---
title: Whitney Extension Theorem
url: https://www.emergentmind.com/topics/whitney-extension-theorem
type: topic
---

# Whitney Extension Theorem

The Whitney Extension Theorem concerns extending differentiable data prescribed on a closed set \(F \subseteq \mathbb{R}^n\) to a globally \(C^m\) function. In its classical form, it gives necessary and sufficient compatibility conditions on a finite family of continuous functions \(\{f^{(\alpha)}\}_{|\alpha|\le m}\) on \(F\), interpreted as the partial derivatives of the sought extension. Over time, the theorem has become a central organizing principle for problems in jets, interpolation, geometric analysis, sub-Riemannian geometry, ultradifferentiable classes, Banach-space smoothness, and computable analysis [2507.02113].

## 1. Classical Euclidean formulation

Let \(E \subset \mathbb{R}^n\) be closed and \(m \in \mathbb{N}\). A Whitney jet of order \(m\) on \(E\) is a family \(\{f_\alpha:E\to\mathbb{R}\}_{|\alpha|\le m}\), indexed by multiindices, intended to represent the restrictions of the derivatives \(D^\alpha F|_E\) of some \(F \in C^m(\mathbb{R}^n)\). If \(x,y\in E\), the associated Taylor polynomial at \(x\) is
\[
P_x(y)=\sum_{|\alpha|\le m} f^{(\alpha)}(x)\frac{(y-x)^\alpha}{\alpha!}.
\]
The Whitney compatibility conditions require continuity of the \(f_\alpha\) and a uniform Taylor-remainder estimate. One standard form is
\[
\left| f^{(\alpha)}(y)-\sum_{|\beta|\le m-|\alpha|}\frac{f^{(\alpha+\beta)}(x)}{\beta!}(y-x)^\beta \right|
\le C \|y-x\|^{m-|\alpha|},
\]
for all \(x,y\in E\) and \(|\alpha|\le m\). Classically, these conditions are necessary and sufficient for the existence of \(g\in C^m(\mathbb{R}^n)\) such that \(\partial^\alpha g|_E=f^{(\alpha)}\) for all \(|\alpha|\le m\) [2507.02113].

The same framework has a \(C^{m,\alpha}\) version, where the remainder is required to satisfy a Hölder bound of order \(m-|\alpha|+\alpha\). In the vector-valued case with target \(\mathbb{R}^d\), the theorem remains valid componentwise. An equivalent jet formulation uses polynomial fields \(J=\{P_x\}_{x\in E}\) of degree at most \(m\), and the Whitney–Glaeser compatibility may be written through quantities such as
\[
\max_{|\alpha|\le m}\frac{|D^\alpha P_x(x)-D^\alpha P_y(x)|}{|x-y|^{m-|\alpha|}},
\]
which, for \(p=\infty\), leads to the classical \(C^{m,1}\) refinement [1607.01660].

## 2. Linear extension operators and quantitative refinements

A major structural feature of Whitney theory is the existence of linear extension operators. For fixed closed \(F\subseteq\mathbb{R}^n\), one may choose a linear operator
\[
E_{m,n,F}: J^m(F)\to C^m(\mathbb{R}^n),
\]
mapping a jet to an extension, with operator norm controlled in terms of \(m\) and \(n\); Stein and Fefferman are explicitly cited in this context in the computability literature [2507.02113].

Quantitative control becomes especially მნიშვნელოვანი in high dimension. A variant proved for \(\mathcal C^m(\mathbb{R}^n)\) constructs a linear extension operator whose norm grows polynomially rather than exponentially with the ambient dimension: the operator norm is at most \(C n^{5m/2}\), with \(C\) depending only on \(m\). The construction starts from a Whitney decomposition of \(\mathbb{R}^n\setminus E\), introduces a partition of unity on enlarged Whitney cubes, and then averages the resulting extension over translated dyadic grids to suppress the exponential dependence on \(n\) that appears in the standard proof [1508.01779].

The classical operator is also robust outside the purely \(C^m\) setting. For \(p>n\), Shvartsman gives an intrinsic characterization of traces of jets generated by \(L^{m+1}_p(\mathbb{R}^n)\) functions by means of variational seminorms over \(y\)-sparse families of pairs, and shows that the very same classical linear Whitney extension operator yields an almost optimal Sobolev extension. In the limit \(p=\infty\), this reduces to the Whitney–Glaeser theorem [1607.01660].

## 3. Structured Euclidean variants

Several important branches of the theory impose extra geometric or regularity constraints on the extension.

For convexity, the data are a pair \((f,G)\) on \(C\subseteq\mathbb{R}^n\), with \(f:C\to\mathbb{R}\) and \(G:C\to\mathbb{R}^n\). The extension problem is no longer governed solely by local Taylor compatibility. In the \(C^{1,\omega}\) setting, Azagra and Mudarra identify a necessary and sufficient convex Whitney condition
\[
f(x)-f(y)-\langle G(y),x-y\rangle
\ge \eta \,\|G(x)-G(y)\|\,\omega^{-1}\!\Big(\frac{1}{2M}\|G(x)-G(y)\|\Big),
\]
and show that it characterizes the existence of a convex \(F\in C^{1,\omega}(\mathbb{R}^n)\) with \(F=f\) and \(\nabla F=G\) on \(C\). For compact \(C\), the \(C^1\) convex theory is described by the supporting inequality \((C)\) and the flat-piece condition \((CW1)\), rather than by classical Whitney conditions alone [1507.03931].

In ultradifferentiable Roumieu classes, extension is controlled by weight matrices. For an admissible weight matrix \(\mathbb{M}\), one can explicitly compute a descendant class \(\mathcal C'\) such that every Whitney jet of class \(\mathcal C'\) on an arbitrary compact set extends to a function in \(\mathcal C\). The descendant is defined from a weight sequence \(N\) by
\[
T_k=1+\sum_{j>k}\frac{1}{v_j},\qquad o_k=T_k^k,\qquad S_k=\prod_{j=0}^k o_j,
\]
where \(v_k=N_k/N_{k-1}\), and the extension operator preserves controlled growth [1607.01206]. In the Roumieu weight-function formulation, controlled loss from \(\sigma\) to \(\omega\) is characterized by the integral condition
\[
\int_1^\infty \omega(tu)\,\frac{du}{u^2}\le C\sigma(t)+C,
\]
and one result of Rainer–Schindl is precisely that a previously imposed extra matrix condition can be removed [1808.10253]. In the Beurling setting, a corresponding theorem requires an \(r\)-strong pair \((\omega,\sigma)\), namely
\[
\int_1^\infty \omega(tu)\,\frac{du}{u^{1+r}}\le C\sigma(t)+C,
\]
for some \(r\in(0,1)\), and yields extension with controlled loss of regularity [2011.02178].

There are also vector-valued and scale-valued analogues. For a real Hausdorff locally convex target space \(E\), Whitney \(k\)-jets on a closed \(A\subseteq\mathbb{R}^n\) admit a continuous linear extension operator for finite \(k\), and if \(E\) is metrizable then Whitney \(\infty\)-jets also extend [2307.03473]. For scales of Banach spaces with smoothing operators, Baldi modifies the classical Whitney formula by inserting a smoothing \(S_{\theta_Q}\) on each Whitney cube \(Q\), with \(\theta_Q \asymp 1/\operatorname{diam}(Q)\), so that the jet components may live in different levels of the scale while the remainder is measured in the weakest space [2010.07236].

## 4. Banach-space, manifold, and nonlinear generalizations

The \(C^1\) Banach-space theory replaces finite-dimensional Taylor polynomials with strict derivatives. If \(X\) and \(Y\) are Banach spaces, \(A\subset X\) is closed, and \(G:A\to\mathcal L(X,Y)\) is continuous, a Whitney-type hypothesis may be written as
\[
\|f(z)-f(y)-G(x)(z-y)\|\le \varepsilon \|z-y\|,
\]
for nearby \(y,z\in A\), or globally in modulus form
\[
\|f(x)-f(y)-G(y)(x-y)\|\le \|x-y\|\,\omega(\|x-y\|),\qquad
\|G(x)-G(y)\|\le \omega(\|x-y\|).
\]
Under a Lipschitz extension property \((LE)\) and a Lipschitz-approximation property \((LA_1)\), this yields \(g\in C^1(\Omega;Y)\) with \(g|_A=f\) and \(Dg|_A=G\). In the Lipschitz case, one gets quantitative bounds such as \(\operatorname{Lip}(g)\le 16C^3L\), improved later to \(4C^2L\) under a refined constant bookkeeping, and higher regularity \(g\in C^k(X\setminus A;Y)\) when \((LA_k)\) is available [2403.14317].

The higher-order Banach-space picture is markedly more rigid. Johanis proves that vector-valued \(C^{1,\omega}\) extension theorems do hold when the target is injective, for example \(\ell_\infty\), but fail for mappings into “somewhat euclidean” spaces. He also proves negative results for scalar \(C^{2,+}\), \(C^{2,\omega}\), and \(C^3\) extension problems on infinite-dimensional spaces; the scalar \(C^2\) Whitney extension problem on \(\ell_2\) is explicitly left open [2507.09384].

On manifolds, the theorem becomes a statement about restriction maps between function spaces. If \(M\) and \(N\) are smooth manifolds and \(C\subset M\) is either a submanifold with corners or a compact submanifold with rough boundary, the restriction map
\[
\mathrm{res}_C^M:C^\infty(M,N)\to C^\infty(C,N)
\]
is a submersion of locally convex manifolds. At the linear level, if \(E\to M\) is a finite-rank vector bundle and \(C\subset M\) satisfies the cusp condition, the restriction
\[
\Gamma_c(M,E)\to \Gamma_c(C,E)
\]
has a continuous linear splitting [1801.04126]. A different nonlinear direction appears on symmetric spaces: for certain compact and noncompact homogeneous spaces, finite value data at points can be interpolated by analytic, square-integrable, \(K\)-finite functions constructed from spherical harmonics or Flensted–Jensen functions. This is “Whitney-type” in the sense of extension from finite closed sets, but it is not a full jet-extension theorem [2407.21420].

## 5. Carnot groups, horizontal curves, and pliability

In Carnot groups, the noncommutative geometry replaces ordinary derivatives by Pansu differentials and linear Taylor polynomials by homogeneous homomorphisms. If \(G_1,G_2\) are Carnot groups, a \(C_H^1\)-Whitney condition on compact \(K\subset G_1\) for data \((f,L)\), with \(L(p)\in\operatorname{Hom}_{\mathrm{hom}}(G_1,G_2)\), takes the form
\[
d_{G_2}\big(f(q),\,f(p)\cdot L(p)(p^{-1}\!\cdot q)\big)\le \omega\big(d_{G_1}(p,q)\big),
\]
where \(\omega(t)=o(t)\). For curves, \(G_1=\mathbb R\), every homogeneous homomorphism is \(h\mapsto \exp(hX)\), and the condition becomes
\[
r_{K,\eta}:=
\sup_{\tau,t\in K,\ 0<|\tau-t|<\eta}
\frac{d_{G_2}\big(f(t),\,f(\tau)\cdot \exp((t-\tau)X(\tau))\big)}{|\tau-t|}
\to 0.
\]
The central theorem is that \((\mathbb R,G)\) has the \(C_H^1\) extension property if and only if \(G\) is pliable, meaning that every straight horizontal curve \(t\mapsto \exp(tX)\) can be perturbed in a \(C_H^1\)-small way so that endpoint and terminal velocity vary in a neighborhood. All step-2 Carnot groups are pliable, but there are non-pliable groups, including the Engel group, and there are pliable groups of arbitrarily large step [1603.02639].

The Heisenberg-group case exposes the additional compatibility hidden by noncommutativity. For \(C^1\) horizontal curves \(T=(f_1,g_1,\dots,f_n,g_n,h)\) on a compact \(K\subset\mathbb R\), extension requires the ordinary Whitney \(C^1\) conditions on the coordinate jets, the pointwise horizontality relation
\[
h'(s)=2\sum_{i=1}^n\big(f_i'(s)g_i(s)-f_i(s)g_i'(s)\big),
\]
and a quadratic compatibility involving the group law,
\[
\frac{h(b)-h(a)-2\sum_{i=1}^n(f_i(b)g_i(a)-f_i(a)g_i(b))}{|b-a|^2}\to 0
\]
as \(|b-a|\to 0\) with \(a,b\in K\) [1507.02240]. For \(C^m\) horizontal curves in \(\mathbb H^1\), Pinamonti–Speight–Zimmerman show that one must add polynomial identities for all derivatives up to order \(m\) and a scale-sensitive area discrepancy condition \(A(a,b)/V(a,b)\to 0\), where \(A(a,b)\) compares the actual vertical increment with the one predicted by the horizontal Taylor polynomials [1807.07936]. In the \(C^{m,\omega}\) setting, Speight–Zimmerman prove that the correct necessary and sufficient condition involves an \(\omega\)-adapted velocity \(V_\omega\); a more naive analogue based on the \(C^m\) condition is too weak and can fail even to imply a \(C^{m,\omega^\alpha}\) extension for any \(\alpha\in(1/2,1]\) [2204.03544].

For free step-2 Carnot groups \(\mathbb G_r\), a single area condition is no longer enough. Shibahara introduces generalized area/velocity constraints \(E_{ij}(a,b,c,\hat c)\) that encode quantitative linear dependences among the horizontal jets and the interaction of several vertical area components. The resulting \(C^m\) Whitney extension theorem for horizontal curves in \(\mathbb G_r\) extends the Heisenberg case and makes explicit that in higher rank the vertical coordinates cannot be corrected independently [2308.01991].

## 6. Computability, algebraic reformulations, and open directions

One modern line of work asks not only whether an extension exists, but whether it can be produced effectively. In Type-2 Theory of Effectivity, if a closed \(F\subseteq\mathbb R^n\) is given with a representation making the distance function \(x\mapsto d(x,F)\) computable, and if a Whitney jet of order \(m\) is computable together with a Whitney constant, then one can compute an extension \(g\in C^m(\mathbb R^n)\) with matching derivatives. The proof effectivizes Stein’s construction through a computable Whitney decomposition, a computably smooth partition of unity, approximate projections \(r_Q\in F\), and explicit finite local summation formulas [2507.02113].

A second reformulation is algebraic. For a closed \(E\subseteq\mathbb R^n\), the classical theorem may be viewed as the surjectivity of the completion map for the differential-power filtration along \(E\). Belitskii–Kerner extend this to general \(C^\infty\)-rings and general filtrations, giving necessary and sufficient conditions for surjectivity of the completion map and proving that every element of the completion has a \(C^\infty\)-representative that is real-analytic outside the locus of completion, can satisfy prescribed positivity conditions, and may also satisfy compatible linear constraints [1907.10053].

There is also a Schwartz-space analogue on the positive orthant. For
\[
\mathcal S(\mathbb R_+^d)=\{f\in C^\infty(\mathbb R_+^d): D^p f \text{ extends continuously to } [0,\infty)^d,\ \sup_{x\in\mathbb R_+^d}|x^kD^pf(x)|<\infty\},
\]
the restriction map
\[
\mathcal S(\mathbb R^d)\to \mathcal S(\mathbb R_+^d)
\]
is a topological homomorphism onto. This “extension theorem of Whitney type” is obtained from Laguerre expansions and the Schwartz kernel theorem for \(\mathcal S(\mathbb R_+^d)\) [1602.04008].

Several open problems remain explicit in the literature surveyed here. In Carnot geometry, the case of general domain groups \(G_1\) is still open in the non-Abelian target setting, as is a full intrinsic characterization of pliability for all Carnot groups; higher-order \(C_H^m\) extension theory is also open in that setting [1603.02639]. In Banach spaces, the scalar \(C^2\) Whitney extension problem on \(\ell_2\) remains unresolved [2507.09384]. These unresolved cases suggest that the theorem is not a single statement but a family of extension principles whose exact form depends sharply on the ambient geometry, the target regularity class, and the algebraic structure imposed on the data.

Source: https://www.emergentmind.com/topics/whitney-extension-theorem