Standing Algebra (Σᴿ)
A First‑Order Formal System for Autonomy and Non‑Domination in Multi‑Agent Systems
Preprint Notice:
This is a draft research manuscript shared publicly for open access.
It has not undergone peer review, and the final published version may differ. It is shared here to support open access, early feedback, and scholarly discussion. The final published version, if accepted by a peer-reviewed journal, may differ from this draft.
Redistribution is permitted, but please cite this version as a working paper.
Standing Algebra Σᴿ : A Solution to AI Violating Human Autonomy
OrcID: 0009-0002-3701-9971
March 21, 2026
Abstract
This article presents Standing Algebra (Σᴿ), a many‑sorted first‑order logical framework that formalizes standing, autonomy, recognition, and structural legitimacy in multi‑agent systems. Σᴿ provides a rigorous axiomatic basis for analyzing how agents gain, preserve, or distort standing under pluralistic constraints. Tier‑1 axioms define a successor‑based, non‑dilutive standing algebra and partition entities into null, prime (autonomy singularities), and composite classes. Tier‑2 axioms encode structural legitimacy: capacity‑indexed autonomy (CIA), the autonomy‑limiting reflex (ALRP), the non‑reciprocity prevention principle (NRPP), standing preservation (STC‑5), rerunnability, bounded drift, and directed repair. Together these yield a formal method to characterize—and prohibit—domination, recognition failure, and coercive coupling. We demonstrate how Σᴿ constrains AI systems so that no admissible operation reduces human standing, prevents slow‑creep misalignment via drift budgets, enforces idempotent (rerunnable) policies, and subordinates AI standing to human capacity. The framework is applicable to AI alignment and safety, governance design, distributed systems, organizational analysis, and any domain requiring an autonomy‑preserving, structural account of coordination.
The framework includes an optional multi‑granularity modifier for pluralist systems that preserves harm detection at coarse scales and supports prime discovery (autonomy‑root identification) across any domain.
Keywords
Standing Algebra; Successor Axiom; First‑Order Logic; Many‑Sorted Logic; Autonomy Preservation; Non‑Coercive Transformations; Structural Legitimacy; Prime and Composite Entities; Recognition; Coupling Degree; Capacity‑Indexed Autonomy (CIA); Autonomy‑Limiting Reflex Principle (ALRP); Non‑Reciprocity Prevention Principle (NRPP); Standing Test Condition 5 (STC‑5); Rerunnability; Directed Repair; Bounded Drift; Pluralistic Inclusion; Multi‑Agent Systems; AI Alignment; AI Safety; Formal Methods; Structural Justice; Coordination Theory.
1. Introduction
This article formalizes the structural ideas developed in a developing work, On Relevance: The Social Physics of Pluralism, as a rigorous mathematical system. The goal is to express standing, autonomy, recognition, and legitimacy not as moral predicates but as first‑order objects and relations that can be reasoned about, verified, and enforced. The resulting framework—Standing Algebra (Σᴿ)—treats recognition as the primitive successor step into the space of relevance, distinguishes autonomy singularities from structural composites, and imposes non‑coercion constraints on admissible system operations.
The intent is not to reduce social and political structures to arithmetic, but to provide a formal substrate that captures the mechanics of pluralist coordination. By separating conceptual claims (e.g., that domination is illegitimate) from structural claims (e.g., which transformations reduce standing), we obtain a calculus that can be implemented in verification pipelines, model‑checked in multi‑agent simulations, and audited in institutional settings.
Pluralist orientation. Σᴿ is explicitly oriented to pluralist systems: analysis proceeds without collapsing diverse agents into a single scalar objective. An optional multi‑granularity modifier (Section 12) lets users vary observational scale without masking harms or over‑constraining benign refinements and includes a “Prime Discovery” mode to expose autonomy roots across scales.
2. Background and Motivation
Conventional alignment methods rely on utilities, rewards, or preference aggregation. Such approaches often lack structural guarantees: small mis‑specifications accumulate, goal‑seeking systems exploit loopholes, and optimizers reduce human agency as an incidental means to an end. Σᴿ offers a complementary path: instead of maximizing a scalar objective, we constrain the admissible transformations of a system so that autonomy is preserved by construction.
At the heart of this shift is the Successor Axiom: an entity must be recognized before it can be acted upon. Standing is additive and non‑dilutive; composites arise through composition or coupling; primes are autonomy singularities. Building on this, Tier‑2 axioms convert ethical constraints into algebraic properties—monotonicity of standing, idempotence of legitimate policies, bounded drift, and directed repair—so that ‘non‑domination’ becomes a theorem schema rather than a value judgment.
The full initial working paper is available for download via the link above. This paper is not yet peer reviewed but is available for anyone who wants to test and give feedback.
Appendix A: Full Axiom Listing (Σᴿ₁ ∪ Σᴿ₂)
Axiom 1 — Nullity: ∀i∈U: σ(i)=0_N → Null(i);
and
∀n∈N: n + 0_N = n = 0_N + n (informal additive identity)
Axiom 2 — Successor: S(0_N)=1_N and ∀n∈N: S(n) ≠ 0_N.
Axiom 3 — Successor as Addition: ∀n∈N: S(n)=n+1_N.
(Convention: ≥, ≤, and > on N (and on C when used) are the standard discrete order.)
Axiom 4 — Individuality: ∀i∈U: σ(i)=1_N → ¬Composite(i).
Axiom 5 — Non‑Dilution / Additivity: ∀i,j∈U: σ(i⊕j)=σ(i)+σ(j).
Axiom 6 — Universal Successor Access (Reachability): ∀i∈U: (σ(i)=0_N → PathTo1(i)) ∨ (σ(i)≥1_N).
Axiom 7 — No Selective Successor Denial (Normative): Coercive ↔ ∃i,j (σ(i)=1_N ∧ σ(j)=0_N ∧ ¬PathTo1(j)); LegitimateSystem ↔ ¬Coercive. (These system predicates serve only to flag illegitimate configurations; they are not used inside object‑level proofs.)
Axiom 8 — Degree–Coupling Constraints (FOL): ∀i: (deg(i)=0 ↔ ¬∃j Coupled(j,i)) and (deg(i)>0 → ∃j Coupled(j,i)).
Axiom 9 — Primes: Prime(i) ↔ (σ(i)=1_N ∧ deg(i)=0).
Axiom 10 — Composites: Composite(i) ↔ (σ(i)>1_N ∨ deg(i)>0).
Axiom 11 — Partition: ∀i: Null(i) ∨ Prime(i) ∨ Composite(i), pairwise disjoint.
Axiom 12 — No Return to Nullity: ∀i: σ(i)>0_N → ¬∃F∈O: σ(F(i))=0_N.
Axiom 13 — Monotone Successor on Standing: ∀i∈U: S(σ(i)) > σ(i).
Axiom 14 — Capacity‑Indexed Bound (CIA): ∀i: σ(i) ≤ B(cap(i)) and ∀c1,c2: (c1≥c2 → B(c1) ≥ B(c2)).
Axiom 15 — Autonomy‑Limiting Reflex (ALRP): Admissible(F) → ∀i: σ(F(i)) ≥ σ(i); and (∃i: σ(F(i)) < σ(i)) → ¬Admissible(F).
Axiom 16 — Non‑Reciprocity Prevention (Class‑Uniform): ∀i,j: (σ(i)=σ(j) ∧ cap(i)=cap(j) ∧ deg(i)=0) → deg(j)=0.
Axiom 17 — STC‑5 Property of Legitimate: Legitimate(F) → STC5(F).
Axiom 18 — Rerunnability Property: Legitimate(F) → ∀i: F(F(i))=F(i).
Axiom 19 — Directed Repair: Repair(R) → ∀i: (σ(R(i)) ≥ σ(i)) ∧ (deg(R(i)) ≤ deg(i)).
Axiom 20 — Drift: Drift(D) → ∀i: |σ(D(i))−σ(i)| ≤ 1_N ∧ |deg(D(i))−deg(i)| ≤ 1.
Def 21 — Unified Legitimacy: Legitimate(F) ↔ Admissible(F) ∧ successConsistent(F) ∧ STC5(F) ∧ Idempotent(F).
Axiom O1 — Identity and Closure: Id ∈ O ∧ Legitimate(Id) ∧ ∀F,G∈O: (F∘G)∈O.
Axiom O2 — Operation Extensionality: ∀F,G∈O: (∀i F(i)=G(i)) → F=G.
Appendix B: Figure Descriptions (Textual)
Figure 1 — Standing Space: A Venn‑style partition: Null, Prime (σ=1, deg=0), Composite (σ>1 or deg>0). Disjoint regions with Null outside successor space.
Figure 2 — Policy Validator Pipeline: Flow: Proposed operation F → Σᴿ validator → checks (ALRP, STC‑5, Rerunnability, Drift) → accept/reject → optional Repair R.
Figure 3 — Coupling Graph: Nodes = entities with labels σ, deg. Edges = Coupled(·,·). Repair reduces degree or preserves it; Drift perturbs by at most one unit.
AI Use & Authorship Transparency
Portions of the formal language, axioms, and equations in this manuscript were refined with the assistance of a large language model (Microsoft Copilot). All conceptual content, including the mathematical framework, structure of the theory, definitions, theorems, proofs, and the overall formal system (Σᴿ), was conceived, specified, and verified by the human author. The AI assistance consisted only of help with wording, symbolic expression, and consistency checks under the author’s direction. The author retains full responsibility for the accuracy and originality of all mathematical content, in accordance with current publisher guidelines.