1
The Reference Problem of Digitalization

What problem does digitalization solve? If I am right, this question about the reference problem of digitalization has not yet been asked; and how you ask makes a difference. I do not ask, ‘What is digitalization?’ Nor do I ask, ‘What is the problem with digitalization?’, or ‘What problems does digitalization cause?’. Particularly about this last one, we sometimes know more than we do about my main question, on the reference problem. We know for example that digitalization is a threat to privacy, that it destroys jobs through its efficiency, especially in repetitive activities, that it can also be an economic opportunity, that it opens up possibilities of control that did not exist before, and so on. In a way, such statements presuppose digitalization as an independent variable, in order to enquire about its consequences. My question starts at a completely different point. What problem does it solve?

My answer is going to be as follows: the reference problem of digitalization is the complexity and, above all, the regularity of society itself. The argument is that modern society, especially through its digital kind of self-observation, encounters only those regularities, that stubbornness and resistance that make up social relations. True, society is a fluid, fast-moving, accelerated object, yet it is enormously stable, regular, and indeed predictable in many respects. This object contains patterns that are not recognizable at first sight. The second glance, which of course reveals them, is increasingly a digital one.

Should this thesis prove to be sustainable, it has considerable consequences for a sociological theory of digitalization that does not simply examine the consequences of digitalization and the manner of disruption attached to a certain kind of technology and technique but rather starts with the foundations of modern society itself. And this means that we do not see digitalization, but crucial parts of society are already seeing in digital fashion. Digitality is one of society’s crucial self-references. To be on the safe side, I should say right away that here I don’t take the digital as a metaphor. But more on this later.

This much is already clear: in the elaboration of a theory of digital society, methodological questions arise first, that is, questions of theory construction. If these are not answered, the few statements made so far remain simple assertions. The question about the reference problem is a functionalist one. Functionalist questions are not causal; they are about the relation between problem and solution.

Functionalist questions

Perhaps the most important foundation of functionalist thought comes from Ernst Cassirer. In his early book Substance and Function he postulated a transition from concepts of substance to concepts of function, thus presenting a critique not only of the ontological understanding of the world but also of the retrospective ontologization of the epistemic process. For Cassirer, epistemic objects are constituted in and by the cognitive process itself, which thereby becomes an undetermined point (or one to be determined) within a network of relations: ‘Thus we do not know “objects” as if they were already independently determined and given as objects – but we know objectively, by producing certain limitations and by fixating certain permanent elements and connections within the uniform flow of experience.’1

Cassirer calls the fact that cognition identifies ‘objects’ a ‘formula of confirmation’; and the identified objects are thus not so much ‘“signs of something objective” as rather objective signs’,2 whose objectivity is due to the fact that they prove their worth empirically.3

Then, from a mathematical perspective, the observation of something is always a function of this observation – and this conception of functionalism breaks with the idea that the indeterminacy of the world could be resolved or dissolved through unambiguous determinacy. This manner of relating is a feature not only of scientific insight but of practice in general. It makes visible how concrete decisions relate to something and how something appears the way it does as a result of a particular practice. Thus functionalism always has to do with indeterminacy, or, better still, with the practical production of determinacy – both on the side of cognition and on the side of the object.

Now this is not about epistemological questions around functionalism but actually about the question of the problem–solution constellation of what we call digitality or digitalization. So the thesis could be that we do not see the digital but that we see digitally, so that something like the digital may successfully become visible or emerge. One of the pioneers who do not just describe the digital but demonstrate how we should see digitally to describe the modern world is Dirk Helbing. This trained physicist describes for example the automatization of areas of society not as a mere disturbance that comes from the outside to a certain extent, but, on the contrary, as a part of the social structure that makes possible in the first place the description of disturbances such as the cascade effects in complex systems (e.g. energy supply).4 Consequently Helbing describes the digital revolution as a revolution of the complexity of society itself.

Connecting data: offline

If a functionalist way of thinking is characterized by the need to keep both the cognition side and the object side contingent, it is worth starting with a phenomenological description of digital technologies. In other words, we must adopt an offline view to begin with and completely dispense with the description of the thing itself – that is, of digital technology – so as to get the fundamental structure of society into perspective. If there is anything that the entire digital shares, it is the capability of connecting data with data, that is, the capability of apparatuses to connect data points to one another. The raw material consists of data that exist in counted or countable form and whose form has such a low threshold that they can actually combine and recombine among themselves.

One of the earliest forms of digital – that is, countable – data processing was certainly government social statistics, which developed together with the establishment of the modern state. Thus the ‘social physicist’ Adolphe Quetelet (1796–1874) was one of the first to apply statistical methods to society and social planning. He wondered how much regularity is to be found in human behaviour, for example in marriage. Marital behaviour is in each case an individual kind of behaviour. Concretely, you are dealing with two people who decide to get married. But, numerically, what you are dealing with is slightly different. The piece of information ‘marriage’ recombines with other features, to then make visible what was not really visible before. Now, we must concede that even our everyday understanding works with probability assumptions concerning which marriage is to be expected and which is not. The idea of stratification in society, of cultural or denominational matching, of age distribution within couples, of economic and biographic preconditions, and so on is already rooted in our perceptual schemes and stereotypes, so that we recognize the regularity assumptions of our perceptions only in cases of deviation. But social statistics is able to conceptualize such regularities and ultimately to make them manageable; and through its quantitative capacity it is able to identify rather invisible regularities, which to everyday reasoning look like coincidences or like the contingent result of individual decision-makings.5

The prerequisite for all this is the form of data, a form that has become countable through the recoding of typical features, and a form of digital data, which can be recombined among themselves. Thus the material of data processing or digitality is items that in principle can be recombined among themselves and whose informational value consists precisely in the limited nature of possible combinations. More specifically, if any possible element were linked to any other possible element, the data could not provide any information at all, and so they could not make any difference. By the way, Quetelet regarded deviations from the normal distribution as a disturbance and ended up being fascinated by the idea of an homme moyen, an average person who can be calculated accordingly and who at the same time constitutes the foundation of all those practices through which humans are shaped as self-responsible individuals.

Thus the raw material of the digital consists in lists of coded numerical values, and the solution consists of information about everything that is possible on the basis of data. More specifically, these are probability statements about combinatorics, about how individual items are related – or, even more precisely, about the limits of combinatorics, because only data in which not everything combines with everything else can contain or generate information. This can be information of an entirely different nature:

• the intelligent steering of a machine that is able to adjust itself to changing environmental conditions and to process data in such a way as to react to them through its own specifications;
• the intelligent self-monitoring of...