Parallel Distributed Processing Challenges the StrongModularity Hypothesis, Not the Local- ity Assumption [Commentary on M. J. Farah, Neuropsychological inference with an interactive brain: A critique of the “locality” assumption]. Behavioral and Brain Sciences, 17, 77–78