The problem of double contingency and its accompanying parable of the black boxes informs Luhmann’s conception of meaning and frequently provides illustration at critical junctures. In life, (we) ‘psychic systems’ and (societies) ‘social systems’ are constantly faced with situations that require choices. Our meanings develop from those choices. Each choice that we make is an element of our meaning. Autopoiesis forms the background of Luhmann’s theory. Psychic and social autopoietic systems live by constantly maintaining their reproduction as a closed system. At the same time, they constantly interact with their environment by incorporating elements from it and releasing unneeded elements back into it. In this process, what remains the same is the reproductive process, which incorporates those elements that foster its life and evolution. Interpenetration describes how closed autopoietic systems come to share meaning and come to cooperatre and understand each other.

Open peer commentary on the article “Subsystem Formation Driven by Double Contingency” by Bernd Porr & Paolo Di Prodi. Upshot: I acknowledge the value of Porr & Di Prodi’s piece for simulating Luhmann’s key process of subsystem formation and exploring how the concepts of “differentiation” and “binary code” relate to their model.

Discourse about smart algorithms and digital social agents still refers primarily to the construction of artificial intelligence that reproduces the faculties of individuals. Recent developments, however, show that algorithms are more efficient when they abandon this goal and try instead to reproduce the ability to communicate. Algorithms that do not “think” like people can affect the ability to obtain and process information in society. Referring to the concept of communication in Niklas Luhmann’s theory of social systems, this paper critically reconstructs the debate on the computational turn of big data as the artificial reproduction not of intelligence but of communication. Self-learning algorithms parasitically take advantage – be it consciously or unaware – of the contribution of web users to a “virtual double contingency.” This provides society with information that is not part of the thoughts of anyone, but, nevertheless, enters the communication circuit and raises its complexity. The concept of communication should be reconsidered to take account of these developments, including (or not) the possibility of communicating with algorithms.

Key words: social algorithms, big data, artificial intelligence, neural networks, deep learning, double contingency, systems theory.

Open peer commentary on the article “Subsystem Formation Driven by Double Contingency” by Bernd Porr & Paolo Di Prodi. Upshot: Agents that learn from perturbations of closed control loops are considered constructivist by virtue of the fact that their input (the perturbation) does not convey ontological information about the environment. That is, they learn by actively experiencing their environment through interaction, as opposed to learning by registering directly input data characterizing the environment. Generalizing this idea, the notion of learning by experiencing provides a broader conceptual framework than cybernetic control theory for studying the double contingency problem, and may yield more progress in constructivist agent design.

Open peer commentary on the article “The Circular Conditions of Second-order Science Sporadically Illustrated with Agent-based Experiments at the Roots of Observation” by Manfred Füllsack. Upshot: Observers construct an internal, local state of order for the perspective of their observations, but in doing so they increase the overall entropy of the system they belong to, e.g., society, by adding more options for potential courses of action. Thus, the Second Law of Thermodynamics is not just satisfied in the circular condition of second-order observations, but the increase in entropy is actually exploited as a resource by a system such as society, where it serves as a prerequisite for double contingency.

Open peer commentary on the article “Ontology, Reality and Construction in Niklas Luhmann’s Theory” by Krzysztof C. Matuszek. Upshot: The question of whether contingency can be limited concerns the foundations of sociological systems theory as a theory of cognition. This commentary argues that while such limits may seem plausible and apparent at first, they would consequentially give rise to an ontological notion of society within society. Rather, the commentary proposes to understand the limits identified in the target article as social functions structuring expectations in situations of double contingency.

Open peer commentary on the article “Subsystem Formation Driven by Double Contingency” by Bernd Porr & Paolo Di Prodi. Upshot: By analyzing Porr and Di Prodi’s model for addressing the double contingency problem, I try to take a step further by questioning the importance and implications of the loop concept in the constructivist approach.

Open peer commentary on the article “Communication Emerging? On Simulating Structural Coupling in Multiple Contingency” by Manfred Füllsack. Upshot: Our criticism aims at the premises of Füllsack’s simulation model, i.e., we claim that his interpretation of the Luhmannian concept of double contingency contradicts the systems theoretical approach in fundamental ways. Neither the view of communication as an emergent system, nor the theory of double contingency is addressed in an adequate manner. Thus Füllsack in fact does not simulate a systems theoretical approach to double contingency but simulates a mere reduction of the social to the individual psyches

Open peer commentary on the article “Subsystem Formation Driven by Double Contingency” by Bernd Porr & Paolo Di Prodi. Upshot: There is merit to the continuous-signal-space homeostatic viewpoint on subsystem formation presented by Bernd Porr and Paolo Di Prodi; many of their ideas also align well with a heterostatic constructivist perspective, and specifically developments in the field of reinforcement learning. This commentary therefore aims to identify and clarify some of the linkages made by the authors, and highlight ways in which these interdisciplinary connections may be leveraged to enable future progress.

Purpose: This article investigates the emergence of subsystems in societies as a solution to the double contingency problem. Context: There are two underlying paradigms: one is radical constructivism in the sense that perturbations are at the centre of the self-organising processes; the other is Luhmann’s double contingency problem, where agents learn anticipations from each other. Approach: Central to our investigation is a computer simulation where we place agents into an arena. These agents can learn to (a) collect food and/or (b) steal food from other agents. In order to analyse subsystem formation, we investigate whether agents use both behaviours or just one of these, which is equivalent to determining the number of self-referential loops. This is detected with a novel measure that we call “prediction utilisation.” Results: During the simulation, symmetry breaking is observed. The system of agents divides itself up into two subsystems: one where agents just collect food and another one where agents just steal food from other agents. The ratio between these two populations is determined by the amount of food available.

Key words: Social systems, constructivist paradigm, cybernetics, double contingency, symmetry breaking, emergence