Cybernetics as an Observational Stance
The theoretical spine of this cluster is the move away from treating digital phenomena as effects produced by discrete causes, and toward treating them as observed by systems that are themselves part of what they observe. Giglietto2019-e9be81c1 makes this move explicit: drawing on second-order cybernetics (von Foerster, Bateson) and Luhmann, it reframes “fake news” not as a property fixed by a creator’s intent but as an emergent, observer-dependent cascade produced through chains of propagators’ judgments — a micro/meso/macro architecture in which global patterns of misinformation are genuinely emergent rather than aggregated from individual acts. This is sociocybernetics in its classical register: information is “a difference that makes a difference,” and truth-claims are second-order observations of first-order observations. Anicker2024-vp extends the same anti-essentialist logic to agency itself, arguing that whether an entity (human, institution, or AI) counts as an agent is not an inherent technical property but a socially granted and monitored status — a “demarcation problem” that only social theory, not engineering, can resolve. Together these two papers establish the foundational move of the whole topic: internet phenomena (information, agency, publics) are constituted by social attribution and self-referential observation, not by fixed substrates.
Complexity, Multi-Level Systems, and Formalization
A second strand takes this systems-theoretic sensibility and pushes it toward formal, multi-level modeling. Frischlich2025-vn argues that misinformation cannot be captured by any single analogy (virus, weapon, pollutant) because it is a complex adaptive system coupling micro-cognitive, meso-social, meso-platform, and macro-societal levels with feedback and tipping points — explicitly borrowing the epistemic posture of climate science rather than linear causal inference. Ng2026-og radicalizes this same intuition for agentic AI, formalizing “Multi-Agent Social Systems” as a dynamical triple of information exchange, influence dynamics, and interaction network, and deriving four structural priors (strategic heterogeneity, network-constrained dependence, co-evolution, distributional instability) that echo core sociocybernetic themes — feedback, structural coupling, self-organization — but render them testable against large-scale empirical agent networks. Read alongside Frischlich2025-vn, this paper shows the topic’s arc extending from describing complexity qualitatively to operationalizing it formally, positioning social theory (rather than single-agent AI paradigms) as the necessary “structural prior” for understanding both human and machine-populated systems.
Social Forms and the Morphology of Publics
A third thread applies systems/social-form thinking to the changing shape of online publics themselves. Gerbaudo2026-fo explicitly invokes Simmel’s notion of “social forms” and Weberian ideal-types to theorize TikTok’s shift from people-centric “networked publics” to item-centric, algorithmically inferred “clustered publics” — a morphological transformation in how the social is constituted online, echoing Luhmannian concerns with how systems produce their own boundaries and relevancies. Boyd2026-op narrates a parallel, complementary transformation at the level of terminology and lived practice: “social media” becoming “parasocial media” as reciprocal peer sociality gives way to one-sided, algorithmically amplified attention to distant others. Both papers can be read as tracing the same underlying process — the reconfiguration of “the social” as a system property rather than a fixed substance — from two angles: Gerbaudo’s structural/algorithmic register and boyd’s experiential/political-economic one. Sofia2012-802a21dc anticipates this concern from an earlier, more infrastructural vantage: arguing that Internet architecture itself must be redesigned around social-capital properties (trust, reciprocity, centrality) rather than treated as a neutral technical substrate, it is an early, pre-platform articulation of the idea — central to the whole topic — that social structure and technical structure are mutually constitutive rather than separable.
Beyond Causality: Infrastructure, Deliberation, and Normative Reflexivity
A final set of papers uses this systems-theoretic sensibility to critique dominant methodological paradigms in adjacent fields — chiefly platform and polarization research — for their reductive, causal-individualist assumptions. Bechmann2026-dr argues that platform-democracy scholarship should move beyond narrow causal-effects designs toward treating “platform collective behavior” as democratic infrastructure, a reframing that only makes sense once platforms are understood as systems rather than as bundles of discrete stimulus-response mechanisms. Esau2025-tf and van-Eck2026-xg extend this reflexivity to polarization research specifically: Esau theorizes “destructive polarization” as a systemic property of online discourse addressable through deliberative reciprocity and inclusive listening, while van Eck and colleagues expose, via systematic review, how depolarization recommendations in the literature are typically decoupled from the evidence generated and instead smuggle in unexamined normative ideals (chiefly deliberative democracy) — a second-order observation of the polarization field’s own observational practices. Falkenberg2026-ka closes the loop by diagnosing the field’s fragmentation and Western skew, calling for the kind of integrated, cross-contextual framework that sociocybernetic thinking — attentive to systems, boundaries, and observer-position — is well suited to supply. Taken together, these papers do not so much apply systems theory directly as demonstrate, empirically and reflexively, the costs of not doing so, thereby motivating the shift toward complexity- and systems-aware research that the rest of the topic works out in more explicit theoretical terms.