An investigation of false memory in Turkish-English bilinguals-

False recognition across languages is investigated using the Deese-Roediger-McDermott (DRM) paradigm. English-Turkish balanced bilinguals studied word lists which are associates of a critical lure. Some lists were received by the participants in English while others in Turkish. After the study phase, participants performed a recognition test that included studied words, nonstudied critical lures, and nonstudied new items whose language were either the same with or different from those used in the study phase. Half of the participants were instructed to recognize only if the test word was in the same language as the studied word while the remaining received no such instruction. The results demonstrated a strong false memory effect for both within-language and across-language conditions. False recognition decreased when the study and test languages differed compared to when they were the same, and this effect was similar for both groups which had an instruction and no instruction. These results were interpreted in such a way that even though false memories can occur across different languages, they are not completely independent from the language assuming that the participants encode not only the meaning of the words but also the language of the words (i.e., the phonological form of a word in that particular language) that they were presented during the study phase, so that they were encoded as distinct and presented less false alarms. In terms of bilingual language representation, the results are consistent with models that assume a separate lexical representation for each language and yet a common single conceptual representation for both languages.