Recent human-monkey comparative data based on new state-of-the-art test procedures have provided overwhelming evidence that the monkey is typically NOT a good model of the adult human listener in speech perception. The present long-term objective is to determine if the monkey can model the preverbal human """""""" infant before it tunes into a native language. The assumption is made that both the monkey AND the newborn human infant perceive speech using similar general primate psychoacoustic mechanisms.
The specific aims are to use a """"""""monkey-model"""""""" to investigate two different theories describing how the human infant tunes into native speech sounds, and tunes out non-native sounds. First, Universal/Loss Theory holds that the infant is born with a rich universal ability to perceive all speech sounds, but perceptual loss occurs for non-native sounds. Second, Perceptual Learning Theory holds that the infant is born with little or no ability to perceive speech sounds, and must actively learn to perceive native sounds. The proposed experiments will use a symmetrical """"""""go-left/go-right"""""""" behavioral test procedure and a combined comparative/cross-language design to precisely compare human adults (both native and non-native listeners) and monkeys in categorizing certain phonetic contrasts that have traditionally posed problems for non-natives, e.g. the English liquid contrast or the Hindi dental-retroflex contrast. In cases where the monkey better matches the native over the non-native listener, a Universal/Loss model would be supported in the case of the non-native listener. In contrast, in cases where the monkey better matches the non-native over the native listener, a Perceptual Learning model would be supported in the case of the native listener. As far as health-relatedness of the overall project, """"""""monkey-model"""""""" data may contribute to the verification of currently proposed models of human infant speech perception development.
Showing the most recent 10 out of 14 publications
