home | about | graduate | undergraduate | events | people | labs | papers | jobs | contact

Events

Department Events Calendar

Friday Colloquium Series

The Phonetics/Phonology Lecture Series

The Semantics Lecture Series

The Syntax Brown Bag

The Working Group in Urban Sociolinguistics

The LANYU Forum

Workshop on Romance Object Clitics

Recent Conferences

Events in the Tri-State Region

Other NYU Events

The universality of the prosodic hierarchy: Experimental evidence from Japanese

Shigeto Kawahara,

Thursday, March 13, 2008, 9:00 AM

---

The prosodic hierarchy defines the domains of phonological processes and determines the distributions of boundary tones. The prosodic hierarchy also affects articulatory controls, which are particularly manifested in domain-edge strengthening and final lengthening. In the words of Hayes, therefore, prosodic hierarchy serves as "a general organizing principle for the phonology" (1995: 82). Against this background, I address the following question: does Universal Grammar admit language-particular variations in the number of prosodic levels? Following Selkirk (2005), I answer negatively to this question, suggesting that all languages should have all prosodic levels admitted in the Universal Grammar.

Pursuing the universality of the prosodic hierarchy, Selkirk (2005) advances a general theory of syntax-phonology mapping where prosodic levels are "phonological reflexes" of syntactic levels, and hence universal. In particular, Selkirk asserts that a phonological Intonational Phrase (IntP) is aligned with a syntactic CommaP (i.e. ALIGN(CommaP, IntP)), where a CommaP includes a syntactic clause, an epithet, a parenthetical phrase, etc (Potts 2003). This proposal appears to contradict previous studies on Japanese intonation, in which Japanese was thought to lack an IntP (Beckman and Pierrehumbert 1986; Pierrehumbert and Beckman 1988; Venditti 1995). The experiments I present in this talk show that, contrary to the previous claims, we do find evidence for the IntP in Japanese, when we look at the intonation of various examples of the CommaP.

Experiment 1 uses the intonation of gapping and coordination, which are typical multi-clause constructions in Japanese (Kawahara and Shinya in press). The results show that the edges of clauses exhibit phonetic properties that differ from those observed at the edges of Major Phrases (MaP). In particular, the IntP is characterized by a distinctive large initial rise and pitch reset in initial position as well as tonal lowering, creakiness and a pause in final position. In short, the experiment shows that each syntactic clause-or CommaP-corresponds to a phonological IntP.

Experiment 2 investigates the intonation of nominal parentheticals, another instance of a syntactic CommaP. The results again show that the edges of parentheticals exhibit prosodic properties that differ from the MaP edges. The left edges of nominal parentheticals show strong pitch reset and a large initial rise. In addition, their right edges are signaled by a pause and creakiness. These properties correspond to the properties of the IntP identified in Experiment 1.

Overall, the experiments demonstrate that various instances of the syntactic CommaP are aligned with IntP boundaries, showing that a general syntax-phonology mapping principle-ALIGN(CommaP, IntP)-is at work in Japanese. I conclude that theoretical considerations such as Selkirk's can help guide experimental design by showing us what to look at in our experiments. In turn, experiments can contribute to addressing theoretical questions.

---

nyu | gsas