Are Associative and Propositional Processes Qualitatively Distinct? Comment on Gawronski and Bodenhausen (2006) Arie W. Kruglanski University of Maryland Mark Dechesne Radboud University of Nijmegen The authors comment on B. Gawronski and G. V. Bodenhausen’s (2006) associative-propositional evaluation model of implicit and explicit attitudes by examining the claims that (a) truth value is attached to propositions but not to associations; (b) pattern activation is qualitatively different from syllogistic structure of arguments; and (c) Pavlovian conditioning may be propositional, whereas evaluative conditioning is not. They conclude that despite surface dissimilarities between implicit and explicit attitudes both may be mediated by the same underlying process. Keywords: implicit and explicit attitudes, reasoning, conditioning The distinction between implicit and explicit attitudes has re- cently been of considerable interest to social cognition researchers (Fazio & Olson, 2003; Greenwald & Banaji, 1995; Wilson, Lind- sey, & Schooler, 2000). Not only are explicit and implicit attitudes measured differently but often they do not correspond to one another and are differentially influenced by the same variables (e.g., Dasgupta & Greenwald, 2001; Karpinski & Hilton, 2001). In the accompanying article, Gawronski and Bodenhausen (2006) set out to explain the differences between implicit and explicit atti- tudes with respect to their associative-propositional evaluation, or APE, model. The basic claim of the APE model is that implicit attitudes derive from associative processes that are qualitatively distinct from propositional processes. The differences are assumed to incorporate several features. According to the APE model, asso- ciative processes do not involve the assignment of truth value, whereas propositional processes do. Associative processes are automatic, whereas propositional processes are controlled. Ac- cordingly, associative processes do not require resources, whereas propositional processes do. Associative processes are based on pattern activation, whereas propositional processes are based on syllogistic reasoning. Finally, “true” associative processes are ex- emplified by evaluative conditioning (EC) rather than by Pavlov- ian conditioning (PC), which has been portrayed as propositional (e.g., Holyoak, Koh, & Nisbett, 1989). In the present article, we comment on the APE model by examining its claim concerning the distinctions between associa- tive and propositional processes. Whereas implicit attitudes are at times different from explicit attitudes and may respond to different manipulations, this, in and of itself, is not proof that they derive from different processes. For instance, it has been claimed that attitudes can be constructed on the spot, hence specific attitudes toward the same object can be (but do not have to be) different on different occasions even though the general construction process was the same (Erber, Hodges, & Wilson, 1995; Potter, 1998; Schwarz & Bohner, 2001). It is likewise possible that measure- ment of implicit and explicit attitudes toward the same object in a given instance may activate different “pieces” from which the attitude is constructed. In fact, different ways of measuring im- plicit attitudes also have been known to yield different results (e.g., measurement of attitudes via the Implicit Association Test yielded different results from measurement via affective priming; Olson & Fazio, 2003). In short, the fact that implicit attitudes toward the same attitude object often differ from explicit attitudes does not necessarily mean that the two types of attitude were arrived at via different processes. Simply, there exists substantial within (im- plicit or explicit) categories variability to question the validity of such a claim. Admittedly, the foregoing argument does not constitute com- pelling evidence against the possibility that implicit and explicit attitudes are mediated by different processes. To evaluate such a claim with any degree of cogency, one must consider more spe- cifically the characterization of these processes in the APE model. In what follows we do so by examining in turn the notions that (a) truth value is attached to propositions only and not to associations; (b) pattern activation is qualitatively different from syllogistic structure of arguments; and (c) whereas Pavlovian conditioning may be propositional, evaluative conditioning is not. The Issue of Truth Value According to the APE model, “The most important feature [of associative processes] . . . is that associative evaluations are inde- pendent of the assignment of truth values” (Gawronski & Boden- hausen, 2006, p. 693). By contrast, “the most important feature that distinguishes propositional from associative processes is their de- pendency on truth values” (p. 697). Yet, the authors’ statements repeatedly suggested that associations do have truth value attached Arie W. Kruglanski, Department of Psychology, University of Mary- land; Mark Dechesne, Department of Social Psychology, Radboud Uni- versity of Nijmegen, the Netherlands. This work was supported by National Science Foundation Grant SBR- 9417422 and Nederlandse Organisatie voor Wetenschappelijk onderzoek Grant 45102080. Correspondence concerning this article should be addressed to Arie W. Kruglanski, Department of Psychology, University of Maryland, College Park, MD 20742-4411. E-mail: Arie@psyc.umd.edu Psychological Bulletin Copyright 2006 by the American Psychological Association 2006, Vol. 132, No. 5, 736 –739 0033-2909/06/$12.00 DOI: 10.1037/0033-2909.132.5.736 736 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.