LINGUIST List 7.1227

Tue Sep 3 1996

Disc: Grammaticalization

Editor for this issue: Anthony Rodrigues Aristar <>


  1. David Pesetsky, Re: 7.1208, Disc: Grammaticalization
  2. Dan Everett, Grammaticalization

Message 1: Re: 7.1208, Disc: Grammaticalization

Date: Sun, 01 Sep 1996 14:29:43 EDT
From: David Pesetsky <pesetskMIT.EDU>
Subject: Re: 7.1208, Disc: Grammaticalization
Perhaps one enters this discussion at one's peril. Part of the problem is
the notion "Chomskyan paradigm", which is a bit vague. Let me just make a
few observations and then leave the discussion to the experts.

I expressed puzzlement at Dahl's remark because it seemed to suggest that a
"Chomskyan paradigm" insists on a uniform explanation (UG) for all
apparently universal facts about grammatical structure. That is not the
Chomskyan paradigm that I know. Indeed, I think it is quite *universally*
acknowledged that facts about grammars -- including interesting and
significant facts -- may have grammar-external explanations.

=46or example, Grimshaw's 1981 paper "Form, Function and the Language
Acquisition Device" [in "The Logical Problem of Language Acquisition",
Baker and McCarthy, eds. MIT Press] argues that a particular gap in
syntactic selection does not reflect a property of UG, but arises as an
artifact of the way in which categorization is acquired by the child.

Likewise, I doubt that anyone would find it implausible that certain facts
about grammar *might* arise from (possibly inevitable) paths taken by
historical change -- or from aspects of language use. If by "Chomskyan"
one refers to Chomsky, there are lots of places in which he's made just
this point. If by "Chomskyan" one refers more generally to a research
community, one can find lots more to support my general point.

Now about the case at hand, Trask writes:

> The study of
> grammaticalization has seemingly revealed powerful regularities in the
> way speakers use languages, and, at the very least, the Chomskyan
> approach simply has nothing to say about these regularities.
> Moreover, it appears difficult to conceive of how a Chomskyan approach
> could ever have anything interesting to say about them, since the very
> concepts and categories required to characterize these regularities
> (or at least many of them) do not exist in Chomskyan linguistics

Let's suppose the facts of the matter are as Trask states. It would be
interesting to know whether one can actually show redundancies or
incompatibilities between the "concepts and categories" mentioned by Trask
and concepts of Chomskyan linguistics -- or whether (I think I am agreeing
with Croft here) distinct bodies of work are examining interacting, but
distinct aspects of language. If the former, then researchers might try to
tease out the truth of the matter.

Others can speak much better than I to the specific historical questions in
dispute here. I don't understand, frankly, the contrast drawn by Trask
between a view of language change as change in a "more-or-less well-defined
grammar" and facts that might indicate both social and individual
gradualness to language change. But in any case, the logically prior
issues are the ones raised by Newmeyer -- is the research on
grammaticization correct. I don't want to deflect attention from that
question, since it comes first.

-David Pesetsky

P.S. In his interesting message, Croft quotes a characterization by Chomsky
('idiosyncratic factors, as contrasted with the more significant reality of
UG' [LGB p.8]) as a possible example of why "the regularity and potential
explanatory significance of grammaticalization is in itself perceived as a
challenge to generative grammar". Actually, the context of that quote
makes quite the opposite point.

Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Grammaticalization

Date: Tue, 03 Sep 1996 12:31:17 EDT
From: Dan Everett <>
Subject: Grammaticalization
With regard to the compatibility between grammaticalization and
Chomskyan grammar, although some have said that grammaticalization is
compatible with Chomskyan theory, others (Bill Croft and Larry Trask,
in particular) claim that if it is, it is just barely so: their claim
is that grammaticalization and Chomskyan theory don't "match", like
orange and red. This is supposed to be because grammaticalization is
performance-based, while Chomsky's model is competence-based, and
because grammaticalization implies that people take the linguistic
sign as a whole, rather than as separate phonological, semantic, and
morphosyntactic bits. I think that they *are* compatible, but that
they are not supposed to *match* (in fact their compatibility relies
on them *not matching*), since they study different kinds of
things. It should not be held against either view that it is unable to
account for phenomena treated by the other. Grammatical theory is
*not* about trying to account for all *linguistic behavior*, to quote
from Larry Trask's otherwise useful discussion.

One thing is clear, Chomsky's theory doesn't have much to say about
grammaticalization. To the degree that one theory should handle all
facts of human language, that is a defect, just as it is a defect for
grammaticalization theories that they are unable to handle concepts
like, inter alia, "minimal domain", the ECP, and spell-out rules, i.e.
concepts that require separate domains of grammatical knowledge,
e.g. phonology, syntax, morphology. But who says or seriously thinks
that any one theory is supposed to handle all the facts of human

What is it that Chomsky's theory is after? What does it say it is
accountable for? These are the principal questions in determining
whether or not it is or is not compatible with findings and proposals
on grammaticalization. The basic research question of Chomskyan
grammar is, it seems to me, "What do people know when they know the
grammar of their language?" (Secondarily and ideally, the theory ought
also to tell us something about how they got this knowledge, hence the
priority placed on language acquisition.) In this regard, people DO
know about phonology, syntax, morphology, and structural semantic
questions *as separate systems of principles* - at least they seem
separate after decades of effort by some of the best linguists in the
field. For example, people's knowledge of underlying phonemes is
independent from their knowledge of case assignment, which is
independent of their knowledge of quantifier scope (to some degree),

Now, it also turns out that people know about *SIGNS* (form-meaning
composites, to put the question in Bill Croft's way) after they have
emerged from these separate knowledge domains, reanalyzing them and
resubmitting them to the grammar in the grammaticalization process,
based on performance errors, stylistic preferences, etc. The question
is whether or not these matters are part of *grammatical* knowledge. I
think that they are not. My reason for saying this is that in the
explanation of grammaticalization, extralinguistic knowledge (or
"factors", to use a more neutral term) is causally implicated, as is
typical of performance-based skills. This is not the case with the
kinds of principles crucial to grammar as conceived by Chomsky,
e.g. those pertaining to extraction domains or, say, to the
relationship between agreement and phrasal projections. There thus
seems to me to be a rather clear epistemological dividing line between
the lines of research represented by generative grammar and
grammaticalization, a line vital to their complementarity.

Setting aside rather wild claims from both sides (attributed dubiously
to various authors), e.g. "there is no grammar, only pieces created by
grammaticalization" or "all historical change is parametric change" or
"all historical change results from innovations in the child's
grammar", it doesn't seem to me that there is any incompatibility at
all between grammaticalization and Chomskyan theory. If those working
in grammaticalization or, say, Minimalist Theory were to claim that
their particular perspective was all that was needed to understand the
interaction of grammatical components in diachrony and synchrony, then
the perspectives would be incompatible for these people. But I must
agree with David Pesetsky in seeing no fundamental conflict once
rhetorical excesses have been discarded. Again, however, the view that
they are complementary is only possible if we see them as separate
domains of inquiry, epistemologically and ontologically. Without that
perspective, then they may indeed seem contradictory.

Dan Everett
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue