Is grammar derivable?

George Lakoff lakoff at COGSCI.BERKELEY.EDU
Sat Mar 6 01:06:02 UTC 1999


At 9:46 AM +0100 3/5/99, Östen Dahl wrote:
>In a recent Swedish introduction to generative grammar, it is said that
>generative grammar, in the miminalist version, postulates an "internal
>grammar", acting as an independent cognitive module. The author then goes on
>(my translation): "There are other current theories of grammar that do not
>assume an independent internal grammar, such as functional grammar,
>according to which grammar is derivable from language use..., and cognitive
>semantics, according to which grammar is derivable from meaning...."
>
>It would be interesting to know whether FUNKNET subscribers agree with these
>characterizations.
>
>
>Östen Dahl

Dear Östen,

That is, as I'm sure you know, a thoroughly inaccurate statement. A more
accurate statement would be:

Cognitive-functional grammar claims, on empirical grounds, that principles
of grammar commonly make reference to, and are often motivated by, aspects
of communicative function in context and embodied meaning in context (as
characterized via embodied cognition).

Here "motivated" is used as characterized in Women, Fire, and Dangerous
Things.
Motivation is neither derivability nor arbritrariness.

That oversimplified statement, of course, does not say a whole lot. Reading
the literature of both fields is required to get a sophisticated idea of
what all this is about. For example, in the Neural Theory of Language, a
great deal of language makes use of neural mechanisms external to language
and neural mechanisms in general.

"As characterized via embodied cognition" makes reference to a huge
literature on cognitive semantics -- image-schemas, metaphors, blends,
force-dynamics, frames, X-schemas, radial categories, types of prototypes,
and so on.

For a discussion see Case study 3 in Women Fire and Chapter 22 in
Philosophy in the Flesh, as well as Volume 1 of Langacker's Foundations of
Cognitive Grammar.


The statement you quoted is characteristic of the sort of innacurate things
Chomskyans used to say about generative semantics.

Best wishes,

George



More information about the Funknet mailing list