|Title:||Constraints on Subjects: An Optimality Theoretic Analysis|
|Comment:||1996 Rutgers University Dissertation, 237pp, in 7 files|
|Abstract:|| This dissertation argues for an Optimality Theoretic analysis of
null subjecthood, subject inversion, agreement and structural case
assignment. It does so on the basis of the hypothesis that an analysis
in terms of the interaction of violable, conflicting constraints adds
to the deductive structure of linguistic explanations while simplifying
the definition of the relevant syntactic modules.
Among the most relevant results is a unified analysis of the
crosslinguistic and language-internal distribution of null and
inverted subjects. An initial investigation shows that subjects are
null when referring to antecedents with topic status, and inverted
when focused, a result formalized through the constraints DropTopic
and AlignFocus. The interaction between these constraints and the
constraints Subject and Parse, favoring subjects in preverbal subject
position, determines the distribution of null subjects language-
internally and crosslinguistically, eliminating the need for an
independent pro-drop parameter (Grimshaw & Samek-Lodovici 1995).
A second result concerns expletives, whose language specific
inventories are shown to follow to a high degree from the interaction
between the above constraints and Full-Int, a constraint requiring
that all constituents be interpreted. This shows that expletive
inventories can be derived by way of grammar, with no recourse to
lexical stipulation (PrinceÊ&ÊSmolensky 1993, Grimshaw 1995, Grimshaw
& Samek-Lodovici 1995). The analysis also predicts the universal ban
on overt expletives in null subject languages.
A similar result is pursued with respect to agreement, which is
derived by means of three general agreement constraint-schemata.
Finally, the position of subjects and their case assignment
configuration in Italian declaratives, gerundives and subjunctives
are derived from the interaction between CaseGov, a constraint
requiring case assignment under proper government, and the other
constraints of UG. Once reranked, the same constraints derive
declaratives in Arabic and infinitivals with overt subjects in
English and Portuguese, with no appeal to a parametric account of
abstract case assignment.
Crucially, the analysis of crosslinguistic variation consistently
turns out to be closely tied with the analysis of language-internal
variation, as predicted by an Optimality Theoretic approach to Syntax.