Prof: Christopher Cherniak
|
PHIL 209H / HONR 258N PHILOSOPHY AND COMPUTERS: From Logic to Thinking Machines? Tu, Th 2:00-3:15 Spring 2005 |
Christopher Cherniak Office: 1103B Skinner Office Hrs: by appt, & Tu 3:30-4:30, Th 3:15-4:00 |
Summary:
During the past century, formal logic made its
greatest progress since
Aristotle.
Its achievements are perhaps comparable to better-known
ones of our era--for instance, of relativity
theory or quantum
mechanics.
Paradoxically, the main results of mathematical logic are
negative, demonstrating absolute as well as
practical limits on all
computing:
simple, clear problems that an ideal machine the size of
the Universe could never solve.
This research on the abstract theory of
computation led directly to
the engineering technology of real-world
digital computers
(principally in connection with weapons
projects of WW II and the Cold
War; practical large-scale computation was
essential in the design of
nuclear weapons). A more positive outcome of the emergence of real-
world computing hardware has been research on
machine intelligence.
Indeed, the computer model of the human mind
serves as a central
unifying conceptual framework of the
cognitive sciences.
This course proceeds from an introduction to
computation theory, to
some philosophy of mind--that is, from the
unsolvability results of
computation theory to questions regarding
whether machines can (ever)
think.
The first half of the course is organized around the key
concept of computation theory, that of the
algorithm or program-
schematic.
The second half focusses on the philosophical adequacy of
computational psychology, the
information-processing model of mind;
more concretely, what are the possibilities,
prospects, and
limitations of artificial intelligence?
PHIL 209H / HONR 258N is in the area of
Mathematics and the Sciences--
Mathematics and Formal Reasoning (Non-Lab)-MS
in the CORE Distributive
Studies category. There will be homework problem assignments in
connection with the first part of the course
(some of them to be
worked out using PC software developed for
the course by the
University of Maryland Philosophy
Department). For the second half of
the course, the main requirement will be
three short papers (about 5
pages each: typed, double-spaced, with
1" margins) on assigned,
well-defined questions. This is not a programming course, nor does
it
require any programming background. Students not majoring in
philosophy are welcome.
Principal texts (in order of use):
D.
Harel, Algorithmics, 3rd ed, 2004.
(A. Dewdney, The (New) Turing Omnibus, 1993.)
R. & D. Cummins, eds, Minds, Brains, and Computers, 2000.
H. Dreyfus, What Computers Still Can't Do, "3rd" ed, 1992.
(W. Lycan, ed., Mind and Cognition, 1990/1999.)
(J. Haugeland, ed., Mind Design II, 2nd ed, 1997.) Copies of some additional material will be
distributed in class. Other suggested readings:
M. Minsky, Computation: Finite and Infinite Machines, 1967.
R. Courant and H. Robbins, What is Mathematics?, 1969.
S. Kleene, Mathematical Logic, 1967.
D. Dennett, Brainstorms, 1978.
P. Winston, Artificial Intelligence, 2nd ed., 1984. Software
packages. Downloadable from:
Outline I.
The Nature of Computation 1.
The concept of an algorithm: a finite, completely specified set of
instructions. Harel, Part I, especially ch 1 (Minsky, ch 5) (Kleene, pp. 223‑231) 2.
Turing machines: a standard format for algorithms. Harel, ch 9, especially 219-231 [= 223‑238, 2nd ed] TM* PC onscreen Turing machine
simulator/interpreter (on PCOMP) 3.
The universal Turing machine: an idealized computer that can execute any
algorithm. Harel, ch 9, especially 236-238 [= 242‑244] (Minsky, ch 7) UTM on TM* on PCOMP
4.
Unsolvability: the Halting Problem cannot be solved by any Turing machine. Harel, ch 8, especially 198-207 [= 203‑211] (Minsky, ch 8) Kleene, pp. 175‑183, 246 (Courant & Robbins, pp. 77‑88 (countability &
uncountability)) W. Quine, "The Ways of Paradox," in The Ways
of Paradox, 1966 [Computational
intractability vs absolute unsolvability.]
H. Lewis & C. Papadimitriou, "The
Efficiency of Algorithms," Scientific American, Jan 1978 L. Stockmeyer & A. Chandra,
"Intrinsically Difficult Problems,"
Scientific American, May
1979 Harel, ch 7
5.
The organization of von Neumann machines: a simple model of a random‑access
memory computer.
RAM* PC onscreen von Neumann machine
simulator/interpreter (on PCOMP)
www.glue.umd.edu/~cherniak/philcomp/
MIT Encyclopedia of Cognitive Sciences [ cognet.mit.edu ]
PCOMP logic engine simulator (TM* & RAM*) UM THERMO parallel computation modeller CogSci Micro-SHRDLU blocks-world natural-language "understander" Courseware
ALICE [2000, 2001, 2004 winner of Loebner "Turing Test" Prize] www.alicebot.org
(ELIZA "psychiatrist" program, J. Weizenbaum)
(RACTER conversationalist program, INRAC Associates)
Dictionary of Algorithms [ www.nist.gov ]
History
Stanford Museum: www.computerhistory.org
Turing: www.turing.org.uk/turing/
Bletchley Park: www.codesandciphers.org.uk
[6. Parallel computation: out of the von Neumann bottleneck?]
S. Kirkpatrick, C. Gelatt, & M. Vecchi, "Optimization by Simulated Annealing," Science 220, 1983
THERMO parallel computation modeller
D. Tank & J. Hopfield, "Collective Computation in Neuronlike Circuits," Scientific American 257, 1987
(Harel, ch 10)
II. The Computer as Model of the Mind
1. Are minds machines: implications of unsolvability theorems?
D. Dennett, "The Abilities of Men and Machines," in Brainstorms
(J. Lucas, "Minds, Machines, and Godel," (1961) www.jstor.org)
Lucas, The Freedom of the Will (1970) chs 24-26
http://users.ox.ac.uk/~jrlucas/mmg.html
(R. Penrose, The Emperor's New Mind (1989) ch 4)
2. Can machines think: Turing's test.
Harel, ch 15 [= 12]
A. Turing, "Computing Machinery and Intelligence," in Cummins [& jstor]
ELIZA "psychiatrist" program & RACTER conversationalist program vs M‑SHRDLU "blocks‑world" natural language understander
(Winograd, ch. 1, Understanding Natural Language; cf Cummins)
3. What are mental states: computational psychology.
H. Putnam, "Robots: Machines or Artificially Created Life?" 1964
(Putnam, "Minds and Machines," in Cummins) - on jstor
(J. Fodor, "The Mind‑Body Problem," Scientific American, Jan 1981)
N. Block, "The Computer Model of the Mind"
C. Cherniak, "The Wager," AI Magazine, Aug 1986 - on www.glue.umd.edu/~cherniak/
4. Against artificial intelligence: a phenomenological critique.
H. Dreyfus, Introduction to 2nd Edition (1979), What Computers Still Can't Do; also, Introduction to "3rd" Edition (1992)
(Dreyfus, pp. 155‑227, 285‑305)
[C. Cherniak, "Undebuggability and Cognitive Science," Communications of the Association for Computing Machinery, April 1988] www.glue.umd.edu/~cherniak/