Ruipeng Shi

In the cellular-automaton programme for quantum mechanics (including 't Hooft's Cellular Automaton Interpretation and related Hamiltonian-automaton and stochastic-clock approaches), a core question still awaiting operationalization is how to define, inside the automaton, a subsystem that acts as an observer — how measurement is physically implemented, and how an embedded subsystem can operate at an effective resolution below the global update rate. We define an observer by its registration loop rather than by labelled cells: a local process counts as part of an observer S iff it enters S's registration, comparison, maintenance, or response-selection loop. S is a proper subset of the system it inhabits, with no cost-free access to resources beyond itself; since registering a difference requires holding a current item against a maintained contrast item, the distinctions S can concurrently maintain within one operational step are bounded by a registration capacity C(S). Measurement is then a finite map π_S: Ω → R_S, inducing a partition ∼_S on configuration space. A resolution asymmetry follows: S can register, at full grain, structures generated more coarsely than itself, but not structures as fine as itself — including any domain containing S, whose registration is a registration of S's own distinction structure (a second-order, self-referential measurement). Since S cannot internally generate a strict refinement of ∼_S, on the full-grain self-referential task π_S is necessarily non-injective relative to any finer comparison standpoint: a forced, internally undetectable coarse-graining (a quotient) — S cannot tell from a registered value whether its preimage is one fine-grained configuration or several. Other-directed quotienting is conditional, occurring when the target's grain outruns the observer's resolving grain or when concurrent maintenance demand exceeds C(S); self-directed quotienting is unconditional at the resolution level; and any finer standpoint invoked here is itself finite. Information thus forms only downstream of resolvability and concurrent maintenance: what S holds is either a full-grain registration or a quotient — never the pre-registration event structure itself. Two consequences follow. Probability need not enter the update rule: the substrate remains deterministic, in line with the Cellular Automaton Interpretation, while for the embedded observer probabilistic description appears as a normalized valuation on the quotient outcome space — classes corresponding to preimages it cannot unfold — a natural operational form for registration under insufficient capacity, offering a complementary, registration-level route to stochastic-clock-type behaviour. And a "slower observer" requires no modified dynamics: slowness is a property of registration structure — step size, operational window, concurrent capacity — not of the update rule. A minimal simulation can make the contrast between the two standpoints explicit: the modeller reads the full configuration and lattice structure cost-free within the simulation, while the embedded observer holds only the quotient induced by its capacity — a standpoint unavailable to any physical observer. Measurement is thus constitutive rather than read-out. All results follow from counting and partition arguments, independent of the specific update rule.

This paper takes a single registered difference as its starting point, works back to the minimal conditions it requires, and derives classification, re-identification over time and questions of identity only where further premises are explicitly declared. Proposition A states three necessary conditions of a registered difference: two relational positions, resolution relative to the configuration and the task, and joint availability within a single operation. Finitude and embeddedness are treated as restrictions on the type of system studied: relational, operationally constitutive and index-local. Every proposition is marked for its argumentative status and for its standpoint. The marking separates the modeller's statements, events on the observer's side, and bridging claims that need a model–implementation condition. On this basis the paper introduces symmetric negation. Under a comparison condition and an agreement premise, a negative registration licenses only "at least one side has judged incorrectly", without saying which. Where the declared realization class contains candidates compatible with the registration under both attributions of failure, the negation is symmetric, and breaking the symmetry requires further discriminating evidence or an explicit asymmetric restriction. A symmetric negation may put in question not only the content of a claim but also the claimant's qualification to produce it: the correctness of a judgement and the conformity of the operation that produces it are two different propositions. Registered blindness can then be characterized as a configuration's negative registration that calls its own attribution of qualification into question. A difference of direction follows: a registration can call qualification into question, while positive qualification can be certified only conditionally under declared realization conditions, and cannot be established without presupposition from the outputs of the operation being certified. The paper further proves that, at a symmetric registration of disagreement, every sound single attribution of failure rests on a declared restriction the registration cannot certify; adding sources in separated dependency classes cannot remove this restriction; it can only transform the restriction from an asymmetric premise about identity into a symmetric structural premise requiring at least 2f + 1 dependency classes. The partition into dependency classes is itself a modeller's declaration: agreement cannot certify it, and disagreement can negate it only symmetrically. A finite embedded observer can approach the partition asymptotically, with no terminal certification. The paper does not claim that distinction creates underlying existence, nor that description achieves ultimate self-grounding. The declarations it uses but does not establish are collected in the D list (§9.2). By the paper's own classification, that list records the author's registered blindness about the paper's premises, and cannot claim to be complete.

This paper develops an operational model of the limits of self-grounding in finite embedded systems. It distinguishes a system’s capacity to describe its own rules from its capacity to certify that its actual realisation follows those rules, and separates certification from recorded evidence from certification carried out by the system’s own run. An evidence-fibre criterion characterises when a positive verdict is sound relative to a specified class of possible realisations. Within this framework, uninvoked transitions are not determined by past history under local freedom, while certifications that depend on a restrictive realisation assumption cannot establish that assumption from the very same evidence. For realised audits, masking constructions identify conditions under which an internal positive verdict is unsound, alongside a counterexample showing that model-relative positive self-certification is possible. Finite non-circular ultimate grounding is then shown not to close under either per-credential masking or an expressly non-primitive grounding semantics; the latter is a task assumption, not a consequence of finitude alone. The analysis also distinguishes diachronic audit from synchronic mutual checking and derives fault-tolerance bounds in terms of independent dependency classes rather than raw channel counts. The resulting framework clarifies both the uses and the stated limits of self-audit, formal verification and epistemic authority.

Jump Epistemics develops a meta-epistemology of knowledge generation for finite embedded epistemic agents. It represents epistemic activity as a forward-growing chain of costly loops and introduces five basic elements—memory–computation, distinction, record, index, and presupposition—together with derived judgments of evidence, reliability, qualification, and grounding. On this basis, the paper distinguishes legitimate local closure from final closure, recorded forward openness from backward and endless opening, and classifies pathologies of closure, openness, configuration, and order. A configurational account of inquiry and targeting shows why qualification depends on actual generative conditions rather than on the name or numerical value of an output. From this account the paper derives a strict form of Goodhart’s law: targeting can preserve an indicator’s symbolic identity while destroying its referential identity relative to the assertions it supports (support equivalence). A proposition on identity follows: sameness of name is not self-certifying, since the loss of referential identity cannot be registered by the records of the identically named indicator itself; identity is supplied by generative history, which cannot be recovered from the value. Applications include quotient formation, fraction reduction, diachronic reconstruction, numerical approximation, and the logical form of quantum-measurement puzzles. The framework permits conditional, corrigible knowledge without self-grounding, while specifying the additional evidence, declared restrictions, weakened targets, or derivations required for legitimate epistemic promotion.

Every measurement by a finite embedded observer is a quotient: the observer holds its record, not the fibre behind it. A registered value is relational, constitutive and local (A1). Three prohibitions follow, one for each property (A2): no assertion without its presuppositions, since a value holds only relative to its anchors; no bare-value read-out, since a value is constituted by its arrangement and carries its index; and no refinement of local values into a global ontology without new evidence. Whether new evidence can refine a registration depends on the direction of measurement. For other-referential measurement, a finer embedded observer can in principle resolve what a coarser one merged. For full-grain registration of a domain containing the observer, every embedded observer of any size meets a pigeonhole bound (A3). The quantum puzzles are located by the refinement each demands. A single measurement is an ordinary registration; the puzzles arise when its value is read bare, or when values from different arrangements or times are assembled into one global assignment — a refinement of a parent-system property into local values. Values in different arrangements are constitutively independent; their correlation is a parent-system property registrable only at a comparison point. By Fine's theorem a Bell violation shows that no global joint distribution reproduces the statistics, and a Leggett–Garg violation shows the corresponding failure for histories, given non-invasiveness. These are obstructions of a different kind from A3, which bounds full-state registration and implies neither result. Light-speed invariance and gravitational curvature show the same structure of complete local measurements and parent-system properties registered by comparison. Wave–particle duality, delayed choice and Wigner's friend are analysed on this basis. The paper does not derive the Hilbert-space formalism or the Born rule. It explains the paradoxical form of the measurement puzzles, states complementarity as a hypothesis about a parent-system bound, and leaves the quantitative form of interference statistics to further work.

Political authority does not operate over a pre-given field of facts, objections, and historical comparisons. It also participates in maintaining the infrastructures through which relevant differences become adjudicable, challenges acquire institutional force, and past exercises of authority remain comparable with present ones. This paper develops a narrow theory of legitimacy verification rather than a general theory of legitimate authority. It distinguishes actor-indexed adjudicability (Δ), the dependency topology of challenge routes (Γ), and diachronic trace structures that preserve rule and jurisdictional comparability. The central claim is conditional: when an authority invokes an empirical justificatory property—such as competence, responsiveness, contestability, or non-arbitrariness—the continuing evidential force of that claim depends on maintaining institutional conditions capable of discriminating its failure. Formal plurality of oversight is therefore insufficient when review channels share critical upstream dependencies. Because de-correlated challenge is costly to maintain, challenge architectures may become more correlated through locally reasonable consolidation even without capture, deception, or anti-democratic intent. Such contraction can reduce adverse findings and thereby create apparent success that further weakens the case for restoring independent challenge. The paper calls this a self-obscuring hollowing mechanism and derives diagnostic implications for accountability, transparency, institutional design, and diachronic claims of authority continuity.

Finite observers do not merely update beliefs on a fixed stream of evidence. They must continuously reproduce the distinctions, historical anchors, and de-correlated challenges that make correction possible. Once an established belief influences the investment in those very structures, success and corrigibility can diverge: reasons continue to accumulate while the capacity to generate reasons for revision contracts. This reflexive divergence is epistemic hollowing. This paper proposes a dynamic framework for belief formation, justification maintenance, and epistemic hollowing in finite observers. Its point of departure is not probability, uncertainty, or error rates, but the more upstream condition of registration: a worldly difference must first be resolved by a finite interface and jointly sustained within a comparative window with retrievable anchors before it becomes a registrable event for the observer; observer-internal probability assignments only acquire an operational domain once this outcome space has been constituted. Consequently, historical identifiability is not treated as a background automatically provided by time, but as an achievement of registration that requires continuous resource investment. The paper further repositions JTB (Justified True Belief) as an entitlement architecture of 'continued taking-as-true,' rather than as a final definition of knowledge; shows that a substantial family of Gettier-like luck can be represented as saturation failure of predicates relative to registration equivalence relations; and argues that any belief-revision rule defined solely on the current registration space cannot recover task-relevant distinctions that have not entered its domain. Dynamically, the success history S and the challenge structure Γ can grow apart: success can be capitalized through memory into new justification, but only when the internal warrant-updating rule underestimates the dependency structure of the generating path; when it does not, repeated success does not inflate justification. Once a belief acquires entitlement, it in turn shapes future action, sampling, archives, classification, and challenge investment, producing a self-reinforcing trajectory in which success reproduction can crowd out failure-space reproduction. This gives rise to three hollowing mechanisms—distinction collapse, challenge collapse, and identity drift—corresponding to the three dimensions of corrigibility: registration capacity (Δ), de-correlated challenge capacity (Γ), and cross-temporal rule invariance (I). In dialogue with core literatures, the paper operationalizes Bateson's 'difference that makes a difference' as 'registered difference that makes a registered difference,' prefixes Ashby's requisite variety with 'requisite registrable variety,' shows how a substantial family of Gettier-like cases can be represented as registration saturation failure, pushes Hume's problem of induction from insufficient exogenous samples to endogenous sampling distortion, and engages the organizational decoupling literature to specify an endogenous persistence mechanism: success is capitalized into entitlement, entitlement reduces challenge reproduction, and the resulting evidential thinning is then read as further success. Institutions instantiate the same abstract architecture, but introduce additional parameters of delegation, conflict, and transferability.

A finite observer can be genuinely adequate to a declared task; what it cannot obtain from the same current interface is a final certificate that its registration exhausts every task-relevant finer distinction. This paper develops the framework behind that claim by separating three levels that earlier versions partially conflated. P0 states the constitutive structure of realised distinction: a distinction requires first-term uptake, maintenance across a comparison span, and the independent resolution of a second term; these are not three interchangeable resources but three irreducible aspects of one distinction-making act. P1 states that their physical realisation in any current observer is finite and non-zero; P2 states that every resource actually used in that realisation belongs to the embedded observer. The number two therefore enters before Shannon: one anchor can register presence, whereas two are the minimum arity of a non-trivial relation and first permit same/different and, for live succession, before/after. Measurement is consequently quotient formation rather than cost-free read-out. We distinguish six quotienting mechanisms and isolate a physical self-referential limit based on generative-grain / resolving-grain asymmetry: full-grain current self-resolution cannot be implemented at the same current registration grain. The same minimal relational structure constitutes registered temporal order without purporting to derive substrate time. Epistemically, the framework separates theoretical awareness, instance adjudication, and final certification; it is non-sceptical because refinement can correct earlier blindness even though no current finite interface thereby acquires a final certificate of exhaustive grain. Two domain-general corollaries follow: every registered value is loop-constituted rather than implementation-free read-out (a1), and no quotient value may be asserted as grain-exhaustive (a2). Consequences include the operational locus of probability without substrate randomness, the underdetermination of resolving grain by measured rate, and — at the quantum interface — the claims that the EPR reality inference is not licensed by finite embedded registration alone, that measurement independence is finitely defeasible but not finitely exhaustively confirmable, and that Bell’s theorem remains untouched.

Regulatory compliance is often treated as evidence that the practices or capabilities a rule is meant to secure remain in place. This article develops a theory of structural hollowing: an accepted qualification persists while task-relevant realization declines without deception or rule breaking. The argument separates realization, adjudication, qualification, and institutional consequence; distinguishes observational from normative thickness in accepted classes; and treats adjudicability as a maintained organizational capacity rather than a fixed monitoring parameter. When qualification has independent value, lower realization is reachable and locally less burdensome, recurring selection opportunities create downward pressure. Hollowing becomes cumulative only when hysteresis makes restoration harder and may weaken the challenge, evidence, or memory capacities that would reveal the loss. The framework yields failure-locus-specific diagnostics and explains why adverse findings can decline either because performance improved or because the system became less capable of generating findings against its own accepted states.

Around 450 BCE, Zeno of Elea posed a paradox that has haunted Western thought for two and a half millennia: an arrow in flight must occupy a definite position at each instant, yet at any such instant it is motionless. The standard mathematical response — that calculus resolves the paradox by summing infinitely many intervals — is formally correct, yet it leaves the deeper physical question untouched. Can a durationless instant ever function as a real existence event for an embedded observer? Within the framework of Subsystem Epistemics — the study of what knowledge embedded subsystems can acquire under finite resources — the answer is no. An event exists physically only if it admits observational closure by at least one embedded subsystem. Since closure requires finite registration, comparison, and distinction, a zero-duration instant cannot itself be a closed existence event. The arrow's motion is therefore not a sequence of static instantaneous positions, but an ordered succession of discrete closure events — what we call xent — anchored within a minimal stepped structure. This paper develops the Information-Spacetime Stepping Framework (ISSF): a structure discrete at the ontic level of closure events and continuous at the observational level of interval relations. Continuity is not rejected; it is relocated. Discreteness belongs to the closure events that make existence registerable; continuity belongs to the measurement relations that embedded observers recover from those events. The framework therefore returns to Zeno with a different answer. Zeno's paradox does not merely call for calculus. It calls for a theory of observational closure. The ISSF supplies precisely that: a unified continuous-discrete spacetime in which existence occurs through discrete xent events, while embedded observers reconstruct continuous relations through intervals between them. What appeared as an ancient paradox thereby becomes the structural condition for any stable system capable of supporting knowledge from within.

Aristotle, in Physics III.5–8, argued that nature admits only potential infinity — endless process, never completed totality. An actual infinity, he held, could possess no determinate place or operation within the cosmos. Two millennia later, continuous spacetime and uncountably dense truth-value structures became the default background of mathematical physics, and Aristotle's argument was set aside. The question he raised, however, has never been answered. It has only been bracketed. This paper removes the bracket. We present a reductio showing that any embedded subsystem with bounded energy and the capacity to acquire approximately stable knowledge cannot operationally distinguish a continuous truth-value structure from a discrete one of finite cardinality. The argument rests on two empirical premises — that energy within any bounded region is finite, and that subsystems do in fact form stable records — together with the derived consequence, via Landauer's principle, that each registration carries a non-zero energetic cost. The continuum, if it exists at all, is thereby operationally inaccessible: a substrate posit with no consequence for any energy-bounded epistemic system. Four available counter-strategies — denying the operational claim, denying registration cost, denying bounded energy, or asserting stable knowledge in a continuous regime — are enumerated and foreclosed. Each is either physically untenable, metaphysically unobservable, or structurally equivalent to the discreteness conclusion under different terminology. The argument proceeds independently of quantum mechanics, general relativity, or any specific commitment to spacetime structure. It uses only what no working physical theory currently denies. Where modern physics treats the continuum as a calculational background while leaving its operational status unexamined, we argue that the operational status is structurally determined: any subsystem that actually knows anything must operate on a finite quotient. Aristotle's rejection of actual infinity in nature is recovered, not as metaphysical dogma, but as a theorem about the necessary conditions for stable knowledge under finite resources. The path to actual infinity, like the path to the Ding an sich, leads back through the very resource constraints that make knowledge possible — and stops there.

In 1632, Galileo observed that no experiment performed within a closed cabin can detect the ship's uniform motion relative to the shore. This observation was elevated by Einstein in 1905 to a postulate of fundamental physics: the laws of physics are the same in all inertial frames, and the speed of light is invariant. The mathematical consequences — Lorentz transformations, time dilation, length contraction, the spacetime interval — have been confirmed to extraordinary precision. Yet the question of why uniform motion is undetectable, and why c is invariant, remains answered only by stipulation. This paper proposes a structural account that recovers Galileo's intuition through the information-theoretic tools unavailable in either Galileo's or Einstein's time. Within the Subsystem Epistemics (SE) framework, information is the registration of a difference between two distinguishable events, observation is necessarily performed by subsystems embedded in the system they observe, and all measurement must proceed through finite-time closure using the system's own propagation rate r = l/t. Measurement is therefore self-referential: the same r that constrains access to physical quantities is used to measure them. Three levels of temporal structure follow. The ontic level (L1) is the absolute xent count on the underlying lattice. The local level (L2) is each subsystem's accumulation of existence events, bounded above by the resolution limit r/2 inherited from the two-event structure of information. The assigned level (L3) is constructed through signal-mediated measurement procedures such as the radar protocol. Under these constraints, the radar protocol is shown to be the unique L3 measurement procedure compatible with resource-bounded observation, and the geometric mean assignment t_B = sqrt(t_1 · t_2) is uniquely determined by accessibility, reciprocity, scale invariance, and closure consistency. From this foundation, relativistic effects emerge as algebraic consequences of self-referential measurement rather than physical deformations of local processes. The Lorentz factor γ arises from the geometric reconstruction of remote events; absolute velocity is eliminated from every observable quantity through the dimensionless form v/r; the preferred frame (the lattice background) exists but is structurally inaccessible to embedded observers — not by conspiracy, but by the algebra of measurement itself. Applied to the twin scenario, the framework yields equal local xent counts at reunion under symmetric trajectories: the apparent age difference is a computational output at L3 rather than a physical accumulation at L2, and arises from conflating the two levels. Where Galileo articulated the phenomenological observation and Einstein elevated it to a postulate, the present framework provides a structural mechanism: self-referential measurement — the use of r to measure r — algebraically eliminates absolute velocity from every observable quantity available to an embedded subsystem. The paper does not claim that special relativity is incorrect; its mathematical structure is fully reproduced. The contribution is interpretive: a reduction of brute postulate to structural necessity, with empirical content concentrated in cellular automaton simulations where the embedded-observer / simulator distinction is concretely realisable. This paper is a working paper. Its purpose is to isolate a structural possibility rather than to present a completed physical theory. The central claims are the three-layer distinction of time, the self-referential character of embedded measurement, and the resulting inaccessibility of absolute motion through v/r-normalised observables. Several technical components, including the full discrete realisation of the three-axis virtual ruler, the CA implementation, and the physical identification of clock readings with L2 tick counts, remain open problems.

In 1814, Laplace articulated the ideal of an intellect that, knowing all forces and all positions at a single instant, could predict the entire future and retrodict the entire past with certainty. The demon's power rests on a presupposition that has rarely been challenged: that an observer can stand outside the system, reading its complete state without consuming resources or perturbing the system itself. Subsystem Epistemics (SE) identifies this presupposition as structurally untenable. The relationship between observer and system is not that of a container and its separable contents, but that of a Klein bottle — any attempt to occupy an external vantage point leads back to the interior. A Laplacian demon is therefore not merely practically unavailable but structurally self-defeating: the act of registration is itself a physical operation within the system, subject to the same constraints as everything it seeks to describe. This paper develops SE as a minimal structural framework for analysing information and observation under conditions of embeddedness. The framework rests on two foundational identifications. First, information consists in the registration of a difference between two events: a single isolated event does not constitute information, since no comparison is available. Second, measurement is the unique operational process through which such information is acquired. All observation is performed by subsystems operating under finite resource constraints, with no external vantage point available. From these conditions, minimal temporal and configurational separations — denoted t and l — are introduced as hidden constants: structurally real, but operationally inaccessible. From these we derive a structural rate r = l/t and a maximal subsystem resolution bound r/2. This bound is universal, recursive, and implies an intrinsic coarse-graining that is structural rather than epistemic. Three structural difficulties for continuous spacetime models are identified — indeterminacy of closure completion, limitations on invariance confirmation, and loss of distinguishability — and it is shown that standard mathematical tools address these formally but not operationally for a resource-bounded embedded subsystem. Minimal step structures resolve all three difficulties while preserving the foundational assumptions. The framework is interpretatively neutral. It provides a shared operational floor for both Relational Quantum Mechanics (by grounding the inaccessibility of observer-independent descriptions) and the Cellular Automaton Interpretation (by motivating discrete substrate structures from operational conditions alone). Where Laplace assumed that an external observer could in principle grasp the totality of a system from outside, the SE framework demonstrates that embeddedness is not a practical limitation but a structural condition: observation, resource consumption, and constraint are three descriptions of a single operation. The path to the outside leads back to the inside.

“You cannot step twice into the same river,” Heraclitus observed, raising a question that has remained unresolved for over two millennia: under what conditions can identity be maintained across change? If a system is continuously different at every moment, what makes it the same system at all? This question has been addressed in multiple traditions without receiving a unified structural formulation. In early modern philosophy, it appears as Hume’s problem of uniformity and habit; in Kant, as the presupposition of persistent substance; in control theory and engineering, as system stability; and in complex systems, as self-organisation and steady states. In each case, stability is treated either as a presupposition or as a dynamical outcome, but not as a structural condition for the possibility of observation itself. This paper develops a minimal framework — subsystem epistemics — in which stable systems are taken as the prerequisite for observational closure. Within this framework, observation is not performed by arbitrary embedded subsystems; rather, it requires subsystems situated within systems that support stable structure, in which Law, Validation, and Manifestation can be consistently invoked and jointly realised. We show that observational closure is possible if and only if this triadic structural condition — the TECUM equivalence — is satisfied. On this basis, we adopt two minimal conditions: (i) information consists in the registration of a difference between two events; and (ii) all observation is performed by subsystems embedded within the system they observe, subject to the same resource constraints. Together with the principle that logical operations are physical processes consuming system resources, these conditions imply that observation is possible only in stable systems that support consistent closure operations. We then show that continuous spacetime models fail to sustain this structure. Three independent structural tensions arise: indeterminacy of closure completion, limits on invariance confirmation, and loss of informational uniqueness. These tensions jointly prevent the stable registration of identity across change. The minimal resolution is the introduction of minimal temporal and configurational steps, giving rise to the existence event unit (xent) and to the Information-Spacetime Stepping Framework (ISSF) as the operational background within which observational closure can be consistently defined. Within this framework, we introduce the Informational Uniqueness Principle (IUP) and show that it arises as a structural consequence of closure conditions, rather than as an independent axiom. We further introduce an information field density, ρ_info ∈ [0,1], as a structural quantity emerging from the verification congestion mechanism, under which the holographic principle appears as a boundary condition rather than as an independent postulate. The framework is structurally realist and does not presuppose any specific dynamics. It identifies the minimal structural conditions under which identity, information, and observation are possible for embedded subsystems. Two classes of testable predictions are proposed: lattice-level anisotropy detectable in cellular automaton simulations, and verification congestion effects providing a unified structural account of light propagation in media, slow-light phenomena, and horizon freezing. Where Heraclitus articulated the instability of identity in continuous flux, the present framework offers a structural resolution: identity is maintainable only where observational closure is possible, and observational closure is possible only in systems that support stable structure and minimal discrete stepping.

The most basic precondition of knowledge is rarely articulated as a question: how is distinction itself possible? Aristotle took this question to its limit when he identified the principle of non-contradiction as the most absolute presupposition of thought (Metaphysics IV.4) — A cannot at once be X and not-X. But the principle itself rests on a prior condition: X and not-X must be Aristotle distinguishable. Without distinguishability, the principle is empty. Distinguishability is therefore prior to thought, and the conditions under which distinguishability can be achieved are the conditions under which knowledge can take any form at all. This paper develops a minimal structural framework addressing that condition, under the constraint that the entity attempting to distinguish is itself part of what it distinguishes. Two assumptions are introduced: (i) information consists in the registration of a difference between two events, and (ii) all observation is performed by subsystems embedded within the system they observe. From these assumptions, together with the existence of minimal temporal and spatial steps t and l, a structural resolution bound r/2 is derived, where r = l/t. Since registering a difference requires at least two events, and each event occupies at least one minimal step, the minimal closure time is 2t and the maximal resolvable rate is r/2. This bound does not depend on measurement precision or specific physical implementations. It follows from the minimal requirements for observational closure under resource constraints, and is therefore structurally distinct from familiar resolution limits in signal processing, which arise from external sampling considerations rather than from the structure of distinguishability itself. Observation is characterised as a coarse-grained process: multiple configurations map to indistinguishable outcomes, and differences that cannot be registered do not enter any internal description. The result is a structural form of epistemic incompleteness — not a deficit of knowledge that better instruments might remove, but a feature of what registration can be when the registering subsystem operates within the system it registers. Possible connections to statistical structure, observation under varying conditions, and the formation of knowledge from accumulated regularities are briefly indicated, without commitment to specific physical interpretations. Where Aristotle established that thought presupposes the principle of non-contradiction, the present framework articulates the prior condition: the structural circumstances under which two events can be registered as distinct, and the bound on what can be so registered when the observer is embedded. form at all.