Towards a model of intonational phonology of Turkish: Neutral intonation

Abstract

This study proposes an Autosegmental-Metrical model of Turkish intonation based on sentences produced in neutral focus, as part of our ongoing research investigating Turkish intonational phonology. Tonal patterns of utterances were examined by varying the length of a word and a phrase, the location of stress, syntactic structures, and sentence types. Preliminary results suggest that Turkish has a H* pitch accent, realized on the stressed syllable of most content words. Each content word forms one Prosodic Word (PW) whose left edge is marked by an L tone. There are two prosodic units higher than PW: an Intermediate Phrase (ip) marked by a final rising (LH) tone and an Intonational Phrase (IP) marked by various types of a final boundary tone. These three prosodic units are also distinguished by the degree of juncture. Interestingly, the ip-final LH boundary tone marks the right edge of a heavy syntactic constituent regardless of the length of the unit. Furthermore, the left edge of a nuclear pitch accent is also marked by a rising tone (LH) which is realized on the last syllable of the immediately preceding PW. The ip-final LH tone and the pre-nuclear LH tone are phonetically different and perceptually distinct.

Constraints on syntax-prosody correspondence: The case of clausal and subclausal parentheticals in Turkish

“…6 shows that the parenthetical is the same level as surrounding non-final-ws. The final rise of the w that was adjacent to the nucleus was often higher than the final rise of the preceding pre-nuclear ws, confirming the observations of İpek and Jun (2013). Fig.…”

“…The accented word (left) was elicited after the following prompt: ''Please say out loud the town name that is written here.'' 10 See İpek and Jun (2013) for a similar point of view where ws that end with a finally stressed v are marked with an accented edge tone; H*-. Whether H-is actually an accented tone in the case of finally stressed words requires further investigation, and quantitative support.…”

“…The question how an edge tone (without an accent) is also a spreading tone lacks theoretical motive and requires a detailed quantitative support. İpek and Jun (2013) is another study that addresses this issue. These authors assume a post-lexical downstepped high pitch accent docked on the right edge (final syllable) of the head of the final-w, !H*.…”

“…These authors assume a post-lexical downstepped high pitch accent docked on the right edge (final syllable) of the head of the final-w, !H*. For İpek and Jun's (2013) account, that there is no acoustic evidence to suggest a high tone associated with a pitch accent (Kamali, 2011;İpek, 2011;Güneş, 2013a) is a crucial issue that should be addressed in detail considering data. Additionally, a tone can only be downstepped relative to another preceding pitch accent in the same w. In the current annotation, for the cases of finally stressed vs, no accent precedes the final syllable of the nucleus in its w. How such a downstep is motivated without a preceding T* is another issue concerning their account.…”

“…a H-in the case of the arguments and presumably H-, H% or L% in the case of parentheticals). For the case of H-, the presence of a non-final accent does not affect the edge tones (Kamali, 2011;İpek and Jun, 2013). The current paper focuses only on the use of boundary or edge tones.…”