Reductionism & Subjectivism Defined & Defended

Behavioral and Brain Sciences, 15 (1), 1992, 32-33.

A Commentary on "Ways of Coloring: Comparative color vision as a case study for cognitive science" by Evan Thompson, Adrian Palacios, & Francisco Varela. Behavioral and Brain Sciences, 15 (1), 1992.

Austen Clark
Department of Philosophy U-54
University of Connecticut
Storrs, CT 06269-2054

As a reductionist and a subjectivist I find little to dispute, and much to cheer, in the use of the comparative argument against objectivism. The best available form of objectivism is anthropocentric realism, and at the very least the comparative argument dispels much of the motivation for that position. But I have qualms with the arguments arrayed against "neurophysiological subjectivism." I do not think they demonstrate failures either in neurophysiological reductionism or in subjectivism.

Thompson, Palacios, and Varela clearly deny neurophysiological reductionism. They write that chromatic experiences "do not yield to analysis purely in terms of the neurophysiological structure of the perceiver" (14); that "...a purely neural explanation for color vision is incomplete" (30); that color "experience can be understood only in the context of its ecological embodiment" (14-15) and that any "adequate account" of color must be ecological (1). Furthermore, the embrace of the preferred ecological and "enactive" perspective is claimed in several places to be contrary to subjectivism, presumably meaning that they cannot both be true (1, 14-15, 37).

In outline the argument against neurophysiological subjectivism proceeds as follows:

  1. Animal capacities for color vision in many cases co-evolved with the very colors that those capacities enable the animal to discriminate.
  2. Hence color vision contributes to the determination of the environment of an animal.
  3. Hence the qualities perceived in color vision are not purely internal and "in the head", as subjectivism claims, but are ecological.

A neurophysiological reductionist will happily agree to (1), dispute the terms in (2), and steadfastly deny the inference to (3). As examples of (1) the authors cite polymorphism in the color vision of squirrel monkeys, the co-evolution of honey bee color vision and flower colors, and the evolution of animal coloration for camouflage and signal purposes (pp. 29-30). None are inconsistent with neurophysiological reductionism. The neurophysiological reductionist claims that the chromatic experiences a species has today can be explained without remainder in terms of the visual nervous system that species has today (Clark 1980, 1992). Curious humans can go on to ask: why do these various species have the visual systems that they do? One hopes evolutionary theory will someday provide an answer. But whatever the answer is, it will not contradict neurophysiological reductionism as described. The neurophysiological story is, admittedly, "incomplete", but only in the sense that it does not ask--and so does not answer--the question of the origins of the systems to which it refers. Presumably no one thinks that evolutionary history affected color vision independently of its effects on nervous systems!

In the second stage of the argument the authors shift from claims of animal-environment co-evolution to what they call "co-determination." The term is paradoxical. At least in the ordinary sense of "environment," animals with distinct perceptual capacities can share the same environment. Which animals determined its features? Flowering plants are parts of the environment of many different animals. According to Thompson, Palacios, and Varela, some have incommensurable color spaces. If the color of flowers were determined by animal-environment co-evolution with one of those species, it therefore could not have been determined by the history of its interactions with other species. Contrary properties cannot both be determinative. If the bees did it, it can't have been the pigeons.

At some points the authors use the words "environment" and "perceptual object" in ways that would lead them to deny that environments or perceptual objects are (or even could be) shared by species with different perceptual capacities (cf. 25, 27, 37). Such usages are salient in the third stage of the argument, where it is urged that we adopt an "ecological" and "enactive" approach to color vision. The gaps left in neurophysiological accounts are to be plugged by treating perception always as perceptually guided activity (34, 35). Colors are "properties of the world that result from animal-environment co-determination" (33). There is much in this talk that I do not understand or that seems open to immediate and devastating counter-examples. Try star gazing. On a moonless night, find a dark field, sit still, and you will see the colors of the stars. Although there are those who believe the stars determine the fate of humans, the stirring history of our species couldn't have affected any feature of the light entering one's eyes. It is also unclear what sort of "perceptually guided activity" is going on in our motionless star-gazer.

Perhaps the "ecological" perspective on the flower in my garden is as follows. That flower is part of my environment, and by hypothesis its colors are in part determined by the interactions its ancestors had with mine. That same flower is not part of the environment of honey bees; instead they perceive a different object, with different colors, determined in part by the history of interactions between that object and the ancestors of the bee (37).

One problem with talking this way is that evolutionary theory could not corroborate any claims about that kind of "environment." For purposes of evolutionary theory we must be able to talk about the one flower out there in the garden, the one with genes, etc., that can interact with many different species.

A subjectivist and reductionist description of the situation seems much more straightforward. The one flower out in the garden has various properties. When suitably illuminated it causes in me various kinds of chromatic experiences, and other kinds, perhaps incommensurable kinds, in the honey bees and local pigeons. The character of those experiences can in principle be entirely explained by--reduced to--the neurophysiology of our various visual systems. We happily turn to evolution for hints as to why the species have the different systems that they do. The same account accommodates star-gazing most elegantly. Our visual system evolved, and we can admit it may have co-evolved with a panoply of animal and plant colors in biological environments. Nevertheless, it now gives us certain capacities, and we can exercise those capacities even when confronted with stimuli (such as starlight) remote from biological environments and for which a co-determination story is completely implausible.

Finally, only straw-man versions of subjectivism are contrary to claims of animal-environment co-evolution. Subjectivists need not define colors as "internally generated qualities that the animal simply projects onto the world" (33), and do not endorse the view that colors are simply "in the head" (37). We can admit the existence of stimuli. The point is rather that the "collecting principles"--or that in virtue of which two stimuli are perceived to be the same color--cannot be framed in extra-dermal terms. The stimuli are indeed out there in the world--note they must be, if genes can ever control them. But the principles that collect those stimuli into color classes can only be found in the innards of visual nervous systems. That different species operate in such radically different ways is just sauce for our (putatively pentachromatic) goose.


Clark, A. G. (1980). Psychological Models and Neural Mechanisms: An Examination of Reductionism in Psychology. Oxford University Press.

Clark, A. G. (1993) Sensory Qualities. Oxford University Press.

Return to Austen Clark's online papers .

Return to the Philosophy Department home page.