While Theodor Adorno was exiled in Los Angeles, he wrote The Stars Down to Earth, a short book about the “pseudo-rationality” of mid-20th century American culture drawing on his study of “Astrological Forecasts,” the Los Angeles Times’s astrology column. Adorno uses the column to demonstrate how the capitalist culture industry in 1950s America sold quasi-scientific posturing to help an audience “excluded from educational privileges” nonetheless feel in the know.
Though the differences between Adorno’s time and ours are vast, his concept of pseudo-rationality still has something to tell us about the “rationality” of contemporary algorithmic culture, social media, and big data. The pseudo-rationality Adorno identifies in the astrology column shares key features with the data-driven “science” of forecasting that Nate Silver describes in his 2012 book The Signal and the Noise. For both Adorno and Silver, forecasting is a “down to earth” activity, a matter of applied knowledge that helps people figure out what to do in their daily lives. Both kinds of forecasting use profiles to explain the past and predict the future choices we will make: Both astrological signs and psychographic categories derived from demographic data (like, say, “college-educated women who tweet about Scandal and buy shoes online”) similarly forecast individual behavior. Adorno describes astrology’s capacity to fulfill “the longings of people who are thoroughly convinced that others (or some unknown agency) ought to know more about themselves and what they should do than they can decide for themselves.” Data-driven algorithms fulfill a comparable function, but now the secret to our identity and our future happiness and success lies not in the stars but in the cloud.
When personal identity is experienced and understood as a matter of forecasting, “the adage ‘be yourself’ assumes an ironical meaning,” Adorno claims. Such forecasting doesn’t predict the future; Adorno argues that it crafts the future in the image of “the established ways of life,” “the life of those whom it embraces.” The profiles are designed to produce the identity or frame of mind that they supposedly just describe.
Astrology, in Adorno’s account, maintains its reliability by avoiding easily falsifiable concrete details. Astrologers rely on their “knowledge of the most frequently recurring problems prescribed by the set-up of modern life and the characterological patterns [they] had frequent occasion to observe,” figuring out “a number of typical situations” that many followers might find themselves in. That description is equally applicable to big-data methods. The big-data algorithm, like the astrologer, observes patterns (of behavior, of interactivity, etc.) across populations and ties its forecasts to this input. Both kinds of forecasts “modulate” (to use scholar John Cheney-Lippold’s term) themselves to user behavior.
Whereas mass media tries to mass-produce standardized audiences, algorithmic media adapts to users—music-streaming services and Facebook’s Timeline algorithm “learn” what content optimizes individual users’ engagement and “tailor results according to user categorizations based on the observed web habits of ‘typical’ women and men,” Cheney-Lippold argues. Through this feedback loop of observation and adjustment, social media produce the identity categories—like “typical” men and women—it claims to merely observe.
Forecasting itself cannot be an exact science, and it doesn’t even aspire to be. Its goal isn’t accuracy, per se, but the avoidance of noisy dissonances with real life that call the reliability of the underlying ideological framework into doubt. What makes forecasting pseudo-rational is its offer of a nominally objective, systematic account of the “delusions” (in Adorno’s words) necessary to live in a capitalist society. Pseudo-rationality obscures the irrationality of social norms and makes what ought to feel outrageous seem completely down-to-earth. In its current big-data form, it rationalizes what Charles W. Mills, in The Racial Contract, calls the “cognitive dysfunctions” that make white-supremacist society fully functional. As Latanya Sweeny has shown in her study “Discrimination in Online Ad Delivery,” Google searches for names “racially associated” with black people return promoted results that imply the person you’re searching has an arrest record. On one website, a “black-identifying name was 25% more likely to get an ad suggestive of an arrest record.” This both expresses and reinforces anti-black racism. As Cecilia Esther Rabess argues in “Can Big Data Be Racist?,” Google AdSense “translates cultural clichés and stereotypes into empirically verifiable data sets.”
Though, as Adorno pointed out, a newspaper column could only “pretend” to tailor the content of each sign’s horoscope to users’ needs, wants, wishes, and demands, big data and social media overcome the limitations of mass media and allow forecasting to fully realize its capacity to tailor categories and output to observed user behavior. Scaled up in size and in processing power, big data could be the realization of what Adorno called “the potential danger represented by astrology as a mass phenomenon.”
One aspect of that danger is the “abstract authority” of astrologers, now mirrored by the black-box algorithms of the cloud. The opacity of the analytic method lends forecasts their appearance of authoritative objectivity. In “Astrological Forecasts”, Adorno notes “the mechanics of the astrological system are never divulged and the readers are presented only with the alleged results of astrological reasoning.” “Treated as impersonal and thing-like,” stars appear “entirely abstract, unapproachable, and anonymous” and thus more objective than mere fallible human reason. Similarly, as Kate Crawford pointed out in an essay about fitness trackers for the Atlantic, “analytics companies aren’t required to reveal which data sets they are using and how they are being analyzed.” The inaccessible logic of their proprietary algorithms is imposed on us, and their inscrutability masquerades as proof of their objectivity. As Crawford argues, “Prioritizing data—irregular, unreliable data—over human reporting, means putting power in the hands of an algorithm.” As Adorno puts it, “The cult of God has been replaced by the cult of facts.”
The apparent objectivity of the stars or the data cloud intensifies forecasters’ existing biases, allowing them to be passed off as neutral and matter-of-fact. Adorno argues that astrology rearticulates unfashionable superstitions in the occult, in mysticism, and so on, by presenting them in empirical rather than supernatural terms—star charts and tables, for example. Upgrading the medium in which they are expressed, obsolete social myths gain new life as apparent fact.
Similarly, big data can rearticulate “unfashionable” beliefs in, say, eugenics, by presenting them in supposedly more advanced and accurate empirical terms. Crawford points this out: When fitness-tracking devices (like FitBit) are “used to represent objective truth for insurers or courtrooms,” this treats their inconsistent and unreliable measurement of both what counts as exercise and what counts as a “‘normal’ healthy body” to pass as hard evidence. Fitness-tracking systems thereby build dominant ideas of health, embodiment, ability, and activity into the hardware, the software, and the algorithms embedded within them.
Forecasting repackages old-fashioned ideas as unprecedentedly objective knowledge, in part by sweeping inconsistencies under the rug of “individual responsibility.” To pass the social system off as an objective artifact determined by (quasi-)scientific processes, forecasting has to scapegoat “irresponsible” individuals for failing to live up to the terms of the forecast. Adorno writes that “the constant appeal of the column to find fault with oneself rather than with given conditions” is evidence of “the implicit but ubiquitous rule that one has to adjust oneself continuously to commands of the stars at a given time.” When forecasts end up being inaccurate, the fault lies not in the prediction methodology but the individual’s failure to adjust to the forecaster’s advice.
Big-data-enabled self-tracking foregrounds this same sort of adjustment and tries to make it seem really easy. As Whitney Erin Boesel argues in “Data Occupations,” self-tracking apps are “a single-serving ‘solution’ to a much larger collective problem”—they encourage individuals to fix themselves rather than collectively address problematic social norms.
Adorno explains how this can seem empowering but really isn’t: “The idea that the stars, if only one reads them correctly, offer some advice mitigates the very same fear of the inexorability of social processes the stargazer himself creates.” It reinforces the neoliberal myth of individual responsibility for social problems and misdirects our attention toward dumbed-down superficial solutions to complex social problems. For example, framing problems of political economy, class, and race as an “obesity epidemic” assumes both that obesity is a problem and that it is a problem that can be solved by modifying individual behavior (diet, exercise).
Though Adorno wasn’t thinking explicitly in these terms, Stars Down to Earth helps us see that neoliberalism’s ideal subject, homo economicus, embodies the same pseudo-rationality found in both astrology and big data, and that this economic pseudo-rationality is itself a trendy, supposedly more objective upgrade to unfashionable superstitions.
Homo economicus is the name for the view, held by neoliberal economists like Gary Becker, that, as Jason Read explains in “A Genealogy of Homo Economicus,” “everything for which human beings attempt to realize their ends, from marriage, to crime, to expenditures on children, can be understood ‘economically’ according to a particular calculation of cost for benefit.” Humans are beings who make choices, and every choice is ultimately a cost-benefit analysis: Everything I might do, from college education to cosmetic surgery, is an investment of time and resources, so I must decide which investments best serve my interests and give me relatively better choices in the future. Because, as Read notes, “the operative terms” in this theory of human life “are no longer rights and laws but interest, investment, and competition,” cost-benefit calculus updates old unfashionable beliefs in things like human rights with a supposedly much more objective and effective belief in the market.
Adorno finds this sort of economistic rationality throughout “Astrological Forecasts.” First, like neoliberal economic theory, it stresses the significance and efficacy of individual choice. Astrology’s “basic presumption” is “that everyone has to make up his mind at every moment.” Through the column’s pseudoscience, “instinctual needs contrary to the rule of rational interests appear to be commandeered by rational interests.” “The addressee,” for example, “has to ‘calculate’ very carefully his relationships with his family. He has to pay for the help and solidarity he expects.” The column advises readers, in line with a familiar stereotype, “to send flowers to one’s wife not because one feels an urge to do so, but because one is afraid of the scene she makes if one forgets.” In this example, “‘to be rational’ means not questioning irrational conditions,” like the fetishized commodification of love or heterosexual marriage, “but to make the best of them from the viewpoint of one’s private interests.”
Adorno echoes political theorist Andrew Dilts’s claim in “From ‘Entrepreneur of the Self’ to ‘Care of the Self’,” that this cost-benefit calculus itself has a price—it “sacrifice[s] any possibility of being critical.” Cost-benefit calculus works because everything is reduced to the common denominator of “private interest,” so the big-picture factors that would call into question why some choices seem better than others are necessarily factored out of this equation.
Like neoliberal economic theory, in which, as Read writes, “individualized, market-based solutions appear in lieu of collective political solutions,” “Astrological Forecasts,” Adorno writes, “implies that all problems due to objective circumstances … can be solved in terms of private individual behavior or by psychological insight, particularly into oneself, but also into others.”
And this is where big data comes in. It can provide us with the unprecedented—and supposedly more objective—insight into ourselves and others that we need to solve life’s problems. It can identify patterns of behavior, in individuals and across populations, that can then be monitored and managed. If you know enough about someone’s material, social, and psychological situation, their past habits and choices, you ought to be able to predict which future choices they will make, which alternatives will seem like the most economically “rational” ones from their perspective.
As Dilts emphasizes, as long as we can more or less successfully predict homo economicus’s behavior, he doesn’t have to actually behave rationally—his choices don’t have to be the result of well-reasoned, logical thinking. “Becker insists that economic analysis does not require ‘actual rationality’ at all, but is perfectly consistent with a wide array of irrational behavior. All that matters is if firms, households, or individuals act (drawing directly from Milton Friedman) ‘as if’ they are rational. That is so long as they respond to ‘reality’ and adjust their (even irrational) behavior, it is ‘as if’ they had in fact made a rational calculation.” Rationality, from the neoliberal point of view, is simply another word for predictability.
Homo economicus’s cost-benefit analysis is thus a type of pseudo-rational forecasting. Like the forecaster, who talks “as if he knew and as if the constellations of the stars provided him with satisfactory, sufficient and unequivocal answers,” homo economicus makes a prediction about the outcome of an investment, and as long as that choice appears, from the outside, to reflect “an unquestioning common-sense attitude” that pragmatically weighs costs and benefits in line with “accepted values,” then that choice is, for all and intents and purposes, rational.
In this way, homo economicus is a microcosm of big data: Both embody the same pseudo-rationality, a type of calculation that brings even the most irrational choices, behaviors, and patterns “down to earth.” Just as Astrological Forecasts makes its readers in the image of its own pseudo-rationality, big data makes its prosumers in the image of its pseudo-rationality.
Down-to-earthness is precisely the problem with forecasting: It only ever reproduces society and its most conventional norms, values, and practices. All that data up in the cloud opens no new vistas; it just repackages tired social, political, and economic institutions (white supremacy, capitalism, patriarchy) in new, hip abodes on more seemingly solid ground. Could we use big data and social media to shoot for the stars, to produce knowledge and types of sociality that transport us from this unjust world to a better one? That’s hard to predict.