At long last, Alex Fine‘s paper on syntactic adaptation expectation is about to appear in PLOS One. You can download the pre-proof from our academia.edu page (the final version will be linked there as soon as it’s available):
- Fine, A. B., Jaeger, T. F., Farmer, T. , and Qian, T. 2013. Rapid expectation adaptation during syntactic comprehension. PLoS One.
The paper presents a novel framework that ties together syntactic comprehension and implicit learning. We tie together work on expectation-based sentence understanding, syntactic priming in comprehension, statistical learning, and speaker-specificity in syntactic comprehension.In two self-paced reading studies, we show that readers rapidly adjust their expectations for specific syntactic structures to converge on the statistics of the current environment. They do so based on both previous experience and recent experience within the experiment. This leads readers to overcome life-long biases in reading within this specific environment (the experiment). Alex has replicated these findings for different syntactic structures and under different conditions (if you’re interested, please see his thesis for now, though a follow-up paper is in the works). I think this work raises a number of interesting questions for future work. We’ve tried to spell out at least the first couple of steps of this program in our discussion.
Here are some slides from a recent talk I gave at the Donders Institute of Cognition (DCC) on these and one one other experiment. Perhaps they are useful for a quick glance of what’s going on:
For Fine et al , we intentionally kept the framework in informal terms. For modeling implementations, please see Fine, Qian, Jaeger and Jacobs, 2010; Kleinschmidt, Fine, and Jaeger, 2012. Also keep your eyes out for some really interesting work by Mark Myslin and Roger Levy that builds on and extend this work and the ideas developed in Jaeger and Snider (2013).
If you’re interested in the broader computational framework that we have in mind for the type of results discussed in Fine et al , you might also enjoy reading Ting Qian’s paper about learning across multiple environments, when the statistics of interest are variable across environments. It’s a perspective piece that also provides an overview of the relevant human and animal learning literature:
- Qian, T., Jaeger, T. F., and Aslin, R. 2012. Learning to Represent a Multi-Context Environment: More than Detecting Changes. Frontiers in Psychology 3, 228.
Some of you have seen Alex, Thomas, or me present this work before. Thank you for all your feedback. We have tried to incorporate as much as possible of it.