Copyright Sociological Research Online, 1999


Silicon Second Nature: Culturing Artificial Life in a Digital World

Stefan Helmreich
University of California Press: Berkeley, CA
1998
0520207998 (hb)
US$29.99 (hb)
xii + 314 pp.

Order this book

This book is fascinating and frustrating in equal measure. It is fascinating for two reasons. Firstly, it deals with the study of Artificial Life, a research area of considerable intrinsic interest and importance. Artificial Life is the study, using computational methods, of 'possible Biologies', the investigation of interactions between simulated 'creatures' and 'environments' that may not exist on earth and possibly could not exist, given scientific laws as we currently understand them. One important implication of this approach, which has become part of the Artificial Life manifesto, is that studying possible worlds gives us a far better understanding of the world as it is. Despite the possibilities of the comparative approach, there are certain phenomena, like the evolution of language, for which history appears to have provided only a single example. (In other cases, such as the evolution of life, we are not quite sure how many independent examples we have got.) In such circumstances, consideration of possible worlds is the only way to broaden our knowledge of process. The relevance of this aspect of Artificial Life for sociology is clear. There is a long tradition of naturalism and essentialism in social science, often supporting highly questionable politics. If we are limited to the world as it is, there is a strong tendency to mistake what is universal for what is necessary. The study of possible worlds allows us to challenge this inference. (This approach also informs the field of social simulation or the study of Artificial Societies (Gilbert and Doran 1994; Gilbert and Conte 1995; Journal of Artificial Societies and Social Simulation 1998). The second reason why Helmreich's book is interesting is that he has used anthropological techniques to study the Santa Fe Institute, a key institution 'producing' Artificial Life research. SFI is unique in a number of respects. It is not closely tied to an existing university, it is profoundly inter-disciplinary, it has very rapidly become recognised as the leading centre in this field and it constitutes, through extensive use of summer schools, visiting fellowships and electronic communication, an 'invisible college' of a very modern kind. The combination of interesting subject matter and a key research site ought to be a foolproof recipe for success.

Unfortunately, as the old saw has it, fools are so ingenious. This is not a gratuitous insult. This is the first book I have ever read where I wished for fewer references. As the blurbs suggest, it is clear that the author is very widely read, fecund with ideas and has a refreshingly clear and easy writing style. (This last fact is particularly valuable during the necessary explanations of difficult and technical ideas on the borderland of biology, computing and philosophy.) The key difficulty with the book is that it almost wilfully resists any clear methodological or theoretical stance that could guide the selection of subject matter, order the analysis of data and provide an overall sense of direction. In a manner of speaking, this approach is a methodological choice, a bricolage of narratives, research strategies and critical social (de)construction, but the overall impact of the book does not serve as a very good advertisement for it, at least in the opinion of this reviewer.

It is not that the book lacks a thesis. In fact it has three, all of which are interesting. The first is that, in practice, the space of possible worlds explored at Santa Fe is considerably narrowed by its cultural, spatial and historical position. The second is that old myths and narratives are not simply being reiterated but also recast in Artificial Life research. The third is that, in the longer term, this process may also work in the opposite direction, with concepts from Artificial Life informing (and perhaps reconstituting) everyday notions such as 'life' and 'nature'. Unfortunately, the absence of a clear methodological stance often obscures the connection between particular discussions and the overall treatment of these theses. The lack of focus reproduces itself at a number of different levels.

Firstly, it is very hard to get a grip on what the author means by a distinctively anthropological study. In consequence, it is difficult to know what evidential criteria to apply to the material presented. Perhaps this is the intention. Indicatively, the term anthropology is used interchangeably with ethnography throughout. The author cursorily dismisses the folk view of anthropology as the study of isolated, geographically located tribes possessing shared and stable understandings (p. 37) without reflecting on the extent to which the effectiveness of anthropological (or ethnographic) techniques relies on the assumption of shared practices and perspectives in a bounded space. As a trivial example, Helmreich discusses the implications of a picture (p. 32) to be found in the stairwell at SFI. He cites this picture as evidence of a 'white imagination' at work. In fact, the survival of the picture could say just as much about a lack of observation and reflection in the workaday environment (particularly among transient academics), the vagaries of institutional decor or the attempts of some individual to project a corporate (or other) image. In any event, there is no reason (and none offered) to suppose that such institutional details are 'owned' by some, or indeed any, of the occupants of the building.

In the same way, Helmreich sums up his methodological stance of 'implication' in the Artificial Life community in a single sentence which approves a quote from Haraway. He says nothing about the extent to which a white, male and (by inference) middle class college graduate can expect to make effectively distanced observations concerning the cultural practices in a society dominated by white, male and middle class college graduates. It is not as if anthropology has given no attention to the notion of difference! Indeed, the author seems to say rather little that is particular or surprising about the SFI community, as distinct from other societies of computer programmers, research scientists or middle class men. (Where the book is different and innovative is in discussing the contents of the debates, although this raises problems of its own, discussed shortly.)

Secondly, the book covers an extremely broad canvas (owing partly, in fairness, to its ground breaking nature), but this means that the reasoning behind any single assertion is seldom supported by much evidence or analysis. This withdraws yet another evidential check, that of triangulation, from the reader. With the exception of the debate about whether simulated environments in computers can be considered as in some sense 'real' worlds capable of supporting 'life', very few of the issues thrown up are pursued in adequate depth, at least in my opinion. (In this at least, the book follows the old anthropological tradition of hurried jottings on first contact, perhaps before a 'native' form of life is swept away forever. Presumably, this threat was not what motivated the approach to study at Santa Fe.) There is enough matter in some of the issues raised - like the cultural and social implications of a predominantly male group working enthusiastically on a form of creation that both literally and metaphorically doesn't require women (pp. 113-122) - that a chapter could easily have been written on these alone, with the benefit of far greater focus on detailed analysis of relevant anthropological literature and provision of supporting evidence. With a sharp methodological focus, breadth can be exhilarating. Without, it can easily descend in superficiality and confusion.

Finally, the lack of methodological orientation makes it hard to know to treat Helmreich's reporting of the Artificial Life debates. (How well do ethnography and anthropology serve us as tools when beliefs are, by their nature, highly contested?) Sometimes the tone of the book is very close to popular science journalism. At other times, the author seeks to participate in the debates (usually critically) without engaging in the tedious task of actually doing any Artificial Life research for himself. For example, Helmreich rightly points out (pp. 83-88) that simulators routinely talk about being able to 'play god' with simulations, meaning an omnipotent, omniscient, unitary Judeo-Christian god. But what follows? Will researchers resist any simulations designed with different cultural pre-conceptions? I doubt it, and little evidence is provided that this might be happening. Similarly, will Artificial Life prescribe notions of theology among the wider public? If this is occuring, it would be nice to hear more about it. Without strong supporting evidence, much of the argument suffers from post hoc propter hoc: sexism results in all male white communities, this is such a community, therefore, somewhere, sexism must be at work. I sometimes got the feeling that Helmreich wanted his comments to be taken seriously both as standard critiques within Artificial Life, but also as meta-critiques, immune to criticism from within the community he is studying. There is an ironic resonance between this ambiguity and discussions of 'playing god'.

Surprisingly, the one kind of analysis that is conspicuous by its absence, in view of the strong ethnographic tradition in the Sociology of Scientific Knowledge, is any detailed account of the actual process or history of particular research at SFI, the mixture of seminars, programming, inspirations, discussion papers, "bull sessions" and other public or private activities that might construct the "Santa Fe view" in a particular context. (Such an approach would be particularly interesting given the existence of independent Artificial Life narratives in Europe and Japan.) The history of a single (collectively developed) simulation (like Tierra) might also have provided much needed focus. This would have increased the value of the book considerably, since one criticism of ethnographies in SSK is that they sometimes give insufficient weight to the relationship between research content and research practices.

To sum up, this book can be broadly recommended to non-specialist social scientists interested in the field of Artificial Life, its arguments and some aspects of its culture of production. It is full of suggestive ideas, clearly expressed and makes enjoyable reading. However, the methodological ambiguities mean that the analytical (and even empirical) contribution of the book is frustratingly less than it might be.

Edmund Chattoe
University of Surrey

References

GILBERT, NIGEL and CONTE, ROSARIA (editors) (1995) Artificial Societies: The Computer Simulation of Social Life. London: UCL Press.

GILBERT, NIGEL and DORAN, JIM (editors) (1994) Simulating Societies: The Computer Simulation of Social Phenomena. London: UCL Press.

Copyright Sociological Research Online, 1999