Reply to carnie from LLJ

[Shalom Lappin]

We will pass over the offensive style and tone of Andrew Carnie's
recent notes and limit ourselves to correcting the more obvious
misrepresentations of our work with which his comments are replete.
Johnson and Lappin (1997), "A Critique of the Minimalist Program",
Linguistics and Philosophy, and Johnson and Lappin (1999), Local
Constraints and Economy, CSLI, Stanford, are not polemical pieces
but detailed critical studies of the Minimalist Program, primarily
as it was presented in Chomsky (1995). There is not a trace of
rancour in these publications. Instead, we did our best to
accurately describe the defining assumptions and the primary claims
of the MP, and to subject them to a thorough examination. We made
every effort to formulate this model (program, assembly of ideas,
or however one wishes to describe it) as generously as possible,
frequently supplying what we took to be missing premises and
reformulating questionable definitions in order to strengthen the
proposal. After a careful consideration of the computational and
formal properties of the MP (of that period) and an examination of
the main empirical arguments in support of it we concluded that it
was an untenable framework that had not been properly motivated.

To the best of our knowledge, these arguments were never taken up in
print. Most of the critical replies in NLLT to our Topic-Comment did
not refer to this work but dealt only with the T-C piece and our
subsequent reply. This is also true of Freidin and Vergnaud (2001),
"Exquisite Connections: Some Remarks on the Revolution of Linguistic
Theory", Lingua, which also addresses only our T-C note and discards
the debate that followed in NLLT, as well as Johnson and Lappn
(1997) and (1999).

Our T-C piece and subsequent replies were indeed polemical, as this
forum is intended to be. They were written after the extended absence
of any discussion of what we took to be the deep foundational
problems that we had identified. Our contributions to the debate
were not "nasty", nor were they intended to give offence. They
raised questions concerning the nature of scientific discourse
in the field. We did not accuse the adherents of being "mindless drones"
or "slaves" (Carnie's terms,not ours). We presented
evidence for asserting that the transition from GB to the MP
had been made rapidly and without the sort of detailed examination
of the relative advantages of the two theories that one would
expect in the case of a major theory change. This evidence
consisted in a comparison of the percentage of syntax pieces
in major linguistics journals that were formulated within GB at
the beginning of the 1990's with the percentage of MP-inspired
syntax papers in the same journals at the end of the 1990's. We
found a move from approximately 75% in GB mode to approximately
75% in MP orientation in less than ten years. Most strikingly, we
did not find a single article or monograph that provided a
systematic argument for abandoning GB and its core theoretical
notions of government, theta-theory, Case theory, DS/SS, etc. in
favour of the MP, which discards all of these elements. This led
us to the conclusion of unmotivated theory change that we
presented in the context of the NLLT debate. So far neither Carnie
nor anyone else has said anything of substance to challenge this
conclusion. It is understandable that one might feel uncomfortable
or embarrassed by these facts, but to point them out is not "nasty"
or "insulting". It is a legitimate part of inquiry into the way in
which theory change occurs within our field. If we are wrong in
either our empirical observations or the conclusions that we draw
from them, then the reasonable response is to show us where we
have gone wrong. Unfortunately, this is not how the debate has
been conducted until now.

Rather than addressing the substantive issues raised in the
NLLT debate, Carnie returns to the question of whether the move
to the MP was actually a paradigm change or simply a natural
progression within the P&P framework. Two of the MP respondents
in the NLLT debate, Juan Uriagareka and Ian Roberts, strongly
endorsed the paradigm change view of the MP, with Roberts
going so far as to characterize it as a new mode of explanation
in linguistics. Eric Reuland dissented, adopting Carnie's
incremental revision approach. The other two participants did not
address the issue directly. As far as we can see, there is no
real point at issue in the choice of terminology here. Regardless
of whether one regards the MP as a paradigm change or an
incremental revision of the P&P assumptions underlying GB, it is
clearly a fairly radical departure from the latter theory, given
that it dispenses with most of its architecture (levels of
representation and constraints defining those levels) and
introduces either global or local economy conditions
(transderivational constraints). It also postulates a (vague)
notion of grammar as a "perfect computational system" and the
related (exceedingly vague) idea of "virtual conceptual necessity",
both of which are frequently invoked to justify the machinery of
the theory. We discuss all of these concepts in considerable detail
in both our technical work and within the NLLT. None of our main
concerns with them have been answered. We invite readers of this
list to revisit the NLLT debate to reach their own conclusions on
this matter. Of course we could have missed something that appeared
recently, and, if this is the case, we would be interested in hearing
about it.

We invoked the term "postmodernist", which Carnie is so fond of
brandishing as an ironic self-ascribed epithet, when discussing
the metaphorical use of terms from physics in some MP publications,
particularly Uriagereka's. We had in mind his penchant for
expressions like "entropy" and "event horizon ". We attempted to
make sense of this use of "entropy" in one of our replies in the
NLLT debate, but without apparent success. In fact, as we showed
and Uriagerka seemed to concede, the principes which he based on
this notion of entropy turned out to predict exactly the
opposite of what they were intended to. In using the term
“postmodernist” we were referring to Sokal and Bricmont's
(1999) masterful study of the peculiar insistence of postmodernist
cultural theorists on using terminology from physics and
mathematics in their work without providing any sense of how they
intended this terminology to be interpreted. Nothing that we have
seen since the NLLT debate was published has persuaded us that this
reference is inappropriate.

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?  Was our monograph
an exercise in redundancy?  More seriously, where are the
solutions to these problems? Collins (1996) and Yang (1996) did
indeed propose substituting local for global economy conditions.
We discuss these suggestions in detail in a chapter of our
monograph, where we argue that both proposals are computationally
problematic and empirically unmotivated. Specifically, they do not
appear to allow for a coherent integrated economy metric in which
all the local economy conditions are consistently applied to a
derivation to yield a single economy value. We point out numerous
other difficulties with these proposals that we will pass over
here. Readers are invited to consult the book for the relevant
arguments.

Oddly, in his first note Carnie tells us "Any good scientist
attempts to be aware of the real problems in their own work,
even if they choose to temporarily ignore them for practical
reasons. The MP world doesn't need HPSG critics, it has its
own. What it needs are solutions to those problems." This is a
remarkable assertion. It appears to deny the desirability of
criticism from non-adherents and limit their role to offering
fixes for the MP's difficulties. Chomsky has frequently pointed
out the importance of evaluating criticism solely on the basis
of its merits and not its source. Surely discussion of a
theorie’s flaws is not a privilege enjoyed only by its advocates.
If MP theorists are not inclined to be receptive to critical
comment coming from outside the confines of their assumptions,
as Carnie suggests, then this would indicate that there is
indeed a serious problem in the way in which scientific
debate is conducted within the field, and this is, of course,
precisely the point that we were trying to make in our NLLT
pieces.

Finally, it was not our intention, either in our technical work
or in the NLLT debate, to promote one particular theoretical
paradigm over another. We were concerned to examine the foundations
of an influential theory which has been shaping work in the theory
of grammar for the past fifteen years. Any model that purports to
offer a serious framework for describing and explaining
the facts of natural language should be subjected to rigorous
scrutiny on a continuing basis. In our view the dominance of tribal
theoretical wars has rendered linguistics an increasingly barren
field over the past several decades. Theories are instruments of
explanation, not religious affiliations. Our focus should be not
on advancing the hegemony of a particular paradigm banner but
arriving at the best possible account of the properties of natural
language. This is, after all, what linguistics is supposed to be
about.

       Shalom Lappin, Robert Levine, and David Johnson