The Kantian Peace Machine: Toward a Mechanism-Design Framework for Institutional Peace

Abstract (English) How can institutional design transform a security dilemma into a stable peace equilibrium? This paper proposes a Kantian mechanism-design framework that maps Kant's three Definitive Articles of Perpetual Peace onto a four-layered architecture: objective function, feasibility constraints, a publicity meta-constraint on the design process itself, and a cosmopolitan dual layer. Its bilateral result states the two strict dominance margins under which peace is the unique Nash equilibrium; an admissible institutional configuration is a separate feasibility question. A constructed compact capacity model gives both a positive witness and a budget countercase, without identifying the real institutional feasibility set. The exit-rights mechanism is a bounded proof of concept, while the remaining parameters form a research agenda. A relative-gains transformation requires two model-specific margins; no universal critical α* follows. A reduced feedback system has conditional fixed-point results, not empirically validated dynamics. The contribution is architectural and conditional; empirical calibration, cross-module composition and extension to more than two players remain open. Zusammenfassung (Deutsch) Wie kann institutionelles Design ein Sicherheitsdilemma in ein stabiles Friedensgleichgewicht transformieren? Dieser Beitrag schlägt ein Kantisches Mechanism-Design-Framework vor, das Kants drei Definitivartikel des Ewigen Friedens auf eine vierschichtige Architektur abbildet: Zielfunktion, Zulässigkeitsnebenbedingungen, eine Publizitäts-Metabedingung für den Designprozess selbst und eine kosmopolitische Doppelschicht. Das bilaterale Ergebnis gibt die beiden strikten Dominanzmargen an, unter denen Frieden das einzige Nash-Gleichgewicht ist; ob eine zulässige institutionelle Konfiguration existiert, ist eine getrennte Frage. Ein konstruiertes kompaktes Kapazitätsmodell liefert sowohl einen positiven Zeugen als auch einen Budgetgegenfall, ohne die reale institutionelle Zulässigkeitsmenge zu identifizieren. Exit Rights bleiben ein begrenzter Proof of Concept; die übrigen Parameter bilden eine Forschungsagenda. Eine Transformation relativer Gewinne erfordert zwei modellspezifische Margen; daraus folgt kein universelles kritisches α*. Ein reduziertes Rückkopplungssystem liefert bedingte Fixpunktergebnisse, keine empirisch validierte Dynamik. Der Beitrag ist architektonisch und bedingt; empirische Kalibrierung, Modulkomposition und die Erweiterung auf mehr als zwei Spieler bleiben offen. This is an advanced draft. Major revisions are not excluded. CHANGELOG Changes Version 6.6 (conditional bilateral capacity and current scope, October 2026) Bilateral design condition: The untransformed symmetric two-player base game requires both strict peace-dominance margins at one common point of the specified feasible set on D_0. Monotonicity alone does not establish that such a point exists. Constructed capacity result and countercases: For a specified compact capacity set, a common strict-dominance witness exists exactly when its max-min margin is positive. The illustrative eight-channel budget family yields H > 24/7; the boundary and a separate budget-triangle case limit that conclusion. The 200 mechanical cases are diagnostics, not a universal proof or evidence of real institutional capacity. Consumer and domain alignment: The English and German texts distinguish partial parameter channels from the joint two-margin gate, delimit the other-regarding payoff transformation and relative-gains comparison, and state the reduced feedback-system conclusions conditionally. The additive model is neither a general lower nor an upper bound under interactions. Bilingual files and open work: The current English and German PDFs have 36 and 38 pages. The combined 74-page PDF is an internal control copy only. Empirical calibration, cross-module composition, real feasibility and games with more than two players remain open. Historical context: Earlier changelog entries describe their respective earlier versions; the current abstracts and scoped manuscripts state the present results. Changes Version 6.5 (native German UTF-8 text layer and bilingual style audit, July 2026) German UTF-8 text layer: Replaced TeX-escaped German umlauts in the active German manuscript with native UTF-8 characters, improving searchability, indexing, and copy-paste behavior in the generated PDF without changing the scientific claims. Bilingual style audit: Completed a German/English technical language and translation audit for the v8 manuscript, including terminology parity, structural parity, and epistemic claim-boundary checks. PDF refresh: Rebuilt the English and German PDFs after the audit; the combined PDF remains a local control artifact and is not part of the two-file Zenodo upload set. Scope: This maintenance version is a presentation, encoding, and bilingual-readability update. It does not close the open Phase-2 calibration, data acquisition, or broader model-extension gates. Changes Version 6.4 (formal refinement of Proposition 1 restricted domain and deviation sensitivity, July 2026) Domain Restriction: Proposition 1 (peace as unique Nash equilibrium) is formally specified on the restricted domain D_0 = {Theta in [0,1]^10 : theta_8 = 0}, clarifying that for strong altruistic preferences (theta_8 > 0) mutual war can emerge as a second equilibrium when federation burden exceeds defense gain. Binding Sensitivity Analysis: Corrected the binding role of federation defense burden B versus deviation cost C: B enters only unilateral aggression constraints, whereas C is the sole instrument destabilizing mutual war. Joint Satisfiability Anchor: Added explicit limitation commentary addressing the joint satisfiability (conjunction problem) of idealization assumptions across empirical dyads. DE/EN Synchronization: Synchronized English and German v8 PDF manuscripts cleanly; zero LaTeX compilation errors or missing citations. Changes Version 6.3 (three critical formal fixes from the adversarial review chain, July 2026) Critical: Proposition 1 (peace as unique Nash equilibrium) is now stated on the restricted domain D_0 = {Theta in [0,1]^10 : theta_8 = 0}, because for theta_8 > 0 the payoff function itself is transformed and the proof does not apply. A new remark after the proof shows this restriction is necessary, not merely cautious: for strong other-regarding preferences (alpha(theta_8) near 1) and federation burden B > v(W,P) - v(W,W), mutual war becomes a second Nash equilibrium of the transformed game, because joining a war shields the opponent from the federation burden -- a newly documented theta_8 x theta_2 interaction. Critical: The claim that the federation defense burden B is "the critical design variable" was incorrect and has been replaced by the correct binding analysis: B enters only the unilateral-aggression condition and never activates at mutual war, so only the institutional deviation costs C can destabilize (War, War); B is a complementary instrument. A new sensitivity note in the numerical illustration makes this concrete (both thresholds equal 3; C = 7 alone suffices with B = 0, while C = 2 with B = 4 leaves mutual war as a second equilibrium). Critical: New limitation paragraph "Joint satisfiability of the assumptions (conjunction problem)": Proposition 1 holds under the conjunction of all idealizations, and the paper does not establish that their intersection is non-empty for any empirically relevant dyad. The proposition is now explicitly framed as a characterization of the design target, not as a claim about existing dyads. Minor: DE/EN synchronization of the limitations summary: the German novelty claim and proof-of-concept wording were aligned with the previously softened English formulations. DE/EN: English and German v8 manuscripts rebuilt after the fixes; both compile with zero errors and zero undefined references. Changes Version 6.2 (prepared from local v8 post-v6.0 maintenance, source check and same-day disclosure-policy hotfix, May 2026) Source metadata: Follow-up source check against Crossref, DataCite, Cambridge Core and V-Dem records corrected Kleingeld's Kant and Cosmopolitanism to 2011 and replaced the V-Dem 2026 corporate author with the official author line Nord, Altman, Fernandes, Good God and Lindberg. Major: German synchronization after the v6.0 upload: the theta_3 proof-of-concept now includes the missing companion-paper and sensitivity-reference framing, and the European illustration now carries the intended worst-case and moral-guardrail framing. Minor: Citation integrity rechecked after the literature and DE-sync updates: EN and GER now share 89 citation occurrences, 66 unique citation keys, and 66 bibliography entries, with no missing, duplicate, or unreferenced active keys. Minor: Design and PDF-maintenance pass: language-prefixed Hyperref destinations, improved Unicode text extraction via glyphtounicode/pdfgentounicode, and refreshed local combination PDF metadata. Minor: German AI-disclosure wording was tightened to refer to documented source and citation checks instead of an overbroad independent-verification claim. DE/EN: English and German v8 PDFs rebuilt after the source, maintenance and disclosure-policy checks. Local v6.2 file MD5s: EN 8a07e4b2ef903c4c82240a2b6c9dea3a; GER 69738f6a61328f87994ac3c8bc2d25a9. This version supersedes the same-day v6.1 record 10.5281/zenodo.20338521, which contained the same EN file and the pre-hotfix GER file. Changes Version 6.0 (prepared from local v7, May 2026) Critical: Bibliographic and factual source check updated against DOI, publisher, Crossref and web records. Corrected or finalized entries include Tomz & Weeks (2026), Debre & Sommerer (2026), Davies et al. (2025), Mangini (2024), Banta (2025), and V-Dem (2026). Major: Democratic-peace, autocratization, international-organization,