Tydskrif vir Geesteswetenskappe
On-line version ISSN 2224-7912
WISSING, Daan. Derounding in Afrikaans. Tydskr. geesteswet. [online]. 2011, vol.51, n.1, pp. 01-20. ISSN 2224-7912.
Afrikaans is a Germanic language, characterised by a rich vowel system, comprising twelve phonemic vowels, among them three abnormal (marked) rounded non-back vowels, viz., in SAMPA notational system /9, y, 2/, corresponding to the more familiar IPA symbols /œ, y, ø/. These sounds are the rounded counterparts of the three normal unrounded non-back vowels /@, i, e/ (SAMPA), in IPA convention transcribed as /ə, i, e/. Rounded non-back vowels are a rarity among the languages of the world. Maddieson (1995) mentions only about 7% of the 535 languages investigated by him. No other case of extensive presence of derounding in languages other than Afrikaans could be traced. It is a well-known view that Afrikaans is typified by a process of active derounding of /œ, y, ø/. This process is an important distinguisher between Afrikaans on the one hand, and the other Germanic languages such as Dutch and Swedish on the other hand. Initially the result of this process was viewed in strong negative terms as being substandard; more recently the presence thereof is described in neutral terms as a typical feature of Afrikaans as language over a broad spectrum. The present contribution concentrated on two additional facets, viz. details concerning production and perception of the vowels in question. In this mixed method study a triangulation design was utilised. This entails quantitative as well as qualitative techniques of data collection and analysis; production over and above perception studies were incorporated with a view to a better understanding of this phenomenon and its characteristics. Apart from a variety of smaller investigations, the main experiment focused on the speech productions of 16 adult speakers of Afrikaans, two male and two female persons each of the age brackets 20-30 years, 40-50 years, 60-70 years, and 80-90 years old. A detailed description is presented of the main acoustic features in terms of the second vowel formant (F2) of these six vowels as produced by the participants. They read a word list as well as some full sentences comprising a balanced set of words containing the vowels in question mentioned above. Furthermore, they produced in a spontaneous manner some frequently used words or phrases that include some of the same vowels. Interesting results were found with respect to a battery of listening tests carried out. A group of listeners (ranging from 14 to 20 participants) had to discriminate between members of roundedunrounded vowels in one instance, and to identify some others in two other tasks. Whereas they generally had no difficulty with auditory stimuli produced by speakers of the oldest group, the same cannot be said for vowels of younger speakers. Most telling is the following fi nding: A word containing either a rounded vowel or its unrounded counterpart (e.g. klere vs. kleure - resp. clothing, colours) when read by a 23 year old speaker were presented auditorily to 14 listeners. Their success rate in identifying the vowels correctly was only 43%. These results too suggest less derounding in the case of the older speaker, and, consequently, also point in the direction of the existence of two chronolects. This in turn also could be a sign of a diachronic shift in the case of the vowel system of Afrikaans. Indications of similar modifications to other vowels of Afrikaans seem relatively clear, but defi nite results are not yet available. This study mainly reveals the following: derounding indeed seems to be a robust phenomenon of Afrikaans. Nontheless, seemingly a complete neutralisation is not present (yet). Even in cases where no salient derounding takes place the phonemic rounded vowels /9, y, 2/ turn out to be much closer in acoustic terms to their unrounded phonemic counterparts /@, i, e/. Typical F2 values that are associated with the presence of derounding are in the case of full sentences: 1819 Hz for unrounded /@/ compared to 1782 Hz for its rounded counterpart /9/, and, likewise, 2298 Hz for unrounded /e/ but 2250 Hz for the rounded member of this pair, viz. /2/. Unrounded /i/ shows a similar small difference with the rounded /y/, namely 2328 Hz and 2349 Hz respectively. These differences are clearly not significant, seemingly sharply different to languages such as Swedish and Dutch, where F2 values for /i/ and /y/ are highly different - values for Swedish were measured as F2's 2747 Hz for (/i/), and only 2477 Hz for (/y/). Respective measurements Dutch (F2 = 2895 Hz; 2327 Hz) show a similar clear difference. Comparatively similar results for vowels in the case of isolated words were found. In this regard I found some indication of the presence of two different chronolects: in some instances the 80-90 year old speakers tend not to deround to the same degree as especially the youngest group (20-30). Older speakers also showed a presence of derounding in their spontaneous speech, but not to the same extent as in the case of the more formal tasks. Furthermore, the youngest group shows a smaller difference between rounded and unrounded vowels than the other groups. Again, the comparative difference in the case of the oldest group is more salient. The findings may have some relevance for language typology and the sociolinguistics of Afrikaans. As the main focus of this study lies on the descriptive level, I will not elaborate on this facet here.
Keywords : Marked rounded back vowels; Chronolect; Derounding.