Perceptual and categorical judgements of colour similarity

Abstract

Consideration is given to the tasks that make judgements of colour similarity based on perceptual similarity rather than categorical similarity. Irrespective of whether colour categories are taken to be universal (Berlin & Kay, 1969) or language induced (Davidoff, Davies, & Roberson, 1999), it is widely assumed that colour boundaries, and hence categorical similarity, would be used when categorising colours. However, we argue that categorical similarity is more reliably used in implicit than in explicit categorisation. Thus, in Experiment 1, we found that category boundaries may be overridden in the explicit task of matching-to-sample: There was a similar strong tendency to ignore colour boundaries and to divide the range of coloured stimuli into two equal groups in both Westerners and in a remote population (Himba). In Experiment 2, we showed that a distinctive stimulus (focal colour) in the range affected the equal division in a matching-to-sample task. Experiment 3 tested the stability of a category boundary in an implicit task (visual search) that assessed categorical perception; only for this task was categorisation largely immune to range effects and largely based on categorical similarity. It is concluded that, even after colour categories are acquired, perceptual rather than categorical similarity is commonly used in judgements of colour similarity.

Keywords:

• Colour similarity
• Categorical similarity
• Perceptual similarity
• Colour naming

Acknowledgments

This work was supported by EC 6th Framework grant 012984 Stages in the Evolution and Development of Sign Use (SEDSU). JG was supported by an ESRC studentship. We would like to acknowledge Alan D. Pickering for his advice on the decision bound modelling. We are grateful to K. Jakurama for helping in the field.

Notes

¹It has been argued that infants show colour CP (e.g., Franklin & Davies, 2004) and, in that case, language would not be a prerequisite. The arguments pro and con infant colour CP are given in Davidoff and Fagot (2010). However, as it is agreed that these infant colour categories are not necessarily present in the adult, they are not essential for consideration of the present data.

²Brown et al. (2011), Fonteneau and Davidoff (2007), Webster and Kay (2011), and Witzel and Gegenfurtner (2011) were unable to replicate the many findings (e.g., Gilbert et al., 2006) that show CP for green compared to blue only for stimuli presented to the language hemisphere. They all found the advantage across the boundary for both visual fields. Brown et al. argue for a physiological account for categorical similarity at that boundary.

³Motivation for Experiment 3 came in part from a talk given by Dr. Oliver Wright at the Progress in Colour Studies 2008 conference held in Glasgow, UK (Wright, 2011). He noted that the midpoint of the range of blue/green colours used in the visual search study of Gilbert et al. (2006) did not correspond precisely to the usual boundary given for that colour range.