This project studies the production and perception of phonological sequences. Our guiding assumption is that the phonological processing system is constantly changing. It adapts to recent experience, while continuing to reflect the accumulated experience of a lifetime of speaking and listening. The proposed research seeks to understand the adaptability of the phonological processing system, with an emphasis on the malleability of the system's use of general phonological patterns, such as phonotactic constraints. It is well known that speech errors or """"""""slips of the tongue"""""""" exhibit the phonotactic regularity effect: Errors rarely create combinations of sounds that are illegal in the language being spoken. For example, an English speaker might mispronounce """"""""nun"""""""" as """"""""nung"""""""", but not """"""""ngung"""""""" because that violates the constraint of English that """"""""ng"""""""" can never begin a syllable. We created a laboratory analog to this effect and used it to demonstrate learning. Experimental subjects recited strings of syllables in which consonants were artificially restricted to particular positions in syllables; for example, /f/ always occurred as a syllable onset. As subjects experienced these syllables, their slips became strongly sensitive to the distributions of the consonants. For example, nearly every misplacement of /f/ obeyed the rule that /f/ is an onset for this experiment. These data suggest that subjects implicitly learned the sound distributions, and this learning affected their productions. We demonstrated the same effect in perception. Subjects listened to syllables with consonant-position restrictions; in a later speeded repetition task, new syllables that were consistent with the phonotactics of the experiment were repeated more rapidly than syllables that were inconsistent. The proposed research uses these techniques to investigate learning mechanisms in the phonological processing system. The studies manipulate the phonological patterns to be acquired and make direct comparisons between production and perception and between learning by adults and infants. These experiments are informed by psychological theories of learning and linguistic theories of representation, and the obtained data will be used to constrain development of computational models of implicit learning in the phonological processing system. The research will yield a new understanding of how the processing system responds to experience, which ultimately will contribute to the treatment of language disorders and to language pedagogy.
Showing the most recent 10 out of 25 publications