This short reply to O'Reilly and Farah (1999) highlights three issues that continue to divide our approaches. We disagree about what needs to be modelled, the theoretical importance of distributed representations, and how to account for provoked overt recognition. However, the exchange has helped to show why different groups favour different approaches, and has highlighted some testable empirical predictions.
|Number of pages||7|
|Publication status||Published - 1999|