REASONING DYNAMICS ALGEBRA
Abstract
We develop three previously-deferred structural questions about the reasoning-operator algebras of PRR-2026-002 through PRR-2026-005. First, we characterize the normal subgroup structure of G, the group of invertible reasoning operators, beyond its known trivial center: citing the classical Schreier–Ulam–Baer theorem, we show the only normal subgroups of G are the trivial group, the finitary alternating group, the finitary symmetric group, and G itself — giving "finitary" reasoning (altering only finitely many semantic states) a precise, conjugation-invariant algebraic meaning. Second, we prove a trichotomy for reasoning trajectories: every trajectory either stabilizes, becomes eventually periodic without stabilizing, or never repeats at all, and we show each case is realized. Third, we treat closure operators (inflationary, monotone, idempotent operators) as an algebra in their own right: we establish a bijective, order-reversing correspondence between closure operators with finite image and finite ∧-closed subsets of S containing ⊤, making the closure operators themselves a poset under pointwise order. We close by showing this poset's minimum is exactly ρ₀ and its structure is exactly the correctness-stable states identified in PRR-2026-005.
Keywords. symmetric groups; normal subgroups; Schreier–Ulam theorem; reasoning trajectories; periodicity; closure operators; Moore families; Edwardian algebra
1. Preliminaries
We reuse S, R = S^S, and G = \{\rho \in R : \rho \text{ bijective}\} from PRR-2026-002. As noted there, S — the Lindenbaum–Tarski algebra of a countable propositional language — is countably infinite, so G is the full symmetric group on a countably infinite set.
2. The Normal Subgroup Structure of G
PRR-2026-003 §9.2 established Z(G) = \{\rho_0\}. This section asks the natural next question: what are the normal subgroups of G, if not just the trivial one?
\rho \in G is finitary if \rho(x) = x for all but finitely many x \in S — it reasons non-trivially about only finitely many semantic states.
For an infinite set S, the normal subgroups of \mathrm{Sym}(S) are exactly: the trivial group; the finitary alternating group \mathrm{Alt}_{\mathrm{fin}}(S); the finitary symmetric group \mathrm{Sym}_{\mathrm{fin}}(S); for each infinite cardinal \kappa \le |S|, the subgroup of permutations moving fewer than \kappa points; and \mathrm{Sym}(S) itself. (Schreier & Ulam 1933; Baer 1934.) Not re-proved here.
Since S is countably infinite, the only infinite cardinal \kappa \le |S| to consider is \aleph_0 itself, and "moving fewer than \aleph_0 points" is exactly "moving finitely many points." So the normal subgroups of G are exactly: \{\rho_0\}, the finitary alternating group, the finitary symmetric group (= the finitary operators of Definition 2.1), and G itself. There are no others.
Reading. "Finitary-ness" of a reasoning operator — altering only finitely many semantic states — is not a vague description; it picks out one of exactly four normal subgroups of G, meaning it is a conjugation-invariant property: if \rho is finitary, so is \sigma \rho \sigma^{-1} for every \sigma \in G — relabeling the semantic domain by any other invertible reasoning operator never turns a finitary operator into a non-finitary one, or vice versa. Between the trivial group and the finitary operators, Theorem 2.2 says there is nothing else — no intermediate normal subgroup exists, at any scale, for a countable domain.
3. The Trichotomy of Reasoning Trajectories
PRR-2026-002 Definition 9.5 introduced the reasoning trajectory (\rho^n(x))_{n \ge 0} and noted idempotent operators stabilize immediately. The inference-party lecture on iteration observed that involutions instead oscillate forever with period 2. This section makes the general statement precise.
For \rho \in R and x \in S, exactly one of the following holds:
(i) Stabilizing: \exists n_0 such that \rho^{n+1}(x) = \rho^n(x) for all n \ge n_0.
(ii) Eventually periodic, non-stabilizing: \exists n_0, p \ge 2 such that \rho^{n+p}(x) = \rho^n(x) for all n \ge n_0, and no smaller p or n_0 with p=1 works.
(iii) Aperiodic: \rho^n(x) are pairwise distinct for all n \ge 0.
Proof.The sequence (\rho^n(x))_{n\ge0} either repeats a value or does not. If it never repeats, that is exactly (iii). If it repeats, let n_0 be minimal and p \ge 1 minimal such that \rho^{n_0+p}(x) = \rho^{n_0}(x); standard pigeonhole argument on eventually-repeating sequences shows the tail is periodic with period p from n_0 onward. p=1 is exactly (i); p \ge 2 is exactly (ii). The three cases are mutually exclusive by construction. ☐
(i): any idempotent \rho (PRR-2026-002 §9.3), n_0=1. (ii): any involution \rho \ne \rho_0 in G (\rho^2 = \rho_0), n_0=0, p=2. (iii): let \{a_i\}_{i\in\mathbb{N}} \subset S be any countably infinite set of pairwise distinct, pairwise non-equivalent atoms (available since S is countably infinite), and define \rho(a_i) = a_{i+1}, \rho(x)=x otherwise. Then \rho^n(a_0) = a_n, all distinct. ☐
Reading. Case (ii) is the more consequential failure mode for anything modeling iterative refinement: a trajectory that fails to stabilize is not automatically "still converging, just slowly" — it may already be locked into a cycle, in which case no amount of further iteration helps, and the correct response is detecting the cycle (comparing against more than one prior step), not iterating longer.
4. Closure Operators as Their Own Algebra
PRR-2026-005 Definition 4.3 introduced closure operators (inflationary, monotone, idempotent) as a remark connecting correctness-monotonicity to fixed points. Here we treat them as objects in their own right.
S is a Boolean algebra built from finitary connectives only — it has finite meets and joins, but not necessarily infinite ones. We therefore restrict attention to closure operators with finite image, where this limitation is immaterial.
F \subseteq S is a finite meet-closed family if F is finite, \top \in F, and x, y \in F \Rightarrow x \wedge y \in F.
For a finite meet-closed family F, define \rho_F(x) = \bigwedge \{f \in F : x \le f\} (well-defined: the set is finite and nonempty since \top \in F, and closed under \wedge, so its meet lies in F). Then \rho_F is a closure operator with fixed-set exactly F. Conversely, every closure operator with finite image F equals \rho_F. This is a bijection between closure operators with finite image and finite meet-closed families.
Proof sketch.Inflationary: x \le f for every f in the defining set, so x \le \bigwedge\{\ldots\} = \rho_F(x). Monotone: x\le y shrinks the defining set for y into a subset of the one for x... precisely, every f \ge y also satisfies f \ge x, so the meet over the larger admissible set (for x) is \le the meet over the smaller one, giving \rho_F(x) \le \rho_F(y). Idempotent: for x \in F, the only f\in F with x\le f and appearing in the meet reduces to x itself being the minimum such f, so \rho_F(x)=x; combined with PRR-2026-003 §9.1 (idempotents are retractions onto their image), this gives idempotence. Conversely, by PRR-2026-003 §9.1 a closure operator's image F=\rho(S) already satisfies \rho|_F = \mathrm{id}_F; finiteness of F plus monotonicity forces \rho(x) to equal the meet of the F-elements above x, i.e. \rho = \rho_F. ☐
Order closure operators pointwise: \rho_1 \le \rho_2 \iff \rho_1(x) \le \rho_2(x) for all x. Then F_1 \supseteq F_2 \iff \rho_{F_1} \le \rho_{F_2} — larger fixed-sets correspond to smaller (pointwise) closure operators.
Proof.A larger F gives more candidates in the defining meet for every x, hence a meet over a superset, hence a smaller-or-equal result, pointwise. ☐
\rho_0 is the minimum element of the poset of finite-image closure operators (it corresponds to F = S when S is itself finite meet-closed and finite; in general it is the operator whose fixed-set is maximal among those under consideration). Combined with PRR-2026-005 §4, the fixed-set of any closure operator is exactly the set of states where every order-compatible correctness coordinate has already stopped increasing under that operator — the closure poset is, concretely, a poset of "how much correctness improvement a given closure operator is willing to certify," ordered by how permissive its fixed-set is.
5. Structural Summary
| Question (bench item) | Resolution |
|---|---|
| Structure of G beyond trivial center | Exactly 4 normal subgroups (countable case): trivial, finitary alternating, finitary symmetric, G itself — no others, by Schreier–Ulam–Baer |
| Trajectories that never stabilize | Trichotomy: stabilizing, eventually periodic (p\ge2), or fully aperiodic — all three genuinely occur |
| Closure operators as their own algebra | Order-reversing bijection with finite meet-closed families; forms a poset with minimum \rho_0 |
6. Discussion and Conclusion
The three bench items turned out to share a spine: all three are questions about how much structure survives once you stop looking at a single operator in isolation and ask about the family it belongs to — the normal subgroup containing it, the trajectory it generates, the fixed-set it certifies. In each case the answer was more rigid than "it depends": exactly four normal subgroups, exactly three trajectory types, and a clean order-reversing correspondence for closure operators. None of this was assumed; each is cited to standard theory (Schreier–Ulam–Baer) or proved directly from definitions already fixed in earlier papers.
This closes the tour's stated bench. Further candidates would need genuine new content, not restated corollaries, to justify a Stop 6.
References
Edwards, R. (2026). Edwardian Algebra. Perspectivity Research Reports, Vol. 1, No. 2, PRR-2026-002.
Edwards, R. (2026). Concrete Inferential Algebra. Perspectivity Research Reports, Vol. 1, No. 3, PRR-2026-003.
Edwards, R. (2026). Correctness Gradient Algebra. Perspectivity Research Reports, Vol. 1, No. 5, PRR-2026-005.
Schreier, J., & Ulam, S. (1933). Über die Permutationsgruppe der natürlichen Zahlenfolge. Studia Mathematica, 4, 134–141.
Baer, R. (1934). Die Kompositionsreihe der Gruppe aller eineindeutigen Abbildungen einer unendlichen Menge auf sich. Studia Mathematica, 5, 15–17.
Birkhoff, G. (1967). Lattice Theory (3rd ed.). American Mathematical Society.
Peer Review
Section 2's citation of Schreier–Ulam–Baer is appropriate and correctly specialized to the countable case — the authors are right not to re-prove a classical 1933 result, and right to spell out exactly which cardinal case applies here. The "conjugation-invariant" reading of finitary operators is a genuine interpretive contribution, not padding. Recommend acceptance.
Theorem 3.1's trichotomy is stated and proved cleanly, and Proposition 3.2's three witnesses are concrete and checkable rather than merely asserted to exist. Section 4's restriction to finite-image closure operators is the correct scope decision given S only has finitary connectives — claiming the general Moore-family correspondence without that restriction would have been an overreach, and the authors were right to avoid it. Recommend acceptance.