![]() ![]() The key idea is that, if the timing of syllables follows any kind of pattern, this temporal pattern might be helpful for infants acquiring speech (Bialek et al., 2001 Nazzi and Ramus, 2003 Saffran et al., 2006) by providing infants with clues to predict where units begin and end (Trehub and Thorpe, 1989 Trainor and Adams, 2000). In this paper, we test whether this problem might be made more tractable by predictability in the temporal structure of speech. However, it is at present unclear-both theoretically and in terms of building speech technologies-which properties of speech allow this highly underconstrained inductive problem to be solved. Segmentation problems recur at multiple levels of linguistic structure, and must be solved either before or in tandem with higher-level inferences or generalizations that are defined over these units (e.g., syntactic, morphosyntactic, and phonotactic rules). One of the earliest and most basic of these component problems is to segment continuous speech into distinct units, such as words, syllables or phonemes. To acquire a language, human infants must solve a range of intertwined inductive problems which, taken together, represent one of the most demanding computational challenges a child will ever face. We conclude that temporal predictability in speech may well arise from a combination of individually weak perceptual cues at multiple structural levels, but is challenging to pinpoint. Together, our analyses provide limited evidence for predictability at different time scales, though higher-order predictability is difficult to reliably infer. Second, we model higher-order temporal structure-regularities arising in an ordered series of syllable timings-testing the hypothesis that non-adjacent temporal structures may explain the gap between subjectively-perceived temporal regularities, and the absence of universally-accepted lower-order objective measures. First, we analyse distributional regularities using two novel techniques: a Bayesian ideal learner analysis, and a simple distributional measure. Rather than looking for differences between languages, we aim to find across languages (using clearly defined acoustic, rather than orthographic, measures), temporal predictability in the speech signal which could be exploited by a language learner. Here, we compare several statistical methods on a sample of 18 languages, testing whether syllable occurrence is predictable over time. Existing measures of speech timing tend to focus on first-order regularities among adjacent units, and are overly sensitive to idiosyncrasies in the data they describe. This hypothesis tacitly assumes that learners exploit predictability in the temporal structure of speech. ![]() By providing on-line clues to the location and duration of upcoming syllables, temporal structure may aid segmentation and clustering of continuous speech into separable units. Temporal regularities in speech, such as interdependencies in the timing of speech events, are thought to scaffold early acquisition of the building blocks in speech. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |