Semantic category learning through explicit and implicit knowledge of complex sequences

Abstract

Objects with similar functions are judged to be similar "kinds of things," even when they have few directly-observable properties in common. How do people learn such "functional" similarity? One possibility is that objects with similar functions occur in similar contexts: there is overlap in the sequence of events that precede and follow their use. To test this hypothesis, participants learnt a novel sequence with multiple interchangeable objects at each position and subsequently were taught to name the objects. As expected, participants learnt category structures consistent with the positions in the sequence. However, contrary to hypothesis, this learning did not transfer to naming and participants did not learn names consistent with the category structure more quickly than random names. Interestingly, an interaction between naming condition (consistent and random) and learning (participants who learned the sequence and participants who did not) was found also and warrants further investigation.