---
title: One-Dimensional Flat Chain Conjecture
url: https://www.emergentmind.com/topics/one-dimensional-flat-chain-conjecture
type: topic
---

# One-Dimensional Flat Chain Conjecture

The one-dimensional flat chain conjecture concerns the relation between Ambrosio–Kirchheim metric \(1\)-currents and classical Federer–Fleming flat chains. In its Euclidean finite-mass form, it asserts that for every metric \(1\)-current \(T\in M_1(\mathbb R^d)\), the associated de Rham current \(\mathbf T\) is a flat \(1\)-chain of finite mass [2603.29683]. In broader metric-space formulations, the conjecture becomes a density statement for normal currents and is characterized by a geometric curve-covering property [2508.08212]. A distinct variant due to Lang removes the finite-mass requirement; in that formulation, the conjecture fails in \(\mathbb R^d\) whenever \(d\ge 2\) [2506.13718].

## 1. Basic definitions and principal formulations

Let \((E,d)\) be a complete metric space. A metric \(1\)-current is a bilinear map
\[
T\colon B^\infty(E)\times\Lip(E)\to\mathbb R
\]
satisfying locality, continuity under pointwise convergence of equi-Lipschitz test functions, and a finite-mass estimate
\[
|T(f,\pi)|\le \Lip(\pi)\int_E |f|\,d\|T\|.
\]
The space of such currents is denoted \(M_1(E)\). If, in addition, there is a finite Borel measure \(\|\partial T\|\) such that
\[
|T(1,f)|\le \int_E |f|\,d\|\partial T\|,\qquad \forall f\in\Lip(E),
\]
then \(T\) is normal, and the corresponding subspace is \(N_1(E)\) [2603.29683].

In the classical Euclidean setting, a de Rham \(m\)-current is a continuous linear functional on smooth compactly supported \(m\)-forms. Its mass is defined by the comass norm, its boundary by \(\partial \mathbf T(\omega)=\mathbf T(d\omega)\), and its flat norm by
\[
F(\mathbf T)=\sup\Bigl\{\mathbf T(\omega):\|\omega\|_\infty\le 1,\ \|d\omega\|_\infty\le 1\Bigr\}.
\]
A flat \(m\)-chain is any \(\mathbf T\) admitting approximation in flat norm by normal currents [2603.29683].

For one-dimensional metric currents on a complete metric space \(X\), the boundary is the \(0\)-current \(\partial T(f)=T(1,f)\), and the homogeneous flat norm can be written as
\[
\mathcal F(T)=\inf\bigl\{\mathbf M(R):R\in\mathbf N_1(X),\ \partial R=\partial T\bigr\}
\]
in the one-dimensional case [2508.08212].

The principal formulations appearing in the literature are summarized below.

| Setting | Assertion | Status |
|---|---|---|
| \(T\in M_1(\mathbb R^d)\) with finite mass | The associated de Rham current \(\mathbf T\) is a flat \(1\)-chain of finite mass | Proved [2603.29683] |
| \(T\in \mathbf M_1(X)\) in a complete, separable metric space | Every \(T\) can be approximated arbitrarily well in the mass norm by normal currents iff \(X\) is curve-rectifiable | Characterized [2508.08212] |
| Lang’s formulation, without requiring finite mass of the underlying currents | Compactly supported metric \(1\)-currents in \(\mathbb R^d\) are flat \(1\)-chains | False for every \(d\ge 2\) [2506.13718] |

The index \(1\) refers to the dimension of the current; the ambient space may be \(\mathbb R^d\) or a general complete metric space.

## 2. Euclidean finite-mass theorem and the short functional-analytic proof

If \(T\in M_1(\mathbb R^d)\), one associates a de Rham \(1\)-current by
\[
\mathbf T\Bigl(\sum_{k=1}^d f_k\,dx_k\Bigr)=\sum_{k=1}^d T(f_k,x_k).
\]
Standard estimates give
\[
M(\mathbf T)\le \sqrt d\,\|T\|(\mathbb R^d)<\infty,
\]
and if \(T\) is normal then \(\mathbf T\) is a normal de Rham current with
\[
M(\partial \mathbf T)\le \|\partial T\|(\mathbb R^d).
\]
The one-dimensional flat chain conjecture in this form states that every \(T\in M_1(\mathbb R^d)\) yields a de Rham current \(\mathbf T\) that is a flat \(1\)-chain of finite mass [2603.29683].

The short proof of Bouafia–De Pauw uses the Lipschitz-free space \(F(E)\). Fixing a base point \(o\in E\), one defines
\[
\Lip_0(E)=\{f\in\Lip(E):f(o)=0\},
\]
and \(F(E)\) as the closed linear span of the evaluation functionals \(\delta_x\) in \(\Lip_0(E)^*\). The map
\[
\Upsilon:F(E)^*\to \Lip_0(E),\qquad \alpha\mapsto \bigl(x\mapsto \alpha(\delta_x)\bigr),
\]
is an isometric isomorphism. Moreover, a net \(f_i\in\Lip_0(E)\) converges weak\(^*\) to \(f\) if and only if \(\sup_i\Lip(f_i)<\infty\) and \(f_i(x)\to f(x)\) for every \(x\) [2603.29683].

The proof then factors the boundary operator through divergence:
\[
L^1(\mathbb R^d;\mathbb R^d)\xrightarrow{[\cdot]} M_1(\mathbb R^d)\xrightarrow{\partial} F(\mathbb R^d),
\qquad \partial([u])=-\Div(u).
\]
For \(u\in L^1(\mathbb R^d;\mathbb R^d)\),
\[
[u](f,\pi)=\int_{\mathbb R^d} (u\,f)\cdot \nabla\pi,
\]
and the associated de Rham current is \(L^d\wedge u\), which is a flat \(1\)-chain by smooth approximation in \(L^1\). The adjoint of
\[
-\Div:L^1\to F(\mathbb R^d)
\]
is
\[
-\Div^*:\Lip_0(\mathbb R^d)\to L^\infty(\mathbb R^d;\mathbb R^d),\qquad f\mapsto \nabla f,
\]
which is an isometric embedding. By the closed-range theorem, \(-\Div\) has closed and dense, hence surjective, range. Therefore for any \(T\in M_1(\mathbb R^d)\) there exists \(u\in L^1\) such that
\[
\partial T=-\Div(u)=\partial([u]).
\]
It follows that \(T-[u]\) has zero boundary and lies in \(N_1(\mathbb R^d)\), so its associated de Rham current is a normal Federer–Fleming current; adding back \(L^d\wedge u\) yields that \(\mathbf T\) is a flat \(1\)-chain of finite mass [2603.29683].

The paper emphasizes that this argument is very short (one page) and relies only on the identification \(F(\mathbb R^d)^*\simeq \Lip_0(\mathbb R^d)\), the closed-range theorem from functional analysis, the classical fact that normal metric currents give normal de Rham currents, and the observation that \(L^1\)-vector fields yield flat chains by convolution [2603.29683].

## 3. Alternative elementary proofs and analytic reformulations

A distinct elementary proof proceeds by decomposing a metric \(1\)-current in \(\mathbb R^n\) into an absolutely continuous part and a singular part. The argument introduces the notion of a purely non-flat current: a classical \(k\)-current \(T\) with finite mass is purely non-flat if whenever
\[
T=T_1+T_2,\qquad M(T)=M(T_1)+M(T_2),
\]
and \(T_1\) is a flat chain, then \(T_1=0\). For such currents, the flat norm equals the mass,
\[
F(T)=M(T),
\]
and also equals the closed flat norm
\[
F_0(T)=\sup\{\langle T,\omega\rangle:\|\omega\|_{\mathrm{comass}}\le 1,\ d\omega=0\}.
\]
A key lemma states that every smooth closed \(1\)-form on \(\mathbb R^n\) with comass at most \(1\) is exact with a global Lipschitz primitive \(\pi\) satisfying \(d\pi=\omega\) and \(\Lip(\pi)\le 1\) [2411.15019].

The singular part is then handled by translation arguments. If \(\eta\) is a finite Radon measure and \(\mu\) is singular to Lebesgue measure, then for almost every small \(v\in\mathbb R^n\), the translates \(\eta\) and \((\tau_v)_\#\mu\) are mutually singular. For a metric \(1\)-current whose associated classical current is purely non-flat,
\[
\limsup_{|v|\to 0} F(\mathbf T-(\tau_v)_\#\mathbf T)=0.
\]
Iterating this small-translation estimate produces a telescoping decomposition into a flat summand and a remainder of vanishing mass, forcing the singular part to be flat as well [2411.15019].

The PDE viewpoint isolates the one-dimensional mechanism behind flatness. In the survey treatment on \(\mathbb R\) or an open interval \(\Omega\subset\mathbb R\), every bounded measurable closed \(1\)-form \(\omega\) satisfies the Poincaré-type estimate
\[
d\phi=\omega,\qquad \phi\in \Lip(\Omega),\qquad \Lip(\phi)\le \|\omega\|_{L^\infty(\Omega)}.
\]
In that setting, the following are equivalent: every metric \(1\)-current in \(\mathbb R\) with finite mass is a flat chain, and for every closed \(\omega\in L^\infty(\Omega)\) there exists \(\phi\in\Lip(\Omega)\) with \(d\phi=\omega\) and \(\Lip(\phi)\le \|\omega\|_{L^\infty}\) [2511.06822].

This analytic reformulation clarifies why one-dimensional arguments are unusually tractable. The proof described in [2411.15019] explicitly contrasts with earlier proofs by Schioppa, Alberti–Bate–Marchese, and De Masi–Marchese, which rely on Alberti representations, width functions, and measure-disintegration techniques [2411.15019].

## 4. Metric-space characterization via curve rectifiability

In a complete metric space \(X\), the one-dimensional Flat–Chain Conjecture can be formulated as a mass-density statement:
\[
\text{for every }T\in\mathbf M_1(X)\text{ and every }\varepsilon>0\text{ there exists }N\in\mathbf N_1(X)\text{ with }\mathbf M(T-N)<\varepsilon.
\]
Arroyo-Rabasa–Bouchitté identify the geometric hypothesis under which this holds. They call \(X\) curve-rectifiable if every \(1\)-rectifiable set is covered, up to \(\mathcal H^1\)-null, by countably many Lipschitz curves. Their Theorem A states that for a complete, separable metric space, the following are equivalent: \(X\) is curve-rectifiable, and the \(1\)-FCC holds in \(X\) [2508.08212].

The proof is organized around an SBV representation. Any \(T\in\mathbf M_1(X)\) admits a decomposition
\[
T=\int_{\Theta_{\mathrm{SBV}}(X)} \llbracket u\rrbracket^a\,d\eta(u),
\]
where \(\Theta_{\mathrm{SBV}}(X)\) is the set of injective curves \(u\in \mathrm{SBV}([0,1],X)\) of constant metric speed, \(\llbracket u\rrbracket^a\) is the current of its absolutely continuous part, and \(\eta\) is a finite Borel measure on the space of càdlàg curves. Moreover,
\[
\mathbf M(T)=\int \ell(u)\,d\eta(u),
\]
and the total jump mass satisfies
\[
\int |D^j u|\,d\eta(u)\le \|\partial T\|_{\mathrm{KR}}+\varepsilon.
\]
In arbitrary complete metric spaces, this yields a Smirnov-type decomposition for one-dimensional currents as a superposition, without mass cancellation, of currents associated with curves of bounded variation that have a vanishing Cantor part [2508.08212].

Optimal-transport methods enter through the Kantorovich–Rubinstein norm of the boundary,
\[
\|\partial T\|_{\mathrm{KR}}=\sup\{\partial T(f):f\in\Lip_1(X)\},
\]
for which one always has \(\|\partial T\|_{\mathrm{KR}}\le \mathbf M(T)\). In a piecewise–quasiconvex space, every molecule \(m\in\AE(X)\) admits a rectifiable normal filling \(R\) with \(\partial R=m\) and
\[
\mathbf M(R)\le c(1+\delta)\|m\|_{\mathrm{KR}}.
\]
This leads to approximation by normal currents modulo a cycle and, in Banach spaces, to polyhedral approximation modulo a cycle [2508.08212].

A Banach-space filling corollary states that if \(T\in\mathbf M_1(E)\), \(E\) Banach, \(\dim E\ge 2\), then there is a rectifiable cycle \(S\in\mathbf N_1(E)\) with \(\partial S=\partial T\) and
\[
\mathbf M(S)\le \|\partial T\|_{\mathrm{KR}}+\varepsilon.
\]
The same framework proves the necessity direction: if some rectifiable set cannot be covered by countably many curves, its associated fragment current cannot be mass-approximated by normals [2508.08212].

## 5. Lang’s formulation and failure without finite mass

Lang’s formulation extends Ambrosio–Kirchheim currents to allow infinite-mass currents on locally compact metric spaces. In this setting, a \(k\)-current is a multilinear functional
\[
T:\Lip_c(X)\times[\Lip(X)]^k\to \mathbb R
\]
satisfying locality and joint continuity. If \(X=\mathbb R^d\) and \(T\) has compact support, it may be identified with a classical de Rham current \(C_k(T)\) [2506.13718].

For \(k=1\), Lang’s conjecture states: let \(T\) be any compactly supported metric \(1\)-current in \(\mathbb R^d\). Then its comparison \(C_1(T)\) belongs to the space of flat \(1\)-chains \(\mathfrak F_1(K)\) for some compact \(K\); equivalently, \(T\) is a flat \(1\)-chain. The main theorem of [2506.13718] shows that for every \(d\ge 2\) there exists a compactly supported metric \(1\)-current in \(\mathbb R^d\) which is not a flat chain. Equivalently, there is a metric \(1\)-current with infinite flat norm.

The proof is based on a reformulation as a regularity problem for the prescribed Jacobian equation near \(L^\infty\). The nonlinear span of Jacobians of uniformly Lipschitz maps is studied through sets of the form
\[
K_L=\{\,f\,\det D\pi:\max\{\Lip f,\|f\|_\infty\}\cdot \prod_j \Lip \pi^j\le 1\,\},
\]
and the convex hull \(\operatorname{conv} K_L\subset L^\infty(Q)\) is shown to have empty interior, indeed to be nowhere dense, in the weak-\(^*\) topology. A Baire-category argument then yields non-surjectivity of the relevant map \(Em\), which in turn implies failure of the conjecture [2506.13718].

This failure result concerns Lang’s formulation, that is, without requiring finite mass of the underlying currents. The same paper states that in all other cases the conjecture holds; for \(k=1\) in \(d=1\), every metric \(1\)-current in \(\mathbb R\) is of finite mass and hence flat, and for \(k=0\) the statement is trivial [2506.13718]. The finite-mass Euclidean theorem and the infinite-mass counterexample therefore address different formulations.

## 6. Consequences, neighboring results, and related variants

In the Euclidean finite-mass setting, the short proof has several immediate consequences. It identifies normal metric \(1\)-currents with Federer–Fleming normal de Rham currents and yields a transparent decomposition of any metric \(1\)-current into a divergence-free part plus an \(L^1\)-gradient part. The paper also notes that the methods may extend to other contexts, for example quasiconvex metric spaces, and suggest new functional-analytic approaches to higher-dimensional versions [2603.29683].

In the general metric-space theory, the resolution of the \(1\)-FCC produces more than a density theorem. It gives controlled fillings via the Kantorovich–Rubinstein norm, approximation by polyhedral currents modulo a cycle in Banach spaces, and a new SBV-curve decomposition. The abstract formulates the central equivalence as follows: metric currents can be approximated in the mass norm by normal currents if and only if every \(1\)-rectifiable set can be covered by countably many Lipschitz curves up to an \(\mathscr H^1\)-negligible set [2508.08212].

A distinct, but historically adjacent, use of the same label appears in the theory of flat chains modulo \(p\). In that quotient-group setting, the conjecture asserts that every class \([T]\in I_{1,K}^p(\mathbb R^n)\) contains a genuine integral \(1\)-current. Marchese–Stuvard prove that for \(p\ge 2\), \(n\ge 1\), and compact \(K\subset \mathbb R^n\), there exists \(C=p-1\) such that every class \([S]\in R_{1,K}^p(\mathbb R^n)\) has a representative \(S'\in R_{1,K}(\mathbb R^n)\) with
\[
M(S')\le C\,M^p([S]),\qquad M(\partial S')\le C\,M^p(\partial [S]).
\]
In particular, every \(1\)-dimensional class admits an integral representative, and the natural map
\[
I_{1,K}(\mathbb R^n)\to I_{1,K}^p(\mathbb R^n),\qquad T\mapsto [T],
\]
is surjective [1607.05138].

Taken together, these developments show that the phrase “one-dimensional flat chain conjecture” does not refer to a single universal statement. In finite-mass Euclidean space it is a theorem; in complete separable metric spaces it is equivalent to curve-rectifiability; in Lang’s infinite-mass formulation it fails in every nontrivial Euclidean case \(d\ge 2\); and in the modulo-\(p\) setting it becomes a surjectivity statement for integral representatives [2603.29683].

Source: https://www.emergentmind.com/topics/one-dimensional-flat-chain-conjecture