Object Bias Disrupts Rule-Based Generalization in Adults Across Domains

AbstractHumans are remarkably adept at abstract rule learning, but little is known about when learners apply this knowledge. We investigated a fundamental constraint in rule generalization: attention to featural similarity (object bias). Across two experiments in different domains, we asked whether adults’ abstract rule generalization is constrained by superficial matches to the concrete exemplars present during learning, as is known to be the case for analogical reasoning (Gentner & Toupin, 1986). In the present studies, participants were exposed to a series of sequences following a simple rule and were asked to generalize to novel instances of either the same rule or a new rule. In one condition, an individual element present during initial learning was inserted into the new, unfamiliar pattern. Results showed that adults often chose this object match over the rule match, suggesting that abstract rule generalization, like analogical reasoning, is impacted by concrete features of the input.

Return to previous page