The present study investigated how pragmatic information is integrated during L2 sentence comprehension. We put forward that the differences often observed between L1 and L2 sentence processing may reflect differences on how various types of information are used to process a sentence, and not necessarily differences between native and non-native linguistic systems. Based on the idea that when a cue is missing or distorted, one relies more on other cues available, we hypothesised that late bilinguals favour the cues that they master during sentence processing. To verify this hypothesis we investigated whether late bilinguals take the speaker's identity (inferred by the voice) into account when incrementally processing speech and whether this affects their online interpretation of the sentence. To do so, we adapted Van Berkum, J.J.A., Van den Brink, D., Tesink, C.M.J.Y., Kos, M., Hagoort, P., 2008. J. Cogn. Neurosci. 20(4), 580-591, study in which sentences with either semantic violations or pragmatic inconsistencies were presented. While both the native and the non-native groups showed a similar response to semantic violations (N400), their response to speakers' inconsistencies slightly diverged; late bilinguals showed a positivity much earlier than native speakers (LPP). These results suggest that, like native speakers, late bilinguals process semantic and pragmatic information incrementally; however, what seems to differ between L1 and L2 processing is the time-course of the different processes. We propose that this difference may originate from late bilinguals' sensitivity to pragmatic information and/or their ability to efficiently make use of the information provided by the sentence context to generate expectations in relation to pragmatic information during L2 sentence comprehension. In other words, late bilinguals may rely more on speaker identity than native speakers when they face semantic integration difficulties.

In two experiments, male and female rats were trained in a Morris pool in the presence of 1 (Experiment 1) or 2 (Experiment 2) landmarks, which were placed relatively close in relation to a hidden platform. Experiment 1 established the relative salience of 3 landmarks. Two of them revealed a similar salience, and smaller than a third one, the most salient landmark, both in training and on a test trial without the platform. Then in Experiment 2 rats were extensively trained to find a hidden platform in the presence of a configuration formed by 2 landmarks and the effects of varying the salience of one of the landmarks were studied. Subsequent test trials without the platform revealed that finding the platform was controlled by different strategies and that the rats were taking advantage of this redundancy depending on the nature of the test trials. Surprisingly, in Experiment 2 a clear sex difference was found on escape trials only, with males reaching the platform faster than females.