Prosodic representations (Section on tonal alignment) - Archive ouverte HAL Accéder directement au contenu
Chapitre D'ouvrage Année : 2011

Prosodic representations (Section on tonal alignment)

Mariapaola d'Imperio

Résumé

In the last 20-30 years, there have been numerous arguments from cognition, linguistics, and technology for assuming that the words, on one side, and the tonal specification of an utterance, on the other side, are stored and created independently of each other (Liberman 1978, Goldsmith 1979, Pierrehumbert 1980; see Ladd 2008 for a synthesis of the autosegmental-metrical approach of intonation, or AM theory). But this independence comes at a price: tune and words have to be synchronised with each other in time. Tonal alignment is concerned with how target tones are temporally coordinated, or synchronized, with prosodic units (e.g., syllables) and their constituents (the segments that make up syllables). A large body of alignment studies in laboratory phonology have therefore explicitly tested the basic tenet of the AM theory, i.e. that underlying tonal structure is reflected in the signal through the presence of well-defined fundamental frequency (f0) targets, which are specified both in terms of tonal alignment and according to some well-defined melodic value (scaling). Among the two “coordinates”, alignment with the segmental string (i.e., the temporal distance from edges of segments or syllables) has been the most studied aspect so far, including acoustic, production and perception studies, all of which will be reviewed below. Earlier alignment studies set out to question the controversial nature of intonational primitives: finding evidence for the existence of static tonal targets characterized by stable temporal and melodic coordinates calls into question the validity of the primacy of intonational movements (traditionally represented by the British School approaches and the IPO system (cf. 'O Connor & Arnold 1961, 't Hart & Collier 1975). Indeed, the question of the alignment of intonation contours and their impact on perception was first investigated in contour approaches to intonation. For instance, the IPO school, while developing a rule-based generative approach to intonation, extensively investigated the alignment of rises and falls and found that an early vs. late alignment with the segmental string could induce different interpretations of the contour, hence different meanings (cf. §9.3 below). Here I shall more closely concentrate on the insights of AM theory in stimulating laboratory investigations of tonal alignment. Despite the growing attention of the prosody community to alignment studies, there is still a great deal of controversy about the forces as well as the exact mechanism driving the alignment of tonal targets with segmental gestures, and many issues remain unsettled, as will be shown in this section. Among the most significant points in question are how best to tease apart the phonetics and the phonology of alignment, and how to identify universally motivated tendencies (both from a production and a perception standpoint) in both intonational and tonal languages. It should be noted that the choice of concentrating upon tonal alignment regularities relative to specific segmental landmarks or regions is generally not theoretically (acoustically, perceptually and/or articulatorily) driven, making the study of alignment mostly an exploratory enterprise at this point, though certainly not devoid of interest.
Fichier non déposé

Dates et versions

hal-01482606 , version 1 (03-03-2017)

Identifiants

  • HAL Id : hal-01482606 , version 1

Citer

Mariapaola d'Imperio. Prosodic representations (Section on tonal alignment). Abigail Cohn, Cécile Fougeron, and Marie Huffman. The Oxford Handbook of Laboratory Phonology, Oxford University Press, pp.275-287, 2011, Oxford Handbooks. ⟨hal-01482606⟩
140 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More