Recurrent Neural Networks in Linguistic Theory: Revisiting Pinker and Prince (1988) and the Past Tense Debate
This work is incremental, as it applies existing NLP methods to a classic linguistic problem to reassess their utility in cognitive science.
The paper tackles the problem of whether modern NLP advances can improve cognitive modeling by revisiting the past tense debate, showing that encoder-decoder networks address criticisms of earlier neural models without simplifying the mapping task.
Can advances in NLP help advance cognitive modeling? We examine the role of artificial neural networks, the current state of the art in many common NLP tasks, by returning to a classic case study. In 1986, Rumelhart and McClelland famously introduced a neural architecture that learned to transduce English verb stems to their past tense forms. Shortly thereafter, Pinker & Prince (1988) presented a comprehensive rebuttal of many of Rumelhart and McClelland's claims. Much of the force of their attack centered on the empirical inadequacy of the Rumelhart and McClelland (1986) model. Today, however, that model is severely outmoded. We show that the Encoder-Decoder network architectures used in modern NLP systems obviate most of Pinker and Prince's criticisms without requiring any simplication of the past tense mapping problem. We suggest that the empirical performance of modern networks warrants a re-examination of their utility in linguistic and cognitive modeling.