This commentary suggests how Norman's dual control model of vision can be fitted into a broader general model of the control of behaviour by direct (on-line) and indirect (off-line) processes. Some general principles of behavioural organization, development, and competition are described and their specific application to vision is noted.

My response and reactions to the quite diverse commentaries are presented. Among the topics covered are a response to holders of the ecological viewpoint; memory and learning in the two perceptual systems; development of the two systems; biological motion; size and distance perception; illusion and the two systems; and several others. It is suggested that the dual-process approach is a viable working theory of space perception and, perhaps, of other types of perception as well. Hopefully, future research will enhance it with added refinements and variations on the original theme.

In her Behavioral and Brain Sciences target article, Greenfield (1991) proposed that early in a child's development Broca's area may serve the dual function of coordinating object assembly and organizing the production of structured utterances. As development progresses, the upper and lower regions of Broca's area become increasingly specialized for motor coordination and speech, respectively. This commentary presents a connectionist simulation of aspects of this proposal. The results of the simulation confirm the main thrust of Greenfield's argument and suggest that an important impetus for the developmental differentiation in Broca's area may be the increasing complexity of the computational demands made upon it.

Ronan Reilly's connectionist simulation both strengthens and advances the theoretical model presented in my 1991 target article, “Language, Tools, and Brain: The Ontogeny and Phylogeny of Hierarchically Organized Sequential Behavior.” Reilly has tested the whole ontogenetic model with a single simulation study explicitly planned for this purpose. His methodology has established that the various components of the theoretical model imply and are compatible with one another. It has also indicated how learning can actualize a pre-established ontogenetic sequence of combining lingusitic symbols and objects. His simulation suggests that the acquisition of linguistic speech may be facilitated by experience with object manipulation, but not vice versa. This hypothesis can and should be empirically tested through research on behavioral development in the two domains. Finally, Reilly has simulated brain architecture, as well as neural learning. His simulation therefore shows how the development of language and object manipulation can result from an interaction between preprogrammed neural architecture (analogous to network architecture) and experience (analogous to the network's training cycles).