This study examines the integration of negation in sentences. It compares the affirmative forms with two forms of negation: 1. Prefixal negation (unauthorized) and 2. Sentential negation (not authorized). The aim is to determine (i) whether there is a delay in the integration of negation, and (ii) whether prefixal negation is processed in a similar way to the negated form or the affirmative form.

Previous studies using event-related brain potentials (ERPs) have shown that negation is ignored in early processing in the presence of semantic priming effects and incongruent world knowledge (Ferguson, Sanford & Leuthold, 2008, Fischler, Bloom, Childers, Roucos & Perry, 1983; Lüdtke, Friedrich, De Filippis & Kaup, 2008). Based on these findings, the “two-step simulation hypothesis” was developed (Kaup, Lüdtke, & Zwaan, 2006; Lüdtke et al., 2008). According to this hypothesis, language users first simulate the affirmative concept and only later integrate negation (e.g. ‘open door’ and ‘closed door’, respectively, in The door is not open) (Kaup et al., 2006). Other studies have provided evidence suggesting negation can be integrated immediately if the context in which it occurs is optimal and negation fulfills its most natural function of rejecting a plausible statement (Nieuwland & Kuperberg, 2008; Nieuwland & Martin, 2012).  

The present study. Using ERPs, this study revisited this issue by investigating the integration of negation in a sentence comprehension task. Participants (N=26) read sentences such as The White House announced that the new Obama biography was authorized/unauthorized/not authorized therefore the details in the book were correct/wrong in actual fact, where the first part of the sentence contained the negated adjective and the second part contained one member of an antonym pair (correct/wrong), according to which the sentence was either congruent or incongruent. ERPs were time-locked to the antonym in the second part of the sentence and amplitudes were analyzed in two time-windows of 300-400-msec and 500-700-msec.

Results and discussion. In affirmative sentences, Incongruent condition resulted in a larger N400 followed by a larger P600 in the Parietal region and Central region. For Prefixal negation, a larger negativity was observed in both time-windows in the Frontal and Central regions. For sentential negation, no effect of Congruency was found between 300-400 msec. However, in the 500-700-msec time-window, a larger negativity was observed for Incongruent compared to Congruent sentences in the Parietal region.

Conclusion. These findings suggest that while participants react to anomalies in affirmative sentences, they have difficulty processing sentences with prefixal and sentential negation. Both negation types elicit a larger negativity different from the typical N400 which suggests that negation has not been fully integrated at that point in time.