Fonologiese en fonetiese aspekte van slotkonsonantontstemming in die Engels van Afrikaans- en Tswanasprekendes

Abstract

The relationship between die phonetic and phonological aspects of final consonant devoicing in the English pronunciation of mother tongue speakers of Afrikaans and Tswana is investigated in this dissertation. It is also compared to the pronunciation of mother tongue speakers of English. Two hypotheses are set, that the pronunciation of non-sonorant final consonants by Afrikaans and Tswana speakers will be influenced by the neutralization of the voicing distinction, and that both phonetic and phonological aspects will be necessary for the explanation of the differences. In preparation for the empirical study and the theoretical explanations, a literary survey is conducted in Chapter 2. The phenomenon of devoicing is investigated firstly. It is shown that devoicing is incomplete in a number of languages (German, Polish, Catalan, Dutch, Afrikaans and Russian) with putative devoicing rules. Subtle phonetic differences are detected in a combination of parameters such as vowel length, closure duration, voicing into closure and burst duration. No satisfactory explanation of the data is found, because researchers are unable to reconcile phonetic and phonological aspects in the framework of current theories. The pronunciation of final consonants in English is investigated next. It is shown that acoustic differences occur in the vowel length, spectral structure of the vowels, consonant closures, voicing into closure and burst duration for plosives and in the vowel length and fricative duration for fricatives. It is further shown that differences can be explained with reference to the articulatory apparatus utilized for the production of the final consonants' difference in voice quality. This discussion is supplemented with a short reference to the problem of standards of English in South Africa. It is argued that the standards debate does not take pronunciation into account and if it is incorporated in the arguments, no insight in the relationship between phonetics and phonology is displayed. Devoicing is discussed as a second language phenomenon next. It is shown that devoicing is a tendency of second language pronunciation of English and that second

language speakers are not as able as first language speakers when they do produce a difference. It is further shown that second language teaching in South Africa does not pay much attention to pronunciation and that problems are consequently not corrected. It is also shown that phonology theory awards less weight to insights derived at by studies in second language pronunciation. These fields of inquiry, second language pronunciation teaching and second language phonology, are both left behind, but it is shown that both are important for the study of human speech and that they are worthwhile enterprises as long as they take into account the differences between phonetics and phonology, something that doesn't happen frequently at present. In Chapter 3 the results of a production study is presented. .It is shown that English speakers utilize six parameters to produce a difference between underlying voiced and voiceless final consonants: vowel length, transition duration, fricative duration, closure duration, voicing into closure and burst duration. Afrikaans speakers do not use transitions, produce no differences between words ending on underlying voiced and voiceless fricatives and only utilize two parameters, voicing into closure and closure duration to distinguish words ending on long vowels and plosives. Tswana speakers distinguish only between underling voiced and voiceless stops in words with short vowels, utilizing the parameters closure duration and voicing into closure. In Chapter 4 it is firstly indicated that these differences cannot be explained fully using the linear generative phonology. Various limitations are identified, most importantly the inability of the generative analysis to relate diversified phonetic data to phonological categories. Most of the word in this dissertation points to the central importance of the relationship between phonetics and phonology. It is consequently discussed next. A viewpoint that propagates the strict separation between the two disciplines, in addition to a preference for phonology is distinguished from a viewpoint that the two are interdependent. It is argued that the latter viewpoint is the more valid one. On the basis of this, the Integrated Representation System of Clements and Hertz (1994) is discussed with reference to earlier developments such as auto segmental phonology and feature geometry. It is indicated that this approach represents phonological and phonetic information within a single framework by utilizing the .hierarchic ordering principles developed by non-linear phonology. In conclusion, this theory succeeds in explaining the data that eludes explanation within the linear generative model. It is indicated that voiced final consonants spread to neighbouring duration segments during a process of co-articulation. During devoicing, one or more of these association lines are broken. Complete neutralization of the distinction between voiced and voiceless final consonants occurs when all association lines are broken, but various degrees of partial neutralization are also possible. It is proposed that a devoicing parameter is responsible for the extent of the neutralization during devoicing. This parameter is sensitive for the phonological, grammatical and pragmatic context, as well as the extent of second language competence of the speaker. In the final chapter it is indicated that the findings of this study have important implications for the theory of phonology and research in phonology. It is suggested that phonology cannot be studied without taking into account acoustic, articulatory and perceptual information. Further implications are pointed out briefly for the fields of second language teaching and the debate on standards. These possibilities still require a lot of research and no definite conclusions are drawn in this regard.