Validating the truth of propositions: behavioral and ERP indicators of truth evaluation processes

doi:10.1093/scan/nss042 SCAN (2013) 8, 647^ 653 Validating the truth of propositions: behavioral and ERP indicators of truth evaluation processes Daniel Wiswede,1,2 Nicolas Koranyi,1 Florian Müller,1 Oliver Langner,1 and Klaus Rothermund1 1 Friedrich-Schiller Universität Jena, Department of General Psychology, Am Steiger 3, Haus 1, D-07743 Jena, and 2Department of Neurology, University of Lübeck, Ratzeburger Allee 160, 23538 Lübeck, Germany We investigated processes of truth validation during reading. Participants responded to true and false probes after reading simple true or false sentences. Compatible sentence/probe combinations (true/true, false/false) facilitated responding compared with incompatible combinations (true/ false, false/true), indicating truth validation. Evidence for truth validation was obtained after inducing an evaluative mindset but not after inducing a non-evaluative mindset, using additional intermixed tasks requiring true/false decisions or sentence comparisons, respectively. Event-related potentials revealed an increased late negativity (500–1000 ms after onset of the last word of sentences) for false compared with true sentences. Paralleling behavioral results, this electroencephalographic marker only obtained in the evaluative mindset condition. Further, mere semantic mismatches between subject and object of sentences led to an elevated N400 for both mindset conditions. Taken together, our findings suggest that truth validation is a conditionally automatic process that is dependent on the current task demands and resulting mindset, whereas the processing of word meaning and semantic relations between words proceeds in an unconditionally automatic fashion. INTRODUCTION For experienced readers, reading is a more or less effortless process. Research on reading comprehension has shown that the meanings of words, as well as the semantic and syntactic relations between the words of a sentence are processed automatically when we are reading text (e.g. Balota et al., 1990). For example, semantic (‘A robin is a truck.’) or syntactic mismatches (‘A robin is a quickly.’) are detected automatically and produce elevated Event-related potentials (ERPs) (N200, N400, P600) while reading the last word of a mismatching sentence (Kutas and Hillyard, 1980, 1984; Fischler et al., 1983, 1984; Friederici et al., 1993; Hahne and Friederici, 1999, 2002). The fact that processes of language comprehension and meaning extraction proceed in an automatic fashion is well established in the literature; it is less clear, however, whether mere reading also produces automatic truth validation going beyond what is explicitly stated in a sentence. The decisive question of this study is whether sentencerelated factual world knowledge that is stored in long-term memory also becomes automatically activated upon reading and understanding the sentence, and whether this knowledge is used to evaluate the truth status of the respective sentence. Previous ERP studies have shown that semantic mismatches between the subject and the object of a sentence produced elevated ERPs that occurred irrespective of the truth value of the proposition (Fischler et al., 1983): a strong N400 of equal magnitude was observed both for the false sentence ‘A robin is a truck.’ and for the true sentence ‘A robin is not a truck.’, with no systematic differences in ERPs between false and true sentences. This seems to suggest that the semantic consistencies between all words within a sentence are checked as part of an automatic comprehension process, but that the detection Received 18 March 2012; Accepted 24 March 2012 Advance Access publication 29 March 2012 This research was conducted at the University of Jena. We want to express our thanks to Nils Meier for support in programming the experiment. This study was supported by regular funding of the Department of General Psychology (Chair: Prof. Dr. Klaus Rothermund) by the FSU Jena. Correspondence should be addressed to Dr Daniel Wiswede, Department of Neurology, University of Lübeck, Ratzeburger Allee 160, 23538 Lübeck, Germany, or to Prof. Dr. Klaus Rothermund, Department of General Psychology, FSU Jena, Am Steiger 3, Haus 1, D-07743 Jena, Germany. E-mail: ; . of semantic matches or mismatches does not automatically lead to validation of whether sentences are true or false. Recently, however, this conclusion has been called into question by a study investigating whether reading simple sentences that were either obviously true (e.g. ‘Perfume contains scents.’) or obviously false (e.g. ‘Soft soap is edible.’) led to an automatic activation of their corresponding truth values (Richter et al., 2009).1 In support of such an automatic truth validation, Richter et al. (2009) have shown that reading true (false) sentences activated ‘correct’ (‘wrong’) responses that either facilitated or interfered with orthographic decisions (pressing a ‘correct’ or ‘wrong’ button) regarding the last word of the respective sentence. Apparently, both the truth values of sentences and the corresponding responses were activated during reading, even though a truth validation was not required by the orthographic task. Critically, however, it is possible that performing correct/wrong orthography decisions produced an evaluative mindset that may be necessary to trigger the encoding of propositional truth values, which would render the reported finding as support for a conditionally (i.e. task dependent) automatic truth validation process. Furthermore, because the material used by Richter et al. (2009) did not contain negated sentences, truth/falsity and matches/mismatches between sentence subject and object/predicate were confounded. It thus remains unclear whether participants really evaluated the truth value of the sentences or whether ‘true’ and ‘false’ response tendencies were simply triggered by matches/mismatches between the word components of the sentences (Fischler et al., 1983). The present study investigated processes of evaluating the truth of simple sentences, using behavioral (irrelevant stimulus–response compatibility), as well as electrophysiological (ERP) indicators of truth evaluation processes. Specifically, participants had to read short sentences that were either true or false. Orthogonal to their truth value, half of the sentences contained a negation, whereas the other half did 1 Richter et al. (2009) used sentences that ascribed features to objects (‘An OBJECT is/has FEATURE’), which differs from the sentences that were used in many previous studies that employed sentences describing category-exemplar relations (‘An EXEMPLAR is a CATEGORY’). Still, the materials used by Richter and colleagues were comparable with previous studies in that even wrong sentences were always combinations of objects and features that violated participants’ world knowledge but did not contain meaningless combinations of objects and features (e.g. ‘Ideas sleep in car (...truncated)