Adam's message

[Carl Pollard]

Hi Shalom,

It seems that the usage of the term "generative grammar" is divided
between (a) a mathematical system for specifying a natural language,
or aspects of one (strings, structural representations, semantic
interpretations, or some combination thereof), and (b) a colloquial
synonym for "Chomskyan linguistics". Neither usage is slanted more
toward generation than parsing/interpretation (at least, I don't think
so). So my inclination is to stick with the term "generative grammar"
under an (a)-ish usage.

The term "formal grammar" strikes me as more general, not necessarily
connected withy natural language.

Best regards for the millenium,

Carl

>
Hi Bob,
   Many thanks for the illuminating passage on generative grammar from
Chomsky's early paper. It is interesting to see that the term was
originally used to denote the class of formal grammars, and now,
apparently, has been appropriated by Neil Smith to refer to
transformational grammar. Computational linguists tend to restrict
generation to the procedures involved in actually constructing a
string with associated syntactic structure and to contrast this process
with parsing. Therefore, it might promote clarity of discussion if we
simply used the expression 'formal grammar' to describe a grammar or
linguistic theory that is intended to provide a precise mathematical
model of the formal properties of a natural language (or of the set of
possible natural languages). We can then ask whether Chomsky's
Government Binding and Minimalist frameworks are alternatives to formal
grammar. Happy new year. With best regards.
                               Shalom
>