LINGUIST List 4.722

Sat 18 Sep 1993

Disc: WARS and constraints: author's response

Editor for this issue: <>

Directory

1. Randy Allen Harris, Ross, Constraints & The Linguistic Wars (author's response)

I'm very pleased to see the discussion generated by _The Linguistics Wars_
(LW), particularly around the importance of Ross. His reputation has
suffered tremendously since the interpretive-generative wars, and very much
deserves the positive attention that Dry, Fraser, Schaufele, Ingria,
Davison and Anonymous, are giving it (LINGUIST 4.688). Goldsmith also
deserves credit here, special credit, for the 1989 review from which Dry
quotes, but I was also quite interested in Rohrbacher's remarks that
Williams's regard for Ross is higher in the classroom than it is in his
introductory text with van Riemsdijk.

I would, however, like to address a few issues around the messy notions of
constraints and restrictiveness that Pauline Jacobson raised recently
(LINGUIST 4.689). Before I do, though, there are three points of context
for her remarks which are particularly important.

First, please note how reluctant (like Murray before her) she is to beat up
on the book:

 >I don't mean to trash Harris' book since I liked a lot about it.
 >However, ...

I appreciate the good will and the subordinate compliment.

Second, please note that her beef is more clearly with others than with me
or LW:

 >people are constantly proposing constraints on the class of
 >possible grammars without worrying about whether these
 >constraints may be vacuous.) The Peters and Ritchie proof ...

The bulk of her posting is about such people and situations, not about LW.

Third, and most importantly, please note that her reading of the book was
not comprehensive:

 >Having looked through (just part of) Harris' book (THE
 >LINGUISTICS WARS), and having read some of the comments on
 >Linguist ...

Consequently, she is not entirely fair to the book, implying oversights or
errors that in fact don't exist.

Jacobson's post is dense and interesting, and I'm sure it will stir others
to discuss the theoretical fine points of the tangled semantic growth
around terms like "constraint" and "restrictiveness". I will confine my
remarks to the places where my book is actually invoked.

It is invoked with respect to the following, very useful, three-way
taxonomy of the usage of "constraint" by generative linguists:

1. Constraining the set of languages a grammar can describe.
2. Constraining the set of grammars that can do the describing.
3. Constraining what transformations can do.

Most everybody, Jacobson says, conflates these three usages in various
ways, and she takes LW to be symptomatic, making

 >this conflation several times in [its] discussion on pp. 180-
 >181. The gist of the discussion is that Generative Semantics did
 >indeed worry about restricting the power of the grammar [in the
 >first sense] since it paid a lot of attention to constraints [in the
 >third sense].

It's true that my discussion is not as scrupulous about the various uses of
"constraint" as Jacobson's post is. But what she identifies as the gist of
my discussion is not exactly the gist of my discussion, mostly because she
completely ignores the historical context (the early seventies). More
specifically, the discussion concerns the charge of descriptive profligacy
that was recurrently flung at generative semantics, beginning with
Chomsky's observation that "the gravest defect of the theory of
transformational grammar is its enormous latitude and descriptive power
(1972 [1969]:125; quoted on page 179 of LW). This grave defect was part
and parcel of all transformational models, but the most "uninteresting" of
such models, the early form of generative semantics, became "still more
uninteresting by permitting still further latitude, for example, by
allowing rules other than transformations that can be used to constrain
derivations" (1972 [1969]:126, quoted on p.179 of LW).

The issues were even less clear then than now, but the focus of the charge
from Chomsky was that (early) generative semantics was bad because it had
only transformations and PS rules; further, that it was made even worse
when it added other descriptive devices, the dreaded global rules. Adding
global rules, he argued, made a theory which described too many languages
(or rather, too many things, only some of which we would want to call
languages). The proper way to go, Chomsky said, was "to restrict the
category of admissible phrase markers, admissible transformations, and
admissible derivations" (1972 [1969]:125). What Chomsky sidestepped, and
what seemed to be missed in all the ensuing brouhaha, is that generative
semanticists had contributed, and were contributing, toward exactly these
ends. Certainly Lakoff's call for a "global grammar" (1972) included such
restrictions. For instance, rule ordering, something not usually
associated with globality, is the first example of global derivational
constraints that Lakoff cites (1970a [1969]:234).

Rule ordering is a useful example with respect to Jacobson's taxonomy as
well, since it clearly belongs in her first two categories, and her
discussion implies (perhaps unintentionally) that generative semantics was
not interested in such constraints. The cycle, something which
particularly concerned Lakoff and Ross early on, is another such example.
So is Ross's gapping work. So is McCawley's work on the base, which
developed, on the one hand, toward the universal base hypothesis, and, on
the other, toward Emonds's structure-preserving constraint. Indeed, the
Katz-Postal principle (the first obvious step toward generative semantics)
is a category 2 constraint (and maybe a category 1 constraint).

I am guilty of not untangling the various usages of "constraint" that
Jacobson catalogues, and my primary example in the discussion is Ross's
island work (from category 3), since it proved massively the most
influential research of the period into constraints of any category. But
the general point, I think, stands: that Chomsky's assault on the notion
of globality, and his complementary invocation of restrictiveness,
pointedly ignored much of what generative semantics was up to. By "global
rule" Chomsky meant "any rule imaginable" (1972 [1969]:133). That was
decidedly not what the generative semanticists meant by the term (see, in
particular, McCawley's remarks in Parret, 1974 [1972]:268).

(It might also be worth pointing out here that--retroactively adopting
Jacobson's terminology--I do draw attention in the book to the fact that
positing a category 1 or category 2 constraint does not necessarily
constrain the grammar, in the category 3 sense. See LW 293n29. Jakobson
implies that the book is not aware of this situation, though, again,
perhaps inadvertently.)

After rebuking me for my confusion, Jakobson generously suggests a better
tack for the book to take over the issue of restrictiveness:

 >I think a more appropriate response to the attacks on Generative
 >Semantics being "unrestrictive" would be to note that it had never
 >been shown that Interpretive Semantics was any more
 >"restrictive" (in the first sense).

This would indeed be an appropriate response, and, in fact, *was* an
appropriate response, one of the most common the generative semanticists
offered. They regularly complained that Chomsky was cynically disguising
just how powerful his model was. Here is Postal, as quoted in the book,
just a few pages before the section that Jacobson jumps on:

 Chomsky had these what did George [Lakoff] call them? these
 wild cards that he could pull out of his hat whenever he wanted
 and somehow they didn't count when it came to talking about
 restrictiveness.

 Whenever he was doing something descriptive, where he needed to
 describe facts that generative semantics would talk about in terms
 of transformations linking meanings to deep structures, or to
 other kinds of structures, by way of global rules Chomsky would
 appeal to semantic interpretation rules. (LW, 178)

In fact, they claimed that interpretive semantics might be *less
restrictive* than generative semantics, because of its regular appeal to
"overly powerful devices" like syntactic features and semantic
interpretation rules (Postal, 1972:215; quoted in LW 181).

Jacobson has another suggestion for a response to Chomsky's argument:

 >Nor was it shown that adding things like global rules in any way
 >increased the power of the theory in the sense of allowing for
 >more possible languages. (Indeed, in light of the Peters and
 >Ritchie results, one might argue that adding global rules and other
 >devices to the theory couldn't possibly make it "more powerful"
 >since it already was all-powerful.)

I agree, though, of course, Chomsky insists that the Peters-Ritchie results
have nothing interesting to say about restrictiveness.

_________
(P.S. For anyone following the various threads of the book discussion
closely, my apologies for not addressing publicly Rick Wojcik's follow-up
question to my exchange with Murray--"What makes you think Sapir was in a
position to buck the positivist tide?"--in LINGUIST 4.658. I have traded a
few subsequent notes with Wojcik on the topic and, since I don't assert
anything to the contrary in my book, and since Sapir is not a big part of
my story, Wojcik has agreed with my decision to treat the question as
rhetorical, and we are both waiting to see if anyone else wants to engage
it. Any takers?)

(P.P.S. If the contrastive stress in her posted willingness in LINGUIST
4.703 to explain "what life was REALLY like during the linguistic wars" is
meant to imply that LW is way off the mark, I invite Vicki Fromkin to
explain how. If not, not.)

_________
References

Chomsky, Noam. 1972 [1967-69]. _Studies on semantics in generative
grammar_ . The Hague: Mouton.

Lakoff, George. 1970a [1969]. On generative grammar. In _Semantics: An
interdisciplinary reader_. Edited by Danny Steinberg and Leon Jakobovits.
Cambridge: CUP.

Lakoff, George. 1970b [1969]. Global rules. _Language_ 46.

McCawley, James D. 1968. Concerning the base component of a
transformational grammar. _Foundations of language_ 4.

Parret, Herman. 1974 [1972]. _Discussing language. The Hague: Mouton.

Postal, Paul M. 1972. On some rules that are not successive cyclic.
_Linguistic inquiry_ 3.

Ross, John Robert. 1970 [1967]. Gapping and the order of constituents.
_Progress in linguistics: Its development, methods and problems_. Edited
by Manfred Bierwisch and Karl Heidolph. The Hague: Mouton.

Randy Allen Harris rahawatarts.uwaterloo.ca
Rhetoric and Professional Writing 519 885-1211, x5362
English, U of Waterloo FAX: 519 884-8995
Waterloo ON, CANADA, N2L 3G1