Infinite Pliability Across Disciplines

• Infinite Pliability is a multifaceted concept describing non‐rigidity and the capacity for multiple reversible configurations in systems ranging from elastic materials to geometrical and algebraic structures.
• It involves coordinated mechanisms such as metastable elastic minima, fold sequences, and control-theoretic openness that enable systems to access an exponentially large or continuous set of states.
• The concept underpins diverse fields, manifesting as measurable shape retention in metamaterials, NP-complete foldability in origami, finite versus infinite Mori fibre spaces in birational geometry, and treewidth-pliability in approximation theory.

Infinite pliability is not a single standardized concept. In the literature, it denotes several technically distinct notions of flexibility, non-rigidity, or multiplicity of attainable configurations. In elastic sheets and metamaterials it refers to the ability to retain many shapes through coupled metastable elastic minima without plastic deformation (Oppenheimer et al., 2015). In origami theory it can be interpreted as the existence of a sequence of legal global simple folds under the infinite all-layers model (Akitaya et al., 2019). In Carnot groups it is interpreted through pliability, strong pliability, and openness of multi-exponential maps (Juillet et al., 2016, Jean et al., 17 Jul 2025). In birational geometry it denotes infinitude of Mori fibre space structures up to square equivalence (Abban, 2013, Kurz et al., 29 Jul 2025). In approximation theory it appears as treewidth-pliability, a structural condition enabling PTASes for large classes of Max-CSPs (Romero et al., 2019). In orthorhombic flexible membranes it refers to an infinite set of flat phases connected by emergent continuous symmetry (Burmistrov et al., 2021).

1. Elastic shape memory and metastable mechanical shapeability

In elastic shapeable materials, infinite pliability is approached through many metastable elastic minima rather than through plastic flow. A concrete model is a 2D triangular lattice of $N$ nodes embedded in 3D, with programmed rest-length heterogeneity and optional edge or bending regularization. Two realizations are emphasized: a random lattice with spring rest lengths drawn from $L=a(1+0.1R)$ for uniform $R\in[-0.5,0.5]$, and a non-symmetric puckered lattice built from “hexagonal pyramids” with $a_l=1.15a$ and $a_m=1.05a$, with elongated links rotated by $60^\circ$ column by column to create frustration and coupling across hexagons. The elastic energy is written as $U=U_s+U_b$, with

$$U_s=\sum_s \frac12 k_s\big(|r_i-r_j|-\ell_{0,s}\big)^2,$$

and an optional bending penalty

$$U_b=C\sum_{\langle \mu,\nu\rangle}\frac{1}{1.1+\hat n_\mu\cdot \hat n_\nu},$$

with $C=ka^2/3000$ (Oppenheimer et al., 2015).

Shape retention is quantified against target cylindrical surfaces by the projection error

$L=a(1+0.1R)$0

The reported retention is strongest at moderate system size: for a random lattice with edge springs, the best retention occurs at $L=a(1+0.1R)$1 with $L=a(1+0.1R)$2, and across 15 random realizations shaped as half-cylinders $L=a(1+0.1R)$3 with $L=a(1+0.1R)$4; at $L=a(1+0.1R)$5 the error rises to $L=a(1+0.1R)$6. For the puckered lattice with three rest lengths and edge springs, the strongest retention occurs at $L=a(1+0.1R)$7 with $L=a(1+0.1R)$8, while at $L=a(1+0.1R)$9 it degrades to $R\in[-0.5,0.5]$0. By contrast, symmetric puckering or equal spring lengths lead to flattening on relaxation, with $R\in[-0.5,0.5]$1 (Oppenheimer et al., 2015).

The mechanism is not simple local bistability alone. An extensive fraction of nodes is bistable, with approximately $R\in[-0.5,0.5]$2 bistable apex nodes in the flat puckered state and a naive microstate count of $R\in[-0.5,0.5]$3. Yet bending the sheet changes which nodes are bistable, so local bistability is modulated by global shape and neighbor state. Cooperative transitions are diagnosed by the inverse participation ratio

$R\in[-0.5,0.5]$4

applied to dihedral-angle changes; just beyond threshold in a shaped $R\in[-0.5,0.5]$5 puckered sheet, $R\in[-0.5,0.5]$6, indicating a cooperative cluster of about 12 edges or sites rather than a single-site flip (Oppenheimer et al., 2015).

Elastic reversibility is substantial but bounded. A puckered $R\in[-0.5,0.5]$7 sheet shaped as an almost closed cylinder can undergo about $R\in[-0.5,0.5]$8 deformation and recover its stored shape upon release; even at about $R\in[-0.5,0.5]$9 imposed deformation, the sheet still returns, with reported residual strain $a_l=1.15a$0. Beyond the elastic limit, hysteresis loops show jumps between minima and irreversible steps. The same study emphasizes that shape memory decreases with lattice size because barriers between adjacent minima decrease, weak curvatures $a_l=1.15a$1 relax elastically, and effective thickness does not scale with $a_l=1.15a$2 in the simple model. Proposed countermeasures include edge reinforcement, adding $a_l=1.15a$3, using three rest lengths with rotated elongations, reducing apex height, and introducing hierarchical puckers so that effective bending thickness scales with sheet size (Oppenheimer et al., 2015).

This usage of infinite pliability is therefore asymptotic and conditional rather than literal. The model exhibits exponential multiplicity of elastic microstates and reversible access to many remembered shapes, but macroscopic distinctness is constrained by geometric compatibility, barrier heights, and Gauss’s theorem. The paper explicitly states that near-infinite pliability is approachable only if multiscale geometric frustration, dense coupled bistabilities, and folding-capable elements are designed into the material (Oppenheimer et al., 2015).

2. Foldability, semidiscrete surfaces, and infinite frameworks

In origami theory, infinite pliability acquires an algorithmic and combinatorial meaning. Under the infinite all-layers model, each simple fold is along an infinite line and must fold all layers intersecting that line. In this setting, infinite pliability of a crease pattern can be interpreted as the existence of a sequence of legal global simple folds reducing the pattern to zero creases, respecting non-self-penetration and non-stretching constraints; the paper identifies this with the decision problem of infinite all-layers simple foldability. The main positive results are a deterministic $a_l=1.15a$4-time algorithm for 1D crease patterns and a linear-time decision procedure for unassigned axis-aligned orthogonal crease patterns on axis-aligned orthogonal paper. The negative result is that simple foldability is strongly NP-complete when a subset of creases has a mountain-valley assignment, even for an axis-aligned rectangle with square-grid creases (Akitaya et al., 2019).

For semidiscrete surfaces, by contrast, infinite pliability is largely ruled out in the generic multi-ribbon regime. A semidiscrete $a_l=1.15a$5-ribbon surface is a piecewise ruled surface built from curves $a_l=1.15a$6, with discrete edges $a_l=1.15a$7. The paper proves that generic 2-ribbon semidiscrete surfaces have exactly one degree of infinitesimal flexibility and one degree of finite flexibility, while every generic $a_l=1.15a$8-ribbon surface with $a_l=1.15a$9 has at most one degree of finite or infinitesimal flexibility. It also gives a necessary compatibility condition for infinitesimal flexibility of 3-ribbon surfaces and shows that flexibility of general $a_m=1.05a$0-ribbon surfaces reduces to flexibility of their 3-ribbon subsurfaces. The explicit exception noted in the paper is the 1-ribbon case, which can have infinitely many degrees of flexibility; this lies outside the generic $a_m=1.05a$1 regime (Karpenkov, 2010).

Infinite bar-joint frameworks introduce a third meaning. Here the key distinction is between infinitesimal rigidity in $a_m=1.05a$2 and genuine deformability of an infinite structure. The infinite rigidity matrix $a_m=1.05a$3 defines infinitesimal flexes by

$a_m=1.05a$4

for each edge $a_m=1.05a$5, and the paper studies operator-theoretic forms of rigidity such as square-summably infinitesimal rigidity. For crystal frameworks, $a_m=1.05a$6 becomes, after Fourier transform, multiplication by a matrix-valued symbol $a_m=1.05a$7 on the torus. The rigid unit mode set is

$a_m=1.05a$8

and in the Maxwell case it is the zero set of $a_m=1.05a$9. The paper proves that several generic crystal frameworks are square-summably infinitesimally rigid and yet smoothly deformable in infinitely many ways. The kagome framework is a central example: it is $60^\circ$0-isostatic, but there are uncountably many flow-periodic flexes. In this setting, infinite pliability means that rich smooth deformation families can coexist with vanishing $60^\circ$1 infinitesimal flexibility because the relevant motions are non-$60^\circ$2, wave-like, or affine-cell deformations (Owen et al., 2010).

Taken together, these three lines of work show that “infinite” can refer to fold-sequence existence, to infinite-dimensional deformation space, or to the failure of finite-rank rigidity diagnostics to capture large-scale nonlocal motions. The phrase does not encode a uniform notion of softness.

3. Carnot groups, strong pliability, and directional openness

In Carnot groups, pliability is a first-order control-theoretic and geometric property of horizontal directions. Let $60^\circ$3 be a Carnot group with stratified Lie algebra

$60^\circ$4

A horizontal vector $60^\circ$5 is pliable if the endpoint map for horizontal curves is open at the constant control $60^\circ$6, equivalently if the straight horizontal curve $60^\circ$7 can be perturbed in $60^\circ$8 so that its endpoint data fill a neighborhood. The paper characterizing the Whitney extension theorem for curves proves that the pair $60^\circ$9 has the $U=U_s+U_b$0 extension property if and only if $U=U_s+U_b$1 is pliable. It also proves that every step-2 Carnot group is pliable and constructs pliable Carnot groups of arbitrarily large step, so pliability is not confined to small-step geometry (Juillet et al., 2016).

A later note compares several non-rigidity notions and proves a precise equivalence theorem. For $U=U_s+U_b$2, the following are equivalent: pliability (P), strong pliability (SP), openness of a multi-exponential map $U=U_s+U_b$3 at $U=U_s+U_b$4, called the (H)-condition, and the strong (H)-condition. In symbols,

$U=U_s+U_b$5

Moreover, regularity of the endpoint map and the submersive (H)-condition are equivalent and imply all four properties. The same note explains two natural interpretations of “infinite pliability.” Under Interpretation A, it means scale-robust non-rigidity, which is exactly strong pliability: for every $U=U_s+U_b$6 one can find a control perturbation $U=U_s+U_b$7 with $U=U_s+U_b$8, the same endpoint, and surjective differential. Under Interpretation B, it means openness of multi-exponential maps under arbitrarily long concatenations; if $U=U_s+U_b$9 is open at $$U_s=\sum_s \frac12 k_s\big(|r_i-r_j|-\ell_{0,s}\big)^2,$$0 for some $$U_s=\sum_s \frac12 k_s\big(|r_i-r_j|-\ell_{0,s}\big)^2,$$1, then $$U_s=\sum_s \frac12 k_s\big(|r_i-r_j|-\ell_{0,s}\big)^2,$$2 is open at $$U_s=\sum_s \frac12 k_s\big(|r_i-r_j|-\ell_{0,s}\big)^2,$$3 for every $$U_s=\sum_s \frac12 k_s\big(|r_i-r_j|-\ell_{0,s}\big)^2,$$4 (Jean et al., 17 Jul 2025).

Directional versions weaken the requirement from all of $$U_s=\sum_s \frac12 k_s\big(|r_i-r_j|-\ell_{0,s}\big)^2,$$5 to a subset of directions. The directional Whitney extension theorem states that if every vector in a subset $$U_s=\sum_s \frac12 k_s\big(|r_i-r_j|-\ell_{0,s}\big)^2,$$6 is pliable, then $$U_s=\sum_s \frac12 k_s\big(|r_i-r_j|-\ell_{0,s}\big)^2,$$7 Whitney extension holds for compact data with tangent field taking values in $$U_s=\sum_s \frac12 k_s\big(|r_i-r_j|-\ell_{0,s}\big)^2,$$8. The Engel group provides the decisive example. Its horizontal layer is spanned by $$U_s=\sum_s \frac12 k_s\big(|r_i-r_j|-\ell_{0,s}\big)^2,$$9, with

$$U_b=C\sum_{\langle \mu,\nu\rangle}\frac{1}{1.1+\hat n_\mu\cdot \hat n_\nu},$$0

For $$U_b=C\sum_{\langle \mu,\nu\rangle}\frac{1}{1.1+\hat n_\mu\cdot \hat n_\nu},$$1, the paper proves that $$U_b=C\sum_{\langle \mu,\nu\rangle}\frac{1}{1.1+\hat n_\mu\cdot \hat n_\nu},$$2 is pliable if and only if $$U_b=C\sum_{\langle \mu,\nu\rangle}\frac{1}{1.1+\hat n_\mu\cdot \hat n_\nu},$$3 or $$U_b=C\sum_{\langle \mu,\nu\rangle}\frac{1}{1.1+\hat n_\mu\cdot \hat n_\nu},$$4; the only non-pliable directions are nonzero multiples of $$U_b=C\sum_{\langle \mu,\nu\rangle}\frac{1}{1.1+\hat n_\mu\cdot \hat n_\nu},$$5. Consequently, Whitney extension holds on $$U_b=C\sum_{\langle \mu,\nu\rangle}\frac{1}{1.1+\hat n_\mu\cdot \hat n_\nu},$$6. The same paper then proves that every horizontal curve in the Engel group intersects a $$U_b=C\sum_{\langle \mu,\nu\rangle}\frac{1}{1.1+\hat n_\mu\cdot \hat n_\nu},$$7 horizontal curve on a set of positive measure (Speight et al., 20 May 2025).

This body of work fixes a recurring ambiguity. In sub-Riemannian analysis, infinite pliability is not a cardinality of shapes or models. It is an openness and accessibility property, and its strongest verified forms are either equivalence with multi-exponential openness or existence of pliable groups of unbounded step.

4. Birational geometry and infinite pliability of Mori fibre spaces

In birational geometry, pliability is a property of Mori fibre space structures up to square birational equivalence. If $$U_b=C\sum_{\langle \mu,\nu\rangle}\frac{1}{1.1+\hat n_\mu\cdot \hat n_\nu},$$8 is a Mori fibre space, its pliability is

$$U_b=C\sum_{\langle \mu,\nu\rangle}\frac{1}{1.1+\hat n_\mu\cdot \hat n_\nu},$$9

where $C=ka^2/3000$0 denotes square birationality. Infinite pliability means that this set has infinite cardinality. The notion is classical in birational rigidity theory, where rational varieties such as $C=ka^2/3000$1 have infinite pliability, while rigid varieties have pliability $C=ka^2/3000$2 (Abban, 2013).

Low-rank Cox ring and VGIT techniques show that infinite pliability does not arise automatically even when several Sarkisov links exist. For a degree-4 del Pezzo fibration $C=ka^2/3000$3 constructed as a smooth complete intersection

$C=ka^2/3000$4

with

$C=ka^2/3000$5

the paper proves that $C=ka^2/3000$6 and that $C=ka^2/3000$7 is not rational, with $C=ka^2/3000$8. Nevertheless, the same paper explicitly states that infinite pliability is not expected in this class: the relevant Cox rings have finite rank, VGIT chamber decompositions are finite, and the Picard rank is too small to generate infinitely many non-square-equivalent models (Abban, 2013).

An equivariant analogue for $C=ka^2/3000$9 reaches the same conclusion. For finite subgroups $L=a(1+0.1R)$00, the paper classifies the cases in which $L=a(1+0.1R)$01 is not $L=a(1+0.1R)$02-birational to conic bundles or del Pezzo fibrations and describes all $L=a(1+0.1R)$03-Mori fibre spaces $L=a(1+0.1R)$04-birational to $L=a(1+0.1R)$05. In all cases analyzed, the equivariant pliability $L=a(1+0.1R)$06 is finite. For many solid cases, $L=a(1+0.1R)$07. For the exceptional group $L=a(1+0.1R)$08, the paper proves that $L=a(1+0.1R)$09 consists exactly of three models: $L=a(1+0.1R)$10, the terminal Fano-Enriques threefold $L=a(1+0.1R)$11, and a del Pezzo fibration $L=a(1+0.1R)$12 (Cheltsov et al., 2022).

The sharp positive result is now known for surfaces. A paper on sextic del Pezzo surfaces proves that degree 6 del Pezzo surfaces are the only solid surfaces that admit infinite pliability. More precisely, if $L=a(1+0.1R)$13 is a birationally solid del Pezzo surface over a perfect field and $L=a(1+0.1R)$14, then $L=a(1+0.1R)$15 is a sextic del Pezzo surface and $L=a(1+0.1R)$16. Conversely, there exist a perfect field $L=a(1+0.1R)$17 and a solid sextic del Pezzo surface $L=a(1+0.1R)$18 over $L=a(1+0.1R)$19 such that $L=a(1+0.1R)$20. The proof uses the hexagon of six $L=a(1+0.1R)$21-curves on the geometric model

$L=a(1+0.1R)$22

a detailed classification of degree-2 and degree-3 closed points in general position, and explicit Sarkisov links of type II centered at such points. These links produce infinitely many pairwise non-isomorphic minimal del Pezzo models up to square equivalence (Kurz et al., 29 Jul 2025).

Within birational geometry, then, infinite pliability is literal. It counts infinitely many Mori fibre space structures. The recent sextic theorem also shows that this phenomenon is highly exceptional rather than generic.

5. Treewidth-pliability and approximation of Max-CSPs

In approximation theory, pliability is a structural approximation property rather than a flexibility property. For a graph parameter $L=a(1+0.1R)$23, a class $L=a(1+0.1R)$24 of structures is $L=a(1+0.1R)$25-pliable if for every $L=a(1+0.1R)$26 there exists $L=a(1+0.1R)$27 such that every $L=a(1+0.1R)$28 is within opt-distance at most $L=a(1+0.1R)$29 of some $L=a(1+0.1R)$30 with $L=a(1+0.1R)$31. The opt-distance is defined by

$L=a(1+0.1R)$32

and equivalently through overcasts, where $L=a(1+0.1R)$33 means $L=a(1+0.1R)$34 for all $L=a(1+0.1R)$35 (Romero et al., 2019).

The central result is that treewidth-pliability implies a PTAS for the optimal value of maximum homomorphism problems, hence for large classes of Max-2-CSPs. If $L=a(1+0.1R)$36 is a tw-pliable class of bounded-arity structures, then the level-$L=a(1+0.1R)$37 Sherali–Adams relaxation yields

$L=a(1+0.1R)$38

with $L=a(1+0.1R)$39. The same framework unifies two previously separate PTAS paradigms. On the sparse side, fractional-treewidth-fragility implies tw-pliability for Gaifman-restricted classes. On the dense side, dense undirected graph classes with $L=a(1+0.1R)$40 edges are tw-pliable via regularity partitions and constant-size quotients (Romero et al., 2019).

The paper also proves several equivalences: $L=a(1+0.1R)$41 and, for bounded signatures, these are also equivalent to size-pliability and cc-pliability. It further shows that for a monotone graph class, hyperfiniteness is equivalent to being fractionally-treewidth-fragile and of bounded degree (Romero et al., 2019).

This use of the term reverses the usual intuition. Pliability here does not mean having more solutions or being easier to solve exactly. The paper explicitly proves that treewidth-pliability guarantees approximation of the optimal value only, not recovery of a near-optimal assignment in general. It also states that “infinite pliability” is not a defined numerical quantity in this framework; the mathematically faithful interpretation is non-pliability, meaning that for some $L=a(1+0.1R)$42 no finite $L=a(1+0.1R)$43 exists. Tournaments and orientations of graphs with unbounded average degree are given as such non-pliable classes (Romero et al., 2019).

6. Continuum families of anisotropic flat phases and general limitations

A distinct continuum interpretation appears in flexible two-dimensional materials with orthorhombic symmetry. For a membrane with out-of-plane field $L=a(1+0.1R)$44, in-plane displacement $L=a(1+0.1R)$45, nonlinear strain

$L=a(1+0.1R)$46

and orthorhombic elastic constants $L=a(1+0.1R)$47, the long-wavelength RG flow produces a stable line of fixed points rather than an isolated one. The running parameters are

$L=a(1+0.1R)$48

and the RG equations imply $L=a(1+0.1R)$49, $L=a(1+0.1R)$50, and $L=a(1+0.1R)$51. Thus $L=a(1+0.1R)$52 while $L=a(1+0.1R)$53 labels a stable line of flat phases. Each phase is mechanically distinct yet governed by the same universal anomalous elasticity exponent $L=a(1+0.1R)$54, with

$L=a(1+0.1R)$55

and $L=a(1+0.1R)$56 in $L=a(1+0.1R)$57. The same paper states that these anisotropic flat phases are uniquely labeled by the ratio of absolute Poisson’s ratios,

$L=a(1+0.1R)$58

Because $L=a(1+0.1R)$59 varies continuously, the system possesses an infinite set of flat phases connected by emergent continuous symmetry (Burmistrov et al., 2021).

This is one of the few settings where “infinite” means a genuine continuum of phases rather than a discrete infinity of models or an exponential count of microstates. The paper estimates a Ginzburg length $L=a(1+0.1R)$60 and applies the theory to phosphorene, with $L=a(1+0.1R)$61, $L=a(1+0.1R)$62, $L=a(1+0.1R)$63, $L=a(1+0.1R)$64, $L=a(1+0.1R)$65, and $L=a(1+0.1R)$66, giving $L=a(1+0.1R)$67 and thus $L=a(1+0.1R)$68 (Burmistrov et al., 2021).

Across the broader literature, however, several recurring limits remain. In elastic spring lattices, shape memory degrades with system size and compound curvature is constrained by geometric compatibility (Oppenheimer et al., 2015). In generic semidiscrete $L=a(1+0.1R)$69-ribbon surfaces with $L=a(1+0.1R)$70, the degree of flexibility is sharply bounded and infinite pliability does not occur (Karpenkov, 2010). In origami, global-fold existence becomes strongly NP-complete once partial mountain-valley data are imposed (Akitaya et al., 2019). In Carnot groups, global pliability can fail even when directional pliability holds on a full-measure subset, as in the Engel group (Speight et al., 20 May 2025). In birational geometry, infinite pliability is exceptional enough that among solid surfaces it characterizes sextic del Pezzo surfaces (Kurz et al., 29 Jul 2025).

The term therefore has family resemblance rather than uniformity. What persists across fields is the idea of non-rigidity beyond a single prescribed configuration; what changes is the object being counted or opened up—elastic minima, fold sequences, admissible control directions, Mori fibre structures, bounded-treewidth approximants, or RG fixed points.