Direct Perception Requires an Animal-Dependent Concept of Specificity and of Information

Abstract

Specificity and information are at center stage in ecological psychology. Nevertheless, the usual theorizing on these concepts may have made the problem of accounting for perception and action more difficult by so far underestimating the role of animals as both meaning-detectors and meaning-determiners. The usual understanding of information and specificity in ecological psychology seems neither necessary nor even compatible with ecological premises and empirical findings. I argue that a reframing of these concepts to fully take animals into account is necessary to explain perception of action-specific meanings. The reframing proposed converges on ideas from developmental systems theory and in no way concedes to inputs-followed-by-processing-followed-by-representation models. Fully acknowledging the animal for properly defining information over the animal-environment system poses no threat lawfulness, realism, or direct perception. It also invites serious consideration of self-organization and interactivism as sources for further development of ecological science.

Acknowledgments

I thank Claire Michaels, Paula L. P. Silva, Christopher Palmer, Damian G. Kelty-Stephen, and Mark H. Bickhard for very helpful discussions and comments on earlier drafts.

Notes

¹ Haptic information about the body obviously cannot be argued to be “out there” in ambient arrays. However the discussion is meant to call attention to information being defined as “out there” not in the geographical sense of being outside the boundaries of the animal body but in the sense of existing outside and independent of perceptual processes.

² In the sense of pragmatic achievement, an act of perceiving. I do not want to suggest that perceptual processes are distinctively mental. “Perception is not a mental act. Neither is it a bodily act. Perceiving is a psychosomatic act, not of the mind or of the body but of a living observer” (Gibson, 1979/1986, pp. 239–240).

³ As I suggested before, perception is not primarily of action-neutral attributes (length) but of affordances (e.g., wieldability or some other affordance related to length), to which the moments of inertia could be argued to be specific. However, just as tau (time to contact) cannot per se specify the affordance of intercepting an object (Stoffregen, 2000), moments of inertia per se cannot specify “wieldability.” Neither tau nor moments of inertia (or any variable defined in terms extrinsic to the animal) can inform about an actor's capabilities and their fit to the environment.

⁴ These variables might be treated (only as) potential information (Michaels & de Vries, 1998). Jacobs and Michaels (2007) treated potentially informative patterns in terms of their usefulness in a certain context. The term usefulness is interesting as it alludes to actual use by an active perceiver.

⁵ Kugler and Turvey (1987) treated informational properties as emerging from the dynamics of the movement system: “New informational properties emerge … from the creation and annihilation of singular states; these map back onto the system as constraints for local, muscle-joint kinetic spaces—that is, the movement system is internally self-organizing” (p. 425). I think there is no principled justification not to treat all information for perception and action in like terms.

Additional information

Funding

This research was supported by Grant 200676/2009-1 of Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq; Brazilian National Council for Scientific and Technological Development).