Abstract
The language of thought hypothesis is one of Fodor’s seminal contributions to cognitive science. Prominent among the objections to it has been the argument that there is no neurobiological evidence for materially realized symbols in the brain. If memory is materially realized by enduring alterations in synaptic conductances, then this is true, because the synaptic-conductance hypothesis is simply the ancient associative learning hypothesis couched in neurobiological language. Associations are not symbols and cannot readily be made to function as such, thus neurobiologists are unable to say how simple information—for example, the durations of intervals in simple Pavlovian conditioning paradigms—are stored in altered synaptic conductances. Recent results from several laboratories converge, strongly suggesting that memories do not reside in altered synaptic conductances but rather at the molecular level inside neurons. The chapter reviews the experimental evidence for this revolutionary conclusion, as well as the plausibility arguments for it.