Abstract: In recent years, there has been an explosion of research on false memory with Deese–Roediger-McDermott paradigm (DRM). But few of these studies have involved bilingual participants and cross-linguistic effects in false memory. Furthermore, there were some mixed results between Sahlin, Harding & Seamon’s (2005) and Marmolejo, Diliberto-Macaluso & Altarriba’s (2009). In the present experiments, we used 16 DRM lists (8 English and 8 Chinese lists) as study materials and manipulated the DRM list lengths (7 item-list and 15 item-list)to explore the effect of different activation strength in encoding phase on cross-language false memory among less proficient Chinese - English bilinguals. 2 experiments were conducted to investigate the effect of item-specific processing and relational processing on cross-language false memory: Experiment1 studied cross-language false memory under item-specific processing condition, namely the participants were instructed to recognize an item as the old item according to both the meaning and language of the item ; a 2× 2 × 3mixed design was used, with two within-subject variables (language congruence: the same languages, the different languages; list type: studied words, critical lures, unrelated words), one between-subject variable (list length:15item- list,7item- list). Experiment 2 studied cross-language false memory under relational processing condition, study materials and experiment procedure in Experiment 2 were same as that in experiment 1, the only difference was to emphasize the relational processing in experiment 2, namely the participants were asked to judge an item as the old item only according to the meaning regardless of its language. Results of 2 experiments indicated that: 1)cross linguistic false recognition also existed among less proficient Chinese-English bilinguals; 2) In item-specific processing condition, false recognition of critical lures was more higher in the same languages than different languages regardless of the list length. In relational processing condition, false recognition of critical lures was more higher in the different languages than the same language regardless of the list length; 3) only in item-specific processing condition, false recognition of critical lures was more frequent in long list than that in short list, regardless of whether is the same language or different languages. The outcome of 2 experiments is accordance to activation/ monitoring theory.

Key words: false memory, DRM paradigm, cross-linguistic, language congruence, list length