Special alternating links of minimal unlinking number
Duncan McCoy, JungHwan Park · 2026
Read Special alternating links of minimal unlinking number by Duncan McCoy, JungHwan Park (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-10732.txt.
Introduction
Duncan McCoy and JungHwan Park study special alternating links whose unlinking number meets the lower bound set by the classical signature. They prove a rigidity result about sharpness. When the bound is sharp for such a link, its unlinking number can be attained by crossing changes in any alternating diagram, not just a chosen one. As an application, they compute new unknotting numbers for some knots of this kind, with eleven and twelve crossings.
Catalog record
| CREATOR | Duncan McCoy, JungHwan Park |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 26974 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.10732
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Special alternating links of minimal unlinking number Duncan McCoy, JungHwan Park Abstract . For any link in the 3-sphere, there is a natural lower bound for the unlinking number in terms of the classical signature. We prove that if this lower bound is sharp for a special alternating link L L , then the unlinking number of L L is necessarily realized by crossing changes in any alternating diagram for L L . As an application, we compute new values of the unknotting numbers for some special alternating knots with crossing number 11 and 12. 1. Introduction Given a link L in the 3-sphere S^{3} , its unlinking number u(L) is defined to be the minimal number of crossing changes required in any diagram of L to obtain an unlink. Despite the simple definition, this invariant is often rather challenging to compute since one typically does not know which diagrams will realize the unlinking number. Indeed results which predict exactly where the unknotting or unlinking numbers will be realized are rare. For example, among the few general results available, it is known that alternating knots with unknotting number one can always be unknotted by a single crossing change in any alternating diagram 1 1 1 Recall that a reduced alternating diagram is minimal , meaning that it realizes the crossing number. [ 19 ] and 2-bridge links with unlinking number one can always be unlinked with a single crossing change in any alternating diagram [ 15 ] (see also [ 13 , 11 , 12 ] ). However, higher unknotting numbers can not necessarily be computed using minimal diagrams, even for alternating knots [ 3 , 2 , 4 ] . The recent striking examples of Brittenham and Hermiller suggest that the unknotting number is far more subtle than was previously imagined [ 5 , 6 ] . For every oriented non-split k -component alternating link L there is a lower bound for the unlinking number of the form (1.1) u(L)\geq\frac{|\sigma(L)|+k-1}{2}, where \sigma(L) denotes the classical signature. 2 2 2 Links in this paper are always assumed to be oriented. The main result of this paper is to show that if L is a special alternating link for which this lower bound is sharp, then the unlinking number is realized by crossing changes in any alternating diagram for L . Recall that a special alternating link is one admitting an alternating diagram in which one of the planar spanning surfaces is a Seifert surface for L . We obtain a similar result for the 4-ball crossing number c_{4}(L) of a special alternating link. By definition, c_{4}(L) is the minimal number of transverse double points among all properly immersed unions of disks in the 4 -ball B^{4} with boundary L . Theorem 1 . Let L be an oriented non-split special alternating link with k components. Then the following are equivalent: (i) u(L)=\frac{|\sigma(L)|+k-1}{2} ; (ii) c_{4}(L)=\frac{|\sigma(L)|+k-1}{2} ; (iii) L can be unlinked by \frac{|\sigma(L)|+k-1}{2} crossing changes in any alternating or minimal diagram. Although every minimal diagram for a prime alternating link is alternating [ 14 , 22 , 28 ] , non-prime alternating links admit non-alternating minimal diagrams. Hence there is no redundancy in the inclusion of both alternating and minimal diagrams in Theorem 1 ( iii ). The most interesting applications of Theorem 1 concern the unknotting number. Corollary 2 . Let K be a special alternating knot. Then u(K)=|\sigma(K)|/2 if and only if K can be unknotted by |\sigma(K)|/2 crossing changes in any alternating or minimal diagram. ∎ For a special alternating knot K , Corollary 2 gives an explicit, implementable procedure to decide whether u(K)=|\sigma(K)|/2 . As described in Section 4 , this allows us to compute u(K) for many small examples in the knot tables, including several knots of crossing number 11 and 12 for which u(K) was previously undetermined. We remark that Theorem 1 and Corollary 2 do not hold for all alternating knots and links. Indeed, there exist alternating knots K with u(K)=|\sigma(K)|/2 such that this value cannot be realized by crossing changes in any alternating diagram [ 3 , 2 ] . We note also the following corollary on the additivity of the unknotting number. An analogous result holds for links, but for simplicity we state it only for knots. Corollary 3 . Let K be a special alternating knot with K=K_{1}\#K_{2} and u(K)=|\sigma(K)|/2 . Then u(K_{1}\#K_{2})=u(K_{1})+u(K_{2}).\qed This corollary is obtained by applying Corollary 2 to a reduced alternating diagram for K , and recalling that such a diagram is necessarily a connected sum of diagrams for K_{1} and K_{2} [ 21 ] . Finally, we note that Theorem 1 can be extended to a statement about split links. Since the unlinking number and the 4 -ball crossing number are additive under split union, we have the following. Corollary 4 . Let L be the split union of oriented non-split special alternating links L_{1},\dots,L_{\ell} and let m=\sum_{i=1}^{\ell}\frac{|\sigma(L_{i})|+k_{i}-1}{2}, where k_{i} denotes the number of components of L_{i} . Then the following are equivalent: (i) u(L)=m ; (ii) c_{4}(L)=m ; (iii) L can be unlinked by m crossing changes in any alternating or minimal diagram. ∎ Structure of proof For any oriented k -component link L , there is a lower bound on c_{4}(L) in terms of classical invariants: (1.2) c_{4}(L)\geq\frac{|\sigma(L)|-\eta(L)+k-1}{2}, where \eta(L) denotes the nullity of L [ 24 , Lemma 2.2] . Since the nullity \eta(L) of a non-split alternating link is always zero, the implications ( iii ) \Rightarrow ( i ) and ( i ) \Rightarrow ( ii ) in Theorem 1 follow immediately from ( 1.2 ) and the fact that u(L)\geq c_{4}(L) . Consequently, the content of Theorem 1 is the implication ( ii ) \Rightarrow ( iii ). This implication is established by using techniques from smooth 4-manifold topology. In particular, for alternating links we derive a lattice-theoretic obstruction to equality in ( 1.2 ) using Donaldson’s diagonalization theorem [ 8 ] (see Theorem 6 for the precise statement). Although similar obstructions have appeared previously in the literature [ 24 , 9 , 26 ] , we obtain a particularly refined obstruction by exploiting the fact that one of the 4 -manifolds arising in the proof is spin (see Proposition 5 ). The proof of Theorem 1 is completed by analyzing this lattice-theoretic obstruction in the case of special alternating links. This analysis is carried out in Section 3 . 2. Obstructing minimal unlinking numbers The aim of this section is to derive the obstruction to an alternating link realizing equality in ( 1.2 ). The following construction is a specialization of that of Nagel–Owens [ 24 , Proposition 2.3] , which was in turn inspired by Cochran–Lickorish [ 7 ] . The key observation is that, in our setup, the construction yields a spin 4 -manifold. Throughout the article, we work in the smooth category. Proposition 5 . Let L be an oriented k -component link in S^{3} with (2.1) c_{4}(L)=\frac{-\sigma(L)-\eta(L)+k-1}{2}. Then the double branched cover \Sigma(L) bounds a smooth, spin, negative-definite 4-manifold W with b_{2}(W)=-\sigma(L) . Moreover H_{2}(W;\mathbb{Z}) contains c_{4}(L) pairwise disjoint spherical classes of self-intersection -2 . Proof. Let p=c_{4}(L) . By [ 24 , Lemma