Nov 27, 2014

2014/11/27

Pluymaekers, M., Ernestus, M., & Baayen, R. H. (2005). Lexical frequency and acoustic reduction in spoken Dutch. The Journal of the Acoustical Society of America, 118(4), 2561–2569.

Presentation: Hsiang-Yu
Summary: Sheng-Fu

In everyday speech, words are often pronounced with a shorter duration than they are in citation forms. Such reductions have been linked to word frequency. Since words with different frequencies often also differ in length, some studies have focused on the effect of word frequency on phonemes or words matched on length in letters. However, for these studies, factors such as the morphological complexity of the words in interest are still not controlled. The present study examines the effects of word frequency on Dutch affixes. Materials were words containing affixes ge-, ver-, ont-, and -ljkl, taken from the Corpus of Spoken Dutch. Two probabilistic factors were included, namely word frequency and morphological complexity (word-stem ratio). Non-probabilistic factors included speech rate (syllable per second), sex, age, and word position (in the carrier utterance). Results of a series of least squares regression provide strong evidence for the relationship between acoustic reduction and lexical frequency: higher word frequency led to shorter duration of not only the affixes as whole, but also the individual segments. Other significant predictors for acoustic reduction include the presence of disfluency following the target word and a higher speech rate. On the other hand, lower morphological complexity (i.e., a higher word-stem ratio) caused a longer segmental duration because the affixes of morphologically less complex words are more likely to receive stress. Overall, the results showed that models of speech production should incorporate word frequency as a factor affecting the acoustic realization of lower-level units such as individual speech segments in affixes.