The emphasis that Principia Cybernetica places on consensus about fundamental concepts and
principles can be criticized as risking the dangers of formalism and
foundationalism, of adopting a deductive strategy in which cybernetic
theory is linearly derived or proved from axioms. The risks of such
approaches are obvious: either the development of an ossified, static
philosophy which cannot adapt to new information; or an endless, futile
search for the ultimate, jointly necessary and sufficient axiom set from
which "truth" could be derived.
We are well aware of these dangers, but on the other hand we are also aware
of the risk of the converse, of a failure to generate any firm
foundations on which theory can be constructed. We believe that it is this
latter condition that Cybernetics and Systems Science has indeed found itself in today. Even a
cursory examination of current systems literature will reveal a veritable
zoo of advanced, highly sophisticated theories which have only a loose and
metaphorical relation to each other. A clear and elegant underlying theory
on which they could be reconciled is simply lacking.
Rather the approach that we adopt aims to steer a middle ground between
both extremes. It does so through the reliance on the general method we
adopt throughout: a balance between the freedom of variation and the
constraint of selection in a hierarchically organized system of control. In
this case the multiple components of the hierarchy are foundations,
axiomatic sets which reciprocally and irreducibly support each other. While
each component is itself a stable foundation, the overall metasystem is
a-foundational: the choice of an axiom set is ultimately either arbitrary
or non-theoretical (pragmatic).
In this sense, the philosophy we propose is
anti-foundational. Yet a constructive philosophy can be considered
foundational in the sense that it takes the principle of constructive
evolution itself as a foundation. This principle is different from other
foundations, however, because it is empty (anything can be constructed,
natural selection is a tautology), but also because it is situated at a
higher, "meta" level of description. Indeed, constructivism allows us to
interrelate and inter-transform different foundational organizations or
systems, by showing how two different foundational schemes can be
reconstructed from the same, more primitive organization.
