7
Color manipulation and comparative color

They’re not all compatible

Derek H. Brown

Introduction

Comparative color vision has had a rich, positive impact on basic issues in the philosophy of color and philosophy of perception.1 Most centrally, it has demonstrated an unexpected variety of color visual architecture and of uses to which color vision has and can be put by various animals. Architectural differences include differences in the numbers of cones, cone sensitivities (differences within the same range of the electromagnetic spectrum and by virtue of extending to different ranges), the processing of cone outputs, the presence and absence of oil droplets, and so on. The uses of color vision across species have some broad, uninformative commonalities (color vision helps creatures see) and differences that contain important lessons about both humans and nonhumans (see below). This knowledge has forced us to: more broadly conceive of what the function of color vision is within an organism or species; recognize substantive differences in color perceptions across species, fundamentally broader differences than we are forced to recognize within humanity; speculatively conclude that various nonhuman animals have categorically different color experiences from our own; and push color ontology, as many phenomena do, away from any simple form of color objectivism.

In this brief chapter, I must leave to one side many fascinating questions that have been studied in this area over the last twenty-five years. To facilitate this, I will make some simplifying assumptions.

Do nonhuman animals perceive colors at all? I suppose that at least some of them do. Justification for this stems from numerous factors, including the existence of visual systems that are not only wavelength sensitive (i.e., have numerous cones) but also exhibit opponent processing and color constancy. Those who wish to emphasize the importance of consciousness for color perception may wish to examine the animals with these kinds of visual systems to see which ones otherwise satisfy their preferred account of consciousness, and conclude that only these ones see colors. I am not opposed to this, but the issue will not substantively impact this work. For ease of discussion, I will be liberal and suppose that many of the species with these kinds of visual systems do have consciousness and see colors.

Are animal colors like human colors? Fortunately, this is treated in Mohan Matthen’s Chapter 6 in this volume. As a simplifying assumption, I will suppose that Matthen is right that bird colors are distinct from human colors. More generally, I presume:

Human colors are different from and do not overlap with pigeon colors, and so on for different pairs of species. The reader is invited to consult Matthen’s chapter for a supporting argument.

Although Disunity takes no stance on the reality or instantiation of color, it raises difficult questions about it. One easy means of appreciating this is by appeal to the supposition that color experiences are of or about colors. Perceptual relationalists (/acquaintance theorists) presume, minimally, that the colors experiences are about are instantiated in our world. Representationalists presume, minimally, that the colors veridical experiences are about are instantiated in our world. I will work with the latter framework, though what follows is generalizable to the former.

Supposing Disunity obtains, is one of these color spaces uniquely correct of our world? More generally, are some spaces more correct than others? It is exceedingly difficult to single out a preferred space, in large part because it is difficult to single out a preferred color visual system. Different species are sensitive to different wavelength information and in different ways, and use it for different purposes (e.g., to learn different things about the world, to variously interact with it). If there is some unique color space of our reality, comparative color vision does a good job of hiding it from us. To take one fascinating class of examples (Akins and Hahn 2014: 154), a functioning visual system for underwater creatures helps them survive (eat and not be eaten) in their normal, water-laden environments. This means that the system can adequately interpret the light arrays coming through water in front of them. But water – even more so than air – acts as a kind of filter, and thus the wavelengths of light passing through shallower depths is more varied than those passing through deeper depths (where only the intermediate “blue” wavelengths are left). Interestingly, color visual systems are often suitably varied depending on whether the creature typically lives in shallower or deeper waters. Relatedly, creatures who typically look down into the ocean look into a dark background. To facilitate the perception of other life, their color visual systems often systematically normalize or “ignore” darkness and enhance lightness, thus increasing their sensitivity to food and prey. By contrast, creatures who typically look up look into a bright background. Their systems often normalize or “ignore” lightness and enhance darkness, so as to be sensitive to the darkened shadows things above them project downward.

The idea that one color visual system is uniquely correct, or even that one is more correct than another, stands in tension with the kinds of lessons we’ve learned from comparative studies. There is much to discuss here, but for simplicity I assume:

Finally, given that my interest is primarily in what we can learn from cross-species studies, I will also for simplicity suppose a known falsity:

Philosophers of color have emphasized the importance of “normal” systemic variations in human color perception (the plethora of work on unique hue variations is a great source for this; see Tye 2006 and various Analysis articles that followed), and new ground has recently been broken on our understanding of less “normal” variations like color blindness (Broackes 2010). Working these topics into what follows would be a distraction. By contrast, nonsystemic, local variations in color experience within a species cannot be avoided without risking incoherence: different individuals often view things from different locations, and in different conditions, and these factors can impact experienced and perhaps perceived color. These variations are assumed, though they will not play a substantive role in this discussion. What is excluded are variations that remain when members of the same species view the same thing at the same time in the same condition.

1 Pluralism and selectionism

Assume that one species’ color experiences are generally veridical. Given Ecumenicism, this entails that others’ are too, and given Disunity, it follows that there are numerous veridical color experiences from different color spaces. As above, I presume that a minimal condition on a color experience being veridical is that the color the experience is of or about is instantiated in our world. We now have:

While Pluralism places a significant constraint on color ontology, it nonetheless leaves many critical aspects of it underdetermined. For example, assuming that minds are in our world, Pluralism is mute with regard to the mind-(in)dependence of color. For example, color mentalists believe that colors are properties instantiated by minds (e.g., properties of sense-data, qualia, or neural states). Pluralism can be accommodated by holding that the minds of different species instantiate colors from different spaces. Assume that the minds of different species, while located in our universe, occupy distinct, non-overlapping, spatiotemporal regions (e.g., a human mind is never in the same place at the same time as a bird mind, and vice versa). It follows that, while Pluralism obtains, no part of the world instantiates more than one color (on a suitably restricted reading of “part”).4

My interest is in the application of these considerations to Color Objectivism, which asserts that colors are mind-independent properties. Various species visually experience overlapping parts of the world (e.g., humans and birds both see at least some of the same trees). Since by hypothesis these experiences are generally veridical with regard to color, accommodating Pluralism now requires that the same part of the world can simultaneously instantiate more than one color. Put succinctly, a uniform object can simultaneously instantiate different colors. Is this plausible?

The matter concerns our commitments to color (in)compatibilities. Familiar color incompatibilities include:

  1. (1) No uniform thing is simultaneously blue and green.

Our focus is on colors from distinct spaces, something like:

  1. (2) No uniform thing is simultaneously blue and shmeen.

One can generally reject color incompatibilities, in which case one rejects both (1) and (2). Alternatively, one might reject (2) while accepting (1). Given the above assumptions, there is merit in exploring this latter possibility: the assumptions do not conflict with (1); there are independent reasons for accepting (1); and prima facie one might be unsure how to approach (2) and thus remain open-minded about it.

Rejecting (2) while accepting (1) is a commitment of Pluralist Selectionism (hereafter Selectionism, its adherents being Selectionists). The view asserts Pluralism and Objectivism (hence rejects (2)), and holds that during color perception, a given species selects, from among the many colors that a sample uniform thing instantiates, colors from within the color space its vision system is attuned to (Kalderon 2007, Matthen 2005, Allen 2009; this is arguably a consequence of Byrne and Hilbert 2003). Thus, a uniform thing can simultaneously be both objectively blue and shmeen, but while human vision selects blue (resulting in us experiencing the thing’s blue color), pigeon vision selects shmeen (resulting in pigeons experiencing the thing’s shmeen color). A central goal of Selectionists is to nonetheless maintain (1), and they view the kind of incompatibility contained in (1) as not merely a feature of color experiences but as a feature of colors themselves. The aforementioned works achieve these aims by appeal to the determinable-determinate distinction. Colors within, say, human color space bear (in)compatibilities with one another by virtue of being determinates relative to the determinable human color space. But, it is argued, a given species’ color space is not merely determinable but super-determinable, meaning that colors from different spaces avoid being mutually incompatible by being fundamentally not comparable. They are “just different” (Allen 2009: 208). The extent of non-comparability between families is a total, categorical, commitment:

Given Non-comparability, one has seeming grounds to reject (2) but is not pressured to reject (1).

Non-comparability is an extreme position in regards to relations between different color spaces. An equally extreme position in the opposite direction is:

If Incompatibility is true, then (2) is true and there is a thorny relationship between Objectivism, Ecumenicism, and Disunity. I aim to defend an intermediary position, something between Non-comparability and Incompatibility. It is not clear to me that this is where we will end up, but I do believe that Selectionists cannot safely endorse Non-comparability, and a successful argument in favor of an intermediary position that still adheres to Objectivism secures this outcome.

2 Pure and applied Non-comparability

Debates about color (in)compatibilities are notoriously tricky. Evidential sources include experience, language, sensory physiology, optics, and so on. Arguments can appeal to color theory (e.g., opponent-process theory), abstract metaphysics (e.g., the determinable-determinate relation), analogies with relevant cases (e.g., between geometric and color spaces), et cetera. Each of these has at least some value and should contribute to a considered view. I must focus the discussion.

One thing that strikes me about using the determinable-determinate relation to defend Non-comparability is the abstractness of the reasoning. In my judgment, we can use this relation to argue that there is no incoherence in Non-comparability: each color space can be viewed as an internally consistent, self-contained “system of colors” that in principle does not invade or conflict with other systems. This is an argument for an abstract or pure kind of non-comparability. But Selectionists need something stronger. They defend a claim about concrete or applied non-comparability that, for example, a uniform thing in our world can simultaneously be blue and shmeen. This is a very different idea. Consider an analogy with geometric spaces (i.e., Euclidean and non-Euclidean spaces).

That there are numerous (in)compatibilities within a geometric space is adequately clear, as is the idea that each geometric space is self-contained and thus doesn’t place constraints on other spaces. They are non-comparable in this rough sense.6 The difference I wish to highlight is between pure and applied geometry, for even if we exclude (in)compatibilities between geometries in the abstract, it does not mean that they don’t obtain in concrete realities. Indeed, one of the great achievements of Einstein’s General Theory of Relativity was an argument to the effect that the universe has or obeys the rules of a fairly specific non-Euclidean geometric space. The other geometric spaces, though independent from this one in the abstract, are nonetheless inconsistent with this one when applied to our world.

This interpretation of general relativity is no doubt debatable, and the analogy between pure and applied geometric spaces and pure and applied color spaces deserves extended treatment. Regardless, the lesson is simple: non-comparability in the abstract need not translate into non-comparability in the concrete. In my judgment, the idea that there are distinct color spaces that are categorically not comparable (and hence avoid all incompatibilities), and motives for this idea drawn, for example, from the determinable-determinate distinction, are foremost about abstract color knowledge. When applied to concrete worlds, an entirely different set of considerations, notably empirical ones, come to the fore. In other words, the Selectionist argument for Non-comparability conflates two senses of “(in)compatibilities,” one articulable in the abstract domain of inquiry, and the other articulable in the concrete domain. The determinable-determinate distinction can ground conclusions about the former, but not about the latter – and conclusions about the latter are what Selectionists seek to draw.

Primary sources of empirical evidence regarding concrete (in)compatibilities, including Non-comparability, are color experience, sensory physiology, and optics. By virtue of endorsing Objectivism, Selectionists cannot explain concrete (in)compatibilities in terms of sensory physiology, for colors are, by definition, independent of perceiver peculiarities. They may use sensory physiology to help uncover the empirical ground of concrete (in)compatibilities, but the ground must ultimately be in objective features. This is a nontrivial constraint.

Human color experience can be used to ground concrete (in)compatibilities within human color space, but at present not across human and nonhuman spaces, and so also for the indirect knowledge we have of the color experiences of other species. In theory, optics on its own might suffice, but we are not in a position to develop a robust theory of various objective color spaces merely through the study of interactions between light and substance. We need to coordinate the likes of optics, color experience, and sensory physiology to uncover objective concrete (in)compatibilities, and then the relevant parts of optics (/objective reality) to help ground them. Only then can we assess the status of Non-comparability in the applied domain of Objectivism. I will briefly sketch such an approach, one drawn from color manipulation7 and designed to strictly work within Selectionist commitments. Even within these constraints, we should resist Non-comparability.

3 Color manipulation and Non-comparability

When we manipulate colors through objective processes such as painting, dying, and heating, we often witness color replacement, where the color at the end of the process is different from and has replaced the one at the beginning. For example, we turn a blue fabric into a green one via dyes and turn black into red coal via heat. Manipulations that do not result in replacement include ones that achieve color layering (both the original and new color are present, one before the other), failed attempts (e.g., the paint doesn’t stick), and so on. Regardless of the exceptions, it is clear that manipulation routinely results in replacement.

Manipulation is a well-tested and tangible source for knowledge about color (in)compatibilities. When we dye the blue fabric green, we witness a kind of incompatibility between the two colors. By witnessing many such manipulations, we induce that there are numerous color incompatibilities, and that roughly, colors from the same level of determinacy are all mutually incompatible, that is: categorical colors like blue, green, and orange are all mutually incompatible; fine-grained colors like royal blue, aqua blue, and sky blue are all mutually incompatible. Colors across levels of determinacy can be compatible: blue and aqua blue are mutually compatible.8

At first pass, this source of information is confined only to the human color space: via manipulation, we only directly see human colors turn into human colors, and hence our resulting beliefs about (in)compatibilities are limited to that family. One might argue that because of this, it is impermissible to infer anything about animal colors from manipulation. Such arguments should be resisted. Our beliefs about animal colors (e.g., Disunity) derive from a decent understanding of the workings of the color visual systems and environments of various species and not from any direct observation of animal colors. If that source of information is available to justify Disunity or Pluralism, then surely it is available to critique it.

Beliefs about human color (in)compatibilities drawn from manipulation don’t stem only from our color experiences, they also stem from robust knowledge of the impact manipulation has on the physical features of objects and how that is relevant to our color experiences and color perceptions more generally. This can be exploited to break into the debate over Non-comparability and Incompatibility precisely because the very same (or similar) physical features are equally (or similarly) relevant to the color perceptions of nonhumans. The result is that there are compatibilities and incompatibilities between human and animal colors within the broad Selectionist framework. Let me sketch the argument.

Manipulation changes human colors primarily because it changes the way objects reflect (transmit, produce9) light. The way objects reflect light is as critical an environmental factor for understanding animal color vision as it is for human color vision. If, like the Selectionist, one endorses Objectivism, then light reflectance is, if not definitive of color (e.g., Byrne and Hilbert 2003), at least something on which color supervenes (Allen 2009: 215). In saying that for Objectivism color supervenes on reflectance, I minimally mean that if one changes the color of something, then one changes its reflectance profile in some way, and that the reverse need not obtain, though often will. (Below, I consider an Objectivism that denies this supervenience.)

Suppose, with the Selectionist, that there is a uniform fabric that is blue and shmue, the former being the color human vision selects and the latter being the color pigeon vision selects. When we dye the fabric green, we change its reflectance properties, and as such there is every expectation that we might change it from shmue to (say) shmeen. Suppose we do not. Perhaps it is shmue because of a portion of its reflectance profile that was unchanged by the dye. Indeed, that portion may lie inside or outside the visible spectrum, since pigeon vision extends into both the ultraviolet and infrared ranges. In this case, we have prima facie evidence for a kind of compatibility between not only blue & shmue, but also between green & shmue. Suppose, for illustration, that all manipulations preserve the fabric’s shmue, that is, that no manipulation in human color affects the fabric’s shmue color: it is shmue no matter its human color. In this case, calling shmue “compatible” with blue, green, and the other human colors is strictly true, but seems too weak a conclusion. A more appropriate conclusion is that shmue is maximally compatible with, indeed independent from, human colors. This is the kind of outcome predicted by Non-comparability. Anything less would be an admission that shmue is incompatible with some human color, violating Non-comparability.

The problem with this picture should already be apparent. It stems from the implausibility of the idea that manipulating human colors will not impact animal ones, for the reverse will often, if not typically, obtain. Thus, suppose when we dye the blue fabric green, the fabric’s shmue is replaced with shmeen. We now have pima facie evidence of compatibilities between blue & shmue and between green & shmeen, and evidence of incompatibilities between blue & green, and between shmue & shmeen. And even this does not exhaust what we’ve learned, for we also have pima facie evidence of an incompatibility between blue & shmeen and between green & shmue. This is nothing like the idea behind Non-comparability.

A fuller picture of this argument can be detailed with actual reflectance values and illuminants, and actual color visual states induced in pigeons, humans, and other species. We have working knowledge of how different reflectances interact with the color visual systems of different species, and of what various manipulations do to those reflectances, so we can figure out how various manipulations will and will not impact the color perceptions of different species – at least in partial, rough detail. These specifics I leave to scientists working in these fields. I predict that a complicated picture will emerge: some manipulations will not impact the color perceptions of some species but will impact those of others; some manipulations will similarly impact the color perceptions of two species and differently impact the perceptions of two other species; and so on. All of this is to say that, given the tools we have, and a set of assumptions that are friendly to Selectionism, we can develop an empirically informed model of various compatibilities and incompatibilities between colors across spaces, pace Non-comparability.

In reply, one might deny that color manipulation is an evidential source for color (in)compatibilities across color spaces. This is a bumpy road. As stated above, color manipulation is a very familiar and tangible source for knowledge about color (in)compatibilities within human color space, and we can work out (at least roughly) how manipulations will impact color perceptions across species. It is question-begging to assert that in principle, manipulations are irrelevant to assessing Non-comparability. We are owed an argument. Importantly, one means of mounting that argument is unavailable to adherents of Objectivism, namely, attempts to explain (in)-compatibilities across color spaces by appeal to sensory physiology. Another approach would be to free different color spaces from the physical underpinnings of human color. For example, one might think of pigeon colors as not only nonphysical features, but as features that do not supervene or in any important way correlate with reflectances. In this case, manipulating human colors will leave pigeon colors unaffected. Alas, this reply has numerous drawbacks: it is question-begging, makes colors rather mysterious, and stands in tension with the very evidence that underpins Selectionism (i.e., the well-studied relationships between reflectances and animal color vision). I thus submit that the above argument against Non-comparability has merit.

4 Conclusion: how bad is the problem?

In reply, the Selectionist might argue that the picture sketched in Section 3 admits that uniform objects can and often do have multiple colors (e.g., the uniform object is both blue and shmue at the outset, and both green and shmeen at the conclusion). In this sense, although it speaks against Non-comparability, it is consistent with some form of Objectivist Pluralism. That is, it certainly does not justify Incompatibility.

This misses the point. The argument wasn’t intended to defeat Selectionism simpliciter, it was intended to undermine a common form of Selectionism, namely one that proposes Non-comparability to explain how Pluralism can obtain within Objectivism. That goal, I believe, the argument achieves. Beyond that, the above picture does feed into some well-known other challenges to Selectionism. In what remains, I sketch that impact.

Two central pressures against color Objectivism, including Selectionism, stem from the relativity of perception and the structure of color.10 Regarding the former, the worry is that the color of a given objective thing can vary along dimensions that are extrinsic to that thing, such as the nature of the perceivers’ perceptual apparati, and hence that the color at issue cannot be intrinsic to that thing. If that color in fact depends on perceptual apparati, then it is not only not intrinsic to the thing, it is at least somewhat subjective. The worry stemming from the structure of color typically proceeds by specifying various dimensions of colors (e.g., hues, saturations, and lightnesses) and various relations between colors (e.g., relations of composition, exclusion, similarity). The challenge to Objectivism advocates is to explain the resulting relations or overall structure between colors in terms of objective properties (e.g., in terms of reflectances and their features). This turns out to be a nontrival task, for various aspects of color structure seem to be well explained by appeal to perceivers’ perceptual apparati, and less well explained solely by appeal to objective properties. While I am tempted by both arguments, I will not presume their soundness. Instead, I will conclude by remarking on how the above fits into them.

Since (in)compatibilities between colors are typically included within the structure of color, there is a natural connection between our topic and the argument from structure. Suppose there are (in)compatibilities between human and pigeon colors. Color ontologists must explain these in terms of their preferred ontology. Selectionists must therefore explain these in terms of relations between objective colors (or explain why they are not essential to the nature of human and pigeon colors – set this option aside). In one respect, we already have a hint of a solution. For illustration, suppose that something is blue because it reflects light from the short end of the visible electromagnetic spectrum and absorbs it from the long end, and something is shmeen because it reflects light from the long end and absorbs it from the short. We then have the roots of an explanation for why blue and shmeen are incompatible. In this regard, these considerations can assist Objectivism. This again illustrates that my target is not Objectivism so much as the Non-comparability explication of Selectionism. But the details matter, if for no other reason than that this “explanation” of color (in)compatibilities presumes the existence of a viable account of colors in terms of reflectances, and it is far from clear that there is one. Thus, even if some (in)compatibilities between color spaces can be given an objective explanation, it does not follow that all can. Nor does it follow that other aspects of the structures of color spaces fit into the ideals of Objectivism.

With regard to the relativity of color, Selectionism has been offered in part to undermine such arguments, which tend to rely on the premise that a uniform object cannot be one and another color at the same time (Kalderon 2007, Allen 2009). The Non-comparability approach can be used to resist that premise. However, given the above considerations, that approach should be resisted, for a more flexible view of the relations between applied color spaces seems more accurate. If such a “flexible” view can overcome the tension between Pluralism and Objectivism, then at least this challenge from the relativity of interspecies color perceptions would be contained. The challenges raised by other forms of relativity (see, e.g., Section 0) would remain unaddressed.11

Notes

1 Allen (forthcoming) provides a wonderful, accessible overview – one to which I am indebted.

2 See Matthen (1999) for an extended defense. “Exclusion ranges” is from Matthen (2005: 103–5) and “family” is from Kalderon (forthcoming). “Spaces” is used in various publications, but my main source is Allen (2009).

3 Kalderon (2007) contains a beautiful defense of a similar idea. Note that his preferred perceptual theory is relationalist as opposed to representationalist, and that his pluralism is specifically designed to preserve color objectivism, a matter to which I will turn below.

4This conclusion may, but need not, be reached via dispositional color ontologies, views that assert that colors are powers of objects to produce color sensations in perceivers. The matter crucially depends on how “color dispositions” are fleshed out, something space prevents me from discussing.

5 This is a clear commitment of Matthen’s exclusion ranges (Matthen 2005: 103–5), and arguably the Selectionist views of Kalderon (2007) and Allen (2009).

6 The matter is more complicated than this because we can articulate numerous mappings between geometries and discuss various ways in which items in one can be mapped into items in the other, and vice versa. Let us set aside whether or not these mappings can yield a relevant sense of (in)compatibilities between abstract geometries.

7 Campbell (2006) offers a great discussion of color manipulation, but doesn’t consider its application to Pluralism.

8 There are numerous details about color (in)compatibilities within a color space that are worthy of exposition and analysis. Given length constraints, I must leave these details aside.

9 I suppress this qualification throughout.

10 Arguments from perceptual relativity against perceptual objectivity have well-known and ancient roots. A recent and influential use of this reasoning in color is Cohen (2009). Hardin (1988) is a well-known source for the argument from structure; another is Pautz (2006).

11 I am indebted to Jake Beck and Mohan Matthen for helpful comments on an earlier draft. Sincere thanks to Brandon University for granting the sabbatical leave during which this article was written, Clare Hall and The University of Cambridge for providing welcoming accommodations, and the Templeton Foundation and Tim Crane’s New Directions in the Study of Mind project for financial assistance. This publication was made possible through the support of a grant from the John Templeton Foundation. The opinions expressed in this publication are those of the author and do not necessarily reflect the views of the John Templeton Foundation.

Further reading

Thompson, E., A. Palacios, and F. Varela (1992) “Ways of coloring: Comparative color vision as a case study for cognitive science,” Behavioural and Brain Sciences, 15: 1–74, was a ground-breaking work on comparative color vision. Hardin, C. L. (1992) “The virtues of illusion,” Philosophical Studies, 68(3): 371–82, argues that color vision evolved to help organisms engage with and think about their environments by creating colors for visual experience and thought. It is worth reading alongside Hilbert, D. (1992) “What is colour vision?” Philosophical Studies, 68(3): 351–70, which argues that color constancy is central to color vision. Mollon, J. (1989) “‘Tho’ she kneel’d in that place where they grew… ’: The uses and origins of primate colour vision,” Journal of Experimental Biology, 146: 21–38, is an influential statement of the frugivory hypothesis, the idea that the purpose of color vision is to distinguish fruit from foliage.

References

Akins, K., and M. Hahn (2014) “More than mere colouring: The role of spectral information in human vision,” British Journal for the Philosophy of Science, 65: 125–171.

Allen, K. (2009) “Inter-species variation in colour perception,” Philosophical Studies, 142: 197–220.

——— (forthcoming) “Interspecies variations,” in D. H. Brown and F. Macpherson, eds. The Routledge Handbook on the Philosophy of Colour, Milton: Routledge.

Broackes, J. (2010) “What do the color-blind see?” in J. Cohen and M. Matthen (eds.) Color Ontology and Color Science, Cambridge, MA: MIT Press.

Byrne, A., and D. Hilbert (2003) “Colour realism and colour science,” Behavioral and Brain Sciences, 26: 3–64.

Campbell, J. (2006) “Manipulating colour: Pounding an almond,” in T. Gendler and J. Hawthorne (eds.) Perceptual Experience, pp. 31–49, Oxford: Oxford University Press.

Cohen, J. (2009) The Red and the Real: An Essay on Colour Ontology, Oxford: Oxford University Press.

Hardin, C. L. (1988) Color for Philosophers, Indianapolis, IN: Hackett.

Kalderon, M. (2007) “Color pluralism,” The Philosophical Review, 116: 563–601.

——— (forthcoming) “Monism and pluralism,” in D. H. Brown and F. Macpherson, eds. The Routledge Handbook on the Philosophy of Colour, Milton: Routledge.

Matthen, M. (1999) “The disunity of color,” The Philosophical Review, 108: 47–84.

——— (2005) Seeing, Doing, and Knowing: A Philosophical Theory of Sense Perception, Oxford: Oxford University Press.

——— “Novel colours in animal perception,” Chapter 6 in this volume.

Pautz, A. (2006) “Can the physicalist explain colour structure in terms of colour experience?” Australasian Journal of Philosophy, 84(4): 535–564.

Tye, M. (2006) “The puzzle of true blue,” Analysis, 66: 173–178.