Pith. sign in

REVIEW 2 major objections 4 minor 38 references

Higher-order energy-landscape singularities organize path-transition graphs that fully describe multistable materials under multi-parameter driving, and reduce to ordinary transition graphs when driving is restricted to one dimension.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-14 06:27 UTC pith:UYRB5BG2

load-bearing objection Solid singularity analysis that turns the authors’ own pt-graphs into a usable design language and cleanly recovers t-graphs as the 1-D limit. the 2 major comments →

arxiv 2607.11179 v1 pith:UYRB5BG2 submitted 2026-07-13 cond-mat.soft

Vectorial driving of multistable materials: singularities, pt-graphs, and non-generic paths

Colin M. Meulblok , Martin van Hecke This is my paper
classification cond-mat.soft
keywords multistable materialspath-transition graphsvectorial drivingfold bifurcationscusp singularitiesbutterfly singularitymetamaterialshysteresis
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Multistable materials change state when driven by several control parameters at once, and the outcome depends on the path taken through that multi-dimensional control space. The paper shows that these path-dependent responses are completely captured by path-transition graphs whose building blocks are ancestor-descendant-sibling triplets generated by ordinary fold bifurcations. Higher-order singularities—cusps, pitchforks and a codimension-four butterfly—organize those folds into distinct strain maps, thereby fixing the topology of the graphs. Experiments on silicone metamaterial chains and a matching spring model confirm the construction, including the spontaneous symmetry breaking that occurs along non-generic paths through the higher singularities. Restricting the same graphs to one-dimensional driving paths recovers the familiar transition graphs of scalar driving, unifying both regimes inside a single singularity-based framework. The result supplies a concrete design language for metamaterials whose sequential behavior can be programmed for sensing, soft robotics and in-material computation.

Core claim

Higher-order singularities of the energy landscape (cusps that produce pitchforks, and a codimension-four butterfly that separates strong- and weak-coupling regimes) organize the fold curves that generate ADS triplets; those triplets in turn determine the strain maps and path-transition graphs of multistable systems under vectorial driving, while ordinary transition graphs arise exactly as the one-dimensional restriction of the same path-transition graphs.

What carries the argument

Path-transition graphs (pt-graphs): mixed graphs whose nodes are equivalence classes of configurations that can be deformed into one another without crossing a fold, and whose edges (reversible undirected or irreversible directed) are labelled by the fold curves of the strain map; each fold curve contributes one ancestor-descendant-sibling triplet.

Load-bearing premise

The quasistatic motion of both the silicone samples and the spring model is completely determined by the static energy landscape and its singularities, so that dynamical rules for unstable states never change which descendant an ancestor reaches.

What would settle it

Construct a two-unit metamaterial whose measured strain map contains an intersection of fold curves, drive it through that intersection from a multistable domain, and check whether the final state differs from the unique descendant predicted by merging the two ADS triplets of the corresponding path-transition graph.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The manuscript develops a singularity-based framework for path-dependent responses of multistable materials under vectorial driving. Combining experiments on two carefully characterized two-unit silicone metamaterials (samples A and B with distinct stiffness ratios k) and a minimal spring model (Eq. 1 / B4), the authors show that codimension-one fold bifurcations generate ADS triplets that form the elementary building blocks of path-transition graphs (pt-graphs) and associated strain maps. Higher-order singularities (cusps corresponding to pitchforks, and a codimension-four butterfly at critical coupling k*/δ^{2} ≈ 10) organize the global arrangement of fold curves, thereby shaping the distinct pt-graphs of the strong- and weak-coupling regimes. Non-generic paths through fold intersections or cusps remain captured by the pt-graph (with spontaneous symmetry breaking at pitchforks), and t-graphs for scalar driving are recovered exactly as the restriction of pt-graphs to one-dimensional driving paths (Sec. III C).

Significance. If the results hold, the work supplies a unified graph-theoretic and singularity language for both scalar and vectorial driving of multistable systems, elevating path dependence from an obstacle to a design resource for programmable metamaterials, soft robotics, and in-materia computation. Strengths include independent experimental strain maps for two samples, a transparent spring model whose lowest-order reduction yields analytic estimates for the critical strains ε_s and ε_pf that match numerics (Fig. 10 and Appendix B), and an explicit, non-tautological reduction of t-graphs to 1-D slices of pt-graphs. The framework rests on the authors’ prior definition of pt-graphs yet adds new singularity classifications and experimental support that are not circular.

major comments (2)
  1. [Sec. III A 2] Sec. III A 2: The non-local definition of states (illustrated by the restricted-strain-range example for sample A) is load-bearing for the predictive power of any finite strain map. The manuscript should state more explicitly how a practitioner decides that a given strain domain has been exhaustively mapped, and whether extending the domain can retroactively split previously identified states; without this, the completeness claim for the pt-graphs in Fig. 3 remains slightly underspecified.
  2. [Sec. III A 2] Sec. III A 2 and III B: The authors correctly flag that dynamical selection rules may be required when an irreversible transition leaves a multi-state domain, yet assert that this does not arise in the presented data. For the two-unit systems studied every irreversible edge terminates in a monostable domain, so the ADS triplets are unique; however, the central claim that pt-graphs are fully determined by the static fold loci would be strengthened by a short explicit statement of the dynamical assumption (or a reference to the companion work) that is used to resolve descendants in general.
minor comments (4)
  1. [Fig. 3 / Appendix A] Fig. 3 and experimental methods (Appendix A): Experimental fold loci are shown without error bars or uncertainty bands, despite the stated strain accuracy of ~10^{-3}. Adding modest error estimates (or a statement that topology is robust within the accuracy) would improve quantitative comparison with the model.
  2. [throughout] Typographical errors: “These is only one stable configuration” (p. 3), “inddimensional strain-space” (p. 3), “a priory clear” (p. 4), and “in preperation” in Ref. [20].
  3. [Eq. (1) / Sec. III B 3] Notation: the same symbol k is used both for the stiffness ratio and (with asterisk) for the critical coupling; a brief clarifying sentence near Eq. (1) would help.
  4. [Fig. 5] Fig. 5 caption and panels: the experimental pitchfork is described as “near-perfect”; a quantitative measure of left–right asymmetry (or residual bias) would make the comparison with the ideal numerical pitchfork more precise.

Circularity Check

1 steps flagged

Minor non-load-bearing self-citation of the authors' prior pt-graph definition; singularity analysis, strain maps, butterfly critical coupling, and t-graph reduction are independently derived from experiment and the spring model.

specific steps
  1. self citation load bearing [Abstract / Sec. I (Introduction)]
    "To describe these responses, we recently introduced a framework based on path-transition graphs (pt-graphs) for multistable materials under vectorial driving [17]."

    The graph-theoretic language used throughout (nodes = states, directed/undirected edges = irreversible/reversible transitions across fold loci) is taken from the authors' own prior work. This is a self-citation, yet it is not load-bearing for the novel results: the locations of the folds, the cusp and butterfly singularities, the experimental strain maps of samples A and B, the value of k*, and the explicit mapping from pt-graphs to t-graphs are all derived independently from the energy landscape and new data.

full rationale

The paper's central claims concern how higher-order singularities (cusps, pitchforks, codimension-four butterfly) organize fold curves into strain maps and pt-graphs, plus the explicit reduction of t-graphs to one-dimensional slices of pt-graphs. These are obtained from new experimental strain maps on samples A/B, recursive construction of states from fold loci, numerical continuation of the spring energy (Eq. 1/B4) versus k, and analytic estimates of ε_s and ε_pf that locate the butterfly at k*/δ²≈9–10. The sole self-citation of note is the prior introduction of the pt-graph formalism itself ([17]); that citation supplies the graph language but is not used to force any of the singularity classifications, the measured critical coupling, or the topological reduction in Sec. III C. No parameter is fitted to data and then re-labeled a prediction of the same data; no uniqueness theorem is imported to forbid alternatives; ADS triplets and reversible pairs follow directly from the observed fold curves rather than by definitional tautology. The quasistatic assumption is flagged by the authors themselves and does not create a circular loop. Hence only a minor, non-load-bearing self-citation is present.

Axiom & Free-Parameter Ledger

3 free parameters · 4 axioms · 2 invented entities

The central claims rest on standard singularity theory of gradient systems, the authors’ prior definition of pt-graphs, a minimal elastic energy whose lowest-order form is a set of coupled quartics, and the experimental identification of fold loci. Free parameters are the measured stiffness ratio k and geometric bias δ of each sample; the critical coupling k* is derived rather than freely fitted. No new particles or forces are postulated.

free parameters (3)
  • stiffness ratio k = k_h / k_v = 0.089 (A), 0.036 (B)
    Measured independently for each sample from force-extension curves of isolated springs (Appendix A.2); values k_A≈0.089, k_B≈0.036 set the two distinct strain-map topologies.
  • geometric bias δ = ≈0.08
    Rest-position offset of the internal nodes; measured from sample geometry and used both in experiment and in the spring model; effective interaction scales as k/δ².
  • critical coupling k*/δ² = ≈10
    Location of the butterfly singularity that separates strong- and weak-coupling regimes; obtained numerically and approximated analytically as ≈9–10; not a free fit but a derived threshold.
axioms (4)
  • domain assumption Transitions under quasistatic driving occur at fold (saddle-node) bifurcations of the elastic energy; higher-codimension singularities organize the global arrangement of those folds.
    Standard catastrophe/singularity theory for gradient systems; invoked throughout Sec. III and Appendix B.
  • ad hoc to paper States are equivalence classes of configurations that can be smoothly deformed into one another without crossing a fold hypersurface.
    Definition introduced for pt-graphs (Sec. III A); differs from the coarser equivalence used in ordinary t-graphs and is essential to the mixed reversible/irreversible edge structure.
  • domain assumption The lowest-order spring energy (Eq. B4 / B9) captures the same singularity structure as the experimental silicone samples.
    Justified by the observed match of fold loci and pitchforks; higher-order geometric nonlinearities are neglected.
  • domain assumption When an unstable configuration appears, the subsequent dynamics select a unique descendant that is consistent with the ADS triplet of the crossed fold.
    Acknowledged as non-trivial (Sec. III A 2, citing Jin & van Hecke 2025); claimed not to produce ambiguity in the data presented.
invented entities (2)
  • ADS triplet (ancestor–descendant–sibling) independent evidence
    purpose: Elementary building block that associates each fold curve segment with one irreversible and one reversible transition, generating the mixed edges of a pt-graph.
    Defined and used systematically in this paper (and the companion ‘in preparation’ note); independent evidence is the experimental observation of the three configurations linked by the predicted transitions.
  • path-transition graph (pt-graph) independent evidence
    purpose: Mixed directed/undirected graph that encodes both reversible and irreversible transitions under multi-parameter driving.
    Introduced in the authors’ prior work; the present paper supplies the singularity organization and the reduction to t-graphs. Independent experimental realizations are given here.

pith-pipeline@v1.1.0-grok45 · 19340 in / 3079 out tokens · 33434 ms · 2026-07-14T06:27:05.523677+00:00 · methodology

0 comments
read the original abstract

Describing and predicting the response of multistable materials to external driving is central to memory formation, programmable metamaterials, soft robotics, and in-materia computing. While scalar driving is captured by transition graphs (t-graphs), vectorial driving produces path-dependent responses that require the recently introduced path-transition graphs (pt-graphs). In both cases, transitions are governed by singularities in the energy landscape: for scalar driving these correspond to saddle-node bifurcations, but for vectorial driving, higher-order singularities become important. Combining experiments on chain-like metamaterials with a minimal spring model, we investigate how higher-order singularities shape pt-graphs and the resulting path-dependent responses. We moreover discuss the role of non-generic driving paths through higher-order singularities, where the response is governed by spontaneous symmetry breaking. Finally, we demonstrate how t-graphs emerge as the one-dimensional limit of pt-graphs, unifying scalar and vectorial driving within a common graph-based framework. These results establish a singularity-based approach to path-dependent responses and provide a foundation for designing multistable materials with programmable sequential functionality for smart sensing, soft robotics, and in-materia computation.

Figures

Figures reproduced from arXiv: 2607.11179 by Colin M. Meulblok, Martin van Hecke.

Figure 1
Figure 1. Figure 1: FIG. 1. Geometry. (a) Staggered chain of two left-right sym [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Strain maps and path-transition (pt) graphs for [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Pitchfork bifurcation observed along a non-generic [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. Evolution of strain map as a function of coupling [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
Figure 9
Figure 9. Figure 9: FIG. 9. (a-c) Force-extension curves of the vertical spring, [PITH_FULL_IMAGE:figures/full_fig_p009_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: FIG. 10. Evolution of bifurcation points with [PITH_FULL_IMAGE:figures/full_fig_p011_10.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

38 extracted references · 1 linked inside Pith

  1. [1]

    Sample A—These is only one stable configuration at zero strain

    Examples of pt-graphs To illustrate the procedure of obtaining the strain map and pt-graph, we experimentally determine these for samples A and B, which differ in their stiffness ratio k(see appendix A). Sample A—These is only one stable configuration at zero strain. By following its evolution under clockwise and counter clockwise strain loops (see append...

  2. [2]

    However, there are two subtleties

    Subtleties The combination of strain map and pt-graph gives a powerful tool to characterize and distinguish path depen- dencies in multistable systems. However, there are two subtleties. First, as path dependencies lead to non-local proper- ties, extending the strain range of the strain map may lead to new states and transitions even in the regime that wa...

  3. [3]

    Intersections of fold lines To illustrate that the pt-graph fully describes strain path that cross through intersections of fold bifurcations, we consider sample B, subject to strain paths through intersection pointI(Fig. 3c). For each path linking the monostable stateMto one of the three tristable statesT i, the ’aggregate’ transition can be unambigu- ou...

  4. [4]

    Cusp singularities Next, we consider strain paths through the cusp sin- gularities in the strain maps which correspond to pitch- fork bifurcations (Fig. 3). To do so, we revisit sample (a) (b) (c) (d) diag diag (c) (d) FIG. 5. Pitchfork bifurcation observed along a non-generic strain path through a cusp singularity. (a) The collective coordinatex 1 +x 2 a...

  5. [5]

    We numerically track the strain map as a function of the coupling coefficientkfor fixedδ- recall that the ef- fective interaction scales ask/δ 2

    Butterfly singularity: the transition between strong and weak coupling Higher order singularities may also arise when we con- sider the path-dependent response of a family of systems; for example, the strain maps of samples A and B reflect different organizations of fold singularities in strain space and originate from a co-dimension four butterfly singu-...

  6. [6]

    N. C. Keim, J. D. Paulsen, Z. Zeravcic, S. Sastry, and S. R. Nagel, Rev. Mod. Phys.91, 035002 (2019)

    2019

  7. [7]

    J. D. Paulsen and N. C. Keim, Annual Review of Con- densed Matter Physics16, 61 (2025)

    2025

  8. [8]

    Mungan and M

    M. Mungan and M. M. Terzi, Annales Henri Poincar´ e 20, 2819 (2019)

    2019

  9. [9]

    N. C. Keim, J. Hass, B. Kroger, and D. Wieker, Phys. Rev. Res.2, 012004 (2020)

    2020

  10. [10]

    M. M. Terzi and M. Mungan, Phys. Rev. E102, 012122 (2020)

    2020

  11. [11]

    Regev, I

    I. Regev, I. Attia, K. Dahmen, S. Sastry, and M. Mungan, Phys. Rev. E103, 062614 (2021)

    2021

  12. [12]

    N. C. Keim and J. D. Paulsen, Science Advances7, eabg7685 (2021)

    2021

  13. [13]

    C. W. Lindeman and S. R. Nagel, Science Advances7, eabg7133 (2021)

    2021

  14. [14]

    C. W. Lindeman, T. R. Jalowiec, and N. C. Keim, Science Advances11, eadr5933 (2025)

    2025

  15. [15]

    Bense and M

    H. Bense and M. van Hecke, Proceedings of the National Academy of Sciences118, e2111436118 (2021)

    2021

  16. [16]

    Shohat, D

    D. Shohat, D. Hexner, and Y. Lahini, Proceedings of the National Academy of Sciences119, e2200028119 (2022)

    2022

  17. [17]

    Jules, A

    T. Jules, A. Reid, K. E. Daniels, M. Mungan, and F. Lechenault, Phys. Rev. Res.4, 013128 (2022)

    2022

  18. [18]

    J. D. Paulsen and N. C. Keim, Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sci- ences475, 20180874 (2019)

    2019

  19. [19]

    Florijn, C

    B. Florijn, C. Coulais, and M. Van Hecke, Phys. Rev. Lett.113, 175503 (2014), ISSN 0031-9007, 1079-7114

    2014

  20. [20]

    Sirote-Katz, D

    C. Sirote-Katz, D. Shohat, C. Merrigan, Y. Lahini, C. Nisoli, and Y. Shokef, Nat Commun15, 4008 (2024), ISSN 2041-1723

    2024

  21. [21]

    X. Guo, M. Guzm´ an, D. Carpentier, D. Bartolo, and C. Coulais, Nature618, 506 (2023)

    2023

  22. [22]

    C. M. Meulblok, A. Singh, M. Labousse, and M. van Hecke,Path-dependency and emergent computing under vectorial driving(2025), 2503.07764

    Pith/arXiv arXiv 2025

  23. [23]

    Muhaxheri, V

    G. Muhaxheri, V. Antonetti, and C. D. Santangelo, New Journal of Physics27, 125004 (2025)

    2025

  24. [24]

    T. Yang, D. Hathcock, Y. Chen, P. L. McEuen, J. P. Sethna, I. Cohen, and I. Griniasty, Proceedings of the National Academy of Sciences120, e2300081120 (2023)

    2023

  25. [25]

    C. M. Meulblok and M. van Hecke,Path dependency: generic and non-generic paths, in preperation

  26. [26]

    Jin and M

    L. Jin and M. van Hecke, Phys. Rev. Lett.135, 218201 (2025)

    2025

  27. [27]

    Arnol’d, G

    V. Arnol’d, G. Wassermann, and R. Thomas,Catastro- phe Theory(Springer Berlin Heidelberg, 2003), ISBN 9783540548119

    2003

  28. [28]

    J. Liu, M. Teunisse, G. Korovin, I. R. Vermaire, L. Jin, H. Bense, and M. van Hecke, Proceedings of the National Academy of Sciences121, e2308414121 (2024)

    2024

  29. [29]

    Mungan, S

    M. Mungan, S. Sastry, K. Dahmen, and I. Regev, Phys. Rev. Lett.123, 178002 (2019)

    2019

  30. [30]

    Melancon, A

    D. Melancon, A. Forte, L. Kamp, B. Gorissen, and K. Bertoldi, Advanced Functional Materials32(2022)

    2022

  31. [31]

    Ding and M

    J. Ding and M. van Hecke, The Journal of Chemical Physics156(2022)

    2022

  32. [32]

    M. H. Teunisse and M. van Hecke,Transition graphs of interacting hysterons: Structure, design, organization 8 and statistics(2024), 2404.11344

    arXiv 2024

  33. [33]

    van Hecke, Phys

    M. van Hecke, Phys. Rev. E104, 054608 (2021). 9 A. Experimental details

    2021

  34. [34]

    The vertical beams in each geometry have thickness 3.5 mm and have a ta- pered connection to minimize torsional effects (Fig

    Geometry and Fabrication Experimentally, we focus on two distinct samples that differ only in the geometry of the horizontal beams result- ing in distinct effective couplingk. The vertical beams in each geometry have thickness 3.5 mm and have a ta- pered connection to minimize torsional effects (Fig. 8a). All our samples have a thickness of 10 mm to preve...

  35. [35]

    Measuring Effective Coupling We measure the mechanical response of the verti- cal and both horizontal springs to determine the effec- tive coupling of sample Ak A =k h,A/kv and sample B kB =k h,B/kv (the ratio between the horizontal and ver- tical spring constants). We fabricate both springs discon- nected from the system using the same 3d-printing and mo...

  36. [36]

    Each beam is connected at both ends to a DC servo motor actuator (Thorlabs Z825B)

    Driving We use a custom-made actuation device to indepen- dently compress each beam. Each beam is connected at both ends to a DC servo motor actuator (Thorlabs Z825B). To ensure quasi-static driving conditions, the motors operate at a constant velocity of 0.1 mm/s (cor- responding to a strain rate of 5×10 −4 s−1); accelerations are set to 3 mm/s. Each mot...

  37. [37]

    Hence, our model system consists of coupled quartic potentials and thus represents a general class of multistable (disordered) (meta)materials

    Remarks on Spring Model In the lowest order, the energy simplifies to: E= nX i=1 x2 i 2 −γ i 2 + n−1X i=1 k 2 xi −x i+1 −2δ(−1) i 2 , (B4) whereγ i :=ε i +δ 2/2−ε 2 i /2≈ε i +δ 2/2. Hence, our model system consists of coupled quartic potentials and thus represents a general class of multistable (disordered) (meta)materials. Additionally, in the small stra...

  38. [38]

    For typical values ofkandδ, e.g.,k= 0.09 andδ= 0.08, k′ ≫1, andk ′ ≫ε ′, so in the remainder we sety=ε ′/k′

    We thus setε ′ 1 =ε ′ 2 =ε ′ andx ′ 1 =−1−yand x′ 2 = 1 +y, and obtain: E′(y) = 1 2 y2 + 2y−2ε ′ 2 + 2k′y2 .(B12) The equilibrium position of the system is given by: ∂yE′ = 2 y2 + 2y−2ε ′ (y+ 1) + 4k ′y= 0,(B13) which yields that in lowest order iny,y≈ε ′/(1+k ′ −ε ′). For typical values ofkandδ, e.g.,k= 0.09 andδ= 0.08, k′ ≫1, andk ′ ≫ε ′, so in the rema...