Nous accueillons à partir du 20 mai 2019 le professeur Brian Dillon (University of Massachusetts – Amherst (USA)) pour une série de quatre séminaires sur le thème « Acceptability, Decision-making and Parsing ».
Ces séminaires se dérouleront les lundis 20, 27 mai et 03 juin et le mardi 11 juin de 16h00 à 18h00 à l’université Paris Diderot – bâtiment Olympe de Gouges – 8, rue Albert Einstein – 75013 Paris – Salle Laplanche, 576F (5e étage : Demander un badge à l’accueil pour utiliser l’ascenseur).
Abstract :
In this seminar we will ask: where does an acceptability judgment come from? I will introduce two different perspectives on this question, borrowing frameworks for modeling decisionmaking under uncertainty from cognitive psychology. These models realize formal linking hypotheses between linguistic processing and acceptability judgments, and in the context of these linking hypotheses, offer unique insight into the underlying linguistic processing that acceptability judgments reflect. One goal of this seminar will be to give EFL students a chance to apply these models in the context of their own research. Outside of the seminar, I will spend some time with students to help them to find an insightful way to apply these techniques to their own research.
Seminar one:
We will explore why, and how, to use Signal Detection Theory (SDT) to model acceptability judgment data. I will specifically discuss i) how Receiver Operating Characteristic (ROC) modeling can be applied to rating data, ii) how SDT can be used to distinguish bimodal from unimodal response patterns, and iii) the potential pitfalls of analyzing rating data using standard statistical techniques.
Readings: – Bader, M., & Häussler, J. (2010). Toward a model of grammaticality judgments. Journal of Linguistics, 46(2), 273-330. – Dillon, B. & Wagers, M. (in press). Approaching gradience in acceptability with the tools of signal detection theory. Also recommended: – Langsford, S., Perfors, A., Hendrickson, A. T., Kennedy, L. A., & Navarro, D. J. (2018). Quantifying sentence acceptability measures: Reliability, bias, and variability. Glossa: a journal of general linguistics, 3(1).
Seminar two: We will continue our discussion of the use of SDT for acceptability judgment data. We will focus on the design of acceptability judgment experiments for SDT analysis, focusing on practical issues of experimental design. Students should come prepared to do hands-on analysis of data using the pROC package in R.
Readings: – Dillon, Andrews, Rotello & Wagers (2018). A new argument for co-active parses during language comprehension. Journal of Experimental Psychology: Learning, Memory and Cognition.
Seminar three: In seminar three we will introduce the Drift Diffusion Model (DDM), which models both the reaction time and accuracy of rapid two-alternative forced choice decisions. The DDM offers a dynamic perspective on the decision processes that underly acceptability judgments by decomposing the decision process into distinct parameters, including how clearly a string is perceived as grammatical and how long it takes a perceiver to compute that analysis. We will see examples of how this aspect of the DDM has been used to investigate the processing of subject-verb agreement, and the nature of the memory mechanisms that support syntactic processing.
Readings: – Ratcliff, R., Smith, P. L., Brown, S. D., & McKoon, G. (2016). Diffusion decision model: Current issues and history. Trends in cognitive sciences, 20(4), 260-281. – Hammerly,C., Dillon, B., & Staub, A. (in press, Cognitive Psychology). The grammaticality asymmetry in agreement attraction reflects response bias: Experimental and modeling evidence.
Seminar four: We will continue our discussion of the DDM, focusing on practical aspects of fitting the diffusion model. Students should come prepared to do hands-on analysis of data using the fast-DM software.
Readings: – Chen, S. & Husband, M. (2018). Comprehending anaphoric presuppositions involves memory retrieval too. Proceedings of the Linguistic Society of America 3(44), 1-11.
Comments