In this target article the following hypotheses are discussed: (1) Colour is autonomous: a perceptuolinguistic and behavioural universal. (2) It is completely described by three independent attributes: hue, brightness, and saturation: (3) Phenomenologically and psychophysically there are four unique hues: red, green, blue, and yellow; (4) The unique hues are underpinned by two opponent psychophysical and/or neuronal channels: red/green, blue/yellow. The relevant literature is reviewed. We conclude: (i) Psychophysics and neurophysiology fail to set nontrivial constraints on colour categorization. (ii) Linguistic evidence provides no grounds for the universality of basic colour categories. (iii) Neither the opponent hues red/green, blue/yellow nor hue, brightness, and saturation are intrinsic to a universal concept of colour. (iv) Colour is not autonomous.

No crucial experiment demonstrates that four hue categories are needed to describe color appearance. Instead, converging lines of evidence suggest that the terms red, yellow, green, and blue are sufficient and precise enough for deriving color discrimination functions and for a useful model constraining relations between color appearance and neuronal responses. Such a model need not be based on linguistic universals. Until something better is available, this “standard model” holds.

The physics of color and the psychology of color naming are not isomorphic. Physically, the spectrum is continuous with regard to wavelength – one point in the spectrum differs from another only by the amount of wavelength difference. Psychologically, hue is categorical – colors change qualitatively from one wavelength region to another. The psychological characterization of hue that characterizes color vision has been revealed in a series of modern psychophysical studies with human adults and infants and with various infrahuman species, including vertebrates and invertebrates. These biopsychological data supplant an older psycholinguistic and anthropological literature that posited that language and culture alone influence perceptual processes; language and culture may modify color naming beyond basic categorizations.

We argue that the confusing pattern of evidence concerning colour categorization reported by Saunders & van Brakel is unsurprising. On a perspectival view, categorization may follow semantic or pragmatic attributes. Colour lacks clear semantic attributes; as a result categorization is necessarily pragmatic and context-sensitive. This view of colour categorization helps explain the developmental delay in colour naming.

This commentary expresses basic agreement with Saunders & van Brakel, gives explanations for some of the frailties they see in color science, and suggests that the conditioning forces of modern technology may render color categories increasingly useful even if not initially “real.”

Saunders and van Brakel question whether the special status of red, green, yellow, and blue in our perceptual organization is anything more than a shadow cast by the English language. I suggest that it is more than this. We can hardly imagine treating lime, purple, orange, and teal as unique hues, and the reason does not lie in special training. To settle the issue, I suggest some lines for psychological experiment and anthropological investigation.

At the close of their searching critique, Saunders & van Brakel raise, but do not address, the question: “Whither colour research?” There are two distinct traditions of colour research, one based on disembodied coloured lights and another on surface colour. The coherence and integrity of both these traditions are challenged by the nonautonomy of colour.

Saunders & van Brakel argue, inter alia, that there is “little evidence” for the claim that there are four unique hues (red, green, blue, and yellow), and that there are two corresponding opponent processes. We argue that this is quite mistaken.

The question of whether colour categorisation is determined by nontrivial constraints (i.e., universal neurophysiological properties of visual neurons) is an empirical issue concerning the organisation of the internal colour space. Rosch has provided psychological evidence that categories are organised around focal colours and that the organisation is universal; this commentary reconsiders that evidence.

We acknowledge that empirical support for universal colour categories in colour cognition is insufficient: it relies too heavily on Rosch-Heider's work with the Dani. We offer new evidence supporting universal perceptual-cognitive colour categories. The same data also support language modulating colour-cognition: Universal structures are fine-tuned by language.

We need to reconsider and reconceive the path that will take us from innate perceptual saliencies to basic (and perhaps other) colour language. There is a space between the perceptual and the linguistic levels that needs to be filled by an account of the rules that people use to generate relatively stable reference classes in a social context.

The following provides further arguments for the nonuniversality of color as an autonomous dimension. Research on odors suggests that there are cultural constraints on the abstraction of dimensions for objects. Color vision analysis leads to an overemphasis on the role of perceptual processes in categorization. The study of odors points to human activities as a more important principle of categorization that drives the perceptual processing and suggests a reconsideration of vision itself.

Despite the heat of their attack, Saunders & van Brakel do illuminate various shortcomings of color research in the tradition of Berlin & Kay. Berlin and Kay elicit a pan-cultural pattern in color language, but the pattern does not provide much insight into the human mind.

In adult humans, conscious visual experience – including that of colour – is shaped by particular cultural practices, as evidenced in the cross-cultural literature. In addition, the practices of our own culture already inform attempts to assess the “natural” experience of newborns or other animals.

If, as Saunders & van Brakel assert, hue, lightness, and saturation characterize artificial color spaces and not the colors of everyday life, one would expect those color spaces to have limited relevance to our understanding of color phenomena and to be of little practical application. This is not the case. Although people perceive these and equivalent color dimensions holistically rather than analytically, they are able to use such triples to categorize the colors of their environment.

This commentary starts with a simplified Cartesian vector space of the tristimulus theory of color. This vector space is then further simplified so that bitstrings are used to represent the vector space. The Commission Internationale de l'Eclairage (CIE) diagram is shown to follow directly and simply from this vector space. The Berlin & Kay results are shown to agree quite well with the vector space and the two-dimensional version of it, especially if the dimensions are normalized to take into account the sensitivity of the eye to the different wavelengths comprising color. There is asymmetry with respect to the colors and a similar asymmetry in vocalic phonemes; these effects can be explained in terms of physiology. The (in)famous problem of the color channels is given a unified treatment at various levels. An eight-valued color algebra is created, with addition and multiplication corresponding to additive and subtractive blending of colors. Finally, it is shown that the discrete Hamming metric for colors has a natural toroidal topology.

Zone theories of color vision transform cone sensitivities to channel sensitivities before transmitting these signals to the brain. The concepts of “unique hue” and “color opponency” are fundamental to an understanding of this transformation. Saunders & van Brakel question the objectivity of these concepts. Statements in their target article indicate that the reason for this questioning stems from a failure to appreciate the constraints inherent in the definitions of these concepts.

Saunders & van Brakel set out to review color science research and to topple the belief that color-vision neurophysiology sets strong deterministic constraints on the cognitive processing of color. Although their skeptism and mission are worthwhile, they fail to give proper treatment to (1) findings that dramatically support some positions they aim to tear down, (2) existing research that anticipates criticisms presented in their target article, and (3) the progress made in the area toward understanding the phenomenon. At the very minimum these oversights weaken the credibility of their arguments and leave the reader to wonder why their discussion ignores what is clearly omitted.

Saunders & van Brakel's claim that Berlin and Kay (1969) assumed a language/vision correlation in the area of color categorization and disguised this assumption as a finding is shown to be false. The methodology of the World Color Survey, now nearing completion, is discussed and the possibility of an additional language/vision correlation in color categorization is suggested.

There is no natural and pretheoretical classification of colour appearances into hue, saturation, brightness, unique hues, and so on. Rather, our theoretical insights into the coding of colour have reciprocally shaped the way we talk about colour appearances. Opponency is only one of many fundamental aspects of colour coding, and we are hardly justified in ascribing some theoretical prominance to it.