nl-kr-request@CS.ROCHESTER.EDU (NL-KR Moderator Brad Miller) (02/02/88)
NL-KR Digest (2/01/88 23:55:15) Volume 4 Number 13
Today's Topics:
Re: failure of TM
Gross's article
the Gospel according to Newmeyer
the imaginary worlds of grammarians
Submissions: NL-KR@CS.ROCHESTER.EDU
Requests, policy: NL-KR-REQUEST@CS.ROCHESTER.EDU
----------------------------------------------------------------------
Date: Mon, 25 Jan 88 17:46 EST
From: Rick Wojcik <rwojcik@bcsaic.UUCP>
Subject: Re: failure of TM
I want to thank Bill Poser for taking the time to respond at length to
my earlier defense of Peril Slob's parody. I don't think I missed your
point, Bill. Note that I never attacked generative grammar or defended
Maurice Gross' paper. I merely said that his paper deserved a response.
I do not believe that all generative grammarians avoid discussing
issues, but there is a pernicious attitude in the field that generative
grammar can afford to rest on its laurels. You yourself seem quite
ambivalent about whether or not generativists should ignore criticisms.
You bring up many examples of earlier exchanges between generativists
and their critics. What you don't seem to recognize is that the debate
will never end. And it shouldn't. There is a constant flow of
newcomers to the field who need to be exposed to the same old 'tired'
arguments. The foe is never vanquished. Jim Higginbothm once put it
very nicely to a conference on formal syntax when he pointed out that we
are all wrong in the long run. The foundations crumble, and they get
replaced. So you need to keep shoring up the foundations.
Thanks for pointing out Newmeyer's response to Gross. That is a much
better way to handle this matter than to say that the paper doesn't
merit a response. Yet you still defend this arrogant nonsense that
generative grammarians can ignore frontal attacks published in a major
journal like Language. We are not talking about Voprosy Jazykoznanija,
but about the linguistic journal with the widest readership in the
world. Look at your wording:
>Back in the days when generative grammar was new and
>it was necessary to win people over, it was necessary to defend generative
>grammar, and this was done. Now that generative grammar is by far the dominant
>school among theoreticians (though arguably not among all who call themselves
>linguists) it isn't necessary. Maybe this is arrogant, maybe not. You don't
>see physicists responding to every crank attack on relativity. Biologists
>...
>I think that generative grammarians feel that the relevant
>issues have already been debated and that now that they have won the field
>it isn't worth their while to respond all the time. When new challenges
Articles like the Gross paper don't appear in every issue of Language, and
someone could surely spend the time to take the editorial board to task
for permitting nonsense. Physicists have more respect for their
colleagues than to let their journals get away with irresponsible clap.
The lack of a response suggests widespread approval and/or the inability
to respond. You said that Language should not be taken as an authority.
I agree. It should be taken as a forum for discussion.
Here is a passage that puzzles me greatly:
> Let me add that I don't entirely agree with the strategy of not
>replying to criticism. One reason is that it is unfortunate for incorrect
>views to spread, even if they have no real intellectual impact. A second
>is that some people may honestly not know why what they are saying is silly
>and will feel that it is unfair if they do not receive a response. So
>I think that it is generally a good idea for people to publish replies to
>criticism. On the other hand, I have to agree that it often isn't worth it
>and that it is better to get on with real work.
Your last sentence seems totally at odds with the rest of the paragraph.
You can't have it both ways, you know. In any case, I agree with most
of the paragraph, and I would add another reason to disagree with the
attitude that criticism can be stonewalled. You can't really defend
ideas by refusing to let them be challenged. You end up playing
power politics--intellectual king of the hill--and you lose touch with
the rationality that allowed your own ideas to gain credence in the first
place.
I have not tried to defend the Gross article here, although I think that
you have sorely misrepresented it in your discussion. Gross did not
defend traditional grammar, for example. He only lamented the narrow
scope of generative endeavors in comparison to it. But the real issue
is not Gross, it is the strange attitude that generative grammar doesn't
need a defense. (Actually, you didn't really say this--you seemed to
say that it does need a defense, but only when you think it needs a
defense, which is to say not often, and surely not in this case.) I
don't always agree with the editorial board of Language, but they are
not linguistic mountebanks. Gross intended an honest intellectual
challenge, and he got the intellectual finger in return.
===========
Rick Wojcik rwojcik@boeing.com
------------------------------
Date: Wed, 27 Jan 88 18:25 EST
From: Bill Poser <poser@russell.STANFORD.EDU>
Subject: Re: failure of TM
Rick,
I am curious as to how I might have misrepresented Gross, at least on
his main points. That would be pretty tough unless you believe that he
wrote a misleading abstract for his own paper. As for the one point
you mention, that I mis-represent Gross as advocating traditional grammar,
I don't think I misrepresented him at all. To begin with, I didn't say
exactly that. What I said was that Gross compared generative grammar
unfavorably with traditional grammar in claiming that generative grammar
did not lead to grammars as comprehensive as traditional ones, and that
Gross considers construction of comprehensive grammars a virtue. Both of
these points are quite explicit in the body of Gross' paper as well as
the abstract. I did not say that Gross actually advocated traditional
grammar.
The apparent contradiction that Rick sees in my views on whether
generative grammarians should respond to criticism is easily resolved.
In an ideal world everybody would know all of the literature and would
have infinite time in which to reply to everything with which he or she
disagreed. Under such circumstances there can be no doubt that it would
be best to reply to every criticism. In practice, however, we don't have
time to do all that, and so have to be selective. In deciding what to
reply to, two criteria come to mind. First, reply when a significant
issue is raised, since in this case debating the issue is most likely
to lead to advances in knowledge and since the propagation of a dubious
view is most likely to set people off on the wrong track. Second, reply
when there is danger that failure to reply will have unfortunate
political consequences, either in the broader sense of impact on society
or in the narrower sense of effects on grants, faculty appointments, and
so forth. For the most part, replies to crank literature are not motivated
under either criterion.
In the particular case of Gross' paper, I would classify it as marginal.
I think that its argument is so confused that it doesn't really require
response under the first criterion, though it is better than Ian
Robinson's book, for example. Under the second criterion, however,
it arguably does. Although _Language_ is not one of the top theoretical
journals, it is widely read and so articles in it can have a big impact.
So if I had been more advanced in my career in 1980 (and of course, if
I thought then as I do now) I might well have replied to it.
I note, however, that whether or not someone should have replied
does not bear on what inferences can be drawn from failure to reply.
It is not legitimate to infer approval or inability to reply.
(By the way, this raises a pet peeve of mine, which is that some journals,
_Language_ among them, do not accept letters to the editor or other short
communications. In the case of something like Gross' paper, this is probably
the most suitable response, and of course if it is possible to present a
short reply someone is more likely to take the trouble to write one.)
There is one major issue on which I was disappointed to see no response.
I denied that in general generative grammarians had a greater tendency
than anyone else to refuse to reply to criticism. Where are all the
examples of the criticism to which generative grammarians have failed
to respond? Is Gross's paper the only example? I maintain that the serious
issues are being discussed all the time. One example I gave in my
previous message was the connectionist assault on generative views of rules
and representations in morphology, to which Pinker and Prince have replied.
You seem to think that it was only in the early days that generative
grammarians defended their ideas, but even the examples I already
gave, including this one belie that. Fritz Newmeyer's book (1983) hardly
dates back to the beginnings. Even Georgia Green's review of Ian Robinson's
book (1977) is not that long ago.
Recent examples are easily multiplied.
One example is to be found in a book I have sitting on my desk,
entitled _Modularity in Knowledge Representation and Natural Language
Understanding_, edited by Jay L. Garfield (MIT Press, 1987). Modularity, and
associated issues of the autonomy of grammar, syntax, etc., is controversial
and associated with generative grammar. Here's a book containing both
attacks on this position (e.g. the chapter entitled "Against Modularity"
by Marslen-Wilson and Tyler) and defenses by such people as Fodor,
Higginbotham, and Frazier.
Here's another example. The status of statistical generalizations about
surface word order has been a bone of contention between main-stream
generative grammarians and Greenbergian typologists. _Language_ 60.55-69
(1984) contains a paper by Peter Coopmans, "Surface Word-Order Typology
and Universal Grammar" attacking the Greenbergian approach.
Here's still another. Langendoen & Postal's book _The Vastness of Natural
Languages_ argues that natural languages cannot even in principle
be described by generative grammars because they are not recursively
enumerable. _Language_ 62.154-156 (1986) contains a review by Barbara
Abbott.
Where is the refusal to debate the issues?
Let's not get bogged down on the question of under what circumstances a reply
is warranted. I agree that ideally one should reply to everything.
My point has to do with (a) why, whether rightly or wrongly, people don't
always reply; (b) what are reasonable criteria to apply given that one has to
allocate scarce time and energy. But that is not the central issue. The
central issue is whether it is even true that generative grammarians are
worse than other people in failing to reply to criticism and defend their
ideas. I have denied that this is true, and so far all we have is the
single example of Gross' paper. On what basis do you claim that generative
grammarians generally refuse to defend themselves? What makes you think
that anybody else is better?
Bill
------------------------------
Date: Thu, 28 Jan 88 11:38 EST
From: Alan Lovejoy <alan@pdn.UUCP>
Subject: Re: failure of TM
In article <1860@russell.STANFORD.EDU> poser@russell.UUCP (Bill Poser) writes:
> Consider first Gross' paper. In an attempt to give a fair
>summary I will quote the abstract of this paper, written by Gross,
>in its entirety:
>
> An attempt to construct a generative grammar of French
> with a coverage comparable to that of traditional grammars
> has failed. A description has been arrived at in the
> course of this work, however; it is much more complex
> than expected, and turns out to be entirely taxonomic.
> This result calls into question the validity of the
> so-called theory of generative grammar.
>
>In other words, because generative grammarians have thus far not
>produced comprehensive generative grammars, the theory must be wrong.
>This is a non-sequitur. Generative grammar does not purport to
>provide an especially efficient means of constructing grammars.
>Generative grammarians have devoted their efforts primarily to
>constructing theories, to testing them, and to deep studies of
>particular phenomena that they hope will elucidate the nature of
>language in general. Since they have not tried to construct
>comprehensive grammars and since the theory makes no claim of
>efficiency at doing so, how can the fact that generative grammarians
>have not constructed such grammars bear on the validity of the theory?
There is a fundamental and fatal misunderstanding of just what
generative grammar, formal language theory, mathematical linguistics,
denotational semantics and semiotics ARE. People tend to see them
as theories about the real world instead of a metalanguage in which
to express theories about real AND/OR IMAGINARY worlds.
The attacks against generative grammar tend to be about as valid as
an attack against Euclidean Geometry based on the premise that Einstein
discovered that space is curved. Euclidean Geometry is an ABSTRACTION
whose "truth" or "validity" have nothing to do with how well it maps
onto Einsteinian space-time.
Generative grammar is anologous to an algebra. Just because no one can
produce equations in the algebra that correctly describe the behaviour
of French speakers proves only that the present equations are wrong, not
that the algebra is flawed. Physicists cannot write a "grammar" (single
set of equations) which describe the behaviour of all known physical
objects. Does this prove that Mathematics is bunk?
Transformational grammar is analogous to a Calculus. There are many
dynamical systems whose behaviour can be described but not predicted with
differential equations because no one knows how to solve them. Does
this prove that Calculus is bunk? No way. And the fact that a TG
has not YET been produced which completely describes any natural
language should not be surprising: look how many CENTURIES passed
from the invention of calculus to QED theory!!! TG is a YOUNG branch
of mathematics. It may be centuries before some applied linguist will
be able to use it to completely describe a natural language.
--alan@pdn
------------------------------
Date: Fri, 29 Jan 88 08:59 EST
From: merrill@iuvax.cs.indiana.edu
Subject: Re: failure of TM
/* Written 11:38 am Jan 28, 1988 by alan@pdn in iuvax:sci.lang */
Transformational grammar is analogous [sic] to a Calculus. There are many
dynamical systems whose behaviour can be described but not predicted with
differential equations because no one knows how to solve them. Does
this prove that Calculus is bunk? No way. And the fact that a TG
has not YET been produced which completely describes any natural
language should not be surprising: look how many CENTURIES passed
from the invention of calculus to QED theory!!! TG is a YOUNG branch
of mathematics. It may be centuries before some applied linguist will
be able to use it to completely describe a natural language.
/* End of text from iuvax:sci.lang */
There is a fundamental difference between Calculus and TG. Calculus
was invented to solve problems that it solved; transformational
grammar was created to solve problems that it has failed to solve. To
be sure, syntacticians can create grammars for small fragments of the
written language, but there is no language---not even Turkish---for
which an adequate grammar has been constructed for a large fragment of
the written language. No spoken language has been even roughly
approximated. Wisely, I suspect, since, like this sentence, spoken
language tends to be ungrammatical.
The claim that "look how many centuries passed &c" is a traditional
dodge, as commonly used in my own field of AI as it is in Linguistics.
In either case, the appropriate response is: "if your tools are not
yielding interesting results in the fields for which they were
designed, you've got the wrong tools." Tools are designed to solve
problems, and those which are designed, and then don't work
immediately, should be discarded.
------------------------------
Date: Fri, 29 Jan 88 11:55 EST
From: Greg Lee <lee@uhccux.UUCP>
Subject: Re: failure of TM
In article <2132@pdn.UUCP> alan@pdn.UUCP (0000-Alan Lovejoy) writes:
+In article <1860@russell.STANFORD.EDU> poser@russell.UUCP (Bill Poser) writes:
+>...
+>constructing theories, to testing them, and to deep studies of
+>...
+There is a fundamental and fatal misunderstanding of just what
+generative grammar, formal language theory, mathematical linguistics,
+denotational semantics and semiotics ARE. People tend to see them
+as theories about the real world instead of a metalanguage in which
+to express theories about real AND/OR IMAGINARY worlds.
+...
+Generative grammar is anologous to an algebra. Just because no one can
+...
What an awful idea. How comforting to a generative grammarian who
is unable to find evidence to support his views. Let us solicit the
opinion of a mathematician as to whether transformational grammar
is worth pursuing for its mathematical interest.
Greg, lee@uhccux.uhcc.hawaii.edu
------------------------------
Date: Wed, 27 Jan 88 12:43 EST
From: DAVID PESETSKY <PESETSKY@cs.umass.edu>
Subject: Gross's article
Speaking as a (generative) linguist, it seems to me that the recent
discussion about Gross's old paper in Language, and the absence of a
published generativist's reply leaves one important point out. Of
course, whether someone giggled or not at the paper makes no difference
if the work giggled at is of high quality or theoretical importance.
The point is that there are papers published, particularly in our young
and somewhat inchoate field, to which the appropriate reply would be a
few sentences -- and there is no printed forum for such a reply.
Perhaps that is where e-mail comes in. Here is one perhaps appropriate
reply -- with the caveat that I do not have the paper by Gross at hand:
Gross decided to undertake a massive project on the lexicon-syntax
interaction at a time when our understanding of either the lexicon or
syntax was not at a level to carry it out. In fact, such a project
would be only slightly less premature today. Having made this premature
attempt, and failed, he pronounced generative grammar a failure in his
opening sentence.
It's rather as if I decided to travel to another solar system, discovered
that it could not be done with present technology, and thereby pronounce
engineering, or physics a failure. Or perhaps I decide to produce a full
computer simulation of the human being, from biochemistry to gross anatomy;
discovering that I cannot do it, I declare biology a failure. It's because
of the patent absurdity of such conclusions that the article went without a
reply. Sure, his attempt failed -- it was bound to; but who was Gross to
decide for generative grammar what the field should be capable of at some
given moment. Remember, generative grammar has existed in its present form
for quite a short time. We know lots more than we used to (cf. Halle and
Higginbotham's reply to Wasow on "Linguistics as a Science" in Natural
Language and Linguistic Theory from a year or so back), but a lot less than
we hope to. If we had all the answers, we wouldn't be asking any questions,
and we are asking lots of them. It seems quite odd to blame a science for
not having the answers to certain problems -- even fundamental ones. It's by
asking questions and struggling to find answers to them -- yes, even
fundamental ones -- that a living field stays alive.
David Pesetsky
Dept. of Linguistics
South College
University of Massachusetts
Amherst, MA 01003
pesetsk@umass.bitnet
------------------------------
Date: Fri, 29 Jan 88 13:41 EST
From: Bruce E. Nevin <bnevin@cch.bbn.com>
Subject: the Gospel according to Newmeyer
BP> From: Bill Poser <poser@russell.STANFORD.EDU>
BP> Subject: Re: failure of TM
BP> But what I responded to was not Gross, it was a reference
BP> to Gross. In particular, I responded to two points:
BP> (a) the appeal to authority of citing Gross and referring to the
BP> prestige of the journal in which his paper was published.
BP> Saying that the authority is not so authoritative is
BP> a perfectly appropriate response to an appeal to authority.
BP> It is not I who have refused to discuss ideas.
It is a little hard to tell who is responding to whom. I regret that
you have evidently taken my reference to the the prestige of _Language_
as an appeal to authority. What I meant by noting the recognition
accorded the journal of the Linguistic Society of America was only that
the lack of response could not have been because it was printed in some
obscure place and therefore overlooked. I did and do inveigh against
appeals to authority.
The question of perceived standing and circulation of journals is
important. _Language has been criticized as being anti-Generativist
because it publishes non-Generativist and even on occasion
anti-Generativist contributions. (I suppose I could find some
references if you question that.) Some of the other journals cited
recently in this forum do not. This makes it difficult for alternative
perspectives to get a hearing. For an assessment of some aspects of
this difficulty and its implications for the health of the field, see
Sternberg, Crisis in Linguistic Inquiry, _Forum Linguisticum_
3.3:189-207 (1973). This journal may be hard to find, of course. I
assume that the discussion and advocacy of pluralism in linguistics
(interestingly parallel the pluralist/analytical political controversy
in philosophy) continues in more recent literature, but I have not seen
it because I do not have access to non-"mainstream" (political term) journals.
BP> (b) the false inference that failure to reply to Gross in print
BP> constitutes a tacit admission of defeat on the part of
BP> generative grammarians.
Look, Maurice Gross is no dummy. He is not to be compared with
creationists in biology, nor are his writings to be compared with "every
crank attack on relativity" in physics. His book _Mathematical Models
in Linguistics_ (Prentice-Hall 1972) is still both useful and used. His
detailed publications coming out of ongoing research in the grammar of
French at the LADL and CERIL at the University of Paris VII since the
mid 1960s are available for anyone who wants to evaluate "the putatively
comprehensive grammars that Gross holds up as the standard". How can it
possibly be appropriate to publish no rejoinder in a comparable place?
Gross was Chomsky's darling, his hope of further conquests on the
Continent, until Gross began pressing harder for ways to produce a real
grammar that is actually useful for something (a matter of some concern
to computational linguists), and began getting answers to his questions
from Chomsky's teacher because Generativists had none. Now, he is an
un-person. Only the Party Historian, Newmeyer, has anything to say.
And what does Newmeyer have to say? By your account:
BP> Newmeyer's main point is that Gross errs
BP> in assuming that a generative grammar must account for all of the
BP> facts about a particular construction by means of a single rule.
BP> He illustrates this assumption of Gross regarding the English Passive
BP> construction, then summarizes a modular approach in which different
BP> aspects of grammatical theory together provide an account of the Passive.
BP> He also points to extragrammatical factors that play a role in determining
BP> the acceptability of passive sentences.
What Gross is concerned with is not the Passive (though that is the
first of four examples), but rather characteristics of the Generative
paradigm or programme (Lakatos's term) that make it unworkable. In
particular, he is concerned with the Generativist redefinition of data
as existing in the metalanguage rather than in the language. (Please
read that sentence twice.) In the section in question, 1.1, he points
to difficulties with the passive that have been known, yes, since before
Chomsky (who learned about them from Harris), and says "the severity of
these [difficulties] renders surprising the tenacity put into
formalizing it." Because for over thirty years continuing efforts >have<
been made to treat it under a single rule or operation, and it >has<
continued to be the stock example illustrating what transformations are
for the neophyte.
And Gross refers to Harris's proposals for the passive as a resultant of
combinations of rules that also apply to different constructions,
proposals that were soon to be published (in 1982) in section 8.4 of
Harris's _Grammar of English on Mathematical Principles_. They are thus
integrated in a comprehensive grammar. The "modular" proposals (ah,
what a buzzword there!) that Newmeyer puts forward are isolated. We
have no idea whether they will fit consistently together with other
latest and greatest proposals to make a comprehensive grammar.
Generativist linguists are exactly in the position of AI researchers who
have made toy systems for somall domains and then discover that they
cannot generalize from them to larger domains or general coverage. This
is one of Gross's main points. But Newmeyer makes not note of it, but
instead provides yet another example as his suggested counterexample
regarding the isolated problem of the passive!
Now as to synonymy, yes, Chomsky (following Quine, Harris, Bloomfield)
rejects it as undefineable. I believe what Gross is referring to here
is heuristic practice, not theoretical principle. In practice, when
Generative linguists consider paradigms (sets of examples) they
frequently refer to differences of meaning to demonstrate that two
sentences are derived or do not derived from a common underlying source.
Often, this is put in terms of consequences deducible from each
sentence, and the claim is sometimes made that two sentences are
synonyms iff one may deduce from each the same set of consequences.
In any case, his substantive point, which follows immediately upon the
passage that Newmeyer excises for ad hominem carping, is that one must
proceed, as in any science, with an adequate range of data
systematically arranged. He inveighs against the almost exclusive use
of anecdotal data in Generativist literature. About this, Newmeyer says
nothing.
Nor does he say anything about the substantive results of Gross's
large-scale, systematic study of distributional facts of French.
Instead, he limits himself to the vacuous superciliousness of "It is
hard to imagine the source of such a peculiar idea," an almost comical
parotting of Chomsky's polemical style.
Note that he had no need to "imagine" the source of the idea, the source
was set before him in the very article he was allegedly discussing.
Gross was referring to some problematic consequences of the the way the
competence/performance dichotomy (a very slippery, equivocal concept)
is used in Generative theory. The sentence Newmeyer picks out of
context immediately follows this quote from Chomsky's 1962 Texas
Conference paper:
There are in fact exceptions to many of the transformational
rules given above, perhaps all. These will have to be
separately listed, unless some more general formulation can be
found to account for them as well. The discovery of such
exceptions is in itself of little interest or importance
(although the discovery of an alternative formulation in which
the exceptions disappear would be highly important) . . . But
discovery of exceptions to grammatical generalizations is of no
consequence in itself, except when it leads to an alternative,
more comprehensive generalization.
It is at this point that Gross says "Under the most favorable
interpretation, Chomsky appears to regard the exception to a linguistic
rule as a physical scientist might regard an experimental result
incompatible with his theory as being caused by some unperceived error
in the experimental apparatus." (865) And after the part that Newmeyer
quotes, he goes on to say:
Certainly a physical scientist who has not explicitly designed
an 'experimentum crucis' is reluctant to abandon his theory on
the basis of one experimental failure. Nevertheless, it is his
responsibility to DEMONSTRATE, either by experimental repetition
under better conditions or by an analysis, that the given
experimental result inconsistent with his hypothesis is in fact
erroneous by reason of experimental error. It will not do to
dismiss a sentence acceptable by those competent in a language,
when theory suggests it should be unacceptable (or conversely),
simply by noting that it is an exception. Of course, any
natural language will have exceptions, i.e. special usages; of
course, one may not be able to demonstrate at once that each is
idiosyncratic; but at the least, one must note all the
exceptions, even if one defers their consideration for later
study. Discussion of Passive over more than twenty years is a
striking example of this methodological error, which consists
basically in denying the requirement that relevant linguistic
instances be enumerated. [A footnote here describes the
derivation proposed by Harris, mentioned above, which avoids the
"complex problems of notational formalization" {reference to
Chomsky 1975:106-17} that have been >seriously< entertained in
Generativist literature.--BN]
It will be hard for the specialist in a natural science to
believe that such investigations have never been undertaken for
questions of English syntax. Traditional grammarians did not
construct syntactic inventories either. [I believe your claim
was that Gross was holding Traditional grammar up as a standard
for comparison? It is difficult to see how you could hold this
view if you had read past the abstract.--BN] But today we know they
did not possess methods and motivations that might have enabled
them to succeed. Transformational methods [NB, not Generative
methods--BN] made this form of research possible; and for
practically all problems of syntax, accumulating data appears to
be at least as necessary as it is for Passive. Since new
insights into the nature of syntactic phenomena are likely to
arise from systematic exploration with the aid of a dictionary,
it is all the more surprising that GG has ignored this aspect of
linguistics.
Accumulating data is obviously not an aim in itself. But in all
natural sciences it is a fundamental activity, a necessary
condition for evaluating the generality of phenomena. Such a
concept of generality or of importance of facts is totally
absent from GG, where sentences acquire significance only with
respect to formalism. In GG, a linguistic example appears to be
significant only if it allows one to choose between competing
theories. . . .
I quote this extensively to show that Gross is making non-trivial and
substantive claims that Newmeyer simply ignores with a bit of
hand-waving--precisely the criticism others here have recently laid
against Generative linguistics.
But it served its intended purpose, because the party faithful can read
and quote Newmeyer and know that they don't have to bother to read and
understand Gross. This appears to be what you have done:
BP> In an attempt to give a fair
BP> summary I will quote the abstract of this paper, written by Gross
Internal evidence, such as your misconstrual of the status Gross accords
to Traditional grammar, suggests that you read only the abstract and
Newmeyer's critique. Please correct me if I am wrong by addressing some
of the substantive issues in the article.
BP> In sum, Gross's paper badly misrepresents the claims of generative grammar
BP> and proposes totally inappropriate tests of its validity. That's why
BP> it wasn't taken very seriously at MIT.
You go on to say:
BP> Let me now respond to the claim that what arouses the
BP> ire of anti-generative people is the putative tendency of generative
BP> grammarians to dismiss all criticism and refuse to discuss the issues.
BP> I don't really believe that this is true in general, and I would be
BP> most interested to hear of examples.
Thank you for providing two.
I must say I was not surprised at the shallowness of Newmeyer's supposed
critique of Gross. I had previously been exposed to the shoddy
scholarship of his _Linguistic Theory In America_ (1980). Just a couple
of howlers from the early pages of the first chapter:
The dominant intellectual force in the United States from the
1930s to the 1960s was empiricism. (3)
This statement is just incredible. He uses the term "empiricist" as a
mere epithet for rhetorical effect, faithfully following Chomsky's claim
that he, the Rationalist, split with Harris, the Empiricist (he is
none), over their irreconcilable philosophical differences. An example
of Newmeyer's distortion for rhetorical effect is his use of the
relatively unknown and unused name "Logical Empiricists" for the
philosophical school much more generally known as Logical Positivists,
and his implication without quite saying that Harris, Bloch, and the
rest were all Logical Positivists (patent nonsense).
The goal of structural linguistics was to "discover" a grammar
by performing a set of operations on a corpus of data. (6)
The "taxonomicist" quest for a discovery procedure is another part of
the standard mythology of Generative linguistics. Newmeyer says he
rejects the contrary view of Hymes and Fought ("American Structuralism",
in _Current Trends in Linguistics_, the volume on Linguistic
Historiography (1975), pp. 904-1176), as indeed he must since the facts
of history contradict the myth. Consider for example the testimony of
Harris (1951), which Newmeyer like everyone else takes to be the locus
classicus of the "taxonomic" straw man:
These procedures are not a plan for obtaining data or for field
work. . . . The procedures also do not constitute a necessary
laboratory schedule in the sense that each procedure should be
completed before the next is entered upon. In practice,
linguists take unnumbered shortcuts and intuitive or heuristic
guesses, and keep many problems about a particular language
before them at the same time: . . . The chief usefulness of the
precedures listed below is therefore as a reminder in the course
of the original research, and as a form for checking or
presenting the results, where it may be desirable to make sure
that all the information called for in these procedures has been
validly obtained.
Harris, _Methods in Structural Linguistics_,
1-2 (1951, ms completed January 1947)
Elsewhere (I cannot find the quote), Newmeyer makes the stock claim that
"taxonomic" linguistics cannot account for novel utterances and so
cannot account for the "generative" capacities (one of the many senses
of that highly equivocal term!) of native speakers. Contrast again the
following from our locus classicus:
The work of analysis leads right up to the statements which
enable anyone to synthesize or predict utterances in the
language. The elements form a deductive system with
axiomatically defined initial elements and with theorems
concerning the relations between them. The final theorems would
indicate the structure of the utterances of the language in
terms of the preceding parts of the system.
Harris, 1951, 372-3
Please note that Harris's current operator-argument grammar, which
claims to attain this aim, is a product of continuous development from,
and not revolutionary refutation of, structural methods. From the point
of view of linguistics as a science, the Generativist revolution was
unnecessary, although from the point of view of politics it has had its
own sufficiency.
Finally, a couple of closing observations. You say:
BP> Generative grammar does not purport to
BP> provide an especially efficient means of constructing grammars.
BP> . . . they have not tried to construct
BP> comprehensive grammars and . . . the theory makes no claim of
BP> efficiency at doing so
From the point of view of someone who wants to apply linguistics to
practical ends, people like computational linguists and Maurice Gross,
this makes Generative Linguistics pretty useless. Generative Grammar by
your own admission then does fail to serve such ends. It doesn't matter
that Generative linguists have not taken up such aims (is that your
claim, really?) and can therefore be said technically not to have
"failed". Such sophistry does nothing for a computer scientist with
problems to solve. If Generative Linguistics does not address those
problems, then the computer scientists will re-invent linguistics, and
Generative Linguistics will be just plain irrelevant to the enterprise.
This is in fact what we see around us today.
BP> In sum, I do not think that it is true that generative grammarians
BP> have, in general, ignored the serious challenges to their approach.
BP> They have ignored criticism that was plainly silly, and they have ignored
BP> repetitions of criticisms involving issues they consider to have been
BP> settled long ago or to reflect irreconcilable philosophical differences.
Gross is not silly, and he is raising new issues. Even if he is raising
issues that you think were "settled long ago", are you saying that
issues cannot be reopened in light of new evidence (of which he presents
an abundance)? And are you saying that "irreconcilable philosophical
difficulties" are on a par with these two criteria for ignoring
criticism? Then inter-paradigmatic discussion is not merely difficult,
it is foreclosed, and by your choice.
More than once I have seen a linguist of a non-Generativist persuasion
rise at the conclusion of a talk by a Generativist to comment that the
knotty problems just presented by the speaker simply do not arise at all
in their alternative paradigm. Gross points to several such phenomena
in his discussions of the Passive, Raising, Relativization, and
so-called French "aspirated h" in phonology. The affordance of
"parallax" between two perspectives can be extremely useful in
disclosing phenomena that are mere artifacts of the notation or other
conventions used by one or the other paradigm. Another example is the
X-bar notation taken by Chomsky from Harris's 1946 paper. The problem
that it addresses simply does not arise with center-and-adjunct grammar
(string grammar). This was the motivation for Joshi's exploration of
grammars with mixed types of rules, and currently of tree-adjunction
grammars (TAGs).
But the neo-Talmudism (Michael Kac's father proposed the term) of
Generative linguistics seems to preclude this kind of exploration and
discourse. In consequence, the field is becoming increasingly insular,
provincial, and irrelevant. The drying up of funding and the
shrivelling or merger of linguistics departments is one overt
consequence that we have seen for a number of years.
Bruce Nevin
bn@cch.bbn.com
<usual_disclaimer>
------------------------------
Date: Sat, 30 Jan 88 00:54 EST
From: rolandi <rolandi@gollum.Columbia.NCR.COM>
Subject: the imaginary worlds of grammarians
Regarding the following dribble....
>+There is a fundamental and fatal misunderstanding of just what
>+generative grammar, formal language theory, mathematical linguistics,
>+denotational semantics and semiotics ARE. People tend to see them
>+as theories about the real world instead of a metalanguage in which
>+to express theories about real AND/OR IMAGINARY worlds.
>+...
>What an awful idea. How comforting to a generative grammarian who
>is unable to find evidence to support his views. Let us solicit the
>opinion of a mathematician as to whether transformational grammar
>is worth pursuing for its mathematical interest.
>Greg, lee@uhccux.uhcc.hawaii.edu
Yo! Greg! Give it up! Why would you want to question the meanderings of this
grammarian! What do you want? Words with referents? Statements that are,
in some small way, related to facts? My God man, do you fancy a theory about
the real world?
Don't challenge this guy. Let the philosopher sleep! He may be dreaming that
you and I exist! We, and the language we use to describe our existence, may
only exist in the metalanguage of his imaginary world! An imaginary world
entirely defined in terms of his meaningless terminology!! Walk softly man!
walter rolandi
rolandi@gollum.UUCP ()
NCR Advanced Systems, Columbia, SC
u.s.carolina dept. of psychology and linguistics
------------------------------
End of NL-KR Digest
*******************