Almost Finsler Manifolds Overview

• Almost Finsler manifolds are geometric structures that relax classical Finsler conditions by employing slit-based exclusions and partial domains.
• They incorporate diverse formulations including partial Minkowski models, pseudo-Finsler metrics, and even local almost hypercomplex structures.
• The framework enables new classifications via almost isometries and practical applications in Randers-type models and rigidity theorem derivations.

Almost Finsler manifolds arise in several closely related but non-identical extensions of classical Finsler geometry. In one direct recent formulation, a partial or almost Finsler manifold is a triple $(M,S,F)$ with $M$ smooth, $S\subset TM$ a closed conelike subset containing the zero section, and $F:TM-S\to(0,\infty)$ smooth and positively $1$-homogeneous, such that each fiber $(T_xM,S\cap T_xM,F|_x)$ is a partial or almost Minkowski space; the almost case requires positive definiteness of the fiberwise fundamental tensor away from the slit (Davis et al., 29 Aug 2025). Other works use the expression for even-dimensional pseudo-Finsler manifolds carrying local almost hypercomplex structures (Moghaddam, 2013), or for Finsler manifolds considered up to almost isometry through their forward quasi-metric and smooth semi-Lipschitz cone (Daniilidis et al., 2019). This suggests that the term is context-dependent, but in each usage it marks a controlled weakening, enrichment, or equivalence-based reformulation of the standard Finsler framework.

1. Foundational setting and competing definitions

A standard Finsler metric on a smooth manifold $M$ is a function $F:TM\setminus\{0\}\to(0,\infty)$ that is positively $1$-homogeneous in the fiber variables and whose fundamental tensor

$$g_{ij}(x,y)=\frac{1}{2}\frac{\partial^2(F^2)(x,y)}{\partial y^i\partial y^j}$$

is positive definite for every nonzero $M$0. In the formulation of Javaloyes–Sánchez, two basic relaxations are separated: a conic pseudo-Finsler metric allows the domain to be a conic open subset $M$1 and permits the fundamental tensor to be indefinite, while a conic Finsler metric is a conic pseudo-Finsler metric with positive-definite fundamental tensor on $M$2; a pseudo-Finsler metric is the full-domain case $M$3 with no positivity requirement, and a standard Finsler metric is recovered when $M$4 and $M$5 (Javaloyes et al., 2011).

The more recent partial/almost formalism changes the domain language from a conic open set to a slit $M$6. A partial Finsler manifold $M$7 is defined on $M$8, where $M$9 is closed, conelike, nonempty, and contains the zero section; an almost Finsler manifold is the corresponding positive-definite case. The fiberwise model is a partial or almost Minkowski space, so the fundamental tensor is still obtained from the second fiber derivatives of $S\subset TM$0, but positivity is only required away from the slit (Davis et al., 29 Aug 2025).

These definitions are not interchangeable. In the even-dimensional pseudo-Finsler literature, an “almost Finsler manifold” may instead mean an even-dimensional pseudo-Finsler manifold $S\subset TM$1 with $S\subset TM$2 open, $S\subset TM$3, $S\subset TM$4, and $S\subset TM$5 positively $S\subset TM$6-homogeneous, together with local almost hypercomplex structures built from a non-linear connection (Moghaddam, 2013). In the Banach–Stone-type theory of almost isometries, “almost Finsler manifolds” are effectively Finsler manifolds studied modulo almost isometry, with the forward quasi-metric and the cone $S\subset TM$7 playing the central role (Daniilidis et al., 2019).

2. Slit geometry, positivity, and generalized local models

For a partial Minkowski space $S\subset TM$8, the slit $S\subset TM$9 is conelike, meaning $F:TM-S\to(0,\infty)$0 for all $F:TM-S\to(0,\infty)$1. The associated bilinear form at $F:TM-S\to(0,\infty)$2,

$F:TM-S\to(0,\infty)$3

is the fiberwise Hessian of $F:TM-S\to(0,\infty)$4; positive definiteness of this form is exactly the additional condition defining an almost Minkowski space. The indicatrix is the hypersurface $F:TM-S\to(0,\infty)$5, and the solid indicatrix is $F:TM-S\to(0,\infty)$6. Because $F:TM-S\to(0,\infty)$7 is $F:TM-S\to(0,\infty)$8-homogeneous, the indicatrix determines the norm (Davis et al., 29 Aug 2025).

A partial Finsler manifold need not remain positive-definite on all of $F:TM-S\to(0,\infty)$9. To isolate the genuinely positive region, the extended slit is defined by

$1$0

Then $1$1 is an almost Finsler manifold, called the truncation of $1$2. This construction makes explicit that nonpositive eigenvalues of $1$3 are treated as excluded directions rather than admissible anisotropies in the almost category (Davis et al., 29 Aug 2025).

The conic pseudo-Finsler framework clarifies why this truncation is geometrically significant. For Minkowski conic pseudo-norms, the paper of Javaloyes–Sánchez develops a unit-ball/indicatrix characterization in terms of the closed unit ball $1$4 and indicatrix $1$5: strong convexity of the indicatrix characterizes conic Minkowski norms, convexity yields the triangle inequality, and strict convexity yields the strict triangle inequality. At the same time, conic and pseudo-Finsler generalizations exhibit genuine pathologies. For conic Finsler metrics, forward and backward balls are open, but they may fail to form a basis unless the metric is lower bounded; the induced separation $1$6 may be discontinuous or even vanish between distinct points. In pseudo-Finsler geometry, if the structure is not Finsler, then degeneracy of $1$7 must occur somewhere on every tangent space where positivity fails (Javaloyes et al., 2011).

This combination of slit-based exclusion and conic-domain restriction shows that “almost” in Finsler geometry is not merely terminological. It identifies a regime in which one keeps homogeneity and differential tensor calculus but relaxes full-domain availability or strong convexity, often at the price of subtler topology of balls, weaker distance behavior, or singular directions.

3. Bipartite spaces, Randers-type models, and characteristic tensors

A particularly explicit class of almost/partial Finsler structures is furnished by bipartite spaces. Let $1$8 be Riemannian, with

$1$9

and let $(T_xM,S\cap T_xM,F|_x)$0 be a symmetric nonnegative $(T_xM,S\cap T_xM,F|_x)$1-tensor whose eigenvalues lie in $(T_xM,S\cap T_xM,F|_x)$2. Writing

$(T_xM,S\cap T_xM,F|_x)$3

the bipartite norms are

$(T_xM,S\cap T_xM,F|_x)$4

The pairs $(T_xM,S\cap T_xM,F|_x)$5 and $(T_xM,S\cap T_xM,F|_x)$6 are partial Finsler manifolds; their indicatrices are called the lemon and the apple, respectively. The slit coincides with the space of fixed points under scaling in the set-theoretic sense used in the paper, and $(T_xM,S\cap T_xM,F|_x)$7 is a sphere of dimension $(T_xM,S\cap T_xM,F|_x)$8 (Davis et al., 29 Aug 2025).

Two especially important special cases are the $(T_xM,S\cap T_xM,F|_x)$9- and $M$0-spaces. Given a nonzero vector $M$1, decompose $M$2 into the components parallel and perpendicular to $M$3. The $M$4-spaces are defined by

$M$5

with slit $M$6; they are reversible, and on manifolds the corresponding slit is $M$7. The $M$8-spaces are defined by

$M$9

with slit $F:TM\setminus\{0\}\to(0,\infty)$0; $F:TM\setminus\{0\}\to(0,\infty)$1 is always an almost Minkowski norm, while $F:TM\setminus\{0\}\to(0,\infty)$2 is almost Minkowski only for $F:TM\setminus\{0\}\to(0,\infty)$3 (Davis et al., 29 Aug 2025).

The relation with Randers geometry is exact at the level of indicatrix unions. If $F:TM\setminus\{0\}\to(0,\infty)$4 and $F:TM\setminus\{0\}\to(0,\infty)$5, then $F:TM\setminus\{0\}\to(0,\infty)$6 is a Randers metric under the usual norm bound. For the $F:TM\setminus\{0\}\to(0,\infty)$7-spaces, when $F:TM\setminus\{0\}\to(0,\infty)$8 one has $F:TM\setminus\{0\}\to(0,\infty)$9, and when $1$0 one has $1$1. Hence the indicatrix union of the almost Finsler $1$2-manifolds equals the union of the two Randers indicatrices. By contrast, the $1$3-space indicatrix union is a spindle toroid; in dimension $1$4 it is a $1$5-dimensional spindle toroid formed by the apple and lemon glued along the fixed-point circle $1$6 (Davis et al., 29 Aug 2025).

The central tensorial contribution of this framework is the construction of characteristic $1$7-tensors generalizing the Matsumoto tensor. With

$1$8

and $1$9, the bipartite characteristic tensor is

$$g_{ij}(x,y)=\frac{1}{2}\frac{\partial^2(F^2)(x,y)}{\partial y^i\partial y^j}$$0

with

$$g_{ij}(x,y)=\frac{1}{2}\frac{\partial^2(F^2)(x,y)}{\partial y^i\partial y^j}$$1

For bipartite almost Finsler manifolds, $$g_{ij}(x,y)=\frac{1}{2}\frac{\partial^2(F^2)(x,y)}{\partial y^i\partial y^j}$$2. In the $$g_{ij}(x,y)=\frac{1}{2}\frac{\partial^2(F^2)(x,y)}{\partial y^i\partial y^j}$$3-space case there is a simpler vanishing tensor,

$$g_{ij}(x,y)=\frac{1}{2}\frac{\partial^2(F^2)(x,y)}{\partial y^i\partial y^j}$$4

and $$g_{ij}(x,y)=\frac{1}{2}\frac{\partial^2(F^2)(x,y)}{\partial y^i\partial y^j}$$5 on all $$g_{ij}(x,y)=\frac{1}{2}\frac{\partial^2(F^2)(x,y)}{\partial y^i\partial y^j}$$6-spaces. In the Randers and $$g_{ij}(x,y)=\frac{1}{2}\frac{\partial^2(F^2)(x,y)}{\partial y^i\partial y^j}$$7-space limit, $$g_{ij}(x,y)=\frac{1}{2}\frac{\partial^2(F^2)(x,y)}{\partial y^i\partial y^j}$$8, so $$g_{ij}(x,y)=\frac{1}{2}\frac{\partial^2(F^2)(x,y)}{\partial y^i\partial y^j}$$9 reduces to the Matsumoto tensor

$M$00

recovering the classical vanishing criterion for Randers geometry (Davis et al., 29 Aug 2025).

4. Almost isometries and the functional reconstruction viewpoint

A different use of “almost” keeps the underlying object a connected Finsler manifold $M$01 but weakens the morphisms. In the nonreversible case the fundamental metric object is the forward quasi-metric

$M$02

with backward distance $M$03 and symmetrized metric $M$04. The dual norm on cotangent spaces is

$M$05

A smooth function is forward semi-Lipschitz with constant $M$06 if

$M$07

and for $M$08 functions this is equivalent to the pointwise differential bound $M$09 (Daniilidis et al., 2019).

The relevant function space is the convex partially ordered cone

$M$10

together with its open subcone

$M$11

This cone simultaneously records differentiable information through $M$12 and quasi-metric information through $M$13. The fundamental global-local identity is

$M$14

Almost isometries are defined by preservation of the triangular function

$M$15

A bijection $M$16 is an almost isometry if it preserves $M$17, equivalently if there exists a potential $M$18 such that

$M$19

It is strict if it also satisfies a two-sided multiplicative control

$M$20

for some $M$21. In the smooth Finsler setting, every almost isometry arises from a diffeomorphism and a smooth potential by

$M$22

and strictness is equivalent to the associated potentials having semi-Lipschitz norm strictly less than $M$23 (Daniilidis et al., 2019).

The main Banach–Stone-type theorem states that if $M$24 and $M$25 are connected, second countable, bicomplete Finsler manifolds and

$M$26

is an isomorphism of convex partially ordered sets, then there exist a constant $M$27, a smooth $M$28 with $M$29, and a diffeomorphism $M$30 such that $M$31 is almost isometric to $M$32, $M$33 is an isometry between $M$34 and $M$35, and

$M$36

In compact connected Finsler manifolds every almost isometry is strict, while in the reversible case the potential must be constant, so the classification collapses to composition operators with global scaling. In this sense, almost Finsler manifolds understood up to almost isometry are classified by the order-convex structure of $M$37 (Daniilidis et al., 2019).

5. Even-dimensional pseudo-Finsler manifolds and local almost hypercomplex structures

In another precise tradition, an almost Finsler manifold is an even-dimensional pseudo-Finsler manifold $M$38. Here $M$39 is a smooth $M$40-manifold, $M$41 is an open submanifold of $M$42 satisfying $M$43 and $M$44, and $M$45 is smooth, positively $M$46-homogeneous, and has associated quadratic form of signature $M$47 with $M$48. The fundamental tensor is

$M$49

and the Cartan tensor is

$M$50

The geometry is organized by a non-linear connection $M$51, giving the decomposition

$M$52

with local adapted frame

$M$53

The Ehresmann curvature is encoded by the vertical part of $M$54 (Moghaddam, 2013).

Because $M$55, indices can be paired as $M$56. Using these pairs and the adapted frame, one constructs three local $M$57-tensor fields $M$58 by block rotations and horizontal-vertical swaps. They satisfy the quaternionic identities

$M$59

and therefore define an almost hypercomplex structure on each chart domain $M$60. The construction is local and depends on the chosen non-linear connection; integrability is not asserted (Moghaddam, 2013).

Starting from a Finsler connection $M$61, two linear connections $M$62 and $M$63 on $M$64 are defined by

$M$65

and satisfy $M$66. Their torsion decomposes naturally into horizontal and vertical parts. For fixed skew-symmetric Finsler tensor fields $M$67 and $M$68 of type $M$69, there exists on each chart a unique linear connection $M$70 on the vertical bundle that is metric-compatible, $M$71, and has prescribed torsion components

$M$72

These connections are determined by vertical and horizontal Koszul-type formulas, and they extend the $M$73-Cartan connections of Bejancu–Farran (Moghaddam, 2013).

A second theorem constructs a linear connection

$M$74

such that

$M$75

On each chart, this yields an almost hyper-Hermitian object adapted to the Finsler splitting. Even dimensionality is essential, since the index pairing $M$76 is the combinatorial basis of the construction (Moghaddam, 2013).

6. Related “almost” structures, rigidity, and limits of the terminology

The breadth of the term is further illustrated by almost rational metrics. Taha and Tiwari define an Almost Rational Finsler metric by requiring the fundamental tensor to factor as

$M$77

where $M$78 is smooth, $M$79 are symmetric, positive-definite, and rational in the fiber variables, and $M$80 is $M$81-homogeneous in $M$82. In this setting the quantity $M$83 controls rationality of the spray, $M$84-curvature, Douglas curvature, Landsberg curvature, mean Landsberg curvature, and Riemann/Weyl/X-curvatures. The main rigidity statements are that, under the paper’s hypotheses, isotropic $M$85-curvature forces $M$86, isotropic mean Landsberg curvature forces weakly Landsberg behavior, and Einstein almost rational metrics are Ricci-flat when $M$87 is not rational in $M$88. The same paper shows that Randers metrics cannot be AR-Finsler metrics, while generalized Kropina metrics, $M$89-th root metrics, extended $M$90-th root metrics, and their generalized Kropina changes furnish AR examples (Taha et al., 2021).

An additional and independent “almost” theme is almost Ricci solitons on Finsler measure spaces $M$91. These are defined by the soliton equation

$M$92

with soliton scalar $M$93. In the gradient case, the paper proves that $M$94 is a gradient almost Ricci soliton if and only if

$M$95

when $M$96 is compact. For Randers metrics $M$97, every almost Ricci soliton and every gradient almost Ricci soliton has isotropic $M$98-curvature, and navigation data $M$99 reduce the classification to Einstein or gradient almost Ricci data on the Riemannian side together with conformal or homothetic conditions on $S\subset TM$00. The paper also gives rigidity theorems for compact Randers Ricci solitons and constructs several Randers gradient Ricci solitons, described there as the first nontrivial examples of gradient Ricci solitons in Finsler geometry (Xia, 2024).

The coexistence of these uses shows that “almost Finsler manifold” has no single canonical meaning across the literature. This suggests that the expression should always be read relative to an explicit framework: slit-based partial/almost Finsler geometry, even-dimensional pseudo-Finsler geometry with almost hypercomplex structure, or equivalence classes under almost isometry. The technical assumptions are correspondingly framework-specific: bicompleteness and second countability in the Banach–Stone classification, conic-domain and positivity restrictions in conic/pseudo-Finsler geometry, and the extended slit $S\subset TM$01 in the recent partial/almost theory. The main open directions stated in these works include converse characterizations for the new characteristic tensors, extensions beyond smooth cones or beyond $S\subset TM$02 semi-Lipschitz spaces, and sharper understanding of incomplete or pathological slit geometries (Davis et al., 29 Aug 2025).