Jarvilehto's target article argues from a claim about the nature of sensory receptors to the view that the organism and environment should be regarded as a single dynamic system. In this commentary I endorse the claim about sensory receptors but argue that the inferences drawn from that claim are controversial.
2. As I understand it, Jarvilehto's argument runs as follows:
(1) If the organism and environment constitute two distinct systems, then sensory receptors must function as passive transducers of information about the environment -- "channels through which knowledge arrives into the organism from the environment" (paragraph 2).
(2) However, empirical evidence about efferent effects on sensory receptors, and a thought-experiment about an organism without any receptors at all, show that receptors are not passive transducers of information about the environment.
(3) Therefore, it is not the case that the organism and environment constitute two distinct systems. They are better understood as a single dynamic system.
3. It is very plausible that Jarvilehto is right in claiming that sensory receptors are not simply passive transducers of environmental information. A familiar phenomenon, such as dark adaptation in vision (Hood & Finkelstein, 1986), demonstrates that receptors can adapt to changes in the environment so as to make receptor states maximally useful in guiding the behavior of the organism. Similar comments might be made about various top-down effects on perception, such as the effects of attention mentioned in the target article.
4. Jarvilehto makes a significant point when he demonstrates the possibility that other cognitive states may also affect sensory receptors in cognitively relevant ways. Indeed, Jarvilehto's thought-experiment suggests that receptors may, in certain limited behavioral conditions, be unnecessary for the development of adaptive behavior. For the purpose of this commentary, I will assume that Jarvilehto is right in claiming that there are such efferent effects on sensory receptors and, thus, that they are not passive transducers of environmental information. My concern will be to ask what can be inferred from that assumption.
5. I begin by considering the first premise of Jarvilehto's argument, to the effect that if the organism and environment constitute two systems, then sensory receptors must be transducers of environmental information. In order to evaluate this claim, consider the case of color. On the standard view of color vision, color experience tracks the property of surface spectral reflectance; the disposition of a surface to reflect a fixed proportion of light at every wavelength of the visible spectrum. In addition, the fundamental fact about the psychology of color vision is that colors are organized phenomenologically within a space of three opponencies: red-green, blue-yellow, and black-white (Hurvich, 1981). Crucially, however, the abstract structure of reflectance does not seem to exhibit the opponency manifested in color experience (Hardin 1988). Now, let us suppose that this is indeed the correct description of the relevant facts of color vision. On this description, whatever role the sensory receptors (i.e., the cones) play in color vision, it cannot be the mere passive reception of information about reflectance, for the simple reason that information about reflectance is not in fact passively received ("channelled") into the organism from the environment. Whatever information the light carries about the reflectance of object surfaces is restructured by the properties of the visual system that cause color vision to have an opponent structure. This restructuring is adaptive, because, for example, it makes red fruit more salient against a green background than it would otherwise be, but it is not informative in the strict sense. Hence the function of the cones is not to act as passive transducers; because the representation of color is not one that embodies anything like passively transduced information.
6. What is relevant about this case to the present argument is that it shows that the claim that the cones are not passive transducers of information is quite independent of any view about whether the organism and environment should be conceived as distinct or unitary. It is quite consistent with the view of color vision just described that the organism and environment are separate systems. But if one can reject the view that sensory receptors are passive channels of information while retaining the organism-environment distinction, then the first premise of Jarvilehto's argument -- the claim that if the organism and environment are distinct the receptors must be passive transducers of information -- is false.
7. I turn next to the question of whether we should infer from the evidence Jarvilehto cites, and from the thought-experiment he develops, that the organism and environment constitute a single dynamic system. The answer to this question naturally depends in part on what is meant by "single dynamic system," and one cannot expect a full development of this idea in the short space of Jarvilehto's target article. We can, however, briefly consider what the claim -- call it the "single system claim" -- that the organism and environment constitute a single dynamic system might amount to.
8. At least three interpretations of the single system claim are suggested by Jarvilehto's discussion. On the first interpretation, advocating the view that organism and environment constitute a single system is asserting that there is no passive transfer of information between the two. This is the view just discussed. Because there is no transfer of information, so the claim goes, it makes little sense to conceive of the organism and environment as distinct.
9. On a second interpretation of the single system claim that seems to be implicit in Jarvilehto's thought-experiment, the notion of information ought to be abandoned altogether in favor of conceiving causal interaction between the organism and environment as resulting in an increase in behavioral adaptiveness. On this interpretation, the single system claim might be defended as follows: If our concern is with causal processes rather than informational ones, there is no principled organism-environment distinction to be made. This is the case because there are innumerable causal processes involved in behavioral adaptation, some occurring within the organism, some within the environment, and some between the two, and it is likely that a revealing description of the process of adaptation will appeal to a subset of these processes that cuts across the organism-environment divide and renders the distinction irrelevant.
10. On a third interpretation suggested by Jarvilehto's discussion of the empirical results he cites, the single system claim is to be taken as a methodological one concerning the best way to theorize about, or model, perceptual and other processes relevant to behavior and behavioral adaptation. On this interpretation, it is a theoretical fact that treating the organism and environment as a single system produces the best scientific understanding of at least some perceptual or cognitive processes.
11. I have already argued that the first of these interpretations makes the single system claim controversial because it is possible to reject the claim of passive information transfer without the more radical rejection of the traditional distinction between organism and environment. A similar case can be made regarding the second interpretation. The traditional view maintains that there is an information transfer between the organism and the environment although the transfer may not be passive in the sense discussed above. No traditionalist, however, denies that the transfer of information, whatever its features, depends in some way upon the causal processes of sensory transduction and neural activity brought about thereby. Indeed, the classical computational view (Marr, 1982) explicitly assigns a place in cognitive theory for a description of the implementation of cognitive processes, and at the level of implementation one finds a causal story. Hence on this interpretation too, one need not opt for the radical denial of the distinction between organism and environment in order to affirm the essence of the causal claim being made.
12. The third interpretation of the single system claim is much more conservative because it holds only that the best scientific approach to perception is one in which the organism and environment are treated conceptually and mathematically as a single (dynamic) system. This is a conservative claim because it defers to the final judgement of scientific progress: if the dynamic single system approach produces the better theory, then its conceptual framework is the one to be adopted. In particular, the distinction between the organism and environment is, in the relevant sense, to be denied. If the dynamical approach does not produce the best theory, then the traditional distinction can be maintained. On this interpretation of the single system claim, however, one cannot defend it solely by appeal to a theory of the sensory receptors, but only by appeal to successful perceptual and cognitive theories that make use of the single system claim in practice. Jarvilehto cites some of the relevant evidence, and further evidence is also available (see, e.g., Port & van Gelder 1995). A final decision awaits the outcome of further research into perception and cognition.
Hardin, C.L. 1988. Color for Philosophers. Indianapolis: Hackett.
Hood, D.C. & Finkelstein, M.A. 1986. Sensitivity to light. In K. Boff, L. Kaufman, & J. Thomas (eds.), Handbook of Perception and Human Performance, volume 1. New York: Wiley Inter-Science.
Hurvich, L.M. 1981. Color Vision. Sunderland: Sinauer Associates.
Jarvilehto, T. (1998) Efferent Influences on Receptors in Knowledge Formation. PSYCOLOQUY 9(41) ftp://ftp.princeton.edu/pub/harnad/Psycoloquy/1998.volume.9/ psyc.98.9.41.efference-knowledge.1.jarvilehto http://www.cogsci.soton.ac.uk/cgi/psyc/newpsy?9.41
Marr, D. 1982. Vision. San Francisco: Freeman.
Port, R.F. & van Gelder, T. 1995. Mind as Motion. Cambridge: MIT.