Article révisé par les pairs
Résumé : Prior research suggests an egocentric bias in the ability to adopt a third-person perspective in sarcastic statements. However, it remains unclear whether (1) this bias is genuinely due to egocentric anchoring or to the cost of the activation of the sarcastic interpretation; (2) context-based, allocentric processing of sarcasm can be by-passed by cheaper strategies, such as prosody processing. To settle the first question, two sarcastic conditions were compared: one, ‘egocentric’, where the favored interpretation was sarcastic only from the participant's perspective, and another, ‘allocentric’, where the sarcastic interpretation was salient from both the addressee's and the participant's perspectives. To address the second question, performance in the egocentric and allocentric conditions were compared when salient prosodic cues were added. To show direct evidence for serial adjustment and to minimize the possibility of parallel processing of prosodic and contextual cues, we compare two experiments: In the first experiment, French-speaking participants had no time limit to respond, while time pressure was added in the second experiment. Results confirm that perspective-shifting is egocentrically anchored (i.e. slower reaction times and poorer accuracy for egocentric condition than allocentric one); furthermore, this egocentric bias is already evident in early stages of processing (within 3 s). We also show that perspectival assessment of contextual cues is not triggered in the presence of salient prosodic cues. Since perspective-taking is time consuming, using the non-contextual, prosodic cue is an efficient strategy to make an accurate judgment with the least processing effort.