J. C. Roux

Voice quality differences associated with stops and clicks in Xhosa

Abstract Voiced stops/clicks in Xhosa cause f0 depression in the following vowel. Some authors also claim that these sounds are followed by breathy phonation. An appropriate set of words with different stop and click categories was read by eight speakers of Xhosa. Voice quality was measured asH*1 –H*2 andH*1 –A*3. f0 andF1 were determined as well. Measurements were made for four periods early in the following vowel and one close to its center. The occurrence of breathy voice after voiced stops/clicks was inferred from levels ofH*1 –H2 andH*1 –A*3 about as high or higher than those generally expected closely after voiceless aspirated stops. The existence of f0 depression after voiced stops/clicks could be confirmed. F1 was usually highest after aspirated and lowest after voiced stops/clicks. Indications of breathy voice after voiced stops/clicks were found for some speakers only. It is argued that extensive larynx lowering and vocal fold slackening can explain the specifics of the voicing feature in Xhosa. A similar situation occurs in Shanghai Chinese. Based on that comparison it is suggested that “slack voice” is a more appropriate term for the relevant Xhosa sounds than “breathy voice”.

Phonetic analysis of Afrikaans, English, Xhosa and Zulu using South African speech databases

We present a corpus-based analysis of the Afrikaans, English, Xhosa and Zulu languages, comparing these in terms of phonetic content, diversity and mutual overlap. Our aim is to shed light on the fundamental phonetic interrelationships between these languages, with a view to furthering progress in multilingual automatic speech recognition in general, and in the South African region in particular.

A phonological process analysis of the acquisition and loss of clicks in Xhosa

The aim of this article is to describe the nature and extent of phonological tendencies evident in the acquisition and loss of clicks in the Xhosa language. Based on these tendencies numerous phonological processes are identified that have not been documented in previous research. 41 Xhosa speaking children aged between 1;6 and 5;5; 18 English and Afrikaans speaking children acquiring Xhosa as a third/foreign language in the primary school; and three Xhosa speaking aphasics were tested. The results firstly indicate that four major phonological processes appear to be operative in the acquisition and loss of clicks by the three subject groups. Secondly, certain results also indicate that the three subject groups tend to prefer certain phonological processes when producing these sounds.

On the perception and production of tone in Xhosa

This article is concerned with the perception of tone in speech and its consequences for tonal descriptions and the formulation of phonological theories. Confusions in the perception of Xhosa tonal data are discussed. The role of relatively high intensity levels and its influence on tonal perception is illustrated. The existence of a mid tone in Xhosa is critically evaluated in view of experimental phonetic data. An experiment regarding strategies in the production of tonal sequences in Xhosa is described in an attempt to explain discrepancies in tonal data.

Prosodic data and phonological analyses in Zulu and Xhosa

In this article the status of primary data in phonetic descriptions and phonological analyses is assessed. The nature of tonal data found in descriptions in Nguni and Sotho languages is discussed and it is indicated that this data more than often lack authenticity and reliability and are subject to questionable methods of acquistion. Inconsistencies in various data sets are described and an acoustic analysis of Xhosa data illustrate that tone may not be the only disambiguating factor in lexical items. The relationship between phonetic data and phonological analyses is demonstrated with reference to an autosegmental analysis of Zulu nouns. An integrated methodological approach is proposed for the study of speech phenomena.

Computer-assisted phonetic research in African languages

The computer may be used in various ways in phonetic and phonological research. This article focuses on the use of the computer in speech signal analysis and the description of the phonemic structure of different African languages. S. Afr. J. Afr. Lang. 1986, 6: 42–45

On vowel identification and phonological theory

This paper focuses on some recent calls for a more substance-based approach to phonetic and phonological descriptions, and presents some . 2) concrete examples from Sotho to demonstrate some of the inadequacies found in traditional approaches.

Grapheme-to-phoneme conversions in Xhosa

Grapheme-to-phoneme conversions underlie, inter alia, computer-assisted investigations into the nature of the phonotactic structures and into the occurrence and distribution of phonetic and phonemic elements of a specific language. Prior to any such investigation, large text corpora need to be transcribed phonetically and/or phonemically. In this article a computer-assisted system developed to transcribe Xhosa texts phonetically is described. Prerequisites for setting up such a system are described, specific transcription rules for Xhosa are presented, and the effectiveness of the system is discussed.

Acoustic and perceptual qualities of queclaratives in Xhosa

This paper reports on a study of the acoustic and perceptual qualities of queclaratives in Xhosa. Queclaratives, or statements which are question-like in function, have been studied in a number of languages of the world. Many of these studies have implemented experimental techniques at acoustic and perceptual levels, however, prior to this study, this has not yet been done for a language such as Xhosa. In this paper prevalent impressionistic and experimentally based claims, differentiating statements from questions, such as word and penultimate syllable duration, speech tempo and tonal register to name but a few will be listed. It will be demonstrated that although these descriptions all refer to some kind of physical difference between statements and queclaratives each description seems to highlight a different aspect of what is supposed to represent the distinctive quality.

Perceived duration in vowel-length based Civili minimal pairs

Several authors have established the existence of short and long vowels on the basis of minimal pairs occurring in Civili. These works however simply refer to impressionistic phonetic observations and do not rely on experimentally-gathered and verified data. This article focuses on the perceived duration of components of minimal pairs. It is intended to assess the validity of claims on vowel length and minimal pairs in the language. The article gives an account of a perception experiment that was conducted and describes the administration of perception tests. The perceptual data gathered are then statistically analysed and discussed in terms of phonological implications.