NETtalk is a connectionist system able to negotiate a certain problem domain. It may even be said, somewhat loosely, to have knowledge of that domain Likewise, various lower animals know their way around certain problem domains. But, in a sense to be explained, such systems lack structured knowledge of the domain. They lack explicit abstractions which represent constraints on properties common to a whole class of inputs all of which they can nonetheless deal with appropriately. NETtalk's knowledge of the entity `vowel' is a good example. The paper draws together and criticises a number of perspectives on this basic idea. It investigates a developmental model which identifies the higher level abstractions with classical representations (Karmiloff-Smith(1987) and a philosophical speculation (Cussins (forthcoming)) which identifies them with evolved connectionist representations. It is argued that any such choice (between connectionism and classicism) is currently premature, though some suggestive reasons exist for favouring the classical option. Nonetheless there is an underlying philosophical insight which unites all such approaches and which concerns some necessary conditions for genuine conceptual thought. The paper aims to display this insight while distancing itself from any premature computational speculation.
This paper is not available online