Summary of Steven Pinker's (1999) book "Words and Rules"

Aim: to illuminate the nature of language and mind by focusing on the study of regular and irregular verbs.

Chapters:

  1. The Infinite Library - generative grammars and the distinction between lexicon and grammar
  2. Dissection by Linguistics - all about regular verbs
  3. Broken Telephone - all about irregular verbs
  4. In Single Combat - Chomsky and Halle versus Rumelhart and McClelland
  5. Word Nerds - evidence from psycholinguistics - language production and recognition
  6. Of Mice and Men - evidence from neologisms
  7. Kids say the Darnedest Things - evidence from language acquisition
  8. The Horrors of the German Language - evidence from other languages
  9. The Black Box - all about brains
  10. A Digital Mind in an Analog World - how this all links up to other mental faculties

I think this book is great. If it were written in textbook-style rather than popular-science-style, it would be perfect. It covers a lot of stuff - associative memory, cross-modal priming, rationalism versus empiricism, naturalism versus conventionalism, brain scans, connectionism versus symbolic processing, and on and on. Better still, it really focuses on the details, which is great for this course - showing students that something simple that they already know unconsciously offers a huge amount of potential for conscious study. It's relatively unbiased, illustrating why a theory of mind needs both symbolic and connectionist mechanisms.

In addition, it covers a lots of tangential "linguistic literacy" topics that educated adults should know about but as often as not don't, like historical linguistics, the inevitability of language change, the lack of obvious clear-cut distinctions between "correct" and "incorrect" English (gradient grammaticality).

Here is a six-page summary of the book. And a critical review.

The obvious way of approaching this topic would be (following the book):

  1. Basic data on regular versus irregular verb inflection
  2. The simple words-and-rules model (generative grammars, symbolic processing)
  3. More data on the semi-systematic nature of irregular verb inflection
  4. The Chomsky and Halle model and its pros and cons (rationalism)
  5. The Rumelhart and McClelland model and its pros and cons (empiricism, neural networks, associative memory)
  6. Detailed data illustrating why we need both: (a) language processing, (b) language acquisition, (c) ...
  7. More recent research/approaches?

In other words, explore the phenomena first, then describe and evaluate the models.

If we have 13 lectures, how might this play out:

1 for baby linguistics in general

1 for morphology in particular

2 for 1, 2 & 3 above

1 for 4

1 for neural nets

1 for 5

1 for 6

1 guest (Richard Shillcock, maybe) for 7

That's 9, leaves us 3 for side-topics we need for labs (e.g. probabilty/significance)

-- HenrySThompson - 02 Sep 2010

Topic revision: r3 - 02 Sep 2010 - 12:28:32 - HenrySThompson
 
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback
This Wiki uses Cookies