[Note: what follows is my Gravity’s Rainbow, a metatextual meditation on the subject of entropy… my fancy way of saying that it starts fairly structured and then degenerates into a lot of notes and questions I’m asking myself. One of my personal purposes for this newsletter is to let go of perfectionism and just put stuff out there even if it’s unfinished, tentative, sketchy, or otherwise imperfect. So, that’s what’s happening with the marked shift in tone and deterioration of the prose structure after the block quote towards the end.]
*************************************************************
For some time now I have been vaguely alluding to the “discursive power of data” (I think we can all agree that data and numbers do have something like “discursive power”) and trying to locate its origins somewhere in the productive structure of the real economy, i.e., being a good little Marxist. This strategy has been unsatisfying, because it limits the discursive power of data to a ratifying or legitimating role vis-a-vis some kind of ideology that is encoded in and enacted through the productive structure. Now, however, I think I have been focusing on the wrong aspect, and I have (record scratch) Jean Baudrillard to thank for that. In thinking so much about the social construction (production) of data, I have been trying to link the social processes of data construction with the various functions that data perform out in the world, but sort of missing the bridge to be able to do it. Is it because statistics is eugenicist? Not exactly. Is it because of “methodological individualism?” Also not exactly. All to say, in thinking so much about the production of data, I have neglected to think about its consumption. Borrowing some language from semiotics and a lot of conceptual framework from Baudrillard, I want to venture the hypothesis that the discursive authority of data issues from the sign function of data and its operation in the symbolic logic of consumption.
This is not a base/superstructure argument. The symbolic economy and social logic of consumption (which Baudrillard develops extensively, and which I will not recapitulate here) is not some kind of epiphenomenon of the productive economy; rather, production/consumption are two chambers of the same heart, two aspects of an organic whole. To borrow from Ursula Le Guin, consumption is the “left hand” of production. Here, social and psychic logic operate according to the code of signs. (In semiotics language, the sign is the atomic unit of meaning composed of signified – a thing itself – and the signifier – the concept referring to the thing.) The meaning or value of signs depend on their relationships to other signs. Signs perform particular “signifying” functions within the symbolic and psychic logic of consumption – this is one of Baudrillard’s major (early) hypotheses.
We tend to think the important thing about data is the content. But data, inference from data, and the whole mode of statistical reasoning as form are powerful communication media themselves. (Symbolic exchange?) Data as a sign – what is signified? The concepts to which statistical objects and the results of statistical analysis refer can really be anything; data/statistics signify a method and an approach. This is encoded in the technical structure of data and products of analysis – born out of a different phase of capitalism than the one we’re in now, the industrial revolution and subsequent movements of anthropometry and eugenics. Literally (as Baudrillard liberally borrows from McLuhan), “the medium is the message.” This is a neater way of referring to what I have been calling “the social construction and function of data.” “The medium is the message” helps understand how data functions in sign logic as a consequence of its technical structure and organization, both at the level of communications (statistical discourses we use) and at the level of the logical/mathematical substructure of statistics itself – statistics as communications technology, different simulacra, like the “population.” The differentiating function of the code of signs is important here.
How do data and mathematical objects function in the system of signification that makes up, for example, the health disparities discourse? Let’s consider the adjusted rate as an example. Adjustment signals a statistical procedure of normalization, of making different kinds of quantities “comparable” in their sub-numeric dimensions (the size of the reference population, how the data were collected, and so on).
As another example, the p-value. As a piece of mathematics, it’s really ugly. There’s a whole crisis industry dedicated to the deconstruction of the p-value (how do you interpret a given p-value in the context of “researcher degrees of freedom,” p-hacking, pearl-clutching over inbuilt distortions due to the mathematical structure of the object, like how a very large sample will always yield a very small — and thus “significant” — p-value). But p-values are tenacious in biomedical research, probably not least because of their sign function. A “significant” p-value is a sign indicating “reality” of the underlying “difference.” The sign is “this finding is “real.” Here we have the epistemological enterprise, modern science, searching for noumena, and finding them – under the sign of statistical significance. The p-value as sign – produces differentiation that is statistically sanctioned. “Are these two groups/sample means different or not?” “Is this difference real?”
As another example, David Leonhardt… what is he really doing? He’s trying to manipulate a bunch of form signifiers to make an argument about content. We all got so mad about that column about vaccines and unadjusted rates but… the joke is kind of on us, he’s just manipulating the signs. We are the naive ones, for thinking that he was trying to make some substantive argument on the plane of the real (to use Baudrillard’s terminology), or in the realm of content and action. From page 51 of my library copy of The Consumer Society:
But the figures do not speak for themselves, and they never provide any counterargument. Only interpretations speak, sometimes to one side of, sometimes against, the figures.
Now, I’m not going to pretend to have a total grasp on what good old J.B. was thinking, or to pretend at a synthesis of the full sweep of his thought. However, I have caught some vibes, and want to hazard a few embryonic ideas based on these vibes. Such as: “personalized medicine” as a hyperreality phenomenon? Individuation – is it necessary to the symbolic order? Consumption of personalized bits of information… who does the work of it? Are parts of that work being automated and to what social, ideological, or material end? Or – the logic of automation is to combinatorially increase the personalized bits of information? Personalized medicine is trying to operationalize this at the level of the body but it’s again a mathematical language of signification within a defined hierarchy. (Actual human bodies are way too different, way too variable and unconformable to the hierarchy.)
If Baudrillard can be extended here, I think he’s right about how statistical indicators assume real functions, a form of hyperreality. (This is what “listening to the data” really means; Sandro, with his ham-fisted comment equating masks and MAGA hats, is ironically kind of right that data function as signs.) “Biopolitics” is one version of this, “science communication” is another (the form of science learning corresponding to consumption). Numbers in the field of consumption – this is where data do in fact “speak,” by functioning as signs.
COVID as the consumption pandemic, consumable pandemic, “The Accident.” Materiality of sign/signification; again, I’ve only been thinking about productive forces per se in kind of an old-fashioned way. Things that are self-evident in the material plane or on rational terms – not so in the sign economy, which has a different, unconscious logic (like a Grateful Dead song, psychic logic). Connected to the technical organization of the media, the institutional organization of science as well, and of course to the organization of the productive real economy. Vehicle for the myth of formal equality? (See below?)
Knowledge/data are “produced.” Even socially produced – okay! But the missing link for me is how this sits within “production/consumption” as a social order. Who are consumers of facts, data; how does this order the signs; what groups are ordered/signified; what is communicated in terms of social difference/differentiation and hierarchy? (Personalized medicine and personalized risk assessment here, ultimate tool of social differentiation in the face of COVID! All we need to look at to prove that data operate as a sign as well.) “Consumption is an order of significations” – p. 79. Personalization as a “constraint of differentiation” (p. 87). Differences effaced in the real to be recodified and ranked as signs. P. 94: “What is politically effective is the creation not of a situation in which contradiction is replaced by equality and equilibrium, but of one in which contradiction is replaced by difference.” The concept of risk/risk scores/risk stratification fits here.
Science is socially produced, yes! AND socially consumed! In the consumption function, we see the importance of differentiation – replacing contradiction by difference. COVID discourse… “the solution to social contradiction is not equalization, but differentiation.” (Health disparities: the contradiction inherent in America’s pandemic response recast as difference; same especially with “medical vulnerability,” “underlying conditions,” etc.)