Feedback Loops as an Observational Framework: Describing Homeostatic Structures Without Teleology

This paper develops an observational framework for describing structures that maintain themselves under external perturbation. The framework uses the feedback loop—a circulation of input, output, and re-input—as its primary unit of description, and treats the "process" connecting input to output not as a claim about underlying mechanism but as an explanatory hypothesis introduced to render the observed circulation coherent. This treatment of process as hypothesis is the central methodological commitment of the paper, and is what distinguishes it from prior cybernetic and systems-theoretic approaches. The paper has three aims. The first is to establish the observational framework itself, in which any object whose behavior can be described as a circulation of input, output, and re-input is treated as a feedback loop, and in which the diverse dynamic behaviors recognized in control engineering—asymptotic stability, neutral stability, limit cycles, multi-stability, chaos—are unified under the single description of homeostatic mechanisms: structures that preserve some persistent mode of behavior. Crucially, the framework does not require that a goal function, an optimization criterion, or a purpose be posited as a precondition for the loop's existence; the process hypothesis is introduced only after the circulation is observed. The second aim is to demonstrate that, for any given observed circulation, multiple process hypotheses and multiple readings of homeostasis can stand simultaneously. This is shown through three observations of software maintenance work in a corporate setting, each adopting a different process hypothesis (event classification, specification consistency evaluation, prioritization) and yielding a different reading of what is being maintained (operational viability, specification-implementation alignment, processing order). These readings are not mutually exclusive; they are presented as parallel descriptions of the same circulation under different observational stances. The paper argues that this multiplicity is not a defect but a feature: the framework explicitly permits the listing and comparison of competing hypotheses rather than fixing a single one in advance. The third aim is to apply the framework to three substrates of differing kinds—the biological nervous system, language, and the methodology of scholarly inquiry—and to organize what can be observed about each as a feedback loop. In each case, the paper observes that the range over which homeostasis is maintained has expanded over time along three dimensions: temporal span, spatial span, and degree of abstraction. The expansion of homeostasis observed in these three substrates is then explained through the Pan-Structural Theory of Adaptation (Vaernes, 2025b): structures subjected to opposing pressures of variation and conservation come to carry an internal distribution of mutation rates, and through this distribution become multi-layered, manifesting functional characteristics that allow homeostasis to expand. Finally, the multi-layering dynamics of the Pan-Structural Theory of Adaptation are positioned as continuous with the stabilization tendency described in the Precursor Theory of Stabilization (Vaernes, 2025a), placing the entire account within the broader Persistence Selection Theory framework. The paper does not claim that feedback loops, as objects in the world, exist independently of the act of observation. It claims only that, when one adopts the observational stance described here, certain regularities become describable, and that these regularities are sufficient to organize a wide range of phenomena—biological, linguistic, scholarly, computational—under a single non-teleological vocabulary.