LAST LECTURE - PowerPoint PPT Presentation

About This Presentation
Title:

LAST LECTURE

Description:

If we say that a mental state is a functional kind, then we say that to be 'in ... What the mental states have in common is the functional profile of relating ... – PowerPoint PPT presentation

Number of Views:21
Avg rating:3.0/5.0
Slides: 29
Provided by: rwb9
Learn more at: http://home.olemiss.edu
Category:
Tags: last | lecture | mental | state

less

Transcript and Presenter's Notes

Title: LAST LECTURE


1
LAST LECTURE
2
Functionalism
  • Functionalism in philosophy of mind is the view
    that mental states should be identified with and
    differentiated in terms of functional roles.
  • A function (as in mathematics) relates inputs to
    outputs. (e.g. multiplication vs. addition)
  • A functional role would determine a specific kind
    of function.

3
Functional Concepts (1)
  • Part of FUNCTIONALISM is conceptual.
  • In a stuff concept, what a thing is and what it
    does is a matter of the nature of the stuff. To
    be water is to be liquid and wet. To be wood is
    to be, to be a sheep is to be .
  • Most STUFF concepts are the concepts of certain
    kinds in nature (or natural kinds).

4
Functional Concepts (2)
  • A functional concept identifies what a thing IS
    with what it does.
  • To be a knife, is to be a thing that cuts
  • To be money, is to be the kind of thing that
    permits economic exchange
  • To be an umpire, is to be the person who
    adjudicates events in a baseball game.
  • To be a chess queen, is to be a piece that moves
    and captures in a specified manner.

5
Functional Concepts (3)
  • If we say that a mental state is a functional
    kind, then we say that to be in pain is to be
    in that state that relates injury to behaviors of
    type T.
  • What the mental states have in common is the
    functional profile of relating input to output.
    If two functional states relate the same inputs
    to the same outputs then they are the same
    function.

6
Examples
  • Two word processing programs will have different
    computer code, but will perform the same work.
    Same function different realization.
  • A Calculator and an abacus will both perform
    simple mathematical calculations, but one is
    digital and one is analog.
  • A digital clock and a wind-up cuckoo clock both
    keep time, etc.
  • Same function different form

7
Function and Multiple Realization
  • One of the primary reasons people entertain
    functionalism about mental states is that
    functional states are multiply realizable.
  • Recall MR was a problem for identity theory.
    Now it is a virtue of Functionalism.

8
Functionalism Review
  • 1) Functionalism in philosophy of mind is the
    view that mental states should be identified with
    and differentiated in terms of functional roles.
  • 2) A function (as in mathematics) relates inputs
    to outputs. (e.g. multiplication vs. addition).
    A functional role would determine a specific kind
    of function.
  • 3) Functional Concepts define things relative to
    functions
  • 4) Functions are multiply realizable.
  • According to Functionalism
  • To be in a mental state S is to have your brain
    realize a functional state F.

9
Two models
Observable
Not observable
  • Behavioral Model
  • Mind Plays a role(?) model

Stimuli
WHY ISNT FUNCTIONALISM JUST A VERSION OF
BEHAVIORISM
Stimuli
ConditionedResponses
MENTALEVENTS
Determines Type
Behavior
Behavior
10
More Models
Observable
Not observable
  • Functional Model
  • Mind Plays a role(?) model

Stimuli
Stimuli
Because the functional role determines what the
state is, not the input/output pair
Functional State
MENTALEVENTS
Behavior
that means, e.g., pain and fake-pain are
different states
Behavior
11
Hardware Software Analogy
  • According to one version of Functionalism
  • The brain is like a computer and your mental
    states are like programs that run on that
    computer. The same computer can run different
    programs at different times and the same (or
    similar) programs can be run on different
    computers.

12
Objection to Functionalism
  • Blocks Liberalism Objection
  • If having a functional organization F is
    sufficient for being in mental state M, then
    anything which realizes F will be M.
  • So if F1 is being in love with Laura Bush, and
    a crashing wave contains water molecules that
    realize F1, then the wave loves Laura Bush.

13
See the unhappy rainbow?
14
The Qualia Objection
  • Similar to the Conscious experience argument for
    dualism.
  • 1) Conscious experience contains more information
    than the physical facts provide What its like
    to X or Qualia
  • 2) Qualia has no functional role.
  • 3) If functionalism is true there is no
    requirement that functional states have
    associated qualia. Thus mental states need not
    have qualia. (Qualia Zombies)
  • 4) But the presence of qualia is paradigmatic of
    consciousness. And consciousness is a mental
    state.-------------------------------------------
    ----------------------
  • 5) Therefore, functionalism is not a complete
    account of mental states.

15
Functionalism and Artificial Intelligence (AI)
  • Non-Biological Realization
  • If a functional state can be realized outside of
    a biological system, then can we build a system
    (e.g. in a computer) that has REAL mental states?
  • Such a system would be a form of Artificial
    Intelligence. (AI)

16
Types of AI
  • There are two conceptions of AI
  • Strong AI non-biological systems can realize
    real mental states.
  • Weak AI non-biological systems can only
    simulate mental states. We can learn about the
    mind by studying these systems, but they are not
    real minds.

17
Is Simulation just Simulation?
  • Simulated Weather
  • Simulated Sunlight
  • Simulated Flight
  • Simulated Clocks
  • What kind of simulation is an AI system?

18
Can Machines (really) Think The Turing Test
  • In a famous 1950 paper Philosopher-Mathematician
    Alan Turing asked
  • Can a machine think?
  • To answer the question he proposed an experiement
    called the Imitation Game or (later) Turing Test

19
Turing and Turing Machines
ALAN M. TURING He developed a way to resolve
complex mathematical problems using imaginary
machines (digital computers)
One kind of Turing Machine A Turing Bombe
code-breaker
20
A Basic Turing Machine Diagram
INPUT
OUTPUT
PROCESSOR
One way to make a TM is to write a program that
wins the Turing Test
MEMORY
21
Turing Machines are MR
A Turing machine can be diagrammed or built of
metal or of blueberry muffins and
frosting. AND STILL BE A Turing Machine
22
The Imitation Game
A woman in a chat-room answers questions as she
normally would
A Man in a chat-room pretends to be a woman
You may ask any question you want for 15 minutes
via computer. If you cannot tell who the real
woman is, then the man wins. He is
indistinguishable from the woman.
23
The Turing Test A computer tries to imitate a
human
COMPUTER
WHICH IS WHICH? IF YOU CANT TELL, IS THERE A
DIFFERENCE?
24
Try the Turing Test on your friends
  • How do you know that they can think?
  • Imagine that you could only talk to them in a
    chat room or in some other setting where behavior
    was not a factor.
  • You would ask questions and judge on the basis of
    answers.

25
Objections to Turings Claim
  • Turing doesnt answer his own question. He gives
    a test, and suggests that computers will be able
    to pass it at some point
  • Here are some traditional objections
  • 1) The Soul you must have a soul to think.
  • 2) Originality Computers must obey programs so
    they cannot do anything original.
  • 3) Humor Thinking yields humor, machines dont
    make jokes.
  • 4) Head in the Sand Oh dear, wouldnt be awful
    if they did think?
  • 5) Learning Machines cant learn, but thinkers
    can.

26
Turings Replies
  • 1) The Soul you must have a soul to think.
  • -- God could make a thinking machine
  • 2) Originality Computers must obey programs so
    they cannot do anything original.-- Programs
    could incorporate randomness
  • 3) Humor Thinking yields humor, machines dont
    make jokes.-- Why must thinking and humor
    coincide?
  • 4) Head in the Sand Oh dear, wouldnt be awful
    if they did think?
  • -- A worry is not an objection.
  • 5) Learning Machines cant learn, but thinkers
    can.
  • -- Complex machines can acquire new inputs from
    the environment and previous processing.

27
Searles Chinese Room
  • The best known objection to strong AI is John
    Searles CHINESE ROOM objection.

The objection is this If a system can pass the
Turing Test then it understands language, but in
a Chinese Room there is no part that understands
language. So at best a TT simulates thinking.
28
The Chinese Room
MEMORY
OUTPUT
INPUT
PROCESSOR
Write a Comment
User Comments (0)
About PowerShow.com