Producing Less Preferred Structures: More Gestures, Less Fluency

Producing Less Preferred Structures: More Gestures, Less Fluency Susan Wagner Cook (susan-cook@uiowa.edu) Department of Psychology, E11 Seashore Hall Iowa City, IA 52242 USA T. Florian Jaeger (fjaeger@bcs.rochester.edu Department of Brain and Cognitive Science, University of Rochester Rochester, NY 14627 USA Michael K. Tanenhaus (mtan@bcs.rochester.edu) Department of Brain and Cognitive Science, University of Rochester Rochester, NY 14627 USA Abstract Preference and Production Speakers often have choices about how to structure their utterances. However, even though multiple alternatives may be acceptable in theory, often one of them will be preferred over the others. The question we explored here was what happens when speakers produce less preferred alternatives. We developed a new experimental paradigm to reliably elicit the propositional or double object dative with varying degrees of preference. We then used this paradigm to investigate how, given properties of the message, an individual speaker's preference for a particular structure affects how that utterance is produced. Speakers gestured more and were more likely to be disfluent when they chose less preferred structures. Thus, having a choice per se does not guarantee more successful production. Instead, production is facilitated when speakers choose more preferred alternatives. Alternative theories of language production appear to make opposing predictions about why speakers produce less preferred structures, leading to different predictions about how these structures should be produced. Availability-based accounts of production (e.g. Bock & Warren, 1985; Ferreira & Dell, 2000) argue that speakers produce less preferred alternatives because of the material that was more available at the moment of choice. Although the material consistent with the preferred choice is generally more likely to be available at the choice point, when it is not, speakers will instead produce the less preferred alternative, allowing them to take advantage of the material that is currently available. On this account, speakers should experience no more or less difficulty when producing less preferred or preferred structures, because these structures are produced in precisely those environments where they are beneficial to the speaker. Competition-based models of production (e.g. Haskell & MacDonald, 2003; Solomon & Pearlmutter, 2004) assume that speakers produce less preferred alternatives because this structure is more active at the moment of choice, given interactions among all operating constraints. However, when the less-preferred structure is chosen, more often than not, the preferred structure will be more active than the less preferred structure will be when the preferred structure is chosen. Accordingly, speakers should be more likely to run into difficulty when producing less-preferred structures, due to greater ongoing competition from partially active preferred structures. Such difficulty may show in a higher rate of disfluencies and gestures, as well as slowed speech. These accounts offer explanations into the mechanism underlying production difficulty during less preferred structures. Computational level accounts, such as accounts of rational production (in the sense of Anderson, 1990), also make predictions about the production of less preferred structures. Since less preferred structures have a lower probability of occurring in the context where they are less preferred, they necessarily encode more information. Accordingly, under the principle of Uniform Information Keywords: speech production; gesture; disfluencies; dative alternation. Introduction Grammatical constraints restrict the structures speakers can use to convey their message – not all structures are grammatically acceptable. But grammar also provides speakers with some flexibility, by offering more than one grammatically acceptable way of encoding a proposition linguistically. Moreover, speakers appear to benefit from flexibility. When speakers can choose how to structure an utterance (e.g. when choosing between a double object, DO, or prepositional dative, PD to communicate a transitive event), they on average produce those utterances faster and with fewer errors in comparison with when they can not choose (Ferreira, 1996). However, even though multiple alternatives may be acceptable in theory, often one of them will be preferred over the others. For example, in Ferreira’s (1996) study, in the conditions where choice was available, speakers preferred (65-77%) one of the possible alternatives. The question we explored here was what happens when speakers choose to produce the less preferred alternative.