LINGUIST List 4.688

Sat 11 Sep 1993

Disc: Ross and The Linguistic Wars

Editor for this issue: <>

Directory

1. Helen Dry, Ross
2. , 4.671 The Linguistic Wars
3. ALICE DAVISON, Ross and syntactic theory

I've received numerous messages--from linguists of all
ages and persuasions--agreeing with Harris (and me)
about the theoretical significance of Ross's work.
Most of the messages were personal messages sent to me
rather than LINGUIST; but many made points that may be
of general interest, so I'll summarize and quote a few
below.

Almost everyone deplored the tendency of new schools of
thought to ignore or fail to appreciate previous work-
some singling out GB (but admitting that they too were
guilty of the same thing in their days at MIT, or UCLA,
or wherever ...)

Several suggested that the van Riemsdijk & Williams text
itself is primarily responsible for the opinion
that Ross's work is just "descriptive." I was
directed to John Goldsmith's incisive review of van
Riemsdijk & Williams (_Language_, March 1989), which
discusses this point at length:

"...when vR&W do turn to Ross's positive proposals,
there is a strong suggestion in their wording that
Ross's proposals were based only on English, and
intended as language-particular properties. . . . All
syntacticians know, of course, that Ross's thesis was
the first (and, at the time, mind-blowing) massively
crosslinguistic study of an abstract grammatical
property, and his conclusions were stated at the level
of theory, not that of language-particular property."
 (Goldsmith, 1989: 151)

Everyone--no dissenters--asserted that Ross's work was
theoretical. To quote a few:

"Ross deserves credit for discovering these constraints,
though later GB tries to explain them in terms of
bounding nodes and/or Barriers. So his stuff wasn't
fully translated into Universal Principles and
Parameters and all of that, but it certainly wasn't
"just descriptive". Rather it was theoretical, but not
reduced to atomic principles as the later GB versions
have been (which don't work all that well anyway)." -
anonymous

"If you look at the literature of the first 20 years of
generative theory, the only person more quoted than Ross
is Chomsky, and I suspect he contributed more innovative
ideas to thinking about syntax than did Chomsky. Of
course he was working within a larger theory, it was TG,
then revised ST, etc. No, it wasn't the now-current
universal thing (which I 'm not belittling by that
comment) but it certainly was a larger theory than just
English. . . .
 Incidently, I think if you look at the issues before
GB and its progeny, you will find almost exactly the
same issues that TG was struggling with 20 years ago."
 --Bruce Fraser

"...in my dissertation i note that the significance of
Ross' dissertation for the history of Chomskyan theory
is that it signals a heuristic shift from seeking to
indentify and describe transformations to trying to
constrain their power. To my mind, this is TERRIBLY
IMPORTANT, since the whole agenda of EST, REST, GB, PPA,
call it what you will, has since the early 70's been to
constrain the power of the transformational component,
ultimately boiled down to a single, generic
transformation 'Move Alpha'. So what if Ross' work was
'merely descriptive'? God lives in the details!"
 --Steven Schaufele

My thanks to all who replied.

-Helen

I would bet dollars to donuts that the explanation for the
dismissal of Ross as ``merely descriptive'' is two-fold:
(1) general historical ignorance, which is pandemic among
linguists. (I'll bet real money that few recent syntacticans have
actually read Ross; rather they've gotten their impressions of him
only second hand. In fact, I'd bet pocket-change that many haven't
even read _Aspects_; when I was at MIT, around 1980 and 1981, you'd hear
people saying dumb things like ``N, V, and A is not a natural class'' which
displays what would be regarded as, in a more reasonable field, a
massive ignorance of previous theory.); (2) a difference in
perspective: when Ross was writing, coming up with general constraints
on transformations, since there were various particular
transformations, WAS considered theoretical work. Now, such
constraints are always viewed as by-products of more general
principles, so ``just'' coming up with a new island configuration,
say, is considered as ``merely descriptive''. But, within its
context, Ross' work clearly was theoretical. And to not acknowledge
that, by imposing much later ``standards'' is as historically myopic
as criticizing George Washington for doing nothing to limit nuclear
proliferation.

As for my evaluation of Ross' work, I believe you are correct. When I
entered MIT, Ross's dissertation was still something everybody felt
compelled to read, simply because it had started the then-current
trend towards general constraints in syntax, which was culminating in
the ``constraints on transformations'' framework and its descendants.
You can view the relation between Ross' thesis and Chomsky's work as a
kind of elaborate waltz: Ross' dissertation can be seen as an attempt
to rectify problems with the applicability of the A-over-A constraint
by decomposing it into a number of more particular constraints.
``Constraints on Transformations'' can be seen as a counter-move to
split up the theoretical pie differently, by reducing the currently
known island effects to the interaction of three general principles:
Tensed-S, (Specified) Subject, and Subjacency.

``merely descriptive'', indeed! To paraphrase Frege: ``How can
anybody spout such hair-raising nonsense?''

-30-
Bob Ingria

I'm quite surprised at the comments on Ross's work. The
thesis is still being mined for data which other people
can theorize about, and the constraints which Ross first
articulated have STILL not been accounted for in current
theories at least not in a clearly more general and
more insightful way. (Ie subjacency accounts for Complex
NP and wh- island, but not Sentential Subject, without
special assumptions, nor Coordinate Structure). Your
syntacticians may have meant that one can know Ross' work
and not already know much about GB, though GB as a theory
is meant to account for Ross' facts (which remain as
central issues). Also I suspect there is discomfort
about Ross and his departure from MIT, which is in-
dependent of the great importance of Ross' dissertation.
Actually the thesis is hardly descriptive, as the ungram-
matical sentences it accounts for were discovered only
as a consequence of 1960s generative theories of syntax.
Island violations tend NOT to be produced spontaneously
but rather come to light when one asks why movement rules
are notabsolutely general.
I wonder what orhers think! I'm looking forward to
reading the book.