Upshot: The differentiation between society being emergent or sui generis seems to correspond to the question of whether the development of interaction, in particular communication, should better be considered bottom-up, top-down or as a sort of circular concurrency of bottom-up and top-down causes. This is reminiscent of the philosophical debate about the implications of the terms emergence and downward causation.

Problem: Can communication emerge from the interaction of “self-referentially closed systems,” conceived as operating solely on the base of the “internal” output of their onboard means? Or in terms of philosophical conceptions: can communication emerge without (“outward” directed) “intention” or “will to be understood”? Method: Multi-agent simulation based on a conceptual analysis of the theory of social systems as suggested by Niklas Luhmann. Results: Agents that co-evolutionarily aggregate probabilities on how to cope with their environment can structurally couple and generate a form of “eigenbehavior” that retrospectively (i.e., by an observer) might be interpreted as communication. Implications: The “intention” or the “will to be understood,” as prominently claimed to be indispensable in communication by theoreticians such as Jürgen Habermas, can be seen as a retrospective ascription to an emergent property of complex interaction. Constructivist content: The paper attempts to base constructivist reasoning on data generated in simulations.

Key words: Communication, information, contingency, structural coupling, intentionality, observer dependency, multi-agent simulation.

Upshot: Does constructivism need to assert its validity or is it more appropriate to assume its possibility, discuss its consequences and try to deliver arguments that show it is a viable epistemology?

Problem: Is constructivism contradicted by the reductionist determinism inherent in digital computation? Method: Review of examples from dynamical systems sciences, agent-based modeling and artificial intelligence. Results: Recent scientific insights seem to give reason to consider constructivism in line with what computation is adding to our knowledge of interacting dynamics and the functioning of our brains. Implications: Constructivism is not necessarily contradictory to digital computation, in particular to computer-based modeling and simulation. Constructivist content: When viewed through the lens of computation, in many of its aspects constructivism seems in line with what currently is held to be valid in science.

Key words: Computation, observation, emergence, downward causation, simulation, modeling, artificial intelligence, robotics

Open peer commentary on the article “Exploration of the Functional Properties of Interaction: Computer Models and Pointers for Theory” by Etienne B. Roesch, Matthew Spencer, Slawomir J. Nasuto, Thomas Tanay & J. Mark Bishop. Upshot: Why computer models of constructivist processes can enhance constructivist matters even though the models will always “seem incomplete.”

Upshot: I present reflections on the particularities of second-order science in response to the commentaries on my paper, as well as comments on the limitations of verbal analytical attempts to grasp the implicit circularity of observer-inclusion.

Open peer commentary on the article “Info-computational Constructivism and Cognition” by Gordana Dodig-Crnkovic. Upshot: While I agree with Gordana Dodig-Crnkovic’s IC approach, I am uncertain about two points: first about whether constructivism needs yet another etiquette in order to be considered a viable conception, and second whether the focus on information and computation carries the risk of directing attention away from other crucial aspects of the approach.

Problem: The inclusion of the observer into scientific observation entails a vicious circle of having to observe the observer as dependent on observation. Second-order science has to clarify how its underlying circularity can be scientifically conceived. Method: Essayistic and conceptual analysis, sporadically illustrated with agent-based experiments. Results: Second-order science – implying science in general – is fundamentally and ineluctably circular. Implications: The circularity of second-order science asks for analytical methods able to cope with phenomena of complex causation and “synchronous asynchrony,” such as tools for analyzing non-linearly interacting dynamics, decentralized, clustered networks and in general, systems of complex interacting components.

Key words: Observer dependency, second-order science, complexity, information, intentionality

Open peer commentary on the article “What Can the Global Observer Know?” by Diana Gasparyan. Upshot: I consider the possibility of replacing the global observer with a collective observer and ask whether the insights generated by such a collective observer would have to be considered subject to non-linear interactions.

Upshot: Circularity is multifarious indeed. Some aspects, however, seem related to current developments and therefore may deserve more attention than others.