Computational Theory of Mind

topsalmonIA et Robotique

23 févr. 2014 (il y a 3 années et 5 mois)

203 vue(s)

Computational Theory of Mind

Philosophy Course proposed to the Overseas Young Chinese Forum

Syllabus and Schedule


James Blackmon

jcblackmon@gmail.com


Introduction

A guiding principle in cognitive science is that mental processes can be studied as we woul
d study
computational process. According to the computational theory of mind (CTM), this is because cognitive
processes
are

computational processes. In short, the mind is a program

however special and elaborate
this program may be. After a contextualizing
introduction to the philosophy of mind and a brief survey
of the history of the study of cognition, our course will investigate CTM and its theoretical and empirical
implications.

Course Requirements

Student grades will reflect their performance on one pap
er assignment, one exam, two oral
presentations, and class participation according to the percentages listed.


Paper Assignment

20%


Exam



20%


Oral Presentation 1

20%


Oral Presentation 2

20%


Participation


20%

The paper assignment will be a response t
o one of a few prompts I will give or, if you so choose, a
prompt that you propose and I approve. There is no strict page minimum, but about six to eight pages of
focused analysis and argumentation is typical.

The oral presentations will be fairly short (
8
-
12 minutes each), requiring that you present an important
element of a current reading and give some reasoned evaluation of it. You will also be expected to field
questions and comments.

While the paper and the oral presentations emphasize critical evalu
ation, the exam will mainly test your
knowledge of the views and issues we have studied.

Participation involves
productive engagement in discussion and regular attendance.

Course Schedule

Each unit is 2 hours. Typically, we will begin with a lecture prese
ntation by me elaborating the current
readings and follow up with some short student presentations and

guided discussion. Readings
designated as
Recommended

are intended to extend your knowledge of these topics at your discretion,
not to distract from the
required reading. Recommended readings will usually be referenced in my
lectures, but you will not be required to say anything about them (on the exam or in your paper) outside
of what I cover in class.

Unit 1: The Mind
-
Body Problem.

What is the relation b
etween the conscious mind and the material
body, especially the brain? Are “they” strictly identical to the point that we can even say the mind is
about 3.5 pounds of organic material located inside the cranium? Or are mind and brain distinct things
even i
f they may interact or hold strong dependency relations? This unit introduces the responses to the
mind
-
body problem from Descartes’ dualism to the main alternatives with us today. In the process, we
will acquire a useful vocabulary for the course.



Descart
es’
Meditations

II

and
VI

[9]



‘Sensations and Brain Processes’, J. J. C. Smart [8]



Recommended: ‘Foundations’,
Philosophy of Mind: Classical and Contemporary Readings
, David
J. Chalmers [9]

Unit
2
:
Related Problems:
Consciousness and Intentionality.

A comm
on and intuitive view is that any
science of the mind must explain not only the “external appearance” of cognition, but also
consciousness and intentionality. We have qualitative subjective experiences (consciousness), and our
thoughts are about things in
the world (intentionality). This unit lays the groundwork for these
challenges which appear in explicit form later in the course.



‘Epiphenomenal Qualia’, Frank Jackson [8]



‘An Unfortunate Dualist’, Raymond Smullyan [1]



Recommended: ‘A Recipe for Thought’,
Fred Dretske [9]

Unit

3
:
Foundations for CTM
.

As problems for dualism and standard physicalist views mounted,
functionalism was hailed as a promising alternative, for not only did it avoid the objections that had
stalled the competing theories, but also it

made the pursuit of artificial intelligence coherent. This unit
presents foundational work on functionalism and the (more specific) idea that the mind is a computer.
An introduction to formal concepts of computation theory is included: Turing Machines, FS
As, Mealy
and Moore models, machine tables, isomorphism.



Recommended: ‘A History of Thinking’, Denise Dellarosa Cummins [10]



‘Computing Machinery and Intelligence’ Alan Turing [14]



‘The Nature of Mental States’, Hilary Putnam [7]



‘A Short Primer on Finite
State Automata’, J. Blackmon [2]

Unit

4
:
The Language of Thought
.

CTM gained additional support and elaboration from Jerry Fodor’s
language of thought hypothesis. This unit examines the rationale behind it and some of the main
criticism that continue to ch
allenge the view today. The recommended readings will be outlined in class
as contrast cases.



‘The Language of Thought: First Approximations’, Jerry A. Fodor [13]



Recommended: ‘Stalking the Wild Epistemic Engine’ Paul M. Churchland and Patricia Smith
Churc
hland [8]

Unit

5
:
Top
-
Down Computational Approaches.

Though Fodor has done so much to advance CTM, he
has lately come to raise serious objections to the idea that we are developing a general theory of
cognition and to the optimism that surrounds it. In thi
s unit, we will focus on the tensions between
Steven Pinker’s book and Fodor’s response. The recommended additional readings are examples of
originating approaches in CTM.



Excerpt from
The Mind Doesn’t Work that Way
, Jerry Fodor [12]



‘So how does the mind
work?’, Steven Pinker [22]



Recommended: ‘Vision’, David Marr



Recommended: ‘GPS, a Program that Simulates Human Thought’, Allen Newell and H. A. Simon.

Unit
6
:
John Searle on Simulation.

Searle’s Chinese Room thought experiment has generated decades of
live
ly controversy that persists till this day. Searle argues that running some AI program that allows one
to perfectly simulate Chinese conversation is not sufficient for actually understanding Chinese, since
Searle could run the program, but he nevertheless
does not understand any Chinese. If Searle is right,
then CTM fails.



‘Minds, Brains, and Programs’, John Searle [12]

Unit 7. Ned Block on Simulation.

No account of CTM or any kind of functionalism is complete without
addressing the problems Block surveys i
n this important and widely
-
cited piece. We will focus
particularly on Block’s homunculi
-
headed robot, extending our previous work on ideas brought forth in
Searle’s Chinese Room.



Excerpt from ‘Troubles with Functionalism’, Ned Block [4]



Recommended: ‘Trou
bles with Functionalism’, Ned Block [28]

Unit 8:
The Problem of Universal Realization.

What if whether a physical thing implements a
computation is just arbitrary, subjective, or “observer relative”? Or what if, to the extent that it is
objective, it’s tri
vial? Recently, more attention has been given to the idea that, logically speaking,
anything can be interpreted as running any program because programs are mere mathematical
abstractions and, except for complexity constraints, there is nothing beyond our o
wn interpretive
decisions to determine whether a physical thing runs a program. The problem is potentially deadly to
CTM. For, unless there is more to say about programs or program implementation, CTM inadvertently
attributes cognition to almost everything
, rocks included. This is the problem of universal realization.
Here, we investigate this general charge, made by Hilary Putnam, John Searle, and Paul Teller, and we
look at a response by David Chalmers that first strengthens their claims then defends agai
nst them by
proposing a new kind of automaton.



Excerpt from
Representation and Reality
, Hilary Putnam [2]



Excerpt from
The Rediscovery of Mind
, John Searle [3]



‘Teller’s Rock’, J. Blackmon [5]



‘Does a rock implement every finite
-
state automaton?’, David Ch
almers []

Unit 9: The Nature of Program Imp
lementation.

In response to the problem of universal realization,
many people have argued that role of program implementation has been seriously neglected. It is
argued that a robust and empirically
-
based notion o
f program implementation avoids the absurd results
Putnam and Searle find.



‘When Physical Systems Realize Functions’ Matthias Scheutz []



Recommended: ‘Searle’s Wall’, J. Blackmon [12]

Unit 10:
Review.

This unit is dedicated, first, to any issues or topics

we, as a class, have decided warrant
re
-
examination and to synthesizing the course material and its broader implications. Second, students
will have time to share and workshop their paper thesis. Third, I will guide students in preparation for
the exam.



R
eview reading of some previous material will be assigned pending a class discussion on what
has been most important, difficult, interesting, and promising.