B-complex manifolds with generalized corners. I. Newlander–Nirenberg Theorems

Hülya Argüz, Dominic Joyce · 2026

Read B-complex manifolds with generalized corners. I. Newlander–Nirenberg Theorems by Hülya Argüz, Dominic Joyce (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-2604-22642.txt.

Introduction

Argüz and Joyce extend complex geometry to spaces with corners and generalized corners, using complex structures on the logarithmic tangent bundle. They prove a Newlander-Nirenberg type result showing the structure matches a standard model along each corner layer to infinite order. The theorem then gives exact conditions for such a space to arise as a Kato-Nakayama space.

Catalog record

CREATORHülya Argüz, Dominic Joyce
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE139768
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.22642

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

B-complex manifolds with generalized corners. I. Newlander–Nirenberg Theorems Hülya Argüz, Dominic Joyce Abstract We generalize complex manifolds to manifolds with corners X X , and to manifolds with generalized corners (g-corners) in the sense of the second author [ 8 ] , using complex structures on the b-tangent bundle (log tangent bundle) T b ​ X {}^{b}TX . We prove a formal Newlander–Nirenberg type theorem showing that along each corner stratum of X X , the b-complex structure agrees with a standard model to infinite order. In the sequel [ 1 ] we show that if S S is a log smooth log ℂ {\mathbin{\mathbb{C}}} -scheme, or log smooth log complex analytic space, then the Kato–Nakayama space S KN S^{\rm KN} has the structure of a b-complex manifold with g-corners. Using our Newlander–Nirenberg theorem we give necessary and sufficient conditions for a b-complex manifold with g-corners to be a Kato–Nakayama space. 1 Introduction 1.1 Overview Manifolds with corners, locally modelled on [0,\infty)^{k}\times{\mathbin{\mathbb{R}}}^{m-k} , provide a natural extension of smooth manifolds that accommodates boundary and corner strata. They are particularly useful in situations where non-compact manifolds admit compactifications by adding boundary faces with controlled geometric structure, allowing one to study asymptotic behaviour within a differential-geometric framework. Manifolds with generalized corners (g-corners), introduced by the second author [ 8 ] , further extend this notion by allowing more flexible local models governed by ‘weakly toric monoids’. The aim of this paper is to extend the theory of complex manifolds to the setting of manifolds with corners and, more generally, manifolds with g-corners. Classically, complex manifolds may be defined either via holomorphic coordinate charts or via integrable almost complex structures, with the equivalence of these approaches ensured by the Newlander–Nirenberg theorem. In this work, we develop a theory of b-complex manifolds with g-corners following the latter perspective, based on b-almost complex structures and their integrability. Our main result establishes a formal analogue of the Newlander–Nirenberg theorem in this context. A sequel [ 1 ] applies these ideas to log geometry: we show that if S is a log smooth log {\mathbin{\mathbb{C}}} -scheme or log complex analytic space then the ‘Kato–Nakayama space’ S^{\rm KN} is a b-complex manifold with g-corners, and characterizes the b-complex manifolds which arise this way. 1.2 Background and context We recall some well known material, see e.g. Kobayashi and Nomizu [ 10 , §IX] . Definition 1.1 . Let X be a smooth manifold of real dimension 2n , and TX\rightarrow X the tangent bundle of X . An almost complex structure J is a morphism of vector bundles J:TX\rightarrow TX such that J^{2}=-{\mathop{\rm id}\nolimits}_{TX} . We call (X,J) an almost complex manifold . A smooth function f:X\rightarrow{\mathbin{\mathbb{C}}} is called holomorphic if J({\rm d}f)=i\,{\rm d}f in \Gamma^{\infty}(T^{*}X\otimes_{\mathbin{\mathbb{R}}}{\mathbin{\mathbb{C}}}), where \Gamma^{\infty}(E) denotes the vector space of smooth sections of E . An almost complex structure has a Nijenhuis tensor N_{J}\in\Gamma^{\infty}(T^{*}X\otimes T^{*}X\otimes TX) characterized by the fact that for all vector fields v,w\in\Gamma^{\infty}(TX) , we have N_{J}(v,w)=[v,w]+J\bigl([Jv,w]+[v,Jw]\bigr)-[Jv,Jw], (1.1) where [\,,\,] is the Lie bracket of vector fields. We call J integrable if N_{J}=0 , and then we call (X,J) a complex manifold . Equivalently, the complexification TX\otimes_{\mathbin{\mathbb{R}}}{\mathbin{\mathbb{C}}} has a splitting TX\otimes_{\mathbin{\mathbb{R}}}{\mathbin{\mathbb{C}}}=T^{1,0}X\oplus T^{0,1}X , where T^{1,0}X is the eigenspace of J with eigenvalue i , and T^{0,1}X the eigenspace of J with eigenvalue -i . Then J is integrable if and only if \bigl[\Gamma^{\infty}(T^{1,0}X),\Gamma^{\infty}(T^{1,0}X)\bigr]\subseteq\Gamma^{\infty}(T^{1,0}X)\subseteq\Gamma^{\infty}(TX\otimes_{\mathbin{\mathbb{R}}}{\mathbin{\mathbb{C}}}) . Here is the Newlander–Nirenberg Theorem [ 19 ] , a central result. Theorem 1.2 . Let (X,J) be an almost complex manifold of real dimension 2n . Then J is integrable if and only if near each x\in X there exist local complex coordinates (z_{1},\ldots,z_{n}) defined on an open neighbourhood U of x in X, such that z_{j}:U\rightarrow{\mathbin{\mathbb{C}}} is holomorphic for j=1,\ldots,n . In these coordinates, J is locally standard. Writing z_{j}=x_{j}+iy_{j} , we have TX|_{U}=\langle\frac{\partial}{\partial x_{j}},\frac{\partial}{\partial y_{j}}:1\leqslant\penalty j\leqslant\penalty n\rangle_{\mathbin{\mathbb{R}}}, with J\Bigl(\frac{\partial}{\partial x_{j}}\Bigr)=\frac{\partial}{\partial y_{j}},\qquad J\Bigl(\frac{\partial}{\partial y_{j}}\Bigr)=-\frac{\partial}{\partial x_{j}},\qquad j=1,\ldots,n. This paper will generalize these ideas from ordinary manifolds X , locally modelled on {\mathbin{\mathbb{R}}}^{m} , to manifolds with corners , locally modelled on [0,\infty)^{k}\times{\mathbin{\mathbb{R}}}^{m-k} , and to manifolds with generalized corners (or manifolds with g-corners ) from the second author [ 8 ] , which have a larger set of local models X_{P} depending on a ‘weakly toric monoid’ P , described in § 2.2 . As in § 2.1 , a manifold with corners X has two notions of tangent bundle, the ordinary tangent bundle TX , and the b-tangent bundle {}^{b}TX , introduced by Melrose [ 14 , 15 ] . If (x_{1},\ldots,x_{m})\in[0,\infty)^{k}\times{\mathbin{\mathbb{R}}}^{m-k} are local coordinates on open U\subset X , so that x_{1},\ldots,x_{k} take values in [0,\infty) , then TX|_{U}\!\cong\!\Bigl\langle\frac{\partial}{\partial x_{1}},\ldots,\frac{\partial}{\partial x_{m}}\Bigr\rangle_{\mathbin{\mathbb{R}}},\;\>{}^{b}TX|_{U}\!\cong\!\Bigl\langle x_{1}\frac{\partial}{\partial x_{1}},\ldots,x_{k}\frac{\partial}{\partial x_{k}},\frac{\partial}{\partial x_{k+1}},\ldots,\frac{\partial}{\partial x_{m}}\Bigr\rangle_{\mathbin{\mathbb{R}}}. Manifolds with g-corners X should probably be considered singular from the point of view of ordinary differential geometry — for example, the ordinary tangent bundle TX may not exist as a vector bundle — but they have well behaved b-tangent bundles {}^{b}TX . Here is part of Definition 3.1 below. Definition 1.3 . Let X be a manifold with corners or g-corners, of real dimension 2n . Then X has a depth stratification X=\coprod_{k=0}^{2n}S^{k}(X) , where S^{k}(X) is a manifold without boundary of dimension 2n-k , and is roughly the points in X which lie in the intersection of k boundary hypersurfaces. There is a natural exact sequence of vector bundles on S^{k}(X) \textstyle{0\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{{}^{b}N_{S^{k}(X)}\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{{}^{b}i} \textstyle{{}^{b}TX|_{S^{k}(X)}\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{{}^{b}\pi} \textstyle{T(S^{k}(X))\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{0,} where {}^{b}N_{S^{k}(X)} is the b-normal bundle of S^{k}(X) in X , a rank k vector bundle which has a natural flat connection \nabla^{{}^{b}N_{S^{k}(X)}} . A b-almost complex structure J on X is a morphism of vector bundles J:{}^{b}TX\rightarrow{}^{b}TX such that J^{2}=-{\mathop{\rm id}\nolimits}_{TX} , and for all 0\leqslant\penalty k\leqslant\penalty 2n we have {}^{b}i\bigl({}^{b}N_{S^{k}(X)}\bigr)\cap J\bigl[{}^{b}i\bigl({}^{b}N_{S^{k}(X)}\bigr)\bigr]=0 in subbundles of {}^{b}TX|_{S^{k}(X)} . This is only possible if S^{k}(X)=\emptyset when n<k\leqslant\penalty 2n . It holds automatically if k=0 or 1. As in Definition 1.1 , a b-almost complex structure has a Nijenhuis tensor N_{J}\in\Gamma^{\infty}({}^{b}T^{*}X\otimes{}^{b}T^{*}X\otimes{}^{b}TX) such that ( 1.1 ) holds for all b-vector fields v,w on X . We call J integrable if N_{J}=0 , and then we call (X,J) a b-complex manifold with (g-)corners . 1.3 Main results This paper will prove Theorems 3.6 and 3.7 below, to which readers are referred for detailed statements. Theorem 3.6 shows that if (X,J) is b-complex manifold with (g-)corners then on each stratum S^{k}(X) we can find local coordinates (\theta_{1},\ldots,\theta_{k},z_{k+1},\ldots,z_{n}) in {\mathbin{\mathbb{R}}}^{k}\times{\mathbin{\mathbb{C}}}^{n-k} which put J|_{S^{k}(X)}:{}^{b}TX|_{S^{k}(X)}\rightarrow{}^{b}TX|_{S^{k}(X)} , and the natural Lie algebroid structure on {}^{b}TX|_{S^{k}(X)} , into a standard form; essentially, z_{k+1},\ldots,z_{n} are holomorphic functions on S^{k}(X) . Theorem 3.7 is a formal analogue of Theorem 1.2 for b-complex manifolds with (g-)corners. For simplicity, we state here a result equivalent to Theorem 3.7 for manifolds with ordinary corners rather than g-corners: Theorem 1.4 . Let (X,J) be a b-complex manifold with corners, of real dimension 2n, and let x\in S^{k}(X) for 0\leqslant\penalty k\leqslant\penalty n . Then we can choose an open neighbourhood U of x in X and global coordinates (\lambda_{1},\ldots,\lambda_{k},\theta_{1},\ldots,\theta_{k},x_{k+1},\allowbreak\ldots,\allowbreak x_{n},\allowbreak y_{k+1},\allowbreak\ldots,y_{n})\in[0,\infty)^{k}\times{\mathbin{\mathbb{R}}}^{2n-k} on U, where \lambda_{1},\ldots,\lambda_{k} take values in [0,\infty), with S^{k}(X) locally defined by \lambda_{1}=\cdots=\lambda_{k}=0, and \lambda_{1}(x)=\cdots=\lambda_{k}(x)=0, such that defining the ‘standard’ b-complex structure J_{\rm st} on U by J_{\rm st}\Bigl(\lambda_{j}\frac{\partial}{\partial\lambda_{j}}\Bigr)\!=\!\frac{\partial}{\partial\theta_{j}},\;\>J_{\rm st}\Bigl(\frac{\partial}{\partial\theta_{j}}\Bigr)\!=\!-\lambda_{j}\frac{\partial}{\partial\lambda_{j}},\;\>J_{\rm st}\Bigl(\frac{\partial}{\partial x_{j}}\Bigr)\!=\!\frac{\partial}{\partial y_{j}},\;\>J_{\rm st}\Bigl(\frac{\partial}{\partial

Read the complete text