Abstract
Poort, Warren and Rodd (2016) showed that bilinguals profit from recent experience with an identical cognate in their native language when they encounter the same word in their second language. We conducted two experiments employing the same cross-lingual long-term priming paradigm to determine whether this is also the case for non-identical cognates, as this would indicate they share an orthographic representation in the bilingual lexicon. In Experiment 1, Dutch–English bilinguals read Dutch sentences containing identical cognates (e.g. “winter”–“winter”), non-identical cognates (e.g. “baard”–“beard”) or the Dutch translations (e.g. “fiets”) of English control words (e.g. “bike”). These words were presented again in an English lexical decision task approximately 19 minutes later. The analysis revealed only weak evidence, based both on p-values and Bayes factors, for a small 6-9 ms facilitative priming effect. Experiment 2 aimed to determine whether including interlingual homographs (e.g. “angel”–“angel”) in the experiment modulates the size of the priming effect. This time, the analysis revealed no evidence for a priming effect, either based on p-values or Bayes factors, in either version of the experiment for either the cognates or the interlingual homographs. In line with previous findings (Poort & Rodd, 2017, May 9), we did find strong evidence for an interlingual homograph inhibition effect and no evidence for a cognate facilitation effect. We conclude that, since the cross-lingual long-term priming effect is largely semantic in nature, the lexical decision tasks we used were not sensitive enough to detect an effect of priming.