Main Article Content

Die Afrikaanse diftong /E+/: ‘n Eksperimentele ondersoek


Daan Wissing

Abstract

The nature of the diphthong /E+/ (orthographically presented as ‘ei'/‘y'; in IPA terms transcribed as /ci/; in SAMPA1 /@i]) was acoustically investigated in this experiment. The null hypothesis ‘underlying diphthongs are only realised diphthongally' was made and verified from the speech of eight female speakers. This hypothesis reflects the current view in scientific literature that true diphthongs are phonemes consisting of two existing sounds, which is questioned on the basis of observation. Secondly, the traditional transcription [@i] was also tested. The following spectral parameters, which are relevant for vocal analysis, were analysed and measured: the first two formants, the beginning and end of these formants, the percentage gliding of these, and the gradient of the central 50% of the relevant segment (including the standard deviations of these). Based on the beginning and end values of the first two formants, the percentage gliding (%V) of the two formants was calculated. The duration of the relevant elements (vocal or diphthong) was also measured. From the results, it appears that the perception that the Afrikaans diphthong /E+/ (written as ‘ei' and ‘y') is always realised as a double vowel, is wrong — in almost half of all cases it is monophthongised. In addition, it was found that the transcription [@i] is not a correct version of the diphthong /E+/ and that it should rather be transcribed as [{E].
Die aard van die diftong /E+/ (ortografies weergegee met ‘ei'/‘y'; in IPA-simboolstelsel getranskribeer as /ci/; in SAMPA2 /@i/) is in hierdie eksperiment akoesties ondersoek. Die nulhipotese ‘Onderliggende diftonge word slegs diftongies gerealiseer' is gestel, en aan die spraak van agt vroulike sprekers geverifieer. Hierdie hipotese gee die huidige siening in die wetenskaplike literatuur weer dat egte diftonge foneme is wat uit twee bestaande klanke bestaan. Hierdie aanname word op grond van waarneming bevraagteken. Tweedens, is die tradisionele transkripsie /ci/ of /@i/ ook getoets. Die volgende spektrale parameters wat relevant is vir vokaalanalises is geanaliseer en gemeet: die eerste twee formante, die begin en einde van hierdie formante, die persentasie verglyding daarvan, en die gradiënt van die sentrale 50% van die betrokke segment (ook standaardafwykings van genoemde). Op grond van die begin- en eindwaardes van die eerste twee formante is die persentasie verglyding (%V) van die twee formante bepaal. Die duur van die betrokke elemente (vokaal of diftong) is ook gemeet. Dit blyk uit die resultate dat die opvatting foutief is dat die Afrikaanse diftong /E+/3 altyd as tweeklank gerealiseer word; dit word in byna die helfte van alle gevalle gemonoftongeer. Bowendien is bevind dat die transkripsie [@i] ook nie 'n juiste weergawe van die diftong /E+/ is. Dit moet eerder as [{E] getranskribeer word.

Southern African Linguistics and Applied Language Studies 2005, 23(3): 319–334

Journal Identifiers


eISSN: 1727-9461
print ISSN: 1607-3614