Spring 2008

Readings can be found at: http://mind.cog.jhu.edu/courses/326/

Foundations of Cognitive Science 050.326/626

Prof. Paul Smolensky (smolensky@jhu.edu) O.H. TBA

T.A. Simon Fischer-Baum (fischerbaum@cogsci.jhu.edu) O.H. W 12:00PM-1:00PM(and by apmt.)

T.A. Mónica López-González (lopezgonzalez@cogsci.jhu.edu) O.H. W 3:30PM-4:30PM(and by apmt.)

T.A. Michael Oliver (oliver@cogsci.jhu.edu) O.H. T 4:00PM-5:00PM(and by apmt.)

Prerequisites:  At least one course at the 300-level or higher in cognitive science, psychology, computer science, linguistics, or philosophy of mind. Additionally, for 626: graduate standing.

Course goal:  To introduce students to the broad spectrum within cognitive science of: empirical methods, theoretical methods and frameworks, general conceptual issues, and major authors.  The course is short on concrete content, long on understanding and integrating the basic perspectives on cognition provided by multiple disciplines. The key skill in focus is logical argumentation.

Course theme:  Research on ‘neural network’ or ‘connectionist’ computational models will provide one concrete context in which to consider a number of general course themes. Several cognitive domains are considered, but language is the one most frequently discussed.

Coursework:  Reading (mostly major papers from the research literature); daily written work on readings; class attendance, and participation in class discussion. Graduate students will have an additional integrative assignment.

Location: Graduate students and Undergraduate A will meet in the regular classroom. Undergraduate B will meet in the Seminar Room.

Course Requirements

        Attendance and participation in class discussion is extremely important, as this is an entirely discussion-based class. The reading and writing demands of the course are rather high (undergraduates receive ‘W’ credit for the course).

        Students will hand in a write-up for every reading at the beginning of the class meeting during which that reading is to be discussed. Students will have 5 ‘waivers’ of this requirement, to use any time during the semester. For the student, the goal is to gain experience in extracting and evaluating the logic of the arguments presented in cognitive science. For the class, the goal of this requirement is to ensure that students have given sufficient thought to the day’s reading prior to class to be able to contribute to a high-quality discussion. Thoughtful contributions to class discussion is an important part of the course and of the grade. Students may be asked to lead off discussions in some classes.

        Write-ups fall into two categories:  ‘prose’ and ‘diagrams’. 

        The prose write-ups do not have a prescribed format, but they should by and large provide answers to the following sorts of questions:

(1) With what questions is the author mainly concerned?

(2) What are the author’s main claims concerning these questions?

(3) What arguments are given to support these claims? Are they sound?

(4) On what methodologies are these arguments based?

(5) How do these claims relate to the major course issues?

(6) How do these claims relate to those of other authors we’ve (or you’ve) read?

        Prose write-ups should NEVER reiterate the reading. The write-ups address the CONTENT of the paper, not the paper itself. (Not ‘First the author talks about this and then she talks about that and I don’t think she did a good job talking about either one.’) If directly addressed to the point, and concise, such write-ups may require only a single page (single spaced). Two pages may be more realistic much of the time. One well-thought-out point is worth much more than a dozen superficial comments. When several readings are assigned for a single day, even if you choose to focus on one of them, your write-up should address each of the readings to some degree (separately or jointly).

        The format for the ‘diagram’ write-ups is more prescribed. These are formal diagrams of the logic of the reading -- or at least a key portion of it. These argument diagrams will be presented and discussed in class; to get the general idea, check the diagrams on the posters and sheets posted on the wall of the corridor containing the Cog Sci seminar room, Krieger 234C.

        8 of the daily write-ups will be graded on a 5-point scale; the remaining write-ups will be simply checked off as done (assuming a reasonable minimal level of effort has been expended). Each student can choose which assignments are to be graded; these should be clearly marked “for grade” at the top. The graded assignments will consist of 4 prose write-ups and 4 diagrams. Students are always welcome to bring their written work to the professor or TA’s office hours for more detailed feedback.

        The upshot is that for each class, students can choose among these options:

¨       hand in nothing, using up one waiver (no more than 5 total for the semester)

¨       hand in a prose write-up marked “for grade”(for semester, total of 4)

¨       hand in a diagram write-up marked “for grade”(for semester, total of 4)

¨       hand in a prose or diagram write-up not marked “for grade” (all remaining classes)

        Students are strongly encouraged to work together on any of the class assignments. For prose write-ups and essays, each student is expected to do their own write-up; the other students worked with must always be identified. For diagrams, groups can submit a single diagram listing all group members.

        Graduate students will be required to do a final project. The details of this project will be discussed later in the course.

Late work/incompletes. Work handed in late will be marked down (one point per day, out of 3 or 5); work more than a few days late will not be accepted. No incompletes will be given, except for documented medical emergencies.

 Reading Schedule:

This schedule is subject to change.

 Key:

§ = pre-20^th century philosophy: not primary reading; excerpts from B. Russell, A History of Western Philosophy; see below for page numbers

* = in the Cummins & Dellarosa Cummins (eds) collection (Minds, Brains, and Computers)

‡ = in the Pinker & Mehler (eds) collection (Connections and Symbols)

! = note relevant page numbers given below under ‘Readings’ and ‘References’

General Methodological Topics

Research methods

Major distinctions and issues [‘East-pole’ position]

• Theory- vs. data-centering

  • Should the primary goal of cognitive science be to construct a comprehensive, insightful theory or to collect and account for a comprehensive set of data? [theory]

• Theories of brain function vs. theories of behavior vs. theories of mental knowledge

  • What is cognitive science the science of? [mental knowledge]

• Levels of cognitive analysis

  • At what level does the explanation of cognitive phenomena lie: the level of individual neurons, or a more fine-grained, coarse-grained, or abstract level? [highly abstract level]

• Explanation vs. description*

  • What should be the descriptive or explanatory goals of cognitive science? [deep explanation, not superficial description]

• Internal vs. external/functional explanation*

  • Should cognitive explanation derive from principles internal to the cognitive system, or from external factors such as the function served by cognition or the physical constraints under which it operates? [internal explanation]

• Formal vs. non formal theories

  • What is the appropriate level of formality for cognitive theory? Is human knowledge a formal system? [highly formal; yes]

• Computational vs. non computational frameworks

  • Should a cognitive theory be a computational theory, viewing the mind as a machine that takes data in, processes it according to an algorithm, and outputs behavior? [yes]

• Representational vs. non representational frameworks*

  • Does cognition deploy representations of the outside world, on which it computes? [yes]

• Competence vs. performance

  • Should cognitive science seek theories that explain the details of actual human performance, or theories of ‘competence’ — the knowledge underlying performance, ignoring the vagaries of particular instances of use of this knowledge? [competence first]

• Mind as statistical processor vs. mind as structure processor

  • Is the mind primarily a device for processing highly structured, symbolic information, or primarily a device for performing statistical analysis of experience? [structure]

• Nativism vs. empiricism

  • Does knowledge derive from experience? [not the crucial knowledge]

• Conscious vs. unconscious processes

  • What roles do conscious and unconscious processes play in an overall theory of cognition? [unconscious knowledge is key]

• Knowledge/concepts as rules/definitions vs. as examples

  • In the mind, are concepts, or knowledge more generally, constituted of stored specific examples, or of general definitions or rules, as in mathematics? [definitions and rules]

• Inference and decision making: logic based vs. non rational approaches*

  • Are human inferences and decisions based on some kind of logic? [yes]

• Serial vs. parallel processing

  • What roles do serial and parallel processes play in an overall theory of cognition? [abstract seriality]

• Independent mental faculties (‘modules’) vs. interactionism*

  • Is the mind a collection of fairly autonomous faculties or modules, each concerned with some particular cognitive domain and governed by its own idiosyncratic principles, or is the mental processing of information from multiple domains so heavily interactive that decomposition into separate modules is not possible or useful? [modular]

• Localized vs. distributed neural embodiment

  • Are localized bits of representations, knowledge or processes realized in localized bits of the nervous system? [irrelevant]

• Situated vs. non-situated cognition*

  • Must a theory of cognition depend crucially on the way the mind is situated in the body and in the external social and physical world? [are you kidding?]

* Foundations B  

Covert Aphorisms in Cognitive Science:

Philosophy of science

Science = Data        or      Data ≫ Theory

Data = Experiments

Better theory = More data coverage

Serious constraint on theory ⇒ Data         (Therefore, AI, Philosophy ⇒ BS)

Theory = Description

Theory = Empirical generalizations

Unmeasurable ⇒ Unscientific

Precise theory ⇒ Computer implementation

Precise theory ⇒ Mathematical formalism

Implausible (introspective intuition) ⇒ False

Explanation ⇒ External (e.g., functional) justification

Functional fable ⇒ Explanation

Internal explanation ⇒ Circular reasoning

Demystification ⇒ Trivialization/Denigration

Simplification ⇒ Irrelevance

Model ⇒ Theory 

Substantive

Mind = Brain

Cognition = Behavior

Thought ⇒ ¬Logic

Mind ⇒ ¬Formal system

Knowledge = Experience

Knowledge = Rules

Concept = Definition

Category = Categorization

¬[Computational ⇒ Intentional]

Level of representation = Level of cognitive computation

Connectionism ⇒ Associationism

Connectionism ⇒ Empiricism

Neurally informed ⇒ Neural model

Language = X (= Communication; Words; Speech recognition) 

Readings

Books:

Chomsky, N.  1965.  Aspects of the Theory of Syntax. MIT Press. [1/4 chapters]

* Cummins, R. & Dellarosa Cummins, D.  2000.  Minds, Brains, and Computers: The Foundations of Cognitive Science.  Blackwell. [9 primary + 10 secondary chapters]

‡ Pinker, S. & Mehler, J. eds.  1988.  Connections and Symbols.  MIT Press/Bradford Books. [2/3 chapters]

§ Russell, B.  1945.  A History of Western Philosophy.  Simon & Schuster. [9/31 chapters]

Russell readings:

References

Anderson, J. R. 1982. Acquisition of cognitive skill. Psychology Review 89, 369-406; Read 369-382 and skim the rest.

Anderson, John. R., Kline, Paul J., and Beasley, Charles. M, Jr. 1979. A general learning theory and its application to schema abstraction. Psychology of Learning and Motivation 12.

Bechtel, W.  1988.  Philosophy of Mind:  An Overview for Cognitive Science.  Erlbaum. Chapter 1: Some perspectives on philosophy of mind, pp. 1-17. [optional: general background reading]

Chomsky, N. 1957. Syntactic Structures.

Chomsky, N.  1959.  A review of B.F. Skinner's Verbal Behavior.  Language 35:26-58  [Sections 1-4,11; reprinted in Block, N. ed., 1980, Readings in Philosophy of Psychology, Harvard University Press, 48-63.]

Chomsky, N.  1965.  Aspects of the Theory of Syntax. MIT Press. Preface, pp. v-vii, and Chapter 1: Methodological Preliminaries, pp. 1-­62. [read §§1-4, pp. 3-27; skim §§5-7, pp. 27-47; read §8, pp. 47-59; skip §9, pp. 60-62.]

Chomsky, N. 1967. Recent contributions to the theory of innate knowledge. Synthese, 17:2-11.

Chomsky, N. 1969. Linguistics and philosophy. In S. Hook (ed.) Language and Philosophy. NYU Press.

Chomsky, N.  1986.  Knowledge of Language.  Praeger.  Preface, Chapter 1, pp. xxv-13.

Chomsky, N.  1988.  Language and Problems of Knowledge:  The Managua Lectures.  MIT Press.

Dennett, D. 1984/1998. The logical geography of computational approaches : A view from the East Pole. Reprinted in Brainchildren, 215–234. MIT Press.

Elman, Jeffrey L., Bates, Elizabeth A., Johnson, Mark H., Karmiloff-Smith, Annette, Parisi, Domenico, and Plunkett, Kim. 1997. Rethinking Innateness: A Connectionist Perspective on Development. MIT Press.

Fodor, J.A. & Pylyshyn, Z.W.  1988.  Connectionism and cognitive architecture.  Cognition 28:3-71. [Reprinted in S. Pinker & J. Mehler (eds.), Connections and Symbols.]  [Skim §4; leaves 52 pp.]

Fodor, J. A. 1975. The language of thought: First approximations. The language of thought. Harvard University Press.

Fodor, Jerry A. and McLaughlin, Brian P. 1990. Connectionism and the problem of systematicity: Why Smolensky’s solution doesn’t work. Cognition 35.

Haugeland, J. 1996/1997. What is mind design? In J. Haugeland, Mind design II, 1–28. MIT Press.

Hebb, D.O.  1949.  The Organization of Behavior.  Wiley.  Introduction, pp. xi-xix, Chapter 4: The first stage of perception: growth of the assembly, pp. 60-78.  [Reprinted, with new introduction, in J.A. Anderson & E. Rosenfeld eds., Neurocomputing:  Foundations of Research, MIT Press/Bradford Books; pp. 43-56.]

Hofstadter, D. 1979. Gödel, Escher, Bach: An Eternal Golden Braid. Basic Books.

Hofstadter, D.  1985.  Metamagical Themas.  Basic Books. Chapter 26:  Waking up from the Boolean dream, or, subcognition as computation, pp. 630-665.

Lachter, J. , and Bever, T.G. 1988. The relation between linguistic structure and associative theories of language learning : A constructive critique of some connectionist learning models. Cognition 28:195-247. [Reprinted in S. Pinker & J. Mehler (eds.), Connections and Symbols.]

Lakoff, G.  1987.  Women, Fire, and Dangerous Things.  University of Chicago Press.   Excerpts: Preface, pp. xi-xvii; Ch. 1, pp. 5-15; Ch. 2, pp. 12-15, 56-57; Ch. 3, 58-67; Ch. 4, pp. 68-74; Ch. 21, pp. 370-373.

Lashley, K.S.  1950.  In search of the engram.  Society of Experimental Biology Symposium, 4: Psychological Mechanisms in Animal Behavior.  Cambridge University Press.  Pp. 454-455, 468-473, 477-480.  [Reprinted, with new introduction, in J.A. Anderson & E. Rosenfeld eds., Neurocomputing:  Foundations of Research, MIT Press/Bradford Books; pp. 57-67.]

Lettvin, J. Y., Maturana, H. R., McCulloch, W. S., Pitts, W. H. 1959. What the frog’s eye tells the frog’s brain. Proceedings of the Institute of Radio Engineers 47.

Marr, D. 1982. Vision. W. H. Freeman and Company.

McCarthy, J. 1968.  Programs with common sense.  In M. Minsky ed., Semantic Information Processing, MIT Press, 403-418.  [Reprinted, with new introduction, in R. Brachman & H. Levesque eds., 1985, Readings in Knowledge Representation, Morgan Kaufmann, pp. 299-307.]

McClelland, J.L., Rumelhart, D.E., & Hinton, G.E.  1986.  The appeal of parallel distributed processing.  In D.E. Rumelhart, J.L. McClelland & The PDP Research Group, Parallel Distributed Processing: Explorations in the Microstructure of Cognition; Vol. 1: Foundations, 3-44. [Reprinted in Collins & Smith eds., pp. 52-72.]

McCloskey, Michael. 1991.  Networks and theories: The place of connectionism in cognitive science.  Psychological Science, 2:387-395.

McCulloch, Warren S. and Pitts, Walter. 1953. A logical calculus of the ideas immanent in nervous activity. Bulletin of Mathematical Biophysics 5.

Newell, A. & Simon, H.A.  1963.  GPS, A program that simulates human thought.  In E. A. Feigenbaum & J. Feldman eds., Computers and Thought, 279-293.  [Reprinted in Collins & Smith eds., pp. 453-460.]

Newell, A. & Simon, H.A.  1972.  The theory of human problem solving.  Prentiss-Hall,  Chapter 14. [Reprinted in Collins & Smith eds., pp. 33-51.]

Pinker, S. & Prince, A.  1988.  On language and connectionism:  Analysis of a parallel distributed processing model of language acquisition. Cognition 28: 73-194.  [Reprinted in S. Pinker & J. Mehler (eds.), Connections and Symbols.]  [Skim §4, skip §7 and appendix; leaves 47 pp.]

Plunkett, K., Marchman, V. 1993. From rote learning to system building: Acquiring verb morphology in children and connectionist nets. Cognition 48, 21-69.

Putnam, H. 1967. The ‘innateness hypothesis’ and explanatory models in modern linguistics. Synthese 17:12-22

Rosch, E. 1978. Principles of categorization. In E. Rosch & B.B. Lloyd eds., Cognition and Categorization, Erlbaum, pp. 27-48. [Reprinted in Collins & Smith eds., pp. 312-322.]

Rosenblatt, F. 1958. The perceptron: A probabilistic model for information storage and organization in the brain. Psychological Review 65.

Rumelhart, D.E. & McClelland, J.L. 1986. On learning the past tenses of English verbs. In J.L. McClelland, D.E. Rumelhart, & The PDP Research Group, Parallel Distributed Processing: Explorations in the Microstructure of Cognition; Vol. 2: Psychological and Biological Models, pp. 216-271 [Skim pp. 248-260].

Schank, R. & Abelson, H.  1977.  Scripts, Plans, Goals, and Understanding.  Erlbaum.  Chapters 1-3, 1-68.  [Reprinted in Collins & Smith eds., pp. 190-223.]

Searle, J.R.  1980.  Minds, brains, and programs.  In The Behavioral and Brain Sciences,  3:417-424. [Reprinted in Mind Design II, John Haugeland, ed., 183-204. 1997. MIT Press. ]

Sejnowski, Terrence J. and Rosenberg, Charles R. 1987. Parallel networks that learn to pronounce English text. Complex Systems 1.

Skinner, B.F.  1953.  Science and Human Behavior.  Macmillan. Excerpts: pp. 27-35, 62-66, 87-90. [Reprinted in Block, N. ed., 1980, Readings in Philosophy of Psychology, Harvard University Press, pp. 37-47.]

Smolensky, P.  1988.  On the proper treatment of connectionism.  The Behavioral and Brain Sciences, 11:1-23.

Smolensky, P. 1991. Connectionism, constituency, and the language of thought. In B. Loewer & G. Rey (Eds.), Meaning in Mind: Fodor and his Critics. Oxford: Basil Blackwell. 201–227.

Smolensky, P. 1994. Computational theories of mind. In S. Guttenplan (Ed.),  A Companion to the Philosophy of Mind. Blackwell Publishers. 176-185.

Smolensky, P., & Legendre, G. 2006. Harmony optimization and the computational architecture of the mind/brain. In P. Smolensky & G. Legendre, The harmonic mind: From neural computation to optimality-theoretic grammar, Vol. 1, 3–61. Cambridge, MA: MIT Press. [Chapter 1]

Smolensky, P., & Legendre, G. 2006. Principles of the integrated connectionist/symbolic cognitive architecture. In P. Smolensky & G. Legendre, The harmonic mind: From neural computation to optimality-theoretic grammar, Vol. 1, 63–97. Cambridge, MA: MIT Press. [Chapter 2]

Spelke, Elizabeth. 1994. Initial knowledge: Six suggestions. Cognition, 50.

Talmy, L. 1988.  Force dynamics in language and cognition.  Cognitive Science, 12:49-100.

Turing, A.M.  1950.  Computing machinery and intelligence.  Mind, 59:433-460.  [Reprinted in Collins & Smith eds., pp. 6-19.]

Ungerleider, Leslie G. and Mishkin, Mortimer. 1982. Two cortical visual systems. In D. J. Inles (ed.), Analysis of Visual Behavior. MIT Press.