Andy Clark (1998) Could Sensing Play Multiple Roles?. Psycoloquy: 9(80) Efference Knowledge (6)

Volume: 9 (next, prev) Issue: 80 (next, prev) Article: 6 (next prev first) Alternate versions: ASCII Summary
PSYCOLOQUY (ISSN 1055-0143) is sponsored by the American Psychological Association (APA).
Psycoloquy 9(80): Could Sensing Play Multiple Roles?

Commentary on Jarvilehto on Efference-Knowledge

Andy Clark
Philosophy/Neuroscience/Psychology program
Washington University in St. Louis
St. Louis, MO 63130, USA


In his engaging and provocative target article, Jarvilehto (1998) argues that the senses are not transmitters of environmental information so much as connections that support dynamic organism-environment systems. I think there is much to applaud in this treatment, but I want to raise two small doubts: one concerning the nature of the target (why information transmission rather than inner model-building?) and the other concerning the scope of the claim (is sensing always and only used in the way the author suggests?).


afference, artificial life, efference, epistemology, evolution, Gibson, knowledge, motor theory, movement, perception, receptors, robotics, sensation, sensorimotor systems, situatedness
1. Professor Jarvilehto (1998) is absolutely right, I believe, to stress the way sensing can sometimes play a role rather different from that stressed by mainstream cognitive science. A very simple example that captures part of the idea is running to catch a ball (a baseball or a cricket ball, say). Giving perception its standard role, we might assume that the job of the visual system is to take in enough information to project a trajectory so that we can run to where the ball will land. It seems, however, that nature has a better solution: you run so that the tangent of elevation of gaze from the fielder (you) to the ball stays constant (at zero; see Cliff and Noble, in press; Lee 1980). This is a nice, cheap, robust solution, but it also displays, as the roboticist Smithers (1994) has recently pointed out, a somewhat different role for perceptual input. Instead of using sensing to get enough information into the system to allow it to "throw away the world" while it solves the problem internally, it uses the sensor as a conduit allowing environmental magnitudes to exert a constant influence on behavior. As Jarvilehto might put it, the sensor allows the creation of a coupled organism-environment system whose intrinsic dynamics solve the problem. The target treatment then goes further in (nicely) pointing out that efferent signals can alter the response characteristics of receptors. Hurley (1998) presents a strong case for the role of associated motor signals in determining the phenomenological character of certain perceptual events. All these points, it seems to me, tell us true and important things about the way brain, body and world form a unified problem-solving whole. (For an extended defence of just this kind of view, see Clark 1997).

2. Now for the doubts: First, consider the claim that the senses are not transmitters of environmental information. Certainly, all these results suggest that it is OFTEN wrong to see the job of sensing as the creation of a rich inner model that can then be used in place of the real world itself. They also suggest that sensing, acting and perceived content may all be intertwined in complex and important ways. But nothing hereabouts, so far as I can see, shows that the senses NEVER function so as to create an inner model of the world, nor that -- even in the more minimal cases -- they are not transmitting information about external events and processes. Thus, consider catching the baseball. It is true that the best way to understand the role of the sensing is to see it as creating a useful organism-environment loop. But in so doing, the sense channel surely IS carrying information about the location of the ball. What it is NOT doing is presenting such information as might be used to create a rich inner model of the environment. It is selective sensing, selective and minimal information transmission: but it is (or so I would suggest) information transmission all the same.

3. Second, consider the more general claim about the role of the senses in forming knowledge by reorganizing the organism-environment system. Whilst agreeing that this characterization is useful for a great deal of on-line daily problem-solving, I wonder how we should best think of off-line reflection? Certainly one can, and sometimes does, spend time just sitting, reviewing ideas and scenarios in one's imagination. And much of the knowledge that figures in such episodes was acquired by sensory engagement with the world. How, then, should we understand this aspect of the senses role in forming knowledge?

4. Such doubts and quibbles aside, I found this a fascinating target article. We are indeed all too easily seduced by the image of sensing as the brain's way of rebuilding the external world inside the head. Better appreciating the complex interplay between minimal, selective sensing, motor control, and successful thought and action must be high on the agenda for the next century of Cognitive Science.


Clark, A (1997) Being There: PUTTING BRAIN, BODY AND WORLD TOGETHER AGAIN (MIT Press, Cambridge, MA)

Cliff, D and Noble, J (in press) Knowledge-Based Vision and Simple Visual machines. PHILOSOPHICAL TRANSACTIONS OF THE ROYAL SOCIETY

Hurley, S (1998) CONSCIOUSNESS IN ACTION (Harvard Univ. Press, Camb. Ma).

Jarvilehto, T. (1998d) Efferent influences on receptors in knowledge formation. PSYCOLOQUY 9(41)

Lee, D (1980) Visuo-motor Coordination in Space and Time. in G. Stelmach and J. Requin (eds) TUTORIALS IN MOTOR BEHAVIOR (North Holland)

Smithers, T (1994) Why Better Robots Make It Harder. in D. Cliff, P. Husbands, J.A. Meyer, and S. Wilson (Eds.) FROM ANIMALS TO ANIMATS 3 (p. 64-72). Cambridge, MA: MIT Press.

Volume: 9 (next, prev) Issue: 80 (next, prev) Article: 6 (next prev first) Alternate versions: ASCII Summary