The Architecture of Intuition: Fluency and Affect Determine Intuitive Judgments of Semantic and Visual Coherence and Judgments of Grammaticality in Artificial Grammar Learning Sascha Topolinski and Fritz Strack University of Wuerzburg People can intuitively detect whether a word triad has a common remote associate (coherent) or does not have one (incoherent) before and independently of actually retrieving the common associate. The authors argue that semantic coherence increases the processing fluency for coherent triads and that this increased fluency triggers a brief and subtle positive affect, which is the experiential basis of these intuitions. In a series of 11 experiments with 3 different fluency manipulations (figure– ground contrast, repeated exposure, and subliminal visual priming) and 3 different affect inductions (short-timed facial feedback, subliminal facial priming, and affect-laden word triads), high fluency and positive affect independently and additively increased the probability that triads would be judged as coherent, irrespective of actual coherence. The authors could equalize and even reverse coherence judgments (i.e., incoherent triads were judged to be coherent more frequently than were coherent triads). When explicitly instructed, participants were unable to correct their judgments for the influence of affect, although they were aware of the manipulation. The impact of fluency and affect was also generalized to intuitions of visual coherence and intuitions of grammaticality in an artificial grammar learning paradigm. Keywords: intuition, processing fluency, remote associates, artificial grammar learning, visual coherence In modern psychology, there is an ever-increasing interest in intuitive processes, that is, information processes that occur with little awareness of the process itself (e.g., Deutsch & Strack, 2008; Hammond, 1996; Lieberman, 2000; Wilson, Lindsey, & Schooler, 2000), that are fast and effortless (e.g., Epstein, 1991; Gigerenzer, Todd & The ABC Research Group, 1999; Hamm, 2008; Hogarth, 2001; Kahneman & Fred- erick, 2002; Stanovich & West, 2000), that are independent from intention (e.g., Betsch, 2008; Epstein, 1991, 1994; Hog- arth, 2001; Topolinski & Strack, 2008), and that generate cer- tain internal cues, such as an intuitive hunch or gut feeling (e.g., called messages from within: Bless & Forgas, 2000; vibe: Epstein, 1991, 1994, 2008; cognitive feeling: Kahneman & Frederick, 2002; Loewenstein, Weber, Hsee, & Welch, 2001; or understanding by feeling: Bastick, 1982). Researchers have learned a lot about intuition (for a recent, extensive review, see Plessner, Betsch, & Betsch, 2008), its power in integrating vast amounts of complex information (e.g., Betsch, Plessner, Schwieren, & Gu ¨tig 2001; Dijksterhuis, 2004), its flex- ible efficiency (e.g., Gigerenzer et al., 1999), its foresight in guiding the problem-solver (e.g., Bowers, Regehrs, Balthazard, & Parker, 1990; Metcalfe, 1986), its deep connection to affect (e.g., Baumann and Kuhl, 2002; Bolte, Goschke, & Kuhl, 2003), and its shortcomings (e.g., Kahneman & Frederick, 2002; Tversky and Kahneman, 1973). However, there is little known about the un- derlying cognitive and affective processes that lead to intuitive hunches, which prompted Catty and Halberstadt (2008, p. 295) to state that intuition is still the “black box of modern psychology.” Take, for example, the following intuitive competence: When people are confronted with word triads that either share a common remote associate (e.g., salt, deep, foam imply sea; Mednick, 1962; Mednick & Mednick, 1967) or are only random word triads (e.g., dream, ball, book), they can intuitively feel the semantic coherence before and independently of actually retrieving this common as- sociate (Baumann & Kuhl, 2002; Bolte et al., 2003; Bowers et al., 1990). Moreover, people can discriminate between coherent and incoherent word triads above chance in less than 2,000 ms (Bolte & Goschke, 2005). This is an astonishing faculty because partic- ipants feel the existence of something that they do not know or, as Epstein (2008) put it, they know without knowing how they know—and neither do we researchers know how they know it. Although researchers have learned that coherent word triads auto- matically activate their common remote associate (Beeman et al., 1994; Shames, 1994; Topolinski & Strack, 2008) and that coher- ence intuitions are more diagnostic under positive mood than under negative mood (Baumann & Kuhl, 2002; Bolte et al., 2003), the mechanisms producing these intuitions remain inscrutable. Most recently, we opened this black box and connected fairly well-known mechanisms to keep track of this intuitive trace (Topolinski & Strack, in press-b). In a fine-grained analysis of the underlying processes, we traced processing fluency and positive Sascha Topolinski and Fritz Strack, Department of Psychology II, Social Psychology, University of Wuerzburg, Wuerzburg, Germany. We thank Friederike Finger and Rebecca Spatz for so conscientiously gathering the data. We thank Jane Thompson for valuable comments on the article. Correspondence concerning this article should be addressed to Sascha Topolinski, Department of Psychology II, Social Psychology, University of Wuerzburg, Roentgenring 10, 97070 Wuerzburg, Germany. E-mail: sascha.topolinski@psychologie.uni-wuerzburg.de Journal of Experimental Psychology: General © 2009 American Psychological Association 2009, Vol. 138, No. 1, 39 – 63 0096-3445/09/$12.00 DOI: 10.1037/a0014678 39