An associative account of inferences: The development towards the prototype

According to a traditional view, inferences are personal-level entities pertaining to the domain of reasons, and therefore they cannot be accounted for in causal terms – specifically, as mere associations. I intend to argue that this is at the very least a drastic simplification, for two reasons. First, the word “asso-ciation” is polysemous, so we should specify in which of its possible senses an inference is not a mere asso-ciation. Second, personal-level inferences based on formal rules are only the extreme end of a complex de-velopmental trajectory. As the last decades of research in the field have shown, we should refrain from identifying the entire domain of reasoning with that final stage, which is in fact mostly contingent upon extensive logical training. In this paper, I try to disentangle some major stages in the development of full-fledged (prototypical) inferences, and then to show that all of them – till the final one – can be considered associative in appropriate senses of the word.