Non-Human Actors: On Luhmann and Latour

I spent much of last night and early this morning reading through an interesting article by Guenther Teubner, of the Johann Wolfgang Goethe University in Frankfurt. Teubner’s a professor of private law and legal sociology, and he’s written extensively on topics in contract law, corporate regulation, and international law. But he’s also developed a perspective on legal evolution and legal institutions that relies heavily on Niklas Luhmann’s distinctive brand of systems theory, drawing upon familiar Luhmannian concepts such as autopoiesis, structural coupling, and paradoxification and applying them in innovative ways.

In the article, Teubner works at a kind of synthesis or accommodation between Luhmann’s systems theory and Bruno Latour’s actor-network theory, using examples drawn from law. He begins by inquiring into why we would grant legal personhood and assign agency to non-human entities such as animals. Personification of the non-human is a practice with deep historical roots, which sometimes led to what seems to be, from a contemporary perspective, legal absurdities. From medieval to early modern times, for instance, animals were put on trial for various crimes (Teubner notes, drily, “The animals did not always win their case”). This practice disappeared with the spread of modern science - but has come back in recent years, primarily because of but not just as a result of the ecological movement. In particular, the appearance of new information technologies have forced the question of the agency of computers, software, and other electronic actors. Are these new actors agents? Are they persons? 

As Teubner notes, there is still considerable hostility to the notion that we should treat anything other than human individuals (and perhaps more familiar legal entities such as organizations and states) as ‘actors,’ except in a metaphorical sense. Isn’t this just an attempt to anthropomorphise the non-human, assigning to trees, lizards, the wind, computers, water, cows and pigs, etc., human qualities so as to give them legal rights and protections (or otherwise make them subject to legal responsibilities)?

In fact, as Teubner argues, what Latour is doing is stripping the concept of agency of its anthropomorphic properties; human agency may be a special type of agency, but only a special type. For an entity to be an agent, or what Latour calls an ‘actant,’ you don’t necessarily need all of those nice warm things we identify as distinctively human: “Forget reflective capacities, phenomenenal worldviews, empathy, the operation called Verstehen, and the ability to communicate. What is left is the minimal presupposition of double contingency.” Double contingency in this instance refers to the fact that many so-called objects, whether artificial or natural, react to us in unexpected ways. When we exert force or apply our technology to an object, the object doesn’t necessarily respond in predictable ways; it may show resistance and recalcitrance, and as a result we may be required to enter into a frustrating and difficult series of interactions with it . For example, perform an experiment on a virus, and the virus may react in one way, or another, or maybe not even at all. This forces you to devise a new method of eliciting the kind of reaction you want, and the virus may respond yet differently. And so it goes. We may not think that the virus is behaving ‘intentionally,’ but its reaction remains contingent. For Latour, that’s all that is required to establish a basic symmetry between the human and non-human.

But this is a very minimal notion of agency indeed, and that’s why Latour goes on from his concept of actants to talk about hybrids. Hybrids are plural forms of association between the human and the non-human. If we stopped our analysis of the non-human with the idea of double contingency, our interactions with the non-human would consist of us prodding black boxes - natural and artificial objects whose internal dynamics are intransparent and obscure to us - and seeing what kind of reactions we get. In order to go beyond this, there needs to be hybrids, which are more or less stable associations that arise between human and non-human actors. Humans interact, and learn to ’speak,’ for the non-human actors; non-human actors, in turn, ’cooperate’ with the humans. Thus, we humans (or at least those among us who have the appropriate credentials) have learned to speak for the AIDS virus. We communicate what we think the virus is doing, what the virus wants, what the virus plans to do next - even while recognizing that in speaking for the virus, we cannot really fully account for its ’intentions.’ What the virus does in a body, in a population, or even in the controlled circumstances of a lab is contingent, i.e., obscure and not entirely controllable. But we’ve managed to force enough ‘cooperation’ so that we can at least speak for the virus with some authority. But humans could not speak for the virus unless there existed a host of other non-human actors available to aid in this process of translation: the AIDS test, diagnostic equipment, and so on, that allows for a stable set of associations between humans and the virus. AIDS is therefore not just a disease, but a whole set of social arrangements and associations between different actants, in which discourses, technologies, human bodies and mental states, and the actual virus itself are at least provisionally stabilized into a self-perpetuating network.

Teubner does a nice job of giving a schematic of the fundamentals of actor-network theory, albeit in the language of systems theory, and connects up Latour’s analysis to Luhmann’s theory of social differentiation. Luhmann always insisted in his theory of social evolution that society has dissolved from a ’centered’ society organized around political and religious hierarchies, into multiple social systems that engage in a kind of plural or parallel processing of social reality (for Luhmann, a social system is a recursively linked network of communications). Economy, law, religion, science, etc., each have their own distinct internal discourses, codes, and even ‘realities’ and concepts of rationality. There is no central unity tying these all together; everyone lives an economic life, a political life, a religious / cultural life, etc., and these lives make more or less sense depending on their specific systemic logics, but there’s no single vantage point (such as ’solidarity,’ ‘collective conscience,’ ‘value consensus,’ or even an ‘ideal speech situation’) from which all of these logics can be seen at once, much less dictated or controlled.

In light of Latour’s work, however, the question for Teubner becomes: when, how, and to what extent are actants and hybrids incorporated into these social systems / discourses as legal persons? When do we start assigning agency to animals, to software programs, etc.? There is no one right answer to this, but will depend on situational logics - especially risks and dangers. The legal system today is increasingly open to including animals as legal persons, in order to protect them from dangers that other systems present to them - such as the economy, with its predatory relationships with the natural environment. However, the situation is different with software programs and computers in the financial sector. Economic systems can no longer assume that the only thing ‘out there’ are people, corporations, and their more or less aggregated preferences. Instead, we incorporate electronic agents within legal discourses as persons, in order to acknowledge their growing autonomy and with recognition of this autonomy, their simultaneous familiarity and potential for harm: “With the social inclusion of cyborgs and electronic agents, new problems of alienation appear on the horizon of the law.” But the point is not that we are assigning electronic agents a kind of fictitious legal personhood in order to better control them. Rather, by recognizing them as ‘actors’ we acknowledge their dualistic nature. They allow us to cope with uncertainty,  providing us with increasingly sophisticated means and tools for extending human action into ever-widening social realms; and yet, at the same time, because they display their own autonomy they cannot be fully controlled and generate uncertainty of their own kind. This two-fold aspect, of being a source of aid and resistance, of cooperation and conflict, links even electronic agents back to double contingency. Thus, even electronic agents must be co-opted or coerced into behaving in ways that we want, and this is something whose success we cannot always guarantee. And herein rests the ‘actorhood’ of the non-human.

It’s interesting to me to see Luhmann’s students attempting to bridge gaps between systems theory and other kinds of theoretical approaches. Here, Teubner attempts to link up the theory of autopoeitic law to Latour’s actor-network theory, while in a different context - the sociology of science - Rudolph Stichweh has pointed out certain affinities between Luhmann’s writings on globalization and ‘world society’ and the macro-oriented neo-institutionalism of John Meyer and his followers. I suspect that if Luhmann’s writings are to gain a purchase on American sociology, it will be by making these kinds of cross-theoretical exchanges interesting to outsiders.