Speech-intentions and self-monitoring - Manipulating verbal feedback in a single-word production task

Abstract: It is commonly assumed that the speech production process is started and guided by a clear conception of what to say, i.e. an intention or a pre-linguistic message. This intention can also function as a standard of accuracy against which actual performance can be measured. But critique against the idea of such a centrally governed process has been offered, and propositions for how a distributed model can account for the assignment of content to speech acts have been given (Dennett, 1991). In this thesis, the role of the auditory feedback of one's own voice in the understanding of the meaning of self-produced speech is investigated. As participants performed a computerized Stroop test while hearing their own voice exclusively through earphones, certain words were covertly recorded. While the feedback of participants? own voices was blocked out, these words were played back later in the test concurrently with participants uttering another, phonetically similar but semantically dissimilar, word. Results show that, while such manipulations were almost always retrospectively detected, a majority of participants reported in post-test interviews that they had experienced confusion as to the actual source of the manipulated feedback, not being certain if it was self- or other-produced. On a minority of manipulated trials, participants acted towards the manipulated feedback as if it was self-produced.