ABSTRACT
When processing negative sentences without context, participants often represent states of the positive arguments. Why and when does this occur? Using visual world eye-tracking, participants listened to positive and negative sentences in simple or cleft forms (e.g., [It is] Matt [who] hasn’t shut his dad’s window), while looking at scenes containing a target and a competitor (matches or mismatches the implied shape of the final noun). Results show that in the simple but not the cleft condition, there is a difference between negatives and positives: shortly after the verb, there is more looks to the competitor in the simple negatives than the positives. This suggests that the representation of the positive is not a mandatory first step of negation processing (as per rejection accounts). Rather results support the Question Under Discussion (QUD) accommodation account wherein both sentence content and contextual source of relevance are targets of incremental sentence processing.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. “Positive” and “affirmative” are used interchangeably in this paper. We use “argument of negation” to refer to the semantic, conceptual content of the affirmative counterpart of the sentence containing the negative element, “not”.
2. See Geurts (Citation2010) for a discussion of implicatures derived from presuppositions.
3. But note that other visual world studies report a delay in access to implicatures (see Tomlinson et al., Citation2013).
4. We did not use adjectives gap fillers such as “Matt has/hasn't shut the bright new window”, because adjectives (bright new) often imply visual qualities, which can be used to identify the target. In addition, adjective–noun phrases are often used contrastively, implying a salient alternative with different qualities. This will make the sentences less felicitous against our visual scenes.
5. We did not instruct the speaker to insert a post-verb silence. It occurred from natural reading.
6. See Appendix 2 for plots showing the “raw” probability values, separately for target and competitor.
7. The log correction is applied since fixation proportions are bounded by 0 and 1.
8. It is worth noting that the rejection account proposed by Clark and Chase (Citation1972) and Carpenter and Just (Citation1975) was devised in the period before ‘real-time’ incremental processing was accepted in psycholinguistic research.