In case, there’s interest, have a look at the papers to be presented at this year’s Cognitive Science meeting in Boston (July, 20th-23rd). HLP lab will be represented by two talks and four posters. The two talks will presenting work employing artificial language learning to address questions about typological generalizations:
- Masha Fedzechkina(BCS, University of Rochester) will present evidence that language learners are biased to reduced the uncertainty in the mapping from form to meaning. Her work is comparing the acquisition of miniature languages with and without case-marking in terms of to what extent learners tend to regularize or even fix variable word orders for these two types of languages (Fedzechkina, Jaeger, & Newport, 2011). Together with other recent work (e.g. by Newport, by Culbertson), this work provides evidence that language learners deviate from the input provided to them in a predictable manner. In this case, we designed the experiment to directly test the functionalist claim that language learners are biases towards acquiring languages that support communication (cf. Bates and MacWhinney’s early work).
- Hal Tily (BCS, MIT) will present work employing a novel web-based artificial language learning paradigm, in which hundreds of participants can be run within a matter of a few days. Using this paradigm, we first replicated and extended a well-known study on determiner learning (Hudson Kam and Newport, 2004) and then investigate to what extent cross-linguistically observed quantitative patterns in argument and determiner order are replicated by language learners. We discuss how this paradigm will facilitate further tests of typological generalizations (Tily, Frank, & Jaeger, 2011).
- Alex Fine(BCS, University of Rochester) will present evidence that comprehenders are capable to adapt to prolonged exposure to shifted distributions over syntactic structures. Employing a similar paradigm as in Wells et al (2009) seminal study, Alex investigates how implicit changes in the reliability of a cue over several days affect sentence processing (in a garden path environment, cf. Garnsey et al., 1997). The results suggest that comprehenders re-weigh cues to reflect increases and decreases in the cues reliability (cf. Bates and MacWhinney, once again …). This ties work on sentence comprehension to work in other cognitive domains (perception, motor control, etc.). Check it out (Fine & Jaeger, 2011).
- Thomas Farmer and Alex Fine (BCS, University of Rochester) follow up on Fine, Qian, Jaeger, and Jacobs (2010-CMCL), providing further evidence for rapid adaptation based on recently experienced syntactic input. We find that reading times in garden path environments are drastically reduced within the course of a single experiment. The evidence in this study is much clearer than that presented in Fine et al (2010), as the the latter contained several confounds that are addressed in the current work (Farmer, Fine, & Jaeger, 2011).
- Caitie Hilliard (soon to be in Psychology, Iowa) and Katrina Furth (BCS, University of Rochester) will be presenting their experiments on how phonological overlap between adjacent words affects sentence production. Unlike most previous work, their experiments try to investigate phonological encoding in the context of sentence production, although this wasn’t without challenges ;). Two experiments investigate onset and rhyme overlap. Identity (pick – pin) is compared to similarity (pick – bin) and phonologically unrelated conditions. We also investigate whether the interaction between phonological overlap and word frequency observed in Marin and O’Seaghdha (2000) replicates in our setting (Hilliard, Furth, & Jaeger, 2011).