Band 38, Heft 5-6

Laryngeal control occurs mainly along two dimensions. One involves the longitudinal tension of the vocal folds and is used for control of fundamental frequency. The other involves abduction/adduction of the folds. This dimension is used in the vegetative functions of the larynx and in its phonetic function to control voicing and aspiration as well as voice quality. Although fine adjustments in timing of abduction/adduction gestures relative to supralaryngeal events produce contrasts of aspiration in obstruents, variations in the size of these gestures appear to be less significant and less finely controlled. The present experiment explores the control of laryngeal abduction/adduction by examining to what extent speakers can control size of glottal aperture under different conditions, with and without suitable feedback. The results suggest that voluntary control of the size of glottal opening is rather poor, and that subjects are unable to make very fine-graded adjustments along this dimension. Voluntary control of glottal opening in isolation is limited, perhaps because normal activities seldom require separate control of this variable. Instead, control of glottal opening is tightly coupled to such other activities as respiration, swallowing and speech articulation. Even in the context of sound production, glottal aperture is poorly controlled, perhaps because the precise degree of opening (as opposed to precise timing of opening and closing) has comparatively little practical significance over a wide range of openings.

This paper describes investigations of the intrinsic pitch of German vowels, embedded in sentences of equal form. These sentences («lch habe /CVCa/ gesagt»; C = p, t, k; V = various vowels) were spoken several times by 3 speakers. Then the fundamental frequency of the respective vowel V was measured relative to the mean fundamental frequency of each sentence and expressed as a pitch interval. The statistical evaluation concerned the dependence of this intrinsic pitch on the vowel, i. e. its formant frequencies, on the adjacent consonants, and on the speaker. The results are qualitatively consistent with known data and further indicate an (at least partial) central control of intrinsic pitch.

It was the aim of this analysis to obtain information on the importance of such features as quantity and quality of vowels in an emphasized position in German words. In order to do this, the durations of emphasized vowels of spoken words were modified by a digital signal processing technique. The modified words were presented to native speakers for auditory judgments. The results of this research show that there is no uniform answer to the question of the perceptive, and thus phonological relevance of duration of German vowels.

The effects of specific hearer expectations on the categorical perception of speech sound continua were investigated. Subjects were asked to perform identification tests on synthetically produced /ba-da-ga/ continua. Hearer expectations for the /da/ catgory were raised and lowered, respectively, within the experimental group. Only the condition of raised expectation produced a change in the identification of the /da/ category: while the category boundaries themselves did not shift, the perception of /da/ sounds within these boundaries was significantly stabilized. Speech sound expectation apparently adapts to the structure provided by perceptual categories and thus contributes to an optimization of the perceptual process.

The problems of prosodic analysis are discussed within a three-tier framework of phonetic substance, descriptive categories and communicative function. The relationship between the levels is seen as essentially unidirectional, the informational structuring of speech communication being the ultimate determinant. Five communicative functions are specified.

Portuguese (Lisbon dialect) has seven types of tags which do not share, as e.g. in English, a common structural pattern. Their structure is described in the first part of the paper, with special reference to polarity, namely, the common use of constant (negative) polarity questions, and to the setting up of ‘dummy tags’. The second part analyses the consistent use of either a falling or a rising tune only with each tag, in unmarked instances. The forms and meanings of tunes, in attitudinally marked uses are then discussed at some length. Finally, some brief remarks are made as to the use of tags in Portuguese.