Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
To predict or not to predict: The role of context alternatives and truth for the processing of negation
Poster C15 in Poster Session C, Wednesday, October 25, 10:15 am - 12:00 pm CEST, Espace Vieux-Port
Maria Spychalska1, Viviana Haase2, Markus Werning3; 1Max Planck Institute for Psycholinguistics, 2Ruhr University Bochum, 32
Studies on negation processing often report a Polarity-by-Truth interaction: False affirmative sentences show longer response times and larger N400 ERPs relative to true affirmative sentences, whereas the effect is reversed for negative sentences. Furthermore, context has been shown to facilitate negation processing. The Polarity-by-Truth interaction has been linked to variations in lexical associations, predictability, or the need of constructing two subsequent mental representations during the comprehension of negative sentences. To disentangle the context-independent effect of negation as a syntactic operator from the context-dependent effect of predictability we ran five ERP-experiments employing a picture-sentence-verification paradigm. Predictability was manipulated by varying the number of alternative sentence continuations provided by the context and making it equivalent for both sentence polarities. The scenarios depicted three different objects, one or two of which were then selected (framed green) or rejected (framed red) by a virtual player. After the objects were marked, an affirmative or negative sentence (in German) was presented phrase-by-phrase, e.g., “Julia has/has not chosen the plum”. In experiments 1A/B, all stimuli sentences provided a true description of the scenario (false fillers were added). In the strongly constraining context, there was only one object that could complete a true sentence, i.e., one selected object in the affirmative condition or one rejected in the negative condition. In the weakly constraining context, there were two alternative objects that could complete a true sentence. The ERPs were measured on the sentence critical nouns referring to one of the objects. In experiments 2A/B, we tested affirmative and negative false descriptions in the two contexts (true filler sentences were added). In Experiment 3, all conditions were combined to test the Polarity-by-Truth-by-Context interactions. We show that the number of contextually available alternatives modulated the online processing of affirmative and negative sentences similarly. Relative to the weakly constraining context, in the strongly constraining context, where the processor could form a unique prediction for a true sentence continuation, we observed a smaller N400 response when the prediction was confirmed, and a larger N400 when the prediction was violated, for both sentence polarities. Over all experiments, the main effect of negation was observed as a long-lasting positivity occurring around 300 ms post-onset and extending towards the end of the 1000 ms epoch. The effect of Truth was directly dependent on the context rather than polarity: In the strongly constraining context, false sentence continuations elicited larger N400s than the true ones, for both sentence polarities, whereas the effect of Truth was reversed in the weakly constraining context for both negative and affirmative sentences. Additionally, we observed interactions between Truth, Polarity and Context. Our experiments show that after dissociating the cost of predictability, negative sentences still show a processing cost that is not reducible to the cost of an additional word in a sentence. This cost occurs both in the form of an interaction between negation and prediction mechanisms (as shown by negation and context-by-truth interactions in the N400 time-window) as well as in the form of effortful meaning integration mechanisms in negative sentences (positivity).
Topic Areas: Meaning: Discourse and Pragmatics, Syntax and Combinatorial Semantics