Complementarity in categorical quantum mechanics

We relate notions of complementarity in three layers of quantum mechanics: (i) von Neumann algebras, (ii) Hilbert spaces, and (iii) orthomodular lattices. Taking a more general categorical perspective of which the above are instances, we consider dagger monoidal kernel categories for (ii), so that (i) become (sub)endohomsets and (iii) become subobject lattices. By developing a `point-free’ definition of copyability we link (i) commutative von Neumann subalgebras, (ii) classical structures, and (iii) Boolean subalgebras.

💡 Research Summary

The paper investigates the notion of complementarity—a central feature of quantum theory—across three traditional mathematical layers and then unifies them within a single categorical framework. The three layers are (i) von Neumann algebras, where complementary observables are modeled by non‑commuting sub‑algebras; (ii) Hilbert spaces, where complementary measurements correspond to mutually unbiased orthonormal bases; and (iii) orthomodular lattices, where Boolean sub‑lattices represent sets of jointly decidable propositions. Although each layer captures the same physical intuition, the mathematical languages are disparate, making it difficult to compare results or to transfer techniques from one setting to another.

To bridge this gap the authors work in the setting of dagger‑monoidal kernel categories. An object of such a category is interpreted as a physical system, morphisms as quantum processes, the dagger as the adjoint (complex‑conjugate transpose), the monoidal product as the tensor product of systems, and kernels as effect‑like sub‑objects. In this environment the endomorphism hom‑sets of an object recover the algebraic structure of (i), the underlying Hilbert space structure of (ii) appears as a concrete representation of the same category, and the subobject lattice of an object reproduces the orthomodular lattice of (iii).

The central technical contribution is a point‑free definition of copyability. Classical structures (or “copy‑comonoids”) have traditionally been defined by the existence of a comultiplication that copies a distinguished basis vector (a “point”). The authors replace the reliance on a specific point with a categorical condition: a morphism is copyable if it admits a kernel‑based factorisation that behaves like a comultiplication on its own image. This definition is intrinsic to the category and does not presuppose any chosen basis.

Using this notion the paper proves three mutually equivalent characterisations of complementarity:

1. Commutative von Neumann sub‑algebras – Within the endomorphism hom‑set of an object, the collection of copyable morphisms forms exactly a commutative von Neumann sub‑algebra. Thus, copyability captures algebraic commutativity.

2. Classical structures on Hilbert spaces – When the abstract category is instantiated as the category of finite‑dimensional Hilbert spaces, copyable morphisms correspond precisely to the usual comultiplications that copy an orthonormal basis. Consequently, the categorical copyability condition recovers the standard definition of a classical structure (or a “basis”) in categorical quantum mechanics.

3. Boolean sub‑lattices of orthomodular lattices – In the subobject lattice of an object, the copyable sub‑objects generate a Boolean algebra. This mirrors the well‑known fact that a set of jointly decidable propositions in quantum logic forms a Boolean sub‑lattice.

The three results are shown to be bijectively related: selecting a commutative von Neumann sub‑algebra determines a unique classical structure, which in turn determines a unique Boolean sub‑lattice, and vice‑versa. Hence, the three apparently distinct manifestations of complementarity are merely different faces of the same categorical phenomenon.

Beyond the conceptual unification, the authors discuss implications for quantum information theory. The point‑free copyability condition delineates precisely the boundary where the no‑cloning theorem becomes active: any morphism that is not copyable cannot be duplicated, reproducing the familiar impossibility of cloning arbitrary quantum states. Moreover, the framework suggests a systematic way to translate results about commutative sub‑algebras (e.g., spectral theorems) into statements about classical structures in categorical quantum protocols, and conversely to import lattice‑theoretic insights into operator‑algebraic contexts.

The paper concludes by outlining future directions: extending the analysis to infinite‑dimensional settings, exploring non‑dagger monoidal categories that capture more exotic quantum phenomena, and applying the unified perspective to the design of quantum programming languages, error‑correcting codes, and categorical models of measurement‑based quantum computation. In sum, the work provides a robust, mathematically rigorous bridge that aligns three historically separate formulations of complementarity, thereby deepening our structural understanding of quantum theory.