Small is bright and big is dark in synaesthesia

In synaesthesia, certain perceptual or conceptual stimuli (called inducers), trigger an additional concurrent experience. For example, I.S., a digit–colour synaesthete, experiences the colour green whenever he sees the digit 7. Since Galton’s seminal report on synaesthesia [1], it has been a commonly held view that digit–colour synaesthesia is highly idiosyncratic: that is, the same inducer, for example, the digit 7, will evoke different experiences in different synaesthetes. Moreover, the assumption that inducerconcurrent relationships are random is rarely questioned [2] and is based mainly on comparing the salient components of the inducer and the resulting synaesthetic perception. In the case of digit–colour synaesthesia, for example, the name of the colour is compared with the name of the digit. Little or no attention has been paid to other components of the colour or digit, such as luminance, saturation, ordinality or cardinality, which are neither explicit nor cognitively penetrable to the synaesthete. Here we report evidence of a systematic organisation relating luminance and number magnitude in digit– colour synaesthesia. We found that this organisation is based on cardinality rather than ordinality and follows the WeberFechner law, which has been reported previously for numerical representation in humans and monkeys [3]. Our results challenge the underlying assumptions about the mechanisms underlying synaesthesia and its developmental trajectories, and the link between luminance level and numerical magnitude strongly supports the idea of a shared magnitude representation [4]. We analysed triggered colour perceptions of nineteen digit– colour synaesthetes according to hue, saturation and luminance (see Supplemental data available on-line with this issue). A regression analysis showed that the magnitude of an inducing digit can be predicted by the luminance of the synaesthetic experience, and this relationship explained up to 68% of the variance (F(1,9) = 19.89, p = 0.001). This was not the case either with hue (p = 0.86) or saturation (p = 0.56, Figure S1 in the Supplemental data). Moreover, as with numerical representation in humans and monkeys [3], a logarithmic function yielded a better fit to the data than the linear function and explained 77% of the variance (F(1,9) = 30.64, p = 0.0005, Figure 1). In addition, when we examined consistency among individuals, 89% of the synaesthetes (n = 17) showed a negative trend between luminance and numerical magnitude (t(18) = –5.37, p = 0.00004, two-tailed). It is important to establish whether the correspondence between number and luminance is due to an ordinal (place in a sequence) or cardinal (magnitude) representation of numbers. To address this, we analysed the data from eight of the nineteen digit–colour synaesthetes, who also exhibited a coexistence of day-colour and digit-colour synaesthesia. The names of the days which triggered the colour experience were in Hebrew. Importantly, in the Hebrew language the days have a prominent ordinal nature, because they are named in a purely ordinal manner: ‘Sunday’ is called ‘First-day’, ‘Monday’ is called ‘Second-day’, and so forth, with the exception of ‘Shabatday’ for ‘Saturday. This special characteristic of the Hebrew language allowed us to examine the ordinality/cardinality question with the same participants. Neither