You are in: Home » Research at the Unit » Speech and Language
Speech input and phonological processing
Perceptual learning of speech
Although the adult perceptual system is much less adaptable than that of infants, adults retain considerable ability to learn new perceptual skills. For example, although adults who learn a foreign language later in life never reach the same levels of proficiency as native speakers, they can still become highly competent in a second language. With experience we can all improve our ability to understand speaker with strong accents. With practice, deaf individuals can eventually learn to recognise speech given the impoverished signal provided by a cochlear implant. This project investigates this process of perceptual learning.
The starting point for this work is ecent studies of perceptual learning (Johnsrude, Davis, Hervais-Adelman & Brent, 2002; McQueen, Norris & Cutler, 2006; Norris, McQueen & Cutler, 2003) carried out by the Speech and Language Group. The common factor in these studies is that both showed that perceptual learning is strongly influenced by lexical knowledge, and can occur very quickly. Although we have argued elsewhere that lexical feedback is of no value in on-line speech recognition (Norris, McQueen and Cutler, 2000), lexical feedback in learning can be very useful. If listeners can identify a word spoken by someone with an unfamiliar accent, then lexical information can tell them how they should have categorised the speech sounds in that word. The effects of lexical knowledge in our studies is very powerful.
In more recent work we have shown that listeners can also use phonotactic constraints to retune their perception (Cutler, McQueen, Butterfiled & Norris, 2008).