A comparison of lexical-gustatory and grapheme-colour synaesthesia

Abstract

This study compares two different profiles of synaesthesia. One group (N = 7) experiences synaesthetic colour and the other (N = 7) experiences taste. Both groups are significantly more consistent over time than control subjects asked to generate analogous associations. For the colour synaesthetes, almost every word elicits a colour photism and there are systematic relationships between the colours generated by words and those generated by graphemes within the word (hence “grapheme-colour” synaesthesia). For the taste synaesthetes, by contrast, some words elicit no synaesthesia at all, and in those words that do, there is no relationship between the taste attributed to the word and the taste attributed to component graphemes. Word frequency and lexicality (word vs. nonword) appear to be critical in determining the presence of synaesthesia in this group (hence “lexical-gustatory” synaesthesia). Moreover, there are strong phonological links (e.g., cinema tastes of “cinnamon rolls”) suggesting that the synaesthetic associations have been influenced by vocabulary knowledge from the semantic category of food. It is argued that different cognitive mechanisms are responsible for the synaesthesia in each group, which may reflect, at least in part, the different geographical locations of the affected perceptual centres in the brain.

Acknowledgments

This research was supported by the second author’s British Academy Postdoctoral Fellowship (PDF/2001/345) and British Academy research grant (SG-33379), and by a Vacation Scholarship from the Wellcome Trust to the third author.

Notes

Half the control subjects in Ward and Simner (2003) were asked to generate associations that would be easy to remember (e.g., fair → “candy floss”) and were given an additional study period, as well as a monetary incentive, to perform well in the recall taste. The other half were simply asked to free associate and were not warned that their memory would be tested. As both groups performed significantly worse than JIW in the Ward and Simner study we have combined them for this investigation.

Nonparametric tests are carried out because the distribution is heavily skewed towards ceiling for the colour synaesthetes.

Note that nonwords could be considered as words with a frequency count of zero. In this way, one could argue that the lexicality effect might be subsumed within the phenomenon of word frequency.

The grapheme-colour synaesthetes were given 10 colour names. All stimuli were noted to produce a synesthetic colour, but the extent of the alien colour effect was variable, as has been noted before (Gray et al., 2002). The percentage of words eliciting an alien colour were: KZ = 90%, LB = 70%, SG = 50%, DSL = 40%, SJT = 0%, KW = 0%. KA was unavailable for testing on this occasion.