betaFeedback

WorksMathematics

Lattice-ordered Spectral Systems for Morse Representations

Abstract

We show that Morse representations can be algebraically described by spectral systems, a notion introduced by B. Matschke for poset-filtered chain complexes. This provides an algebraic alternative to connection matrices and yields algebraic invariants via R-modules over an arbitrary ring R. This approach applies in particular to Conley index theory for semi-flows. We discuss the theory of spectral systems and, as an application to dynamical systems, explain the relation to R. Franzosa's theory of connection matrices. In this context we introduce the concept of a spectral net. In contrast to traditional spectral sequences, spectral nets over a lattice display a rich network of (higher) differentials. In the application to dynamical systems these differentials provide information about connecting orbits.

Introduction

In this paper three key ingredients come together. Chronologically, the first is the foundational work by R. Franzosa on connection matrices in the setting of dynamical systems, v. [9, 11]. This provides an algebraic theory for Morse representations with information about connecting orbits and bifurcations in flows. The second is the work by J. Robbins & D. Salamon on connection matrix theory for discrete dynamical systems, v. [19]. In this work they utilize the novel observation that the attractors in a dynamical system display a lattice structure. The third is the introduction of spectral systems for filtered chain complexes by B. Matschke, which allow filterings over any poset, v. [16]. In this paper we introduce spectral systems in an abstract setting beyond filtered chain complexes. This abstract setting is applied to the theory of Morse representations and extends the approach via spectral sequences by P. Bartłomiejczyk, v. [1, 2]; cf. [6].

The decomposition of a dynamical system ϕ\phi into a hierarchical structure of invariant sets is a cornerstone of the global analysis of dynamical systems.1 Central to this decomposition is the concept of Morse representation, defined as a finite partially ordered set consisting of pairwise disjoint, compact, non-empty invariant sets — Morse sets — outside of which ϕ\phi has gradient-like behavior, v. Sect. 2 and [5, 11, 15, 13]. Formally, a Morse representation is equivalent to a finite sublattice of attractors A\mathcal{A} of ϕ\phi. To each Morse set MM, i.e. the intersection of an attractor A∈AA \in\mathcal{A} and the dual repeller A′∗A'^* of an attractor A′∈AA' \in\mathcal{A}, with A′⊆AA' \subseteq A, we assign the Conley index HC(M)\mathrm{HC}(M), an algebraic-topological invariant defined as the relative homology H(U,U′)H(U,U'), where U,U′U,U', with U′⊆UU' \subseteq U, are trapping regions for AA and A′A' respectively. The interplay between the Conley indices is captured by a Cartan–Eilenberg system E(M)\mathrm{E}(M) (or E-system), v. [21, 12]. This system provides an exhaustive algebraic framework consisting of RR-modules HC(M)=H(U,U′)\mathrm{HC}(M)=H(U,U') for every pair (A,A′)∈A×A(A,A')\in\mathcal{A}\times\mathcal{A}, with A′⊆AA'\subseteq A. Within this system, every nested pair of attractors (A,A′)(A,A') yields an exact triangle of the form:

⋯⟶H(U′)→iH(U)→jH(U,U′)→kH(U′)⟶⋯(1)\cdots\longrightarrow H(U') \xrightarrow{i} H(U) \xrightarrow{j} H(U,U') \xrightarrow{k} H(U') \longrightarrow\cdots \tag*{(1)}

An E-system serves as a maximal system of invariants, encoding the totality of algebraic (homological) data regarding Morse sets and the connecting orbits between them through a network of commutative squares and exact triangles, v. Sect. 3. In [21] it is shown that E-systems for a Morse representation are equivalent to Franzosa’s homology braids.

An equivalent formulation of this information is via a spectral system2 S(M)\mathrm{S}(M) (or S-system), which represents the data as a collection of RR-modules and differentials. If (Q,≤)(Q,\leq) is a poset and QnQ_n is the poset of ordered nn-tuples α1≤⋯≤αn\alpha_1\leq\cdots\leq\alpha_n in QQ, ordered componentwise and regarded as a thin category, then a spectral system S\mathrm{S} is defined by a functor S:Q4→Mod⁡R\mathrm{S}:Q_4\to\operatorname{Mod}_R, (α,β,γ,δ)↦Sαβγδ(\alpha,\beta,\gamma,\delta)\mapsto\mathrm{S}^{\gamma\delta}_{\alpha\beta}, together with differentials d:Sηζγδ→Sαβγδd:\mathrm{S}^{\gamma\delta}_{\eta\zeta}\to\mathrm{S}^{\gamma\delta}_{\alpha\beta}, (α,β,γ,δ,ϵ,ζ)∈Q6(\alpha,\beta,\gamma,\delta,\epsilon,\zeta)\in Q_6, which satisfy a list of axioms, v. Defn. 4.1. In our setting Q=O(M)Q=\mathrm{O}(M), the lattice of down-sets in (M,≤)(M,\leq), and

Sαβγδ=ker⁡:Eβγ→kEαβim⁡:Eγδ→kEβγ,∀(α,β,γ,δ)∈O(M)4,\mathrm{S}^{\gamma\delta}_{\alpha\beta}= \frac{\ker:E^\gamma_\beta\xrightarrow{k}E^\beta_\alpha}{\operatorname{im}:E^\delta_\gamma\xrightarrow{k}E^\gamma_\beta}, \qquad\forall(\alpha,\beta,\gamma,\delta)\in\mathrm{O}(M)_4,

where Eαβ=H(FβU,FαU)E^\beta_\alpha=H(F_\beta U,F_\alpha U) are the E-terms for E(M)\mathrm{E}(M) — Conley indices — and the sets FαUF_\alpha U define a lattice UU of trapping regions isomorphic to AA, v. Sect. 2. The differentials k:Eβγ→Eαβk:E^\gamma_\beta\to E^\beta_\alpha are the connecting homomorphisms of E(M)\mathrm{E}(M). Spectral systems indexed by a finite distributive lattice, such as O(M)\mathrm{O}(M), will be called lattice-ordered.

The spectral net approach

In the tradition of Franzosa, chain-generated E-systems, v. Sect. 3, can be encoded by an M-graded differential module (C,d)(C,d), also referred to as a connection matrix (or spectral matrix in the context of Cartan–Eilenberg systems). We emphasize that a connection matrix is not a reduction of the data but an alternative, highly structured encoding of the system. Knowing (C,d)(C,d) determines all EE-terms, SS-terms, and differentials. Connection matrices are not unique in general. The existence of connection matrices is proved by Franzosa [11], but not over arbitrary rings. For the extension to arbitrary rings, v. [23].

Because computing the full connection matrix requires global, chain-level information, which is often not available, we introduce the concept of a spectral net. Analogous to spectral sequences in homological algebra, spectral nets provide a framework to deduce structural information about EE-systems and SS-systems from partial, localized data by exploiting the underlying spectral structure.

Spectral systems offer several key advantages over traditional connection matrices. First, they operate over an arbitrary ring RR, providing broad algebraic generality. Second, their differentials are uniquely determined by the EE-system, which makes them robust under perturbations of ϕ\phi, v. [10, 8], and non-trivial differentials carry direct information about the underlying connecting orbit structure, v. Theorems A and B below. In contrast, connection matrix entries are generally non-unique and lack this straightforward geometric interpretation. Finally, spectral systems make information more accessible: while computing a connection matrix requires complete, global knowledge of the entire system up front, spectral systems can be analyzed using localized or partial data via spectral nets. To exploit this accessibility, spectral nets enable an iterative computational approach. Rather than relying on a complete system description, spectral nets incrementally recover information about the spectral system — higher SS-terms and higher differentials — step-by-step from limited initial data: the EE-terms EαβE_{\alpha}^{\beta} with β∖α\beta\setminus\alpha a singleton, and the first-order differentials between them.

An example

We consider an EE-system over the lattice of down-sets O(P)\mathcal{O}(P) of the poset P=N5P = N_{5}. Figure 1.1 gives the Hasse diagrams of both N5N_{5} and O(N5)\mathcal{O}(N_{5}). The

Hasse diagrams of the poset and its lattice of down-sets

Figure 1.1. The poset P=N5={1,2,3,4,5}P = N_{5} = \{1,2,3,4,5\} with covering relations 1<2<4<51 < 2 < 4 < 5 and 1<3<51 < 3 < 5 [left] and its lattice of down-sets O(P)\mathcal{O}(P) [right].

lattice O(N5)\mathcal{O}(N_{5}) has 3 maximal chains with 3 associated linear extensions of N5N_{5}:

K1:0⊆α⊆β⊆ϵ⊆ζ⊆1,⟹1<2<4<3<5;K2:0⊆α⊆β⊆δ⊆ζ⊆1,⟹1<2<3<4<5;K3:0⊆α⊆γ⊆δ⊆ζ⊆1,⟹1<3<2<4<5.\begin{aligned} K_{1}:&\quad0 \subseteq\alpha\subseteq\beta\subseteq\epsilon\subseteq\zeta\subseteq1,\quad\Longrightarrow\quad1<2<4<3<5;\\ K_{2}:&\quad0 \subseteq\alpha\subseteq\beta\subseteq\delta\subseteq\zeta\subseteq1,\quad\Longrightarrow\quad1<2<3<4<5;\\ K_{3}:&\quad0 \subseteq\alpha\subseteq\gamma\subseteq\delta\subseteq\zeta\subseteq1,\quad\Longrightarrow\quad1<3<2<4<5. \end{aligned}

From the structure of the Cartan–Eilenberg system we define page-1 of the spectral net by considering EE-terms corresponding to consecutive down-sets in the maximal chains. We assume that the EE-system is excisive, in particular Eαβ≅Eα′β′E_{\alpha}^{\beta} \cong E_{\alpha'}^{\beta'} for all β\α=β′\α′\beta\backslash\alpha=\beta'\backslash\alpha'.

0⟶Eζ1→d1ζεEεζ→dεβEβε→dβαEαβ→dβα0E0α⟶0↗d1ζε↙dδβα↘dβα0Eδζ→dδεEβδ↘dδγEγδ→dδγαEαγ→dγα0E0α\begin{aligned} 0 &\longrightarrow E^1_{\zeta} \xrightarrow{d_{1\zeta\varepsilon}} E^\zeta_{\varepsilon} \xrightarrow{d_{\varepsilon\beta}} E^\varepsilon_{\beta} \xrightarrow{d_{\beta\alpha}} E^\beta_{\alpha} \xrightarrow{d_{\beta\alpha0}} E^\alpha_{0} \longrightarrow0 \\ &\qquad\qquad\quad\nearrow d_{1\zeta\varepsilon}\qquad\qquad\qquad\swarrow d_{\delta\beta\alpha}\qquad\qquad\qquad\searrow d_{\beta\alpha0}\\ &\qquad\qquad\qquad E^\zeta_{\delta} \xrightarrow{d_{\delta\varepsilon}} E^\delta_{\beta} \qquad\qquad\qquad\qquad\qquad\qquad\qquad\\ &\qquad\qquad\qquad\qquad\qquad\qquad\qquad\searrow d_{\delta\gamma}\qquad\qquad\qquad\qquad\qquad\qquad\\ &\qquad\qquad\qquad\qquad\qquad\qquad\qquad E^\delta_{\gamma} \xrightarrow{d_{\delta\gamma\alpha}} E^\gamma_{\alpha} \xrightarrow{d_{\gamma\alpha0}} E^\alpha_{0} \end{aligned}

If we display each maximal chain we obtain the following diagram:

References

  1. [1]Bartłomiejczyk, P.: The Conley index and spectral sequences. Topol. Methods Nonlinear Anal. 25(1), 195–203 (2005). DOI 10.12775/TMNA.2005.010. URL https://doi.org/10.12775/TMNA.2005.010
  2. [2]Bartłomiejczyk, P.: Spectral sequences and detailed connection matrices. Topol. Methods Nonlinear Anal. 34(1), 187–200 (2009). DOI 10.12775/TMNA.2009.037. URL https://doi.org/10.12775/TMNA.2009.037
  3. [3]Boardman, J.M.: Conditionally convergent spectral sequences. In: Homotopy invariant algebraic structures (Baltimore, MD, 1998), Contemp. Math., vol. 239, pp. 49–84. Amer. Math. Soc., Providence, RI (1999). DOI 10.1090/conm/239/03597. URL https://doi-org.vu-nl.idm.oclc.org/10.1090/conm/239/03597DOI
  4. [4]Castelo-Mourelle, S., Vandervorst, R.: Higher order derived differentials in spectral systems. In preparation (2026)
  5. [5]Conley, C.C.: Isolated invariant sets and the Morse index. 38. American Mathematical Soc. (1978)DOI
  6. [6]Cornea, O., de Rezende, K.A., da Silveira, M.R.: Spectral sequences in Conley’s theory. Ergodic Theory Dynam. Systems 30(4), 1009–1054 (2010). DOI 10.1017/S0143385709000479. URL https://doi.org/10.1017/S0143385709000479
  7. [7]Davey, B.A., Priestley, H.A.: Introduction to lattices and order. Cambridge university press (2002)DOI
  8. [8]Dowling, K.A., Kalies, W.D., Vandervorst, R.C.A.M.: Continuation sheaves in dynamics: sheaf cohomology and bifurcation. J. Differential Equations 367, 124–198 (2023). DOI 10.1016/j.jde.2023.04.041. URL https://doi.org/10.1016/j.jde.2023.04.041
  9. [9]Franzosa, R.: Index filtrations and the homology index braid for partially ordered Morse decompositions. Transactions of the American Mathematical Society 298(1), 193–213 (1986)
  10. [10]Franzosa, R.: The continuation theory for Morse decompositions and connection matrices. Transactions of the American Mathematical Society 310(2), 781–803 (1988)
  11. [11]Franzosa, R.: The connection matrix theory for Morse decompositions. Transactions of the American Mathematical Society 311(2), 561–592 (1989)DOI
  12. [12]Helle, G.O., Rognes, J.: Boardman’s whole-plane obstruction group for Cartan-Eilenberg systems. Doc. Math. 24, 1855–1878 (2019)DOI
  13. [13]Kalies, W., Vandervorst, R.: Priestley duality and representations of global dynamics. Qual. Theory Dyn. Syst. 24(4), Paper No. 186, 57 (2025). DOI 10.1007/s12346-025-01343-6. URL https://doi-org.vu-nl.idm.oclc.org/10.1007/s12346-025-01343-6
  14. [14]Kalies, W.D., Mischaikow, K., Vandervorst, R.: Lattice structures for attractors I. Journal of Computational Dynamics 1, 307–338 (2014)arxiv.org/abs/1307.1737
  15. [15]Kalies, W.D., Mischaikow, K., Vandervorst, R.C.A.M.: Lattice structures for attractors III. J. Dynam. Differential Equations 34(3), 1729–1768 (2022). DOI 10.1007/s10884-021-10056-8. URL https://doi.org/10.1007/s10884-021-10056-8
  16. [16]Matschke, B.: Successive spectral sequences. Trans. Amer. Math. Soc. 375(9), 6205–6254 (2022). DOI 10.1090/tran/8650. URL https://doi-org.vu-nl.idm.oclc.org/10.1090/tran/8650DOI
  17. [17]Mischaikow, K., Reineck, J.F.: A product theorem for connection matrices and the structure of connecting orbits. Nonlinear Anal. 23(10), 1293–1314 (1994). DOI 10.1016/0362-546X(94)90148-1. URL https://doi.org/10.1016/0362-546X(94)90148-1
  18. [18]Miyata, T.: Note on direct summands of modules. J. Math. Kyoto Univ. 7, 65–69 (1967). DOI 10.1215/kjm/1250524308. URL https://doi-org.vu-nl.idm.oclc.org/10.1215/kjm/1250524308DOI
  19. [19]Robbin, J.W., Salamon, D.A.: Lyapunov maps, simplicial complexes and the stone functor. Ergodic Theory and Dynamical Systems 12(1), 153–183 (1992)
  20. [20]Salamon, D.: Connected simple systems and the Conley index of isolated invariant sets. Transactions of the American Mathematical Society 291(1), 1–41 (1985)
  21. [21]Spendlove, K., Vandervorst, R.: Graded differential vector spaces, cartan–eilenberg systems and conjectures in conley index theory. Journal of Dynamics and Differential Equations (2025). URL https://doi.org/10.1007/s10884-025-10421-x
  22. [22]Stanley, R.P.: Enumerative combinatorics. Volume 1, Cambridge Studies in Advanced Mathematics, vol. 49, second edn. Cambridge University Press, Cambridge (2012)
  23. [23]Vandervorst, R.: Representations of spectral systems over finite distributive lattices. In preparation (2026)
  24. [24]Vandervorst, R.: The Conley index for Morse sets. In preparation (2026)
  25. [25]Weibel, C.A.: An introduction to homological algebra. 38. Cambridge university press (1995)DOI

Paper details

vlnc:W_dpf9fsp6
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 10

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents