LLF – Bât. ODG – 5e étage – Salle du conseil (533)
Yanwei Jin and Jean-Pierre Koenig (University at Buffalo)
Expletive negation: From typology to on-line comprehension
Expletive negation (EN) has been discussed extensively within Romance linguistics. In this talk, we explore the hypothesis that EN is due to general properties of language production: the meaning of some lexical items leads to the concurrent activation of a proposition (p) and its dual (not p) and it is this concurrent activation that causes the erroneous production of not p rather than the intended p.
We begin by presenting of a survey of the occurrence of EN across languages of the world (722 languages were surveyed overall); we then present a comprehensive list of EN-triggering contexts collected from French and Mandarin and compare that list with EN triggers in three other languages (4 language families in all). We then propose a language production and semantic account of the similarity of EN-triggering contexts found in these five languages. We propose that the meaning of EN triggers entails or strongly implies ¬p and that the activation of ¬p alongside p is what leads speakers to produce EN. Four semantic licensing conditions for EN triggers are identified and each EN-triggering context is semantically analyzed.
We then show through a corpus study that speakers of languages said not to include EN (English) produce them with some frequency in the same contexts in which they appear in languages where EN is entrenched (French, Mandarin): EN interpretations of negators ranged from almost 0% to 100% in our corpus, depending on the trigger (mean = 28.64%).
Finally, we report an on-line comprehension experiment that tests the hypothesis that native speakers of English interpret negators in complement clauses of EN triggers expletively. To test this hypothesis, we used the logic of the experiments on metaphorical class-inclusion statements reported in Glucksberg, Gildea, & Bookin (1982). We found that, as predicted, EN-trigger continuations elicited less logically accurate answers (p < .01) and longer response times (p < .01). This result suggests that English speakers do interpret negators expletively in the same contexts found to trigger grammaticalized EN in French or Mandarin. We also found that logical inaccuracy for EN continuations correlated with the frequency of EN interpretations in our corpus (r = .66, p < .01), suggesting that the more expletive interpretations of negators for a particular EN-trigger a speaker has encountered, the more likely she is to interpret expletively a new occurrence of that negator for that trigger.
Overall, our research suggests that French EN is not just some odd historical artefact, but is rooted in general cognitive mechanisms. What is special about Romance (French ne, in particular) is the degree to which EN has become entrenched (and grammaticized).