You are viewing the SNL 2017 Archive Website. For the latest information, see the Current Website.

Poster D14, Thursday, November 9, 6:15 – 7:30 pm, Harborview and Loch Raven Ballrooms

Aligning sentence structures in a language game: evidence from healthy aging and aphasia

Jiyeon Lee1, Grace Man1, Victor Ferreira2, Nick Gruberg2;1Purdue University, 2University of California San Diego

Speakers align syntactic structures with their conversational partners. Gruberg et al. (in prep-a) discovered that this syntactic alignment occurs for associations between event content and sentence structures, also known as syntactic entrainment, beyond the level of sentence constituent orders. These effects are shown in both young adults and children, and are viewed as reflecting ongoing prediction error-based ‘tuning’ or language learning throughout the lifespan. Crucially, the prediction errors that cause syntactic alignment are experienced during comprehension, rather than production of sentences (Chang et al., 2006; Jaeger & Snider, 2013), predicting that listening to their interlocutor’s utterances would suffice for speakers to adapt their production preferences. We test this hypothesis in older and aphasic speakers to better understand the mechanisms of syntactic learning. Experiment 1 examined syntactic entrainment in a comprehension-based picture matching game. In Experiment 2, the game was modified so that the participant repeats their partner’s utterances during card matching, obligating prior production of the target structures. For Experiment 1, 20 young, 20 older adults, and 13 adults with aphasia participated in a collaborative language (picture-matching) game. Participants played the ‘matcher’ and subsequently ‘director’ roles with the experimenter, who described pictures using either preferred (active, prepositional dative, and on-variant locatives) or non-preferred structures (passive, double-objective dative, and with-variant locative). When playing the matcher role, the participant placed their cards in the correct order after listening to the experimenter’s sentences. Then the participant, playing the director role, described pictures for the experimenter to match. We measured whether the participant produced the same structures to refer to specific events in the pictures as the experimenter. Results revealed that young adults were more likely to produce preferred structures upon hearing experimenter’s use of preferred vs. non-preferred structures (p < .001) – that is, syntactic entrainment. However, no syntactic entrainment effects were shown in older and aphasic participants (Figure 1a). For Experiment 2, we tested 12 older adults and 8 adults with aphasia. Participants were instructed to verbally repeat the experimenter’s sentences before they select the matching picture card, thus obligating prior production of target sentences. Remaining procedures were the same as Experiment 1. In contrast to the results of Experiment 1, both older (p < .001) and aphasic speakers (p = .01) used preferred structures more frequently following experimenter’s use of preferred vs. non-preferred structures (Figure 1b). Together, our findings show that comprehension-induced prediction error is not sufficient for successful syntactic entrainment effects in older and aphasic speakers, different from what has been shown in young adults and children (Gruberg et al., in prep a, b). This further suggests that as an effect of aging and aphasia, content-structure mapping becomes stabilized so that active production of target content-structure associations has most predictive effects in syntactic learning. In sum, the current study illustrates that the mechanisms of syntactic learning change as a function of age and modality, which may need to be considered in the existing models of error-based language learning (Chang et al., 2006; Jaeger & Snider, 2013).

Topic Area: Language Therapy

Back to Poster Schedule