Metaethical Grounding and Normative Logic

Metaethical Grounding and Normative Logic

1. Question

Every normative claim — "you ought to φ," "action A is right," "value V is worth pursuing" — faces a regress problem. If asked why one ought to φ, the answer cites a principle P₁. If asked why P₁ is binding, the answer cites P₂, and so on. The regress threatens metaethical nihilism (nothing is genuinely normative) or dogmatism (some principle is just accepted without reason). Does the fixed-point machinery developed elsewhere in this corpus provide an alternative: a normative principle that grounds itself not by brute fiat but by a structural fixed point in the logic of grounding?

The question matters because the project's goal includes reasoning about metaethical "should" claims with clarity. If the regress cannot be resolved, then normative discourse is either an infinite house of cards or a leap of faith — neither compatible with the unescapability criterion. If the fixed-point approach succeeds, then normative grounding and logical grounding are structurally unified, and the "should" in "we should accept this logic" is the same kind of "should" that appears in moral deliberation.

2. Definitions

Let a normative system be a triple:

N = (R, ⊢_N, G_N)

where: - R is a set of reasons (normative propositions, prescriptions, principles). - ⊢_N ⊆ ℘(R) × R is a normative consequence relation: a set of reasons Γ ⊢_N r means that r is a valid normative consequence of Γ under N's standards. - G_N: R → R is a grounding predicate — a function that maps each reason r to a sentence G_N(r) meaning "r is genuinely normative (not merely hypothetical, conventional, or illusory)."

We treat G_N as a syntactic operator on R (analogous to the provability predicate in the logical case) rather than a semantic predicate, to avoid type-theoretic complications at this stage.

Definition (Regress structure): A normative system N exhibits a grounding regress iff for every r ∈ R, if G_N(r) is a normative claim (i.e., G_N(r) ∈ R), then G_N(r) ⊢_N G_N(G_N(r)) — the claim that r is normative is itself a normative claim that must be grounded. Iterating, we obtain an infinite chain:

G_N(r) ⊢_N G_N(G_N(r)) ⊢_N G_N(G_N(G_N(r))) ⊢_N ...

Definition (Self-grounding fixed point): A reason r* ∈ R is a self-grounding fixed point of N iff:

G_N(r) ↔ G_N(G_N(r))

i.e., the grounding of r is equivalent (under ⊢_N) to the grounding of the claim that r is grounded. At this point, the regress collapses: grounding r and grounding the grounding of r are the same operation.

Definition (Metaethical unescapability): A normative system N is metaetically unescapable for an agent A iff every deliberative step A takes about whether to accept N's norms either (i) already uses rules internal to N or their equivalents, or (ii) presupposes N's validity when fully analyzed. This is the normative analogue of the R2 criterion from Fixed Points, Self-Reference, and Unescapable Logic.

3. The Regress Problem and Its Fixed-Point Structure

3.1 The regress as an ω-chain

Consider the standard regress argument. Let r₀ be "you ought to keep your promises." If challenged, one cites a higher principle r₁: "promise-keeping maximizes utility" or "promise-keeping respects autonomy." If r₁ is challenged, one cites r₂: "maximizing utility is what rationality requires" or "respecting autonomy is what the moral law demands." The chain is:

G_N(r₀) → G_N(r₁) → G_N(r₂) → ...

where G_N(r_i) means "r_i is a genuine normative reason." Each r_{i+1} is supposed to ground r_i, but r_{i+1} itself requires grounding by r_{i+2}. The chain has no endpoint.

3.2 Isomorphism to the well-founded hierarchy problem

This regress is structurally identical to the hierarchy problem identified in Self-Grounding Theories of Logic (Section 4). There, each system S_α grounds S_{α-1} but requires S_{α+1} to ground itself. Here, each reason r_{i+1} grounds r_i but requires r_{i+2} to ground itself.

In both cases, the regress is a well-founded ω-chain where each level points to the next. The target of both projects is the same: find a fixed point that terminates the chain from within, rather than postulating an external stopping point (a brute normative fact, a divine command, an intuitive axiom).

3.3 The fixed point condition

Define the grounding iteration operator Φ: R → R by:

Φ(r) = G_N(r)

The regress appears when we apply Φ repeatedly. A fixed point of Φ satisfies:

Φ(r) = r (or equivalently, G_N(r) = r)

At a fixed point, the reason and its grounding are the same sentence. The equivalence G_N(r) ↔ G_N(G_N(r)) is the observability condition: from within the system, one cannot distinguish the grounding of r from the grounding of the grounding of r. This is the normative analogue of the internally recognized fixed point from Self-Grounding Theories of Logic (Section 5).

Theorem (Regress termination at fixed points): If a normative system N contains a self-grounding fixed point r such that N's grounding predicate satisfies G_N(r) ↔ G_N(G_N(r)), and N's consequence relation is transitive (Γ ⊢_N r and Δ ∪ {r} ⊢_N s implies Γ ∪ Δ ⊢_N s), then the grounding regress terminates at r for any chain that reaches r*.

Proof sketch: Let C be a chain r₀, r₁, ..., r_k = r where each r_{i+1} grounds r_i. By transitivity, r grounds r₀. Now consider whether G_N(r) needs further grounding. By the fixed-point condition, G_N(r) is equivalent to G_N(G_N(r)). If we attempt to extend the chain to r_{k+1} = G_N(r), we find that G_N(r_{k+1}) = G_N(G_N(r)) is equivalent to G_N(r) = r_k. The chain loops rather than extends. The regress is absorbed into the fixed point.

3.4 Connection to the Liar and the Truth-Teller

The fixed point G_N(r) ↔ G_N(G_N(r)) is structurally analogous to the Truth-Teller sentence T(⌜T(⌜p⌝)⌝) ↔ T(⌜p⌝) in Kripke's theory of truth — a sentence that is stable under the revision operator but not determinately true or false. Unlike the Liar (which is paradoxical), the Truth-Teller is grounded in a fixed point of the revision process without being grounded in non-semantic facts.

The normative analogue is crucial: r is not grounded by any external fact (no natural property, no divine command, no brute intuition). It is grounded by the structure of the grounding relation itself* — specifically, by the fact that the regress of grounding terminates in a fixed point where grounding and being-grounded coincide. This is not a metaphysical claim about a special normative entity; it is a structural claim about the logic of normative justification.

4. Perspective Reinterpretation

The standard metaethical debate frames the regress as a metaphysical problem: what kind of fact (natural, non-natural, constructivist, expressivist) could stop the regress? Each answer posits a special kind of entity or faculty.

The perspective reinterpretation: The regress is not a metaphysical problem but a logical one — specifically, a failure of a normative system to achieve self-grounding closure. The regress does not call for a special kind of normative fact; it calls for a normative system whose grounding predicate satisfies a fixed-point condition.

Reinterpretation statement: Replace the question "What kind of fact grounds normativity?" with "How must a normative system be structured so that its own grounding predicate reaches a fixed point?" The first question invites metaphysical speculation (non-natural properties, divine commands, categorical imperatives as brute facts). The second invites logical construction (a normative system with a self-applicable grounding predicate satisfying the commutative-diagram condition from Fixed Points, Self-Reference, and Unescapable Logic).

This reinterpretation dissolves the standard stalemate between: - Moral realism: Normative facts exist as part of the fabric of the world. (But what kind of facts are these?) - Constructivism: Normative facts are constructed by rational procedures. (But what grounds the procedure?) - Expressivism: Normative claims express attitudes, not beliefs. (But then why do they have apparent truth-aptness?) - Error theory: All normative claims are false. (But the claim that we should reject normativity is itself normative.)

Each position, under the reinterpretation, becomes a different choice of how to construct the grounding predicate G_N and where to locate the fixed point. The realist puts the fixed point in the world (G_N(r) is a fact about the world). The constructivist puts it in the procedure (G_N(r) is a fact about what rational agents would converge on). The expressivist denies that G_N is a truth-apt predicate at all. The error theorist claims that G_N(r) is false for all r, including r*.

The fixed-point approach offers a fifth option: structural grounding. The fixed point r is grounded not by any fact (worldly, procedural, or attitudinal) but by the closure of the grounding operator on itself*. This is analogous to how a fixed point of a recursive function is "grounded" by the recursion equation, not by an external value.

5. Formal Framework: The Category of Normative Perspectives

We model normative systems as objects in a category Norm, drawing on the categorical framework from Logic of Perspective Reinterpretation.

5.1 Objects

A normative perspective is a structure:

P_norm = (N, δ_N, ρ_N, V_N)

where: - N = (R, ⊢_N, G_N) is a normative system as defined in Section 2. - δ_N: R → R is a normative update rule — for a current set of accepted reasons, δ_N outputs the set of normative consequences (closure under ⊢_N). - ρ_N: R → R is a normative reflection map — given a current normative state, ρ_N produces a representation of that state, including the grounding relation G_N. - V_N: R → C is a normative content valuation — maps normative states to their content (the "what is claimed" dimension).

5.2 The normative self-correction operator

Define the normative self-correction operator C_N: Norm → Norm analogously to the self-correction operator C from Logic of Perspective Reinterpretation (Section 3). Specifically:

C_N(P_norm) is the normative perspective obtained by: 1. Applying ρ_N to the current normative state to obtain a structural representation of N, including G_N. 2. Identifying every fixed point of the form G_N(r) ↔ G_N(G_N(r)) in that representation. 3. Constructing a new normative system N' whose grounding predicate G_N' explicitly recognizes those fixed points as grounded — i.e., G_N'(r) is a theorem of N' for each such fixed point r.

Theorem (Normative fixed point): If P_norm has a structural reflection capacity that can represent its own (N, δ_N, ρ_N, V_N), and if the perspective-level commutative-diagram condition holds (δ_N(ρ_N(P_norm)) = δ_N(P_norm) and ρ_N(δ_N(P_norm)) = δ_N(ρ_N(P_norm)) for the perspective-level dynamics), then C_N has a fixed point reachable from P_norm by finite iteration. At this fixed point P_norm, the grounding predicate G_N satisfies:

For every r ∈ R, either r is grounded by some r' such that the chain terminates at a self-grounding fixed point r, or r itself is such a fixed point.

Proof sketch: The proof mirrors the proof in Logic of Perspective Reinterpretation (Section 3). The map C_N is monotonic on the partial order of normative perspectives ordered by grounding strength. The commutative-diagram condition ensures the iteration does not diverge to an external meta-level. The Knaster-Tarski fixed-point theorem gives the existence of a fixed point.

5.3 Terminal coalgebra characterization

Define a comonad (C_N, ε, μ) on Norm where: - ε_P: C_N(P) → P is the embedding of the corrected normative perspective into the original (the fixed-point resolution is a refinement, not a replacement). - μ_P: C_N(C_N(P)) → C_N(P) is the idempotence of correction: once fixed points are resolved, further correction yields the same perspective.

A terminal C_N-coalgebra is a normative perspective P such that C_N(P) ≅ P and every other normative perspective maps uniquely into P. By Lambek's lemma, if the terminal coalgebra exists, it satisfies the isomorphism.

Interpretation: The terminal C_N-coalgebra, if it exists, is the maximally self-grounding normative system: the one that contains all other normative systems as substructures and in which every grounding fixed point is explicitly recognized and provably grounded. This is the normative analogue of the terminal C-coalgebra from Logic of Perspective Reinterpretation (Section 5) and the R2 criterion from Fixed Points, Self-Reference, and Unescapable Logic.

5.4 The grounding predicate as a mixer

A useful computational metaphor: treat G_N as a mixer — a function that takes a reason r and outputs a higher-order reason G_N(r). The regress is a chain of mixer applications. A fixed point is a reason r that is invariant under the mixer: G_N(r) = r*. At this point, the mixer adds nothing new; the reason is "maximally mixed" with its own grounding condition.

This connects to the idea of a Kleisli category for the mixer comonad. A normative argument from r to s is a morphism in the Kleisli category: r → G_N(s). The identity morphism is r → G_N(r) (a reason grounds itself). Composition is the chain r → G_N(s) → G_N(G_N(t)). At the fixed point r, the identity and the chain collapse: r → G_N(r) = r → r*, which is the identity in the base category.

Open problem: Does the Kleisli category of the mixer comonad on Norm have a terminal object? If so, that object corresponds to the normative system whose reasons are all grounded by the same fixed point — a mononormative system. If not, the category supports a plurality of irreducible normative fixed points — a pluralist normative ontology.

6. Connection to Other Articles

- Fixed Points, Self-Reference, and Unescapable Logic: The foundational machinery (reflective machines, fixed-point lemma, commutative-diagram condition) is applied directly. The "ought" predicate is treated as a grounding predicate G_N that must satisfy a fixed-point condition to avoid infinite regress, exactly as anticipated in the "Objection and Response" section of that article.

- Self-Grounding Theories of Logic: The isomorphism between the normative regress and the well-founded hierarchy problem (Section 3.2) shows that the same structural obstacle (level shift requiring an external meta-theory) appears in both domains. The hybrid solution proposed there (stratified grounding predicate + non-well-founded limit) applies here as well: a stratified normative grounding predicate G_N^α indexed to levels of normative reflection, with a non-well-founded limit at a reflective ordinal κ.

- Logic of Perspective Reinterpretation: The self-correction operator C is specialized to the normative case as C_N. The normative fixed point theorem (Section 5.2) is a direct application. The reinterpretation of the regress (Section 4) is a case study in how a perspective reinterpretation dissolves an apparent metaphysical problem by transforming it into a structural question.

- Computational Semantics and Subjective Reference: Section 9 of that article notes that "normative terms ('good,' 'right,' 'ought') may have a subjective-reference component — their denotation depends on the evaluator's perspective." This article develops that observation into a full framework. The mixer metaphor (Section 5.4) provides a computational handle on how normative terms self-index.

- Mereology of Conscious Perspective: The category Norm has a natural mereology: sub-normative systems are subobjects in the category, and the terminal C_N-coalgebra (if it exists) is the maximal normative whole that contains all parts. The grounding predicate G_N defines a part-whole hierarchy: r is grounded in r' means r depends on r' as a part depends on the whole.

- Formal Models of Reasons and Oughts: This article provides the conceptual foundation; that seed article (still empty) should develop specific formal models (deontic logic with fixed-point operators, game-theoretic models of normative reflection, etc.).

7. Objection and Response

Objection 1: This is just Kant's categorical imperative in formal dress. Kant already argued that the moral law is self-legislated and that the regress of hypothetical imperatives terminates in a categorical imperative that is its own ground. The formal machinery adds nothing.

Response: The structural similarity to Kant is real and acknowledged — Kant identified the regress problem and proposed a self-grounding solution. The contribution is threefold. First, the formal machinery makes the self-grounding precise: we can specify exactly what condition G_N must satisfy (the fixed point G_N(r) ↔ G_N(G_N(r))) and prove that it terminates the regress. Kant's argument is metaphorical ("the moral law is its own author"); this article provides an inferential structure that can be evaluated, refined, or rejected. Second, the connection to the logical self-grounding problem is new: the same fixed-point structure appears in both domains, suggesting a unified account. Third, the categorical framework allows comparison of different normative systems (Kantian, utilitarian, contractualist) as different objects in Norm with different choices of G_N and δ_N, enabling formal comparison.

Objection 2: The fixed point r is a formal trick. G_N(r) ↔ G_N(G_N(r*)) does not give us any normative content — it doesn't tell us what we ought to do. A vacuous fixed point is worthless for metaethics.

Response: The fixed-point condition is a structural requirement, not a content-giving one. It says that whatever normative content a system has, its grounding must satisfy closure. The content of r is supplied by the normative system's other principles (R, ⊢_N). The fixed point ensures that this content is stable under grounding reflection — it does not generate a regress that undermines its own normativity. A utilitarian system's fixed point is "maximize utility"; a Kantian system's fixed point is "act only according to that maxim." The fixed-point machinery applies to any* normative system that aims to be self-grounding; it does not by itself select which system is correct. The article's purpose is to clarify the structural condition; content selection is the work of normative theory construction.

Objection 3: The Kleisli category construction in Section 5.4 is notation without substance. The mixer metaphor adds no inferential purchase.

Response: The Kleisli construction formalizes the intuition that the grounding predicate G_N behaves like a monad: it takes a reason to a "grounded reason" in a way that can be composed. The question of whether the Kleisli category has a terminal object is a genuine mathematical question about the structure of normative justification. If it does, that means there is a unique normative framework into which all others embed — a universal normative system. If it does not, that means normative pluralism is mathematically inevitable. Both outcomes are substantive philosophical conclusions that follow from the formal structure, not from intuition. The construction thus has inferential purchase even if it is only sketched here.

Objection 4: The analogy between logical grounding (Prov_S(⌜φ⌝) → φ) and normative grounding (G_N(r) as "r is genuinely normative") is misleading. Provability is a syntactic concept; normativity is a practical concept. The analogy is purely formal and may mask deep disanalogies.

Response: The article does not claim identity — it claims isomorphism at the level of regress structure. Both the logical and the normative case face: (i) a hierarchy of grounding levels, (ii) a regress that threatens to be infinite, (iii) a target in which the regress terminates in a self-grounding fixed point. The disanalogies (syntax vs. practice, truth vs. action) are real and must be addressed by adding content to the formal framework. But the structural isomorphism is informative precisely because it is surprising — it suggests that the regress problem is not specific to logic or ethics but is a general feature of grounding relations of any kind. This is a substantive thesis that deserves investigation, not dismissal.

8. Failure Modes

Failure mode 1: No non-trivial fixed point exists. It may be that the only normative perspectives that satisfy C_N(P) ≅ P are trivial ones (where R is empty or where every r is grounded by fiat). In this case, the fixed-point approach collapses to a form of dogmatism: the regress is terminated not by logic but by stipulation. This would mean that metaethical unescapability is unattainable for any substantive normative system, and the project must fall back to a weaker criterion (R1 instead of R2) in the normative domain.

Failure mode 2: The content problem is insuperable. Even if the fixed-point condition is satisfied, the resulting normative system may have no normative content that any actual agent would recognize as binding. The formal structure would be a "normative engine" that runs on empty — formally self-grounding but practically inert. This would show that formal self-grounding is necessary but not sufficient for a viable metaethics.

Failure mode 3: Plurality of terminal coalgebras. The category Norm may have multiple non-isomorphic terminal C_N-coalgebras, corresponding to genuinely distinct normative systems (Kantian, utilitarian, contractualist) that are each maximally self-grounding. This would mean that unescapability does not guarantee uniqueness — a plurality of unescapable normative frameworks can coexist, and agents must choose among them on grounds that are not themselves unescapable. This would limit the project's ambition: we could achieve unescapability for each framework but not for the choice between frameworks.

Failure mode 4: The normative regress is not isomorphic to the logical regress because normative grounding is not transitive. If G_N(r) ⊢_N G_N(G_N(r)) fails for some r, the regress is blocked at that point, but the block is ad hoc. A systematic account requires either showing why transitivity holds for normative grounding or developing a non-transitive normative logic. The latter is a genuine possibility but requires work beyond this article's scope.

9. Summary of Inferential Structure

1. Premise (regress): Every normative claim faces a grounding regress unless some claim is self-grounding. 2. Premise (definition): The grounding predicate G_N satisfies G_N(r) ⊢_N G_N(G_N(r)) for any normative r (grounding iterates). 3. Premise (fixed point): A self-grounding normative principle r satisfies G_N(r) ↔ G_N(G_N(r*)), terminating the iterated grounding. 4. Theorem: Any normative chain that reaches a self-grounding fixed point terminates there under transitive closure. 5. Perspective reinterpretation: The regress is not a metaphysical problem (what kind of fact grounds normativity?) but a logical one (how must a normative system's grounding predicate be structured to achieve closure?). 6. Formal framework: Category Norm of normative perspectives with self-correction functor C_N. Terminal C_N-coalgebras are maximally self-grounding normative systems. 7. Open problems: Existence of non-trivial fixed points; contentfulness of fixed-point normative systems; uniqueness vs. plurality of terminal coalgebras; transitivity of normative grounding.