The discussion on the basicness of taste at its different levels – perceptual, neurophysiological, psychophysical, linguistic, and cultural – is in many ways reminiscent of the discussion on colour categories. Colour perception is perhaps better understood and less controversial than taste perception. It is accepted that our retina has four types of photoreceptors, three of which are involved in chromatic perception, and the neural coding and transmission of colour perception is well understood (e.g., Gegenfurtner & Sharpe Reference Gegenfurtner and Sharpe1999). However, the cognitive aspects of colour perception and the nature of colour categories in particular have been under debate for more than 50 years (for an overview, see Hardin & Maffi Reference Hardin and Maffi1997).

Colour categories cut up a continuous chromatic experience into concepts that can be associated with linguistic terms. Just as with basic tastes, some colours are considered to be primitive (white, black, red, green, blue, and yellow) and all other colour experiences can be described in terms of these primitive colours. These colours have been identified as being psychologically opponent (Hering Reference Hering1964) and neural coding for these opponent channels has been identified in the brain (De Valois et al. Reference De Valois, Abramov and Jacobs1966). This precludes other colours from serving as primitive colours (but see Jameson & D'Andrade Reference Jameson, D'Andrade, Hardin and Maffi1997).

When considering linguistic colour category systems across cultures, one could at first be tempted by the rather varying colour lexica and their referents across cultures to conclude that the colour continuum is arbitrarily cut up by linguistic categories (e.g., Gleason Reference Gleason1961). This Whorfian view was challenged by Berlin and Kay (Reference Berlin and Kay1969) who demonstrated how a set of different languages spoken in different cultures have basic colour terms of which the referents are remarkably similar. This was reconfirmed almost 30 years later in a statistical study using data from a large set of languages from non-industrialised cultures (Kay & Regier Reference Kay and Regier2003; Lindsey & Brown Reference Lindsey and Brown2006; Regier et al. Reference Regier, Kay and Cook2005). This universalist view holds that these regularities in referents of colour terms result mainly from regularities in the neural coding of colour; and as the neural coding is largely genetically determined, so are colour categories. Nevertheless, there is considerable evidence that colour categories are plastic, that they are learned, and that they change as a result of learning experiences. These learning experiences typically involve language, and, as such, colour cognition is brought into the realm of linguistic relativism. In a series of recent reaction time experiments (Drivonikou et al. Reference Drivonikou, Kay, Regier, Ivry, Gilbert, Franklin and Davies2007; Gilbert et al. Reference Gilbert, Regier, Kay and Ivry2006; Roberson & Hanley Reference Roberson and Hanley2007) subjects were asked to find the odd-one-out in a set of otherwise identical colours. The odd-one-out belongs either to the same linguistic colour category of the distractors (i.e., blue or green) or falls just outside the category; the perceptual distance was, of course, kept identical. When the target appears in the right visual field, subjects were faster at spotting the odd-one-out when it belongs to a different category. This is not observed when the target belongs to the same category, when the target is presented in the right visual field, or when the subjects are distracted with a linguistic task. As language is situated in the left hemisphere, information in the right visual field travelling to the visual cortex via the left hemisphere is more under the influence of lexical representations. This suggests that language, and lexical representations in specific, have an impact on natural categorisation.

These experiments, however, do not inform us on how natural categories are formed; for this, insights can be gained from computer simulations (Belpaeme & Bleys Reference Belpaeme and Bleys2005; Dowman Reference Dowman2006; Jameson Reference Jameson2007; Lammens Reference Lammens1994; Puglisi et al. Reference Puglisi, Baronchelli and Loreto2007; Steels & Belpaeme Reference Steels and Belpaeme2005). These studies typically involve modelling a large population of individuals interacting with each other and trying to reach a consensus on colour terms. As colour terms refer to colour categories, the process of arriving at a linguistic consensus on colour terms affects the categorisation of colour. The interest of this type of simulations lies in the fact that the evolution of colour typology can be studied diachronically and under a varying set of ecological, perceptual, and social conditions.

Simulations in which human-like colour category systems emerge from linguistic interactions about colour are presented in Steels and Belpaeme (Reference Steels and Belpaeme2005) and Belpaeme and Bleys (Reference Belpaeme and Bleys2005). In these simulations colour categories are situated in a semantic colour space, which is implemented using a three-dimensional CIE L*a*b* colour appearance model (Fairchild Reference Fairchild1998). The colour space models colour perception numerically, has an opponent character, and, importantly, is not symmetrical. The colour space thus forms a bias on colour category acquisition, which, together with the linguistic interactions in a situated environment, nudges colour categories towards certain locations in the colour space. It is important to note that language seems to play an important role in binding colour categories of individuals together. Without language, individuals will be able to learn a category for, say, reddish hues, but the categories would be insufficiently coordinated to allow communication about “red.” Lexical terms thus not only serve as a conduit for learning categories, but also serve to coordinate categories between individuals.

In light of what we know about colour categories, it is very likely that the basic tastes are indeed, as Erickson seems to suggest, the product of a cultural agreement made on top of an innately fixed psychophysical substrate.