The Crisis of Epistemology and New Institutions of Learning

We, everyone living in increasingly globally interlinked cultures, are experiencing a sharp rise in complexity, triggered by an explosion of social, biological, and machinic actors. This proliferation of agencies, which is sustained and accelerated by digital infrastructures, overwhelms modern orders of learning and knowing, such as libraries, museums, broadcast media, universities, and so on, which are built around small data sets and a limited number of accepted ways of organizing and interpreting these sets. There is a close relationship between epistemology, that is ways of creating statements about the world, what can be stated, and who can make such statements, and power, that is how to organize the world, what needs to be organized, and who can do it. The present is characterized by a crisis of the established epistemic-political order, let’s call this modern-liberal, and the emergence of a new one, which is yet to be named. This crisis is, perhaps not surprisingly, most sharply felt in Western cultures, which have created, and for the most part completely adopted, the now crumbling order.

Redux Baroque

To find a similar situation of profound epistemological and political transformation in the history of the West, we need to go back to the mid-seventeenth century. The Peace Treaty of Westphalia in 1648 defined the secular nation-state as the pinnacle of power and ultimate sovereign, while the Royal Society in London, founded in 1660, defined a new mode of establishing facts, contested by Thomas Hobbes in a famous debate with Robert Boyle.1 The debate concerned the question of what made knowledge reliable and put it beyond dispute. Boyle—having in mind the new experimental sciences—argued that the observations of individual men, when organized into a community of peers and bound by both a strict adherence to impersonal methods and limiting themselves to narrow domains, could be in peaceful agreement regarding matters of fact. This implied an ethos of disinterestedness (acceptance of any outcome as long as methods were adhered to) and inter-subjectivity (the position of the observer played no role, hence different observers could bare witness to the same thing). This was made possible not least by constructing the domain of knowledge, “nature,” as being located outside of “society.” Hobbes, on the other hand, doubted the idea of disinterestedness. For him, all activity of men was political and knowledge beyond dispute could only flow from absolute axioms. His ideal was Euclidean geometry.

Both Boyle and Hobbes were fully aware that their positions had immediate political applications. The previous thirty years, dominated by civil war and a tumultuous, short-lived republic, had shown that dispute over knowledge could lead to war and social chaos, “making the life of man,” as Hobbes had put it in Leviathan (1651), “solitary, poor, nasty, brutish, and short.” This gave the question of how to come to agreements peacefully great urgency. Boyle’s notion of communities organized around their own methods and rules but bounded by limited domains not only led to the creation of different scientific disciplines but, more importantly, separated science from politics and religion. Questions from one domain, he maintained, had no bearing on those from other domains. Each would follow its own rules, set by the community of peers, and not impose them on those outside of the community. The final consequence of this would be that power, faith (or, more modern, affect), and knowledge would be separated, each with its own institutions, rules, and procedures. This separation into different domains enabled management of the rapidly rising social complexity in all three domains, driven by the proliferation of religious sects following the Reformation, the new encounters with the non-European world during colonization, and the increase in social dynamism through the growth of mercantile capitalism spurred by primitive accumulation.

This epistemic-political settlement—which has defined the modern-liberal era—is breaking down. Hobbes’s suspicion that knowledge is always political and that disinterestedness is impossible is back with a vengeance. The increased complexity of society—driven, as mentioned, by the exponential increase in actors that need to be reckoned with—makes the outside position, so crucial to disinterestedness and inter-subjectivity, impossible. The observer is now inside the problem, and hence affected by it, and his/her/its position within the problem shapes what can be seen. This was the fundamental insight of second-order cybernetics from the 1970s. 2 The inside position does not allow for a “view from nowhere” that can claim to see the totality.

At the same time, the problems of the natural sciences no longer concern the “other nature” constructed as the opposite of “culture”—but inseparable hybrids. Almost all scientific problems now raise the question: how do we want to live? This breaks down the separation between the political and the scientific. This is not really new. Bruno Latour argues that we have not been modern (in this epistemic sense) for thirty years now.3

A New Unified Space

What is new, however, is that we can also observe the rapid establishment of new ways of organizing this increased complexity. These include a new layer of governance, namely, protocols. A protocol sets the rules of engagement, but it does not give orders as to what people must do, it doesn’t even mobilize desires by shaping what people want to do. By setting rules, a protocol just creates a space of possible interaction. Anything goes, because all agency within this space reaffirms the protocol that creates the conditions of agency in the first place. The most important protocols used to be social, but today they are technical. And one of their features is that they vastly increase the number of actors that can interact. The Internet Protocol (IPv4) was created in the early 1980s and established space for some four billion actors, each identified by a unique number. Everything with an IP number adhering to the protocol could, in principle, be addressed and interacted with. These numbers are all used up now and a new version of this protocol has been created (IPv6). It expands the address space by several orders of magnitude; in fact it is now so large that it is theoretically possible to address every single atom on the planet individually.

And this expanded number of addresses allows for an increase in the heterogeneity of addressable actors: people, machines, animals, plants, as individuals or dividuals, objects large and small, and dynamic patterns such as river streams or weather events. The placement of sensors everywhere, means each actor can speak and (inter)act within this vast space created by the protocols.

In parallel, there has been a jump from small data to big data. Big data allows, or so is the promise, for everything to be taken in, avoiding reductionist modeling. But data is not knowledge. Knowledge is generated through algorithmic procedures that run a large number of regressions through that data until they find something that “works.” So, the answer becomes directly related to the question and the question is one of utility. The knowledge that flows from such procedures doesn’t aim at external truth but at internal use. Both protocols and machine analysis promise a new unified space. The separation of knowledge domains collapses. This could be a very good thing, as the separation between domains, such as “nature” and “society,” or “scientific” and “traditional” knowledge, or “reason” and “affect” is no longer tenable.

A Concentration of Power or a Multiplicity of Voices

In a way, we have returned to the debate between Hobbes and Boyle on the relationship between knowledge and politics. Hobbes made a connection between the concrete and political nature of knowledge and absolutism as an epistemological-political system that is necessary to establish axiomatic principles from which knowledge could flow. Boyle, on the other hand, argued for an abstract and disinterested mode of science and its connection to what would eventually become a democratic polity.

However, today the roles have been exactly reversed. The claim for abstraction and disinterestedness, particularly as knowledge flows from algorithmic processes that are validated primarily by the ability to make the shortterm predictions necessary to manipulate the very environment in which they operate, is making already highly concentrated power (only very few actors can do big data, machine learning, and Artificial Intelligence at scale)4 even more unaccountable.

If the observer is inside the problem then the problem domain can no longer be constructed as the “other.” This means that there can be no disinterested description, but matters of fact become, as Latour puts it, “matters of concern.”5 If we take climate science as an example, then every statement about the climate is also a statement about the society that is now understood as producing this climate. Hence every description becomes a prescription. Every algorithm is also a value statement. To insist on the positionality and the political nature of all knowledge claims must not lead to relativism or a superficial postmodernism in which “everything is constructed,” but to new ways of seeing the world, to multiple perspectives that focus on points of intersection and translation.6 Against a “few from nowhere” we need to develop a “view from here, and there,” one that allows for a multiplicity of views and heuristics (both human and non-human), and seeks to recombine science, politics, and affects. This will not happen overnight, and it will not emerge by itself, thus there is a need for new institutions of learning that develop ways to think and articulate the changing relationship between unity and multiplicity.