Analytic Implication
Origin. William Tuthill Parry, "Ein Axiomensystem für eine neue Art von Implikation (analytische Implikation)" (Ergebnisse eines mathematischen Kolloquiums, 1933), written out of Menger's Vienna colloquium: a conditional that holds only when the consequent's content is already contained in the antecedent's. Dunn (1972) gave the first algebraic treatment and a modification of the system. Kit Fine, "Analytic implication" (Notre Dame Journal of Formal Logic, 1986), supplied a semantics and a completeness proof. Angell's analytic containment (1977, 1989) is the sibling relation, and Thomas Macaulay Ferguson's Meaning and Proscription in Formal Logic (2017) places the family alongside Bochvar's logic of nonsense.
Models. Implication as containment of content rather than preservation of truth. A conditional is analytically valid when the antecedent already says everything the consequent says, so nothing in the consequent is new. Parry's formal proxy for content is vocabulary: the consequent may contain no sentence letter absent from the antecedent. This is the strictest of the content-connection requirements — strictly stronger than relevance logic's variable sharing — and it makes the conditional hyperintensional, since classically equivalent consequents can differ in vocabulary.
Formalism.
Proscriptive principle (Parry 1933): If ⊢ A →ᴬᴵ B then var(B) ⊆ var(A). The consequent introduces no new sentence letter. A necessary condition on theoremhood, not a sufficient one.
Comparison with relevance: Variable sharing (Belnap 1960): ⊢ A → B requires var(A) ∩ var(B) ≠ ∅. Proscription implies sharing; sharing does not imply proscription. Proscription is the strictest content condition in the literature.
Valid: A →ᴬᴵ A. A ∧ B →ᴬᴵ A. A →ᴬᴵ A ∨ A. A ∧ ¬A →ᴬᴵ A.
Invalid: A →ᴬᴵ A ∨ B (B new — disjunction introduction fails) A ∧ ¬A →ᴬᴵ B (ex falso fails) A →ᴬᴵ (B → A) (positive paradox fails) A →ᴬᴵ A ∨ (A ∧ B) (though A ≡ A ∨ (A ∧ B) classically)
Fine's semantics (1986): A modal valuation over worlds, plus a content assignment c into a join-semilattice. c of a compound is the join of the contents of its parts; c(¬A) = c(A). So content tracks vocabulary. A →ᴬᴵ B holds iff □(A ⊃ B) and c(B) ⊑ c(A). Two conditions: neither is sufficient alone.
Angell's analytic containment: A ≤ B: B's content is part of A's, read off truthmakers rather than letters. Fine (2016) proves completeness for exact truthmaker content. Not the same relation as Parry's: the criterion is verifier structure, not vocabulary.
Ferguson's connection (2017): Containment logics correspond to Bochvar-style logics of nonsense. A formula with alien vocabulary is not false but meaningless, and what is meaningless cannot be implied.
Symbols.
| Symbol | Unicode | Name | Meaning |
|---|---|---|---|
| →ᴬᴵ | — | Parry's analytic implication | Necessity together with content containment |
| var(A) | — | Vocabulary | The sentence letters occurring in A |
| c(A) | — | Content | The content assigned to A in Fine's models |
| ⊆ | U+2286 | Subset | Vocabulary inclusion; the proscriptive condition |
| ⊑ | U+2291 | Square image of or equal to | Content parthood: c(B) ⊑ c(A) |
| ⊔ | U+2294 | Square cup / Join | Content of a compound from its parts |
| ⊩ | U+22A9 | Forces / Exact verification | State support in the truthmaker reading |
| ⊢ | U+22A2 | Right tack / Derivability | Theoremhood in the calculus |
| □ | U+25A1 | White square / Necessity | The modal half of the truth condition |
| ⊃ | U+2283 | Superset of / Material implication | The truth-functional half |
| ∩ | U+2229 | Intersection | Overlap of vocabularies, in variable sharing |
| ≠ | U+2260 | Not equal to | Nonemptiness of that overlap |
| ≤ | U+2264 | Less-than or equal to | Angell's containment order |
| ≡ | U+2261 | Identical to | Classical equivalence, which the system does not respect |
Metatheory. The proscriptive principle is a syntactic constraint Parry's axiomatization satisfies by construction; the semantic question is what it corresponds to, and Fine (1986) answered it. Models pair a modal valuation with a content assignment into a join-semilattice, A →ᴬᴵ B holding when A necessarily implies B and B's content is a part of A's, and Parry's system is sound and complete for that class. Neither condition alone will do: necessitation without containment is strict implication, containment without necessitation is a merely lexical relation. The consequence relation is therefore subclassical, and the failures are exactly the ones the principle predicts — disjunction introduction, ex falso and the positive paradox all fail, since each licenses a consequent with new vocabulary. Replacement of classical equivalents fails as well, which is what makes the system hyperintensional rather than merely relevant: A →ᴬᴵ A ∨ A is a theorem while A →ᴬᴵ A ∨ (A ∧ B) is not, though the consequents are classically equivalent. Ferguson (2017) showed the containment family corresponds to Bochvar-style logics of nonsense, so proscription and meaninglessness are two readings of one algebra.
Applies to. Content-based accounts of implication and relevance. Truthmaker theory and analytic containment. Subject matter and aboutness. Logics of nonsense and meaninglessness. The analysis of enthymeme and change-of-subject fallacies.
Limitations. The proscriptive principle is a condition on sentence letters, not on content, and the proxy leaks: a conditional whose consequent changes the subject while reusing the antecedent's letters passes unchallenged, and the criterion is not stable under definitional abbreviation, since introducing a defined connective alters the vocabulary and so alters what is derivable. Disjunction introduction, ex falso and the positive paradox all fail, so analytic implication cannot serve as the background consequence relation of any theory whose consequences outrun its premisses' vocabulary. The logic is not congruential — provable equivalents are not intersubstitutable inside the conditional — so the usual Lindenbaum–Tarski quotient is unavailable and algebraic completeness must be obtained by other means. And Parry's, Angell's and Fine's containment relations do not coincide; nothing internal to the idea of content adjudicates between them.
© 2026 Lingenic LLC