UGC/CUR External Academic Reconciliation — Round 3

Version: v0.1
Date: 2026-08-26
作者： Neo.K
機構： EveMissLab／一言諾科技有限公司
Status: External academic stress test / canonical repair proposal
Scope: Grounding & first-cause structure; agency/intervention/causal power; open systems & heterogeneous semantic bridges.

Abstract

Round 3 stress-tests three remaining high-risk regions of UGC/CUR / Unified Closure Theory (UCT):

1. metaphysical grounding, fundamentality, first-cause and regress structure;
2. agency, intervention, causal power, strategic control and responsibility;
3. open-system categorical composition, heterogeneous semantics and institution-level translation.

The result is not a new cosmological conclusion. It is a novelty reclassification and canonical repair package.

The three central repairs are:

and:

rather than a free semantic arrow.

1. Grounding and First-Cause Reconciliation

1.1 Grounding does not collapse into causation or ontological dependence

Rydéhn's work on grounding and ontological dependence argues that, on a standard factive conception of grounding, no known rigid ontological-dependence relation is generally either necessary or sufficient for grounding without substantive bridge assumptions.

UCT should therefore preserve:

A grounding-to-dependence inference is a typed bridge problem.

1.2 A first-cause argument need not be a first-event argument

Pearce's Foundational Grounding and the Argument from Contingency is strong prior art for a move already central to Paper 02: the total causal History should not be explained by simply prepending another causal event to History. Pearce instead proposes a non-causal foundational ground.

This means the following distinction is not by itself novel:

It remains important as a canonical invariant.

1.3 Grounding completeness is not regress termination

Leuenberger distinguishes two conceptions of fundamentality:

• B-fundamentality: being ungrounded;
• A-fundamentality: belonging to every grounding base / participating indispensably in a complete grounding basis.

Oberle further shows that completeness-based fundamentality can be made compatible with self-grounding or mutual grounding if strict-order assumptions on grounding are relaxed.

Therefore the current scalar notion:

is too coarse.

Canonical repair

Replace it by:

Recommended RegressStatus values:

• terminated
• self_grounded
• cyclic
• infinite_candidate
• transfinite_candidate
• open
• unknown

Hence:

and:

1.4 First-Cause Sufficiency Test repair

The First-Cause Sufficiency certificate should not contain an implicit foundationalist thesis.

Recommended structure:

No RegressStatus value automatically proves absolute first causality.

2. Agency, Intervention and Causal Power

2.1 Real-world action is not automatically a model intervention

Jørgensen, Gresele and Weichwald formalize the missing correspondence between actions in the world and mathematical interventions in a causal model. Their central circularity/impossibility result shows that a seemingly natural interpretation can trivialize interventional validity.

UCT should therefore add:

Canonical repair

Introduce:

Without this certificate, an available real-world action cannot serve as an intervention witness merely because the model contains a corresponding $do(\\cdot)$ expression.

2.2 Causal contribution is not strategic or epistemic control

Baier, Funke and Majumdar separate forward responsibility, strategic backward responsibility, and causal backward responsibility. Their examples show that causal responsibility can hold without the knowledge or strategy required for strategic responsibility.

UCT should distinguish:

2.3 Impact is not instrumental control

Carey et al. distinguish response incentives, instrumental control incentives and impact incentives. A variable may be affected by an agent while remaining a side effect rather than an instrumentally controlled objective.

Therefore:

UCT should not infer intentional agency from transformation capability alone.

2.4 Agency is frame-dependent

Abel et al. argue that agency attribution depends on a reference frame containing choices about boundary, causal variables, goal-pursuit criterion, and meaningful adaptation criterion.

Canonical repair

Introduce:

Then:

2.5 Meta-causal agency repair

$\\mathsf{MC}_k$ should be read as rewrite capability relative to a declared law stack and agency frame.

Additional obligations should include:

1. action-to-intervention interpretation;
2. interventional validity;
3. policy/strategy availability;
4. epistemic access;
5. objective/utility relation when intentional control is claimed;
6. agency-frame declaration;
7. rewrite-law provenance.

3. Open Systems, Heterogeneous Composition and Institutions

3.1 Open-system composition is mature prior art

Structured and decorated cospans already provide rigorous interface-based composition of open systems under explicit categorical assumptions. Therefore UCT should not claim categorical open-system composition as a standalone invention.

3.2 Reachability can be strictly coarser than operational semantics

Open Petri-net theory provides a particularly important domain example. Operational semantics retains process structure. Reachability semantics collapses that structure to whether an output marking is reachable from an input marking.

For general composition, reachability is only lax and can miss behaviors involving repeated traversal across the shared interface.

This provides strong external evidence for:

3.3 Open-system bridges remain doctrine-relative

Myers' open dynamical systems and Libkind–Myers' double-operadic theory are highly general, but the specific doctrine, interaction theory, carrier and interfaces remain parameters.

Thus UCT's defensible role is:

rather than a replacement for categorical systems theory.

3.4 Institution-level translation is high-threat prior art

Institution theory and heterogeneous specification already treat different logics as distinct systems connected by explicit morphisms/comorphisms. Such translations carry preservation obligations rather than being free identifications.

Canonical repair

Replace minimal bridge metadata with:

Possible preservation obligations include satisfaction preservation, behavioural-equivalence preservation, reachability soundness/completeness, invariant preservation, provenance preservation, boundary compatibility, conservative extension, compositionality and multi-bridge coherence.

Hence:

and:

4. Novelty Reclassification

4.1 Claims substantially absorbed by prior art

The following should no longer be presented as standalone UCT novelty:

1. grounding as non-causal metaphysical explanation;
2. distinguishing a causal first event from a foundational ground;
3. grounding-base / completeness accounts of fundamentality;
4. intervention-based causal semantics;
5. causal vs strategic vs epistemic responsibility;
6. frame-dependent agency in the broad sense;
7. interface-based categorical composition of open systems;
8. heterogeneous logical translations with preservation conditions.

4.2 Strongest remaining UCT contribution candidate

The candidate novelty is the single audit architecture in which typed promotion across these domains must carry explicit evidence, preservation conditions, obstruction states and outstanding debt.

5. Canonical Repair Table

6. Candidate Patch Obligations for 00–05 / Synthesis

Before a v0.2 theory release:

1. Paper 00: extend bridge schema and proof-obligation matrix.
2. Paper 02: split GroundingComplete; revise $\\mathsf{FCS}_3\\to\\mathsf{FCS}_4$ semantics.
3. Paper 03: add preservation/debt fields for translated ledgers and causal-model interpretation.
4. Paper 04: add AgencyFrame and action/intervention distinction to capability witnesses.
5. Paper 05: index meta-causal agency by frame and law stack; add intervention validity for causal-power claims.
6. Synthesis: redefine Bridge Theory as an evidence/audit meta-protocol that can host categorical, control-theoretic, institution-theoretic and causal-model bridge backends.

7. Core External Literature Used in Round 3

Grounding / first cause

• Stephan Leuenberger, The fundamental: Ungrounded or all-grounding?, DOI 10.1007/s11098-019-01332-x.
• Kenneth L. Pearce, Foundational Grounding and the Argument from Contingency, DOI 10.1093/oso/9780198806967.003.0011.
• Thomas Oberle, Metaphysical Foundationalism and the Principle of Sufficient Reason, DOI 10.1017/S001221732200018X.
• Henrik Rydéhn, Grounding and ontological dependence, DOI 10.1007/s11229-018-1818-4.

Agency / intervention

• Frederik H. Jørgensen, Luigi Gresele, Stefan Weichwald, What is causal about causal models and representations?, arXiv:2501.19335.
• Christel Baier, Florian Funke, Rupak Majumdar, A Game-Theoretic Account of Responsibility Allocation, DOI 10.24963/ijcai.2021/244.
• Ryan Carey et al., Incentives for Responsiveness, Instrumental Control and Impact, DOI 10.1016/j.artint.2025.104408.
• David Abel et al., Agency Is Frame-Dependent, arXiv:2502.04403.

Open systems / heterogeneous translation

• Kenny Courser, Open Systems: A Double Categorical Perspective, arXiv:2008.02394.
• John C. Baez and Jade Master, Open Petri nets, DOI 10.1017/S0960129520000043.
• David Jaz Myers, Double Categories of Open Dynamical Systems, DOI 10.4204/EPTCS.333.11.
• Sophie Libkind and David Jaz Myers, Towards a double operadic theory of systems, arXiv:2505.18329.
• Bartek Klin and Beata Nachyła, Presenting Morphisms of Distributive Laws, DOI 10.4230/LIPIcs.CALCO.2015.190.
• Till Mossakowski, Heterogeneous Theories and the Heterogeneous Tool Set.
• Oliver Kutz and Till Mossakowski, Conservativity in Structured and Heterogeneous Ontologies.

8. Round-3 Assessment

Round 3 reduces novelty claims while increasing formal defensibility.

The major conceptual correction is:

The major agency correction is:

The major bridge correction is:

This should be treated as the canonical research state before any UGC/CUR v0.2 formal extension.