One-Dimensional Flat Chain Conjecture

Updated 8 July 2026

The one-dimensional flat chain conjecture defines the relation between metric 1-currents and classical flat chains, showing that every finite-mass current yields a flat de Rham current.
A short functional-analytic proof uses Lipschitz-free spaces and the closed-range theorem to decompose currents via divergence, establishing the finite-mass result in Euclidean spaces.
In broader metric spaces, the conjecture characterizes flatness through curve rectifiability, while Lang’s variant fails for infinite-mass currents in R^d when d ≥ 2.

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 (Bouafia et al., 31 Mar 2026). In broader metric-space formulations, the conjecture becomes a density statement for normal currents and is characterized by a geometric curve-covering property (Arroyo-Rabasa et al., 11 Aug 2025). A distinct variant due to Lang removes the finite-mass requirement; in that formulation, the conjecture fails in $\mathbb R^d$ whenever $d\ge 2$ (Takáč, 16 Jun 2025).

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

$1$0

The space of such currents is denoted $1$1. If, in addition, there is a finite Borel measure $1$2 such that

$1$3

then $1$4 is normal, and the corresponding subspace is $1$5 (Bouafia et al., 31 Mar 2026).

In the classical Euclidean setting, a de Rham $1$6-current is a continuous linear functional on smooth compactly supported $1$7-forms. Its mass is defined by the comass norm, its boundary by $1$8, and its flat norm by

$1$9

A flat $T\in M_1(\mathbb R^d)$ 0-chain is any $T\in M_1(\mathbb R^d)$ 1 admitting approximation in flat norm by normal currents (Bouafia et al., 31 Mar 2026).

For one-dimensional metric currents on a complete metric space $T\in M_1(\mathbb R^d)$ 2, the boundary is the $T\in M_1(\mathbb R^d)$ 3-current $T\in M_1(\mathbb R^d)$ 4, and the homogeneous flat norm can be written as

$T\in M_1(\mathbb R^d)$ 5

in the one-dimensional case (Arroyo-Rabasa et al., 11 Aug 2025).

The principal formulations appearing in the literature are summarized below.

Setting	Assertion	Status
$T\in M_1(\mathbb R^d)$ 6 with finite mass	The associated de Rham current $T\in M_1(\mathbb R^d)$ 7 is a flat $T\in M_1(\mathbb R^d)$ 8-chain of finite mass	Proved (Bouafia et al., 31 Mar 2026)
$T\in M_1(\mathbb R^d)$ 9 in a complete, separable metric space	Every $\mathbf T$ 0 can be approximated arbitrarily well in the mass norm by normal currents iff $\mathbf T$ 1 is curve-rectifiable	Characterized (Arroyo-Rabasa et al., 11 Aug 2025)
Lang’s formulation, without requiring finite mass of the underlying currents	Compactly supported metric $\mathbf T$ 2-currents in $\mathbf T$ 3 are flat $\mathbf T$ 4-chains	False for every $\mathbf T$ 5 (Takáč, 16 Jun 2025)

The index $\mathbf T$ 6 refers to the dimension of the current; the ambient space may be $\mathbf T$ 7 or a general complete metric space.

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

If $\mathbf T$ 8, one associates a de Rham $\mathbf T$ 9-current by

$1$0

Standard estimates give

$1$1

and if $1$2 is normal then $1$3 is a normal de Rham current with

$1$4

The one-dimensional flat chain conjecture in this form states that every $1$5 yields a de Rham current $1$6 that is a flat $1$7-chain of finite mass (Bouafia et al., 31 Mar 2026).

The short proof of Bouafia–De Pauw uses the Lipschitz-free space $1$8. Fixing a base point $1$9, one defines

$\mathbb R^d$ 0

and $\mathbb R^d$ 1 as the closed linear span of the evaluation functionals $\mathbb R^d$ 2 in $\mathbb R^d$ 3. The map

$\mathbb R^d$ 4

is an isometric isomorphism. Moreover, a net $\mathbb R^d$ 5 converges weak $\mathbb R^d$ 6 to $\mathbb R^d$ 7 if and only if $\mathbb R^d$ 8 and $\mathbb R^d$ 9 for every $d\ge 2$ 0 (Bouafia et al., 31 Mar 2026).

The proof then factors the boundary operator through divergence: $d\ge 2$ 1 For $d\ge 2$ 2,

$d\ge 2$ 3

and the associated de Rham current is $d\ge 2$ 4, which is a flat $d\ge 2$ 5-chain by smooth approximation in $d\ge 2$ 6. The adjoint of

$d\ge 2$ 7

$d\ge 2$ 8

which is an isometric embedding. By the closed-range theorem, $d\ge 2$ 9 has closed and dense, hence surjective, range. Therefore for any $(E,d)$ 0 there exists $(E,d)$ 1 such that

$(E,d)$ 2

It follows that $(E,d)$ 3 has zero boundary and lies in $(E,d)$ 4, so its associated de Rham current is a normal Federer–Fleming current; adding back $(E,d)$ 5 yields that $(E,d)$ 6 is a flat $(E,d)$ 7-chain of finite mass (Bouafia et al., 31 Mar 2026).

The paper emphasizes that this argument is very short (one page) and relies only on the identification $(E,d)$ 8, the closed-range theorem from functional analysis, the classical fact that normal metric currents give normal de Rham currents, and the observation that $(E,d)$ 9-vector fields yield flat chains by convolution (Bouafia et al., 31 Mar 2026).

3. Alternative elementary proofs and analytic reformulations

A distinct elementary proof proceeds by decomposing a metric $1$0-current in $1$1 into an absolutely continuous part and a singular part. The argument introduces the notion of a purely non-flat current: a classical $1$2-current $1$3 with finite mass is purely non-flat if whenever

$1$4

and $1$5 is a flat chain, then $1$6. For such currents, the flat norm equals the mass,

$1$7

and also equals the closed flat norm

$1$8

A key lemma states that every smooth closed $1$9-form on $T\colon B^\infty(E)\times\Lip(E)\to\mathbb R$0 with comass at most $T\colon B^\infty(E)\times\Lip(E)\to\mathbb R$1 is exact with a global Lipschitz primitive $T\colon B^\infty(E)\times\Lip(E)\to\mathbb R$2 satisfying $T\colon B^\infty(E)\times\Lip(E)\to\mathbb R$3 and $T\colon B^\infty(E)\times\Lip(E)\to\mathbb R$4 (Marchese et al., 2024).

The singular part is then handled by translation arguments. If $T\colon B^\infty(E)\times\Lip(E)\to\mathbb R$5 is a finite Radon measure and $T\colon B^\infty(E)\times\Lip(E)\to\mathbb R$6 is singular to Lebesgue measure, then for almost every small $T\colon B^\infty(E)\times\Lip(E)\to\mathbb R$7, the translates $T\colon B^\infty(E)\times\Lip(E)\to\mathbb R$8 and $T\colon B^\infty(E)\times\Lip(E)\to\mathbb R$9 are mutually singular. For a metric $1$00-current whose associated classical current is purely non-flat,

$1$01

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 (Marchese et al., 2024).

The PDE viewpoint isolates the one-dimensional mechanism behind flatness. In the survey treatment on $1$02 or an open interval $1$03, every bounded measurable closed $1$04-form $1$05 satisfies the Poincaré-type estimate

$1$06

In that setting, the following are equivalent: every metric $1$07-current in $1$08 with finite mass is a flat chain, and for every closed $1$09 there exists $1$10 with $1$11 and $1$12 (Marchese, 10 Nov 2025).

This analytic reformulation clarifies why one-dimensional arguments are unusually tractable. The proof described in (Marchese et al., 2024) 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 (Marchese et al., 2024).

4. Metric-space characterization via curve rectifiability

In a complete metric space $1$13, the one-dimensional Flat–Chain Conjecture can be formulated as a mass-density statement: $1$14 Arroyo-Rabasa–Bouchitté identify the geometric hypothesis under which this holds. They call $1$15 curve-rectifiable if every $1$16-rectifiable set is covered, up to $1$17-null, by countably many Lipschitz curves. Their Theorem A states that for a complete, separable metric space, the following are equivalent: $1$18 is curve-rectifiable, and the $1$19-FCC holds in $1$20 (Arroyo-Rabasa et al., 11 Aug 2025).

The proof is organized around an SBV representation. Any $1$21 admits a decomposition

$1$22

where $1$23 is the set of injective curves $1$24 of constant metric speed, $1$25 is the current of its absolutely continuous part, and $1$26 is a finite Borel measure on the space of càdlàg curves. Moreover,

$1$27

and the total jump mass satisfies

$1$28

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 (Arroyo-Rabasa et al., 11 Aug 2025).

Optimal-transport methods enter through the Kantorovich–Rubinstein norm of the boundary,

$1$29

for which one always has $1$30. In a piecewise–quasiconvex space, every molecule $1$31 admits a rectifiable normal filling $1$32 with $1$33 and

$1$34

This leads to approximation by normal currents modulo a cycle and, in Banach spaces, to polyhedral approximation modulo a cycle (Arroyo-Rabasa et al., 11 Aug 2025).

A Banach-space filling corollary states that if $1$35, $1$36 Banach, $1$37, then there is a rectifiable cycle $1$38 with $1$39 and

$1$40

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 (Arroyo-Rabasa et al., 11 Aug 2025).

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 $1$41-current is a multilinear functional

$1$42

satisfying locality and joint continuity. If $1$43 and $1$44 has compact support, it may be identified with a classical de Rham current $1$45 (Takáč, 16 Jun 2025).

For $1$46, Lang’s conjecture states: let $1$47 be any compactly supported metric $1$48-current in $1$49. Then its comparison $1$50 belongs to the space of flat $1$51-chains $1$52 for some compact $1$53; equivalently, $1$54 is a flat $1$55-chain. The main theorem of (Takáč, 16 Jun 2025) shows that for every $1$56 there exists a compactly supported metric $1$57-current in $1$58 which is not a flat chain. Equivalently, there is a metric $1$59-current with infinite flat norm.

The proof is based on a reformulation as a regularity problem for the prescribed Jacobian equation near $1$60. The nonlinear span of Jacobians of uniformly Lipschitz maps is studied through sets of the form

$1$61

and the convex hull $1$62 is shown to have empty interior, indeed to be nowhere dense, in the weak-$1$63 topology. A Baire-category argument then yields non-surjectivity of the relevant map $1$64, which in turn implies failure of the conjecture (Takáč, 16 Jun 2025).

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 $1$65 in $1$66, every metric $1$67-current in $1$68 is of finite mass and hence flat, and for $1$69 the statement is trivial (Takáč, 16 Jun 2025). The finite-mass Euclidean theorem and the infinite-mass counterexample therefore address different formulations.

In the Euclidean finite-mass setting, the short proof has several immediate consequences. It identifies normal metric $1$70-currents with Federer–Fleming normal de Rham currents and yields a transparent decomposition of any metric $1$71-current into a divergence-free part plus an $1$72-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 (Bouafia et al., 31 Mar 2026).

In the general metric-space theory, the resolution of the $1$73-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$74-rectifiable set can be covered by countably many Lipschitz curves up to an $1$75-negligible set (Arroyo-Rabasa et al., 11 Aug 2025).

A distinct, but historically adjacent, use of the same label appears in the theory of flat chains modulo $1$76. In that quotient-group setting, the conjecture asserts that every class $1$77 contains a genuine integral $1$78-current. Marchese–Stuvard prove that for $1$79, $1$80, and compact $1$81, there exists $1$82 such that every class $1$83 has a representative $1$84 with

$1$85

In particular, every $1$86-dimensional class admits an integral representative, and the natural map

$1$87

is surjective (Marchese et al., 2016).

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 $1$88; and in the modulo-$1$89 setting it becomes a surjectivity statement for integral representatives (Bouafia et al., 31 Mar 2026).