Learning complex inflectional paradigms through blended gradient inputs
- Eric R Rosen (Johns Hopkins University)
Abstract
Through Gradient Symbolic Computation (Smolensky and Goldrick, 2016), in which input forms can consist of gradient blends of more than one phonological realization, we propose a way of deriving surface forms in complex inflectional paradigms that dispenses with direct references to inflectional classes and relies solely on relatively simple blends of input expressions.
Keywords: inflectional morphology, learnability, complexity, Gradient Symbolic computation, paradigms
How to Cite:
Rosen, E. R., (2019) “Learning complex inflectional paradigms through blended gradient inputs”, Society for Computation in Linguistics 2(1), 102-112. doi: https://doi.org/10.7275/ccwf-j606
Downloads:
Download PDF