Adaptation and Viability

Tuesday

---

The Tacoma Narrows Bridge completed in 1940 suffered catastrophic failure as a result of aeroelastic flutter.

A few more words about failure, if you will. Some days ago in Complex Systems and Complex Failure I wanted to make the point that, while failures in complex systems may have simple beginnings, the actual collapse of the complex system is as complex as the system itself. There is a sense in which this is logically, even tautologically, true. A complex system can’t be said to have experienced catastrophic failure until it has been compromised across a broad range of functionality.

Here is my formulation from Complex Systems and Complex Failure:

“Complex systems fail in complex ways. Moreover, the scope of a catastrophic failure of a complex system is commensurate with the scope of the complex system. This is easy to see intuitively since a catastrophic cascading failure in a complex system must penetrate through all levels of the system and encompass both core and periphery.”

There are many ways that this might be formulated. I regard the above formulations as tentative, since I haven’t yet thought about this enough to have converged upon an optimal formulation. But I’m sure you get the idea.

Ultimately, the failures of complex adaptive systems are as interesting as the complex adaptive systems themselves, and so these failures merit a careful study. The theoretical justification of the study of catastrophic and cascading failures is not difficult to seek. When searching for materials related to the failure of complex systems I found a lot of intuitive formulations related to the fact that one learns through failures. I believe that this is true. However, one must be careful because one can overstate the case for failure.

Theoretical failure is refutation, and Karl Popper famously made a virtue of refutation by asserting that the difference between science and pseudo-science is that scientific hypotheses can be tested and falsified, whereas non-science cannot be falsified. Thus falsifiability — i.e., theoretical failure — is a theoretical virtue. A theory that cannot be falsified also cannot be much use to us as knowledge.

However, if we look closely we can discern that there are different varieties of falsification. To give Popper an exposition in Kuhnian terms, we can distinguish between falsification in normal science and falsification in revolutionary science. We could also call this local and global falsification. With local falsification, the bulk of the theory remains untouched, and we need only tweak the details in order to set matters right. With global falsification, the bulk of a theory is shown to be false, and a new and perhaps revolutionary theory must be formulated to take its place.

There are pragmatic analogues to these theoretical forms of failure. That is to say, there are “normal,” local failures that do not call a given (social or industrial) infrastructure into question. A flat tire is like this. Most people don’t give up on driving because of a flat tire. But there are “revolutionary” or global failures of infrastructure. For example, chronic traffic congestion may cause someone to give up on driving.

As there are different varieties of failure, so too there are different responses to failure, both on the part of the systems that fail and on the part of those who stand in some relation to the systems that fail.

Complex adaptive systems fail when their ability to adapt to changed conditions is impaired. However, the kind of failure will determine the kind of response that is necessary to the survival of the complex adaptive system. A local failure will call forth a local adaptation; a global failure will call forth a global adaptation. Put otherwise, faced with a local failure, a complex adaptive system will locally adapt or the failure will cascade and perhaps become catastrophic. Faced with a global failure, a complex adaptive system must adapt globally or be annihilated.

These ideas can also be formulated in terms of Toynbee’s challenge and response model for civilizations. If a civilization is faced with a local challenge, it must adapt locally or the failure will cascade; faced with a global failure of its institutions, a civilization must response globally or become extinct.

Change is the price of historical viability; complex adaptive systems “succeed” by adapting to changed conditions, but when the conditions are globally changed a global response is necessary, and a global adaptation means that the complex adaptive system has been changed beyond recognition. In other words, the “success” of a complex adaptive system extrapolated over time means that, in order to survive, a complex adaptive system must change its nature, and if it is an entity of a wholly changed nature that ultimately “survives” as a result of adaptation, it is at least arguable that the entity of the original nature has not survived and therefore the survival strategy of complex adaptive system has failed.

This paradox of survival is widely applicable to real world entities, and bears a strong resemblance to what philosophers call sorites paradoxes. Sorites paradoxes are paradoxes of identity: how many hairs must you remove from a head before it is bald? How many grains of sand must you pile up before they constitute a heap? Do the individual grains of sand lose their identity as individuals in order to become incorporated into the identity of the heap?

There are answers to these questions, but there is no one, universally accepted answer. Any one answer to a philosophical paradox assumes a particular metaphysic, and those who subscribe to another metaphysic will find themselves at odds with such an answer. It would seem that complex adaptive systems are paradoxical, therefore open to distinct metaphysical expositions. Thus the existing literature of complex adaptive systems ought to be interrogated for its metaphysical presuppositions. I will save any such inquiry for another time.

. . . . .

. . . . .

. . . . .