Chaos theory has a lot to bring to archaeology, but this synthesis is not without problems. These are dealt with superficially in Mayntz (1989), but in much greater depth in Francfort (1997), it is on Francfort's competent paper that I shall base this problematisation.
The first point raised is that of the difference between the systems usually dealt with using chaos theory, and their relative simplicity when compared to those present in ancient societies. The scientific models from which we take the principles at least have readily defined quantities with which to deal. The fuzziness of our discipline leads to the danger of us arbitrarily choosing entirely meaningless values for analysis, and hence not gain anything useful from the exercise. He also addresses the interesting point that we have such a free rein in deciding what, where, when and how we study, that whatever we do we are likely to find what we want. In this case however it is very unlikely that any real discoveries will be made either. He states that any society can be painstakingly modelled, but that if the model is constructed to fit the example so precisely that it is without fault, then it will only be a representation of what is already known. It is only if a model has predictive power, or allows us to see an underlying truth that it is of any real beneficial value.
As with all areas of science, chaos can only deal with the factors of society that can be quantified numerically. There are remarkably few valuable numbers that say something useful about society, these are things such as population, spatial distributions, various values related to length of human and animal bones, age of death and such, but there are many dynamic processes that are of interest to archaeologists that can only be numerically represented in the most abstract of terms, things such as technological level and social status.
We also have to address the question of what we want before trying to use chaos, as there are many aspects of the human repertoire to which it is totally blind. We must not treat these areas as unimportant just because the methodology we are using does not recognise them. Too often in the past archaeology has asked only the questions it is equipped to answer, rather than trying to answer the questions that are interesting and important.
Mayntz (1988) deals with 'rules' for the uses of principles outside of their native fields. Dealing explicitly with such rules is beyond the scope of her (and my) paper, but it is essential that we don't fall into the belief that we can pillage concepts from other disciplines recklessly; Einstein's equation relating mass and energy, E=mc2, has absolutely nothing to bring to archaeology even though it is part of the foundation of modern physics. I'm sure, if one tries hard enough, one could find some tenuous analogies with this, but it would be academically inadmissible rubbish. It is hard to tell where the borrowing of a principle is appropriate, and when, in the words of Mayntz (1998) we are being 'seduced by verbal analogies'.
We must also be aware that whilst a civilisation may appear to collapse, it is generally a specific part of that civilisation, and other aspects of the society continue, affected, but not destroyed by the change in the other parts of the whole. Francfort (1997) refers specifically to the example of the French revolution, the collapse of the ancien régime which utterly changed the political situation, but left scientific and industrial progress relatively unaffected. It is doubtful that the average farm-worker knew and cared much about the changes, and it is only history's traditional bias towards the elites that makes us rank the factors of a past society in terms of importance.
Central to the concepts of many chaotic systems is the possibility of equilibria (albeit temporary ones), where the system is stationary due to balances of positive and negative feedback. Can this be the case in a culture - is it even a concept with any meaning? Is it reasonable to co-opt a system into use when one of its basic precepts doesn't appear to cross between seemingly analogous disciplines?
A similar argument can be made for the concept of initial conditions. In all but the most theistic of views, we emerged, rather than simply popping into existence, so the whole concept of initial conditions can be seen as inapplicable to culture, as it is constantly in a state of flux - every part of the system has momentum as well position. To my mind this objection, levelled by Francfort (1997) and others has less weight than some of the others, as we can deal with the conditions at an arbitrary moment in time, and treat these as initial conditions (there is the epistemic point that we have no way of knowing if the whole world was created yesterday, including us, replete with memories - given this, any point in the past can be considered as an initial condition). Even if you accept the possibility of setting initial conditions in a social system, you have to address the almost essentially arbitrary nature of this point, as social phenomena do not just spring into existence, they have their roots deep in the recesses of time.
Given all these problems we must ask how we can tell when a specific application is a reasonable one and when it is a laughable misuse of barely understood principles. This is really a question without an answer, as there is no reductionist technique for validating any method, given the inappropriateness of this approach to archaeology, and the anti-reductionist nature of archaeology. The only way round this is to feel 'a subjective sense of satisfaction' (Francfort, 1997, p. 155) about a method, and by the reaction of one's peers.
Finally we must realise that although past societies are complicated, there is nothing explicitly in their nature that defines them as 'complex' (in the scientific meaning of the phrase). We must be aware that, although many of the traditional problems we have had in systematising are either addressed or side-stepped by the application of chaos, it is still essential as a scientific approach and as such still has the hermeneutic problems inherent in the previous, processual approach to solving archaeological questions. All this means is that whilst non-linear dynamics could bring a lot to archaeology, it is by no means a panacea.
© Joe MacLeod-Iredale 1998