Retrieval of bilingual autobiographical memories: Effects of cue language and cue imageability

Abstract

An important issue in theories of bilingual autobiographical memory is whether linguistically encoded memories are represented in language-specific stores or in a common language-independent store. Previous research has found that autobiographical memory retrieval is facilitated when the language of the cue is the same as the language of encoding, consistent with language-specific memory stores. The present study examined whether this language congruency effect is influenced by cue imageability. Danish-English bilinguals retrieved autobiographical memories in response to Danish and English high- or low-imageability cues. Retrieval latencies were shorter to Danish than English cues and shorter to high- than low-imageability cues. Importantly, the cue language effect was stronger for low-than high-imageability cues. To examine the relationship between cue language and the language of internal retrieval, participants identified the language in which the memories were internally retrieved. More memories were retrieved when the cue language was the same as the internal language than when the cue was in the other language, and more memories were identified as being internally retrieved in Danish than English, regardless of the cue language. These results provide further evidence for language congruency effects in bilingual memory and suggest that this effect is influenced by cue imageability.

Keywords:

• Autobiographical memory
• Bilingualism
• Word imageability

This research was supported by the Danish National Research Foundation [grant number DNRF93]. We thank Marie Laursen Canter for help in recording the stimuli, and Niels Peter Nielsen and Stine Breum Ramsgaard for help in conducting the experiment.

This research was supported by the Danish National Research Foundation [grant number DNRF93]. We thank Marie Laursen Canter for help in recording the stimuli, and Niels Peter Nielsen and Stine Breum Ramsgaard for help in conducting the experiment.

Notes

¹ Some of the studies reviewed in this section varied word imageability, whereas others varied word concreteness. These two measures are closely related, and for most words, they are quite similar in that words that refer to easily imageable concepts typically are concrete, whereas words referring to concepts that are hard to image tend to be abstract. In this section, we will use the same label to refer to this measure (concreteness or imageability) as was used in the study that is reviewed. Because the experimental stimuli used in the present study were selected from corpora of word imageability ratings, the present measure will be referred to as the imageability variable.

² We are not aware of any published study reporting imageability ratings for Danish words.

³ We measured button-press latencies, and not speech onset latencies, because people vary in their speech fluency with some initiating speech in a very disfluent and hesitant way and others speaking more fluently. Such variation is expected particularly when speakers use a non-native language. As a consequence, the onset of the memory description is likely to vary depending on whether speakers are fluent or disfluent in their speech. By measuring button-press latencies, we avoided any influence of speech fluency on the response latencies (for discussion see Marian & Neisser, 2000).