Path Identity: Quantum and Categorical Insights

• Path identity is the identification of indistinguishable quantum or categorical paths that enables coherent superpositions and interference effects.
• In quantum mechanics, overlapping modes from separate sources eliminate which-path information, directly inducing entanglement and high-dimensional coherence.
• In type and category theory, path identity underlies the construction of identity types and higher homotopies, supporting robust models of intensional equality.

Path identity refers to the identification or overlap of quantum, categorical, or syntactic paths such that distinct alternatives—be it in quantum experiments, type theory, or category theory—become indistinguishable at the level of observables or judgments. This indistinguishability fundamentally enables phenomena such as quantum interference and the construction of identity types with homotopical content. The path identity principle has independent and deeply developed manifestations in quantum information (where it induces non-classical entanglement and interference) and in the semantics and syntax of intensional type theory (where it serves as the foundation for identity types and higher-dimensional structures).

1. Path Identity in Quantum Mechanics: Foundational Principle and Experimental Realizations

The quantum version of path identity exploits the indistinguishability of alternative creation or propagation processes for quantum particles. The essential idea is that by overlapping output modes of distinct sources so perfectly that no "which-path" or "which-source" information exists (even in principle), a coherent superposition is enforced and quantum interference or entanglement is induced without the need for direct particle–particle interaction or post-selection at a beam splitter. The process is fundamentally different from a mere erasure of which-way information: in path identity, the distinguishing information is never present to begin with, rather than introduced and subsequently erased as in quantum-eraser protocols (Hochrainer et al., 2021).

A canonical form is realized via overlapping the idler or signal modes of two spontaneous parametric down-conversion (SPDC) crystals. Consider the initial state:

$$|\Psi_0\rangle = \frac{1}{\sqrt{2}}\left(|S_A, I_A\rangle + e^{i\varphi_0}|S_B, I_B\rangle\right).$$

Enforcing path identity in, say, the idler mode $I_A \rightarrow I_B$, yields

$$|\Psi_{\rm pid}\rangle = \frac{1}{\sqrt{2}}(|S_A\rangle + e^{i\varphi_0}|S_B\rangle)\otimes|I_B\rangle,$$

so the signal photon is in a coherent superposition of origins—a state not achievable by simply applying a quantum eraser to marked paths (Hochrainer et al., 2021).

Experimental implementations span induced coherence without induced emission, quantum imaging/spectroscopy with undetected photons, and entanglement engineering for multi-photon, multi-mode, and high-dimensional states (Krenn et al., 2016, Hochrainer et al., 2021, Lahiri, 2018). The critical technical requirement is that the spatial, temporal, spectral, and polarization properties of the overlapped paths must match within their respective coherence lengths (e.g., $|\ell_{S_A}-\ell_{S_B}-\ell_{I}| \ll L_{\text{dc}}$ for single photons) (Hochrainer et al., 2021).

2. Many-Particle Interferometry, Entanglement, and the Path Identity Mechanism

The path identity scheme in many-particle interferometry entails the use of two identical sources, each emitting $N$ particles into $N$ distinct modes. By aligning $M$ outgoing modes of these two sources—so that for some subset $l$, $b_l=e^{i\theta_l}b_l'$—the which-source information for the aligned modes is erased. The resulting quantum state after appropriate mode mixing (e.g., applying 50:50 beam splitters with tunable phases to the non-aligned modes) is a coherent superposition, leading to multipartite entangled output (Lahiri, 2018).

Explicitly, the process constructs states of the form

$$|X_N\rangle = \frac{1}{\sqrt{2}}\left(b_1\ldots b_N + e^{i\varphi_0}b_1'\ldots b_N'\right)|\text{vac}\rangle,$$

with path identity on the last $I_A \rightarrow I_B$0 modes, and transforms under mode-alignment and beam-splitter action into generalized Dicke-state superpositions for the remaining $I_A \rightarrow I_B$1 particles.

Key features include:

• The output state can be written as

$I_A \rightarrow I_B$2

where $I_A \rightarrow I_B$3 is a Dicke state of $I_A \rightarrow I_B$4 qubits with $I_A \rightarrow I_B$5 excitations (Lahiri, 2018).

• By choosing which modes are aligned, the degree and structure of multipartite entanglement (e.g., Bell states, GHZ-class states) can be controlled.
• A key observation is that the entanglement nature of the detected state can be switched via manipulation of global phase variables (e.g., $I_A \rightarrow I_B$6), without any interaction with the entangled particles, enabling interaction-free state control (Lahiri, 2018).

3. Quantitative Metrics: Visibility–Entanglement Relationship and Resource Analysis

The operational signature of path identity is an observable relation between interference visibility and quantum entanglement. When the overlap in a "path identity" mode is imperfect—modeled by an attenuator with amplitude $I_A \rightarrow I_B$7—the single- or multi-photon count visibility becomes $I_A \rightarrow I_B$8, and the concurrence or tangle of the quantum state is matched: $I_A \rightarrow I_B$9 for two-photon Bell states, with similar (though more complex) relations in higher dimensions and parties (Lahiri, 2018, Hochrainer et al., 2021).

From an experimental perspective, path identity offers a resource advantage:

• Entanglement swapping protocols typically require multiple prior entangled sources, active Bell-state measurements, and four-fold coincidence detection.
• Path identity enables the direct generation of entangled states from unentangled sources, without Bell-state measurements or even complete ancilla detection: as little as one ancilla photon measurement is sufficient (with reduced fidelity), lowering the required detector count and post-selection overhead (Wang et al., 2024).
• Quantitative metrics illustrate substantial resource reductions for distributed quantum networks, with simplified interferometric stabilization and scalability to higher-dimensional Hilbert spaces (Wang et al., 2024, Bernecker et al., 19 Aug 2025, Kysela et al., 2019).

4. High-Dimensional and Multipartite Entanglement via Path Identity

Path identity generalizes naturally to the engineering of high-dimensional (HD) and multipartite entangled states. By cascading $$|\Psi_{\rm pid}\rangle = \frac{1}{\sqrt{2}}(|S_A\rangle + e^{i\varphi_0}|S_B\rangle)\otimes|I_B\rangle,$$0 sources, each equipped with optical elements to impart defined OAM or other mode shifts, and perfectly overlapping output paths, one constructs, in the low-pump-power regime,

$$|\Psi_{\rm pid}\rangle = \frac{1}{\sqrt{2}}(|S_A\rangle + e^{i\varphi_0}|S_B\rangle)\otimes|I_B\rangle,$$1

where $$|\Psi_{\rm pid}\rangle = \frac{1}{\sqrt{2}}(|S_A\rangle + e^{i\varphi_0}|S_B\rangle)\otimes|I_B\rangle,$$2 identifies both photons in OAM mode $$|\Psi_{\rm pid}\rangle = \frac{1}{\sqrt{2}}(|S_A\rangle + e^{i\varphi_0}|S_B\rangle)\otimes|I_B\rangle,$$3. In the balanced case, this yields maximally entangled qu-dit states of the form

$$|\Psi_{\rm pid}\rangle = \frac{1}{\sqrt{2}}(|S_A\rangle + e^{i\varphi_0}|S_B\rangle)\otimes|I_B\rangle,$$4

with $$|\Psi_{\rm pid}\rangle = \frac{1}{\sqrt{2}}(|S_A\rangle + e^{i\varphi_0}|S_B\rangle)\otimes|I_B\rangle,$$5 (Kysela et al., 2019).

A critical result is that high-fidelity maximally entangled states in dimension $$|\Psi_{\rm pid}\rangle = \frac{1}{\sqrt{2}}(|S_A\rangle + e^{i\varphi_0}|S_B\rangle)\otimes|I_B\rangle,$$6 require careful engineering of the pump and mode-overlap: only with elementary building blocks (i.e., pump OAM $$|\Psi_{\rm pid}\rangle = \frac{1}{\sqrt{2}}(|S_A\rangle + e^{i\varphi_0}|S_B\rangle)\otimes|I_B\rangle,$$7) can $$|\Psi_{\rm pid}\rangle = \frac{1}{\sqrt{2}}(|S_A\rangle + e^{i\varphi_0}|S_B\rangle)\otimes|I_B\rangle,$$8 be reached. Attempting higher $$|\Psi_{\rm pid}\rangle = \frac{1}{\sqrt{2}}(|S_A\rangle + e^{i\varphi_0}|S_B\rangle)\otimes|I_B\rangle,$$9 injects fixed, untuneable mode clusters that cap the global fidelity, setting a theoretical upper bound on the achievable dimension for maximal entanglement with a given source configuration (Bernecker et al., 19 Aug 2025). This underlies the equivalence between multi-crystal path-identity setups and a single source pumped by an engineered spatial superposition—the two are mathematically equivalent for state construction.

The approach extends to multi-photon GHZ and W states, engineered via overlap and appropriate routing such that only particular combinations of crystals and detector firings can contribute to post-selected events. On-chip realizations exploit path identity for multi-photon sources—enabling the omission of HOM interference—and graph-based techniques for the explicit generation of W-structures (Feng et al., 2020).

5. Path Identity in Type Theory, Category Theory, and Homotopy-Theoretic Models

In the field of type theory and higher category theory, path identity relates to the syntactic and semantic structure of identity types, homotopies, and higher groupoids. At the syntactic level, the identity type $|\ell_{S_A}-\ell_{S_B}-\ell_{I}| \ll L_{\text{dc}}$0 is interpreted as the type of computational paths (finite rewrite sequences) connecting $|\ell_{S_A}-\ell_{S_B}-\ell_{I}| \ll L_{\text{dc}}$1 and $|\ell_{S_A}-\ell_{S_B}-\ell_{I}| \ll L_{\text{dc}}$2 within type $|\ell_{S_A}-\ell_{S_B}-\ell_{I}| \ll L_{\text{dc}}$3 (Ramos et al., 2016, Ramos et al., 2015, Ramos et al., 2018). The computational-path or rewrite-based approaches internalize the formation, introduction, and elimination rules for identity as consequences of rewritability, with normalization and confluence given by explicit term rewrite systems (e.g., LND_EQ–TRS), ensuring each path has a canonical representative.

Semantically, the categorical models—spanning path-object categories, homotopy-theoretic models with weak factorization systems, cubical sets, and the more recent algebraic and Hurewicz-fibration–specified approaches—implement path identity as a structure on objects and morphisms. In path-object categories, for every type (object) there is a functorial path-object $|\ell_{S_A}-\ell_{S_B}-\ell_{I}| \ll L_{\text{dc}}$4 with source, target, and structure maps; the identity type in Martin-Löf theory is modeled via the factorization of the diagonal through these path-objects (Berg et al., 2010, Berg, 2016, Awodey et al., 10 Jan 2026, Swan, 2018). The crucial property is that the syntactic category of a type theory with (at least) propositional identity types and a path-object structure is a path category, in the sense of Brown—satisfying closure, lifting, and stability properties that match the logical structure of path identity (Berg, 2016).

An important nuance is that, though in several settings the path objects $|\ell_{S_A}-\ell_{S_B}-\ell_{I}| \ll L_{\text{dc}}$5 (with $|\ell_{S_A}-\ell_{S_B}-\ell_{I}| \ll L_{\text{dc}}$6 an interval object) formally play the role of identity types, there exist models (notably presheaf assemblies over Kleene algebras) in which path types cannot be used directly as identity types without introducing classical principles like LLPO or EM, thus separating the notions of path type and identity type in constructive settings (Swan, 2018).

6. Applications and Open Problems

Path identity underpins foundational advances in both quantum information theory and homotopy type theory. In quantum optics and communication, it provides resource-efficient schemes for distributed entanglement, quantum remote sensing (where target information is "teleported" via path identity without storage (Dalvit et al., 2024)), and quantum imaging and metrology that exploit undetected-path control (Hochrainer et al., 2021, Dalvit et al., 2024). In the mathematical semantics of identity, it enables the construction of models of type theory that are fully stable under context, support intensional equality, and admit a groupoid or higher-groupoid interpretation (Ramos et al., 2015, Ramos et al., 2016).

Outstanding challenges remain: in type theory, construction of a fully constructive model equating path types and identity types with univalent universes and good computational properties is open (Swan, 2018). In quantum experiments, extension to non-photonic systems, high-gain regimes, and further reduction of detection overhead are under active investigation (Hochrainer et al., 2021).

7. Summary Table: Quantum and Categorical Notions of Path Identity

Context	Core Mechanism	Observable/Structural Consequence
Quantum Optics	Overlap of spatial/temporal modes eliminating which-way information	Multipartite and high-dimensional entanglement; interference visibility equals entanglement metric (Lahiri, 2018, Wang et al., 2024)
Homotopy Type Theory	Syntactic/categorical path objects, computational paths or interval object exponentiation	Identity types, higher groupoids, strictly stable models (Ramos et al., 2016, Berg et al., 2010, Awodey et al., 10 Jan 2026)
Algebraic Models/Cubical Sets	Functorial very good path-objects, Kan liftings & Hurewicz fibrations	Direct (and strictly stable) semantics for intensional identity, compatible with univalence (Swan, 2018, Awodey et al., 10 Jan 2026)

Path identity, as both a physical and abstract formalism, forms the backbone of resource-efficient quantum information processing, robust models of intensional equality, and a unifying principle for interference, indistinguishability, and higher homotopy. Its ongoing theoretical development and experimental realization continue to drive progress at the intersection of quantum physics, logic, and mathematics.