Smolin's Weak Coupling Limit in Euclidean Gravity

• Smolin’s weak coupling limit is a regime where the nonabelian SU(2) gauge structure in Euclidean gravity abelianizes to U(1)^3 by taking G→0.
• The model preserves a nontrivial Hamiltonian constraint algebra with structure functions, offering a simplified yet dynamically rich framework for loop quantum gravity research.
• It supports polymer quantization with charge-network states and anomaly-free dynamics, bridging classical Euclidean gravity with effective U(1)^3 quantum formulations.

Searching arXiv for primary sources on Smolin’s weak coupling limit in Euclidean gravity and LQG. Searching "Smolin weak coupling limit Euclidean gravity U(1)^3 Tomlin Varadarajan" Searching arXiv for: "Smolin weak coupling limit Euclidean gravity U(1)^3" Smolin’s weak coupling limit is the $G\to 0$ limit of Euclidean gravity in which the internal nonabelian $\mathrm{SU}(2)$ gauge structure abelianizes to a generally covariant $U(1)\times U(1)\times U(1)$ gauge theory. In the Ashtekar formulation, one rescales the connection by Newton’s constant and then sends $G_{\mathrm N}\to 0$; the resulting theory retains Gauss, diffeomorphism, and Hamiltonian constraints, and, crucially, its Hamiltonian-constraint algebra still closes with structure functions. For that reason, the model has been used as a particularly useful toy model for the Hamiltonian-constraint problem in loop quantum gravity: it is much simpler kinematically than full gravity, but it preserves the non-Lie character of the constraint algebra that makes the quantum dynamics difficult (Tomlin et al., 2012).

1. Classical definition of the $G\to 0$ limit

The starting point is Euclidean canonical gravity in Ashtekar-like variables $(E_i^a,\mathcal A_a^i)$. Smolin’s construction rescales the connection according to

$A_a^i := G_{\mathrm N}^{-1}\mathcal A_a^i .$

With this rescaling, the curvature becomes

$$\mathcal F_{ab}^i = G_{\mathrm N}\Big(\partial_aA_b^i-\partial_bA_a^i+G_{\mathrm N}\epsilon^i{}_{jk}A_a^jA_b^k\Big),$$

and the gauge-covariant derivative becomes

$$\mathcal D_aE_i^a=\partial_aE_i^a+G_{\mathrm N}\epsilon_{ijk}A_a^jE_k^a.$$

In the limit $G_{\mathrm N}\to 0$, the nonabelian terms disappear. The resulting action is that of a generally covariant $\mathrm{SU}(2)$0 gauge theory with constraints

$\mathrm{SU}(2)$1

$\mathrm{SU}(2)$2

$\mathrm{SU}(2)$3

where

$\mathrm{SU}(2)$4

The limit therefore preserves the canonical form of Euclidean gravity while replacing $\mathrm{SU}(2)$5 by three independent Abelian factors. This is the precise content of the abelianization: the theory is not ordinary electrodynamics, but a generally covariant $\mathrm{SU}(2)$6 theory whose constraints continue to encode gravity-like dynamics (Tomlin et al., 2012).

2. Constraint algebra and the density-$\mathrm{SU}(2)$7 formulation

The importance of Smolin’s limit lies less in its kinematics than in its algebra. In the $\mathrm{SU}(2)$8 model, the Gauss and diffeomorphism brackets take the expected form, while the Hamiltonian brackets still produce a diffeomorphism-type generator with phase-space-dependent structure functions. In the Tomlin–Varadarajan line of work, the Hamiltonian constraint is written with density weight $\mathrm{SU}(2)$9,

$U(1)\times U(1)\times U(1)$0

with electric shifts

$U(1)\times U(1)\times U(1)$1

On the Gauss-law surface, the Hamiltonian bracket can be written schematically as

$U(1)\times U(1)\times U(1)$2

and this can be re-expressed in terms of “electric diffeomorphism” constraints (Varadarajan, 2012).

That reformulation is central. It shows that the model preserves the most difficult structural feature of gravity—the non-Lie nature of the Hamiltonian constraint algebra—even after the internal gauge group has simplified to $U(1)\times U(1)\times U(1)$3. The choice of density $U(1)\times U(1)\times U(1)$4 is equally central: in this quantization program it is not cosmetic, but the device that makes a nontrivial continuum action possible in a polymer representation. The model is therefore simpler than full $U(1)\times U(1)\times U(1)$5 gravity, but not simplified in the one place that matters most for the anomaly problem.

3. Polymer quantization and charge-network dynamics

The quantum kinematics is the $U(1)\times U(1)\times U(1)$6 analogue of the holonomy-flux representation of loop quantum gravity. The basis states are charge-network states, with each edge labeled by an integer triplet $U(1)\times U(1)\times U(1)$7, subject to gauge invariance at vertices. The corresponding holonomy is

$U(1)\times U(1)\times U(1)$8

and gauge invariance requires the net outgoing charge to vanish for each $U(1)\times U(1)\times U(1)$9 factor (Varadarajan, 2019).

In the anomaly-free construction, the Hamiltonian acts at nontrivial vertices and produces graph deformations. At finite triangulation, its action has the form

$G_{\mathrm N}\to 0$0

where $G_{\mathrm N}\to 0$1 is the nontrivial vertex, $G_{\mathrm N}\to 0$2 is a regulating coordinate patch, $G_{\mathrm N}\to 0$3 is the inverse-volume eigenvalue, and $G_{\mathrm N}\to 0$4 is the deformed charge network (Varadarajan, 2012).

The deformation is not merely a local relabeling. The Hamiltonian displaces a vertex a coordinate distance $G_{\mathrm N}\to 0$5 along one of its outgoing edges, flips the charges according to the internal index $G_{\mathrm N}\to 0$6, and inserts kink structures on edges. This is the mechanism by which the second Hamiltonian in a commutator can act nontrivially: the first action creates a new displaced nontrivial vertex, and the second acts there rather than trivially at the original location. In this respect the model departs sharply from older ultralocal pictures of loop-quantum-gravity dynamics.

4. Diffeomorphism covariance, conicality, and anomaly freedom

A major technical difficulty in this program is that the regulated Hamiltonian is defined using coordinate patches. The diffeomorphism-covariance condition is

$G_{\mathrm N}\to 0$7

but the original Tomlin–Varadarajan construction used regulating patches in a way that appeared to conflict with this requirement. The resolution was to choose, for each diffeomorphism class, a reference charge network $G_{\mathrm N}\to 0$8 with reference coordinates $G_{\mathrm N}\to 0$9, and then define the coordinates of any diffeomorphic image $G\to 0$0 by pushforward,

$G\to 0$1

With an appropriate choice of the reference patch, every symmetry of the reference charge network acts on the tangent space as

$G\to 0$2

that is, as a positive scalar times a rotation (Varadarajan, 2012).

This covariant choice of patches generates a new problem: the commutator of two Hamiltonians diverges because the Jacobian between the original coordinates and the once-deformed coordinates degenerates as the regulator is removed,

$G\to 0$3

The cure has three linked ingredients. First, the deformations are modified so that the edge tangents at the displaced vertex become conical. Second, the “scrunching” diffeomorphisms are refined so that this conical structure is preserved under regulator flow. Third, the vertex-smooth algebraic states are replaced by a refined distributional topology in which the vertex function has a controlled short-distance singularity,

$G\to 0$4

With these changes, the continuum Hamiltonian remains diffeomorphism covariant and the continuum commutator becomes finite and anomaly free. The final commutator takes the form

$G\to 0$5

and the quantization of the classical right-hand side gives exactly the same continuum limit (Varadarajan, 2012). The point is not that the construction is free of ambiguities; the point is that, within this model, one can have both diffeomorphism covariance and a nontrivial anomaly-free commutator.

5. Propagation and the response to ultralocality

A separate question is whether the kernel of the constraints encodes propagation. In this framework, propagation is not defined by repeated action of the Hamiltonian alone. Instead, a physical state is a sum over charge-network states, and it “encodes propagation” if among its summands there are pairs related by the migration of a local perturbation from one vertex to another many links away. Varadarajan formulated this in Smolin’s weak coupling limit of $G\to 0$6 Euclidean gravity and argued that the relevant physical states are built from “Ket Sets” closed under Hamiltonian deformations, electric-diffeomorphism deformations, possible-parent relations, and semianalytic diffeomorphisms (Varadarajan, 2019).

In the original construction, the basic graph move is an $G\to 0$7 deformation: an $G\to 0$8-valent vertex deforms to a child with a new displaced $G\to 0$9-valent vertex. Varadarajan argues that these actions are generically too weak to support robust long-range propagation, because the child typically determines its parent uniquely. A slight but crucial modification changes the singular deformation itself. Instead of dragging all $(E_i^a,\mathcal A_a^i)$0 non-axis edges, the modified action picks only three preferred non-axis edges and deforms those conically, while all remaining edges are pulled exactly along the axis edge. The result is an $(E_i^a,\mathcal A_a^i)$1 move: an $(E_i^a,\mathcal A_a^i)$2-valent parent produces a child in which the original vertex drops to valence $(E_i^a,\mathcal A_a^i)$3 and the displaced child vertex has valence $(E_i^a,\mathcal A_a^i)$4 (Varadarajan, 2019).

This modification changes the ancestry structure of deformed graphs. A perturbation emitted at one vertex can be absorbed at another vertex, and the same child can have two distinct parents. Under appropriate conditions, the resulting propagation “merges, separates and entangles vertices of charge network states.” Electric diffeomorphism constraints, originally introduced for anomaly freedom, play a key role in this mechanism. The paper presents the result as a counterpoint to Smolin’s early observation that LQG-style Hamiltonians are too ultralocal to generate propagation. At the same time, it is careful about its status: what is rigorously established is a trivial anomaly-free realization on a class of physical states built from Ket Sets, while a full proof that the modified $(E_i^a,\mathcal A_a^i)$5 action also supports the more elaborate nontrivial off-shell anomaly-free commutators remains open (Varadarajan, 2019).

6. Effective dynamics, relation to full $(E_i^a,\mathcal A_a^i)$6, and scope of the term

Later work asked a sharper question: not only whether Smolin’s weak coupling limit admits an internal quantization program, but whether its effective dynamics agrees with full $(E_i^a,\mathcal A_a^i)$7 loop quantum gravity in the limit $(E_i^a,\mathcal A_a^i)$8. In that setting one rescales the connection as

$(E_i^a,\mathcal A_a^i)$9

keeps $A_a^i := G_{\mathrm N}^{-1}\mathcal A_a^i .$0 and $A_a^i := G_{\mathrm N}^{-1}\mathcal A_a^i .$1 finite, and takes $A_a^i := G_{\mathrm N}^{-1}\mathcal A_a^i .$2. The nonabelian Gauss law reduces to

$A_a^i := G_{\mathrm N}^{-1}\mathcal A_a^i .$3

so the gauge group again becomes $A_a^i := G_{\mathrm N}^{-1}\mathcal A_a^i .$4. The vector and scalar constraints reduce to

$A_a^i := G_{\mathrm N}^{-1}\mathcal A_a^i .$5

with Abelian curvature $A_a^i := G_{\mathrm N}^{-1}\mathcal A_a^i .$6 (Long et al., 2021).

The technical advance of this approach is a parametrization of $A_a^i := G_{\mathrm N}^{-1}\mathcal A_a^i .$7 holonomy-flux variables by $A_a^i := G_{\mathrm N}^{-1}\mathcal A_a^i .$8 holonomy-flux variables. On a fixed graph, the paper defines an $A_a^i := G_{\mathrm N}^{-1}\mathcal A_a^i .$9-valued effective holonomy by

$$\mathcal F_{ab}^i = G_{\mathrm N}\Big(\partial_aA_b^i-\partial_bA_a^i+G_{\mathrm N}\epsilon^i{}_{jk}A_a^jA_b^k\Big),$$0

and transported flux variables by

$$\mathcal F_{ab}^i = G_{\mathrm N}\Big(\partial_aA_b^i-\partial_bA_a^i+G_{\mathrm N}\epsilon^i{}_{jk}A_a^jA_b^k\Big),$$1

These do not reproduce the full $$\mathcal F_{ab}^i = G_{\mathrm N}\Big(\partial_aA_b^i-\partial_bA_a^i+G_{\mathrm N}\epsilon^i{}_{jk}A_a^jA_b^k\Big),$$2 holonomy-flux algebra exactly for arbitrary configurations, but in the small-holonomy regime $$\mathcal F_{ab}^i = G_{\mathrm N}\Big(\partial_aA_b^i-\partial_bA_a^i+G_{\mathrm N}\epsilon^i{}_{jk}A_a^jA_b^k\Big),$$3 they reproduce it at leading order. Using this parametrization, the paper constructs a weak-coupling Hamiltonian on the $$\mathcal F_{ab}^i = G_{\mathrm N}\Big(\partial_aA_b^i-\partial_bA_a^i+G_{\mathrm N}\epsilon^i{}_{jk}A_a^jA_b^k\Big),$$4 Hilbert space and shows that the effective dynamics obtained from coherent-state path integrals in $$\mathcal F_{ab}^i = G_{\mathrm N}\Big(\partial_aA_b^i-\partial_bA_a^i+G_{\mathrm N}\epsilon^i{}_{jk}A_a^jA_b^k\Big),$$5 and $$\mathcal F_{ab}^i = G_{\mathrm N}\Big(\partial_aA_b^i-\partial_bA_a^i+G_{\mathrm N}\epsilon^i{}_{jk}A_a^jA_b^k\Big),$$6 loop quantum gravity are consistent in the weak coupling limit, provided that the coherent-state expectation values of the Hamiltonians reproduce their classical expressions up to $$\mathcal F_{ab}^i = G_{\mathrm N}\Big(\partial_aA_b^i-\partial_bA_a^i+G_{\mathrm N}\epsilon^i{}_{jk}A_a^jA_b^k\Big),$$7 corrections (Long et al., 2021).

This semiclassical result narrows the meaning of Smolin’s weak coupling limit. It is not merely a kinematical Abelian truncation, and it is not an exact quantum equivalence to full $$\mathcal F_{ab}^i = G_{\mathrm N}\Big(\partial_aA_b^i-\partial_bA_a^i+G_{\mathrm N}\epsilon^i{}_{jk}A_a^jA_b^k\Big),$$8 gravity. Rather, it is a controlled small-holonomy and semiclassical regime in which an Abelianized $$\mathcal F_{ab}^i = G_{\mathrm N}\Big(\partial_aA_b^i-\partial_bA_a^i+G_{\mathrm N}\epsilon^i{}_{jk}A_a^jA_b^k\Big),$$9 theory can retain the main structural features of full loop quantum gravity. The phrase is also specific: it should not be conflated with unrelated uses of “weak coupling limit” in quantum annealing, lattice gauge theory, swampland conjectures, or many-body kinetic theory [(Bando et al., 2021); (Anishetty et al., 2018); (Buratti et al., 2020); (Chen et al., 2014)].