One- or two-step? New insights into two-step hypothesis and rainbow-like theory for pitch class-color synesthesia

Abstract

Introduction: This study investigates the mechanisms underlying pitch class-color synesthesia, a cognitive trait in which musical pitches evoke color perceptions. Synesthesia in music particularly involves the association of pitch classes (e.g., do, re, and mi) with specific colors. A previous study introduced the two-step hypothesis, which suggests that pitch class identification precedes color association, and proposed a rainbow-like theory based on color gradients selected by synesthetes. The primary objective is to retest these theories to evaluate their generalizability in explaining pitch class-color synesthesia.

Methods: We employed a dual-task paradigm to assess the robustness of the two-step hypothesis and conducted qualitative interviews to explore the nature of synesthetic experiences.

Results: The results indicated that the two-step hypothesis may not be always applicable, because it effectively accounts for only a subset of synesthetes. The presence of one-step synesthetes, who experience direct pitch-to-color associations without intermediate steps, implied a more varied synesthetic mechanism. Moreover, rainbow-like theory predominantly characterized two-step synesthetes, while one-step synesthetes exhibited distinct color perceptions. Furthermore, we found that the differentiation between two- and one-step synesthesia may be associated with the methods through which participants develop synesthetic associations.

Discussion: The findings highlighted the diversity of synesthetic experiences in pitch class-color synesthesia, which challenges the generalizability of the current theories and poses the need for a further nuanced understanding of this phenomenon.

Keywords: absolute pitch; dual task; pitch class–color synesthesia; rainbow-like theory; synesthesia.

FIGURE 1

Results of the color-selection test in Itoh et al. (2017). Colors selected by individual participants are shown as circles arranged in rows and the averaged colors are shown with square. Figure reproduced from Figure 2 in Itoh et al. (2017).

FIGURE 2

Two-step hypothesis (Itoh et al., 2017). Musical pitches are first identified by their respective pitch class names, and these names are then associated with specific colors. This figure was created by the authors based on Figure 1 in Itoh and Nakada (2018).

FIGURE 3

The word on the left is a word with a closed area, while that on the right is a word without a closed area.

FIGURE 4

Example of a single trial. The participants first saw the “+” symbol for 1 s, then made judgment within 3 s, and, finally, heard a 1-s white noise.

FIGURE 5

(A) RTs under the solo condition were significantly shorter than each of those under the two dual-task conditions, but the study found no significant difference between the visual and verbal interference conditions. (B) Result of paired comparisons. Regardless of the type of response, the study found no difference between the visual and verbal interference conditions. In both panels, error bars indicate standard errors. ***p < 0.001.

FIGURE 6

Colors selected by each participant are shown as circles, and the averaged colors are depicted with squares. The last column illustrates the average results of the color selection test.

FIGURE 7

Numerical representation of the colors in Figure 6, plotted for hue value of color in the HSV space. Each gray line represents an individual case, and the red line represents the average hue. (A) Colors for all sixteen participants. (B) Colors for two-step group. (C) Colors for one-step group.

FIGURE 8

(A) RTs for the two-step group. A significant main effect was found for response. Paired comparison indicates that the color task took significantly more time than the pitch task, and the difference in the average RTs between the pitch and color tasks is greater for the participants in the two-step groups than for the 16 participants under the solo and visual interference conditions. (B) RTs for the two-step group. A significant main effect was found for interference. RTs under the solo task condition were significantly shorter than each of those under the two dual-task conditions. *p < 0.05, ***p < 0.001.

FIGURE 9

The top panel illustrates the average colors for the one-step group, while the bottom depicts those for the two-step group.

FIGURE 10

RTs for the one-step group. No difference was found between the color and pitch tasks. RT under the solo task condition was significantly shorter than each of those under the two dual-task conditions (visual and verbal interference conditions). RTs under the verbal interference condition were significantly shorter than those under the visual interference condition. **p < 0.01, ***p < 0.001.