Quantum Process Theories

Origin. Abramsky and Coecke (2004). Categorical quantum mechanics. Processes as morphisms. Dagger compact categories. Foundation for quantum protocols.

Models. Systems as objects. Processes as morphisms. Sequential composition: g ∘ f. Parallel: f ⊗ g. Dagger: time reversal. Compactness: cups and caps.

Formalism.

Dagger category: †: C^op → C functor f: A → B gives f†: B → A (f†)† = f (g ∘ f)† = f† ∘ g†

Monoidal structure: ⊗: C × C → C (tensor) I: unit object Associativity, unit coherence.

Compact closed: Every object A has dual A*. η: I → A* ⊗ A (cup) ε: A ⊗ A* → I (cap) Yanking: (ε ⊗ id) ∘ (id ⊗ η) = id

Dagger compact: Dagger + compact + compatibility. (η_A)† = ε_A (cups and caps dual)

Quantum interpretation: Objects: Hilbert spaces Morphisms: linear maps Dagger: adjoint Tensor: tensor product Compact: finite dimensional

Graphical calculus: String diagrams. Wires: systems. Boxes: processes. Topology captures equations.

CPM construction: Completely positive maps. Quantum channels. Environment as purification.

Symbols.

Symbol	Unicode	Name	Meaning
†	U+2020	Dagger	Adjoint/reversal; contravariant involutive identity-on-objects functor
⊗	U+2297	Circled times / Tensor	Parallel composition of systems and processes
∘	U+2218	Ring operator / Composition	Sequential composition
η	U+03B7	Eta / Unit	Cup, I → A* ⊗ A
ε	U+03B5	Epsilon / Counit	Cap, A ⊗ A* → I
→	U+2192	Rightwards arrow	Morphism between objects
≠	U+2260	Not equal	Distinctness of morphisms
ℂ	U+2102	Double-struck C / Complexes	Scalars of the quantum model, not forced by the axioms
*	—	Dual	Dual object
I	—	Monoidal unit	Trivial system; scalars are I → I
C	—	Category	Ambient process theory
FHilb	—	Finite-dimensional Hilbert spaces	Standard model
Rel	—	Sets and relations	Non-quantum dagger compact model
CPM	—	Completely positive maps construction	Mixed-state passage

Metatheory. The graphical calculus is not an informal aid but a theorem. Joyal and Street's coherence result makes string diagrams a sound and complete notation for symmetric monoidal categories, and Kelly and Laplaza (1980) extended coherence to compact closed categories, so two diagrams denote the same morphism exactly when one can be deformed into the other without cutting or rejoining wires. Selinger proved the converse direction for the intended model: finite-dimensional Hilbert spaces are complete for dagger compact closed categories, meaning an equation between morphisms of the free dagger compact category holds in FHilb iff it is derivable from the axioms. Abramsky and Coecke (2004) derived teleportation, entanglement swapping, and dense coding from the yanking equations alone. Selinger's CPM construction (2007) takes any dagger compact category to another whose morphisms are the completely positive maps, with the discarded environment recovering Stinespring dilation. Coecke, Pavlovic, and Vicary (2013) showed that special commutative dagger Frobenius algebras in FHilb correspond exactly to orthonormal bases, so classical data is characterized algebraically inside the quantum setting.

Applies to. Quantum computing. Protocol design. Quantum foundations. Compositional semantics.

Limitations. Compactness is a finite-dimensionality assumption: infinite-dimensional Hilbert spaces have no duals, so cups and caps do not exist, and continuous-variable systems, unbounded observables, and von Neumann algebras fall outside the framework. The axioms do not characterize quantum theory. Rel, the category of sets and relations, is dagger compact and supports the same yanking derivations while being entirely non-quantum, so teleportation-style arguments establish less than they appear; the Born rule, probabilities, and the field ℂ are not consequences of dagger compactness and must be added as extra structure. The pure formalism has no measurement and no classical data — these arrive through Frobenius algebras and the CPM or environment-structure constructions, each an additional postulate. Sums of morphisms are not part of the monoidal structure, so superposition needs enrichment or biproducts. The graphical equalities the calculus makes easy are precisely the topological ones; equations depending on specific linear-algebraic content are not diagrammatic.