Working Papers in Linguistics: Volume 36 (September 1987)

Permanent URI for this collection

Papers from the Linguistics Laboratory 1985-1987. Edited by Mary Beckman and Gina Lee


Front Matter
Description | Full Text PDF

Intonation in Cantonese
Johnson, Keith pp. 1-15
Description | Full Text PDF

A Study of Toishan FO
Lee, Gina pp. 16-30
Description | Full Text PDF

The Tonal Behavior of Osaka Japanese: An Interim Report
Kori, Shiro pp. 31-61
Description | Full Text PDF

Perceiving by Syllables or by Segments: Evidence from the Perception of Subcategorical Mismatches
Johnson, Keith pp. 62-74
Description | Full Text PDF

Some Constraints on the Domain of Phrasal Resyllabification in French
Laeufer, Christiane pp. 75-100
Description | Full Text PDF

Phonemic Quantity, Stress, and the Half-Long Vowel in Finnish
Välimaa-Blum, Riitta pp. 101-119
Description | Full Text PDF

Phonetic Characteristics of Levantine Arabic Geminates with Differing Morpheme and Syllable Structures
Miller, Ann M. pp. 120-140
Description | Full Text PDF

Browse

Recent Submissions

Now showing 1 - 8 of 8
  • Item
    Front Matter (Volume 36, September 1987)
    (Ohio State University. Department of Linguistics, 1987-09)
  • Item
    Intonation in Cantonese
    (Ohio State University. Department of Linguistics, 1987-09)
    The experiment in this paper explores the nature of intonation in a language which has lexical tone. In a pilot study it was found that a method of accounting for tone preservation (the identifiability of lexical tones in sentence contexts) which included a declining tone space was better suited to the task than one which assumed a level tone space. The main experiment attempted to separate and observe the contributions to this general downtrend made by boundary effects, tonal interaction and declination. There appears to be evidence for one type of boundary effect (initial raising) and declination. The data of this experiment were not consistent with the presence of the other type of boundary effect (final lowering) or tonal interaction factors. Two important variables were manipulated in this experiment. First, the length of a test sentence was manipulated on the assumption that longer sentences would show a greater decline of FO if there was a declination effect. Second, the discourse position of test sentences was varied (from discourse medial to discourse final) as a test for the effect of discourse final lowering.
  • Item
    A Study of Toishan FO
    (Ohio State University. Department of Linguistics, 1987-09) Lee, Gina
    Like other Chinese languages, Toishan uses tone to signal differences in meaning among words. With the exception of certain morphologically conditioned tone changes, which must be memorized by speakers, Toishan exhibits no tonal modifications; in particular, there appear to be no tonal modifications which are strictly phonologically governed. Given the absence of such changes, Toishan provides an ideal situation for examining the hypothesis that declination--a phonologically unmotivated lowering of FO, independent of tone--is a relevant component in a model of intonation. Three native speakers read paragraphs containing five pragmatically connected sentences of different length. Two target tones within each sentence were measured. No evidence for declination was found. The FO pattern is described in terms of an overall structure involving pitch range expansion for initial sentences and compression (or lowering) for final sentences. The results are discussed in light of evidence from other languages and implications for a model of local FO implementation.
  • Item
    The Tonal Behavior of Osaka Japanese: An Interim Report
    (Ohio State University. Department of Linguistics, 1987-09) Kori, Shiro
    This paper reports the results of a preliminary investigation of the FO characteristics of Osaka Japanese. It describes the realization of word accents in sentences and the interaction of focus with word accents. It discusses the basic unit of phrasing in Osaka Japanese sentences, and suggests that this unit is more closely tied to the word than in Tokyo Japanese because a word does not easily lose its acccentual integrity in larger prosodic contexts. Several other important differences from Tokyo Japanese tonal patterns are also revealed. First, Osaka has two distinct kinds of sharp FO falls, one at the accent nucleus (as in Tokyo) and the other at the boundary of a High-- ending word and a following Low-beginning word. Second, while the FO contour over a sequence of supposedly H-pitched moras has a slow decline, the rate of this downtrend is constant and independent of the size of sequence, suggesting that, unlike in Tokyo, there is H-tone spreading in Osaka Japanese. Uptrends in L-beginning words are also examined, as well as the interaction of L tones with focus.
  • Item
    Perceiving by Syllables or by Segments: Evidence from the Perception of Subcategorical Mismatches
    (Ohio State University. Department of Linguistics, 1987-09) Johnson, Keith
    This paper describes an experiment in which two general hypotheses concerning speech perception are tested. According to the segment perception hypothesis the acoustic signal is interpreted in terms of segments analogous to those used by phoneticians in transcribing speech. The syllable perception hypothesis on the other hand holds that the speech signal is perceived in terms of syllable sized units. The experiment tests these two hypotheses by presenting subjects with a perceptual task for which the two make opposite predictions. Tokens with subcategorical mismatches were produced by cutting the fricatives [s] and [ʃ] from VC syllables (vowels were [i,a,o,u]) and recombining them with vowels which differed from the original context in terms of transitions and rounding. The segment perception hypothesis predicts that in syllables with transition mismatches (ie. transitions for [s] and with [ʃ] actually occurring) coarticulatory rounding on the actually occurring fricative will aid in the perception of [ʃ] and slow the perception of [s], while the lack of rounding on the actually occurring fricative will have the opposite effect. This is because the rounding makes [ʃ] a more extreme example of [ʃ] (and thus easier to categorize as such) while rounding makes an [s] less distinctly an [s]. The syllable perception hypothesis predicts that in syllables with transition mismatches coarticulatory rounding on the actually occurring fricative will aid the perception of [s] and hinder[ʃ] perception. This is because the [s] with rounding is acoustically closer to the prediction made on the basis of the transition on the vowel. Similarly, the [ʃ] with rounding is acoustically further removed from the [s] which is expected as a result of the transitions on the vowel in a mismatched syllable and thus should require more time to be perceived as [ʃ]. The results of the experiment reported here support the segment perception hypothesis. Subjects' perception of [s] in syllables with transition mismatches was inhibited by coarticulatory rounding while their perception of [ʃ] in syllables with transition mismatches was facilitated by coarticulatory rounding.
  • Item
    Some Constraints on the Domain of Phrasal Resyllabification in French
    (Ohio State University. Department of Linguistics, 1987-09) Laeufer, Christiane
    This paper examines the French phrasal resyllabification process known as enchainement 'linking' in the light of three recent proposals about connected speech phenomena and their relation to syntactic structure. It is based on a corpus illustrating linking consonants and initial consonants in minimal or near minimal pairs across different types of prosodic/syntactic boundaries. The sentences were read five times at three different speaking rates by two native speakers. The study shows that (1) with increasing rate of speech, the domain of the rule becomes larger, irrespective of prosodic structure, and (2) there exists a specific degree of disjuncture beyond which it does not apply in a given tempo. The study thus provides evidence that, at least in French, resyllabification belongs to the phonological rules proper, and not to the rules for defining (post-lexical) phonological representation. Research on phrasal resyllabification rules in other languages and based on a larger number of speakers is, however, needed before any generalizations can be made.
  • Item
    Phonemic Quantity, Stress, and the Half-Long Vowel in Finnish
    (Ohio State University. Department of Linguistics, 1987-09) Välimaa-Blum, Riitta
    An experiment was conducted to compare the duration of the final vowels in Finnish disyllabic words which have short stressed initial syllables -CVCV(V)- with that in words with long stressed initial syllables -CVVCV(V), CVCCV(V), CVVCCV(V). It was found that in the first group the final short and long vowels were systematically longer than in the second group. This is taken to suggest that stress interacts with phonemic quantity: on the initial syllable in the CVCV(V)-words there is a conflict between the phonemic duration of the vowel and the use of duration as a stress cue. This is resolved by spreading the stress-induced duration to the second syllable. Thus even if the domain of stress is a syllable, its realization is not independent of the rest of the phonological structure but it operates in concert with it.
  • Item
    Phonetic Characteristics of Levantine Arabic Geminates with Differing Morpheme and Syllable Structures
    (Ohio State University. Department of Linguistics, 1987-09) Miller, Ann M.
    Phonologists distinguish two types of geminates: tautomorphemic (belonging to a single morpheme and composed of one set of phonological features) and heteromorphemic (belonging to separate morphemes and composed of two identical sets of phonological features). If there is a phonetic difference between geminates due to the fact that the second type has two separate sets of features while the first type has only one set, it would likely be cued by differences in duration or by movement during the duration of the consonant. Arabic provides a means for comparing these types of consonants since it has both occurring in a variety of morphological affiliations as well as in contrasting surface phonological positions. A pilot study and a larger experiment were performed with speakers of Levantine Arabic to compare these conditions. No significant differences in duration occurred, and several occurrences of both types of geminates had apparent releases during their durations. These results show that no phonetic difference was found between heteromorphemic and tautomorphemic geminates in Levantine Arabic. On the other hand, a phonetic difference was found linked to syllable structure: tautosyllabic and heterosyllabic geminates had different mean durations as compared to their single counter-parts. Since the structure that has been posited by phonologists (association of phonological features with C and V slots) does not capture these differences, it is therefore necessary to refer to syllable structure(= association of phonetic features with syllable slots) in order to represent these phonetic differences.