CUNY Psycholinguistics Supper: Nazik Dinçtopal-Deniz (The Graduate Center, CUNY)
SEP 24, 2013 | 6:30 PM TO 8:00 PM
The Graduate Center
365 Fifth Avenue
September 24, 2013: 6:30 PM-8:00 PM
The first meeting of CUNY's Psycholinguistics Supper in the Fall 2013 series will be held on Tuesday 24 September:
The speaker will be Nazik Dinçtopal-Deniz (The Graduate Center, CUNY), whose talk has the title "Effects of constituent length on syntactic ambiguity resolution in Turkish: Evidence from reading and listening".
ABSTRACT. This study investigates effects of constituent length on parsing decisions for a Late Closure (LC)/Early Closure (EC) ambiguity in Turkish. In the examples in (1), the homophonous morpheme -u on the noun psikoloğu (psychologist) can be interpreted either as possessive or as accusative, changing the interpretation of the subject of the embedded clause.
(1) LC / EC
Ø (Yaklaşık) yirmi öğrenci-nin psikoloğ-u
Pro Nearly twenty student-GEN psychologist-3SG.POSS/ACC
(oldukça) sev-il-di / sev-diğ-in-I san-dı-m.
much like-PASS-PAST / like-FN-3SG.POSS-ACC think-PAST-1SG
LC I thought that the psychologist of (nearly) twenty students was (much) liked.
EC I thought that (nearly) twenty students liked the psychologist (much).
A late-closed subject ends at the POSS morpheme. An early-closed subject ends at the GEN morpheme, and the following ACC-marked noun is the object. In Turkish the subject is followed by a prosodic boundary, as it is the default topic. So in spoken form, boundary placement can disambiguate the temporary ambiguity. In written form, it is resolved only at the subsequent verb, whose morphology signals an LC-compatible intransitive (passive, was liked), or an EC-compatible transitive (active, liked).
To test whether constituent lengths influence parsing decisions, we included either nearly in the subject or much in the VP (total 6 PWds in all cases).
In the reading experiment, length-averaged underscores replaced the disambiguating morphemes, which participants had to supply as they read aloud. Results revealed significant effects of phrase lengths on readers' placement of prosodic boundaries, accompanied by a significant shift in their syntactic/semantic interpretations as indicated by the morphemes they inserted.
In the listening experiments, six conditions manipulated syntax (LC/EC) and spoken prosody (cooperating/conflicting/neutral). Participants made an end-of-sentence 'got it' judgment to report successful comprehension. Results supported an interplay of constituent length effects and syntactic Late Closure in listeners' interpretation of cooperating and conflicting prosodic cues. Results for the neutral prosody condition suggested that listeners mentally project missing prosodic boundaries using the syntactically-defined Late Closure strategy, with a bias toward balanced phrase lengths. Thus, results overall suggest that constituent lengths have powerful effects; yet, the syntactic Late Closure strategy cannot be eliminated in favor of prosodic interface constraints.