Short Answers
Week 9A: Persons and Reality
Susan Schneider
• Professor of Philosophy and Cognitive Science, University of Connecticut
• Editor of our Textbook
• First philosopher we are reading who I have actually met!
Review
• Transhumanism = enhancement of human by artificial devices
• Schneider raises two scenarios: • Mindscan
• Mental properties are copied onto a computer
• Augmentation • Brain is replaced bit-by-bit by cybernetic components
• Both scenarios raise questions: • Can we be psychologically continuous with a computer? • Also: if our minds are our brains, can our minds then become computers?
• Both questions require a substantive account of what would be the same. Enter: computation!
Computation (and Representation)
• What is it for thought/mental stuff to be computational?
• Computations are transitions from an input and a starting state to an output and an end state, with these transitions mediated by representations.
• That’s a lot! Let’s take it one step at a time
Input/Output Relations
• A device that produces input/output relations is a device that receives a signal (or object) and produces another, different signal (or object).
• Example: light switch.
• Input = Press button up, Output = Turn light on.
• Input = Press button down, Output = Turn light off.
State Transitions
• States describe the way something is at a time.
• Being hungry is a state of you (or your body). • Knowing that you are hungry is a state of your mind.
• Being full is another state of you (or your body).
• When you go from being hungry to being full, you have undergone a state transition.
Input/Starting State to Output/End State
• This one is probably easier to explain by example.
• Picture a coke machine.
• A coke machine has input-output relations: you put in a dollar, you get a coke.
• But it also has states.
• Each time you get a coke, it goes from having that many cokes to one less coke, e.g. 10 to 9 cokes.
• So a better way to describe it is: input a dollar and start at 10 cokes, go to output a coke and end with nine cokes.
• It can be described [I + SS] → [O + ES]
• [Dollar + n cokes] → [Spit out a coke + (n-1 cokes)]
• But what if starting state is zero cokes?
• Then it doesn’t output a coke!
• [Dollar + 0 cokes] → [Return dollar + stay at 0 cokes].
Representations
• A representation is something that stands for, or is about, something else.
• Words are paradigmatic examples of things that represent.
• The word ‘dog’ represents this:
Transitions mediated by representations
• Old-school coke machines (like the one in the picture) didn’t represent how many cokes were there. • You gave it money, in the form of either coins or a dollar bill
• It determined if you had enough money by having a weight get to a certain point (for coins) or having the dollar bill roll over a roller without jamming.
• (The old ones couldn’t tell the difference between a dollar and a $20 bill, they just relied on the commonsense notion that no one would intentionally put a $20 bill in a machine to buy something that cost a dollar).
• The weight or roller triggered a mechanical action that dispensed a coke. • It “knew” whether or not it had cokes to dispense by another weight under the stack of cokes.
• But a new soda machine, where cokes cost $2.50, might scan and create an image of a bill to see if it is a $1 or $5 bill. • The image would be checked for a match in a database, and a match would then send a
signal to dispense a coke. • The words in red all indicate representational functions: scanning, checking, a database, an
image, a signal.
• The new coke machine, but not the old one, is performing computations.
Is Thought Computational?
• It sure seems like it!
• Consider • Starting state = +<DES, SANDWICH> • Input <BEL, ‘There is a sandwich in the fridge’> • Output <ACT, Go to fridge and eat sandwich> • End State -<DES, SANDWICH>
• Transitions are right for computation.
• All of these items plausibly represent: • Desires have objects (SANDWICH) • Beliefs have statements that the person thinks are true • Actions are represented in motor systems before being performed.
Computation and Physicalism
• Possibly (hopefully) during this presentation, you had a thought: • But isn’t this just what Armstrong means by ‘causal roles’?
• Yes and no.
• A system with its components related by causal roles can be a computational/representational system.
• But the basic material that makes up the states with those causal roles is like the hardware.
• And the computational/representational theory cares only about the software. • So, a “hard-core” causal-role theorist might say that thought is computational, but it has to be
implemented by a brain (the hardware matters).
• Armstrong himself is neutral between this position and Schneider:
• His view is that the human mind is in fact implemented by a brain but he is silent on whether it must be implemented by a brain.
Computation and Psychological Continuity
• At the beginning of this unit, I made a remark I promised to explain later. • I said: if you are a mind-brain identity theorist, then you will think that brain continuity is sufficient for
psychological continuity. • But it is a further question whether it is necessary for psychological continuity. • I.e. the brain continuity view, assuming identity theory, isn’t automatically the same as a psychological/brain
continuity hybrid view.
• I can now explain what I meant! • If you are a “hard-core” identity theorist, and think the organic nature of the hardware matters, then brain
continuity is close to necessary for psychological continuity (if the mind is transferred, it must somehow be to another brain).
• But if you are a software-only person like Schneider, or think the software can be run on multiple hardwares, then even if identity is true, you can have psychological continuity without brain continuity.
• Because even though the mind in fact is the brain, it could in the future, be some other physical object.
• The computational theory of thought is also very helpful for psychological continuity in another respect. • It gives a much more substantial and sophisticated view of what psychological continuity amounts to than
‘causal dependence’ which is a very vague notion, and can be interpreted in many different ways. • E.g. my computer’s overheating might cause it to break, its broken state is causally dependent on its prior functioning state.
But that couldn’t be psychological continuity.
Hillary Putnam
• American Philosopher, Lived 1926-2016
• Professor of Philosophy at Harvard (also Northwestern, Princeton, and MIT)
• Close friend of Noam Chomsky
• Refuted Ryle’s Behaviorism
• Created ‘Functionalism,’ which is progenitor to Computational Theory of Mind as well as Armstrong’s Causal Theory of Mind
• Advised Dozens of other prominent philosophers (including my M.A. adviser and Susan Schneider’s Ph.D. adviser)
Putnam’s Brain in the Vat
• Brain is not in a body, BUT
• According to CTM, parts of brain receive signals from (output of) sensory organs. • For instance, the visual cortex receives signals from the eyes.
• These signals are input to the brain.
• Putnam’s Brain in the vat receives precise duplicates of these signals from artificial sources.
• Putnam’s BIV is different from Dennett’s, as Dennett’s either: • Received input from an actual body/sense organs, just over a distance.
• Or received no input at all.
Brains and their Inputs
• Normal: Dennett v.1:
• Dennett v:2 Putnam: