---
title: Cobordism Conjecture in Quantum Gravity
url: https://www.emergentmind.com/topics/cobordism-conjecture
type: topic
---

# Cobordism Conjecture in Quantum Gravity

Searching arXiv for recent and foundational papers on the Cobordism Conjecture and closely related developments.
arXiv_search(query="Cobordism Conjecture swampland quantum gravity", max_results=10, sort_by="submittedDate")
arXiv_search(query="On the Swampland Cobordism Conjecture and Non-Abelian Duality Groups", max_results=5, sort_by="relevance")
arXiv_search(query="2012.00013", max_results=3, sort_by="relevance")
The Cobordism Conjecture, proposed by McNamara and Vafa within the swampland program, states that the bordism group of quantum gravity is trivial, \( \Omega_n^{QG}\equiv 0 \) for all \(n\) [1909.10355]. In this formulation, one considers compact backgrounds endowed with whatever additional structure quantum gravity requires—spin, gauge bundles, duality bundles, fluxes, or related data—and identifies two such backgrounds when they are connected by a higher-dimensional quantum-gravity configuration with matching boundary structure [2204.00021]. The conjecture is therefore a sharpened version of the no-global-symmetries principle: a nonzero cobordism class would define a conserved topological charge, and consistency then demands physical defects, such as branes, orientifolds, boundaries, or more general singular objects, that trivialize every class [1909.10355]. Subsequent work has used this principle to recover familiar 7-brane spectra in type IIB/F-theory, constrain supergravities with 16 supercharges, formulate dynamical end-of-the-world singularities, and organize the relevant tangential and geometric structures entering quantum-gravity bordism [2012.00013].

## 1. Mathematical formulation and swampland motivation

For a candidate quantum gravity theory in total spacetime dimension \(d\), with \(D=d-k\) noncompact dimensions, the conjecture is phrased in terms of a cobordism group \( \Omega_k^{QG} \) of compact \(k\)-dimensional internal backgrounds. Concretely,
\[
\Omega_k^{QG}
=
\{
M^k \text{ closed, compact quantum-gravity backgrounds}
\}/\sim,
\]
where \(M\sim N\) if there exists a finite-energy domain wall, realized as a noncompact \((k+1)\)-dimensional quantum-gravity background \(W\) with
\[
\partial W = M \sqcup (-N).
\]
The group law is disjoint union, the zero element is the empty background, and inverses come from reversing orientations [1909.10355]. In the special case of smooth manifolds with additional \(G\)-structure and a map to a space \(X\), this reduces to the generalized cobordism groups \( \Omega_k^G(X) \) familiar from algebraic topology [1909.10355].

The physical motivation is that a nonzero class \( [M]\in \Omega_k^{QG} \) behaves as a conserved \((d-k-1)\)-form charge. The construction described in the swampland literature forms a localized \((d-k-1)\)-dimensional defect by taking the connected sum \( \mathbb{R}^k \# M \) in the internal directions; the associated topological current \(J\) obeys
\[
d\,{*J}=0,
\qquad
Q=\int_{S^k} {*J} = [M],
\]
for a linking \(k\)-sphere \(S^k\) [1909.10355]. Black-hole evaporation arguments then lead to the standard alternatives of charge nonconservation or an infinite family of Planck-mass remnants, so the conjecture concludes that consistency requires
\[
\Omega_k^{QG}=0 \qquad \forall\,k
\]
[1909.10355].

A related formulation emphasizes that any two closed \(k\)-dimensional QG-manifolds \((M,D_M)\) and \((N,D_N)\) are QG-bordant if there is a compact \((k+1)\)-manifold \(W\) carrying an extension of the required QG-structure and satisfying \( \partial W \simeq M\sqcup N \) as QG-manifolds [2204.00021]. This makes explicit that the conjecture is not only a statement about ordinary topology, but about topology together with the full physical structure of the theory.

## 2. Defect trivialization and the role of explicit quantum-gravity objects

In the conjectural picture, triviality of \( \Omega_n^{QG} \) is implemented dynamically. For a family of \(n\)-manifolds \( \{M_\alpha\} \) generating a bordism class \( [M_\alpha]\in \Omega_n^{QG} \), the statement \( \Omega_n^{QG}=0 \) means there exist \((n+1)\)-manifolds \(W\) with
\[
\partial W = \bigsqcup_\alpha M_\alpha,
\]
realized physically as worldvolumes or boundaries of defects. In practice one writes an exact sequence
\[
0 \to \dots \to H_{n+1}(B\,G;\Omega_*(pt)) \to \Omega_n^{QG}(B\,G) \to H_n(B\,G;\Omega_*(pt)) \to \dots,
\]
and shows that all classes in \(H_n\) are cancelled by explicit brane insertions [2012.00013].

Several canonical examples recur across the literature. For \(p\)-form gauge fields, an approximate group \( \Omega_{p+1}^{SO,U(1)_p} \) is detected by
\[
Q(M,F)=\int_M F_{p+1},
\]
and killing this charge requires a magnetic monopole brane, identified with a D\((d-p-3)\)-brane [1909.10355]. For \( \Omega_4^{Spin}=\mathbb Z \), generated by K3, the conjecture requires orientifold-like objects: in M-theory the map \( \Omega_4^{Spin}=\mathbb Z\to\Omega_4^{Pin^+}=\mathbb Z_{16} \) has kernel \(2\mathbb Z\), and nonorientable \(Pin^+\) manifolds such as the MO5-plane break the corresponding charge; further defects such as the MO1 kill the remaining torsion [1909.10355]. For \( \Omega_1^{Spin}=\mathbb Z_2 \) in type IIB on \(S^1\) with periodic spin structure, the class is killed by an O7-plane background, described as the half-K3 F-theory compactification on the hemisphere or Sen’s orientifold limit [1909.10355].

| Approximate group | Representative | Known killer |
|---|---|---|
| \( \Omega_4^{Spin}=\mathbb Z \) | K3 | MO5, MO1 |
| \( \Omega_1^{Spin}=\mathbb Z_2 \) | \(S^1_p\) | O7-plane background |
| \( \Omega_0^{Spin}=\mathbb Z \) | \(pt\) | MO9, O8 |

The same logic extends beyond supersymmetric defects. The 2019 swampland analysis argues that known supersymmetric branes, NS5-branes, orientifold planes, Horava–Witten walls, and smooth Calabi–Yau blow-up/down processes kill almost all candidate classes, but residual gaps predict genuinely non-supersymmetric objects, such as a domain wall between IIA and IIB in 10d or a junction of 24 non-BPS \((p,q)\) 7-branes killing the last \(Spin^c\) 4-generator [1909.10355]. These defects are topologically stable through a finite cyclic selection rule, \( \text{total defect number} \equiv 0 \mod n \), rather than through an accompanying low-energy gauge field [1909.10355].

A further refinement arises in AdS/CFT. There, domain walls implement cobordisms between asymptotic boundary data, and topological obstructions to direct conformal interfaces can be bypassed by localizing anomaly-absorbing matter on the wall. In this sense, any two consistent AdS duals can be connected by a bulk domain wall carrying sufficient worldvolume fields to absorb the mismatch, and a single AdS background admits an end-of-the-world brane by the same mechanism [2006.13953].

## 3. Type IIB, duality bundles, and non-Abelian 7-brane physics

The type IIB/F-theory realization is the most detailed arena in which the conjecture has been analyzed. In type IIB compactified on a circle, a duality twist \( \gamma\in SL(2,\mathbb Z) \) defines a class in \( \Omega_9^{QG}(B\,SL(2,\mathbb Z)) \), and the conjecture predicts that such a twist must be trivialized by codimension-two defects, namely 7-branes. Concretely,
\[
\Omega_1^{Spin}(B\,SL(2,\mathbb Z)) \simeq \mathbb Z_{12}\oplus\mathbb Z_2,
\]
and the \( \mathbb Z_{12} \) factor is generated by a D7-brane monodromy, so vanishing of the corresponding QG bordism forces the inclusion of at least one D7-brane for each unit of \(T\)-twist [2012.00013].

The non-Abelian structure of the IIB duality group requires more than the D7-brane sector alone. The axio-dilaton
\[
\tau = C_0 + i\,e^{-\phi}
\]
takes values in the upper half-plane \( \mathbb H \), with \( SL(2,\mathbb Z) \) acting by fractional linear transformations, and the physical moduli space is the orbifold
\[
\mathcal M = \mathbb H/SL(2,\mathbb Z).
\]
Since \( \pi_1(\mathbb H)=0 \), one has
\[
\pi_1(\mathbb H/SL(2,\mathbb Z)) \simeq SL(2,\mathbb Z),
\]
so closed paths in \( \mathcal M \) encode the full duality-twist data [2012.00013]. The general \([p,q]\) 7-brane carries monodromy
\[
M_{p,q}=
\begin{pmatrix}
1+pq & p^2 \\
-q^2 & 1-pq
\end{pmatrix}
\in SL(2,\mathbb Z),
\]
and this monodromy agrees with the transition function of the corresponding loop in \( \mathcal M \) [2012.00013].

This geometric description also reproduces non-Abelian braid statistics. For two non-local 7-branes \(A\) and \(B\) with monodromies \(M_A\) and \(M_B\), dragging \(A\) around \(B\) conjugates \(M_A\) by \(M_B\), and the resulting braid relation is
\[
A\,B\,A = B\,A\,B.
\]
Thus bordism triviality supplies the need for 7-branes, while the orbifold fundamental group of axio-dilaton moduli space captures the full noncommutative exchange structure of \([p,q]\) 7-branes [2012.00013].

Later work sharpened this picture by incorporating fermions and orientation reversal. In type IIB with fermions, the duality group lifts from \( SL(2,\mathbb Z) \) to its metaplectic double cover \( Mp(2,\mathbb Z) \), and allowing worldsheet orientation reversal promotes it to the \(Pin^+\) cover of \( GL(2,\mathbb Z) \). In that setting, many non-trivial bordism classes with \(Mp(2,\mathbb Z)\) duality bundles are identified with asymptotic boundaries of known supersymmetric F-theory backgrounds, including \([p,q]\)-7-branes, non-Higgsable clusters, and S-folds. Extending to the \(Pin^+\) cover requires an additional non-supersymmetric “reflection 7-brane,” predicted precisely to kill the extra bordism classes appearing in the enlarged duality structure [2302.00007].

## 4. Moduli-space constraints, anomalies, and restrictions on supergravity

Combining the Cobordism Conjecture with the Ooguri–Vafa expectation that a quantum-gravity moduli space becomes simply connected once all physical defects are included yields sharp constraints on F-theory vacua. For a congruence subgroup \( \Gamma\subset SL(2,\mathbb Z) \), one compactifies
\[
\mathcal M_\Gamma = \mathbb H/\Gamma
\]
by adjoining cusps and orbifold points to obtain the modular curve \(X(\Gamma)\). The condition \( \pi_1(X(\Gamma))=0 \) forces \(X(\Gamma)\) to have genus zero. In 8D F-theory this implies that \( \Gamma \) must be one of the genus-zero congruence subgroups, and the corresponding Mordell–Weil torsion groups are exactly the known ones: \( \Gamma_1(k)\), \(k=2,3,\dots,10,12\), giving \( \mathbb Z_k\); \( \Gamma(k)\), \(k=2,3,4,5\), giving \( \mathbb Z_k\times\mathbb Z_k\); and intersections such as \( \Gamma(2)\cap\Gamma_1(4) \), giving \( \mathbb Z_2\times\mathbb Z_4 \), \( \mathbb Z_2\times\mathbb Z_6 \), and related cases [2012.00013].

A second line of development uses cobordism together with anomaly cancellation to constrain supergravity theories with 16 supercharges in \(d>6\). In the relevant \(Pin^-\) structure, one has
\[
\Omega_2^{Pin^-}=\mathbb Z_8,
\]
generated by \( \mathbb RP^2 \). Triviality requires a singular 3-manifold \(Y_3\) with \( \partial Y_3=\mathbb RP^2 \), realized in string examples as the local orbifold \( \mathbb R^3/\mathbb Z_2 \) or an orientifold quotient; the associated codimension-three defect is called an I5-fold [2008.11729]. Compactification on \(T^3/\mathbb Z_2\) with eight fixed points then turns bordism triviality into arithmetic restrictions. In nine dimensions one obtains
\[
r \equiv 1 \pmod 8,
\]
while in eight dimensions one obtains
\[
r \equiv 2 \pmod 8.
\]
Combined with the distance-conjecture bounds \(r\le 17\) in nine dimensions and \(r\le 18\) in eight dimensions, these congruences give \(r=1,9,17\) and \(r=2,10,18\), respectively, exactly the known string-theoretic constructions [2008.11729].

The same anomaly analysis constrains the global structure of enhanced gauge groups. For non-abelian factors \(G\) whose conjugation is in the Weyl group, anomaly cancellation forces
\[
\dim G + \operatorname{rank} G \equiv 0 \pmod 8.
\]
This excludes, for example, \(F_4\), while \(E_8\), \(E_7\), and appropriate \(SO(2k+1)\) and \(Sp(k)\) factors remain compatible [2008.11729]. The broader significance is that cobordism triviality is not merely a topological existence statement for defects; when combined with anomaly inflow, it constrains the rank and global form of admissible low-energy gauge sectors.

## 5. Dynamical cobordism, singularities, and end-of-the-world branes

A major refinement replaces the static statement \( \Omega_k^{QG}=0 \) by a dynamical criterion for singular solutions. In the generalized Dudas–Mourad and Blumenhagen–Font models, finite-size “rolling” solutions in a \(D\)-dimensional EFT develop a singularity at finite proper distance, and consistency with quantum gravity requires a local on-shell end-of-the-world brane whose near-core behavior matches that singularity. In both models one finds the universal lower bound
\[
\delta_{cr}=2\sqrt{\frac{D-1}{D-2}}
\]
for the critical exponent \(\delta\) controlling the scaling of proper distance and curvature near the wall [2303.03423]. In the codimension-one sector, neutral and charged ETW defects realize the two characteristic values \( \delta=\sqrt{2}\,b_{cr} \) and \( \delta=\sqrt{2}\,|b| \), and BPS orientifold planes appear as special charged ETW branes; in particular, the \(D=10\) case recovers the O8-plane [2303.03423].

The Sharpened Dynamical Cobordism Conjecture formulates this in terms of a structure-dependent allowed range \(R^\xi\) for \(\delta\). For a near-singularity solution
\[
ds_d^2=e^{-2\sigma(z)}\,ds_{d-1}^2+dz^2,
\qquad
\Delta\sim z\sim e^{-\frac{\delta}{2}D},
\qquad
R\sim e^{\delta D},
\]
a singularity with \( \delta\in R^\xi \) is interpreted as a genuine transition-to-nothing, whereas \( \delta\notin R^\xi \) indicates an obstruction, i.e. a non-trivial cobordism global charge, and the EFT must be enlarged by new higher-form fields or explicit brane defects [2605.06793]. In a pure Einstein-dilaton theory this gives
\[
R^\xi=[0,\delta_0],
\qquad
\delta_0=2\sqrt{\frac{d-1}{d-2}}.
\]
The resulting diagnostics are nontrivial: in massive IIA without O8-planes the Romans-mass solution has \( \delta=5/\sqrt2 \), which lies outside the pure Einstein-dilaton range, but after adding the 9-form potential under which the O8 is charged, the allowed range collapses to the single point \( \delta=5/\sqrt2 \), so the O8-plane precisely restores cobordism triviality [2605.06793]. By the same criterion, the Janis–Newman–Winicour naked singularity is “bad,” the extremal Garfinkle–Horowitz–Strominger solution is “good” in Einstein-dilaton-Maxwell theory, and only the problematic \( \sigma_5 \) D3-brane distribution fails both Gubser and sharpened dynamical cobordism [2605.06793].

A geometrically distinct approach studies cobordism through Ricci-flow surgery. If \( \phi:S^p\times D^q\to M \) is an embedding, surgery produces
\[
M'=(M\setminus \operatorname{Int}\phi(S^p\times D^q))\sqcup_\phi(D^{p+1}\times S^{q-1}),
\]
and the trace-cobordism
\[
W=(M\times I)\cup_{\phi\times\{1\}}(D^{p+1}\times D^q)
\]
satisfies \( \partial W=M\sqcup M' \), so surgeries preserve cobordism classes [2209.10297]. In oriented cobordism one has \( \Omega_4^{SO}\cong\mathbb Z \), generated by \( [\mathbb CP^2] \), while \( \sigma(K3)=-16 \), so
\[
[K3]=-16[\mathbb CP^2].
\]
Accordingly,
\[
\hat M = K3 \# 16\,\mathbb CP^2
\]
has trivial class in \( \Omega_4^{SO} \), providing a concrete model in which defect insertion preserves trivial cobordism after surgery [2209.10297].

## 6. Structural refinements, dualities, and neighboring notions

One structural proposal is that the Whitehead tower organizes the tangential data entering the conjecture. For \(X=BO\), the first stages are
\[
BO \leftarrow BSO \leftarrow BSpin \leftarrow BString \leftarrow BFivebrane,
\]
with obstruction classes \(w_1\), \(w_2\), \( \tfrac12 p_1 \), and \(p_2\), respectively [2204.00021]. The same work argues that geometric structures such as higher \(U(1)\)-bundles with connection should be included directly in bordism, using stacks like \( B^nU(1)_{\mathrm{conn}} \), and proves
\[
\Omega_k^\tau(B^nU(1)) \cong \Omega_k^\tau(B^nU(1)_{\mathrm{conn}}).
\]
Allowing magnetic defects modifies the classification by quotienting flux labels,
\[
\Omega_k^{SO}(B^nU(1))/(n_j\cdot\mathbb Z),
\]
and this in turn suggests the necessity of Kaluza–Klein monopoles from the requirement that \( \Omega_2^{SO}(BU(1))=\mathbb Z \) be killed [2204.00021]. The same framework emphasizes that T-duality exchanges NS5-branes, KK monopoles, and Q-branes while preserving the cobordism criterion [2204.00021].

A complementary refinement identifies cobordism charges with open-string charge groups. Using the Atiyah–Bott–Shapiro and Todd orientations, the open-closed correspondence described in the literature gives
\[
KO^{-n}(X)\cong \Omega_n^{Spin}(X),
\qquad
K^{-n}(X)\cong \Omega_n^{Spin^c}(X),
\]
interpreted as a physical manifestation of a generalized Conner–Floyd isomorphism [2112.07678]. In this picture, D-brane K-theory charges and closed-string cobordism charges are the same RR gauge charge, and gauging the diagonal combination recovers type I and F-theory tadpole cancellation conditions, including the familiar 24 seven-branes on \( \mathbb P^1 \) in F-theory on K3 [2112.07678].

More recent work shows that bordism trivialization for discrete symmetry groups can require networks rather than isolated defects. For a discrete group \(G\), the second bordism is controlled by the short exact sequence
\[
0 \to H_1(BG;\Omega_\xi^1(pt)) \to \widetilde\Omega_\xi^2(BG) \to H_2(BG;\mathbb Z) \to 0,
\]
and nontrivial classes in \(H_2(BG;\mathbb Z)\) are trivialized not by codimension-three “holes” but by linked and junctioned networks of ordinary codimension-two defects [2605.18952]. In four-dimensional supergravity with a discrete Heisenberg symmetry acting on axions, this predicts explicit D4-string and fundamental-string junctions whose valence is fixed by the commutator structure of \(H_2(BG;\mathbb Z)\) [2605.18952]. A related G-theory-motivated analysis computes
\[
\Omega_6^{Spin}(B[SL(2,\mathbb Z)\times SL(2,\mathbb Z)])
\cong
G_{(2)}\oplus(\mathbb Z_3)^3,
\]
with \(G_{(2)}\) an abelian 2-group of order 16, most likely \( \mathbb Z_4\oplus\mathbb Z_4 \); 24 \((p,q)\)-branes cancel the perturbative \( \mathbb Z_{12} \)-subgroup, while the extra torsion points toward S-folds and other exotic U-duality defects [2605.09868].

A persistent terminological ambiguity should be removed. In higher category theory, the “cobordism hypothesis” is the statement that fully extended framed \(n\)-dimensional TQFTs are classified by fully dualizable objects in a symmetric monoidal \((\infty,n)\)-category, with
\[
Fun^\otimes(Bord_n^{fr},C)\simeq Obj^{fd}(C)
\]
[1705.02240]. That theorem concerns the bordism category of manifolds and higher morphisms in TQFT. The swampland Cobordism Conjecture instead concerns the vanishing of quantum-gravity bordism groups and the necessity of defects that trivialize them. The shared terminology reflects common topological input, but the two statements address different problems.

Source: https://www.emergentmind.com/topics/cobordism-conjecture