The Krokers are an industry unto themselves...this is from their mammoth monument, Ctheory:
http://www.ctheory.net/printer.aspx?id=383
The interview is, technically speaking, utterly motherfucking massive, but here’s what grabs my attention on a late Saturday night.
I wish I could recommend reading this whole interview but it’s mostly excessively boring theoretical discursion about Marxism, just a bunch of name-dropping for more than 50% of this…
De Landa: One of the themes of the War book was the tendency of military organizations to get “humans out of the loop.” Throughout the book (and in my only live lecture to the military) I have very strongly criticized this, urging for the lowering of decision-making thresholds so that soldiers in the field with access to real time information have more power to make decisions than their superiors at the Pentagon. (This theme, of course, goes beyond the military to any kind of centralized decision-making situation, including economic planning.) The problem you raise is, I believe, related to this. If all technological decisions are made centrally without thinking of issues of maintenance in the field, and if there is no incentive for field soldiers to become “grease monkeys” or “hackers,” the army that results is all the more brittle for that. Flexibility implies that knowledge and know-how are not monopolized by central planners but exist in a more decentralized form.
De Landa: All models, by definition, simplify things. Contagion models can be very useful to study certain propagation effects, whether these are fads, fashions or ideas. Can they also be useful to study the propagation of affect? We can’t tell in advance. What is important to see is that even if they turn out to be useless to study violence that does not affect their usefulness in other areas. Also, contagion models differ in the detail with which they portray agency, from completely mechanical models with no agency at all (a homogeneously socialized population) to models in which some form of virtual agent is included. But the key problem is that no one agrees what agency is: must it include some form of rational choice, and if so optimizing or satisfying rationality? Should all psychological effects be eliminated and only inter-subjective effects taken into account? How socialized and obedient should we assume agents to be, and should these qualities be modeled as homogeneously or heterogeneously distributed? Most of these issues have nothing to do with computers and will haunt any modeling effort however informal.
CTHEORY (Mallavarapu): To ask a related question… In your introduction to War in the Age of Intelligent Machines, you take care to point out that your use of the idea of self-organization is “more analogical than mathematical.” What are the problems and possibilities that arise from the use of analogies from chaos science to describe social phenomena?
De Landa: That remark is a disclaimer to draw attention to the fact that one does not have the legitimate right to postulate an “attractor” until one has some mathematical evidence one may be lurking there. (This, by the way, does not imply possession of a formal model. One can infer the presence of an attractor from an analysis of time series, such as those we have for production prices in economics, or voting patterns in political science). The remark in that book was to the effect that I did not model warfare either directly or through time series. That’s the only way one can use these ideas non-metaphorically. (Then, of course, one has to show evidence that the actual physical or social system has an attractor by giving it a push, for example, injecting some energy or spending some money, and checking whether the system returns to its previous state after a while).
An interesting aside on complexity and pet theories:
CTHEORY (Selinger): A few weeks ago I heard Stephen Wolfram give a lecture based on his book A New Kind of Science. There was a performative element to this talk which I found striking. Have you read this recent book or any of his published material?
De Landa: Though I have not read his recent book, I think his claims have to be wildly exaggerated. In fact, it would seem that each famous scientists in this field would want his own theory or model to be the center of it all. Ilya Prigogine wants everything to be “order through fluctuations”; Roy Bhaskar wants it all to be about self-organized criticality (his sand piles with fractal avalanches); Stuart Kauffmann wants it all to be about “the edge of chaos”, and now of course Wofram wants it all to be about this one CA rule. To me this denies the basic insight of nonlinearity, its plurality of effects. Enrico Fermi once said that to speak of “nonlinear mathematics” made as much sense as to speak of “non-elephant zoology.” In other words, the dichotomy linear-nonlinear is a false one: there are many nonlinear effects and linear ones are one special case of it (so the word nonlinear should eventually disappear). Whenever one opposes chaos and linearity one is bringing back the dichotomy. And so one does when one favors one particular phenomenon at the expense of the large variety of others. Wolfram has done very good work (classifying cellular automata, for example) and his claim to have discovered a special rule is probably serious. But so are the claims by the other scientists I just mentioned.
