Critical theory of technology and STS

The early Frankfurt School proposed a critique of technology which, like that of Heidegger, addressed the epochal crisis of modernity in the 20th century. The major critical theorists, Adorno, Horkheimer, and Marcuse, argued that we live in a technocratic society with a culture colonized by technical rationality. Not only is the culture technological but the actual technology that we employ is adapted to technocratic control of the underlying population. These concerns of early Critical Theory have been left behind by Habermas and his successors. They ignore technology and focus on other issues. But in recent years technical issues have invaded the public sphere. Critical Theory risks finding itself irrelevant if it does not renew its critique of technology.

I will argue that Science and Technology Studies can play a role in that renewal. STS is based on a critique of the very same technocratic assumptions against which Critical Theory had previously argued. Its critique of positivism and determinism has political implications. But at its origins STS took what Wiebe Bijker called the ‘detour into the academy’ in order to institutionalize itself as a social science (Bijker, 1995: 5). It adopted empirical methods, developed case histories, and limited its scope, avoiding politically controversial issues. Its latent political critique has become explicit in recent years as STS has responded to the rise of technical politics by broadening its concerns and reaching a wider audience both within and outside the academy. I want to begin with that political background. Its wide scope justifies my continuing interest in the equally encompassing Critical Theory of the early Frankfurt School.

The New Left first introduced the general public to the politics of technology. In the late 1960s students protested technocratic control and demanded participatory democracy. The most important New Left movement occurred in France in 1968. The May Events rejected both Soviet communism and Western capitalism and demanded self-management in all the institutions of society. In opposition to technocratic rationality, its slogan was ‘All power to the imagination’. Although none of the New Left movements succeeded in transforming their society, they inaugurated a fundamental shift in the culture of advanced capitalism. The 1950s consensus was shattered and controversies arose in many different areas of social and political life.

Many of these controversies concerned technology. Let me remind you of some highpoints in this history. In the 1970s feminists demanded changes in the over-medicalized childbirth procedures of the post-war period. Although this movement involved millions of women and actually instituted new medical practices, it is largely forgotten today. But women still enjoy the right to have partners accompany them in labor and delivery rooms, a direct consequence of those protests. Also in the 1970s environmentalism emerged as an important political movement. Concern with air and water pollution, toxic wastes, lead paint in poor neighborhoods, and many other issues mobilized large publics and changed both technological design and the political agenda. In the 1980s AIDS patients challenged the organization of clinical research. This was an aspect of a larger move against paternalistic practices in medicine which has had permanent consequences. In the 1990s the internet was transformed into a social medium by its users. And in the 2000s we have been preoccupied by climate change.

I call public activity in these domains ‘democratic interventions’. They respond to two different types of motivations: on the one hand, the public intervenes to address problems such as discrimination and pollution; on the other hand, interventions aim to realize potentials unsupported by existing designs, as in the case of social uses of the internet. The exercise of these new forms of sociotechnical agency takes novel forms. There are three main types of interventions: first, controversies, which include protests, boycotts, litigation, online petitions, and so on; second, creative appropriations such as hacking in which technologies are repurposed for different uses than those intended by their creators or owners; finally there are various sorts of participatory dialogue between lay and expert actors. These democratic interventions have enlarged the public sphere (Feenberg, 1999: 120–1).

This is the background against which STS has become academically influential, but it is a background which STS itself has never attempted to analyze. Remarkably, STS investigates everything except the sources of its own success. The question is open: what does the emergence of technical politics mean? Does STS have the means to analyze such a vast phenomenon after rejecting what it considers speculative approaches such as Critical Theory? Case histories cannot explain it, but it is true that a speculative answer is no longer convincing.

STS has adjusted to the new facts of technical life within the limits of its essentially empirical approach. It has studied new domains of controversy, new actors and new modes of action. For example, there is now a considerable literature on user agency and on the role of subordinate social groups such as women and colonized peoples. These studies show new layers of functionality delegated to technologies in the course of democratic interventions from below. STS researchers have also studied hybrid forums, citizen juries and other similar experiments in public input into science and technology policy. The concept of co-production of science and society has been applied to such cases. Science is no longer viewed as a neutral resource, above the fray of competing social interests, but is recognized as a social institution influenced by public debate (Latour and Weibel, 2005; Jasanoff, 2004; Wynne, 2011; Chilvers and Kearnes, 2016; Callon et al., 2011).

It is hard to formulate a general argument for democratic intervention on the basis of this research, but the opposition to public participation is itself highly generalized. Rationality and efficiency are constantly invoked in defense of the existing system. This cannot be dismissed as mere ideology since technical disciplines often validate the rejection of democratic demands. Thus critical theory was not wrong to emphasize the importance of what Marcuse called ‘technological rationality’. The central insight of critical theory, that rationality has become a cultural form, remains of great interest. What is new is that now we are confronted with multiple rationalities instead of a single technocratic rationality.

The challenge is to work out that insight concretely. For that we need new concepts as a basis for a synthesis of the contributions of STS and Critical Theory. In the rest of this paper I will explore this proposal in two stages, first through a generalization of some ideas about co-production and networks drawn from STS, and second through a concretization of the Critical Theory approach to technocratic rationality. This will be an eclectic exercise, so be warned that every orthodoxy will be offended.

To begin, I will develop an STS-inspired perspective. As I said earlier, STS is implicitly anti-technocratic through its rejection of determinism and the exaggerated role attributed to efficiency as a technical criterion. These concepts were intended to place technical decision-making in the heaven of pure reason alongside mathematics and natural science. Instead, STS proposes that technical decisions and designs are underdetermined by purely technical considerations. STS argues for the role of social groups, ‘actors’, in interpreting the meaning of technological artifacts. Assigning one or another meaning influences design choices. In STS terms, efficiency is not an absolute criterion of development but is context-relative (Pinch and Bijker, 1987). The social relativity of design explains how it can become the legitimate object of political struggle.

The concepts of under-determination and interpretative flexibility drawn from social constructivism can be applied to the notions of co-production and networks, loosely inspired by actor-network theory (Latour, 1992). Networks are assemblages of people and things, joined together by causal and symbolic bonds. Individuals acquire a perspective on the technical networks in which they are enrolled as well as what I call ‘participant interests’. These interests reflect the new needs or problems that may arise from their participation. So, for example, people who must commute to work acquire an interest in good roads, while those whose homes are polluted by the cars’ exhaust acquire an interest in better pollution controls, and so on.

Participant interests are articulated in programs and anti-programs. Programs consist in the dominant actors’ designs for the network. They are able to organize a large portion of the nodes around their intentions, but it would be an elementary mistake to imagine that their program coincides with the entire network. Other nodes become active around consequences of network participation that the dominant actors fail to foresee and control. Where these uncontrolled nodes are human beings, they may propose intentional anti-programs that conflict with the dominant program. For example, a factory may appear to its owners as a business with an economic purpose. The corresponding program will consist in a business plan. Its workers may have other priorities. Their anti-program may include a strike for higher pay. Meanwhile those living near the factory may regard it as a polluting nuisance. Their anti-program may involve a protest and a lawsuit (Feenberg, 1999: 114–19).

Note that I am only interested in programs and anti-programs articulated by human actors. This deviation from Actor-Network Theory is necessary to make sense of what I call the ‘symmetry of program and anti-program’. The dominant program is materialized in actual technologies through designed-in values and purposes. The dominant actors thus always have the ‘facts’ on their side. The anti-program may be confined at first to discursive expressions such as protests and demands articulating values different from those of the dominant actors. The subordinate actors’ demands usually appear to be unrealistic, ideological, in the face of the ‘facts’. The charge of irrationality haunts all challenges to the technical status quo. But STS argues that values are routinely delegated to technology; thus there is no obstacle in principle to these discursively formulated values eventually reshaping the technologies. This in fact happens more and more frequently, as I argued in the brief history of technical controversy with which we began. Values are thus in some sense the (possible) facts of the future (Feenberg, 2017: Prologue).

The correlate of this counter-intuitive proposition is Marcuse’s equally counter-intuitive claim that technology is ideological (Marcuse, 1964: 11). This brings me to a consideration of Critical Theory. Its early critique of technological rationality might be reconceived as an attempt to provide the anti-programs that mobilize around technical networks with a response to the charge of irrationality.

Consider Marcuse’s original 1964 critique of technocratic liberalism, an ideology that was at its height in the 1950s and ‘60s. This was the apotheosis of technocracy, an ideology with roots in Saint-Simon and Comte. The modern version was based on intuitions that emerged at the end of the 19th and the beginning of the 20th century around the development of large-scale technical systems such as the railroads and the electrical system. These systems appeared as direct interventions of science in the social world, transcending the many normative disagreements and conflicts of interest that prevailed in the political sphere. The idea that technocratic rationality could unite and improve society took concrete shape then, but it only achieved widespread popularity after the Second World War.

The great contest between Western democratic capitalism and the ‘totalitarian’ alternatives – fascism and communism – was over and the West had won. This was the ‘end of ideology’; all significant social problems were treated as occasions for technical interventions based on the advancing social and natural sciences. Political debate inspired by conflicting values was expected to give way to social technology around which a rational consensus could be built. Extrapolating to the limit, one could argue for the obsolescence of politics and its replacement by scientific administration. Although today few would defend that extreme position, technocratic arguments reappear whenever the autonomy of science and technology is challenged.

One-Dimensional Man (1964) attacked the assumptions behind these arguments. Marcuse claimed that technological rationality had become a new legitimating ideology, replacing the outmoded notion of free markets. His critique had three main elements. First, he argued that the mass media and individual consumption were powerful integrative forces. They had successfully reduced the resistance of the proletariat to capitalism. Second, he showed that technological rationality had become a cultural form, a kind of a priori of experience that redefined social conflicts as technical problems. He criticized the substitution of technical solutions for normative discussion of fundamental issues. A society based on artificial scarcities and competitive struggle was protected from challenge by minor reforms that silenced criticism. Third, Marcuse argued that technological designs such as the assembly line conform to the requirements of top-down control and reinforce the capitalist system. He claimed that a redesigned technology under socialism would embody different values.

Marcuse’s critique is ambivalent. In the first pages of his book he says he will waver between two hypotheses. According to the first hypothesis American society is moving toward total integration. On this account it will be able to cancel its own contradictions and maintain itself fundamentally unchanged for the foreseeable future. The other hypothesis holds that new resistances may emerge from the widespread realization that competition and scarcity have long since become technologically obsolete in the affluent society (Marcuse, 1964: xv).

The first hypothesis is reminiscent of dystopian fictions such as Brave New World. In fact dystopian ideas and images were commonplace in the mass culture and politics of the late 1960s, especially on the left. The second hypothesis promises a breakdown of the ideological consensus shaped in the 1950s around the technocratic legitimation of capitalism. Marcuse called this the ‘Great Refusal’. Presumably, it will lead to the creation of a new and more humane technology. But no concrete theory of technical politics joins these latter two aspects of Marcuse’s theory.

The emergence of technical politics in recent years seems to confirm the second hypothesis. Marcuse’s abstract hope is fulfilled in new political realities. Technical politics revives popular agency and puts dystopian fears to rest. The system is no longer total, no longer wholly integrated. But the picture differs from Marcuse’s anticipations. Democratic interventions are inspired by participant interests rather than by a general disillusionment with capitalism. And they induce technical change not through a Great Refusal but through negotiations, some conflictual, others cooperative, between lay actors or technical outsiders and those in command of the institutions.

In sum, we are even further from Marxism than was Marcuse, who jettisoned the notion that the proletariat was inherently revolutionary under capitalism. But some elements of the original Marxist theory retain their validity in this new context. Marx believed that workers, as technological insiders, could modify the technological foundations of society. There is thus in Marxism a notion of imminent resistance to the given technical system. Class interests, like my generalized concept of participant interests, lie behind the conflict over technology. The idea that those subordinated to the technology might be capable of understanding and improving it is still of interest in our new situation even though the movements that challenge the dominant technological rationality are no longer found mainly in factories.

Marx’s concept of resistance implied a notion of rationality superior to that of the market. He defined it in terms of conscious control and planning. This is dangerously vague, as the Soviet attempt to implement it has demonstrated. From the standpoint of critical theory, the problem now is to achieve a more concrete understanding of that alternative rationality from below. This requires new strategies of critique, more fine-grained than Marx’s concept of exploitation.

Adorno formulated the goal of critical theory as a ‘rational critique of reason’ (Adorno, 1973: 85). This seems right: an irrational critique of reason, that is, hostility to reason as such, reaches a wide public only in the undesirable form of religious or nationalistic fanaticism. Adorno’s approach to critique implies another rational standpoint, other than the technocratic standpoint. He found that standpoint in art rather than the proletariat. The utopian impulse is preserved there. But art makes no explicit claims against the norm of efficiency or the technical disciplines that govern social life. No wonder Adorno ended up comparing Critical Theory to a message in a bottle tossed into the sea.

We need a rational critique of reason that validates the rationality of actual resistance to alienated technical arrangements. There is every reason to believe in the existence of that alternative rationality. Its validity has been proven over and over as democratic interventions succeed in placing problems on the agenda and forcing their resolution. What critical strategy corresponds to this new situation?

The chief obstacle to answering this question is the usual understanding of rationality. The legitimating power of the dominant rationality stems from its implicit claim to represent humanity in general. Critical theory must refute that claim and demonstrate the social relativity of rationality as it is applied in the real world. This requires a method for showing the bias of rational disciplines, systems and artifacts.

This is challenging because our conception of bias was formed in the Enlightenment. The philosophers of the 18th century who inspired the French Revolution struggled against a regime and a religion based on narrative legitimations, stories about the past, and pseudo-facts that had no rational basis. They challenged those legitimations with facts and scientific theories. I call this a critique of ‘substantive bias’ because it addresses a substantive content of belief. This is still a valid critique of racism and other forms of prejudice. However, it is not effective against systems that claim a rational pedigree. The chief examples of such systems are what I call the ‘technosystem’, the market, the administration, and technology. They call for a different kind of critique.

It was Marx who invented the first such critique. His theory of surplus value is a model from a methodological standpoint. He showed how the market produced discriminatory outcomes despite its basis in equal exchange, enriching the capitalist and impoverishing the worker. Thus a rational principle – and what could be more rational than mathematical equality? – was in fact biased to privilege one of the parties to the exchange. I call this ‘formal bias’ because the outcome is prejudiced by the rational form of the social arrangements rather than any particular content of belief.

Critique of this type of bias works by enlarging the context within which claims to rationality are evaluated. For example, the narrow context of the labor market justifies the capitalists’ claims of equity, but extending the context to include the production process reveals an asymmetry: the workers’ productive capacity exceeds the cost of their labor; the equal exchange of labor for wages leaves the excess to the capitalist without compensating the worker. The economic claim is less interesting today than the method which can be applied wherever rational systems discriminate against the participant interests of some of those they enroll.

The ‘formal bias’ of technology is reflected in what I call the ‘technical code’ that translates between specific social demands and technical specifications (Feenberg, 1999: 87–9). Artifacts have standard forms that reflect the social world in which they are situated. These forms do not conflict with efficiency but are the framework within which efficiency is achieved. Let me give some simple examples. These examples show the ubiquity of formal bias which is inherent in all practical applications of rationality and not confined to politically salient forms of discrimination.

Example 1: Certain tools such as scissors are hard for the left-handed to use. These tools exhibit a formal bias independent of myth and prejudice. It is contained in the design of technology. No prejudice presides over that design, but there is a problem nevertheless. Specialized stores have been created to market tools for the left-handed.

Example 2: The sidewalk ramp offers a politically significant example. Sidewalks with high curbs block the movement of wheelchairs. The disabled experience normal curbs as discriminatory. They demanded the right to freedom of movement and their claim was eventually recognized by law. Recognition now takes the form of a specification for the position and angle of sidewalk ramps.

There are also technical codes that cover whole realms of technology. I call these ‘domain codes’. At the highest level of abstraction, domain codes are epochal forms defining an idea of progress. For example, 19th-century capitalism introduced notions of deskilling, commodification and centralized management that presided over hundreds of different inventions and rationalizations of production and administration. These domain codes express the formal bias of capitalist technology. Marcuse’s argument could be reformulated on these terms as advocating an alternative domain code for a society that is no longer dominated by capitalism.

The theory of formal bias gives us a way of thinking about the context-dependent character of rationality. And that in turn suggests an evaluation of multiple rationalities emerging from different perspectives on the formal bias of technical systems. On the one hand there is knowledge from above which is conditioned by problems of technical control of potentially uncooperative subordinates. These problems first appear with wage labor. Until capitalism reorganized production most workers worked for themselves and their family. They were productive on their own account without the need for management. Capitalism expropriates the workers and so must find ways of coercing or incentivizing them into producing even though they have no obvious interest in the success of the firm. This leads to the development of new forms of administration that eventually spread wherever technical mediation supports large-scale organization.

Marcuse argued that advanced capitalist administration is characterized by the systematic substitution of manipulation for coercion. The operational components of resistances, problems, and protests are discovered by specialized social sciences employing empirical methods and quantitative measurements. This is technological rationality at work.

But technical politics reveals the existence of another kind of knowledge, a knowledge from below reflecting the experience of subordinate participants in technical networks. This is a qualitative knowledge that is not formalized in specialized technical disciplines. In fact it often crosses the line between specializations where they ignore salient connections in the real world. Experiential knowledge is responsive to a broad range of values, not simply efficiency and control. It inspires resistance to the dominant organization of technical networks (Feenberg, 1999: 110–14).

What are the prospects for a democratic alternative based on knowledge from below? It seems likely that technical politics will become a commonplace phenomenon in the future. Under the impact of democratic interventions, capitalism might become more humane, better able to take into account the needs of the underlying population. Democratic legitimations might replace technocratic forms. To some extent this is already happening, as witness the rhetoric of democratic management and flattened hierarchies.

But there is an inherent tension in a democratic legitimation of capitalism. The appeal to democracy undermines the centralized technical power that seems to be essential to capitalism. Growing economic inequality testifies to the fact that this power has not diminished. Democratic legitimation might liberate the potential for a new form of technologically advanced society based on power relations incompatible with capitalism. At this early stage in the development of technical politics, there is no way to know if this is a realistic prospect, but it is at least interesting to contemplate.

Contemplating it, I am reminded of a suggestive remark in Foucault’s lectures on neoliberalism. He is describing the difference between the early modern state which attempted to regulate every aspect of social life, and the liberal state that rules through opening a domain of free economic exchange. These are two ‘arts of government’ or systems of ‘governmentality’, in his terminology. Suddenly, Foucault digresses, as he often does in these lectures, to exclaim that socialism lacks a comparable art of government. It merely imitates the attempt at universal regulation of the early modern state, or limits to some extent the domain of free exchange established by the liberal state. But it has no original art of its own (Foucault, 2004: 93).

After the failure of the Soviet Union, those who continue to speculate on socialist alternatives have indeed mostly sought regulatory and market-based alternatives, confirming Foucault’s critique. But perhaps technical politics adumbrates a specifically socialist art of government present today under capitalism in embryonic form. Foucault himself described struggles around medicine and prisons that resemble technical politics in counter-posing a ‘subjugated knowledge’ from below to official administrative and medical sciences. Such struggles were confined to a few institutions in Foucault’s day, but now they have spread in more or less conflictual forms to a great many more. These struggles are, to be sure, non-revolutionary. They are far less disruptive than classic labor struggles, but they have significant effects on the quality of life. Most importantly, they bring different forms of knowledge into communication. That is something new for societies with advanced technology and administrative systems.

Can these struggles be dismissed as ‘reformist?’ Of course they do generate reforms, not revolutions. It is normal that they be absorbed by a still vigorous capitalist system, able to adjust to new constraints. But only a total failure of imagination can dismiss epoch-making changes such as the new roles of women, environmental reforms, and the appropriation of the internet by its users for human communication simply because the capitalist framework persists.

The existence of democratic interventions suggests still more far-reaching consequences in a socialist society organized to prevent the rigid bureaucratization that destroyed the Soviet experiment. Some of these anti-bureaucratic measures are familiar from experiments in workers’ control. While necessary, democratization of the enterprise is insufficient. Its relevance is challenged by the growing role of specialized knowledge in every area of modern life. Furthermore, it does not address the whole range of increasingly threatening externalities of modern technology.

But suppose that an anti-bureaucratic socialism opened the administrations to challenges such as those Foucault analyzed. This would transform technical politics from a disruption of technical normality into a standard aspect of technical life. Under these conditions might a socialist society develop an original art of government based on the fluid interactions between lay and expert network participants? I will leave you with this question, which intrigues me even though I do not yet know how to answer it.