Embodiment theories suggest that Affective Priming (AP) may derive from somatic processes namely simulation of facial expressions triggered by stimulus valence. The current study used the AP paradigm in the context of a native (L1) or learned (L2) language and examined the general hypothesis that L2 is not emotionally anchored by somatic processes. Two sets of positive, negative and neutral words were presented as primes and as targets, either in Portuguese (L1) or English (L2). Participants had to classify target words as positive or negative. For L1 stimuli we observed the classic AP effect ‐ faster reaction times in congruent trials (prime+/target + and prime‐/target‐) than in incongruent trials (prime+/target‐ and prime‐/target+). As predicted these effects were not observed in L2. Error rates were similar in L1 and L2 discarding differences in comprehension but participants were faster in L1 suggesting that the lack of simulation in L2 slows down the process.

--