Presentation: Thomas
Summary: Sally
The aim of this study is to
examine whether intonation is “post-lexical,” as proposed by the autosegmental-metric
theory and generally assumed, or there is lexical storage of intonation, as
proposed by the exemplar model. The Parametrisation of pitch accent shape (the
PaIntE model) was adopted to analyze the data from a speech corpus, and three
experiments were then designed. The first one investigated the absolute frequency
of occurrence of a given word with a certain type of pitch accent and its
effect on the shape of accents on that word; the second looked at the effect of
both the absolute and the relative frequencies of a target word with a
particular accent type and their effect on the variability of the shape of these
accents; the third examined the effect of the relative frequency of a word in
its lexical context on the variability in its prosodic context as well as on accent
shape variability. Result of the first experiment showed that the higher
frequency of the appearance of a word with a certain accent type, the greater
the amplitude of that accent. Moreover, as shown in the second experiment, when
both absolute and relative frequencies were considered, the former increased
the variability in pitch accent shape, whereas the latter reduced this
variability. In the last experiment, both the variability in the word’s
prosodic context and the accent shape variability decreased as the relative
frequency of the word in its lexical context increased. The authors concluded
that all the aforementioned findings support the exemplar model, which proposes
storage of tonal information with individual lexical items. And following this
train of thought, language units are believed to be concrete and highly
specified instances and simultaneously contain both the tonal and the lexical
properties.
