Maximality Inheritance Principle

• Maximality Inheritance Principle is a framework that defines maximal elements by combining local choice selectors on strict upper cones with global least-upper-bound conditions in posets.
• It formalizes the construction of Bourbaki towers to derive fixed points, showing that the presence of a strictly progressive map leads to a contradiction in a maximality-free poset.
• The approach bridges classical Zorn’s Lemma with modern forcing and game theory, offering a reusable proof architecture in order theory, set theory, and stochastic contexts.

The Maximality Inheritance Principle is, in its order-theoretic formulation, a least-upper-bound version of a Zorn-type maximality argument: if a nonempty poset $(P,\leq)$ has least upper bounds for all well-ordered subsets and admits a choice selector on strict upper cones, then $P$ has a maximal element (Liu, 8 May 2026). In the framework of "Bourbaki--Zorn Normal Forms for Maximality Arguments" (Liu, 8 May 2026), this conclusion is obtained by passing through a fixed-point mechanism for progressive self-maps. In other parts of the literature, the same expression is used for the inheritance of forceably necessary truths in forcing and resurrection theory, and for a gap-propagation proof method in stochastic-game verification (Minden, 2020, Phalakarn et al., 8 Aug 2025).

1. Order-theoretic formulation

A partially ordered set, or poset, is a pair $(P,\leq)$ where $\leq$ is reflexive, antisymmetric, and transitive. Writing $x<y$ means $x\leq y$ and $x\neq y$. A map $f:P\to P$ is progressive if

$\forall x\in P,\quad x\leq f(x),$

and strictly progressive if

$\forall x\in P,\quad x<f(x).$

For $P$0, the strict upper cone is

$P$1

The formal Maximality Inheritance Principle isolated in (Liu, 8 May 2026) is the following: suppose $P$2 is nonempty, every well-ordered subset of $P$3 has a least upper bound in $P$4, and there exists a choice selector $P$5 on strict upper cones, meaning that whenever $P$6, one has

$P$7

Then $P$8 has a maximal element.

The principle is called an “inheritance” mechanism because the global least-upper-bound property of the poset, combined with local choice of strict successors, forces maximality. The paper emphasizes that the contribution is methodological rather than axiomatic: it makes explicit a reusable proof architecture connecting Bourbaki–Witt fixed points, strict progression obstructions, and least-upper-bound versions of Zorn-type maximality arguments (Liu, 8 May 2026).

2. Bourbaki towers and the normal form

The central construction is the Bourbaki $P$9-tower generated by a progressive map $(P,\leq)$0. Fix $(P,\leq)$1. A Bourbaki $(P,\leq)$2-tower based at $(P,\leq)$3 is a well-ordered subset $(P,\leq)$4 such that $(P,\leq)$5 is the least element of $(P,\leq)$6; if $(P,\leq)$7 is not the largest element, then its successor in $(P,\leq)$8 is

$(P,\leq)$9

and for each $\leq$0, the least upper bound

$\leq$1

exists in $\leq$2 and belongs to $\leq$3, where

$\leq$4

Thus the tower is closed under successor steps and least upper bounds of initial segments (Liu, 8 May 2026).

Under the hypothesis that every well-ordered subset of $\leq$5 admits a least upper bound, Bourbaki towers based at the same $\leq$6 are linearly ordered by initial segment. The comparison lemma states that if $\leq$7 and $\leq$8 are Bourbaki $\leq$9-towers based at the same $x<y$0, then one is an initial segment of the other. The proof uses the common weak initial part

$x<y$1

where $x<y$2, and rules out branching both at successor stages and at limit stages by uniqueness of least upper bounds (Liu, 8 May 2026).

Because the class $x<y$3 of all such towers is linearly ordered by inclusion, one may define

$x<y$4

This union is itself a Bourbaki $x<y$5-tower. It is well-ordered, successor steps are still generated by $x<y$6, and limit stages remain canonical because the least upper bound of each initial segment is determined inside some constituent tower and therefore also lies in the union. The object $x<y$7 is therefore the largest Bourbaki tower generated by $x<y$8 from $x<y$9 (Liu, 8 May 2026).

3. Terminal fixed points, strict progression, and maximality

The largest tower has a terminal least upper bound that is a fixed point. Writing

$x\leq y$0

the argument in (Liu, 8 May 2026) shows first that $x\leq y$1: otherwise adjoining $x\leq y$2 after all elements of the tower would produce a larger Bourbaki tower, contradicting maximality. Second, $x\leq y$3: progression gives $x\leq y$4, but if $x\leq y$5, adjoining $x\leq y$6 as successor of $x\leq y$7 again enlarges the tower, a contradiction. This is the Bourbaki Tower Normal Form.

An optional strengthening applies when $x\leq y$8 is monotone. If the tower is described transfinally by

$x\leq y$9

and $x\neq y$0 is the stage at which maximality prevents further extension, then

$x\neq y$1

satisfies $x\neq y$2 by a direct Bourbaki–Witt style limit-stage calculation. The paper stresses, however, that its tower-maximality proof obtains the same conclusion without assuming monotonicity (Liu, 8 May 2026).

The fixed-point conclusion immediately yields the strict progression obstruction. If every well-ordered subset of $x\neq y$3 has a least upper bound, then there is no strictly progressive self-map $x\neq y$4. Indeed, a strictly progressive map would generate a largest Bourbaki tower with terminal least upper bound $x\neq y$5 satisfying $x\neq y$6, contradicting $x\neq y$7 (Liu, 8 May 2026).

The Maximality Inheritance Principle then follows by contradiction. Assume $x\neq y$8 has no maximal element. Then $x\neq y$9 for all $f:P\to P$0, so a choice selector $f:P\to P$1 on strict upper cones defines a map $f:P\to P$2. This map is strictly progressive, which is impossible under the least-upper-bound hypothesis. Therefore $f:P\to P$3 has a maximal element. The paper summarizes the proof architecture in the sequence: choice selector $f:P\to P$4 strict progression $f:P\to P$5 Bourbaki tower $f:P\to P$6 fixed point $f:P\to P$7 contradiction (Liu, 8 May 2026).

4. Relation to Zorn’s lemma and Bourbaki–Witt

The framework is explicitly compared with both classical Zorn’s Lemma and the Bourbaki–Witt fixed-point theorem. Classical Zorn assumes that every chain has an upper bound and concludes that there is a maximal element. The Bourbaki–Zorn normal form instead assumes the stronger, canonical condition that every well-ordered subset has a least upper bound. This canonicity at limit stages is what drives the tower comparison lemma: different towers cannot branch because the limit value is uniquely determined by the initial segment (Liu, 8 May 2026).

The similarity to Zorn’s Lemma is that both arguments produce maximal elements from chain-like transfinite constructions. The difference is that the Bourbaki–Zorn normal form uses canonical least upper bounds at well-ordered limits, thereby avoiding global choices at limit stages and simplifying comparison. In Zorn’s setting, only upper bounds are guaranteed, not least upper bounds, so different choices at limits may lead to incompatible extensions unless additional choice or bookkeeping is supplied (Liu, 8 May 2026).

Relative to Bourbaki–Witt, the paper isolates a normal form in a slightly different hypothesis regime. Bourbaki–Witt is usually stated for posets in which every chain has a supremum and for monotone progressive maps, yielding fixed points. Here, least upper bounds for well-ordered subsets suffice to build a largest Bourbaki tower and force a fixed point for progressive $f:P\to P$8, without explicit monotonicity assumptions, by an extremal argument on towers. When $f:P\to P$9 is monotone, the familiar identity

$\forall x\in P,\quad x\leq f(x),$0

reappears as a direct limit-stage computation (Liu, 8 May 2026).

The paper also records typical settings in which the framework applies. These include complete lattices, where every subset has a least upper bound; posets closed under suprema of increasing transfinite sequences; and classical algebraic contexts in which unions of well-ordered chains of substructures provide least upper bounds. By contrast, $\forall x\in P,\quad x\leq f(x),$1 fails the least-upper-bound condition for the well-ordered subset $\forall x\in P,\quad x\leq f(x),$2 itself, so the framework cannot conclude maximality there. The choice requirement is local: full Axiom of Choice is not needed; any choice sufficient to select a strict successor for each non-maximal element suffices (Liu, 8 May 2026).

5. Modal and resurrection-theoretic interpretations

In set theory, “maximality inheritance” is used in a distinct modal sense. For a definable forcing class $\forall x\in P,\quad x\leq f(x),$3, the Maximality Principle is the schema

$\forall x\in P,\quad x\leq f(x),$4

or equivalently, under the usual closure assumptions on $\forall x\in P,\quad x\leq f(x),$5, “every $\forall x\in P,\quad x\leq f(x),$6-forceably $\forall x\in P,\quad x\leq f(x),$7-necessary sentence is already $\forall x\in P,\quad x\leq f(x),$8-necessary” (Minden, 2020). In "Combining Resurrection and Maximality" (Minden, 2020), the Maximality Inheritance Principle is the phenomenon whereby forceably-necessary truths, relative to $\forall x\in P,\quad x\leq f(x),$9, are inherited back to the ground model and preserved across $\forall x\in P,\quad x<f(x).$0-extensions via elementarity “resurrection.”

The formal chain emphasized there is

$\forall x\in P,\quad x<f(x).$1

Here $\forall x\in P,\quad x<f(x).$2 is a resurrection axiom asserting that after any $\forall x\in P,\quad x<f(x).$3-forcing one can force further so that $\forall x\in P,\quad x<f(x).$4, and $\forall x\in P,\quad x<f(x).$5 is a local maximality principle over $\forall x\in P,\quad x<f(x).$6. The implication $\forall x\in P,\quad x<f(x).$7 expresses the inheritance step: if a statement about $\forall x\in P,\quad x<f(x).$8 becomes forceably necessary, then elementarity in a resurrected extension reflects it back to the ground structure (Minden, 2020).

Several other papers place this modal reading in a broader hierarchy. "A Mid Version of Hamkins' Maximality Principle" formulates $\forall x\in P,\quad x<f(x).$9, where $P$00 uses unrestricted forcing modalities but is required to hold in every c.c.c. extension; the paper describes this as a robust inheritance principle for reals under c.c.c. forcing and proves it equiconsistent with the existence of a weakly compact cardinal $P$01 such that $P$02 (Mohammadpour, 2015). "Separating Maximality Principles" shows that no implication between $P$03-$P$04 and $P$05-$P$06 is provable in $P$07, so inheritance across complexity levels fails in general (Gappo et al., 22 Aug 2025). In the potentialist setting, S5-validity at a world is itself called a potentialist maximality principle, and its persistence is highly system-dependent: in countable-transitive-model potentialism it is persistent for a fixed countable language, whereas in rank, Grothendieck–Zermelo, and transitive-set potentialism it is not generally inherited upward because the required correctness properties need not persist (Hamkins et al., 2017).

6. Algorithmic and reverse-mathematical analogues

A further use of the expression appears in stochastic-game verification. "Widest Path Games and Maximality Inheritance in Bounded Value Iteration for Stochastic Games" isolates a proof method called the Maximality Inheritance Principle and applies it to a bounded value iteration algorithm, 2WP-BVI, for finite two-player turn-based stochastic games with reachability objectives (Phalakarn et al., 8 Aug 2025). There the relevant operators are the reachability Bellman operator $P$08, a widest-path Bellman operator $P$09, and the widest-path operator $P$10. The key theorem proves

$P$11

so $P$12 has a unique fixed point at the value function $P$13 (Phalakarn et al., 8 Aug 2025).

The proof uses a gap-maximizer argument. Assuming a gap between $P$14 and $P$15, one chooses a state $P$16 maximizing $P$17. By the max-vs-average lemma, if a maximum equals an average under a probability distribution, then all positively supported successors have the same value. This forces maximality of the gap to be inherited to successors under suitable optimal actions. In the resulting subgame, the inherited positive gap implies that the target set is unreachable, contradicting the positivity of the widest-path width. The same paper presents this as an abstraction of a classical Markov-chain uniqueness proof associated with Baier–Katoen’s Theorem 10.19 (Phalakarn et al., 8 Aug 2025).

In reverse mathematics, the phrase itself does not appear in "Reverse mathematics and equivalents of the axiom of choice", but the paper identifies the closest countable analogues of such maximality principles as the Finite Character Principle (FCP), the Finite Intersection Property maximality principle (FIP), and the closure extension principles CE and NCE (Dzhafarov et al., 2010). In that setting, FCP is the general scheme that every finite-character property admits $P$18-maximal subsets, while FIP and $P$19IP are intersection-maximality schemata. The paper shows that these principles disperse widely across the reverse-mathematical hierarchy: for example, $P$20, FIP is strictly weaker than $P$21 and incomparable with $P$22, full FCP is equivalent to $P$23, $P$24-CE is equivalent to $P$25, and $P$26-NCE is equivalent to $P$27-$P$28 (Dzhafarov et al., 2010).

Across these settings, the common structural theme is not a single theorem but a recurring template: a local extension or persistence condition is converted into a global maximality or uniqueness conclusion. In the order-theoretic normal form, local choice on strict upper cones together with least upper bounds for well-ordered subsets yields a maximal element (Liu, 8 May 2026). In forcing, resurrection turns forceable necessity into truth in the ground model (Minden, 2020). In stochastic games, maximality of a hypothesized fixed-point gap propagates to successors until it contradicts reachability structure (Phalakarn et al., 8 Aug 2025).