Search

Colors and odors are associated; for instance, people typically match the smell of strawberries to the color pink or red. These associations are forms of crossmodal correspondences. Recently, there has been discussion about the extent to which these correspondences arise for structural reasons (i.e., an inherent mapping between color and odor), statistical reasons (i.e., covariance in experience), and/or semantically-mediated reasons (i.e., stemming from language). The present study probed this question by testing color-odor correspondences in 6 different cultural groups (Dutch, Netherlands-residing-Chinese, German, Malay, Malaysian-Chinese, and US residents), using the same set of 14 odors and asking participants to make congruent and incongruent color choices for each odor. We found consistent patterns in color choices for each odor within each culture, showing that participants were making non-random color-odor matches. We used representational dissimilarity analysis to probe for variations in the patterns of color-odor associations across cultures; we found that US and German participants had the most similar patterns of associations, followed by German and Malay participants. The largest group differences were between Malay and Netherlands-resident Chinese participants and between Dutch and Malaysian-Chinese participants. We conclude that culture plays a role in color-odor crossmodal associations, which likely arise, at least in part, through experience.

Strong associations exist between specific odours and colours, and these associations have been found to be both consistent within populations and over time (Gilbert et al., 1996). Experimental manipulations of these associations have shown that both taste and odour perception rely heavily upon visual cues (e.g., Blackwell, 1995; Sakai et al., 2005); participants often make errors in odour judgements when stimuli have been artificially coloured (Morrot et al., 2001), and the presence of a strongly-associated colour can greatly enhance the detection of an odour and the intensity of aromas or flavours (Zellner and Kautz, 1990; Zellner and Whitten, 1999), as well as preference and enjoyment (Herz, 2001; Herz and Beland, 2004). Such associations between colour and odour appear to be based on prior experience (Blackwell, 1995; Morrot et al., 2001; Sakai et al., 2005; Stevenson and Oaten, 2008), and odours are usually perceived alongside visual, taste and tactile sensations, as well as higher order cues such as shape, size and object labelling. As such, an odour maybe perceived quite differently depending upon its current multisensory context, and experiencing an odour without these additional cues is likely to be different from experiencing the odour in a natural multisensory environment. Here we explore if odour-evocative words, rich in semantic connotations, differ in their colour associations compared to those associated with just odour. Twenty individuals were tested in each of four geographical locations; Germany, Malaysia, The Netherlands and the United States of America. Participants chose the three colours they most closely associated with both odours and odour-words from a chart of 36, using Xperiment software (www.xperiment.mobi). Preliminary results indicate that there were differences between odour-evocative words and odour cues in terms of the associated colours, for all populations.

Associations between colour and odour are likely culturally specific. Exposure to a new culture’s cuisine and food customs however may alter these associations. Here we test for cultural colour–odour association differences and whether exposure to a new culture impacts upon these associations. Participants were given 14 odours and choose from a chart of 36 randomly presented colours which of 3 colours they most associated with each odorant, and which of 3 colours was least associated with each. Data collection was done on Android and Apple iPod devices using Xperiment software (see www.xperiment.mobi). In the first study, we tested equal numbers of Dutch and Chinese participants living in the Netherlands (n = 20 per group). In the second study we tested equal numbers of Malay and Malaysian-Chinese participants living in Malaysia (n = 20 per group). Preliminary results indeed indicate that the cultures in both studies did differ from each other in terms of colour–odour pairings. Further analyses are ongoing.