Abstract:
It has been shown that speakers adjust their articulation according to the amount of semantically biasing information in the whole utterance. Furthermore, subjects modify their speech production for the benefit of a listener. In the present investigation, speakers' production of high- and low-frequency words was examined when produced in two semantic contexts (semantically biasing and semantically neutral) and in two pragmatic contexts (presence or absence of an overt listener). In addition, speakers were instructed to produce the target words in a production mode versus a perception mode of experience. For example, speakers were instructed to interpret the words according to how frequently they ``say'' the word or how frequently they ``hear'' the word. The results of acoustic measurements showed that subjects' production of high- versus low-frequency words was affected by the semantic and pragmatic contexts. However, subjects modified their articulation differently depending on the mode of experience in which they interpreted the words' frequency. These results will be discussed according to pragmatic compensation and semantic activation in speech production and, in particular, how speech production may be affected by the speakers' representation of lexical items according to their production versus perception mode of experience. [Work supported by NIDCD.]