l'm trying to read and understand Varela's "Calculus for Self-Reference" .
l thought l'd do myself a favour and have a wee joint before l tackled it, big mistake!
In a tradition of different origin, the idea of self- programming, or self-modification has emerged as a possible remedy to the often noticed stiffness of formal systems. At the root of the idea we find the intuition that a system should sometimes become something else than it was before; in other words, that its defining symbols can be both too vague and too narrow in comparison to processes of interest. Part of what in the seventies were called "goal-seeking systems", "catastrophes", or "dissipative structures" (characterized by a computational bottom-up approach that permits flexibility at a higher level) lived further in the eighties in connectionism and in the subsymbolic paradigm, or in the notion of "swarm intelligence". A good introduction to the flexibility aspect of subsysmbolic computations is A. Clark's Micro Cognition . Self-programming computers that could directly alter their own codes were already studied by von Neumann , and several authors afterwards. Whether and to what extent self-programming can be carried out algorithmically, was always unclear, however.
l should have had a couple of beers!!