Cross-cultural color-odor associations

Abstract

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.

Figure 1. Color choices.

36 colors were arranged in random order for each trial.

Figure 2. Color congruency for each odor in each culture.

Colors per odorant per country are ordered by frequency (most frequent are shown lowest on their respective y-axis). Frequency is represented by the height of each color box; the box on the right of the figure shows the height a given box must be for there to be 10, 9, 8 etc. ratings of that color for a given odorant. Boxes have been given a slight shadow to help with the perception of harder to see light colors. The background bars are only colored so as to help with reading the figure.

Figure 3. Odor representation by culture.

One Representational Dissimilarity Matrix (RDM) for each of the six populations sampled. Both axes of each matrix represent the 14 odors. Each cell in the matrix indicates the degree of dissimilarity between the color-patterns of the respective odors in that row and column. Warmer colors indicate higher dissimilarity. The dark blue diagonal indicates the perfect similarity of the odors with themselves. The representational geometry, or the spatial configuration of clusters of high and low dissimilarities, shows differences and commonalities in each culture. Along the color legend, a line graph shows a bootstrapped null distribution of pattern dissimilarities.

Figure 4. Crosscultural dissimilarity of representation.

Representational Dissimilarity Matrix comparing cultures. Both axes represent the six cultures. Each cell in the matrix indicates the degree of dissimilarity between the respective cultures' odor representation geometry (Figure 3). The dark blue diagonal indicates the perfect similarity of the cultures with themselves.