We conducted three experiments, which gathered (electro-)physiological and behavioural measures. The overall objective and results of each experiment are reported below.
Experiment 1.
Foucart, A., Santamaria-Garcia, H., Hartsuiker, R.. Short exposure to a foreign accent impacts subsequent cognitive processes. (in press). Neuropsychologia.
Although speaking a foreign language is undoubtedly an asset, foreign-accented speakers are usually perceived negatively. It is unknown, however, to what extent this bias impacts cognitive processes. We used ERPs and pupillometry to investigate whether the negative bias generated by a short exposure to a foreign accent influences the overall perception of a speaker, even when the person is not speaking. We compared responses to written sentence comprehension, memory and visual perception, associated with native speakers (high and low social status) and a foreign-accented speaker (high social status). The foreign-accented speaker consistently fell in-between the high-status native speaker and the low-status native speaker. This study was the first physiological demonstration that short exposure to a foreign accent impacts subsequent cognitive processes, and that foreign-accented speakers seem to be considered less reliable than native speakers, even with equally high social status. Awareness of this bias is essential to avoid discriminations in our multilingual society.
Experiment 2
Foucart, A., Costa, A., Hartsuiker, R. (in prep.). Foreign-accented speakers are not that ‘incredible’.
Previous results suggest that foreign-accented speakers are less credible than native speakers. This negative bias has been attributed to reduced processing fluency; the more difficult one is to understand, the less credible. We collected behavioural responses to replicate results and to evaluate the impact of accent on decisional processes. We used ERPs as a more fine-grained measure to examine the impact of accent on online (automatic) processes during sentence comprehension. We presented Dutch native speakers with audio sentences containing true (and known) information, unknown (but true) information or information violating world knowledge, spoken either in a native or a foreign accent. The sentences of particular interest were those containing unknown information because participants had to rely on the speaker’s knowledge to evaluate the veracity of the sentence. Behavioural measures did not show differences in the judgement of sentences spoken by the native and foreign speakers. As expected from the language literature, a larger N400 component was found for sentences containing unknown information and world knowledge violations than for true sentences, but it was not modulated by accent. Interestingly, accent impacted the perception of speaker’s status and the affect towards that speaker. The findings suggest that processing fluency may not have a determining role in the negative bias usually observed against foreign-accented speaker that is probably more strongly driven by social categorisation.
Experiment 3.
Foucart, A., Costa, A., Morís Fernández, L., Hartsuiker, R.. Foreignness or processing fluency? On understanding the negative bias towards foreign-accented speakers. (Registered Report, In-Principle-Acceptance). Language Learning.
There is a negative bias towards foreign-accented speakers. For example, trivia statements spoken in a foreign accent are usually assessed as less true than when spoken in a native accent. However, it is not clear to what extent this bias originates from social categorisation (in/out-group categorisation) and from processing fluency (ease with which information is processed). It has been argued that accent first induces a social identity effect and that processing fluency later modifies the impact of this effect during communication. Using ERPs, the study tested this hypothesis by looking at the effect of social categorisation and processing fluency on sentence comprehension (taking the N400 component as an index of information acceptability). The study aimed at reproducing behavioural results, and further investigates the impact of accent on trust at neural levels. The results revealed no difference across speakers at neural level but showed that accent modulates the status of the speaker and the affect towards her. The findings suggest that an accent triggers an immediate categorization of the speaker as an out-group member, which negatively affects the perception of the speaker. This negative perception is increased by the difficulty to process foreign-accented speech.
The development of the scientific aspects of this project has been combined with the accomplishment of other training and career development objectives.
