This study tests the prediction made by the Parasitic Model of vocabulary acquisition (Hall, 2002), that upon initial exposure to a second language (L2), learners benefit from the existence of formal and conceptual overlap that is shared between pairs of cognates. The accuracy and response time performance of two groups of native English-speaking learners of French, one with less French exposure and one with more French exposure, was compared in a translation recognition task. In the task, participants had to decide whether two words were correct translation pairs when the first word presented was French and the second word presented was English. In the trials where participants were required to answer “no”, the English words were incorrect but were similar to the French word in form. Half of the incorrect pairs included a French word that is a true French-English cognate (e.g. gris–grim) and half of the incorrect pairs included a French word that is a true French-English noncognate (e.g. lapin – latin). In the trials where participants were required to answer “yes”, half the pairs were correct cognates (e.g. gris – grey) and the remaining pairs were correctnoncognates (e.g. lapin – latin). In the “no” trials there was no effect of cognate status of the French prime, but an overall effect of orthographic relatedness of the distracter item, which was present in both less and more L2 exposure groups. This result is taken as support of the theory of language non-selective access (Dijkstra& Van Heuven, 1998). In the “yes” trials there was an effect of cognate status, which provides support for the Parasitic Model. Moreover, the accuracy results show that the rate of the lexicon’s integration of orthographic knowledge is greater for novel noncognate words compared to novel cognate words. This result is taken as support for the hypothesis that the excitatory connections that exist between cognate representations remain permanent throughout second language acquisition.