Work in Progress Doc. 1 Cross-Modal Perceptual Effects on Consumption Experience Work In Progress Document David Gal

Work in Progress Doc. 2 Summary Marketers use product packaging—including brand names and package designs—to convey information about their products to consumers. For instance, both the Go-Gurt brand name and its package design—featuring a kid riding on skates—suggest a brand that is for active kids on the go. Likewise, according to naming consultants involved in naming the erectile dysfunction drug Viagra, the name Viagra is intended to evoke semantic associations through its phonetic resemblance to the words “vigor” and “Niagara.” However, many product names and package designs have no clear logical or semantic associations with information relevant to the products. For instance, the Crest mouthwash bottle features many corners and hard edges. Why would Crest choose this particular bottle shape over any other bottle shape? Can the mere shape of a mouthwash container influence consumer perceptions of the mouthwash’s flavor or effectiveness? Indeed, there is both neuroanatomical and experimental evidence to suggest that it could. For instance, Ramachandran and Hubbard (2001) showed subjects two shapes—one featuring jagged edges and one featuring rounded edges. The subjects were informed that the shapes were Martian characters and that one of the shapes was named “Bouba” and that the other was named “Kiki;” the subjects were asked to guess which was which. Almost invariably, the experimental participants associated the Kiki name with the jagged shape and the Bouba name with the round-cornered shape. In my dissertation research, I propose that the result of this Bouba/Kiki experiment and similar evidence of cross-modal associations reflect an underlying biological correspondence between different sensory modalities and categories. That is, I propose that stimuli that people logically group into distinct perceptual categories often affect perceptions across modalities, and that these cross-modal perceptions reflect an inherent correspondence

Work in Progress Doc. 3 between phenomenological experiences in the brain rather than any logically mediated or learned associations. In accord with this account, I hypothesize that stimuli that are thought to invoke one perceptual modality (e.g., visual) can impact sensory inputs to other modalities (e.g., gustatory or somatosensory). That is, I posit that sensory inputs that are specific to one modality (the “inducing stimulus”) can influence perceptions of an unrelated stimulus in a separate modality (the “target stimulus”), in the absence of any logical or semantic associations connecting the two stimuli. In a first demonstration study, I show that participants rate the flavor of a mouthwash as sharper when the mouthwash is contained in a bottle with an angular shape than when the same mouthwash is contained in a bottle with a rounded shape. Although there is no obvious logical basis for this result, this experiment cannot entirely preclude an explanation based on learned associations. Therefore, in a second study, I use insights from research on perception to show that the effect observed in the demonstration study can be flipped. That is, I show that, under some conditions, the effect of an inducing stimulus on judgments of a target stimulus will be opposite people’s a priori associatons. In particular, I show that under some conditions people rate a towel with the brand name “bouba” to be harsher in texture than one named “kiki” even though people predict, a priori, that bouba is softer than kiki. In a third study, I show that perceptual spillover effects can arise through incidental exposure, in the absence of any association between the inducing and target stimulus; i.e. when the stimuli are not both attributes of the same product. In particular, I show that participants primed with a set of shapes featuring jagged edges judge a cheddar cheese to taste sharper than participants primed with a set of shapes featuring rounded edges.

Work in Progress Doc. 4 In a series of additional studies, I test the notion that cross-modal perceptions can impact behavior. For instance, in one experiment, I show that bright lighting decreases the duration that subjects are able to maintain their hand in chilled water relative to dim lighting. I attribute this result to the cross-modal nature of intensity, whereby the intensity of the bright lighting is transferred to the intensity of cold/pain experienced by the participants. In a subsequent experiment, I rule out an alternative account that the bright light was merely distracting to participants, and that this distraction was the source of the enhancement of subjects’ pain. Preliminary data from another study in this series suggest that the character of lighting also impacts participants’ taste of coffee and the quantity of coffee they drink. Specifically, the preliminary data suggest that participants who prefer mildly flavored coffee drink less coffee under bright lighting conditions, whereas those who prefer intensely flavored coffee drink more coffee under such lighting conditions. The findings of these experiments are important for consumer research, because they suggest that cross-modality can influence consumers’ perceptions and behaviors associated with a consumption experience. More broadly, the notion that inputs to one sensory modality can influence actual perceptions in seemingly distinct sensory modalities, provides insight into the nature of perception itself. In particular, these findings suggest that seemingly distinct sensory modalities are not distinct after all, but share an underlying biological correspondence that may account, in part, for the origins of metaphor, art, and language.

Work in Progress Doc. 5

Cross-Modal Perceptual Effects on Consumption Experience

Marketers use product packaging—including brand names and package designs— to convey information about their products to consumers. For instance, both the Go-Gurt brand name and its package design—featuring a kid riding on skates—suggest a brand that is for active kids on the go. Likewise, according to naming consultants involved in naming the erectile dysfunction drug Viagra, the name Viagra is intended to evoke semantic associations through its phonetic resemblance to the words “vigor” and “Niagara” (Wall Street Journal). However, many product names and package designs have no clear logical or semantic associations with information relevant to the products. For instance, Crestbrand mouthwash comes in a bottle featuring many corners and hard edges. Why would Crest choose this particular bottle shape over any other bottle shape? Can the mere shape of a mouthwash container influence consumer perceptions of the mouthwash’s flavor or effectiveness? Lay theories of perception, namely that different sensory modalities, such as vision and taste, are categorically distinct, would seem to preclude such an influence of bottle shape on taste. Likewise, researchers modeling the brain have tended to assume that perception arises from distinct modules in the brain for different types of sensory inputs. In particular, researchers have presumed that lower-level sensory processing is unisensory, whereas multisensory integration—where one sense may inform another— occurs only during higher level processing (e.g. Felleman & Van Essen, 1991). Thus, unless a particular bottle shape conveys some learned or other semantic associations,

Work in Progress Doc. 6 traditional models of sensory processing appear to preclude an influence of bottle shape on mouthwash taste. However, is the prevalent model of sensory processing correct? The prevalent model of discrete processing modules for different sensory modalities was supported by early anatomical studies of primate brains that indicated that auditory, visual, and somatosensory cortices lacked interconnections (Kuypers et al, 1965), and by studies in which experimental lesions in primate cortical areas led to modal-specific deficits (Massopust et al., 1965). However, a growing stream of more recent neuroanatomical evidence challenges the notion that sensory modalities are truly independent (e.g. Falchier et al., 2002; Rockland & Ojima). For instance, Poremba et al. (2003) found that cortical areas in monkeys thought to be strictly associated with visual processing were activated by auditory stimuli. The growing evidence for interconnections amongst sensory modalities is highlighted by a recent review of the neuroanatomical evidence for such interconnections, which concludes that “...the work published to date may reveal only the ‘tip of the iceberg’, as ongoing studies continue to reveal extensive interactions among low-level sensory areas” (Ghazanfar and Shroeder, 2006). This emerging neuroanatomical evidence for sensory interconnections appears consistent with experimental findings showing that people are able to make congruency judgments across seemingly distinct sensory domains. For instance, in one study, modeled on earlier studies by Kohler (1929, 1947) and Werner (1934, 1957), Ramachandran and Hubbard (2001) reported showing experimental participants two shapes, one featuring jagged edges and the other featuring soft, rounded edges (see Figure 1). Participants were told to assume that the shapes were Martian characters

Work in Progress Doc. 7 named “bouba” and “kiki,” and were asked to identify which name they thought was associated with each shape. Over 95% of participants identified the soft, rounded shape as bouba and the jagged shape as kiki. Koehler (1929) had earlier obtained similar results when the same stimuli were named takete and baluma. Similar experiments to the bouba/kiki experiments of Ramachandran and Hubbard were conducted by Yorkston and Menon (2004). In a between-subject design, these researchers asked participants to read the description of an ice cream that was named either Frish or Frosh. Their results showed that participants tended to infer that Frosh was a richer and creamier ice cream than Frish. Their findings were consistent with prior research in linguistics suggesting that the position of the tongue in the mouth during the production of vowel sounds conveys symbolic meaning. In particular, this research suggests that the vowels produced when the tongue is in the low, back position of the mouth (e.g. [oo]) are associated, inter alia, with the qualities of large size, power, broadness, and richness; conversely, vowels produced when the tongue is in the high, frontal position of the mouth (e.g. [ee], [i]) are associated, inter alia, with the qualities of small size, lightness, mildness, fastness, and bitterness (e.g. Hinton et al., 1994; Makino et al., 1999). Although evidence for cross-modal congruency judgments suggests an intrinsic correspondence between what are traditionally regarded as distinct sensory modalities, such a conclusion is far from clear. In the bouba/kiki experiment, participants merely assigned names to target shapes; in the frish/frosh experiments, participants evaluated ice cream based on a written description and manipulated name. From these experiments, it is unclear whether sensory inputs to one modality (i.e. the auditory modality) directly

Work in Progress Doc. 8 influenced perceptual experience in the target modality, or if participants merely inferred meaning from the auditory phonemes and subsequently judged stimuli in the target modality based on these inferred meanings. Stated differently, it is unclear from the existing experiments whether participants’ perceptual experience in one sensory modality was affected by inputs to the other sensory modality through cross-modal percepts or if participants were using the symbolic meaning inherent in phonemes as information to deduce the qualities of the target shapes or ice cream flavors. This distinction between a cognitive-inferential-linguistic account and an intrinsic-biological-perceptual account is a manifestation of one of the most vigorous debates in cognitive science: the basis of cognition itself. In particular, this debate concerns whether cognition arises purely from the abstract data manipulation capabilities of the brain—a view in accord with traditional models (e.g. Elman et al., 1996) and the Whorfian hypothesis (e.g. Whorf, 1956)—or whether the basis of cognition is embodied, reflecting the structure of the brain itself, with its various modules and specific interconnections between them (e.g. Lakoff and Johnson, 1980; Damasio, 1994; Dehaene et al., 2006). In my dissertation research, I propose an account of cross-modal associations that is consistent with the latter, embodied-cognition view. That is, I propose that crossmodal associations reflect an underlying biological correspondence between different sensory modalities and categories, rather than any logically mediated or learned associations. More specifically, I propose that stimuli that people logically group into distinct sensory categories often give rise to percepts that are not specific to a single modality, but that affect processing of inputs across multiple sensory modalities.

Work in Progress Doc. 9 In accord with this account, I hypothesize that stimuli that are thought to invoke one perceptual modality (e.g., visual) can influence perceptions of stimuli in other modalities (e.g., gustatory or somatosensory). That is, I posit that sensory inputs that are assumed specific to one modality (the “inducing stimulus”) can influence perceptions of an unrelated stimulus in a separate modality (the “target stimulus”), in the absence of any logical or semantic associations connecting the two stimuli.

Experiment 1 Experiment 1 was designed as a demonstration study to test the notion that inputs to one sensory modality could impact people’s judgments about inputs to another sensory modality. Yorkston and Menon (2004) previously showed that people evaluating a written description of an ice cream use the sound of the ice cream’s brand name to make inferences about its flavor. However, an unanswered question remaining from their research was whether the sound of the brand name would still have had an impact on people’s judgments of the ice cream’s flavor if they had actually tasted the ice cream rather than merely having read a written description of the ice cream. That is, can sensory inputs to one sensory modality actually influence sensory experience in another modality or are Yorkston and Menon’s results a reflection of an inference that is relevant only in lieu of actual sensory experience? To test whether sensory inputs to one sensory modality can affect people’s judgments about actual sensory inputs to another sensory modality, I performed an experiment in which I asked participants to evaluate the sharpness of a mouthwash’s flavor when the mouthwash was contained in either a round-shaped container or an

Work in Progress Doc. 10 angular container. Although, objectively, a sharp flavor and an angular-shaped container are unrelated, I posited that both evoked a nonmodal-specific (i.e. cross-modal) percept of “sharpness,” the existence of which was posited based on the bouba/kiki findings of Ramachandran and Hubbard (2001). Given this proposition, I predicted that participants would evaluate the same mouthwash to have a sharper flavor when it was contained in the angular-shaped container than when it was contained in the round-shaped container.

Procedure Ninety-seven subjects, undergraduates at a large North American university, were randomly assigned to evaluate a mouthwash contained in either a round-shaped container or an angular-shaped container. Prior to the evaluation task, the round-shaped container (which initially contained Safeway’s private label mouthwash) was emptied of its contents and thoroughly rinsed; the bottle was then filled with the same mouthwash as that contained in the angular-shaped container (Crest-brand mouthwash.) All label information on the containers was concealed by dark tape prior to the evaluation task. The mouthwash evaluation task was one in a series of unrelated product evaluations performed by the participants as part of a market research study. Participants were told that they would be evaluating a new brand of mouthwash and instructed to measure out a single dose of mouthwash into a disposable plastic cup. To ensure that doses were equivalent in both conditions, a line was drawn on the plastic cups prior to the experiment to indicate the correct filling dose. Participants were instructed to use the mouthwash as they normally would, and to spit it out back into the plastic cup after they

Work in Progress Doc. 11 had completed their use of the mouthwash. Afterwards, participants deposited the plastic cup into a nearby receptacle. After using the mouthwash, participants answered a brief series of evaluation questions. The main question of interest for the purpose of our study was participants’ ratings of the sharpness of the mouthwash’s flavor. Participants’ provided their rating on a seven point Likert scale ranging from ‘1’ (‘not at all sharp’) to ‘7’ (‘very sharp’).

Results and Discussion Consistent with my prediction, participants judged the same flavor of mouthwash to taste sharper when that mouthwash was contained in an angular-shaped container (M=4.51) than when it was contained in a round-shaped container (M=4.02; t(95)=2.14, p=0.035). Thus, it appears that people’s judgments of actual gustatory experience can be influenced merely by the shape of a product’s packaging. Moreover, although previous research has shown that people’s judgments of sensory inputs can be influenced by information, such as that contained in brand names (e.g. Shiv and Nowlis, 2004), experiment 1 suggests that people’s judgments of sensory inputs can be influenced by product attributes that bear no informational content. Nonetheless, the absence of overt informational content in the shapes of the mouthwash containers does not necessarily imply that participants did not derive meaning from the shapes of the containers—perhaps through learned associations. Therefore, it is possible that participants inferred the flavor of the mouthwash based on the shapes of the mouthwash containers rather than that their perception of the

Work in Progress Doc. 12 mouthwash’s flavor was affected by the shape of the container through the cross-modal perception of sharpness (as proposed in the present research). Moreover, it is possible that an inference as to the sharpness of the mouthwash’s flavor influenced participants’ judgments of the mouthwash’s flavor (i.e. a response bias) without affecting their actual sensory experience. Experiment 2 was designed to juxtapose such an inference-based account more directly with the cross-modal perception account that I proposed.

Experiment 2 Experiment 1 showed that inputs to one sensory modality could influence people’s judgments about inputs to a different sensory modality; however, the findings of experiment 1 did not preclude an inference-based account as an alternative to the proposed cross-modal perceptual account. In experiment 2, I sought to demonstrate that inputs to one sensory modality could influence judgments about inputs to another sensory modality in a manner consistent with a cross-modal perception account, but opposite in direction to that predicted by a logical inference account. In particular, in experiment 2, I paired soft-sounding or harsh-sounding brand names (‘bouba’ or ‘kiki’) with very soft or moderately soft towels. A natural prediction for an inference-based account is that participants’ should judge the towels to be softer when they are associated with a soft-sounding as opposed to a harsh-sounding name. However, I predicted that participants would only perceive that the moderately soft towel was softer when it was paired with a soft-sounding brand name, whereas they would perceive that the very soft towel was softer when it was paired with a harsh-sounding brand name.

Work in Progress Doc. 13 This prediction was based on the perceptual assimilation and contrast literature, the key finding of which is that perceptual assimilation tends to occur when objects are similar whereas perceptual contrast tends to occur when objects are dissimilar (e.g. Tootell et al., 1995). For instance, people tend to perceive that a circle is larger when it is circumscribed by a slightly larger circle (i.e. assimilation), but tend to perceive that it is smaller when it is circumscribed by a much larger circle (i.e. contrast) (e.g. Pressey, 1977). Although such assimilation and contrast effects have typically been observed within a single modality (e.g. object size), I posited that the same effects should be observed across modalities if cross-modal percepts exist. In particular, I posited that the haptic and auditory modalities share a common percept of “softness;” and, accordingly, that the softness of a texture and the softness of a sound were not ‘apples and oranges,’ but could be compared. Therefore, sounds associated with ‘similar’ textures would lead to assimilation (i.e. the texture’s softness would feel more like the softness of the sound), whereas sounds associated with ‘dissimilar’ textures would lead to contrast (i.e. the texture’s softness would feel less like the softness of the sound).

Stimuli The stimuli for this experiment were a moderately soft terry cloth towel and a very soft towel made of fine Egyptian cotton.

Procedure

Work in Progress Doc. 14 This study featured a 2(towel texture: moderately soft vs. very soft) X 2(brand name sound: Bouba vs. Kiki) between subject design. Participants, 144 undergraduates at a large university on the west coast of The United States, were told that they would be asked to evaluate a towel that a company was introducing. Participants completed the evaluation as part of a series of unrelated product evaluations that they were informed were for market research purposes. Participants were randomly assigned to evaluate either the moderately soft or the very soft towel and were informed, also via random assignment, that the towel’s brand name was either Bouba or Kiki. Participants were asked to form an evaluation of the towel and to complete a series of survey questions regarding their evaluation. The focal question of interest was participants’ rating of the towel’s softness on a seven point Likert Scale (from ‘1’ not at all soft to ‘7’ very soft).

Pretests (Stimuli) Prior to performing the experiment, I wanted to ensure that participants perceived that the “very soft” towel was indeed very soft and that the “moderately soft” towel was indeed moderately soft. Similarly, I wanted to ensure that participants perceived that the sound of the Bouba brand name was relatively soft-sounding and that the sound of the Kiki brand name was relatively harsh sounding. Accordingly, I elicited ratings of softness or harshness from ten different individuals for each item (i.e. the “very soft” towel, the “moderately soft” towel, the Bouba brand name, and the Kiki brand name). Respondents rated the towels’ softness on a bipolar 11 point scale ranging from -5 (very harsh texture) to 5 (very soft texture). Likewise, respondents rated the

Work in Progress Doc. 15 sound of the brand names on a bipolar 11 point scale ranging from -5 (very harsh sound) to 5 (very soft sound). Respondents’ ratings were as predicted: The “moderately soft” towel received a mean rating of 1.1, the “very soft” towel received a mean rating of 4.6, the Kiki brand name received a mean rating of (-2.5) and the Bouba brand name received a mean rating of 2.7. Pretests (Inference) The design of experiment 2 was intended to provide evidence for cross-modal perception by showing that an inducing stimulus could influence perceptions of a target stimulus in another modality in a direction opposite to that predicted by an inferencebased account. Therefore, I conducted a pretest to test how experimental participants inferred that the towels should be called. Ten participants evaluated the moderately-soft towel and were asked how wellsuited the name “Bouba” was to describe the texture of the towel and how well-suited the name “Kiki” was to describe the texture of the towel, each on a seven point scale from ‘1’ (not at all) to ‘7’ (very). Ten other participants provided similar evaluations for the very soft towel. Participants judged “Bouba” to be a better suited name than “Kiki” for both the moderately-soft (M = 4.1 vs. M = 3.5) and very-soft towel (M = 4.4 vs. 2.9). Results As predicted, participants perceived the very soft towel to be significantly softer when it was named Kiki (M=5.97) than when it was named Bouba (M=5.55; t(63)=2.23 p=0.03). In contrast, participants perceived the moderately soft towel to be softer when it

Work in Progress Doc. 16 was named Bouba (M=3.51) than when it was named Kiki (M=3.14); this latter difference was marginally significant (t(76)=1.86, p=0.067). The opposite effects of brand names on the very soft and moderately soft towel were reflected by a significant interaction effect (F(1,139)=8.04; p .45. I conducted additional analyses to test whether these taste perceptions would mediate liking of the cheese. As reported above, the shape prime exerted a significant effect on participants’ perceptions of cheese sharpness. To the extent that the participant likes sharp cheeses, these modified perceptions of sharpness should affect liking of the cheese, at least to the extent that sharpness contributes to liking or disliking cheese (above other aspects of the taste not assessed here). Hence, we would expect that the pointy-cornered shape prime would increase perceptions of sharpness, and therefore liking of the cheese, among participants who like sharpness, but decrease liking of the cheese among participants who dislike sharpness. To test this model, I used the procedures described by Preacher, Rucker, and Hayes (2005). Results for the first half of the model were presented above. The second half of the model was highly significant. Specifically, the perceptions of sharpness x liking of sharpness interaction strongly predicted liking of the cheese, B = .22, t = 6.62, p < .0001. I then tested the direction and significance of the mediational path from the shape prime to liking of cheese at high and low levels of liking sharp cheese (at one SD above and below the mean). Results indicated that among participants who like sharp cheese, there was a positive indirect path from the shape prime to liking of the cheese, B = 1.58, z = 1.79, p = .035, one tailed, indicating that the pointy shapes increased liking

Work in Progress Doc. 20 via increasing perceptions of sharpness. Among participants who dislike sharp cheese, this result was reversed and negative, B = 1.58, z = 1.64, p = .05, one tailed, indicating that sharp shapes decreased liking via decreasing perceptions of sharpness. It was clear that participants drew no connection between the two tasks (i.e. between the prime and taste evaluation tasks), and that they believed our cover story. No participants indicated that they drew a connection between the two tasks. Furthermore, after completing the experiment, nearly all participants spontaneously (or after being asked “how do you think you did?” by the experimenter) asked the experimenter to reveal which shape actually contained the largest surface area. Discussion In this study, consistent with my prediction, participants primed with shapes featuring pointy corners judged a cheddar cheese to taste sharper than participants primed with shapes featuring soft, rounded corners. Moreover, those participants who particularly liked sharp cheeses liked the cheese more following the pointy-cornered primes than following the rounded-cornered primes. What was the source of these differences in participants’ judgments of the cheese’s taste following the different primes? It is unlikely that the differences were due merely to cognitive inferences about the cheese’s taste, because participants did not make a connection between the priming and tasting tasks. Rather, the difference in participants’ taste judgments in the different priming conditions appears to imply that sensory inputs to the visual modality impacted participants’ sensation of sharpness in the gustatory modality through the activation of a cross-modal sharpness percept. In other

Work in Progress Doc. 21 words, the ability of sharp shapes to enhance sharp tastes appears to reflect a correspondence between the way human brains process sharp shapes and sharp tastes. What are the potential non-perceptual explanations for this finding? One potential explanation is that the pointy-cornered shapes in the shape-perception task activated the concept of sharpness, and that the shared concept of sharpness with the flavor of cheese influenced participants’ taste judgments. However, sharpness in the context of the shapes means ‘pointy,” and, therefore, does not reflect the same meaning as ‘sharpness’ in the context of taste. After all, a person cannot stab another with a sharp cheese. Indeed, the word sharp in the context of shapes and the word sharp in the context of cheese are homonyms, or more specifically polysemes—words that have a common linguistic origin, but that bear different meanings. And research has shown that activating a word in one context not only does not facilitate access to its polyseme in a different context, but may actually suppress it (e.g. Williams, 1992). Thus, lexical or conceptual priming accounts for the findings of experiment 3 seem unlikely.

Work in Progress Doc. 22 References Bowdle, B.F. & Gentner, D. (2005). The career of metaphor. Psychological Review, 112(1), 193-216. Damasio, A.R. (1994). Descartes' Error: Emotion, Reason, and the Human Brain. Harper Perennial: New York. Dehaene, S., Izard, V., Pica, P., & Spelke, E. (2006). Core knowledge of geometry in an Amazonian indigene group. Science, 311, 381-384. Duncan, J. (1984). Selective attention and the organization of visual information. Journal of Experimental Psychology: General, 113, 501-517. Elman J. L. et al. (1996) Rethinking Innateness: A Connectionist Perspective on Development: MIT Press. Eisenberger, N.I., Lieberman, M.D., & Williams, K.D. (2003). Does rejection hurt? An fMRI study of social exclusion. Science, 302, 290-292. Gardner, H. (1975). The shattered mind: the person after brain damage (New York: Knopf). Green, B.G. (1989) Capsaicin sensitization and desensitization on the tongue produced by brief exposures to a low concentration. Neuroscience Letters, 107(113), 173-178. Harrison, J.E. & Baron-Cohen, S. (1997), Synaesthesia: a review of psychological theories. Synaethesia: Classic and Contemporary Readings eds: Baron-Cohen, S. and Harrison, J.E. (Oxford: Blackwell). Hayes, J.E. (2000). Psychophysical and physiological response of humans to the oral irritant capsaicin. Unpublished Dissertation. Hinton, L., Nichols, J., & Ohala, J. eds. (1994). Sound Symbolism, Cambridge:

Work in Progress Doc. 23 Cambridge University Press. Kohler, W. (1929). Gestalt Psychology. (New York: Liveright). Kohler, W. (1947). Gestalt Psychology (2nd Ed.). (New York: Liveright). Lakoff, G. & Johnson, M.H. (1980). Metaphors We Live By (Chicago: University of Chicago Press). Makino, S., Nakada, S., & Ohso, M. (1999). Kodansha’s Basic English-Japanese Dictionary, Tokyo: Kodansha Int. Moll, J., de Oliveira-Souza, R., Moll, F. T., Ignacio, F.A. Bramati, I., Caparelli-Daquer, E.M., & Eslinger, P.J. (2005). The moral affiliations of disgust: A functional MRI study. Cognitive & Behavioral Neurology, 18(1), 68-78. Ramachandran, V.S. & Hubbard, E.M. (2003). Hearing colors, tasting shapes. Scientific American, May, 52-59. Ramachandran, V.S., & Hubbard, E.M. (2001). Synaesthesia -- a window into perception, thought and language. Journal of Consciousness Studies, 8(12), 3-34. Werner, H. (1934). L’unite des sens [The unity of the senses]. Journal de Psychologie Normale et Pathologique, 31, 190-205. Werner, H. (1957). Comparative Psychology of Mental Development. (Rev. ed.) (New York: International University Press). Whorf B.L. (1956) In: J.B. Carroll, Editor, Language, Thought and Reality: Selected Writings of Benjamin Lee Whorf, MIT Press. Williams, J.N. (1992). Processing polysemous words in context: Evidence for interrelated meanings. Journal of Psycholinguistic Research, 21, 193-218. Yorkston, Eric and Menon, Geeta (2004). A sound idea: phonetic effects of brand names

Work in Progress Doc. 24 on consumer judgments. Journal of Consumer Research, 31, 43-51.