Biological origins of color categorization

Abstract

The biological basis of the commonality in color lexicons across languages has been hotly debated for decades. Prior evidence that infants categorize color could provide support for the hypothesis that color categorization systems are not purely constructed by communication and culture. Here, we investigate the relationship between infants' categorization of color and the commonality across color lexicons, and the potential biological origin of infant color categories. We systematically mapped infants' categorical recognition memory for hue onto a stimulus array used previously to document the color lexicons of 110 nonindustrialized languages. Following familiarization to a given hue, infants' response to a novel hue indicated that their recognition memory parses the hue continuum into red, yellow, green, blue, and purple categories. Infants' categorical distinctions aligned with common distinctions in color lexicons and are organized around hues that are commonly central to lexical categories across languages. The boundaries between infants' categorical distinctions also aligned, relative to the adaptation point, with the cardinal axes that describe the early stages of color representation in retinogeniculate pathways, indicating that infant color categorization may be partly organized by biological mechanisms of color vision. The findings suggest that color categorization in language and thought is partially biologically constrained and have implications for broader debate on how biology, culture, and communication interact in human cognition.

Keywords: categorization; color lexicons; color perception; infant; vision.

Fig. S1.

The average percentage of targets fixated (left side of figure) for the four pairs tested in the target-detection task, with novelty preferences from the main experiment for comparison (right side of figure). The dashed line in each panel indicates chance performance (also indicated by the guess rate in the target fixation panel). Asterisks indicate pairs for which the evidence was in favor of H1 compared with chance.

Fig. S2.

(A) Infant looking time during the familiarization trials averaged across hues (±1 SE). (B) Novelty preference percentages (±1 SE) for pairs of adjacently sampled hues (pairs on left of vertical dashed line in figure) and larger hue pairs which straddle smaller pairs (pairs on right of vertical dashed line in figure). Asterisks indicate novelty preferences that produce a Bayes factor in support of H1 relative to H0 (B > 3), with equal looking at novel and familiar hues indicated by the dashed line at 50%. Munsell codes of hues of each pair are given.

Fig. 1.

Infant color categorization and the relationship to lexical color categorization. (A) Novelty preferences suggest infant recognition memory distinguishes red–yellow, green–yellow, blue–green, and blue–purple hues but not hues within these categories. Sampled stimuli are outlined in black, horizontal lines joining stimuli indicate color pairs that did not elicit novelty preference, gaps indicate novelty preference. Numbers are from the WCS stimulus grid and indicate the different hues. (B) Color naming systems for row G stimuli based on: WCS data for Wobé (3 BCTs), Jicaque (5 BCTs), and Huave (7 BCTs); English color naming (11 BCTs) (42); and the GBPm for row G stimuli (12). Vertical thick black lines indicate category boundaries between stimuli given the same most frequent term within a language. Correspondence between the category boundaries in language and infant novelty preferences can be seen. (C) Frequency of category centroids from the WCS for each hue in row G (8). The gaps in the thick black horizontal bars at the bottom of C indicate hues which were straddled by color pairs which elicited novelty preference, as also shown by the gaps in the black horizontal lines in A. Category centroid frequencies tend to peak in regions which are not distinguished by infant recognition memory and are generally lowest in regions which are distinguished.

Fig. S3.

The full WCS stimulus grid and centroid analysis. (A) WCS stimulus grid for 40 hues at 8 lightness levels. Stimuli in the current experiment were sampled from row G. (B) Centroid analysis from Kay and Regier (8). The black and gray contour lines indicate the number of speaker centroids falling at that point in the stimulus grid from WCS data, with the outermost contour representing 100 centroids and each subsequent inner contour representing an increment in 100 centroids. Infant novelty preferences are indicated by the gaps in the thick black horizontal line at row G.

Fig. 2.

Stimuli plotted a version of the MacLeod–Boynton chromaticity diagram with L/(L+M) and (S/L+M) cardinal axes of color vision that correspond to the retinogeniculate pathways. The dashed vertical and horizontal lines indicate the background (Munsell N5) to which infants were adapted. The Munsell hue codes for stimuli are given and black lines connecting stimuli indicate no novelty preference for that pair. The cross between 7.5R and 5R indicates a pair that was not tested.

Fig. S4.

Stimuli plotted in CIELAB perceptual color space (a*,b*). Stimulus pairs for which there was no novelty preference are indicated with black lines joining the stimuli, and pairs where there was a novelty preference are indicated by the absence of these lines. The cross between 7.5R and 5R indicates a pair that was not tested. Euclidean distances in this space do not predict infants’ novelty preference.