Reading List:

Chomsky, Noam. 1968/72. Language and Mind. New York: Harcourt Brace
Dougherty, Ray C. 1994. Natural Language Computing: An English Generative Grammar in Prolog. Hillsdale, N.J.: Lawrence Eribaum Inc.
Negroponti, Nicholas. 1996. Being Digital. New York: Random House, Vintage.

By the first exam, students should read Negroponti's book.
By the second exam, students should read this book, available over the internet and on reserve:
Darwin, Charles. 1872. The Expression of Emotion in Man and Animals.

All supplemental readings, and all intemet materials, are on reserve in Bobst Library and in the Linguistics Department Library, 719 Broadway, Fifth Floor.

Check the Academic Computing Facility for lab hours:

Computational Linguistics: Symbolic Information Processing

1 . Introduction to Linguistics, Intelligence, and Symbol Processing. Overview of the course, materials, the world wide web, HTML, Prolog, linguistics, animal languages, and sundry details such as exams, final papers, computer accounts, course requirements, and so on.

2. Background and Problem Statement. What is a language? A grammar? What does a linguist do? Can a computer be trained to speak and understand? What is natural intelligence? Artificial intelligence? Anirnal/Hunian intelligence? Why is it difficult to program a grammar (English, Chinese, and so on) onto a computer? How can one study animal languages/communication without knowing what the animals are saying or talking about? Signal structure (syntax) versus message content (semantics). Autonomous syntax.
(b) Dougherty, 1994, Preface, pp. viii-xviii
(c) Dougherty, 1994, Introduction: What is Computational Linguistics?, pp. xix-xxviii
(d) Dougherty, 1994, How to Use this Book: A Pictorial Essay, pp. xxix-xlvi
(e) Chomsky, 1968, Language and Mind.- Preface, pp. I-vii

3. Overview of Intelligence: Natural, Artificial, Animal, and Otherwise. Of all of the various tests and criteria for intelligence offered in the past couple of centuries, we will focus on linguistically based definitions. We will focus on the intelligence required by a human (child and adult) to acquire, understand, and use a human language.
(a) Chomsky, 1968, Linguistic Contributions to the Study of Intelligence: Past, pp. 1-23.
(b) Dougherty, 1994, Natural Intelligence, Linguistics, and Prolog, pp. 1-23.

4. Methods In Linguistic Research. Chomsky's ideas about language learning in children and adults. Observational, descriptive, and explanatory adequacy. Peirce on abduction, deduction, and induction. Using Peirce's logic to show how to choose the 'best' grammar from several alternative proposals. The justification of beliefs and the construction of knowledge.
(b) Dougherty, 1994, 7. 1. The Levels of Human Language Structure, pp. 161-179.
(c) Dougherty, 1994, Chapter 2, How to Read and Write in Prolog, pp. 24-51

Noam Chomskv: Human Intelligence, Language, and Mind

5. Chomsky's Views about Language, Mental Processes, and Grammar. Definitions of morpheme, word, phrase, sentence, part of speech. Constituent structure, functional structure, relational diagrams. Sound-meaning correlations. A language is a huge relational database of elements (sentences) composed of smaller parts (words) combined according to grammatical principles.
(b) Dougherty, 1994, 7.2. Morphological Parsers, pp. 180-194.
(c) Chomsky, 1966, The Formal Nature of Language, pp. II 5-136.
(d) Dougherty, 1994, Chapter 2, How to Read and Write in Prolog, pp. 24-5 1.

6. The Levels of Human Language Structure. Elementary concepts of constituent structure. How to assign/correlate a phrase marker with a string of elements. Top-down parsing. Bottom-up parsing. Parallel processing. Do words, phrases, and clauses have internal structure?
(b) Dougherty, 1994, 7.3. Recursion: Affixes on the Affixes, pp. 195-215.
(c) Dougherty, 1994, Chapter 2, How to Read and Write in Prolog, pp. 24-5 1.
(d) Dougherty, 1994, Chapter 3, How to Load, Run, and Edit a Prolog Program, pp. 52-81.

7. Grammars: Parsers and Generators. Theories and models in linguistic research. Thinking of parsing and derivational processes as part of an intelligence machine that recognizes/decodes patterns in human languages.
(b) Dougherty, 1994, 7.4. Regular and Irregular Morphology, pp. 216-228.
(c) Dougherty, 1994, Chapter 3, How to Load, Run, and Edit a Prolog Program, pp. 52-8 1.

8. A Grammar/Language is Like a Cross-Point Switch. Chomsky's model of an ]-language as a switch. A grammatical sentence defines a path through the switch. Language acquistion consistsof setting the switch connectionstpositions.
(b) Dougherty, See Readingsfor Lecture 8, available in the libraries.
(d) Chomsky, 1968, Linguistic Contributions to the Study of Mind. Present, pp. 24-45.

9. Sentences: Simple, Complex, Coordinate, Subordinate. Big sentences (I know that you believe that she saw io are constituted from smaller sentences (/ know. You believe. She saw it.) Since there is no largest sentence in any language, there must be an unbounded number of them. A grammar is a finite means (computational scheme) to recognize all of the sound-rqeaning pairs (sentences) of a language. Words like and, or, after, that, for, and so on are the glue that join little sentences to make big ones.
(b) Dougherty, 1994, 8.1. Syntax: Representations and Parsers, pp. 239-253
(c) Chormky, 1968, The Fornwl Nature of Language: The Structure of the Syntactic Component, pp. 139-160

First Third Exam: Open Book, 20% of Grade

10. First Third Exam. There will be two sections, one based on readings and one based on lectures. Each will contain several questions. You must answer two from section I and two from Section II. This is an OPEN BOOK EXAM. You may bring any materials you wish to the exam: books, computers, printouts, notes, articles, etc. See the sample exams from another class. The exam will cover:
(a) Chomsky, Language and Mind, The readings listed above..
(b) Dougherty, 1994, pp. i-xliv, Chapters 1, 2, 3, 7. Section 8.1.
(c) Negroponti, 1996, Being Digital, The whole book.

Konrad Lorenz: Animal Intelligence, Releasers, and Automata

11. Animal Intelligence: Konrad Lorenz ande Niko Tinbergen's Nobel Prize Work. The different types of animal communication systems (Ducks/Geese, wolves, birds, bees). Efforts to teach chimpanzees English. The concept of a releaser.
(b) Readings for Lecture I 1: Articles by Lorenz and Tinbergen, etc. in libraries.
(d) Chomsky, 1968, Linguistics and Philosophy, pp. 161-194.
(e) Lorenz, Konrad. 1950. Part and Parcel in Animal and Human Societies: A Methodological Discussion. (81 pages) In libraries.

12. Communication in Bees: Frisch. Bee dances enable one honeybee to communicate information about type, source, direction, distance, and so on about flowers. The temporal and spatial cognitive capacities of the bee and bee communication systems.
(b) Readings for Lecture 12: Articles by Frisch, Wilson, etc. in libraries.
(d) Chomsky, 1968, Form and Meaning in Natural Languages, pp. 100- 1 14.

13. Communication in Ducks and Geese. Bird intelligence in general. The dancing language of ducks and geese. Pairing.
(b) Readings for Lecture 13: Articles on Bird Minds, in libraries.
(d) Chomsky, 1968, Linguistic Contributions to the Study of Mind: Past, pp. 1-24.
(e) Lorenz, Konrad. Preface to Darwin's book, The Expression of Emotion in Man and Animals. In libraries.

14. Communication among Dogs and Wolves. Releasers and egression. Symbolic fighting. Maintaining a hierarchy among the animals.
(b) Chomsky, 1968, Linguistic Contributions to the Study of Mind: Future, pp. 65-99.
(c) Readings for Lecture 14, Releasers and Hierarchy in Dogs, in libraries.
(e) Book extract: How to Judge Deer on the Hoof. In libraries.

Shannon, Wiener, and Turing: Machine lntelligence and Cybernetics

15. Shannon's Information Theory View of Language. Information relates to choices. Markov probabalistic grammars. Finite state models of language. Chess playing machines. A cybernetic machine.
(b) Dougherty, 1994, 8. 1. Syntax.- Representations and Parsers, pp. 230-253.
(c) Dougherty, 1984, 8.2. Rule Governed Creativity: Derivations, pp. 254-265.
(d) Shannon and Weaver, Excerpts about Information Theory, in libraries.
(e) Shannon, Claude. 1950. Chess Playing Machines. In libraries.

16. Wiener's Cybernetic Theory of Learning. The concept of feedback and its role In improving the performance of intellient machines. Feedback mechanisms in living systems, like an ant hill. A cybernetic learning machine.
(b) Dougherty, 1994, 8.3. Parsers Assign Structure to an Ordered String, pp. 266-275.
(c) Chonisky, 1968, The Fo@ Nature of Language: Structure of the Syntactic Component, pp. 65-99.
(d) Wiener, Norbert, Excerpts from Cybernetic Theory, in libraries.
(e) Bateson, Gregory. Cybernetic Explanation. In libraries

17. Turing's Ideas about Code Cracking. A child learning a language is basically cracking the code (acquiring the grammar) used by adults. An adult communicating with another adult is using a code (grammar) that is common knowledge between the two adults.
(b) Dougherty, 1994, 8.4. Top-Down and Bottom- Up Parsing, pp. 276-302.
(c) Chomsky, Reading to be announced.
(d) Turing, Alan, Excerpts from Turing's articles, in libraries.
(e) Hofstadter, Excerpts from book about Turing, in libraries.

18. Autonomous Syntax: The meaning of a sentence, synonymy of two sentences. What does it mean to say Iwo languages are the same'? When can a child be assumed to 'know' the language of its community? The Rubic language can be learned by pattern matching or statistically.
(b) Dougherty, 1994, 8.5. Horizontal Appends: Complement Structures, pp. 303-318.
(c) Gardner. Preface to The Great Debate (between Chomsky and Piaget). In libraries..

19. Formal Computational Models of Language Learning and Code Cracking. When can a code cracker assume that they have discovered the correct signal-message correspondences and cracked a secret code? When can a child assume it has acquired a/the correct grammar and learned the language?
(b) Dougherty, Ray. Cybernetic Models of Language Acquisition. In libraries.
(c) Chomsky, Reading to be announced.
(d) Other Reading to be announced.

20. A Review and Discussion of the Computational Models of Language and Learning. How do Chomsky's models of language as a code, as a switch, and so on relate to the signalmessage processing ideas of Shannon, Wiener, and Turing? How do their ideas of intelligence, game playing machines, problem solving machines, and so on relate to Chomsky's ideas?
(b) Dougherty, 1994, Chapter 7, Computational Toolsfor Language Processing, pp 161-238.
(c) Dougherty, 1994, Chapter 8, Computational Toolsfor Sentence Processing, pp. 239-327.

Second Third Exam Open Book, 2O% of Grade

21. Second Third Exam. There will be two sections, one based on readings and one based on lectures. Each will contain several questions. You must answer two from section I and two from Section 11. The exam is an OPEN BOOK EXAM. You may bring any materials you wish to the exam: books, computers, printouts, notes, articles, etc. See the sample exams from another class. The exam will cover all reading materials listed above from lecture 1 to 20.

Any and all student term papers, computer projects, or HTML projects must be approved by the professor by the 22nd class.

Helen Keller, Alan Turing, and Charles Sanders Peirce

22. Helen Keller and Emanuel Swedenborg. The role of sense information in forming our knowledge of the world. Helen Kellers views on Descarte's cogfto: I think, therefore / am. When Helen Keller was asked how she could know so much given her dual handicaps of deafness and blindness, she responded: How can you know anything, given you are continuously misled by your senses?
(b) Series of three articles from the New Yorker on Genie. In libraries.
(c) Chomsky, Reading to be announced.

23. Alan Turing and his Friend Christopher. We discuss mainly one topic: Alan Turing's views on extra- sensory perception as presented in his Turing test.
(b) Turing, The Turing Test
(c) Hofstadter on Turing
(d) other

24. Charles Sanders Peirce's Views on Human Intelligence. Many people ask: How complex does a computational device have to be before it, perhaps like spontaneous combustion, begins to exhibit intelligence, or think? This question implies that the physical substance (the machine) precedes the thinking. Peirce asked: How complex and organized does thought have to be before it, perhaps spontaneously, concludes that there is a physical world, or at least some thoughts can best be organized by assuming they correspond/correlate with something called a physical world? For Peirce, in the beginning was the word.
(b) Brent, The Wasp in the Bottle. In libraries.

The State of the Art in Computational Models of Language

25. The Evolution of Mind: Darwin. Darwin and Wallace on intelligence and language. The evolution of emotion in man and animals. The Darwinmaterials are available from the http location.
(b) Darwin: The Expression of Emotion in Man and Animals
(c) Darwin and Wallace. 1859. Their original paper is available at the http site.
(d) Lorenz, Konrad. Preface to Darwin's book. In libraries.

26. Chomsky's Minimalist Version of Linguistics. Chomsky's Minimalist Version of Grammar. I- language: A grammar as a mental organ. Language acquistion as parameter setting. Grammar is more like solving aa jigsaw puzzle than like playing chess: parallel processors working on a shared memory space. (a)
(b) Dougherty, 1994, 7.5. The Minimalist Framework, pp. 229-238.
(c) Chomsky, Reading to be announced.

27. Review and Discussion of Major Topics Presented. Review of material for final exam.
(b)The exam will cover all materials (lectures and readings) from lecture I to now.

Final Exam, Open Book, 60% of Grade.The exam is cumulative and covers all material in the course. Term Projects/Papers are Due on Exam Day

The Fine Print: A student may take the final exam, or produce a computer program in Prolog (no other language) that analyzes some aspect of human language, or write a term paper. The final project can be tumed in as a typewritten term paper or placed on-line as a set of HTML pages. Students may pair up and work together on a single project if it is of awesome proponions. All projects must be approved by the professor by the 22nd class meeting.