Bilingual episodic memory: Acquisition and forgetting

Abstract

Two experiments investigated the representation of meaning and input language in bilingual memory. Experiment I used the savings method to investigate the kind of information remaining in a nonrecallable memory trace 5 weeks after original learning. The results showed significant savings for same versus changed meaning (across translation equivalents such as HORSE and CHEVAL) but not for same versus changed language. This result, in conjunction with previous research on savings, suggests that translation equivalents do not function as synonyms but, instead, are mediated by an underlying supralinguistic concept. Experiment II used the depth-of-processing incidental-learning paradigm to investigate the kind of information acquired when words are classified in terms of either meaning (LIVING vs. NONLIVING) or language (FRENCH vs. ENGLISH). The results showed that meaning classifications produce better memory than linguistic classifications, both in terms of recall for meaning and in terms of recognition for the language of presentation; this suggests that memory processing is not directed to a single level but, instead, passes through the linguistic level on the way to the deeper semantic level.