Learning exceptionality and variation with lexically scaled MaxEnt
Coral Hughto, Andrew Lamont, Brandon Prickett, and Gaja Jarosz. 2019. Learning exceptionality and variation with lexically scaled MaxEnt. In Proceedings of the Second Annual Meeting of the Society for Computation in Linguistics (SCiL). 91-101.
A growing body of research in phonology addresses the representation and learning of variable processes and exceptional, lexically conditioned processes. Linzen et al. (2013) present a MaxEnt model with additive lexical scales to account for data exhibiting both vari- ation and exceptionality. In this paper, we im- plement a learning model for lexically scaled MaxEnt grammars which we show to be suc- cessful across a range of data containing pat- terns of variation and exceptionality. We also explore how the model’s parameters and the rate of exceptionality in the data influence its performance and predictions for novel forms.