Over the recent months during Intelligence Design related discussions at UncommonDescent.com I heard many naturalistically-minded colleagues say that we, ID proponents, cannot adequately define intelligence.
I think that as far as cybernetics is concerned intelligence can be defined as anything that is able to impart functionality to systems. This phenomenological definition is analogous to the definition of gravity in Newton's mechanics as a natural phenomenon by which physical bodies attract with a pull proportional to their mass. Intelligence is the only means of creating functionally complex configurations/patterns of matter that we know of in practice. Figuratively speaking, intelligent agency leaves a trace of specified complexity behind, similarly to ants communicating by a feromone trail. Using the above definition we can get away from the curse of infinite regress that comes as a package deal with naturalistic emergence theories (that is, the so-called theories of self-organisation that have little to do with reality or practicality).
Defining intelligence in this way, we can get away from the vicious circularity of naturalistic emergentism present contemporary self-organisation theories.
- Emergentism postulates the hypothetical spontaneous emergence of particular properties of systems under certain conditions. E.g. self-organisation assumes unobserved spontaneous generation of formal functionality and cyberbetic control in complex systems.
Another important issue here is the ability to scientifically describe concrete ways by which intelligence can act on matter. I believe that the action of intelligence is inherently heuristic. It is quite interesting to note in defense of this statement, that animals and humans can sometimes solve practical combinatorial problems very efficiently (see e.g. this paper about the efficiency of solving the travelling salesman problem visually by practitioners).