ABSTRACT
This chapter examines a naturalistic approach to the hypothesis that human perception comprises self-organizing mechanisms for the detection of informational patterns as a result of the mutual interaction between organisms and the environment. The novelty of this approach, though it deals with a well-known hypothesis, resides in the indication of how certain informational patterns may be represented according to self organizing principles applied to neural network models of perceptual activities. Self-organization can be understood as “self-connection,” where a form of conditionality is established between regularities in the environment and states of activation/inhibition of the network’s neuron-like units. A neural network trained in accordance with the Harmony principle will eventually reach a state of minimum energy, which corresponds to its internal state of temporary equilibrium. The harmony principle can be seen as a useful mechanism for the study of self-organizing systems.
