Complete non-singular toric varieties with Picard number 4
Suyoung Choi, Hyeontae Jang, Mathieu Vallée · 2025
Read Complete non-singular toric varieties with Picard number 4 by Suyoung Choi, Hyeontae Jang, Mathieu Vallée (2025) 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-2504-18134.txt.
Introduction
Suyoung Choi, Hyeontae Jang, and Mathieu Vallée classify a family of geometric shapes in this 2025 paper. The shapes, called toric varieties, are complete and smooth, with a Picard number of four. That number is a way to count basic building data for the shape. The authors work through a framework based on fan-like spheres and maps.
Their sorting yields 59 fan-like seeds with Picard number four, plus all the manifolds they support. The work settles a past conjecture by giving the first known cases built on neighborly polytopes. It also answers a question from Batyrev about minimal non-faces of such spheres. The results give pure math a full map of this class.
Catalog record
| CREATOR | Suyoung Choi, Hyeontae Jang, Mathieu Vallée |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 75 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-18134
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Complete non-singular toric varieties with Picard number 4 Thanks: This research was supported by Basic Science Research Program through the National Research Foundation of Korea(NRF) (RS-2021-NR060141 and RS-2025-00521982). Suyoung Choi, Hyeontae Jang, Mathieu Vallée Abstract . We classify all complete non-singular toric varieties with Picard number four via a combinatorial framework based on fanlike simplicial spheres and characteristic maps. This classification yields 59 59 fanlike seeds with Picard number four, along with all toric manifolds supported by them. As a consequence, we resolve a conjecture of Gretenkort, Kleinschmidt, and Sturmfels by presenting the first known examples of toric manifolds supported by neighborly polytopes. We also answer a question of Batyrev concerning minimal non-faces of such spheres. 2020 Mathematics Subject Classification 57S12, 14M25 1. Introduction The classification of algebraic varieties is a fundamental problem in algebraic geometry. Among these, toric varieties serve as a bridge between algebraic geometry, combinatorics, and topology; see [ 8 , 16 , 25 , 32 ] for details. Their explicit and computable nature makes them valuable as testbeds for illustrating theoretical phenomena and verifying conjectures. Despite their rich combinatorial structure, the complete classification of complete non-singular toric varieties, simply referred to as toric manifolds , remains an open and challenging problem, especially in higher dimensions. While all complex two-dimensional toric manifolds are well-known to be obtained through successive blow-ups of either the projective plane or Hirzebruch surfaces [ 33 , 32 ] , an explicit classification in higher dimensions has yet to be fully understood. An alternative approach to classification considers the Picard number , which measures the rank of the Picard group. The classification of toric manifolds has been successfully completed up to Picard number 3 . In their pioneering works, Kleinschmidt [ 30 ] classified those with Picard number 2 , and Batyrev [ 6 ] extended the classification to Picard number 3 . The fundamental theorem for toric geometry establishes a bijection between complex n -dimensional toric manifolds with Picard number p and real n dimensional complete non-singular fans with n+p rays. Each of these fans is characterized by two components: • an (n-1) -dimensional piecewise linear (for short, PL) sphere, called its underlying complex , and • a set of n+p primitive ray vectors in \mathbb{Z}^{n} , one for each vertex of the sphere. More precisely, for an n -dimensional complete non-singular fan with n+p rays, each cone corresponds to a subset of the rays, and the collection of such subsets forms a sphere. For clarity, we call a PL sphere fanlike if it arises as the underlying complex of a complete non-singular fan. Therefore, the classification of complete non-singular toric varieties reduces essentially to the following question; Which pairs of fanlike PL spheres and sets of primitive ray vectors produce complete non-singular fans? However, this problem remains highly nontrivial since, even after fixing either the dimension or the Picard number, infinitely many fanlike spheres exist. A complete characterization of fanlike spheres is not yet known, presenting a significant barrier to classification. Numerous studies such as [ 5 , 21 , 17 , 35 , 36 , 7 , 3 , 1 ] have investigated necessary conditions for a PL sphere to be fanlike, but complete characterizations have remained out of reach, leaving many important open questions and conjectures. In particular, Batyrev [ 6 ] suggested an upper bound on the number of minimal non-faces of a fanlike sphere, and Gretenkort, Kleinschmidt, and Sturmfels [ 26 ] conjectured that no neighborly fanlike sphere exist. Remarkably, this paper resolves both issues, including providing a counterexample to the conjecture of Gretenkort, Kleinschmidt, and Sturmfels. The fact that their conjecture stood unresolved over three decades despite being incorrect underscores the inherent difficulty in classifying toric manifolds. A significant breakthrough toward classification was achieved by Choi and Park [ 12 ] . While the wedge operation was already known, Choi and Park were the first to systematically apply this operation to classify toric manifolds. They showed that if all toric manifolds supported by a given PL sphere K are known, then applying a wedge operation at a vertex of K allows one to determine all toric manifolds supported by the resulting complex, which has one dimension higher while preserving the Picard number. This approach provides a highly practical and concrete construction. By iteratively applying this method, the classification of toric manifolds supported by K extends to include all toric manifolds supported by the PL spheres obtained through successive wedge operations on K . These resulting PL spheres form what is known as the J -construction , introduced by Bahri, Bendersky, Cohen, and Gitler [ 2 ] , and are denoted by K(J) , where J is a tuple of positive integers specifying the number and positions of the wedge operations. If a PL sphere cannot be obtained through wedge operations, we call it a seed . Notably, any PL sphere can be expressed in the form K(J) with K being a seed. Thus, the classification of toric manifolds with Picard number p essentially reduces to identifying all fanlike seeds with Picard number p and determining the toric manifolds supported by them. In their subsequent work [ 14 ] , Choi and Park proved that for a fixed Picard number, there exist only finitely many fanlike seeds. For example, the only fanlike seeds with Picard number 3 are the boundaries of a pentagon and a 3 -crosspolytope, while the only seed with Picard number 2 is the boundary of a square. These results provide immediate reproofs of Kleinschmidt and Batyrev’s classification results; see, for instance, [ 12 , Theorem 6.8] . A major advantage of this approach is that it does not rely on any geometric properties of toric manifolds, whereas Batyrev’s method applies only to projective toric manifolds. Since non-projective toric manifolds exist for Picard number at least 4 , Batyrev’s method is not applicable in this case. However, the approach based on Choi and Park’s work remains viable. Indeed, this paper successfully applies it to classify all toric manifolds with Picard number 4 . The authors of this paper previously classified all potential candidates for fanlike seeds with Picard number 4 in [ 9 ] . They identified all seeds with Picard number 4 that support a non-singular characteristic map, a necessary condition for being fanlike, and justified that there are exactly 3153 such seeds, and included them in a publicly accessible database. In this paper, we provide a complete characterization of the fanlike seeds from this list, presented in Theorem 4.2 . Using this characterization, it is established that there exist exactly 59 fanlike seeds; see Section 4 . Furthermore, all toric manifolds supported on these fanlike seeds are identified. As a direct consequence of these results and Choi-Park’s framework, these lists complete the full classification of toric manifolds with Picard number 4 . As highlighted earlier, our results provide explicit counterexamples to previously unresolved conjectures posed by Batyrev and Gretenkort, Kleinschmidt, and Sturmfels, offering negative answers to their questions in the final section of this paper; see Propositions 5.3 and 5.4 . Such explicit classification results not only resolves a long-standing open problem but also highlights the effectiveness of our classification framework in