Does this sentence make sense? It depends on who said it!

Sentences uttered by foreign-accented speakers are often judged differently compared to those produced by native speakers. For instance, true statements that are not widely known (Ants don’t sleep) tend to be judged as less credible when produced by foreign-accented speakers. This phenomenon has been explained based on different intelligibility whereby foreign accent involves more processing load in the listeners, influencing the judgment of the statement. Interestingly, statements with omission information (Some dogs are mammals) tend to be judged as more acceptable when associated to foreign-accented speakers. This phenomenon has been explained based on different expectations about the speech of foreign and native persons (foreign speakers will commit more grammatical errors and pragmatic violations). In sum, while there is a tendency to forgive the omissions to foreign-accented speakers, they are also considered as less credible. Here we explored for the first time these two phenomena under the same experimental conditions by exploring how unknown and under-informative statements are judged when associated to native and foreign-accented speakers. Statements were presented in a written modality. In the first set of experiments, 200 Italian native participants read a bio description of two persons (an Italian and a foreign-accented fellow). Two critical types of statements were presented: Unknown and Under-informative. In addition, two filler statement conditions were added: True and False. For each participant, half of the statements were associated to the Italian and the other half to the foreign-accent speaker. Participants were required to rate how much each statement makes sense on a five-point scale. Linear mixed-effect models were performed on the ratings. Results showed more acceptable judgments in Under-informative statements when associated to the foreign-accented speaker. Non speaker effect was reported in the Unknown condition. This pattern of results was replicated in a new set of experiments (n = 200) combining facial images with the bio information. The lack of an effect on the Unknown condition seems congruent with those theories that identify cognitive load as the origin of the lower credibility for foreign-accented utterances. Furthermore, our results show pragmatic violations make more sense when associated to a foreign-accented speaker. This result is congruent with theories that postulate that expectations on the linguistic speaker's identity can modulate language processing. The conclusion that the role of expectation in language processing is not restricted to auditory modality but also affects the written modality may have relevant implications for native and non-native social interactions.