Speech act recognition in conversation: Experimental evidence

Speech Act Recognition in Conversation: Experimental Evidence Rosa S. Gisladottir (rosa.gisladottir@mpi.nl) Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands International Max Planck Research School for Language Sciences, Nijmegen, The Netherlands Dorothee J. Chwilla (d.chwilla@donders.ru.nl) Radboud University, Donders Institute for Brain, Cognition and Behaviour, Nijmegen, The Netherlands Herbert Schriefers (h.schriefers@donders.ru.nl) Radboud University, Donders Institute for Brain, Cognition and Behaviour, Nijmegen, The Netherlands Stephen C. Levinson (stephen.levinson@mpi.nl) Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands Abstract Recognizing the speech acts in our interlocutors’ utterances is a crucial prerequisite for conversation. However, it is not a trivial task given that the form and content of utterances is frequently underspecified for this level of meaning. In the present study we investigate participants’ competence in categorizing speech acts in such action-underspecific sentences and explore the time-course of speech act inferencing using a self-paced reading paradigm. The results demonstrate that participants are able to categorize the speech acts with very high accuracy, based on limited context and without any prosodic information. Furthermore, the results show that the exact same sentence is processed differently depending on the speech act it performs, with reading times starting to differ already at the first word. These results indicate that participants are very good at “getting” the speech acts, opening up a new arena for experimental research on action recognition in conversation. Keywords: Action; Speech acts; Implicature; Pre-Offers; Conversation Analysis; Self-paced reading. Introduction Knowing a language doesn’t just require syntax or semantics, but the ability to extract speech acts from our interlocutors’ utterances. This is crucial in conversation, since all actions – be they non-verbal or verbal – have implications for how we should respond (Schegloff, 2007): A greeting calls for another greeting, an offer is followed by an acceptance or declination. Scholars in Conversation Analysis were the first to reveal the systematicity of courses of action in turn-taking (e.g. Sacks, Schegloff, & Jefferson, 1974; Schegloff, 2007), moving away from the single act as the fundamental unit of analysis – the perspective of Speech Act theory (Austin, 1976; Searle, 1969) – to the role of sequential context. One of the main observations from this literature is that turns tend to come in pairs of actions and in such adjacency pairs the first (pair) part sets up powerful constraints on what type of action can follow (Schegloff, 2007). Moreover, not all actions have equal status. Reflecting an orientation towards preference structure, dispreferred actions, such as rejections to invitations, tend to be delivered with inter-turn gaps, turn-initial delay, hedges or other discourse markers (uhm, well) (Schegloff, 2007), while preferred actions (e.g. acceptances) do not. This suggests that conversationalists not only monitor speech for actions but also orient to sequential constraints. Given that action is the sine qua non of conversation, how do we map speech acts onto our interlocutors‘ utterances, bridging the gap between sentence meaning and action? In some cases this is a simple matter. In the utterance Please close the door, for instance, the imperative mood and the adverb please function as “special markers” or “illocutionary force indicating devices” (Levinson, 1983; Clark, 1979; Schegloff, 2007) that clearly indicate this is a request. In most utterances, however, the absence of such dedicated vocabulary leaves the propositional content underspecified for the speech act level of meaning. As an example, an assertion like I have a credit card can deliver different speech acts, depending on context. When responding to a question from our interlocutor (e.g. How are you going to pay for the ticket?), I have a credit card functions as an information-giving answer. If it follows an offer of payment (e.g. I can lend you money for the ticket), it is used to indirectly decline it. In this case it could be characterized as an indirect speech act, in which “one illocutionary act is performed indirectly by way of performing another” (Searle, 1975). In yet another interchange, if our interlocutor has expressed the need or desire for some means of payment (e.g. I don’t have any money to pay for the ticket), the same statement of ownership can function as a prelude to an offer, called a Pre-Offer in Conversation Analysis (Schegloff, 1988; Schegloff, 2007). In all three cases, the form and semantic content is underspecified for the action such that the full import of the utterance I have a credit card can only be ascertained relative to the context, in this case the prior speech act in the conversation. There is some psychological evidence that people do extract speech act information online. Using a recognition probe task and lexical decision task, Holtgraves (2008a) addressed whether the comprehension of a sentence like