Mental representations

CoNGA's third meeting will revolve around mental representations, one of the core concepts of modern Cognitive Science. Our two talks, from a philosopher and a computational psycholinguist, will challenge the traditional view on mental and linguistic representations, pinpointing to pitfalls and shortcomings of theories that assume such constructs as building blocks of cognition.

The meeting will take place in the Annexe to the R building, in the City Campus of the University of Antwerp, from 6 to 8 pm.

The abstracts of the talks follow:

Farid Zahnoun (Centre for Philosophical Psychology): Dereifying Representation

Abstract: The notion of internal representation is without doubt one of the most central explanatory posits within mainstream cognitive science. At the same time, however, there is much controversy about how, exactly, we should conceive of internal representations. Indeed, there seem to be multiple notions of representation at work within cognitive science. This talk focuses on the classic, and still prominent notion of representation, where representations are understood as internal content-carrying items which can in principle be individuated and identified. I refer to this conception as the reified notion of representation. In the first part, I want to lay bare and investigate three possible explanations of what might motivate cognitive scientists to use and reuse this notion. Crucially, none of these motivations derives from scientific practice itself, nor warrants a use within scientific explanations of cognition. The methodological contribution of this first part can be put as a challenge for theorists to show how, first, their conception of the reified notion of representation isn’t derived from the three non-scientific motivations sketched below, and second, to show how, if at all, invoking these representations can be properly motivated.  In the second part of the talk, I will argue that any such motivation is bound to fall short, because the reified notion is ultimately incoherent, and therefore untenable.


Stéphan Tulkens (CLiPS): Lexical representations in theories of word reading

Abstract: Most theories of word reading incorporate some form of mental lexicon, which is a word store that contains representations of all the words a person knows. These theories of word reading embody what I call a dictionary metaphor, in which the orthographic form of a word is the key to an entry containing the semantics of that word.
In this presentation, I will argue that this view is incorrect, and in need of revision. I will use evidence from studies of bilingual processing to show that phonology and semantics greatly influence word reading times, and hence play a role in lexical access, even across languages. The fact that both phonology and semantics play a role in the retrieval of words implies that all information in our mental lexicon is contained in the key of the dictionary. If this is the case, then what is left in the actual entry?
I will propose that we abandon talking about representations of words as being static symbols that receive activation from visual input. Instead, I propose an alternative in which ambiguity, or lack thereof, is the main source of facilitation or inhibition in word reading.