Processing metalinguistic negation: Evidence from eye-tracking experiments

• We conducted eye-tracking experiments on negative sentences in Korean.
• Clarifications of MNs and DNs took similar processing times.
• The issue of MN interpretation is very similar to the issue of GCIs.

This paper investigates the interpretative process of metalinguistic negation (MN), as opposed to descriptive negation (DN), by using eye-tracking experiments on negative sentences in Korean. It has been suggested that negation is interpreted as descriptive by default and that an MN interpretation is taken only after the DN interpretation turns out to be a semantic contradiction to the clarification clause (a semantic account). Another suggestion is that the type of negation is chosen by considering optimal relevance. That is, people take an interpretation that yields greater cognitive effects with less processing effort (a cognitive account). Eye-tracking experiments were conducted on MN–DN pairs that received similarly high ratings on sensicality: the first with external negation and the second with long-form negation. In each pair, the two negative clauses were followed by the same clarification clause. In the experiments, clarification clauses showed no significant differences in the processing time between MNs and DNs. The results provide no evidence of the semantic account that the participants interpreted negation as descriptive by default. We suggest that they decided on the type of negation when they read the clarification clause, so that their processing times at the clarification clauses were not different between MNs and DNs, which is consistent with the cognitive account.