---
title: Non-Trivial Negation in Logic C
url: https://www.emergentmind.com/topics/non-trivial-negation-inconsistent-logic-c
type: topic
---

# Non-Trivial Negation in Logic C

“Non-Trivial Negation Inconsistent Logic C” most precisely denotes the propositional connexive logic \(C\) studied by Wansing and analyzed proof-theoretically in recent work: a logic with strong negation, provable contradictions, and failure of explosion, so that inconsistency does not collapse the consequence relation into triviality [2507.06854]. The designation \(C\), however, is not uniform across the literature. It also appears in da Costa’s \(C\)-systems and in Mortensen-related nomenclature, while closely related research develops Logic \(C\)-style behavior through recovery operators, four-valued LFIs, or calculi with multiple negations [1102.1935]. The unifying theme is controlled reasoning with contradiction: negation remains inferentially robust, but contradictions need not entail arbitrary formulas.

## 1. Nomenclature and family resemblances

The label \(C\) is used for several non-equivalent logical traditions. In Wansing’s usage, \(C\) is a connexive logic with strong negation and a bilateralist proof theory. In da Costa’s tradition, \(C_n\) denotes paraconsistent calculi with explicit control of consistency. In Mortensen-related work, \(C\) occurs in names such as \(C0.2\), and the notation \(cCSL3\) or \(cP2\) marks “connexive variants” rather than da Costa’s hierarchy [2507.06854].

| Usage of \(C\) | Characterization | Source |
|---|---|---|
| Wansing’s \(C\) | Propositional, connexive, negation-inconsistent, yet non-trivial | [2507.06854] |
| da Costa-style \(C\)-family | Paraconsistent calculi with controlled consistency and failure of EFQ | [1102.1935] |
| Mortensen-related \(C\) notations | \(C0.2\), \(cCSL3\), \(cP2\), distinct from da Costa’s \(C_n\) | [2204.06731] |

This terminological plurality matters because the phrase “Logic \(C\)” can otherwise obscure substantive differences. Wansing’s \(C\) is explicitly connexive: it validates Aristotle’s and Boethius’ theses and rejects symmetry of implication. By contrast, da Costa’s systems are organized around paraconsistency and consistency operators, not connexivity as such. Mortensen’s systems introduce yet another axis: truth-functional connexive conditionals over paraconsistent bases [2204.06731].

## 2. Connexive logic \(C\) in the strict sense

Logic \(C\) is a propositional, connexive, negation-inconsistent, yet non-trivial logic extending Nelson’s constructive logic with strong negation \(N4\), but modifying the introduction and elimination of negated implications so as to enforce connexive constraints [2507.06854]. Its primitive signature is
\[
{\sim},\ \wedge,\ \vee,\ \rightarrow.
\]
The connective \({\sim}\) is a strong negation that internalizes a notion of direct refutation.

In the adopted schema, a logic is connexive iff it proves Aristotle’s and Boethius’ theses, together with their converses, and fails symmetry of implication. The characteristic laws of \(C\) are:
\[
{\sim}(A \rightarrow {\sim}A),
\]
\[
{\sim}({\sim}A \rightarrow A),
\]
\[
(A \rightarrow B) \rightarrow {\sim}(A \rightarrow {\sim}B),
\]
\[
(A \rightarrow {\sim}B) \rightarrow {\sim}(A \rightarrow B),
\]
while
\[
(A \rightarrow B) \rightarrow (B \rightarrow A)
\]
is not provable [2507.06854].

The decisive departure from \(N4\) concerns the refutation of implication. In \(N4\), refuting \(A \rightarrow B\) is tied to asserting \(A\) together with refuting \(B\). In \(C\), refuting \(A \rightarrow B\) is conditionalized: one refutes the implication by deriving a refutation of \(B\) from the assumption \(A\). This is the specifically connexive interaction of strong negation with implication.

The system is negation-inconsistent because it proves some instances of both \(A\) and \({\sim}A\), but it is non-trivial because ex contradictione quodlibet fails. A canonical example is the pair
\[
(A \wedge {\sim}A) \rightarrow A,
\]
\[
{\sim}\big((A \wedge {\sim}A)\rightarrow A\big),
\]
both provable in \(C\) [2507.06854]. Thus contradiction is theorematic at some formulas, yet no rule licenses arbitrary \(C\) from \(A\) and \({\sim}A\).

## 3. Bilateral proof theory and proof-theoretic functional completeness

The proof-theoretic analysis of \(C\) is bilateralist. The framework extends the object language \(L\) to higher-order \(R\)-expressions: every formula is an \(R\)-expression; if \(S\) is an \(R\)-expression not of the form \(-T\), then \(-S\) is an \(R\)-expression; and if \(S_1,\dots,S_n,T\) are \(R\)-expressions, then \((S_1,\dots,S_n \Rightarrow T)\) is an \(R\)-expression [2507.06854]. The meta-level symbol “\(-\)” expresses refutation of an \(R\)-expression, whereas object-language strong negation is \({\sim}\). Derivability is written \(\Delta \vdash S\).

The higher-order calculus \(SC_\infty\) contains structural rules such as Reflexivity, Weakening, Permutation, Contraction, and Cut, together with higher-order introduction rules
\[
RI^{+},\ LI^{+},\ RI^{-},\ LI^{-}.
\]
The point of these rules is that refutation is primitive and is not reduced to assertion. This is essential because, in \(C\), connexivity alters the refutation clauses specifically for implication. The key implicational rules are:
\[
R\!\rightarrow^{*}: \text{ from } \Delta,A \Rightarrow B \text{ infer } \Delta \Rightarrow A \rightarrow B,
\]
\[
L\!\rightarrow^{*}: \text{ from } \Delta \Rightarrow A \text{ and } \Delta,B \Rightarrow S \text{ infer } \Delta, A \rightarrow B \Rightarrow S,
\]
and their refutational counterparts
\[
R\!\rightarrow^{*}_{-}: \text{ from } \Delta,A \Rightarrow -B \text{ infer } \Delta \Rightarrow -(A \rightarrow B),
\]
\[
L\!\rightarrow^{*}_{-}: \text{ from } \Delta \Rightarrow A \text{ and } \Delta,-B \Rightarrow S \text{ infer } \Delta, -(A \rightarrow B) \Rightarrow S.
\]

The calculus establishes a strict correspondence between formulas and higher-order sequents:
\[
\vdash A \rightarrow B \Leftrightarrow^{s} A \Rightarrow B,
\qquad
\vdash {\sim}A \Leftrightarrow^{s} -A.
\]
Here \(S \Rightarrow^{s} T\) abbreviates \(S \Rightarrow T\) and \(-T \Rightarrow -S\), and strict equivalence \(S \Leftrightarrow^{s} T\) means mutual \(\Rightarrow^{s}\) [2507.06854].

The functional completeness result is proof-theoretic rather than model-theoretic. The basis \(\{{\sim},\wedge,\vee,\rightarrow\}\) is shown to define any connective \(F\) governed by generalized bilateral rule schemata \((I)\)–\((IV)\). The central device is an encoding \(\overline{(\cdot)}\) from \(R\)-expressions to formulas:
\[
\overline{S}=S \text{ if } S \text{ is a formula},
\qquad
\overline{-S}={\sim}\overline{S},
\]
\[
\overline{\Delta \Rightarrow S}=\overline{\overline{\Delta}} \rightarrow \overline{S}.
\]
The resulting adequacy theorem states
\[
\vdash \overline{S} \Leftrightarrow^{s} S,
\]
and the functional completeness theorem yields a normal-form representation of each admissible \(F(A_1,\dots,A_n)\) as a disjunction of conjunctions built from \(\overline{S_{ik_i}}\) using only the primitive basis [2507.06854]. In this sense, the meaning of the connectives is fixed by inferential role in a strictly bilateralist setting.

## 4. Controlled inconsistency in Logic \(C\)-style calculi

A broader Logic \(C\) tradition is visible in da Costa-inspired paraconsistent systems. The modified systems \(mZ_n\) and \(mCZ_n\) were obtained by removing the axiom \(\neg 1 \Rightarrow 0\) from earlier systems \(Z_n\) and \(CZ_n\), precisely because that axiom made the systems explosive [1102.1935]. The modified calculi preserve a constructive positive fragment \(IPC^{+}\), with language
\[
0,\ 1,\ \wedge,\ \vee,\ \Rightarrow,\ \neg,
\]
and use only Modus Ponens as inference rule.

Their negation is weakened but remains antitone and additive. The key axioms include
\[
(A \Rightarrow B) \Rightarrow (\neg B \Rightarrow \neg A),
\]
\[
1 \Rightarrow \neg 0,
\]
\[
(\neg A \wedge \neg B) \Rightarrow \neg(A \vee B),
\]
with \(mCZ_n\) adding
\[
\neg(A \wedge B) \Rightarrow (\neg A \vee \neg B).
\]
These systems are paraconsistent: neither proves
\[
(A \wedge \neg A) \Rightarrow B
\]
for arbitrary \(B\), while the derivable negative form
\[
(A \wedge \neg A) \Rightarrow \neg B
\]
is explicitly established in the paper as NEFQ [1102.1935]. They therefore instantiate non-trivial negation: contradiction yields controlled inferential effects, but not unrestricted explosion.

A different realization of the same paradigm appears in the four-valued logic \(BD2\), which expands Belnap–Dunn logic by implication and a primitive weak consistency operator \(\copyright\). \(BD2\) is both a Logic of Formal Inconsistency and a Logic of Formal Underdeterminedness [2212.01677]. Its native negation \( - \) is paraconsistent and paracomplete, while a classical negation \(\sim\) is interdefinable with \(\copyright\). The recovery schemata
\[
\copyright \varphi \rightarrow (\varphi \vee -\varphi),
\qquad
\copyright \varphi \rightarrow (\varphi \rightarrow (-\varphi \rightarrow \psi))
\]
show exactly how classical reasoning is recovered in marked contexts [2212.01677]. This is structurally analogous to the way da Costa-style systems separate contradiction-tolerance from controlled recapture of classicality.

## 5. Semantics, recovery, and algebraic organization

A major semantic generalization of Logic \(C\)-style behavior is given by topological Boolean algebras equipped with closure, interior, exterior, border, and frontier operators [2104.04284]. This framework does not axiomatize da Costa’s \(C_n\) directly, but it develops a uniform semantics for LFIs and LFUs with unary recovery operators and explicitly aligns with systems such as \(mbC\) and \(RmbC\).

Two central non-classical negations are defined topologically:
\[
\neg_{p}A := \mathsf{Cl}(\neg A),
\qquad
\neg_{c}A := \mathsf{Int}(\neg A).
\]
The first is paraconsistent and allows truth-gluts; the second is paracomplete and allows truth-gaps. Explosion fails for \(\neg_p\):
\[
A,\neg_p A \nvdash \psi,
\]
and excluded middle fails for \(\neg_c\):
\[
\nvdash A \vee \neg_c A
\]
[2104.04284].

Recovery is internalized by fixed-point operators. For a unary operator \(\chi\),
\[
(fp\,\chi)(A) \iff \chi(A)=A,
\]
and the fixed-point transformation \(\chi^{fp}\) yields a recovery operator. Concrete instances include
\[
R_I(A):=I^{fp}(A),
\qquad
R_B(A):=B^{fp}(A).
\]
Under recovery, classical principles reappear locally. For paraconsistent negation one gets restricted explosion:
\[
(fp\,I)A \wedge A \wedge {}^{C}A \vdash \bot,
\]
and the paper also records recovery of excluded middle, double negation, and contraposition under suitable recovery assumptions [2104.04284]. The resulting “topological cube of opposition” organizes duality, complement, and fixed-point transformations among these operators.

This semantic program is highly formalized. Boolean algebras are encoded as algebras of sets in higher-order logic, and all results are formally verified in Isabelle/HOL. The framework thereby supplies an automation-ready semantics for non-trivial negation, whereas the functional-completeness result for Wansing’s \(C\) deliberately refrains from model-theoretic semantics and proceeds purely proof-theoretically [2507.06854].

## 6. Comparative systems, variants, and prospective extensions

Mortensen’s \(M3V\) provides a striking connexive example of a negation-inconsistent but non-trivial logic. It is obtained by adding a special conditional \(\rightarrow_E\) to \(LP\), with formulas evaluated into \(\{1\},\{0\},\{1,0\}\) in a Dunn/FDE-style setting [2204.06731]. The logic validates unrestricted Detachment,
\[
A,\ A \rightarrow_E B \vDash_{M3V} B,
\]
is connexive with respect to \( \sim \), and is negation-inconsistent because there are formulas \(A\) such that both \(\vDash_{M3V} A\) and \(\vDash_{M3V} \sim A\). A standard witness is \(A \rightarrow_E A\), since both
\[
\vDash_{M3V} A \rightarrow_E A
\]
and
\[
\vDash_{M3V} \sim (A \rightarrow_E A)
\]
hold [2204.06731]. Yet explosion fails:
\[
A,\sim A \nvdash_{M3V} C.
\]
The same paper studies \(cCSL3\) and \(cP2\), two further connexive variants obtained by adding the same \(E\)-conditional, and emphasizes that these Mortensen-related \(C\)-notations are unrelated to da Costa’s \(C_n\) [2204.06731].

A more recent, explicitly stratified approach is \(CPNn\), a propositional calculus with multiple negations [2411.01627]. It introduces a super-negation \( -^{(n)} \), a family of weak negations \( -_{c(k)} \), and switch constants \(L_{c(k)}\). The system is paraconsistent for weak negations:
\[
\not\vdash_{(n)} (\alpha \wedge^{(n)} -_{c(k)}\alpha) \rightarrow^{(n)} \beta
\]
for \(1 \le k \le n-1\), but explosive for the top negation:
\[
\vdash_{(n)} \alpha,\ -^{(n)}\alpha \vdash_{(n)} \beta.
\]
It also validates a form of gentle explosion:
\[
\vdash_{(n)} -_{c(k)}\alpha \rightarrow^{(n)} (L_{c(k)} \rightarrow^{(n)} \alpha),
\]
\[
\vdash_{(n)} -_{c(k)}\alpha \rightarrow^{(n)} (L_{c(k)} \rightarrow^{(n)} -^{(n)}\alpha),
\]
hence
\[
L_{c(k)},\alpha,-_{c(k)}\alpha \vdash_{(n)} \beta.
\]
The paper explicitly compares this mechanism with da Costa’s \(C\)-systems and presents \(CPNn\) as a framework in which classical and non-classical negations coexist [2411.01627].

Across these developments, a common conclusion emerges. Negation-inconsistency need not coincide with triviality; rather, its logical significance depends on how implication, refutation, and recovery are regimented. In Wansing’s \(C\), the decisive factor is the connexive understanding of refutation and the bilateralist treatment of assertion and denial. In da Costa-style and LFI frameworks, it is the presence of controlled recovery principles. In truth-functional systems such as \(M3V\) and \(CPNn\), it is the design of the conditional or the hierarchy of negations. Open directions mentioned or suggested include further study of the connexive conception of refutation and its consequences for quantifiers and modalities, as well as broader investigation of contradictory but non-trivial logics within proof-theoretic semantics [2507.06854].

Source: https://www.emergentmind.com/topics/non-trivial-negation-inconsistent-logic-c