3.1 The Original Model

Document: Modelling Autopoiesis: Harder Than It May Seem!

3 A Case Study: Computational Autopoiesis

3.1 The Original Model

In the early 1970's (well ahead of the modern re-birth of ALife) two Chilean biologists, Humberto Maturana and Francisco Varela, were struggling with the question of how a living system (such as a bacterium) is different from an arbitrary assemblage of its raw molecular constituents. More precisely, what is it that is so special about the organisation of the molecular constituents of a living system that gives rise to the living phenomenology?

Their tentative answer was that living systems are distinguished by a special class of organisation which they termed autopoiesis , from the Greek roots for ``self-producing''. Roughly speaking, an autopoietic system is a closed network of chemical reactions, such that every molecular constituent of the system is a product of one of the reactions in the network, and the system establishes and maintains a spatial boundary which serves to concentrate the reactants and sustain the network.

This concept is still somewhat vague and it is not clear just how much of the living phenomenology it can illuminate; but it is certainly a creative and original approach to this very difficult problem.

In any case, in an effort to make their ideas more concrete, Varela and Maturana also developed a ``minimal model'' for the kinds of components and interactions that seem to be required to support autopoietic organisation; that is to say, they described an abstract ``artificial chemistry'' in which autopoietic systems could both spontaneously emerge and sustain themselves. Finally, with another colleague, Ricardo Uribe, they developed a specific computer model or realisation of this artificial chemistry, and demonstrated that it did indeed support at least the basic autopoietic phenomenology.

Clearly this work falls under the general class of synthetic ALife work described earlier: there is a claim that certain life-like phenomena (autopoietic organisation) can arise in any system which realises a certain set of abstract ``chemical'' interactions. And equally, although this particular claim cannot be meaningfully tested against natural, biological, data, it is still perfectly scientific in the sense that it can be tested by the synthesis of independent realisations (computer models) of the same abstract chemistry.

Of course, all that remained was for someone to actually do such tests...

3.2 The Problems

3 A Case Study: Computational Autopoiesis

Document: Modelling Autopoiesis: Harder Than It May Seem!

mcmullin@eeng.dcu.ie