Far-from-equilibrium systems. The clearest implication of the interactive model for b i o l o g i c a l f o u n d a t i o n s i s t h a t g e n u i n e r e p r e s e n t a t i o n , t h u s g e n u i n e c o g n i t i o n , c a n e m e r g e
only in far-from-equilibrium systems — more precisely, recursively self-maintenant systems. It is only in far-from-equilibrium systems that function can emerge, and only a
certain kind of interactive indicational function that constitutes emergent representation.
This is already a major shift from computer model or connectionist approaches. If the interactive model is correct, various properties of cognition can be simulated in computers and in connectionist nets, but they cannot be more than simulations. There are actually several deficiencies of computer models. The first is that they are not self- maintenant systems — by most standards, they are not even far-from-equilibrium systems
(their dependence on externally supplied power does keep them at some distance from thermodynamic equilibrium).
Open systems. A second is that computers are not in any relevant sense open systems. They process inputs or data, but do not interact with their environments. Recursive self-maintenance requires environmental interactions that achieve a closure in the sense of circling back to support the far-from-equilibrium conditions that made those interactions possible in the first place (Christensen & Hooker, 2000; Ruiz-Mirazo &
Interactive timing. A third involves issues of interactive timing. Successful interactions require appropriate timing. Mere speed is not sufficient; an interaction can fail from being too fast just as easily as from being too slow. Computers do have timing information in their internal clocks, but there is no timing in the Turing machine formalisms (and equivalents) for computers, and the architecture of the timing in a computer is not an
evolutionary possibility. If the brain involved clock driven processes akin to those in a computer, every evolutionary change in the brain would have to have involved simultaneous adaptive changes in the interactive circuitry and precisely coordinated changes in the timing circuitry. That is vanishingly improbable for any instance, and simply impossible for any extended evolution (Bickhard & Terveen, 1995).
The brain solves the timing problem in a very different way: Put clocks everywhere, and construct all functional relationships out of relationships among those clocks. Since clocks are “just” oscillators, this translates into: build the central nervous system out of oscillators, and construct all functional relationships out of modulatory
relationships among those oscillatory subsystems. This is, in fact, the basic architecture of