LINGUIST List 12.2024

Fri Aug 10 2001

Review: Epstein, Essays in Syntactic Theory

Editor for this issue: Terence Langendoen <terrylinguistlist.org>

Directory

1. Manideepa Patnaik, Review: Epstein, Essays in Syntactic Theory

Epstein, Samuel D. (2000) Essays in Syntactic Theory. Routledge,
hardbound ISBN 0-415-19235-8, 216pp, Routledge Leading Linguists
Series.

Manideepa Patnaik, Harvard University

This book is a collection of nine articles by Samuel D. Epstein which
have been published in leading journals between 1984 and 1997. Each
article is within the generative theory of linguistics, pioneered by
Noam Chomsky. They are as follows:
1. A Note on Functional Determination and Strong Crossover
2. Quantifier-pro and the LF Representation of PROarb
3. The Local Binding Condition and LF Chains
4. Adjunction and Pronominal Variable Binding
5. Quantification in Null Operator Constructions
6. Differentiation and Reduction in Syntactic Theory: A Case Study
7. Derivational Constraints on A-bar Chain Formation
8. Overt Scope Marking and Covert Verb-Second
9. "UN-Principled" Syntax and the Derivation of Syntactic Relations

The first article has something in common with the last one in
presenting arguments that certain representational constructs are
inadequate and that one empirically and conceptually preferable theory
makes reference to properties of the derivation as opposed to
derivation. The first one of these two articles suggests that the
Functional Determination algorithm, coupled with a representational
definition of variable, is in certain respects inadequate to the task
of accounting for Strong Crossover (SCO) configurations, by appeal to
Condition A and Condition B of the Binding theory. That is, these
algorithms which apply to representations in order to determine the
feature content of certain categories appearing in these
representations over generate certain Strong Crossover configurations.
The intermediate representations are in fact implicitly postulated even
in this "representational approach"--namely, those representations
containing empty categories whose featural content is not yet
established. The functional determination algorithms then applies to
these representations and maps them into another representation in
which the feature content of the empty categories is specified.

Thus, implicitly there is D-structure , transformations, another
structure, application of the Functional Determination algorithm, and
ultimately S-Structure. This seems to be derivational, implicitly
postulating transformed intermediate representations (neither D-
Structure nor S-Structure) from which the application of Functional
Determination derives S-Structure representation. An alternative
account is suggested to take care of this problem in which the feature
content of "empty" or phonologically null is determined by formal
properties of the rule that created it. This avoids the problem
confronting an analysis incorporating both Functional Determination and
the representational definition of "variable". This "Intrinsic Feature"
account, whereby the mover determines properties of the trace accords
with current conceptions of the trace theory, in particular the copy
theory as represented in Chomsky(1995).

The last article argues against defining any syntactic relations on
trees/representations, and suggests instead that such relations might
be motivated by partially-ordered rule applications that constitute the
minimalist derivation. This derivational approach to syntactic
relations is explored further in Epstein, Groat, Kawashima and
Kithahara (1998) and is being investigated in ongoing research. This
theory of syntactic relations extends the rule based Minimalist
approach in an attempt to derive syntactic relations from the
properties of the derivation-certainly preferable to defining relations
on trees without explaining why these particular definitions/tree
configurations are syntactically significant. Similarly
filters/principles, that is definitions of what constitutes an ill-
formed property of a tree, are targeted for elimination, and the Y-
model itself, with the two interface levels derived only at the end of
the line--that is, post-transformationally--is significantly revised,
yielding a level-free system with interpretation of each
transformational operation-that is derivational interpretation
interpretation-with the possible elimination of trace theory, chain
theory, or some subparts of these systems of representation. The idea
is that traces and chains (among other representational postulates) are
induced by the Y-model's components postponement of interpretation
until after transformational application is complete, necessitating
representational annotations that allow important aspects of the
derivational history to be encoded in the representational output.

The first article not only relates to the last, but is also related to
the sixth article. The first article discusses potential empirical
problems confronting an elegant and in a sense eliminative analysis of
Strong Crossover, that is, one that eliminates appeal to Condition C of
the Binding Theory in accounting for SCO. The sixth article is also
concerned with problems confronting another elegant eliminative
analysis, namely Visibility analysis seeking to eliminate the Case
Filter-a worthy goal, since filters are, by their very nature, non-
explanatory, describing what is an illicit representation. The article
constitutes ac se study of a broader methodological issue confronting
linguistic research: namely how we might determine which
rules/principles exclude which data.

Scope and Quantification are the central aspects of the second article,
which has two central goals. The first is to try to explicate the
precise meaning of "arbitrary interpretation", a type of interpretation
perhaps unique, borne only by PROarb. The second goal is to try to to
eliminate, or at least reduce, the number of cases presumed to be
instances of uncontrolled PRO. This article was probably the first the
first to suggest that a non-null-subject grammar, namely English, had
pro in its lexical inventory. The basic idea is that uncontrolled
PROarb is in fact controlled PRO, controlled by an implicit argument,
pro, which itself undergoes pronominal coreferent interpretation or, if
free, is interpreted as a universal quantifier, and as such undergoes
Quantifier Raising in LF. Arbitrary interpretation is thereby
reanalyzed as universal quantification, and at least one type of
allegedly uncontrolled PRO is reduced to PRO that undergoes
obligatorily control.

The third article examines different definitions of the central
relation "local binding" and investigates an independently motivated
analysis within which the definitions are empirically distinguishable.
The analysis concerns the well-formedness of the analogous, but overt,
anaphoric cliticization in Italian. Here, as in the eight article, he
adopts the influential and intriguing idea that in certain grammars
there exist LF operations akin tot hose found overtly in other
grammars, the fundamental idea being that the syntactic derivation
which undergoes phonological interpretation is subject to parametric
variation. Under this kind of an analysis the perceived vast or
infinite diversity of human languages or grammars is an observed
illusion. This very same leading idea is adopted in the eighth article
in which he suggests that English incorporates covert operations
generating verb-second configurations in the LF component.

The fourth article examines a potential problem confronting May's
influential segmental theory of adjunction. Perhaps the central facet
of the segmentation analysis, in direct contrast to category-
duplicating adjunction, is that it predicts that an adjoined category
commands a slightly larger domain than is predicted under duplication.
The author suggests that this slightly increased "permissiveness"
results in the over generation of certain cases of quantificational
weak crossover. Scope and quantification re also central topics of
second article.

The fifth article seeks to deduce an unrecognized prediction concerning
quantifier interpretation, from the independently-motivated Null
Operator analysis of certain constructions. The question posed was why
is there no narrow scope reading, derived by Quantifier Lowering of the
subject quantifier. The idea was that although quantifier lowering from
a theta less position is allowed, the lowered quantifier adjoined to
the embedded infinitival projection (IP) would ail to bind the Null
operator occupying the embedded [Spec, CP] at LF. At the same time,
binding of the null operator by the matrix subject, a theta less trace
of quantifier lowering, fails to "strongly bind" the null operator,
which as a result present yet uninterpretable at LF, yielding an
illicit representation at that level, thus providing independent
evidence for the presence of a Null Operator in the LF representation
of such sentences. The article addresses three general topics directly:
Adjunction, Quantification and Local binding.

The sixth article argues against a reduction of the case filter,
suggesting instead that this filter is a distinct, differentiated
principle of Grammar. This form of argumentation is applicable to all
proposals regarding the organization of syntactic theory and thus
provides a method for determining the precise nature of differentiation
and reduction throughout the syntactic component. The methods employed
in this article are presumably applicable to research concerning the
organization of non syntactic components as well.

The seventh article seeks to eliminate a number of descriptive filter
and constraints, by reducing them all to a single derivational economy
condition, last resort, in the sense of Chomsky (1991). The author
investigates properties of certain A-bar chains, and seeks to derive
them from overarching derivational constraints, under which the
"function" of Movement, namely to optimally derive convergent interface
representations containing no illegitimate morpho lexical features;
that is, features that are (naturally) uninterpretable at the
interfaces. Again, the attempt is to explain properties of A-bar chains
merely described by a formal apparatus like filters on representation,
or descriptive, ad hoc constraints on movement and/or chain formation.

The eighth article suggests that English incorporates covert operation
generating verb-second configurations in the LF component. In addition,
overt scope-marking conditions are investigated, and it is also
proposed that certain LF representations (derived by covert V movement
ultimately to C) are category-neutralized VP-recursion structures
lacking functional checking categories and their projections. This
checking-induced deletion and the resulting derived LF constituent
strictures are argued to permit simplification of index-dependent,
head-government ECP requirements. This article also proposes a "mixed"
theory of adjunction (as recently advocated by Lasnik and Saito (1992)
while covert adjunction segments the target (May 1985). Most of the
articles published in this book are well known. Syntacticians working
on LF, Binding, and Control will benefit immensely from this book.
Though some of the articles are within a not-so current framework of
Chomsky's, they nevertheless bear insights that can are important for
current investigations. Chomsky's articles in recent years have
acknowledged inputs from Epstein.

References
Epstein, S. D., E. Groat, R. Kawashima, and H. Kithahara (1998) A
Derivational Approach to Syntactic Relations, New York: Oxford
University Press

Lasnik, H., and M. Saito (1992) Move Alpha, Cambridge, Mass.: MIT
Press.

May, R. (1985) Logical Form: Its Structure and Derivation, Cambridge,
Mass.: MIT Press.

About the reviewer: After working extensively generative theory of
Chomsky, I have got down to writing grammars of individual languages. I
have been primarily working on Indo-Aryan, Tibeto-Burman and Austro-
Asiatic languages. Of late, I have taken an interest in sociolinguistic
research.