Generalized Property R Conjecture: handleslide equivalence of R-links to the unlink and the status of the generalized-square-knot families $L^{p,q}_{c/d}$

The Generalized Property R Conjecture (GPRC), recorded as Problem 1.82 in the standard low-dimensional topology problem list, predicts that every R-link — an n-component link in $S^3$ on which 0-framed surgery produces $\#^n(S^1\times S^2)$ — is handleslide-equivalent to the unlink, extending Gabai's 1987 Property R theorem from knots to links. The conjecture implies the Smooth 4-Dimensional Poincaré Conjecture (S4PC) for geometrically simply-connected homotopy 4-spheres, and its Weak variant is equivalent to that statement. Meier and Zupan's study of generalized square knots $Q_{p,q}=T_{p,q}\#T_{-p,q}$ proved that 2R-links containing such knots satisfy the Weak variant and produced, for every even number of components, explicit families $L^{p,q}_{c/d}$ of R-links that are potential counterexamples — the first known with more than two components. As of August 2026 the core problem is open: the GPRC and its Stable variant remain open in general; the status of the $L^{p,q}_{c/d}$ families is experimentally settled only for one subfamily in small parameter ranges ($L(3,2;4/d)$ stably handleslide-trivial for $d\le 39$); and neither a proof nor a genuine counterexample is known.

Background

The Smooth 4-Dimensional Poincaré Conjecture (S4PC) asserts that every homotopy 4-sphere — a closed smooth 4-manifold homotopy-equivalent to the standard 4-sphere $S^4$ — is diffeomorphic to $S^4$. The topological-category analogue is known, so the difficulty is purely smooth, and S4PC is the final unsettled case of the generalized Poincaré conjecture. The problem considered here is a 3-dimensional reformulation of part of S4PC together with its explicit test cases.

By the Kirby calculus, every smooth 4-manifold built without 1-handles admits a surgery description: attach 2-handles along the components of a framed link in $S^3$ (0-framed means each component carries its Seifert-surface framing) and cap off with 3- and 4-handles. An n-component link $L\subset S^3$ is called an R-link if 0-framed Dehn surgery on $L$ produces $\#^n(S^1\times S^2)$, the connected sum of n copies of $S^1\times S^2$ — the manifold obtained by the 'obvious' surgery on the n-component unlink. Gabai's celebrated Property R theorem (1987) shows the obvious description is the only one for knots: a knot in $S^3$ on which some Dehn surgery yields $S^1\times S^2$ is the unknot. For links the question is subtler because handle slides — sliding one component over another along a band, a local move that changes the link while preserving the surgery output — generate many nontrivial surgery descriptions of the same manifold.

The Generalized Property R Conjecture (GPRC, recorded as Problem 1.82 in the standard low-dimensional topology problem list, as stated for instance by Lidman–Oliveira-Smith–Zupan 2026) asserts that every R-link is handleslide-equivalent to an unlink. Two weaker variants were introduced by Gompf–Scharlemann and Gompf–Scharlemann–Thompson: the Stable GPRC allows first taking the split union of L with a 0-framed unlink U (geometrically, adding canceling 2-handle/3-handle pairs), and the Weak GPRC additionally allows a split collection of Hopf pairs (a 0-framed Hopf link whose dotted component encodes a canceling 1-handle/2-handle pair). The GPRC implies, and by Cerf theory the Weak GPRC is equivalent to, the S4PC for geometrically simply-connected 4-manifolds — those admitting handle decompositions without 1-handles. The GPRC also bears on the topological Schoenflies conjecture (Scharlemann 2008).

Gompf–Scharlemann–Thompson (2010) constructed 2-component R-links $L_n$ containing the square knot as probable counterexamples to the GPRC: if $L_n$ had Property R, the balanced presentation $P_n=\langle x,y\,|\,xyx=yxy,\,x^n=y^{n+1}\rangle$ of the trivial group would be Andrews–Curtis trivial (reducible by the moves of the Andrews–Curtis conjecture, widely believed unable to trivialize these presentations for $n\ge 3$), and these links generate slice knots — knots bounding smoothly embedded disks in the 4-ball — that are not known to be ribbon, that is, not known to bound ribbon disks (disks in the 4-ball whose radial projection to the 3-sphere has only ribbon-shaped singularities), bearing on the Slice-Ribbon Conjecture, which predicts that every slice knot is ribbon. Scharlemann (2016) located the $L_n$ within the slope-indexed classification of 2R-links containing the square knot, and Meier–Schirmer–Zupan (2016) proved the Stable GPRC for c-component links of tunnel number at most c (the tunnel number of a knot being the minimum number of simultaneously embedded arcs in its complement that turn the complement into a handlebody).

Meier and Zupan (2019; published 2022) extended this program. A generalized square knot is $Q_{p,q}=T_{p,q}\#T_{-p,q}$, the connected sum of a $(p,q)$-torus knot with its mirror ($Q_{3,2}$ is the classical square knot); such knots are ribbon and fibered (their complements fiber over $S^1$ with fiber a Seifert surface, the fibration's return map being the monodromy). Analyzing Casson–Gordon derivative links — companion links lying as collections of curves in the fiber surface of $Q_{p,q}$ — Meier–Zupan classified the possible derivatives and proved that every 2R-link containing a generalized square knot has Weak Property R (for $Q_{p,2}$ a single Hopf pair suffices). They also produced, for each pair $0<q<p$ with $n=(p-1)(q-1)$ and every rational slope $c/d$ with $c$ even, an explicit nR-link $L^{p,q}_{c/d}$ contained in a fiber of $Q_{p,q}$; these give potential GPRC counterexamples for every even number of components, the first known with more than two. Through trisections — decompositions of a smooth 4-manifold into three 4-dimensional 1-handlebodies meeting pairwise in 3-dimensional handlebodies, introduced by Gay and Kirby — Meier–Zupan showed that the Stable GPRC for such a link is equivalent to the natural trisection of the associated homotopy 4-sphere $X_{L^{p,q}_{c/d}}$ being $\{2,3\}$-standard — a trisection of $S^4$ is $\{2,3\}$-standard when its connected sum with finitely many copies of the standard genus-one trisections $S_2$ and $S_3$ of $S^4$ is a standard trisection — and they proved that Stable Property R forces a link to be ribbon while leaving open whether Weak Property R does. Whether any $L^{p,q}_{c/d}$ (or the $L_n$) actually fails to have (Stable) Property R — i.e., whether any is a genuine counterexample — was left open, as was the classification of the fibered 2-knots $\mathcal{K}_{c/d}$ in the Casson–Gordon spheres for slopes $c/d\neq\pm 2n/1$.

Problem Statement

Primary question (the Generalized Property R Conjecture in its standard formulation): Let $L$ be an n-component link in $S^3$ such that 0-framed Dehn surgery on $L$ produces $\#^n(S^1\times S^2)$ (an R-link). Is $L$ necessarily handleslide-equivalent to the n-component 0-framed unlink? If not in full generality, does at least the Stable variant hold — is $L\sqcup U$ handleslide-equivalent to an unlink for some 0-framed unlink $U$?

Derived target from the source paper: determine the handleslide status of the explicit potential-counterexample families. For every generalized square knot $Q_{p,q}$ with $0<q<p$, every rational slope $c/d$ with $c$ even, and $n=(p-1)(q-1)$: does the nR-link $L^{p,q}_{c/d}$ have Property R, or at least Stable Property R? Equivalently, by the source paper's Theorem 9.5, is the natural trisection of the homotopy 4-sphere $X_{L^{p,q}_{c/d}}$ $\{2,3\}$-standard?

The conjecture can be resolved by proving the GPRC (or the Stable GPRC) in general, by refuting it via an explicit R-link shown rigorously not to be handleslide-equivalent to any unlink, or, for the $L^{p,q}_{c/d}$ families, by determining, for all parameters $(p,q;c/d)$, whether $L^{p,q}_{c/d}$ (respectively $L^{p,q}_{c/d}\sqcup U$) is handleslide-equivalent to an unlink.

References

1. Jeffrey Meier and Alexander Zupan, Generalized square knots and homotopy 4-spheres, arXiv:1904.08527, DOI 10.48550/arXiv.1904.08527, https://arxiv.org/abs/1904.08527; published as J. Differential Geom. 122 (2022), no. 1, 69–129, DOI 10.4310/jdg/1668186788. (Source paper.)
2. David Gabai, Foliations and the topology of 3-manifolds III, Journal of Differential Geometry 26 (1987), no. 3, 479–536, DOI 10.4310/jdg/1214441488, https://doi.org/10.4310/jdg/1214441488
3. Robert E. Gompf and Martin Scharlemann, Fibered knots and Property 2R, II, arXiv:0908.2795 (2009), DOI 10.48550/arXiv.0908.2795, https://arxiv.org/abs/0908.2795
4. Robert E. Gompf, Martin Scharlemann, and Abigail Thompson, Fibered knots and potential counterexamples to the Property 2R and Slice-Ribbon Conjectures, Geometry & Topology 14 (2010), no. 4, 2305–2347, DOI 10.2140/gt.2010.14.2305, arXiv:1103.1601, https://arxiv.org/abs/1103.1601
5. Martin Scharlemann, Generalized Property R and the Schoenflies Conjecture, Commentarii Mathematici Helvetici 83 (2008), no. 2, 421–449, DOI 10.4171/cmh/131, arXiv:math/0603511, https://arxiv.org/abs/math/0603511
6. Martin Scharlemann, Proposed Property 2R counterexamples examined, Illinois Journal of Mathematics 60 (2016), no. 1, 207–250, DOI 10.1215/ijm/1498032031, arXiv:1208.1299, https://arxiv.org/abs/1208.1299
7. Jeffrey Meier, Trent Schirmer, and Alexander Zupan, Classification of trisections and the Generalized Property R Conjecture, Proceedings of the American Mathematical Society 144 (2016), no. 11, 4983–4997, DOI 10.1090/proc/13105, arXiv:1507.06561, https://arxiv.org/abs/1507.06561
8. David Gay and Robion Kirby, Trisecting 4-manifolds, Geometry & Topology 20 (2016), no. 6, 3097–3132, DOI 10.2140/gt.2016.20.3097, https://doi.org/10.2140/gt.2016.20.3097
9. Maggie Miller and Alexander Zupan, Equivalent characterizations of handle-ribbon knots, Communications in Analysis and Geometry 31 (2023), no. 9, 2157–2193, DOI 10.4310/cag.2023.v31.n9.a1, arXiv:2005.11243, https://arxiv.org/abs/2005.11243
10. Ethan Romary and Alexander Zupan, A family of Andrews–Curtis trivializations via 4-manifold trisections, Geometriae Dedicata 218 (2024), no. 2, Paper No. 45, DOI 10.1007/s10711-024-00891-6, arXiv:2304.03196, https://arxiv.org/abs/2304.03196
11. Jeffrey Meier and Alexander Zupan, Knots bounding nonisotopic ribbon disks, Journal of Topology 18 (2025), no. 4, Paper No. e70047, DOI 10.1112/topo.70047, arXiv:2310.17564, https://arxiv.org/abs/2310.17564
12. Román Aranda and Alexander Zupan, Manifolds with weakly reducible genus-three trisections are standard, arXiv:2503.04607, DOI 10.48550/arXiv.2503.04607, https://arxiv.org/abs/2503.04607
13. Tye Lidman, Trevor Oliveira-Smith, and Alexander Zupan, A homological generalized Property R conjecture is false, arXiv:2603.05664, DOI 10.48550/arXiv.2603.05664, https://arxiv.org/abs/2603.05664
14. Wenjie Diao, Haoqian Pan, and Chunxing Yan, Some experimental results on stable equivalence of GST Links for the Generalized Property R Conjecture, arXiv:2604.17737, DOI 10.48550/arXiv.2604.17737, https://arxiv.org/abs/2604.17737