Thurston's Jiggling Lemma

• Thurston's Jiggling Lemma is a transversality theorem that adjusts smooth triangulations to achieve general position with respect to a given rank-k distribution.
• It employs a two-step procedure—subdivision to refine simplices and C¹-small vertex perturbations—to systematically enforce transversality on all faces.
• The lemma underpins flexible h-principle methods by extending to piecewise-smooth solutions of first-order differential relations without extra homotopical data.

Searching arXiv for recent and relevant papers on Thurston's jiggling lemma and related h-principle work. Thurston’s jiggling lemma is a transversality statement for triangulated smooth manifolds: given a smooth manifold $M$ with a smooth rank-$k$ distribution $\xi \subset TM$, any smooth triangulation can be modified by a two-step procedure—subdivision followed by a $C^1$-small perturbation—so that every simplex and every face is in general position with respect to $\xi$. In the modern formulation, the perturbation can be made arbitrarily small, localized near a prescribed compact set, and relative to a subcomplex where general position already holds (Fokma et al., 23 Jan 2025, Fokma, 26 Aug 2025). The lemma originated in Thurston’s work on foliations and has subsequently been understood both as a combinatorial transversality device and, in generalized form, as a route to an h-principle for piecewise-smooth solutions of open, fiberwise dense first-order differential relations without homotopical assumptions (Fokma et al., 23 Jan 2025).

1. Formal statement and basic setup

In the standard setting, $M$ is a smooth $m$-manifold, $\xi \subset TM$ is a smooth rank-$k$ distribution, and $T\colon |K| \to M$ is a smooth triangulation. A triangulation here means a homeomorphism from the geometric realization of a simplicial complex onto $k$0, with each simplex smoothly embedded (Fokma et al., 23 Jan 2025). A distribution is a smooth subbundle of constant rank $k$1, equivalently a smoothly varying $k$2-plane field in local charts (Fokma, 26 Aug 2025).

Thurston’s jiggling lemma asserts that for every compact $k$3 and every $k$4, there exists a subdivision $k$5 of $k$6 and a $k$7-small homeomorphism $k$8 with $k$9, supported in an arbitrarily small neighborhood of $\xi \subset TM$0, such that the new triangulation $\xi \subset TM$1 is in general position with respect to $\xi \subset TM$2 over $\xi \subset TM$3 (Fokma et al., 23 Jan 2025). In a finite-complex formulation, if $\xi \subset TM$4 is a finite simplicial complex, $\xi \subset TM$5 is a piecewise-smooth embedding into a manifold $\xi \subset TM$6 with distribution $\xi \subset TM$7, and $\xi \subset TM$8 is a subcomplex already in general position, then for any $\xi \subset TM$9 there exists a $C^1$0-jiggling $C^1$1 of $C^1$2 such that $C^1$3 is in general position with respect to $C^1$4 and $C^1$5 (Fokma, 26 Aug 2025).

The perturbation is measured in the $C^1$6-norm. In the source exposition, the condition $C^1$7 is emphasized as controlling both the positional displacement and the tangent map on each simplex (Fokma et al., 23 Jan 2025). This smallness is essential: the lemma is not merely existential transversality but a controlled transversality statement compatible with relative and parametric constructions.

2. General position and the meaning of “jiggling”

The operative notion is stronger than ordinary top-dimensional transversality. For a piecewise-smooth map transverse to a rank-$C^1$8 distribution, being in general position means that transversality holds on every top-simplex and that for every simplex of any dimension, the image meets the leaves of the distribution transversely in the usual sense of lower-dimensional immersions (Fokma et al., 23 Jan 2025). A more explicit formulation given in the later treatment states that a simplex or map is in general position if it is transverse not only in top dimension but also on all faces, and moreover at every point $C^1$9 one has transversality to the constant foliation defined by $\xi$0; equivalently, for every face $\xi$1 and every $\xi$2,

$\xi$3

(Fokma, 26 Aug 2025)

This stronger requirement explains why “jiggling” is more than a generic perturbation argument. The procedure has two explicit stages. First, one subdivides to make simplices small in diameter. Second, one perturbs the images of vertices by small amounts in local charts, thereby tilting simplices and their faces (Fokma et al., 23 Jan 2025). In the later terminology, an $\xi$4-jiggling of $\xi$5 is a pair $\xi$6 where $\xi$7 is a subdivision of $\xi$8 and $\xi$9, measured simplex-by-simplex (Fokma, 26 Aug 2025).

Barycentric subdivision appears as a canonical device in the original-style presentation. Given a simplicial complex $M$0, the barycentric subdivision $M$1 is formed by placing a vertex at the barycenter of each simplex and coning off the subdivisions of the boundary; iteration yields $M$2 (Fokma et al., 23 Jan 2025). The later conceptual proof instead emphasizes crystalline subdivision, in which an ordered simplex is embedded in the unit cube, the cube is partitioned into $M$3 cubes of side $M$4, and each cube is coned into simplices; after $M$5 steps, every simplex has diameter $M$6, while maximal edge length, minimal height, and local combinatorial complexity are quantitatively controlled (Fokma, 26 Aug 2025).

A common misconception is that jiggling is simply a synonym for triangulating in generic position. The cited formulations make clear that the target condition is not only genericity of top-dimensional simplices but uniform control of all faces against a varying distribution, together with relative support and $M$7-smallness (Fokma et al., 23 Jan 2025, Fokma, 26 Aug 2025).

3. Structure of the classical proof

The classical proof proceeds by localization and induction on skeleta. One first covers $M$8 by finitely many coordinate charts in which $M$9 is $m$0-close to a constant $m$1-plane field. Once simplices are sufficiently small, they appear almost linear inside these charts, and the distribution is almost constant over each simplex (Fokma et al., 23 Jan 2025). This reduction permits a finite-dimensional linear transversality argument to be applied simplex by simplex.

The perturbation is then organized dimensionally. The original sketch orders simplices by dimension $m$2: first vertices, then edges relative to the moved vertices, then $m$3-simplices relative to the $m$4-skeleton, and so on up to top dimension (Fokma et al., 23 Jan 2025). At the $m$5-th stage, all simplices of lower dimension have already been arranged in general position. One perturbs barycenters or vertices of $m$6-simplices by tiny vectors chosen so that each $m$7-simplex becomes transverse to the approximating constant plane field, while the perturbation remains too small to destroy transversality already achieved on lower skeleta (Fokma et al., 23 Jan 2025).

Quantitative separation of scales is central. The exposition specifies that one chooses the size of the vertex moves at stage $m$8 smaller than a preassigned $m$9, with $\xi \subset TM$0, so that each stage preserves previous transversality margins (Fokma et al., 23 Jan 2025). This is the combinatorial analog of the standard transversality principle that openness allows previously arranged transverse conditions to survive sufficiently small later perturbations.

The later conceptual proof recasts this mechanism in a more uniform Euclidean framework. After linearizing the map on a refined subdivision, one fixes a small $\xi \subset TM$1 so that any $\xi \subset TM$2-perturbation of vertices remains an embedding. A uniform estimate then shows that any set of $\xi \subset TM$3 transverse simplices through a point can be made $\xi \subset TM$4-semitransverse simultaneously with $\xi \subset TM$5 independent of the subdivision depth, once diameters are sufficiently small (Fokma, 26 Aug 2025). This yields a vertex-by-vertex induction: given perturbed images of earlier vertices, one chooses the next perturbed vertex in a small ball so that every face containing it is $\xi \subset TM$6-semitransverse; after all vertices are moved, full general position follows because semitransversality on all faces, combined with small simplex size, promotes transversality to the actual varying distribution (Fokma, 26 Aug 2025).

4. Conceptual reformulation: crystalline subdivision and projection criteria

A major development in the 2025 treatment is a more conceptual proof based on crystalline subdivision and explicit projection criteria (Fokma, 26 Aug 2025). This proof separates geometric control of simplex shape from the transversality condition itself.

The key linear-algebraic device is the projection criterion. For a simplex $\xi \subset TM$7 in $\xi \subset TM$8 and a constant foliation $\xi \subset TM$9, one has

$k$0

This criterion turns transversality into a non-incidence condition in the quotient by the distribution (Fokma, 26 Aug 2025). On this basis, the proof introduces $k$1-semitransversality at a chosen vertex—meaning that perturbing that vertex within $k$2 preserves transversality to a fixed constant foliation—and $k$3-transversality, meaning that transversality survives any simultaneous $k$4-perturbation of all vertices (Fokma, 26 Aug 2025).

The significance of crystalline subdivision is that it supplies uniformly nondegenerate simplices: diameter, edge length, and minimal height all scale in a controlled manner under refinement (Fokma, 26 Aug 2025). This avoids pathological flattening that could destroy stability of transversality under small perturbations. If every face of a simplex is $k$5-semitransverse in one chosen vertex, then shape control implies an $k$6-ball in the Grassmannian around each face’s tangent plane, and hence uniform $k$7-transversality of the simplex (Fokma, 26 Aug 2025). The argument thereby replaces ad hoc perturbation choices by a systematic stability theory.

This suggests a broader interpretation of jiggling as a controlled discretization method for transversality: shape-regular subdivision and vertex perturbation together create a finite combinatorial framework in which openness of transversality can be quantified. That interpretation is not explicitly stated in the cited texts, but it is a plausible implication of the role played by projection criteria, simplex shape estimates, and uniform perturbation radii.

5. Relative, parametric, and manifold versions

The lemma has relative and parametric forms. In the relative version, if a subcomplex $k$8 is already in general position, the perturbation can be supported away from $k$9, leaving it fixed (Fokma et al., 23 Jan 2025). In the finite-complex formulation, this appears as the condition $T\colon |K| \to M$0 for a prescribed subcomplex $T\colon |K| \to M$1 (Fokma, 26 Aug 2025). Relative control is fundamental for patching local modifications and for inductive constructions over charts.

The parametric version treats continuously varying triangulations $T\colon |K| \to M$2 and distributions $T\colon |K| \to M$3 over a compact parameter space $T\colon |K| \to M$4. The same subdivision and vertex-move construction is carried out uniformly in $T\colon |K| \to M$5, with a uniform $T\colon |K| \to M$6-bound ensuring that the resulting family remains in general position for all parameters (Fokma et al., 23 Jan 2025). This form is what connects the lemma to h-principle arguments, where families rather than isolated solutions are typically decisive.

For arbitrary smooth manifolds, the later treatment emphasizes that the Euclidean argument must be globalized carefully, since general position is not preserved under naive subdivisions (Fokma, 26 Aug 2025). The method is chart-by-chart. One covers the image by finitely many coordinate charts $T\colon |K| \to M$7, then inductively jiggles over each chart while avoiding the already-good region (Fokma, 26 Aug 2025). A relative Euclidean lemma handles the local perturbations.

A further refinement is “jiggling the subdivision itself.” One perturbs the identity map $T\colon |K| \to M$8 to a piecewise-linear triangulation $T\colon |K| \to M$9 so that $k$00 is compatible with the distributions appearing in the charts, using barycentric subdivision and careful interpolation so that certain skeleta remain fixed (Fokma, 26 Aug 2025). Alternating chart-jiggling and subdivision-jiggling in finitely many steps produces a global piecewise-linear homeomorphism that carries the original triangulation into one in global general position and fixes the prescribed initial subcomplex (Fokma, 26 Aug 2025).

6. Generalization to differential relations and the “h-principle without homotopical assumptions”

The most substantial recent extension is from triangulations transverse to distributions to piecewise-smooth solutions of first-order differential relations (Fokma et al., 23 Jan 2025). Let $k$01 be a smooth fiber bundle and $k$02 its first-jet bundle. A first-order differential relation is a subset

$k$03

A formal solution is a section $k$04 with $k$05, while a holonomic solution is a genuine section $k$06 whose jet $k$07 lies in $k$08 (Fokma et al., 23 Jan 2025). The relation is called open if it is open in the standard topology on $k$09, and fiberwise dense if for each $k$10 and each $k$11, the fiber $k$12 is dense in the affine space $k$13 (Fokma et al., 23 Jan 2025).

Under these hypotheses, one has a generalized jiggling theorem: for any smooth triangulation $k$14, any $k$15, and any continuous or piecewise-smooth section $k$16, and for any subcomplex $k$17 on which $k$18 already satisfies $k$19, there exists a subdivision $k$20 and a piecewise-smooth section $k$21 with $k$22, such that $k$23 and $k$24 for all $k$25 (Fokma et al., 23 Jan 2025). In simplicial-set language, the exposition states a weak homotopy equivalence $k$26 (Fokma et al., 23 Jan 2025).

The proof mirrors the geometric jiggling scheme. One first subdivides so that the bundle and relation are almost linear on each simplex. One then linearizes a section on each top-simplex by affine approximation at the vertices, changing the $k$27-norm by $k$28. Using fiberwise density, one perturbs the $k$29-jets at vertices so that they land in $k$30, and extends by the unique affine map on each simplex. A coloring argument then processes top-simplices by finitely many colors, tilting all simplices of one color by a small $k$31-move, then repeating with successively smaller moves $k$32, preserving previously imposed conditions. Final boundary interpolation yields a global piecewise-smooth section $k$33 with

$k$34

(Fokma et al., 23 Jan 2025)

The phrase “without homotopical assumptions” refers to the fact that the resulting solution need only be piecewise-smooth, so its jet $k$35 may be discontinuous across codimension-one walls. Consequently, there need not be an underlying global formal solution class in $k$36; no extra formal data enters the theorem (Fokma et al., 23 Jan 2025). This sharply distinguishes the result from standard h-principle statements, which usually begin with a formal solution and then seek a homotopy to a holonomic one.

7. Applications, examples, and mathematical significance

A standard illustrative example is $k$37 with the horizontal distribution $k$38. Starting from an arbitrary triangulation, subdivision followed by vertical perturbation of vertices yields a triangulation transverse to the horizontal foliation, so in particular every edge becomes non-horizontal (Fokma, 26 Aug 2025). This example captures the essence of the lemma: arbitrarily fine combinatorial refinement converts a global geometric positioning problem into local slope constraints on simplices.

Historically, the lemma is presented as a key step in Thurston’s work on foliations. The 2025 account states that it underlies the approximation of smooth foliations by ones carried by triangulations in general position and is crucial in proving h-principle-type flexibility for foliations (Fokma, 26 Aug 2025). The generalized theorem of 2025 extends this logic from foliated geometry to arbitrary open, fiberwise dense first-order conditions in jet bundles (Fokma et al., 23 Jan 2025).

The later paper also lists corollaries and neighboring domains: Whitehead-type triangulations adapted to a given plane field, symplectic triangulations in the spirit of $k$39, and foundational transversality results in contact and Engel geometry (Fokma, 26 Aug 2025). The exact formulations of those corollaries are not reproduced in the supplied material, but their mention indicates the breadth of contexts in which jiggling functions as a combinatorial transversality mechanism.

Taken together, the two 2025 treatments present Thurston’s jiggling lemma in two complementary roles. In one role, it is a theorem about triangulations and distributions, proved by subdivision and vertex perturbation, with strong relative and parametric control (Fokma, 26 Aug 2025). In the other, it becomes the prototype for a flexible existence theorem for piecewise-smooth solutions of open, fiberwise dense first-order relations, yielding an h-principle that dispenses with prior formal homotopical data (Fokma et al., 23 Jan 2025). The enduring significance of the lemma lies precisely in this dual character: it is at once a geometric perturbation method and a structural bridge between combinatorial topology, transversality theory, and h-principle phenomena.