A simplified proof of the o-minimal Whitney Extension Theorem

Beata Kocel-Cynk, Wiesław Pawłucki and Anna Valette · 2026

Read A simplified proof of the o-minimal Whitney Extension Theorem by Beata Kocel-Cynk, Wiesław Pawłucki and Anna Valette (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-2603-14924.txt.

Introduction

Kocel-Cynk, Pawłucki and Valette simplify the proof of the o-minimal Whitney Extension Theorem. Whitney extension asks when a jet on a closed set extends to a smooth map. The o-minimal version works inside definable geometry. Their proof trims the older arguments to a leaner path. A definable Urysohn lemma for C^q maps is the new hinge. That lemma does the heavy lifting once in place. Extension then follows with less machinery than before. Model theorists and geometers gain a cleaner reference proof.

Catalog record

CREATORBeata Kocel-Cynk, Wiesław Pawłucki and Anna Valette
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE20864
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.14924

Browse the full library · About the Tower of Babel

Full text

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

A simplified proof of the o-minimal Whitney Extension Theorem Beata Kocel-Cynk, Wiesław Pawłucki and Anna Valette Abstract . We give a proof of the o-minimal version of the Whitney Extension Theorem simplified as compared to the original ones. A new simplifying ingredient is a definable variant of Urysohn’s lemma for class 𝒞 q \mathcal{C}^{q} (see Section 3). 1991 Mathematics Subject Classification Primary 14P15. Secondary 03C64, 32B20, 46E35 1. Introduction In this article we give a simplified proof of the following o-minimal version of the classical Whitney Extension Theorem [W], earlier proved in [KP1], [KP2] and in [T] (see also [PR]): Theorem 1.1 . Given an o-minimal extension of the field of real numbers \mathbb{R} , let E be a definable closed subset of \mathbb{R}^{n} and let p and q be positive integers such that p\leq q . Let F(x,X)=\sum_{|\alpha|\leq p}\frac{1}{\alpha!}F^{\alpha}(x)X^{\alpha}\qquad(X=(X_{1},\dots,X_{n})) be a definable \mathcal{C}^{p} -Whitney field on E . ( Definability of F means that all F^{\alpha} are definable functions. ) Then there exists a definable \mathcal{C}^{p} -function f:\mathbb{R}^{n}\longrightarrow\mathbb{R} , of class \mathcal{C}^{q} on \mathbb{R}^{n}\setminus E such that D^{\alpha}f=F^{\alpha} on E whenever \alpha\in\mathbb{N}^{n} , |\alpha|\leq p . As in [KP1], [KP2] and [T], the main tool of the proof are stratifications into \varLambda_{p} -regular cells and the extension procedure when E is a \varLambda_{p} -regular cell. A simplifying ingredient of the present proof is a definable \mathcal{C}^{p} -version of Urysohn lemma, which appeared in our paper [K-CPV2] (see Section 3). Throughout the article, p and q will denote positive integers, p\leq q . 2. \varLambda_{p} -regular maps and stratifications into \varLambda_{p} -regular cells We recall after [KP1] (see also [KP2], [Th] and [PR]), the definition of \varLambda_{p} -regular maps. Definition 2.1 . A map f:\varOmega\longrightarrow\mathbb{R}^{m} defined on an open subset \varOmega of \mathbb{R}^{n} is called \varLambda_{p} -regular if it is of class \mathcal{C}^{p} and \forall\,\alpha\in\mathbb{N}^{n}\,x\in\varOmega:\,1\leq|\alpha|\leq q\,\Longrightarrow\,|D^{\alpha}f(x)|\leq\frac{C}{d(x,\partial\varOmega)^{|\alpha|-1}}, where C is a positive constant. In this definition, we adopt the convention d(x,\emptyset):=1 , to cover the case \partial\varOmega=\emptyset . Remark 2.2 . In particular, all the first order partial derivatives \frac{\partial f}{\partial x_{j}} are bounded. It follows that if A is a closed subset of \varOmega which is quasi-convex (i.e. there exists a positive constant M such that for each pair of points a,b\in A there exists a rectifiable arc \gamma:[0,1]\longrightarrow A such that \gamma(0)=a , \gamma(1)=b and |\gamma|\leq M|a-b| ), then f|A is Lipschitz, hence it admits a continuous extension \overline{f|A}:\overline{A}\longrightarrow\mathbb{R}^{m} to the closure \overline{A} . Definition 2.3 . We say that S is an open \varLambda_{p} -regular cell in \mathbb{R}^{n} if (2.1) S is any open interval in \mathbb{R} , when n=1 ; (2.2) S=\{(x^{\prime},x_{n}):x^{\prime}\in T,\psi_{1}(x^{\prime})<x_{n}<\psi_{2}(x^{\prime})\}, where x^{\prime}=(x_{1},\dots,x_{n-1}) , T is an open \varLambda_{p} -regular cell in \mathbb{R}^{n-1} and every \psi_{i} (i\in\{1,2\}) is either a definable \varLambda_{p} -regular function on T with values in \mathbb{R} , or identically equal to -\infty , or identically equal to +\infty , and \psi_{1}(x^{\prime})<\psi_{2}(x^{\prime}) , for each x^{\prime}\in T , when n>1 . Extending the above definition, we say that S is an m -dimensional \varLambda_{p} -regular cell in \mathbb{R}^{n} , where m\in\{0,\dots,n-1\} , if (2.3) S=\{(u,w):\,u\in T,w=\varphi(u)\}, where u=(x_{1},\dots,x_{m}) , w=(x_{m+1},\dots,x_{n}) , T is an open \varLambda_{p} -regular cell in \mathbb{R}^{m} , and \varphi:T\longrightarrow\mathbb{R}^{n-m} is a definable \varLambda_{p} -regular mapping. Remark 2.4 . It is elementary that if M\geq 0 is a Lipschitz constant of the mapping \varphi , then putting L:=1/\sqrt{1+M^{2}} , we have (2.4) \forall x=(u,w)\in T\times\mathbb{R}^{n-m}:\,L|w-\varphi(u)|\leq d(x,S)\leq|w-\varphi(u)| and (2.5) \forall\,x\in\mathbb{R}^{n}\setminus(T\times\mathbb{R}^{n-m}):\,d(x,S)\geq Ld(x,\partial S). Remark 2.5 . It follows from Remark 2.2 by induction that every \varLambda_{p} -regular cell is quasi-convex and it is Lipschitz in the sense that each of the functions \psi_{i} in (2.2), if finite, as well as the map \varphi in (2.3), are Lipschitz. Besides, every \varLambda_{p} -regular cell in \mathbb{R}^{n} is a definable connected \mathcal{C}^{p} -submanifold of \mathbb{R}^{n} . Definition 2.6 . Let us recall that a ( definable ) \mathcal{C}^{p} - stratification of a (definable) subset E of \mathbb{R}^{n} is a finite decomposition \mathcal{S} of E into (definable) connected \mathcal{C}^{p} -submanifolds of \mathbb{R}^{n} , called strata , such that for each stratum S\in\mathcal{S} , its boundary in E ; i.e. \partial_{E}S:=(\overline{S}\setminus S)\cap E is the union of some strata of dimensions <\dim S . If A_{1},\dots,A_{k} (k\in\mathbb{N}) are subsets of E , we call a stratification \mathcal{S} compatible with the subsets A_{1},\dots,A_{k} , if each A_{j} is a union of some strata. Theorem 2.7 ([P1, Proposition 4 or P2, Theorem 3]) . Given any finite number A_{1},\dots,A_{k} of definable subsets of a definable subset E of \mathbb{R}^{n} , there exists a \mathcal{C}^{p} -stratification \mathcal{S} of E compatible with sets A_{1},\dots,A_{k} and such that each stratum S\in\mathcal{S} , after an orthogonal linear change of coordinates in \mathbb{R}^{n} , is a \varLambda_{p} -regular cell in \mathbb{R}^{n} . Remark 2.8 . By [P2], as orthogonal linear changes of coordinates in the above theorem it suffices to take permutations of Cartesian coordinates. 3. Definable \mathcal{C}^{q} -Urysohn lemma The definable \mathcal{C}^{q} -Urysohn lemma (Theorem 3.2) is a basic new ingredient in our proof of Theorem 1.2. Definition 3.1 . Let Z be a closed definable subset of \mathbb{R}^{n} and let W be a definable closed subset of \mathbb{R}^{n}\setminus Z . We will consider the following open neighborhoods of W in \mathbb{R}^{n} G_{\eta}(W,Z):=\{x\in\mathbb{R}^{n}\setminus Z:\,d(x,W)<\eta d(x,Z)\}, where \eta>0 . Theorem 3.2 ([K-CPV, Proposition 3.9 ]) . Let Z be a definable closed subset of \mathbb{R}^{n} and let W be a definable closed subset of \mathbb{R}^{n}\setminus Z . Then for any positive integer q and for any \eta>0 there exists a definable function \omega:\mathbb{R}^{n}\setminus Z\longrightarrow[0,1] such that (3.1) \forall\,x\in\mathbb{R}^{n}\setminus Z,\,\alpha\in\mathbb{N}^{n}:\quad|\alpha|\leq q\Longrightarrow|D^{\alpha}\omega(x)|\leq\frac{C}{d(x,Z)^{|\alpha|}}, (3.2) \omega\equiv 1 on G_{\rho}(W,Z) , for some \rho\in(0,\eta) , and (3.3) support of \psi in \mathbb{R}^{n}\setminus Z is contained in G_{\eta}(W,Z) . 4. \mathcal{C}^{p} -Whitney fields Let A be a locally closed subset of \mathbb{R}^{n} . By \mathcal{C}(A) we denote the algebra of continuous functions on A with values in \mathbb{R} . A \mathcal{C}^{p} -Whitney field on A is a polynomial F(u,X)=\sum_{|\alpha|\leq p}\frac{1}{\alpha!}F^{\alpha}(u)X^{\alpha}\in\mathcal{C}(A)[X]=\mathcal{C}(A)[X_{1},\dots,X_{n}], which fulfills the following condition ( \ast ) for each c\in A and for each \beta\in\mathbb{N}^{n} such that |\beta|\leq p F^{\beta}(a)-\sum_{|\alpha|\leq p-|\beta|}\frac{1}{\alpha!}F^{\alpha+\beta}(b)(a-b)^{\alpha}=o(|a-b|^{p-|\beta|}),\quad\text{when $A\ni a\rightarrow c$, $A\ni b\rightarrow c$}. We call such a \mathcal{C}^{p} -Whitney field definable , if A is definable and all the functions F^{\alpha} are definable. Remark 4.1 . If F is a \mathcal{C}^{p} -Whitney field on a locally closed subset A\subset\mathbb{R}^{n} , then its restriction F|B to a locally closed subset B\subset A , defined by F|B(u,X)=F(u,X) , for any u\in B , is a \mathcal{C}^{p} -Whitney field on B . Let \pi_{p}:\mathcal{C}(A)[X]\longrightarrow\mathcal{C}(A)[X] denote the natural projection \pi_{p}(\sum_{\alpha}\frac{1}{\alpha!}F^{\alpha}X^{\alpha}):=\sum_{|\alpha|\leq p}\frac{1}{\alpha!}F^{\alpha}X^{\alpha} onto the space of polynomials of degree \leq p . Remark 4.2 . The set \mathcal{E}^{p}(A) of all \mathcal{C}^{p} -Whitney fields on a locally closed subset A\subset\mathbb{R}^{n} with its natural addition and the multiplication defined by FG:=\pi_{p}(FG) is an \mathbb{R} -algebra. Remark 4.3 . It is also natural to define the composition of \mathcal{C}^{p} -Whitney fields as follows. Let F_{1},\dots,F_{m}\in\mathcal{E}^{p}(A) , where A\subset\mathbb{R}^{n} is locally closed and let H\in\mathcal{E}^{p}(B) , where B\subset\mathbb{R}^{m} is locally closed. Assume that (F_{1}^{0},\dots,F_{m}^{0})(A)\subset B . Put F:=(F_{1},\dots,F_{m}) . Then the composition H\circ F defined by the formula (H\circ F)(u,X):=\pi_{p}[H(F_{1}^{0}(u),\dots,F_{m}^{0}(u),F_{1}(u,X)-F_{1}^{0}(u),\dots,F_{m}(u,X)-F_{m}^{0}(u))], for any u\in A , is a \mathcal{C}^{p} -Whitney field. Remark 4.4 .

Read the complete text