Quantum Gromov–Hausdorff Propinquity

• Quantum Gromov–Hausdorff propinquity is a metric framework extending classical Gromov–Hausdorff convergence to noncommutative compact metric spaces while preserving the C*-algebraic structure.
• It employs innovative bridge and tunnel formulations to measure convergence, ensuring that a zero distance signifies a full quantum isometry via *-isomorphisms that maintain Lip-norms.
• Variants such as modular, spectral, and strongly Leibniz propinquity adapt the framework to diverse structures, enabling convergence analysis for quantum tori, AF algebras, and metrized quantum vector bundles.

Quantum Gromov–Hausdorff propinquity is a noncommutative analogue of the Gromov–Hausdorff distance, introduced by Latrémolière to measure convergence of quantum compact metric spaces—unital C$^*$-algebras endowed with Lip-norms—while retaining the C$^*$-algebraic structure. In its dual tunnel formulation, it is a metric up to full quantum isometry; in particular, distance zero is equivalent to a *-isomorphism preserving the Lip-norm, which corrects a central limitation of Rieffel’s original quantum Gromov–Hausdorff distance. The framework has since developed into a family of related metrics, including dual, modular, covariant, spectral, topographic, and strongly Leibniz variants, each adapted to additional geometric or dynamical structure (Latremoliere, 2014, Latremoliere, 2015, Aguilar et al., 2015).

1. Classical origin and quantum compact metric spaces

The classical model is the Gromov–Hausdorff distance $d_{GH}(X,Y)$ between compact metric spaces $(X,d_X)$ and $(Y,d_Y)$, defined by embedding $X$ and $Y$ isometrically into a common metric space and taking the infimum of the Hausdorff distance between their images. In the commutative C$^*$-algebraic picture, a compact metric space $(X,d)$ is encoded by $(C(X),L_d)$, where $^*$0 is the usual Lipschitz seminorm (Latremoliere, 2018, Latremoliere, 2015).

The noncommutative replacement is a Leibniz quantum compact metric space $^*$1, where $^*$2 is a unital C$^*$3-algebra and $^*$4 is a seminorm on a dense Jordan–Lie subalgebra of $^*$5 such that $^*$6 vanishes exactly on the scalars, is lower semicontinuous, satisfies a Leibniz inequality, and induces on the state space $^*$7 the Monge–Kantorovich metric

$^*$8

which metrizes the weak$^*$9 topology (Latremoliere, 2018, Latremoliere, 2014).

Rieffel’s compactness criterion gives equivalent formulations of the weak$d_{GH}(X,Y)$0-metrization property in terms of precompactness of a Lipschitz unit ball modulo scalars. This places noncommutative metric geometry in direct analogy with classical Arzelà–Ascoli compactness and makes Lip-norms the central analytic datum (Latremoliere, 2015).

A persistent motivation for the propinquity is that Rieffel’s quantum Gromov–Hausdorff distance is defined on order-unit spaces and may assign distance zero to C$d_{GH}(X,Y)$1-algebras that are not *-isomorphic. The propinquity was designed to retain the multiplicative structure and to work directly with C$d_{GH}(X,Y)$2-algebras endowed with Leibniz or quasi-Leibniz seminorms (Aguilar et al., 2015, Adams et al., 2023).

2. Bridges, tunnels, journeys, and extent

The bridge-based formulation starts from a bridge

$d_{GH}(X,Y)$3

where $d_{GH}(X,Y)$4 is a unital C$d_{GH}(X,Y)$5-algebra, $d_{GH}(X,Y)$6 and $d_{GH}(X,Y)$7 are unital *-monomorphisms, and the pivot $d_{GH}(X,Y)$8 has nonempty 1-level set. The bridge seminorm is

$d_{GH}(X,Y)$9

Its geometric quality is measured by height and reach, and the bridge length is the maximum of the two (Aguilar et al., 2015, Latremoliere, 2013).

The dual propinquity is formulated with tunnels. A tunnel from $(X,d_X)$0 to $(X,d_X)$1 is a quantum compact metric space $(X,d_X)$2 together with quantum isometries $(X,d_X)$3 and $(X,d_X)$4. The original dual propinquity used reach, depth, and length of tunnels, while the later extent-based variant defines

$(X,d_X)$5

This extent satisfies

$(X,d_X)$6

so it is equivalent to tunnel length up to a factor $(X,d_X)$7 (Latremoliere, 2014).

A technically important point is the evolution from journeys to single-tunnel formulations. The original dual propinquity used finite chains of tunnels, called journeys, to obtain the triangle inequality. The later extent-based variant proves a composition theorem for tunnels: $(X,d_X)$8 which yields the triangle inequality directly and removes the need for journeys in the definition (Latremoliere, 2014).

This evolution clarifies a common misconception. The propinquity is not a single immutable formula; rather, the literature contains equivalent or closely related bridge and tunnel formulations whose differences are motivated by technical control of the Leibniz property, composability, and additional structure such as modules or dynamics (Latremoliere, 2014, Latremoliere, 2015).

3. Foundational properties and relation to classical Gromov–Hausdorff distance

The central structural property is coincidence: the propinquity vanishes exactly at the correct notion of isomorphism. In the dual setting, $(X,d_X)$9 if and only if there exists an isometric isomorphism $(Y,d_Y)$0; in the bridge-based quantum propinquity, $(Y,d_Y)$1 if and only if there exists a *-isomorphism $(Y,d_Y)$2 with $(Y,d_Y)$3 (Latremoliere, 2014, Aguilar et al., 2015).

The tunnel-based and extent-based variants are metrics up to full quantum isometry and satisfy symmetry and the triangle inequality. The newer extent-based metric is bi-Lipschitz equivalent to the original dual propinquity: $(Y,d_Y)$4 Hence the two generate the same topology (Latremoliere, 2014).

Completeness is a major distinguishing feature. The dual propinquity is complete on the class of quasi-Leibniz quantum compact metric spaces for suitable permissible functions, and the equivalent extent-based metric inherits this completeness (Latremoliere, 2015, Latremoliere, 2014). By contrast, the bridge-based quantum propinquity is designed to preserve the C$(Y,d_Y)$5-structure and to dominate Rieffel’s quantum Gromov–Hausdorff distance, but the data block emphasizes completeness most explicitly for the dual formulation (Aguilar et al., 2015, Latremoliere, 2015).

In the commutative case, the propinquity recovers the classical Gromov–Hausdorff topology. For compact metric spaces $(Y,d_Y)$6 and $(Y,d_Y)$7,

$(Y,d_Y)$8

and the map $(Y,d_Y)$9 is a homeomorphism onto its image (Aguilar et al., 2015). Thus the propinquity is not merely analogous to $X$0; it is an actual extension of classical Gromov–Hausdorff convergence to noncommutative compact metric geometry (Latremoliere, 2015).

4. Representative convergence theorems and model families

A large part of the theory is built around explicit convergence results. The following families are canonical.

Family	Quantum metric structure	Propinquity result
Quantum and fuzzy tori	Lip-norms from dual torus actions	Fuzzy tori converge to quantum tori (Latremoliere, 2013)
AF algebras with faithful trace	$X$1	Finite-dimensional stages converge to the AF limit (Aguilar et al., 2015)
AF algebras with Christensen–Ivan metrics	$X$2	Effros–Shen and UHF families are continuous (Adams et al., 2023)
Inductive systems with bridge builders	Uniformly equivalent Lip-norms plus bridge builders	Convergence criterion for inductive limits (Farsi et al., 2022)

For quantum and fuzzy tori, explicit bridges are constructed inside $X$3 using finite-rank diagonal pivots $X$4. The resulting length estimates show that finite-dimensional fuzzy tori converge to infinite-dimensional quantum tori for the quantum Gromov–Hausdorff propinquity (Latremoliere, 2013). This is one of the central explicit bridge constructions in the literature.

For AF algebras with faithful tracial state $X$5, the seminorm

$X$6

is a $X$7-quasi-Leibniz Lip-norm, and the finite-dimensional stages converge to the inductive limit with

$X$8

This gives a systematic finite-dimensional approximation theory for AF algebras and yields the “quantum ultrametric” interpretation in the commutative Cantor case (Aguilar et al., 2015).

The Christensen–Ivan spectral metrics provide a different AF-algebraic model. For a faithful trace $X$9, one defines

$Y$0

and obtains a Leibniz compact quantum metric space. In this setting, the paper on Christensen–Ivan quantum metrics proves continuity, for the propinquity, of both Effros–Shen algebras parametrized by irrationals and UHF algebras parametrized by multiplicity sequences (Adams et al., 2023).

For general inductive systems, a necessary and sufficient condition under uniform equivalence of Lip-norms is formulated in terms of bridge builders, i.e. *-automorphisms of the inductive limit that provide controlled approximation in norm with Lipschitz control. This criterion is sufficient for convergence in the dual propinquity and is used later to pass from quantum compact metric space convergence to spectral propinquity convergence (Farsi et al., 2022).

5. Extensions to modules and spectral data

The modular Gromov–Hausdorff propinquity extends the theory from base algebras to modules endowed with metric data. A metrized quantum vector bundle is a quintuple

$Y$1

where $Y$2 is a Leibniz quantum compact metric space, $Y$3 is a left Hilbert $Y$4-module, and $Y$5 is a norm on a dense subspace with compact unit ball, satisfying inner and modular Leibniz inequalities. Modular bridges add anchors and co-anchors to the base bridge, and the modular propinquity is the infimum of their lengths (Latremoliere, 2018).

Heisenberg modules over quantum $Y$6-tori provide the basic nontrivial example. The paper on metrized quantum vector bundles proves that Heisenberg modules $Y$7, endowed with D-norms built from the canonical Heisenberg action and connection,

$Y$8

are Leibniz metrized quantum vector bundles (Latremoliere, 2017). The continuation of this program proves that, for fixed $Y$9, the corresponding Heisenberg modules form a continuous family for the modular propinquity as $^*$0 varies (Latremoliere, 2018).

The spectral propinquity pushes the framework from Lip-normed C$^*$1-algebras to metric spectral triples. For a metric spectral triple $^*$2, the associated Lip-norm is

$^*$3

Metrical tunnels combine algebraic extents with control of the quantum dynamics induced by the Dirac operators. Under an inductive-limit hypothesis and the existence of a bridge builder that is a full quantum isometry on the limit algebra, convergence in the spectral propinquity follows, implying convergence both of the underlying quantum compact metric spaces and of the unitary dynamics $^*$4 (Farsi et al., 2022).

These module and spectral extensions show that the propinquity is not restricted to “spaces” in a narrow sense. It can also encode noncommutative vector bundles with connection and Dirac-type geometric data, which is why it interfaces naturally with Heisenberg modules, noncommutative solenoids, and Bunce–Deddens algebras (Latremoliere, 2018, Farsi et al., 2022).

6. Proper, covariant, and strongly Leibniz generalizations

For noncompact geometry, the topographic Gromov–Hausdorff quantum hypertopology extends propinquity ideas to pointed proper quantum metric spaces. Here one works with Lipschitz triples $^*$5, where $^*$6 is an abelian topography controlling localization. The resulting global quantity $^*$7 is an infra-metric, vanishes exactly on isometrically isomorphic pointed proper quantum metric spaces, and restricts to the dual propinquity on compact Leibniz quantum compact metric spaces (Latremoliere, 2014).

For dynamical systems, the covariant Gromov–Hausdorff propinquity incorporates actions of groups, monoids, or small categories by Lipschitz maps. One line of work proves that approximate actions of a small category on a convergent sequence of quantum compact metric spaces yield an actual action on the limit space, providing a compactness theorem for actions (Latremoliere, 2017). A later development gives sufficient conditions for convergence of Cauchy sequences in the covariant propinquity and uses them to show completeness of natural classes of Lipschitz dynamical systems (Latremoliere, 2018).

A further refinement is the strongly Leibniz propinquity. Strongly Leibniz Lip-seminorms satisfy an inversion estimate: if $^*$8 is invertible, then $^*$9 stays in the domain and

$(X,d)$0

The strongly Leibniz propinquity $(X,d)$1 restricts tunnels to those whose middle space is itself $(X,d)$2-strongly Leibniz. It is complete on the class of $(X,d)$3-strongly Leibniz quantum compact metric spaces, and it supports an inductive-limit theorem parallel to the ordinary propinquity theory (Aguilar et al., 2023).

This refinement is applied to AF algebras using Frobenius–Rieffel norms associated to faithful conditional expectations. The resulting strongly Leibniz Lip-seminorms make AF inductive systems converge in $(X,d)$4, and they yield continuity, with respect to the irrational parameter, of the Effros–Shen family, as well as continuity of UHF families over the Baire space (Aguilar et al., 2023).

Taken together, these developments show that “quantum Gromov–Hausdorff propinquity” now denotes a coherent metric framework rather than a single isolated distance. Its compact, proper, modular, covariant, spectral, and strongly Leibniz versions organize noncommutative metric geometry into a hierarchy of convergence theories tailored to progressively richer structure, while preserving the decisive coincidence principle: vanishing distance encodes the correct notion of noncommutative isometry.