Question about functional approaches

George Lakoff lakoff at COGSCI.BERKELEY.EDU
Thu Jul 3 09:17:14 UTC 1997


As a cognitivist who shares many functionalist goals, my answer will be
different than
functionalists who are not cognitivists.

>        As I understand it, the functionalist <--> formalist dimension is
>orthogonal to the innate (language-specific) knowledge <--> non-innate
>(general cognitive) knowledge dimension. Presumably one could have:
>(1)     highly functional syntax as well as innate language-specific
>        learning principles;
>(2)     highly functional syntax and no language-specific learning
>        principles;
>(3)     "dysfunctional" syntax as well as innate language-specific
>        learning principles;
>(4)     "dysfunctional" syntax and no language-specific learning
>        principles;

The cognitive response is NONE OF THE ABOVE. The reason is that in the
rapidly evolving Neural Theory of Language, much of syntax uses semantics
and much of linguistic semantics uses embodied structures from the
sensory-motor system.
All innateness issues get blurred here, since the sensory-motor/linguistics
no longer exists and since it is impossible to disentangle exactly what is
and is not there
in the sensory-motor system at birth, in the womb, and as a result of
genetics (which underdetermines neural structure). From a neural/cognitive
persopactive, all these issues look different. Things to read: Terry
Regier's book THE HUMAN SEMANTIC POTENTIAL (MIT Press, 1996), and the
papers, reports and dissertations on the
website at In this tradition, the poverty of the
stimulus is replaced by the richness of the substrate.

>        As befits our East Polar status, weare trained here in generative
>grammatical theory, but some of us have been trying to learn more about
>functional approaches, especially to syntax, since syntax so readily leads
>to debates about innateness of linguistic knowledge. There is a whole
>standard set of problems in generative syntax, whose ultimate explanation
>points towards innateness claims, via poverty of the stimulus, etc. Some of
>these include that-trace effects, subjacency violations, ECP effects,
>wh-islands, what used to be the SSC and NIC, the Head Movement Constraint,
>and so on. Can anyone out there suggest places we could look to see how a
>functionalist approach would answer questions of this type? Do
>functionalist theories of syntax have as a goal this sort of investigation?

Two obvious places to look:

For a Langacker-tradition approach to anaphora, try Karen Van Hoek's thesis
(soon to be published by Chicago) and her Language article.

For a NTL approach to syntactic constraints using Shastri's theory of
temporal binding for structured connectionist model, look at Jamie
Henderson's thesis from Penn a few years back.

>        Having been trained in formalist and nativist theories of syntax,
>we are curious to see whether there is any cross-talk between these two
>areas. Do functionalists consider the questions that formalists ask worth

Cognitive-finctionalists do not find the questions at all interesting,
since they make
assumptions contrary to the evidence from cognitive science and neurosceince.

Is there an entirely different approach and entirelyt different
>issues and questions to be answered that leave aside the traditional
>concerns of generative theories?

Yes. There is very extensive literature. The whole field of cognitive
linguistics as well as
the functionalist literature.

>        If anyone out there can suggest places we can start looking for
>some ways in which similar linguistic data can be viewed through the lenses
>of both theoretical appraoches, we would be happy. If there are really no
>or very few problems common to both approaches, this is interesting (and

There does not seem to be much overlap.

>        If I have gone of the rails here, please let me know where and how
>(why, too, if you feel combative!). If anyone has a response to these or
>related questions, I am happy to take responses by personal e-mail, and
>provide a summary. Your anonymity will be respected if you ask me to do so.
>Thanks in advance,
>David Parkinson
>dp11 at

I'm glad you're interested. There's a whole universe of great linguistics
outside of generative linguistics. It's developing in an incredibly
exciting way.

Best wishes,

George Lakoff

More information about the Funknet mailing list