Sasaki Hook

Origin. Usa Sasaki, "Orthocomplemented lattices satisfying the exchange axiom" (1954), where the projection appeared; Finch (1970) and Mittelstaedt (1972) proposed it as the quantum conditional; Gary Hardegree, "The conditional in quantum logic" (1974) and "Material implication in orthomodular (and Boolean) lattices" (1981), gave the argument that it is the one. Herman, Marsden, and Piziak (1975) classified the alternatives.

Models. Orthomodular lattices have no material conditional: a⊥ ∨ b fails the property one wants, since a ∧ (a⊥ ∨ b) ≤ b holds only in the distributive case. There are exactly five polynomial candidates satisfying the minimal requirement that a → b = 1 iff a ≤ b, and the Sasaki hook is the one that survives the further conditions — which is why quantum logic has a conditional at all, and why it is not the classical one.

Formalism.

The hook: a →_S b = a⊥ ∨ (a ∧ b)

The projection: φ_a(b) = a ∧ (a⊥ ∨ b) In Hilbert space: φ_a is the orthogonal projection onto the subspace a, applied to b.

Residuation — the reason it is the right connective: φ_a(b) ≤ c iff b ≤ a →_S c The Sasaki projection is left adjoint to the Sasaki hook. Orthomodular lattices are, in this sense, residuated — with a non-monotone product.

Modus ponens: a ∧ (a →_S b) ≤ b holds in every orthomodular lattice. This is equivalent to orthomodularity, so the connective and the lattice condition are the same fact.

The five candidates (Kotas 1967, Hardegree): Each polynomial p(a,b) with p(a,b) = 1 iff a ≤ b. They coincide exactly on the distributive lattices — that is, in classical logic all five are material implication. Hardegree's criteria (Birkhoff–von Neumann conditional, the counterfactual reading, the Sasaki adjunction) select →_S uniquely.

What fails: Import-export: a →_S (b →_S c) ≠ (a ∧ b) →_S c Transitivity of →_S fails. Contraposition fails. The deduction theorem holds for a single premise and not in general (Malinowski 1990).

Symbols.

Metatheory. The defining property is residuation: φ_a(b) ≤ c iff b ≤ a →_S c, so a →_S − is right adjoint to the Sasaki projection φ_a. Being a right adjoint, a →_S − preserves all existing meets, which is why a →_S (b ∧ c) = (a →_S b) ∧ (a →_S c) holds even though the lattice is not distributive. The hook satisfies a →_S b = 1 iff a ≤ b, the minimal adequacy condition shared by all five candidates, and among them it is the one that arises as a residual. Modus ponens a ∧ (a →_S b) ≤ b holds in an ortholattice iff that lattice is orthomodular, so the connective and the lattice condition are the same fact. In the subspace lattice of a Hilbert space, φ_a(b) is the closure of the image of b under the orthogonal projection onto a, which is exactly Lüders' rule for state update after a yes-outcome measurement. By the Foulis–Holland theorem all five coincide with the material conditional whenever the elements commute, so the classical case is recovered locally.

Applies to. Quantum logic's conditional and the counterfactual reading of measurement. Orthomodular lattice theory. The algebraizability of quantum logic. Quantum computation's assertion logics, where the Sasaki adjunction supplies the weakest precondition.

Limitations. Import-export, transitivity, and contraposition all fail, so the hook supports almost none of the reasoning a conditional is wanted for — chaining two quantum conditionals is not licensed. The deduction theorem holds for a single premise only (Malinowski 1990), so multi-premise entailment is not internalized. The Sasaki product is not monotone in its left argument, so the residuation is one-sided and the connective is not a residuated-lattice implication in the usual sense; correspondingly, Sasaki projections do not compose, φ_a ∘ φ_b ≠ φ_{a ∧ b}, which is where import-export goes. Hardegree's uniqueness argument depends on the criteria chosen, and the other four polynomials have had their defenders. And the physical interpretation is contested at the root: whether a →_S b says anything about a quantum system, or is an artifact of insisting that the lattice carry a conditional, is the question the quantum-logic programme never settled.

© 2026 Lingenic LLC