Reply to carnie from LLJ

[Shalom Lappin]

I appreciate David Pesetsky's interesting comments. They greatly improve the
atmosphere of the discussion. I will limit myself to two points in reply.
My co-authors may want to respond individually.

In J&L (1999) we did, in fact, discuss some of the proposals
in Chomsky (1998), "Minimalist Inquiries". Specifically, on p. 64 we
take up Chomsky's suggestion that numerations can be specified for the
derivation of embedded phrases and then extended to larger numerations
to yield the matrix clauses that contain these phrases. We argued that, if we
understood this proposal correctly, it renders vacuous the notion of a
numeration as a bounded selection of lexical items. This seems to undermine
the role of a numeration as a well defined initial state in a derivation.

David says that the MP is a loose program rather than a clearly defined
theory or theoretical framework. Therefore our criticisms apply only to
the version of the MP (basically Chapter 4 of Chomsky (1995)) that we
discuss in the book. It may well be the case that many of the technical
issues that we raise in the book are no longer relevant to current
formulations of the MP. However, our main concern was (and remains) with
the foundational assumptions of this program. In particular, we argued
that Chomsky's view of language as a "perfect computational system" is
seriously misguided and has no place in the theory of grammar. We observed
that economy conditions, global or local, are directly inspired by this view
of grammar as an "efficient" system for mapping lexical numerations into
interface representations. We attempted to show that both sorts
of economy condition generate insuperable computational and formal
difficulties. Moreover there is no independent empirical motivation
for such conditions and considerable evidence against them. The only reason
for adopting them appears to be Chomsky's intuition that grammar is "perfect".
We devoted considerable effort to trying to understand this intuition, and
to showing that on any of the credible interpretations that we were able
to assign it, the intuition has no role in linguistic theory. It continues
to strike us as remarkable that this intuition has enjoyed such widespread
acceptance without being subjected to critical examination and clarification.
These concerns also animated our contributions to the NLLT debate. From what
I can see these criticisms continue to apply to the MP as it is pursued today.
I have yet to see a convincing answer. Regards.
                     Shalom Lappin




On Sun, Jun 27, 2004 at 02:24:02PM -0400, David Pesetsky wrote:
> Dear HPSG-L,
>
> I've been a subscriber to this list for a while, but have refrained from
> getting involved in the discussions. I think this was a good policy,
> especially since I didn't want to be seen as intruding. At the same time I
> don't want to let Andrew Carnie be hung out to dry, so I hope I will be
> forgiven for adding a remark or two.
>
> I agree with Lappin, Levine and Johnson (and with Andrew himself) that the
> Johnson & Lappin book was not polemical in tone.  We read through the book
> in a class I taught last Fall, so I am fairly familiar with it (as are a
> number of our students).  The book is impressive in the degree to which it
> takes the opposing views seriously, clarifying these views where necessary,
> offering alternatives not only from an HPSG perspective, but also from
> Minimalist and GB perspectives.
>
> That said, Ivan has characterized some of my impressions accurately, when
> he writes, recalling a recent conversation:
>
> >[Ivan:] I was
> >talking with David Pesetsky in May about the Johnson-Lappin book and why there
> >had been no published response to it. What he said was that (almost all?) the
> >objections they raised in the book were the very issues that had already been
> >recognized as critical problems for MP and which people were already working
> >on. So in David P's view, if I understood him right, the issues raised didn't
> >provide grounds for a rejection of MP at all. David P. also suggested that
> >this mismatch of assumptions played a role in why no one took on their book
> >pointedly. All right. Presumably, David J and Shalom would disagree with this
> >assessment, but let's not go there right now.
>
> Lappin, Levine and Johnson, responding to a similar point made by Andrew,
> wrote in their recent message:
>
> >[L, L & J:]Carnie informs us that the adherents of the MP were well aware
> >of the issues that we raised in our L&P article and our monograph,
> >and they were attempting to address them. This is interesting and
> >important news. Why weren't these efforts published? Where is the
> >record of critical examination of these problems within the MP
> >literature? Are we alone in having missed it?
>
> In fact, much of the literature that calls itself "Minimalist" -- including
> Chomsky's own subsequent articles -- is devoted to "critical examination of
> these problems" and others like them. For example, Chomsky's own proposal
> in "Minimalist Inquiries" that syntax proceeds "phase by phase" (when
> combined with very explicit views about the categories that make up a
> phase) responds to many of the combinatory explosion problems discussed by
> Johnson and Lappin in section 2.3 of their monograph and elsewhere.  It is,
> in fact, I think, close to the patch that Johnson and Lappin themselves
> hint at in the last paragraph (second sentence) of page 14 of their
> monograph.  Likewise, the inconsistency between covert movement analyses of
> the English 'there' construction and other "Minimalist" proposals,
> discussed in section 2.4.4 of J&L's book, is resolved by Chomsky's later
> proposal to distinguish (but link) agreement and movement, when combined
> with his ideas about the particular properties of  "defective" categories
> that lack a full set of phi-features.   Similaly, the Smallest Derivation
> Principle (J&L's terminology) -- invoked by Chomsky to explain the
> impossibility of overtly raising a direct object from the Object Shift
> position to Spec,TP,  and roundly criticized by J&L in section 2.5 -- is
> not needed if Chomsky's "Activity Condition" is invoked.  Indeed, exactly
> this sort of problem *motivated* the Activity Condition, which is argued to
> play a role in explaining other phenomena as well.
>
> I assume that HPSG-L is not the place to argue at length whether these
> responses are on the right track.  All I want to note is (1) that the
> issues were indeed already live when J&L's work appeared; and (2) that the
> responses exist -- thus substantiating Andrew's remark.
>
> Now I don't think it is a bad thing at all that J&L called attention to
> problems that others were also identifying.  It is impressive when
> researchers in different locations, with very different perspectives and
> goals, converge on a conclusion. Indeed, J&L could hardly have been unaware
> of this, since the third chapter of their book (as they note) is devoted to
> critiquing proposals by Collins and Yang that also aimed to solve some of
> the problems presented in J&L's second chapter.  But it is perhaps worth
> asking why one group of researchers  (e.g. Johnson and Lappin) concludes
> from their critique that they have shown the Minimalist Program to be "an
> untenable framework that had not been properly motivated" -- while others
> view it as a *strength* of the approach that "Minimalist" work suggests
> interesting, testable solutions to such problems.
>
> One answer is the obvious difference in the hunches and prejudices of
> different groups of syntacticians.  Another answer, I think, is simply
> terminological confusion.  For better or for worse, the phrase "Minimalist
> Program" has come to mean lots of different things to different people.
> The phrase encompasses everything from Chomsky's speculations about
> language to the research guidelines argued to flow from these speculations,
> to the very specific set of proposals and analyses proposed by Chomsky
> himself -- as well as a much larger (and quite diverse) set of proposals
> and analyses by researchers whose thinking is somehow influenced by
> Minimalist work.  Whatever the term "Minimalist Program" may mean to the
> field (and I personally have no objection to discarding it), it's not a
> "framework".
>
> Not that we probably all agree on what a "framework" is, either!  But I
> suspect that the term "framework" can be meaningfully applied to the
> packages of disparate but connected proposals that one finds in certain
> individual publications or groups of publications,  e.g. Chomsky's
> "Categories and Transformations" (= "Chapter 4") in the case at hand.   One
> might indeed reasonably claim that L&J and others have shown that the
> "framework" presented in that paper of Chomsky's is "untenable" and not
> "properly motivated" --  in that there are problems and internal
> contradictions among the ideas of that paper.  You could probably find
> multitudes of self-styled "Minimalists" who would agree on that point, put
> that way.  A meaningful divergence of views between L&J and Minimalists
> would probably come in answer to the question "what next?", but that's a
> different topic.
>
> Even so, however, I think we can do better than debating frameworks.  The
> wisest remark made so far in this discussion has been Lappin, Levine and
> Johnson's admonition that "theories are instruments of explanation, not
> religious affiliations".  Yes!  It is my personal experience that
> discovering the links and disconnects among specific ideas is more
> productive (and a lot more interesting) than debating the merits or
> demerits of pre-packaged "frameworks" as if they were take-it-or-leave-it
> religions -- which they never really are.
>
> Thanks for hearing me out.
>
> -David Pesetsky
>
>
> *************************************************************************
> David Pesetsky	[pesetsk at mit.edu]
> Department of Linguistics and Philosophy
> 32-D862 Massachusetts Institute of Technology
> 77 Massachusetts Avenue
> Cambridge, MA 02139 USA
> (617) 253-0957 office           (617) 253-5017 fax
> http://web.mit.edu/linguistics/www/pesetsky.home.html