The study seeks to establish regularities in carryover coarticulation induced by the first unstressed vowel and the intervocalic consonant on the second vowel in VCV clusters in the Russian language. The experiment is conducted on nonsense words of the 'etV1CV2te' structure. The V1 vowel is presented by the unreduced unstressed vowels e, a and u; the intervocalic consonant C is presented by the sounds p, t and k; the V2 vowel is presented by the emphatically stressed e, a and u. Two male native speakers read the corpus of 540 words. The PRAAT software was used to measure F2 formants in the V2 vowels. ANOVA was employed to process the data obtained. It was established that in Russian VCV structures both the front/back dimensionality of the unstressed V1 vowels and the place of obstruction of the intervocalic consonants have a clearly pronounced statistically significant effect on the F2 formant values of the emphatically stressed V2 vowels. The interaction of the two factors also proved to be significant. However, the ANOVA procedures did not make it possible to establish any clearly defined regularities in the carryover coarticulation impact either of the unstressed V1 vowel front/back row or the place of obstruction of the intervocalic consonant on the F2 value of the V2 vowel. It was established that speaker-specific differences heavily affect the coarticulation in VCV structures. The results obtained are interpreted using existing findings.
Sergey Batalin (Thu,) studied this question.