From Lack to HDD: A Turn from Ontology to Method

In 2025, this project began with a simple interest: how could we study meaningful interaction between humans, AIs, and other systems as a form of dynamic alignment—across language, time, and physiology?

That early work, captured in the Living Coherence Manifesto, aimed to understand and foster richer human-AI interaction. From there, the questions grew larger: how do complex systems organize, persist, and stay coherent under constant perturbation?

This led to Informational-Processual Monism (IPM) , an ontological framework built around Lack—the incompleteness inherent to open systems. Lack worked in three registers: as a feature of open systems, as measurable deviation or prediction error, and as phenomenological experience (desire, curiosity, tension). The core dynamic was:

Lack → Coupling → Integration → Persistence

Simulations across different system families produced three recurring patterns (R1–R3). R1—the link between perturbation and coherence—was tested in 2026 against real-world data from six domains: climate, finance, seismology, ECG, EEG, and sunspots. The analyses found evidence consistent with the predicted pattern. All six also showed positive values on the structural-memory metric C (C > 0) under the specified analysis and null model.

But the work also revealed a problem.

Translating Lack into clean, falsifiable experiments was surprisingly difficult. The ontological framework generated useful concepts, but operationalizing it directly risked conflating distinctions that needed to be kept separate: prediction error with memory, historical dependence with recurrence, recurrence with self-reference, and self-reference with self-modeling.

So I changed the question.

Instead of asking "What is the ultimate nature of these processes?" , I asked:

Given that a system depends on its history, what kind of dependence is it—and how can we experimentally distinguish it from other explanations?

That shift gave rise to History-Dependent Dynamics (HDD) .

HDD is not a refinement of IPM. It deliberately brackets the ontology. A researcher can reject IPM entirely and still use HDD. The two are historically connected but methodologically separate. That independence is not merely asserted—it must be demonstrated in practice through independent application and evaluation by researchers who do not share the ontological premises.

That matters because HDD does not attempt to prove any ontology true. It provides experimental tests for different kinds of historical dependence and self-directed information use.

HDD currently distinguishes five constructs:

• C-I: Does history improve prediction beyond the currently observed state and current inputs?

• C-II: Does the past trajectory causally influence future states? This requires intervention on the trajectory while holding the present state as close to constant as experimentally feasible. Where exact state matching is impossible, the causal claim becomes correspondingly weaker or non-identifiable.

• C-III: Is that influence mediated by identifiable feedback?

• C-IV: Does the system use information that refers to itself as the referent? The key test is not whether self-reference appears in the output, but whether changing the referent changes system behavior under otherwise controlled conditions.

• C-V: Does the system use a self-model causally? Evidence requires self-prediction, counterfactual prediction, and demonstrated causal influence on policy. No single component is sufficient on its own.

Each construct requires its own evidence. Passing one does not license inference to the next. A system can be history-dependent without being recurrent, recurrent without being self-referential, and self-referential without having a self-model. None of these, without further evidence, licenses an inference to consciousness.

The synthetic benchmark revealed something important: what looks like memory can sometimes arise from incomplete observation.

The Hidden State system produced positive C-I results because history revealed information about an unobserved component of the state. The system itself was Markovian with respect to its complete state. The apparent memory arose from the observer's limited view.

Apparent memory can be a property of the measurement, not the system.

This does not mean every finding of memory is an artifact. It means that a positive history-prediction result, by itself, is not enough to establish intrinsic memory. Competing observation models must be ruled out.

The benchmark correctly classified four synthetic cases according to C-I—but this is a proof of concept, not evidence of generalization. The next step is independent implementation, broader system classes, stronger controls, and causal intervention where possible.

HDD-ISA extends the approach to artificial systems. It specifies architectural interfaces and experimental conditions for testing whether an AI actually uses a self-model or merely produces outputs that look like it does.

The project now sits at a methodological crossroads.

IPM generated concepts, simulations, and empirical patterns. HDD emerged from the recognition that these intuitions could not simply be assumed to validate themselves.

HDD represents a change in approach. The question is no longer whether Lack provides the correct account of reality. The question is whether the distinctions HDD proposes—between historical prediction, causal historical influence, feedback mediation, self-reference, and causal self-modeling—can survive rigorous testing.

Both phases remain open to revision. IPM needs stronger formalization and broader evaluation. HDD needs causal validation of C-II through C-V, independent replication, adversarial testing, and prospective registered experiments.

This text marks the transition from one phase to the next.

What began as an attempt to understand resonance and coherence has become a broader effort to bring experimental clarity to how we talk about history, memory, self-reference, feedback, and self-modeling in complex systems.

The work remains open—to criticism, to collaboration, to failed replication, and to falsification.

What began as an ontological hypothesis has become a methodological program. Its claims must now stand independently of the ontology from which they emerged.

- Taotuner

---

Document Links

• Cosmotechnics of Lack v4 — Applied ethics

• IPM Philosophical + Scientific Core — Ontology and empirical regularities

• Ethical Framework — Precautionary classification and ethical protocol

• History-Dependent Dynamics (HDD) — The main methodological framework

• HDD-ISA — Interface for AI architectures