept, we are trying to transform our intuitive understanding into a theoret- ical one. That is, we are trying to make explicit what is implicit in our under- standing of a concept. Because having the ability to apply a concept correctly doesn't necessarily give us the ability to state the conditions for applying it, dif- ferent people may have different theories about what those conditions are. But because we have an intuitive understanding of the concept, there is a body of data--our "intuitions"-that can be used to adjudicate vario\ls theories of it.

Conceptual intuitions are not the only data that philosophical theories must take into account, however. As we've seen, philosophical problems arise when our intuitions seem to conflict with other beliefs we have. Often these beliefs come from science. Trying to square our philosophical beliefs with our scientific ones has been a major concern of philosophy since its inception. The goal is to arrive at a view of the world that makes sense of it. As American philosopher Wilfred Sellars puts it, "The aim of philosophy, abstractly formu- lated, is to understand how things in the broadest possible sense of the term hang together in the broadest possible sense of the term."30 To accomplish this goal, philosophy can't afford to leave anything out of account.

Criticizing Thought Experiments

The value of any experiment is determined by the amount of control with which it is executed. The more controlled the experiment, the less chance that its results will be misleading. It is not possible to control all the variables in an experiment, however. No one, for example, can control the position of Earth relative to the sun and the other planets. Nevertheless, it is sometimes possible to control all the relevant variables-that is, all the variables that could reason- ably be expected to affect the outcome of the experiment. Criticizing an exper- iment usually involves explaining how it's possible that something other than the variable under investigation could have produced the result.

Some thought experiments describe situations that are physically impossi- ble. That is not necessarily a strike against them, however, for their more

Chapter 1 • The Philosophical Enterprise

. Just as we :tter A ap- determine a thought

~ more dif- . more out- determine iments are

· be able to 1pplying it. hout being

an intui- )retical un- applying a ) a theoret- our under-

;,t correctly lying it, dif- ms are. But is a body of eories of it. :al theories Hems arise )ften these !fs with our . inception. s American ictly formu- of the term accomplish

omrol with chance that iables in an :on of Earth mes possible )Uld reason- -ig an exper- : other than

illy impossi- their more

fantastic aspects may not be relevant to their outcome. Thought experiments examine the logical relations between concepts, and abstracting from physi- cal reality is sometimes necessary to throw those relations into proper relief. Of course, the more outlandish a thought experiment, the more likely it is to alter a variable that is relevant to its outcome. If you doubt the results of an experiment, however, the burden of proof is on you to show where it went wrong by providing an alternative explanation of the results .

There is usually widespread agreement about the outcome of a thought ex- periment.31 Thus thought experiments serve as an objective check on philo- sophical theorizing. When there is disagreement, it usually focuses on the interpretation of the results rather than on the results themselves. In the case of Warren's moral space traveler, for example, there is widespread agreement that persons need not be human beings, and vice versa. There is much less agree- ment, however, about what implications this has for the abortion controversy.

Even if fetuses aren't persons, many claim that fetuses are nonetheless valuable forms of life and thus should be destroyed only if there are good rea- sons for doing so. For example, Daniel Callahan, co-founder of the Hastings Center, an institute devoted to studying biomedical ethical issues, claims, "[Abortion] is not the destruction of a human person-for at no stage of its development does the conceptus fulfill the definition of a person, which implies a developed capacity for reasoning, willing, desiring and relating to others-but it is the destruction of an important and valuable form of human life."32 As a result, Callahan maintains, taking such a life "demands of oneself serious reasons for doing so."33 Just what those reasons are, he doesn't say. Nevertheless, it's clear that Callahan doesn't believe that the nonpersonhood of the fetus justifies abortion on demand. So Warren's moral space traveler thought experiment has not settled the abortion controversy. By clarifying the concept of a person, however, it has raised the level of discussion .

Even if the situation envisioned in a thought experiment is well defined, we may still reject the results of the thought experiment on the grounds that its assumptions are unreasonable. No theory-whether about concepts or physical objects-can be tested in isolation. Theories of any sort have test- able consequences only in the context of certain background assumptions. Assumptions about the nature of human cognition and the nature of the ex- ternal world, for example, lie behind every experiment. Thus if an experi- ment yields an incredible result, the problem may lie with the background assumptions rather than with the theory being tested.

Conceivability and Possibility

To show that a condition is not necessary for the application of a concept, one needs to show only that it's possible for the concept to apply without the condition being met. The best evidence that a situation is possible is that it's conceivable-that is, coherently imaginable. A situation is coherently imaginable when its details can be filled in and its implications drawn out without running into a contradiction. If, on examination, a situation is found to harbor an inconsistency, then it is not conceivable.

t;ool/ reasons must of force, Jive !'lace tu hdter.

-WILLIAM

SHAKESPEARE

/11,uqi11atio11 ru/e5 the world

-NAPOLEON

coherently imaginable A situation is coher- ently imaginable when its details can be filled in and its impli- cations drawn out without running into a contradiction.

The Laboratory of the Mind 51

Nc inJ is a.s far aw'9' a.s one inute aJO,

-JIM BISHOP

(), reason mu.st be con- s· rel as a kin/ of cau.se, of nidt truth is the nat-

e!foct. -DAVID HUME

Consider, for example, time travel. At first glance, traveling backward in time seems perfectly conceivable. It may be technically impossible to build a time machine, but the many science-fiction stories that make use of this notion seem to suggest that it's at least logically possible. This suggestion is mistaken, however, because an event that has already happened cannot also not have happened. Suppose you travel back in time to a town at the turn of the century whose population was exactly 10,000 on January 1, 2000. After you arrive, the town will then have a population of 10,001. But it is logically impossible for a town to have a population of both 10,000 and 10,001 on January 1, 2000. So, appearances to the contrary, traveling backward in time to the same universe is not coherently imaginable because when we fill in the details and examine the consequences, we arrive at a contradiction.

If our time machine takes us back to a different universe, however, the contradiction can be avoided. Science writer Martin Gardner explains. "The basic idea is as simple as it is fantastic. Persons can travel to any point in the future of their universe, with no complications, but the moment they enter the past, the universe splits into two parallel worlds, each with its own time track. Along one track rolls the world as if no looping had occurred. Along the other track spins the newly created universe, its history perma- nently altered."34 If the universe splits when you travel backward in time, there will be no contradiction because in neither universe will something both be and not be the case.

What the time travel example shows is that apparent conceivability doesn't guarantee possibility. From the fact that a situation seems coherently imagi- nable, it doesn't follow that it is, for it may contain a hidden contradiction. Apparent conceivability does provide good evidence for possibility, however, because if, after careful reflection, we haven't found a contradiction in a situ- ation, we're justified in believing that it's possible.

Our conceptual ability can be compared to our perceptual ability. We can seem to perceive something that isn't real, but we can't actually per- ceive something that isn't real. For example, we might seem to perceive a cat in the yard while in actuality it's an old shoe. In that case, we never perceived a cat; we just thought we did. Similarly, we can seem to con- ceive something that's impossible, but we can't actually conceive some- thing that's impossible. To distinguish apparent from real perception, we often gather more perceptual data; we look more closely at the situation or perform additional physical experiments. Similarly, to distinguish apparent from actual conception, we often gather more conceptual data; we look more closely at the logical implications of the situation or perform addi- tional thought experiments. If we doubt the results of a physical experi- ment, we can check them by means of another physical experiment. Similarly, if we doubt the results of a thought experiment, we can check them by means of another thought experiment.

Because our conceptual scheme is an interconnected web of beliefs, every philosophical problem has a bearing on every other. Whatever solution is pro- posed to one problem must be judged in terms of the sorts of solutions it sug- gests to others. Deciding among various solutions to philosophical problems, then, requires appealing to considerations of scope, simplicity, conservatism,

Chapter 1 • The Philosophical Enterprise

backward in ible to build e use of this suggestion is cannot also

.t the tum of 2000. After

it is logically d 10,001 on .,.·ard in time we fill in the on. 1owever, the ler explains. to any point aoment they with its own ad occurred. story perma- 'ard in time, 11 something

bility doesn't :ently imagi- )ntradiction. ity, however, ion in a situ-

ability. We actually per- o perceive a ;e, we never :em to con- ,ceive some- rception, we situation or

ish apparent na; we look erform addi- sical experi- experiment.

e can check

beliefs, every lution is pro- lltions it sug- :al problems, :onservatism,

and fruitfulness. The theory that does best with regard to the criteria of ade- quacy will produce the most understanding.

Thought experiments are just one tool among many that philosophers use to evaluate their theories. But they are an important tool, for not only can they strengthen or weaken existing theories, they can also generate data that any future theory must take into account. Theories at the forefront of philo- sophical research are generally superior to their predecessors because the thought experiments of the past have broadened the evidence base on which future theories must rest.

Scientific Thought Experiments

Thought experiments aren't unique to philosophy. They can also be found in the sciences, where they have helped produce a number of scientific ad- ,·ances. Their use in the sciences is instructive.

One of the hallmarks of a good theory is that it is free from contradiction. Any theory that implies that something both is and is not the case is unaccept- able, for n9t only is it uninformative, it cannot possibly be true. Thought exper- iments are particularly useful in testing for contradictions. Galileo used a thought experiment to demonstrate that Aristotle's theory of motion was self- contradictory, and thereby paved the way for the modem science of mechanics.

Aristotle held that heavier bodies fall faster than lighter ones. Galileo, on the other hand, maintained that all bodies, regardless of their weight, fall at the same rate. To show that his view was superior to Aristotle's, Galileo pro- posed the following thought experiment.

Impossibility of Aristotle's Theory of Motion

i!

1 ~ i;)

Imagine that a heavy cannonball is attached to a light musket ball by means of a rope. Now imagine that both this combined system and an ordinary

Truth is what stantls the fest of CXf&Yience.

-ALBERT EINSTEIN

The Laboratory of the Mind 53

54

cannonball are dropped from a height at the same time. What should happen? According to Aristotle, because lighter objects fall more slowly than heavier ones, the musket ball attached to the cannonball should act as a drag on it. So the combined system should fall more slowly than the cannonball alone. But because the combined system is heavier than the cannonball alone and because heavier objects fall faster than lighter ones, the combined system should also fall faster than the cannonball alone. But it is logically impossible for one object to fall both faster and more slowly than another. So Aristotle's theory cannot be correct. Galileo's theory, however, avoids the contradiction by main- taining that all bodies fall at the same rate. It follows, then, that Galileo's view is more credible than Aristotle's.

By showing that Aristotle's theory harbored an inconsistency, Galileo made the modem science of mechanics possible. The value of thought exper- iments, then, lies not only in their immediate results but also in their long- term consequences.

Summary

Philosophical theories explain how it is possible or why it is impossible for a concept to apply by identifying the conditions for applying it. Thought experiments test these theories by determining whether they hold in all pos- sible situations. If they do not-that is, if there are counterexamples to the theory-there is reason to believe that the theory is mistaken.

Like scientific experiments, thought experiments can go wrong and can be criticized for it. If they are not sufficiently spelled out or if they rest on unrea- sonable assumptions, their value is questionable. If you believe that a thought experiment is problematic, however, the burden of proof is on you to provide an alternative explanation of the results.

The adequacy of a theory is determined by how much understanding it produces, and the amount of understanding produced by a theory is deter- mined by how well it systematizes and unifies our knowledge. Criteria such as conservatism, scope, fruitfulness, and simplicity can be used to gauge the ad- equacy of a theory.

Thought experiments not only help us evaluate theories but also generate data that any future theory must take into account. Theories at the cutting edge of philosophical research are usually superior to their predecessors be- cause previous thought experiments have added important considerations that any future theory must incorporate.

Study Questions

1. What is a thought experiment?

2. How are thought experiments possible?

3. On what grounds can thought experiments be criticized?

Chapter 1 • The Philosophical Enterprise