good news from generative grammarians

Robert Levine levine at ling.ohio-state.edu
Mon May 14 13:38:30 UTC 2001


Re KP Mohanan's posting:

The term `generative' entered the lexicon of linguistic theory with a
well-defined technical meaning, referring to explicit enumeration of
certain sets, in opposition to a taxonomic classification of data
which treated languages as finite corpora. Please note the following:

    a *generative grammar* (that is, an explicit grammar that makes no
    appeal to the reader's `faculte de langage' but rather attempts to
    incorporate the mechanisms of this faculty) is a system of rules
    that relate signals to semantic interpretations of these signals.

(TOPICS IN THE THEORY OF GENERATIVE GRAMMAR, 12.) Note also from the
preface to LSLT:


    a grammar constructed in accord with the principles postulated in
    such a theory [of generative grammar] gives and explicit
    characterization of a language and its structure---and within the
    broader semiotic theory envisioned but not developed here, an
    explicit characterization as well of the meaning and reference of
    expressions and conditions of appropriate use.

I have no idea what your basis for saying that `Most people associate
'generative" with A2'; certainly most people who pay attention to the
meaning of technical vocabulary do not. The fact is that the
well-defined technical meaning of `generative'---in the sense that
Chomsky introduced and has never (as vs. some of his accolytes)
disputed---has become distorted as part of a deliberate attempt to
give a priviledged position to one particular, rather vaguely
formulated story about the human language capacity. Do you believe
that this has any more relevance than, say, the Stalinists'
application of the term `(social) facist' to the social democrats
during the Spanish civil war, so far as the meaning of the term
`facist' is concerned? The ideologically-driven expropriation of the
term `generative' to refer exclusively to whatever current MIT-orbit
linguistics is doing at the momeent is, essentially, restricted to
people working in that orbit. So when you say `most people
associate...', you are at best talking about most people doing P&P or
whatever. And these are just the people whose distortion of the
original meaning serves a particular agenda which they seek to
promote. Exactly how does this practice, which reflects a certain kind
of intellectual dishonest driven by sectarian rather than intellectual
motivation, become legitimate, as you apparently regard it? If a
number of women involved in right-wing politics start referring to the
Republican caucus position on women's issues as `feminism', and claim
that this is the correct definition of `feminism', does this mean that
we now have to take seriously that there are now two different
meanings of the term `feminism' and that `most people' use the term in
the sense `right-wing position on women's issues'? This is very close
to what you seem to be accepting as a legimate use of terminology. The
manufacture of consent goes on unabated, apparently, with it now
becoming legitimate to introduce any politically motivated revision of
mathematically explicit vocabulary as long as one can claim that `most
people' do it.

As far as your distinction between `model', `framework' and `theory'
is concerned, you seem to be saying that as long as I ascribe a claim
of universal laws. Fine, I claim that the Head Feature Principle, the
Valence Principle and Slash Amalgamation are all universal. Now we
have theory, not just a model or a framework. But I can't really say
whether these are distinct from minimalist `laws/constraints', since
these have yet to be formulated explicitly enough that anyone knows
exactly what their denotation is. So how does anyone know just what is
predicted by any version of minimalism? And if it is impossible to
tell just what follows from, say, Chomsky (1995) in the case of a vast
variety of phenomena that the MP simply has no story for, in what
useful sense of the word is the MP a *theory* of anything??

Bob Levine



More information about the HPSG-L mailing list