The Engineering discipline gains tremendous power by taking a reductionist modular approach. An Engineering problem is recursively broken down into component problems with well defined interfaces. Each of the problems are solved in turn and then they are all "bolted" back together into the engineering solution.
However, I think that the traditional Engineering approach is ill suited to constructing an intelligent device.
Overwhelming complexity abounds at every level in a brain; chemistry, cellular, neurological, psycological and social. There is most certainly structure in a brain, but that structure is inextricably intertwined.
Brains have evolved over aeons as an intimate part of organisms, their behaviour, and their reproductive success. Each step of the evolution has necessarily been an adjustment to the genotype at the time which resulted in a reproductive advantage in the environment at the time. Evolutionary structure is in a sense heirachical and incremental in time with each juncture a mutational event.
Each individual brain forms as part of the unique development of an organism. However, DNA simply does not contain sufficient information to explicitly specify every construction detail. Provided the development environment is 'normal', a 'normal' functioning brain results.
If intelligence is necessarily based on structures such as these, then we need different tools to engineer it. It would seem to require a step back to concentrate on specifying the critical junctures and to allow he rest to 'look after itself'. It is a more 'organic' approach focussed on defining the growth process and influencing the growth environment.
Self-Organising systems may provide the techniques required to progress this approach.
Page updated