[lingtalks] TODAY 2pm: Rodrigo Gutierrez Bravo (Linguistics Colloquium)

[Roger Levy]

Today at 2pm, Rodrigo Gutiérrez-Bravo will be giving a Linguistics 
Department colloquium in AP&M 4301.  Title and abstract follow.

Please note that an earlier announcement in which the time was given as 
2:30 was in error.  The colloquium will begin at 2pm!

***

The Subject-initial/Verb-initial Alternation in Optimality Theory

Rodrigo Gutiérrez-Bravo/ CIESAS-Mexico City
Jorge Monforte y Madera / Academia de la Lengua Maya de Yucatán

     Many theories (and especially Optimality Theoretic Syntax) assume a 
one-to-one relation between the constructs of the semantic and 
information structure components and the syntactic component. In other 
words, for every distinct semantic/pragmatic representation, there is 
one (and just one) distinct syntactic representation (for instance, Kuhn 
2001, 2003). In this talk we provide evidence that this is not always 
the case and that there are cases when two different semantic/pragmatic 
representations map into the same syntactic structure, a process akin to 
neutralization in phonology. We argue that this kind of mismatch occurs 
because there are cases when syntactic well-formedness conditions 
override what would be the ideal mapping from the semantics/information 
structure components into the syntax. We illustrate this phenomenon with 
languages that appear to show a “free” subject-initial and verb-initial 
word order alternation (Greek, Venezuelan Spanish, and Yucatec Maya 
(Mayan)) and then develop and Optimality Theoretic account of this 
alternation. First, we show that even though these languages are alike 
in that they display this word order alternation, they cannot be grouped 
together. This is because whereas Greek can plausibly be analyzed as 
having VSO as its unmarked word order (SVO thus resulting from 
topicalization of the subject), the evidence points to the conclusion 
that Spanish and Yucatec are SVO languages, and so verb-initial 
transitive constructions in these languages need to be derived with a 
different mechanism than the one that derives VSO in Greek. Building on 
the distinction between thetic and categorical judgements, we argue that 
the verb-initial transitive constructions in Spanish and Yucatec result 
from the feature [thetic] being a part of their semantic content. We 
formalize the mapping of this feature into the syntax as a violable 
constraint requiring that no referential XP c-command the highest 
inflectional head of the clause. In this analysis, if the semantic input 
lacks the [thetic] feature in Yucatec, the result is an SVO clause, and 
when this feature is present the result is VOS. This same result extends 
to Venezuelan Spanish, but crucially not to varieties of Spanish that 
disallow the SVO/ VSO alternation. Because of a high-ranking EPP 
constraint in these varieties, the inputs with and without the [thetic] 
are neutralized into a single SVO output. Similarly, if the analysis of 
Greek as a VSO langue is correct, then the inputs with and without the 
[thetic] feature result in the same output (VSO), again a case of 
neutralization. We conclude by suggesting that, in essence, the verb 
initial constructions of Yucatec and Venezuelan Spanish are the 
equivalent to transitive expletive constructions of the Germanic 
languages (Alexiadou & Anagnostopoulou 1998). In our OT analysis, the 
difference between them (the presence of an expletive in Germanic 
languages) results not from any cross-linguistic difference in the 
semantic or pragmatic content of these constructions. Rather, the high 
ranking of the EPP constraint in the Germanic languages forces the 
insertion of an expletive in the preverbal position, whereas this 
constraint is lower ranked in Venezuelan Spanish and Yucatec, the result 
being VSO and VOS clauses, respectively.