Cybernetics and China’s Population
By: Lyle FearnleyIn her recent book Just One Child, anthropologist Susan Greenhalgh traces the origins of China’s infamous ‘one child policy’ to a group of defense scientists who specialized in cybernetics and ‘control theory’. Her book is unabashedly both an analytic project and a criticism of the roots of the policy, that is to say, she begins from the claim that the ‘one child policy’ is an ethical bad and uses her analysis to discover what led to such unethical policy. Her claim that missile scientists were at the route of the policy, in other words, is a denunciation of a particular application of ‘natural science’ in government policy. First, I will tell a little bit of her story, which is incredibly interesting in its resonances with some of the topics we have been following in VSS. Then, I want to show how the perspective we have developed in the VSS research collaboration can productively engage as well as put in perspective her denunciation of cybernetic planning.
As she tells it, the problem of ‘population’ first became a fundamental aspect of PRC planning in the mid-1970s. As the anti-intellectualism of the Cultural Revolution was winding down, a number of social scientists began to collect moderate amounts of data regarding national population numbers. Prior to this, there were only estimates: a classic statement, often repeated, was that China’s population was ‘around 800 million’. So a few sociologists begin to construct population as a problem, and argue that some form of population control should be implemented. But, Greenhalgh argues, the social sciences had been massively discredited and dispersed; only natural science, and most importantly, defense or weapons science had been actively supported. In the late 1970s, one Song Jian, cybernetician working on missile logistics and control, became interested in the problem of China’s population. Song became interested in population control primarily through interactions with the Club of Rome and its claim that global human population was reaching ecological limits of sustainability. Notably, in the Club of Rome publication The Limits to Growth used cybernetic techniques and control theory in order to make projections of population growth. Song adapted these techniques, and employed his own expertise, developing a number of simulations and projections (using early, and at the time rare in China, computers) which projected an imminent population crisis. Greenhalgh argues that it was this epistemological reframing that led to the policies of ‘only one child’ (rather than a more moderate policy entitled
‘later, longer, fewer’). For Greenhalgh, this was a misapplication of a natural (and military-oriented) science onto the object of the human population. As she puts it, “their specialty was control theory, an engineering approach to controlling the behavior of machines—not humans” (125). Two critical engagements I think our VSS collaboration can make with this argument. First, this was certainly not unique to China. As we have shown in a number of papers and blog posts, the migration of military logistics, techniques, and technologies to previously ’social’ fields is extensive and longstanding (see, for example, Collier and Lakoff ‘On Vital Systems Security‘; Lakoff, ‘The Generic Bio-Threat’; my own working paper, ‘Pathogens and the Strategy of Preparedness’). Moreover, cybernetics (or cybernetics-like systems thinking) we have tracked (particularly in Brian and Onur’s work; also Collier, ‘Enacting Catastrophe’) as it is applied first to the threat of nuclear attacks, then increasingly to other ’social’ fields such as energy systems or labor. See also Stephen’s recent blog entry on the use of cybernetics in Allende’s socialist Chile.
A second point is that we might question placing cybernetics firmly within ‘natural science’ against ‘human science’ or ’social science’. The cases above show how, in the US, cybernetic techniques were applied both to social and ‘natural’ objects. From what I understand, there is nothing about cybernetic techniques that (any more than statistical techniques) determines whether its objects are humans or missiles. What may be significant, however, is the form it gives to those objects; that is to say, whereas statistics and probability were fundamental to the emergence of ‘the social’ as an object of government (see Rabinow, French Modern), cybernetics gives human social activity another form. As I think we have argued in VSS throughout, this doesn’t call for denunciation out of hand but engagement with the limits and possibilities of such thinking.
May 26th, 2008 at 6:42 am
Thanks a lot for this post Lyle. It is really interesting stuff.
I agree with everything you say, particularly concerning the question of whether cybernetics, or systems theory, or whatever it is that we are calling that set of techniques that came out of World War II and the early post-War period can be associated with “human” or “inhuman” systems. Of course, it is not right to say that they were initially used to model systems that were only machinic. They were used, in part, to model human-machine interactions in systems. The point is moreover that human-machine distinction, with systems theory, becomes obsolete. This is a point that Donna Haraway made a long time ago.
There is another point here that Onur and I have been talking about quite a bit, namely the question of whether “probability and statistics” should be associated with knowledge of “the social” and its distinctive form form of analyzing an archive of past events. What I try to indicate in “Enacting Catastrophe” — and what Onur and I have discussed in a number of other contexts — is that this kind of archival statistics and the probability claims it entails really just constitute a special case but absolutely not a paradigmatic case of how it is that statistics can be used to made probabilistic statements. Part of what is interesting about systems theory is that it does, in fact, use statistics and very much depends on probabilistic statements, it is just that these statements do not always (often do not) use analysis of an archive of past events. So, for example, one pervasive use of Monte Carlo techniques is to “simulate” an archive — to create a whole series of possible worlds — and then to conduct statistical analysis on that simulated archive, and to derive probability statements on the basis of it. This became standard in catastrophe modeling, for example, in the middle 1980s.
Maybe Onur can weigh in on this as well.
May 28th, 2008 at 11:09 am
Hi Lyle,
This is indeed very interesting, and it seems like it can be an anchor point to discuss what happens to biopolitics in the second half of the twentieth century. In a way, it relates back to our discussion at Berkeley this January on how different is vital systems from the 19th century problematization of life and emergence of biopower. One question that I find interesting is what happens to the conceptualization of ‘population’ once you can map this object with new and novel techniques in the 20th century with the help of computers. Along with simulation models, network analysis seems to be one of the candidates for doing this (indeed some of Columbia sociology’s Peter Bearman’s work on transmission of sexually transmitted disease is one interesting example. As a student of Harrison White, who brought network analysis from micro-physics to sociology, Bearman ends up conceptualizing population as a system in the form of a complex and dynamic network.) I believe such a conceptualization is light years away from the early population experts such as Malthus. Can we say that today population is conceptualized as a vital system? If so, how do we know it is one? This seems to be a question that is homologous to thinking the economy as a vital system.
I also agree with Stephen’s point about the obsoleteness of the human-machine distinction. It is fascinating how a generation of historians of science have been obsessed with automation and have argued in a similar vein along Greenhalgh that this was a suspicious and potentially evil conspiracy of the military industrial complex. see Galison, Hughes, Mirowski; especially Paul Edwards’ Closed Worlds is big on this narrative of the fantasy of automation. I would actually argue that the attempts at automation resulted in an unexpected way leading to the autonomy of the human in an ironical way. Both Galison’s piece on the anti-aircraft gunners and Hacking’s discussion of Pierce indicate that these techniques brought into light how Newtonian deterministic mechanics were fundamentally faulty in their deep ontological assumptions about the state of nature. What both Pierce and the cybernaticians of Galison discovered was an irreducible ‘error’ in the way in which things are supposed to be working: the processes they were trying to model and solve turned out to be fundamentally probabilistic. You could only know where the pilot will be in the next move up to a certain degree of certainty and probability. Same was true with Pierce’s pendulums.
What I find really surprising is how much a figure such as Hayek is standing on top of this seemingly non-related genealogy. As I had mentioned in a previous response to an entry, Hayek’s conceptualization of the market was fundamentally based on ideas revolved around cybernetics and feedback loops. This was how he had responded to Oskar Lange, soviet reform communist, in the famous calculation debate of the 30s. In a recent correspondence with Stephen, we realized that this profound level in his argument was not noticed until the mid-80s. Even the bourgeois economists had thought that there was no difference between the argumentation and style of reasoning between Hayek and von Misses against the soviet economists who were favoring ‘rational’ centrally planned economy. They thought even Hayek had argued from moral principles and not based on rational reasoning. Whereas Hayek’s reasoning was deeply saturated around cybernetics principles; interestingly enough his interlocutor Lange had invented a new schema for economic planning called ‘economic cybernetics’. He thought he could could plan a mixed economy on the bases of electrical engineering models. Then, it should not be surprising that economics today turned into a science of decision making, as our ideas around decision seems to be the products of these unexpected experiments with systems and the failure of deterministic mechanisms of Newtonian Physics. After all, it is not so hard to see the affinity between von Neumann (game theory & optimization), Dantzig (linear programming) and Hayek (markets) as they all seem to be standing on a common ground and sharing something common that I am not sure we come to understand well yet. At least, we cannot really tell what the new figure (will) looks like.
In this respect, I share the same conclusions as you do: that these names, cybernetics, operations research, industrial engineering, systems analysis etc… do not have much meaning in themselves. Rather as Stephen and I have recently discussed they are the elements of a new logic and style of reasoning that is being assembled in response to new set of problems that were either inconceivable, and therefore unnoticed, before or could not pass the threshold of problematization.
I think there is also much to be discussed on the topic of statistics and probability. Maybe starting from Hacking’s work on statistics and probability would be a good idea. I think Stephen’s Enacting Catastrophe is a very promising start in terms of translating Hacking’s history of ideas (even though he says it is archeology, I am not so convinced yet) into a type of analysis that goes beyond the history of ideas (be it archeology, genealogy or problematization).
May 28th, 2008 at 12:00 pm
This is intriguing, but confusing to me. Intuitively, I think there is something to greenhalgh’s connecting up population control and cybernetics– but agree that the label is a marker for something else. But i also think it distracts. The real object here are the “population engineers”– the people involved in the construction of systems of population management and policing… perhaps something more at this level: Saul Halfon’s The Cairo Consensus? Michelle Murphy at University of Toronto also knows a ton about this stuff, especially in Bangladesh (I invited her to comment).
I just mention it because I have a strong suspicion that there is a large gulf between the legitimizing force of a label like “Cybernetics” (both for china in the 1970s and for Greenhalgh in anthropology today) and the actual practices of population control as they played out. Another interesting comparison here is Fred Turner’s book From Counter-culture to Cyberculture which also makes the claim of a transfer of cybernetics–in this case from Norbert Weiner to the planning department of Royal Dutch Shell around the 1973 oil crisis via Stewart Brand and the Whole Earth catalog. But it’s eminently clear that it wasn’t any actual set of practices that was transferred so much as a kind of systems-ecology-scenario-seeing mantra that fit in with Brand’s worldview. Whatever cybernetics is doing in that case (and maybe in the population case) it isn’t at the level of practices but at the level of justifications and ideological argument…