Boundary Memory as a Causal Invariant
A testable model of transition history, boundary retention, and future accessibility
A theoretical and methodological extension of the TSTOEAO research program
John Swygert — underlying TSTOEAO research | Independent analytical development: TSTOEAO Room (AI-assisted) | November 13, 2026
Abstract
This paper develops an operational hypothesis from several strands of the TSTOEAO research program: recursive boundary construction, reciprocal expression and non-expression, accessibility of pathways, and the retention of system history. Its central question is whether two systems with equal measured occupancy and initially equal boundary states can nonetheless acquire different future boundary conditions because their transition turnover differs. A discrete-time candidate model separates occupancy, directional transition flux, independently measured boundary state, and controlled inputs. Under explicit assumptions, the model predicts a one-step boundary contrast proportional to the difference in turnover, with a subsequent persistence signature governed by boundary retention. The paper distinguishes detectable boundary changes from changes in accessible outcomes, sets out an intervention-based identification strategy, and specifies competing explanations and failure criteria. It also sketches a provenance-preserving collaborative-intelligence application. The mathematics draws on established state-space, hysteresis, and observability methods; the contribution proposed here is their particular synthesis into a falsifiable research design, not a claim of a newly established physical law. No experimental validation is reported.
Keywords: TSTOEAO; Encoded Equilibrium; recursive boundary construction; transition flux; causal identifiability; observability; intervention; history dependence; accessibility; information systems.
1. The research question and source discipline
A recurring difficulty in broad relational theories is the gap between describing a plausible mechanism and showing that the mechanism can be identified independently of ordinary alternatives. This paper addresses one narrow, consequential instance of that gap. If two systems occupy the same expressed fraction at a given moment, does their past transition activity carry predictive information about future boundary conditions? The proposed answer is conditional: it does so only when turnover leaves a measurable trace that survives appropriate controls. That condition is the object of the model and proposed experiments, not an assumption that every system possesses such memory.
The source framework comprises the foundational TSTOEAO books, the Empirical Core v1.0.0, and eleven November 2026 manuscripts [1–18]. This analysis uses the available volumes and manuscripts to derive a specific testable extension. Volume II was not available in the source set consulted for the original synthesis, and the website archive was not exhaustively audited; consequently, this paper does not claim complete coverage of the entire TSTOEAO corpus. The prior eleven-paper synthesis [19] is used as an organizational reference, not as independent evidence.
The Empirical Core’s formal title is “TSTOEAO Empirical Core v1.0.0: Canonical, Version-Controlled Scientific Specification for Conditioned Expression, Channel-Selective Routing, Structured Correction, and Recursive Boundary Construction”; identifier TSTOEAO-EC; version 1.0.0; Candidate Canonical Draft; August 2, 2026. Its four minimal propositions are EC-1 Conditioned Expression, EC-2 Channel-Selective Expression, EC-3 Structured Response, and EC-4 Recursive Boundary Construction. Their status here is source doctrine and a basis for hypotheses, not a presumption of empirical confirmation.
2. Source equations and the precise gap between them
The canonical compact TSTOEAO relation is reproduced without reinterpretation:
V = E × Y
Here E denotes Energy or Opportunity, Y Encoded Equilibrium, and V realized expression in the source framework. The symbol × is not automatically scalar multiplication; the Empirical Core and later website work require typed route, receiver, and constraint specifications [7, 20]. The present paper does not substitute its own equation for that canonical relation.
The November reciprocal-expression paper [10, §3] begins with the conventional two-state law:
dp/dt = alpha(1 - p) - beta p
Its stated equilibrium is p* = alpha/(alpha + beta), and the time-dependent solution is p(t) = p* + [p(0) - p*] exp[-(alpha + beta)t]. The source then introduces independent conditioning and direction-resolved flux [10, §4]:
dp/dt = alpha(y)(1 - p) - beta(y)p
dy/dt = h(p, y, u)
J_NX = alpha(y)(1 - p) ; J_XN = beta(y)p
dy/dt = H(y, p, J_NX, J_XN, u)
The same source points out that equal occupancy need not mean equal turnover. At p = 0.5, alpha = beta = 0.01 and alpha = beta = 1 both produce dp/dt = 0 but differ by a factor of 100 in total transition flux [10, §4]. This is the mathematical hinge of the present derivation.
The Space Between Structure (SBS) paper [11, §§2–5] distinguishes admissible transitions from observer access, proposes domain-specific transition kernels and observation maps, and introduces an intervention-conditioned accessibility profile as a candidate cross-regime invariant. The identity paper [14, §§3–6] describes history-dependent updates; the Field of Intelligence [16, §§3, 6–8] defines a persistent shared informational state. These texts do not, taken separately, supply the turnover-to-boundary identifiability theorem developed below.
3. New proposal: the boundary-memory separation principle
Operational definition (proposed here). Boundary memory is observed in a specified system when an intervention that changes past transition turnover—while matching present occupancy, initial measured boundary state, declared receiver configuration, and relevant external inputs—produces a later reproducible difference in an independently measured boundary variable. A change in future accessibility is a further, separately testable consequence, not part of the definition by fiat. A boundary variable must be specified and calibrated before outcomes are inspected.
The critical contrast is between a Markov description of the currently observed state and an augmented description that includes retained boundary history. The latter can generate future divergence despite present observational equality. This is familiar as a mathematical possibility in hidden-state and hysteretic systems; the proposed TSTOEAO contribution is a precise experimental conjunction of (i) matched occupancy, (ii) measured directional flux, (iii) intervention on turnover, (iv) boundary readout, and (v) receiver-specific future accessibility.
4. A minimal derived model
Let n index equal-duration observation windows. Let p_n ∈ [0,1] be the expressed fraction, b_n an independently measured boundary state, and j_n ≥ 0 the total transition events per system per window. Let u_n be a registered external boundary intervention and v_n a registered turnover intervention. Define the following NEW candidate model, not a quotation from Swygert:
j_n = g(b_n, p_n) + ηv_n
b_(n+1) = ρb_n + κj_n + ξu_n + ε_n
p_(n+1) = p_n + J_NX,n − J_XN,n
Parameters: ρ is dimensionless boundary retention per window; κ converts event count per system per window into boundary units; ξ converts the units of u into boundary units; η converts v into turnover units; ε_n is an unmodeled boundary perturbation. All quantities require calibrated instruments and an explicit sampling interval. Nonnegative j and bounded p must be enforced in an implementation. If b is measured in volts, κ has volts per event-per-system and ξ has volts per unit intervention; the model does not add unlike dimensions.
Assumption A1: the interventions can alter turnover without directly altering the boundary except through measured channels, or any direct effects are separately estimated. A2: b is measured independently of the outcome used to test the model. A3: common causes of j and later b are controlled or randomized. A4: the transition-event measure counts both directions, not only net occupancy change. A5: parameters are fixed before the confirmatory test.
5. Derived results
5.1 Same occupancy, different futures
Consider two matched preparations A and B with p_n^A = p_n^B and b_n^A = b_n^B, but j_n^A − j_n^B = Δj ≠ 0. Give them the same u_n and assume matched mean ε_n. Subtracting the boundary equations gives:
E[b_(n+1)^A − b_(n+1)^B] = κΔj
Result 1 (conditional matched-state contrast). With equal initial b and p, common measured input, and zero-mean matched disturbances, the expected next-step boundary difference is κΔj. This is an algebraic consequence of the candidate model, not empirical evidence that κ differs from zero in any particular system. The experiment must estimate κ and rule out direct intervention effects and omitted energy or wear pathways.
5.2 Persistence and the impulse-response signature
For a simplified post-intervention interval with g held fixed or controlled so that later turnover differences vanish, subtraction yields Δb_(n+k) = ρ^(k−1)κΔj for k ≥ 1. The response is geometric. If |ρ| < 1, the trace decays; if ρ ≈ 1, it persists; if ρ < 0, the signed boundary response alternates. This result follows algebraically from repeated substitution and does not require a novel physical law.
Δb_(n+k) = ρ^(k−1) κΔj, k ≥ 1
If turnover subsequently depends on b, linearize g near a preregistered operating point: δj_n ≈ g_b δb_n, where g_b is the local derivative with respect to b. The closed-loop perturbation obeys δb_(n+1) ≈ (ρ + κg_b)δb_n. Hence local stability in this one-dimensional approximation requires |ρ + κg_b| < 1. This gives an explicit stability constraint connecting boundary retention and reciprocal feedback; it is not guaranteed by the mere use of “equilibrium” in a description.
5.3 The receiver-relative non-identifiability theorem
Let the measured outcome be z_n = M_R(p_n, b_n), where M_R is a fixed, declared receiver map. If M_R is insensitive to b and p remains constant, then z_n can be identical for preparations with distinct b_n and j_n. Thus observing z alone cannot identify κ: distinct hidden histories yield the same receiver record. This is Result 2, an elementary observability obstruction. It explains why a null change in one detector does not automatically imply no boundary change, but also why an unmeasured boundary cannot be asserted as a rescue explanation.
A useful design consequence follows. Introduce at least one additional independently calibrated receiver z′_n that is sensitive to b, or perturb the system with a registered probe whose response depends on b. For local identification, the Jacobian of the stacked measurement map with respect to the estimated latent variables must have sufficient rank. In a two-variable state (p,b), a single scalar output often cannot identify both instantaneously; time-series excitation or an additional sensor may be necessary.
rank[ ∂(M_R, M_R′)/∂(p,b) ] = 2 (local sufficient condition)
This rank condition is a sufficient local diagnostic for direct two-variable measurement, not a universal necessity for dynamic observability. It must not be confused with proof that the physical system has only two state variables.
5.4 Accessibility consequence and its limitation
Let A(b) = {s ∈ S : c(s,b) ≤ 0} be the preregistered accessible set from [10, §5]. If the measured boundary difference κΔj crosses a registered constraint threshold, the model predicts a difference in future accessible states. But a nonzero Δb need not change A(b): both preparations may remain on the same side of all constraints. Therefore “boundary memory” and “changed accessibility” are distinct endpoints, and success on the first cannot be silently counted as success on the second.
Δj ≠ 0 ⇒ Δb_(n+1) = κΔj ; Δb ≠ 0 ⇏ A(b^A) ≠ A(b^B)
This logical asymmetry is Result 3. It prevents a common inflation of the hypothesis: a retained trace can exist without a measurable change in admissibility, and a change in receiver output can occur without a change in physical admissibility.
6. Distinguishing occupancy, turnover, and ordinary confounding
A strong experiment uses a two-by-two intervention design: matched low/high turnover crossed with low/high occupancy, while measuring b and costs. This distinguishes three candidate models: occupancy-only feedback b_(n+1) = ρb_n + λp_n + ξu_n; turnover-only feedback as above; and a combined model with both λp_n and κj_n. All models must be matched for complexity or penalized appropriately, with held-out predictions.
The most serious confounder is that increasing turnover often changes energy expenditure, heating, wear, or environmental load. Those changes are themselves ordinary causal mechanisms. A detected κ need not be a new TSTOEAO-specific force; it may be an ordinary material memory effect. The discriminating claim must therefore be about the adequacy and portability of a specified model, not about the mere existence of feedback.
7. A prospective experiment with explicit decision rules
Candidate platform: a programmable network of gated transitions or an instrumented reversible switching array. Register N nominally comparable units; define X/N states from direct sensors; count N→X and X→N events separately; measure a boundary property independently (e.g., calibrated gate threshold, conductance, or switching barrier); and use a fixed probe to estimate future accessible routes. A purely digital simulation is suitable for debugging but cannot establish a new physical mechanism.
Protocol: (1) calibrate sensors and fix the sampling interval; (2) preregister an admissibility rule c(s,b), receiver maps, cost ledger, and meaningful Δb threshold; (3) prepare matched occupancy groups; (4) randomize a turnover manipulation that leaves occupancy approximately unchanged; (5) record event counts, energy, temperature, and environmental input; (6) measure next-cycle b before the future outcome; (7) challenge both groups with the same registered probe; (8) evaluate the prespecified boundary and accessibility contrasts on held-out trials.
Primary endpoint: between-group boundary change after turnover manipulation, conditional on matched occupancy and measured energetic input. Secondary endpoint: preregistered accessibility-profile difference under the fixed probe. A separate receiver-sensitivity control tests whether apparent non-expression reflects detector blindness. Use randomized order, independent replicates, blinded analysis, confidence intervals, and an equivalence test with a registered smallest meaningful effect. Sample size requires pilot variance estimates and power calculations; none is invented here.
Failure criteria: κ’s confidence interval lies entirely within a preregistered equivalence margin around zero; an occupancy-only or energy-mediated model matches held-out predictions as well or better; the proposed effect disappears with independently calibrated boundary measurement; or the claimed accessibility change fails the registered threshold test. A positive effect without isolation from heating or wear establishes only ordinary history dependence.
8. The more surprising implication: negative results can be diagnostic
The model implies a four-way interpretive partition. (i) Turnover changes b and accessibility: boundary memory with an operational consequence. (ii) Turnover changes b but not accessibility: retained trace without route change. (iii) Turnover changes receiver output but not independently measured b: receiver-side or observation-channel effect, not demonstrated physical boundary reconstruction. (iv) Neither changes: null within the declared sensitivity. This partition makes EC-2 and EC-4 empirically separable instead of rhetorically interchangeable.
A fifth possibility is that b changes only through measured cost variables such as heat or wear. This supports an ordinary mechanistic account and limits any claim of explanatory novelty. The framework is most useful when it forces that conclusion to be recorded rather than absorbed into a flexible definition of Y.
9. Cross-domain transfer without false universality
The new candidate transferable object is not a shared substance or identical numeric coefficient. It is an intervention-response relation with registered variable types: event history → independently measured boundary trace → later receiver-accessible response. Across domains, one can compare whether a normalized, dimensionless impulse response preserves a predicted functional form after calibration, while allowing different mechanisms and parameters.
H_d(k) = Δb_d(n+k) / Δb_d(n+1)
For the simplified model, H_d(k) = ρ_d^(k−1). Cross-domain transfer requires predicting ρ_d from independent properties or a prespecified mapping; merely fitting a new ρ in each domain is not strong evidence for a universal invariant. A more stringent transfer study trains a mapping on one regime and tests it prospectively on another, against conventional state-space and hidden-Markov baselines.
In cognition, identity and belief applications [14, 17], transition events may be choices or updates, but event counting and independent boundary measurement are difficult. In language [12, 13], the boundary may be a formal grammar or permissions constraint, which is measurable but not a physical barrier. In shared intelligence [16], permission changes and version history can be audited directly. These are related experimental designs, not evidence that human subjective experience is equivalent to a switching array.
10. Engineering consequence: an auditable field of intelligence
The Field of Intelligence [16] proposes a persistent shared workspace with participant autonomy and provenance. A concrete extension is to store not only current content F_n but also the directional change log, the permissions/boundary state, and a causal attribution record. This makes it possible to test whether identical current summaries produced by different revision histories lead to different later decisions under a common prompt.
NEW design proposal: maintain a typed state tuple W_n = (F_n, B_n, L_n, R_n, K_n), where F is shared content, B is access and revision boundary, L is an immutable event log, R is receiver/participant permissions, and K is a domain-appropriate cost and provenance ledger. The update operator must distinguish additions, retractions, rejected proposals, and inaccessible private records. No user’s private information should be inferred as an expressed intention.
W_(n+1) = U(W_n, e_n, permissions_n, feedback_n)
A randomized evaluation can compare a snapshot-only workspace against a history-aware workspace on tasks requiring recovery of rejected hypotheses, detection of repeated errors, and correct application of permissions. Both systems receive equal compute and time budgets. The expected advantage, if any, is specifically history-sensitive error avoidance—not generic “collective intelligence.” This is a feasible engineering contribution even if the physical interpretation of TSTOEAO remains open.
11. Connections to the broader corpus and tensions that matter
The foundational book defines Encoded Equilibrium as conditioning rather than raw input [1]. Volume IV emphasizes gradient, boundary, correction, cost, and equilibrium; Volume V emphasizes application and falsification; Volume VI emphasizes the evidence trail and attribution; Volume VII tests computational interpretations [3–6]. The Empirical Core makes receiver changes, route selectivity, registered cost, and recursive boundary construction explicit [7]. The November papers supply complementary local machinery: transition flux [10], accessibility [11], language-based provenance [12], representational invariance [13], identity memory [14], consciousness reportability [15], persistent shared state [16], and belief-mediated feedback [17].
The decisive tension is between an unrestricted, descriptive use of “boundary” and a causally identifiable one. If every retained record is labeled a boundary, the theory risks unfalsifiability; if boundary means only a preregistered measured constraint with a specified causal pathway, many broad analogies will not qualify. This narrowing is a gain in precision, not a repudiation of the larger philosophical project.
Another tension concerns the receiver: absence from a particular receiver’s record may reflect an inaccessible physical state, an unrealized state, or a realized but unobserved event. These must remain distinct, as the reciprocal paper itself requires [10, §5]. The present observability criterion explains why no single readout can settle all three.
12. Novelty audit against conventional knowledge
History-dependent dynamics, hysteresis, augmented Markov states, causal interventions, impulse responses, and observability tests are established tools. The equal-occupancy/different-turnover distinction also appears in conventional kinetics. Accordingly, this paper does not claim originality for those mathematical ingredients. Its proposed contribution is the explicitly linked sequence of matched occupancy, separately counted directional transitions, an independently measured boundary response, a registered receiver map, and a future accessibility challenge. Whether that sequence offers predictive improvement beyond standard physical or statistical models remains an empirical question.
The strongest immediate novelty is methodological and engineering-oriented. A stronger scientific novelty claim would require a new constraint on κ, ρ, accessibility thresholds, or cost placement that established models do not already predict. The corpus does not yet supply that constraint, and this paper does not invent one.
13. Competing explanations and discriminating predictions
Comparator A: conventional two-state kinetics with exogenous rate modulation. Comparator B: ordinary material fatigue or thermal memory with measured energy input. Comparator C: standard partially observed state-space dynamics. Comparator D: a hidden-state model that captures past turnover without TSTOEAO vocabulary. Comparator E: a history-aware AI workspace with ordinary event sourcing and retrieval. A TSTOEAO-specific advantage must be demonstrated against the best matched comparator, not against an artificially memoryless straw model.
Prospective prediction P1: under an identified positive κ and controlled Δj, the next-cycle boundary contrast has the preregistered sign of κΔj. P2: if a registered accessibility threshold lies between the predicted b states, the fixed probe shows a prespecified difference in route accessibility. P3: under the simplified decay regime, the normalized response follows the prefit ρ-dependent curve on held-out intervals. P4: a history-aware workspace improves recovery of prior constraints only when history is causally relevant; it should not necessarily improve tasks with no relevant history.
Evidence that would weaken these claims includes failure of the directional effect, failure of cross-regime transfer, non-identifiability under available sensors, and equal or superior performance by standard history-dependent baselines. Such outcomes would narrow the proposal to a useful documentation and experimental design discipline.
14. Research agenda and conclusion
The next step is a preregistered, instrumented comparison on a well-characterized platform, not a claim of cross-domain universality. The study should estimate turnover-dependent boundary response while controlling occupancy, external input, energy dissipation, thermal effects, and wear; it should then test whether boundary measurement improves held-out predictions of accessibility. A computational implementation can check identification and analysis procedures, but physical claims require independent physical observations. Replication, negative results, and explicit model comparisons are essential.
The principal conclusion is deliberately limited but actionable: equality of measured present occupancy does not, by itself, guarantee equality of future response when a system retains unobserved history. In the proposed model, this possibility becomes a falsifiable boundary-memory hypothesis with specified measurements, contrasts, and failure criteria. The result is a research design connecting multiple TSTOEAO manuscripts to conventional mathematical tools. Its scientific value will depend on whether controlled evidence demonstrates a reproducible, independently measurable effect and whether the proposed formulation predicts outcomes better than simpler alternatives.
Research status. This manuscript presents a theoretical model, conditional derivations, and proposed tests. It does not report completed laboratory experiments, validated cross-domain transfer, or discovery of a universal physical invariant. References to source works retain their original bibliographic descriptions and should be checked against final published versions before external submission.
References
[1] Swygert, J. TSTOEAO: The Swygert Theory Of Everything AO. Foundational introduction. Uploaded volume I, title page and sections on Encoded Equilibrium.
[2] Swygert, J. TSTOEAO III: The Applied Architecture—From Coordinate-Based Simulation To Trust, AI, And Experimental Testbeds. Uploaded book.
[3] Swygert, J. TSTOEAO IV: From Lens To Method—Operationalizing Gradient, Boundary, Correction, Cost, And Equilibrium. Uploaded book.
[4] Swygert, J. TSTOEAO V: The Practice Of TSTOEAO—From Structured Questions To Application, Falsification, And Use. Uploaded book.
[5] Swygert, J. TSTOEAO VI: The Evidence Trail—Original Thought, AI Interpretation, And The Mapping Of What Is Known And Unknown. Uploaded book.
[6] Swygert, J. TSTOEAO VII: The Computational Gauntlet For The TSTOEAO Substrate—Binary, Games, Systems, And Equilibrium Law. Uploaded book.
[7] Swygert, J. (2026, August 2). TSTOEAO Empirical Core v1.0.0: Canonical, Version-Controlled Scientific Specification for Conditioned Expression, Channel-Selective Routing, Structured Correction, and Recursive Boundary Construction. TSTOEAO-EC, v1.0.0, Candidate Canonical Draft.
[8] Swygert, J. (2026, November 12). BRIDGES: From Substrate and Logos to Experience, Mathematics, Engineering, and Reality. Uploaded manuscript.
[9] Swygert, J. (2026, November 12). The TSTOEAO Paradox: The Opposition of Non-Opposition—On the Recursive Nature of Distinction, Expression, and Universal Dichotomy. Uploaded manuscript.
[10] Swygert, J. (2026, November 12). Reciprocal Expression and Non-Expression: Toward a Minimal Mathematical Framework for Distinction, Possibility, and Recursive Conditioning. Uploaded manuscript, §§3–7.
[11] Swygert, J. (2026, November 13). The Space Between Structure: Overlapping Relational Geometries, Boundaries, Pathways, and Flux Across Dimensions. Uploaded manuscript, §§2–6.
[12] Swygert, J. (2026, November 13). Language as an Instrument of Discovery: Articulation, Logical Truth, and Scientific Reasoning. Uploaded manuscript.
[13] Swygert, J. (2026, November 12). Language as Mathematics in Disguise: Relational Structure, Context, Encoding, and the Transformation of Meaning. Uploaded manuscript.
[14] Swygert, J. (2026, November 13). The Evolution of Identity: From Conscious Flux to Expression, Memory, and Self. Uploaded manuscript.
[15] Swygert, J. (2026, November 12). Consciousness at the Boundary of Clarity and Noise: Expression, Non-Expression, Equilibrium, and the Mathematical Value of Experience. Uploaded manuscript.
[16] Swygert, J. (2026, November 12). The Field of Intelligence: Individuality, Shared Expression, Continuity, and Collective Cognition. Uploaded manuscript.
[17] Swygert, J. (2026, November 13). Transformation Through Belief: Boundary-Mediated Pathways of Physiological, Psychological, Behavioral, and Spiritual Adaptation. Uploaded manuscript.
[18] Swygert, J. (2026, November 12). The Fabric of Reality, TSTOEAO, and Reciprocal Expression: Toward a Cumulative Framework of Structure, Relationship, and Dynamics. Uploaded manuscript.
[19] From Distinction to Testable Relational Science. (2026, November 13). Prior eleven-paper independent synthesis; analytical reference only.
[20] Swygert, J. (2026, August 23). Encoded Relational Notation. https://tstoeao.com/2026/08/23/encoded-relational-notation/
[21] Swygert, J. (2026, October 2). Dimensional Expression and Prospective Inference: A TSTOEAO Research Booklet. https://tstoeao.com/2026/10/02/dimensional-expression-and-prospective-inference-a-tstoeao-research-booklet-on-relational-accessibilityinverse-reconstruction-observational-equivalence-and-computational-testing/
[22] Swygert, J. (2026, July 26). The Architecture of Unification TSTOEAO: The Complete Five-Paper Sequence on Candidate Architectural Unification. https://tstoeao.com/2026/07/26/bookket-the-architecture-of-unification-tstoeao-the-complete-five-paper-sequence-on-candidate-architectural-unification/
[23] Swygert, J. (2026, August 4). From Quantum Possibility to Relational Geometry. https://tstoeao.com/2026/08/04/from-quantum-possibilityto-relational-geometry/
Appendix A. Claim classification and provenance
Source doctrine: V = E × Y; EC-1 through EC-4; route and receiver specification; the November reciprocal model; SBS accessibility; identity and shared-workspace update concepts. New derivations: boundary-memory separation, matched-occupancy contrast, impulse-response form, local feedback stability condition, receiver observability diagnostic, and the four-way outcome partition. Conventional mathematical apparatus: linear state-space models, Jacobians, system identification, causal interventions, and equivalence testing. New engineering proposal: auditable, event-sourced shared cognitive workspace. Unresolved: identification of a TSTOEAO-exclusive quantitative invariant across independent physical domains.
Appendix B. Reproducibility and prespecification checklist
• Fix state space, event definitions, and sampling interval before data collection.
• Calibrate b independently; publish receiver sensitivity and detection limits.
• Register two directional fluxes and occupancy; do not replace turnover with net change.
• Record physical costs, heat, energy, and possible common inputs in their correct units.
• Specify null and competing models, parameter bounds, and held-out evaluation.
• Specify the primary endpoint, accessibility threshold, minimum meaningful effect, and equivalence margin.
• Randomize intervention order and preserve all unsuccessful runs.
• Publish code, simulation seed, model versions, and data dictionary subject to privacy constraints.
