| Researchers investigating speech acts (such as apologies or requests) often elicit data with questionnaires so that they can control the experimental variables. However, questionnaire data may not reflect how people actually speak, the most serious problem being that interaction is absent. In addition, we do not know enough about the possible task effects of directing respondents to produce the speech act being investigated on a questionnaire. These are the problems addressed in this study.;Two tasks were used to elicit data on complaints. One was a traditional Discourse Completion Task (DCT), in which respondents were asked to imagine themselves in a situation and to write down what they would say. The other task was an innovative, interactive version (IDCT), which allowed respondents to take three turns. For both the DCT and IDCT, about half of the respondents were directed to respond to each situation by complaining, while the others were prompted with an open-ended question about what they would say.;The findings suggest that the IDCT elicits a form of data intermediate between questionnaire data and role-play data. Compared to the DCT responses, IDCT responses were significantly longer and showed a greater variation of outcomes by the end of the exchange. Directing respondents to complain did not result in any significant differences in response length or tendency to state the problem. An unexpected finding was that, in two out of ten complaint situations, the respondents directed to complain were significantly more likely not to enter into an exchange at all.;The IDCT combines experimental controls with an interactive context, allowing researchers to elicit quantitative data to test hypotheses having to do with how speech acts are used in conversation. This new method can play a role in helping move speech act theory beyond the single utterance. |
