Modeling Intonational Declination in Estonian

    Eva Liina Asu

    University of Cambridge / University of Tartu

    The present paper aims to shed some light on the issue of declinational properties of Estonian statements. Two models are tested.
    According to the 'global' or 'superpositional' model, the F0 is a combination of a global component stretching over the whole intonational phrase, and a succession of local F0 movements defining accents. In the 'local' or 'linear' model, the F0 of each accent is determined locally with respect to the preceding accent only. The 'global' model predicts the temporal distance between two consecutive peaks to have an effect on the F0 value of the second peak: the greater the distance the lower the second peak. The 'local' model, on the other hand, expects no time-dependent effect on the scaling of peaks but models the downstepping contour by means of a downstepping ratio and a final lowering constant.
    The materials used for testing the effect of the temporal distance on the scaling of the accent peaks consist of a set of tightly controlled utterances each containing four intonational accents. The distance between the accents was varied, comprising either one or three unaccented syllables. The materials were recorded by 5 native speakers of Estonian.
    The results seem to support the 'local' model, because the step between the accented syllables in statements of different length is the same and not larger when the time distance between accented syllables is greater.