ID de réunion : 957 4747 8729
Mot de passe : 466990
Efrat Levant (Tel Aviv University)
Default sarcastic interpretations of attenuated and intensified similes
What causes an utterance to be interpreted as sarcastic? Does it have to do with familiarity, context, or structure? Are some constructions always sarcastic, regardless of other factors? Many theories have attempted to find out what makes an utterance sarcastic. The Defaultness Hypothesis suggests that attenuating highly positive concepts (of novel, ambiguous similes, involving no internal incongruity) will generate sarcastic interpretations by default (i.e. automatically, regardless of context or salience). This theory was previously proven on two constructions – negations (X is not the mostY)and rhetorical questions (Do you really believe he is smart?).
In this work, we test the effect of attenuators (about/almost/at least) and intensifiers (absolutely /really/so) on the sarcasm levels of similes within the framework of the Defaultness Hypothesis. Five offline experiments were performed. Our results support the prediction of the Defaultness Hypothesis, showing that indeed strongly attenuating highly positive concepts generate sarcastic interpretations by default. Additionally, our experimental results diverge from previous corpus-based findings. This prior work found that attenuated similes involving internal incongruity are significantly more sarcastic than their unattenuated counterparts. In contrast to these findings, we show that no significant difference in sarcasm rating was detected between attenuated/unattenuated similes when containing internal incongruity. However, when comparing attenuated/unattenuated similes without internal incongruity, a significant difference in sarcasm rating was found. We conclude that, as predicted by the Defaultness Hypothesis, it is attenuating highly positive concepts that affect sarcasm by default, i.e., even without relying on supportive context or internal incongruity.