Unintentional and efficient relational priming.

If relational priming is responsible for unintentional analogical reasoning, as has been suggested, it too should occur unintentionally. However, results from previous studies are inconclusive - studies that use the sensicality task usually demonstrate unintentional priming, while lexical decision tasks have failed to capture the effect without explicitly instructing participants to note and use the relational similarity in the stimuli. We discuss possible reasons for these contradictory results. Based on this discussion, we aimed to maximize our chances to find an effect by ensuring that the primes and targets elicit the same processing, by using a longer SOA, suitable for the more complex nature of the task, and by ensuring that the stimuli are good exemplars of the relations that hold between them. We present two experiments that obtained unintentional and efficient priming of relations with a lexical decision task. Participants made a lexical decision for a target pair of words more quickly when it was preceded by a similarly related pair of words, compared to an unrelated pair. Participants were not instructed to note and use the relations. Experiment 2 extended those results by showing that the effect is present even when executive working memory resources are occupied by a secondary task. Even though it turned out that the base pairs differed on semantic similarity, co-occurrence and imageability between the two conditions, these differences were not responsible for the effect. Thus, relations can be primed unintentionally and efficiently, even when relational integration is not necessitated by task demands.