Abstract
This study investigated how linguistic predictors such as word frequencies, the difficulty and creativity of problems, and the category of problems contribute to analogical reasoning in L1 and L2. This study also investigated how different types of similarities (i.e., perceptual and relational similarities) are processed in analogical reasoning. In Experiment 1, Japanese participants were asked to solve 100 multiple-choice A:B::C:D analogy problems (e.g., skeleton: bone:: tornado: wind) in their first language, Japanese (L1). In this experiment, participants also rated the difficulty and creativity of problems. In Experiment 2, Japanese participants completed the same tasks, but the problems were shown in their second language, English (L2). The results showed that problems presented in L1 elicited higher accuracies and faster response times than in L2. A significant interaction was found between languages (L1/L2) and the category of problems which indicates that finding a perceptual similarity (e.g., the shape image of word concepts) with verbal stimuli in L2 is more challenging than in L1. Moreover, our results on response times indicated that processing relations between words would be carried out in L1 without any specific instruction while it would not be completed in L2 possibly due to the cognitive demand related to lexical processing. Considering these results, it is advisable in an educational setting to provide L2 learners with enough time and explicit instruction on understanding word relationships when forming analogies.