Shift equivalence implies flow equivalence for shifts of finite type

Mike Boyle

REVIEW 3 major objections 3 minor 24 references

Shift equivalence implies flow equivalence for shifts of finite type

T0 review · 3 major / 3 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Shift equivalence over Z+ implies flow equivalence for shifts of finite type.

desk verdict A long-open implication in SFT classification, proved with explicit PSE equations and a partitioned argument; the main caveat is a justified reliance on the author's earlier classification theorem, but the dependence is transparent. read the letter →

arxiv 2411.14629 v2 pith:A66EIF22 submitted 2024-11-21 math.DS math.OA

Mike Boyle This is my paper · ORCID

classification math.DSmath.OA MSC 37B1046L35

keywords shiftequivalenceflowshiftsoffinitetypeeventualconjugacyPSEequationpartitionedmatricesSL(Z[t])cyclecomponents

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper proves that two fundamental equivalence relations on shifts of finite type are nested: if square nonnilpotent matrices $A$ and $B$ with entries in $\mathbb{Z}_+$ are shift equivalent over $\mathbb{Z}_+$, then the edge shifts of finite type $\sigma_A$ and $\sigma_B$ are flow equivalent. In dynamical language, eventual conjugacy implies flow equivalence for shifts of finite type. The proof is constructive: it turns a shift equivalence into an explicit polynomial matrix equation (the PSE equation), evaluates the equation at $t=1$, and shows the resulting stabilized $\mathrm{SL}(\mathbb{Z})$ equivalence is positive on cycle components, exactly what the flow equivalence classification demands. A closing section shows the implication fails for general topological systems, which can be eventually conjugate without being flow equivalent.

What carries the argument

The engine is the PSE (polynomial shift equivalence) equation. From a lag-$\ell$ shift equivalence $A^\ell=RS$, $B^\ell=SR$, $AR=RB$, $BS=SA$ over a ring, the paper writes an explicit $2\times 2$ block matrix identity over $\mathbb{Z}[t]$ showing that stabilizations of $I-tA$ and $I-tB$ are $\mathrm{SL}(\mathbb{Z}[t])$-equivalent. The proof first reduces $A$ and $B$ to block upper triangular form with irreducible diagonal blocks, repartitions the shift equivalence according to the component poset (Proposition 4.3), and applies the $P$-partitioned version of the PSE equation. Setting $t=1$ produces a stabilized $\mathrm{SL}_P(\mathbb{Z})$-equivalence of $I-A$ and $I-B$; a nonnegativity argument (Proposition 5.8) shows this equivalence is positive on cycle components, which is precisely the hypothesis of the flow equivalence classification theorem (Theorem 5.7) that yields the conclusion.

What would settle it

Exhibit one pair of square nonnilpotent matrices over $\mathbb{Z}_+$ that are shift equivalent over $\mathbb{Z}_+$ but whose edge shifts of finite type are not flow equivalent; no such pair is known, and finding one would refute Theorem 1.1. A direct place to look is the family of reducible shifts of finite type known to be eventually conjugate but not conjugate: computing their flow equivalence invariants (the K-web, equivalently the stabilized $\mathrm{SL}_P(\mathbb{Z})$ class with positivity on cycle components) would decide the theorem for those examples.

Watch

Extended reading notes

Core claim

The central claim is Theorem 1.1: for square nonnilpotent matrices $A$ and $B$ over $\mathbb{Z}_+$, shift equivalence over $\mathbb{Z}_+$ implies that the edge shifts of finite type $\sigma_A$ and $\sigma_B$ are flow equivalent. Since every shift of finite type is conjugate to an edge shift, and shift equivalence over $\mathbb{Z}_+$ characterizes eventual conjugacy, the equivalent Theorem 1.2 states that eventual conjugacy implies flow equivalence for shifts of finite type. The result was already known for irreducible shifts of finite type through the earlier algebraic classification, and it was remarked without proof in a 2000 survey; this paper supplies the full proof for reducible systems. It also records a consequence for C*-algebras: any equivariant stable isomorphism of not necessarily simple Cuntz–Krieger algebras (a class of C*-algebras built from finite 0-1 matrices) can be replaced by a diagonal-preserving one, as follows from an existing diagram of implications.

Load-bearing premise

The proof depends on a previously published classification theorem (Theorem 5.7) that is quoted, not proved here; if that classification is wrong, the conclusion need not follow.

Editorial extensions

If this is right

Theorem 1.2: eventual conjugacy implies flow equivalence for all shifts of finite type, not just irreducible ones.
The proof is an explicit algorithm: from the matrices $R$ and $S$ witnessing a shift equivalence, one writes the PSE equation, specializes at $t=1$, and reads off the stabilized $\mathrm{SL}_P(\mathbb{Z})$ equivalence.
Via the diagram of implications in the C*-algebra literature, Theorem 1.1 implies that any equivariant stable isomorphism of not necessarily simple Cuntz–Krieger algebras can be replaced by a diagonal-preserving one.
The theorem leaves open whether eventual conjugacy implies flow equivalence for arbitrary subshifts; Section 7 shows the implication is false for general (non-subshift) systems.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The explicit PSE equation gives future invariants a canonical test object: to check whether a proposed invariant separates two flow-equivalent SFTs, one can plug in the stabilized $\mathrm{SL}_P(\mathbb{Z})$ equivalence this construction produces, rather than hunting for one.
Because the proof's last step leans on a classification theorem quoted from an earlier paper and patched with a corrected lemma, a careful reader will want to re-verify the two places that earlier lemma was used; if the patch is incomplete, the bridge has a hole.
The same construction might adapt to other semirings (for example, replacing $\mathbb{Z}_+$ by $\mathbb{R}_+$), since the positivity argument only uses nonnegativity of the factors; the paper does not explore this.
Question 6.6—whether eventually conjugate subshifts can fail to be flow equivalent—now looks sharper: if such examples exist, they can only be non-SFT subshifts, since the SFT case is settled by this theorem.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

You should know that this paper closes a long-open gap in the basic equivalences for shifts of finite type: shift equivalence over Z+ implies flow equivalence. We already knew the irreducible case from Franks; the general reducible case is new and the proof is direct.

What the paper does well: it constructs an explicit polynomial shift equivalence (PSE) equation from a given shift equivalence (Section 3), develops a partitioned version that respects the poset of irreducible components (Section 4), and then verifies the positivity-on-cycle-components condition needed for the final classification step (Section 6). The algebra in the PSE equation and the partitioned lemma (Prop 4.3) is explicit and checkable. Proposition 5.8 and Lemma 5.9 are proved in detail and are the right tools for the positivity argument. This is a serious, careful piece of work.

The soft spot is the one the stress-test flags: the final step invokes Theorem 5.7, a flow-equivalence classification taken from the author's 2002 paper [6], which is not re-proved here. And [6] had a known error in Lemma 2.6; the paper patches it (Remark 5.12) by replacing it with Lemma 5.9 and asserts the proofs go through. I do not think this is a load-bearing flaw. The author identifies exactly where Lemma 2.6 was used, the replacement is proved with the needed generality, and Remark 5.13 explains how Theorem 5.7 follows from [6, Theorem 3.1]. The remaining worry is the usual one about depending on a published external theorem; it is not a concrete gap. A referee should still check the two cited uses and the consistency of the positivity definitions, but that is a bounded verification task, not a reason to reject.

Overall: the result is new, the proof is transparent, and I believe the claim is correct. This paper deserves a serious refereeing and will be cited. I would accept it for peer review and expect it to be published, with the erratum discussion kept as is.

Referee Report

3 major / 3 minor

Summary. The paper proves that if A and B are nonnilpotent square matrices with entries in Z+ and are shift equivalent over Z+, then the corresponding edge shifts of finite type are flow equivalent (Theorem 1.1). The proof constructs an explicit polynomial shift equivalence equation (3.2) from a shift equivalence, then a P-partitioned version (4.9) via a unique partitioning result (Proposition 4.3). Setting t = 1 yields an SL_P(Z) equivalence of stabilizations of I - A and I - B (Eq. (6.3)), and the paper verifies the positivity on cycle components required by the flow equivalence classification Theorem 5.7. Theorem 1.2 states the equivalent formulation that eventual conjugacy implies flow equivalence for SFTs, and Section 7 supplies examples showing the implication fails outside the subshift class.

Significance. The result is a significant advance for the classification of shifts of finite type: it proves a long-standing implication that was previously known only in the irreducible case and directly yields that SE over Z+ preserves the algebraic invariants of flow equivalence. The explicit PSE equation is a useful tool that may have further applications, and the paper is careful with the partition structure, especially in Proposition 4.3. The main limitation is that the final classification step uses the author's 2002 theorem [6, Theorem 3.1], whose statement is not proved here and which relies on an erratum in [6] that the paper repairs only by assertion. With a completed verification of this dependence, the paper would be a strong contribution.

major comments (3)

[Theorem 5.7 and Remark 5.12] The proof of Theorem 1.1 in Section 6 depends on Theorem 5.7 as the final classification step, but Theorem 5.7 is not proved in this paper; it is extracted from [6, Theorem 3.1] by Remark 5.13. Since Remark 5.12 states that [6, Lemma 2.6] contains an error and asserts that replacing it by Lemma 5.9 makes the proofs in [6] go through, the validity of this classification result is not self-contained. Please include a proof of Theorem 5.7, or at least a complete verification that Lemma 5.9 repairs every use of Lemma 2.6 that is needed for [6, Theorem 3.1].
[Section 6, after Eq. (6.3)] The matrix M defined to give the induced cokernel isomorphism is printed as (I - R; 0 I)_p (I 0; S 0)_p. This does not match the left-hand side of (6.3), where the multipliers appear as (I 0; S I) followed by (I - R; 0 I), and it conflicts with the subsequent phi_1, phi_2 composition. The correct expression is M = (I 0; S I)_p (I - R; 0 I)_p. The intended argument is clear, but the displayed definition should be corrected.
[Section 6, application of Theorem 5.7] After obtaining the SL_P equivalence (6.3), the paper says that Theorem 5.7 applies provided positivity on cycle components is verified. However, condition (a) of Theorem 5.7, namely that A and B have the same cycle components, is not explicitly established. This follows from the shift equivalence equations: if A{p,p} is essentially cyclic then A^ell{p,p} = R{p,tilde p} S{tilde p,p} is a permutation matrix, forcing B^ell{tilde p,tilde p} = S{tilde p,p} R{p,tilde p} to be a permutation matrix as well; but the proof should state this to make the application of Theorem 5.7 complete.

minor comments (3)

[Proposition 4.3(3)] In the proof of claim (3), the sentence 'Similarly, S{tilde p, r} != 0 implies 0 != B^ell{tilde p, tilde r} != 0, hence p <= r' is garbled; the displayed inequality should be a single nonzero condition and the conclusion about p <= r in P_A needs a clearer derivation.
[Throughout] There are several typos, e.g., 'consideral umprovement' in the acknowledgments and the odd '(T k,k)k' in Section 7; these should be fixed.
[Proposition 2.1] Proposition 2.1 is invoked to reduce to M^{sq,delta}(Z+), but its proof is only sketched; this is acceptable as a standard exercise, but a reference to a full proof would improve the paper.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the proof derives an algebraic equivalence from shift equivalence and applies an external flow-equivalence classification; the cited classification does not presuppose the target implication.

full rationale

The derivation chain is self-contained except for reliance on the external classification Theorem 5.7. Section 3 computes the PSE equation (3.2) directly from the four shift-equivalence equations, and Section 4 reproves the partitioned version (4.9); no step defines shift equivalence in terms of flow equivalence or fits a parameter to the target statement. Section 6 specializes the PSE equation at t=1 to obtain an SL_P(Z) equivalence of stabilizations and verifies the 'positive on cycle components' condition using Proposition 5.8, which is proved in the paper. The final implication is supplied by Theorem 5.7, a biconditional characterization of flow equivalence in terms of such equivalences; the direction used, (2) => (1), has no shift-equivalence hypothesis, so applying it to the constructed algebraic equivalence is not a restatement of the theorem being proved. Theorem 5.7 is quoted from the author's 2002 paper [6], and Remark 5.13 explains the extraction; this is self-citation, but it is not circular because [6]'s classification is an independently published external result rather than a premise that already contains 'SE implies FE'. The paper does flag an external correctness risk: Remark 5.12 reports an error in [6, Lemma 2.6] and asserts that replacing it with Lemma 5.9 makes the proofs in [6] go through. That assertion is an unproved patch, and Theorem 5.7 is not proved in this manuscript; however, a missing proof or an external correctness concern is not the same as the target result being equivalent to its own inputs. No fitted quantity is renamed as a prediction, no ansatz is smuggled in via citation, and no known result is merely renamed. Therefore the circularity score is 0.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The paper introduces no free parameters and no invented entities. The central claim rests on standard definitions, a known equivalence between shift equivalence and eventual conjugacy, and the author's 2002 flow equivalence classification theorem, one lemma of which is corrected here. The main genuinely new input is the explicit PSE equation and its partitioned version, derived inside the paper.

assumptions (5)

standard math Standard definitions and facts: SE_Z+ iff eventual conjugacy; SSE_Z+ iff conjugacy; every SFT is conjugate to an edge SFT.
Section 2, citing Lind and Marcus [20, Theorems 7.5.15 and 2.3.2].
standard math Franks' theorem: for irreducible non-permutation matrices over Z+, flow equivalence is equivalent to isomorphism of cok(Z^n/(I-A)Z^n) and equality of det(I-A).
Theorem 5.1, citing [15]; used for background and for the irreducibility case.
domain assumption Boyle's flow equivalence classification: for A, B in M^{sq,delta}(Z+), flow equivalence of sigma_A and sigma_B is equivalent to existence of a poset P and a stabilized SL_P(Z) equivalence of I-A and I-B that is positive on cycle components.
Theorem 5.7, extracted from [6, Theorem 3.1]; the central external theorem used in the proof of Theorem 1.1.
standard math Proposition 2.1: every nonnilpotent square matrix over Z+ is SSE-Z+ to a block upper triangular matrix with irreducible diagonal blocks.
Section 2, proof sketched as an exercise with an explicit ESSE example; standard result.
domain assumption The validity of [6, Theorem 3.1] after replacing [6, Lemma 2.6] with Lemma 5.9 of this paper.
Remark 5.12: the paper corrects an erratum in [6] and asserts the proofs relying on it go through.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Shift equivalence implies flow equivalence for shifts of finite type." pith.science (2026). https://pith.science/paper/A66EIF22

@misc{pith2026241114629,
  author       = {Pith},
  title        = {Pith review of: Shift equivalence implies flow equivalence for shifts of finite type},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/A66EIF22}},
  note         = {Machine review of arXiv:2411.14629}
}

read the original abstract

Shifts of finite type defined from shift equivalent matrices must be flow equivalent.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 23 canonical work pages

[6]

Flow equivalence of shifts of finite type via positive factorizations.Pacific J

Mike Boyle. Flow equivalence of shifts of finite type via positive factorizations.Pacific J. Math., 204(2):273–317, 2002. SHIFT EQUIV ALENCE IMPLIES FLOW EQUIV ALENCE 19

work page 2002
[1]

Homology for zero-dimensional nonwandering sets

Rufus Bowen and John Franks. Homology for zero-dimensional nonwandering sets. Ann. of Math. (2), 106(1):73–92, 1977

work page 1977
[2]

Boyle and J

M. Boyle and J. B. Wagoner. Positive algebraicK-theory and shifts of finite type. InModern dynamical systems and applications, pages 45–66. Cambridge Univ. Press, Cambridge, 2004

work page 2004
[3]

Shift equivalence and the Jordan form away from zero.Ergodic Theory Dynam

Mike Boyle. Shift equivalence and the Jordan form away from zero.Ergodic Theory Dynam. Systems, 4(3):367–379, 1984

work page 1984
[4]

Symbolic dynamics and matrices

Mike Boyle. Symbolic dynamics and matrices. InCombinatorial and graph-theoretical problems in linear algebra (Minneapolis, MN, 1991), volume 50 ofIMA Vol. Math. Appl., pages 1–38. Springer, New York, 1993

work page 1991
[5]

Algebraic aspects of symbolic dynamics

Mike Boyle. Algebraic aspects of symbolic dynamics. InTopics in symbolic dynamics and applications (Temuco, 1997), volume 279 ofLondon Math. Soc. Lecture Note Ser., pages 57–88. Cambridge Univ. Press, Cambridge, 2000

work page 1997
[7]

Flow equivalence and isotopy for subshifts.Dyn

Mike Boyle, Toke Meier Carlsen, and Søren Eilers. Flow equivalence and isotopy for subshifts.Dyn. Syst., 32(3):305–325, 2017. Corrigendum: ibid, page (ii)

work page 2017
[8]

Poset block equivalence of integral matrices.Trans

Mike Boyle and Danrun Huang. Poset block equivalence of integral matrices.Trans. Amer. Math. Soc., 355(10):3861–3886, 2003

work page 2003

Show all 24 references

[9]

Strong shift equivalence and algebraic K-theory.J

Mike Boyle and Scott Schmieding. Strong shift equivalence and algebraic K-theory.J. Reine Angew. Math., 752:63–104, 2019

2019
[10]

Symbolic dynamics and the stable algebra of ma- trices

Mike Boyle and Scott Schmieding. Symbolic dynamics and the stable algebra of ma- trices. In Peter J. Cameron R.A.Bailey and Yaokun Wu, editors,Groups and graphs, designs and dynamics, volume 491 ofLondon Mathematical Society Lecture Note Se- ries, pages 266–422. Cambridge Univ...

2024
[11]

A class ofC ∗-algebras and topological Markov chains.Invent

Joachim Cuntz and Wolfgang Krieger. A class ofC ∗-algebras and topological Markov chains.Invent. Math., 56(3):251–268, 1980

1980
[12]

The Williams problem through the lens of Cuntz-Krieger algebras.Ober- wolfach Rep., 19(3):2111–2113, 2022

Søren Eilers. The Williams problem through the lens of Cuntz-Krieger algebras.Ober- wolfach Rep., 19(3):2111–2113, 2022. Abstracts from the workshop held August 7–13, 2022, Organized by Dimitri Shlyakhtenko, Andreas Thom, Stefaan Vaes and Wilhelm Winter

2022
[13]

Søren Eilers, Gunnar Restorff, Efren Ruiz, and Adam P. W. Sørensen. The com- plete classification of unital graphC ∗-algebras: geometric and strong.Duke Math. J., 170(11):2421–2517, 2021

2021
[14]

¨Uber den Zusammenhang zwischen dem Begriff der Gleichartigkeit zweier Ideale und dem ¨Aquivalenzbegriff der Elementarteilertheorie.Math

Hans Fitting. ¨Uber den Zusammenhang zwischen dem Begriff der Gleichartigkeit zweier Ideale und dem ¨Aquivalenzbegriff der Elementarteilertheorie.Math. Ann., 112(1):572– 582, 1936

1936
[15]

Flow equivalence of subshifts of finite type.Ergodic Theory Dynam

John Franks. Flow equivalence of subshifts of finite type.Ergodic Theory Dynam. Systems, 4(1):53–66, 1984

1984
[16]

K. H. Kim and F. W. Roush. Williams’s conjecture is false for reducible subshifts.J. Amer. Math. Soc., 5(1):213–215, 1992

1992
[17]

K. H. Kim and F. W. Roush. The Williams conjecture is false for irreducible subshifts. Ann. of Math. (2), 149(2):545–558, 1999

1999
[18]

Kitchens.Symbolic Dynamics

Bruce P. Kitchens.Symbolic Dynamics. Universitext. Springer-Verlag, Berlin, 1998. One-sided, two-sided and countable state Markov shifts

1998
[19]

Cambridge University Press, Cambridge, 1995

Douglas Lind and Brian Marcus.An introduction to symbolic dynamics and coding. Cambridge University Press, Cambridge, 1995

1995
[20]

Cambridge University Press, Cambridge, second edition, 2021

Douglas Lind and Brian Marcus.An Introduction to Symbolic Dynamics and Coding. Cambridge University Press, Cambridge, second edition, 2021

2021
[21]

A topological invariant of flows on 1-dimensional spaces

Bill Parry and Dennis Sullivan. A topological invariant of flows on 1-dimensional spaces. Topology, 14(4):297–299, 1975

1975
[22]

N. A. Vavilov and A. V. Stepanov. Linear groups over general rings. I. Generalities. Zap. Nauchn. Sem. S.-Peterburg. Otdel. Mat. Inst. Steklov. (POMI), 394:33–139, 295, 2011

2011
[23]

R. B. Warfield, Jr. Stable equivalence of matrices and resolutions.Comm. Algebra, 6(17):1811–1828, 1978

1978
[24]

R. F. Williams. Classification of one dimensional attractors. InGlobal Analysis (Proc. Sympos. Pure Math., Vols. XIV, XV, XVI, Berkeley, Calif., 1968), volume XIV-XVI ofProc. Sympos. Pure Math., pages 341–361. Amer. Math. Soc., Providence, RI, 1970. University of Maryland, Col...

1968

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [6]

Flow equivalence of shifts of finite type via positive factorizations.Pacific J

Mike Boyle. Flow equivalence of shifts of finite type via positive factorizations.Pacific J. Math., 204(2):273–317, 2002. SHIFT EQUIV ALENCE IMPLIES FLOW EQUIV ALENCE 19

work page 2002

[2] [1]

Homology for zero-dimensional nonwandering sets

Rufus Bowen and John Franks. Homology for zero-dimensional nonwandering sets. Ann. of Math. (2), 106(1):73–92, 1977

work page 1977

[3] [2]

Boyle and J

M. Boyle and J. B. Wagoner. Positive algebraicK-theory and shifts of finite type. InModern dynamical systems and applications, pages 45–66. Cambridge Univ. Press, Cambridge, 2004

work page 2004

[4] [3]

Shift equivalence and the Jordan form away from zero.Ergodic Theory Dynam

Mike Boyle. Shift equivalence and the Jordan form away from zero.Ergodic Theory Dynam. Systems, 4(3):367–379, 1984

work page 1984

[5] [4]

Symbolic dynamics and matrices

Mike Boyle. Symbolic dynamics and matrices. InCombinatorial and graph-theoretical problems in linear algebra (Minneapolis, MN, 1991), volume 50 ofIMA Vol. Math. Appl., pages 1–38. Springer, New York, 1993

work page 1991

[6] [5]

Algebraic aspects of symbolic dynamics

Mike Boyle. Algebraic aspects of symbolic dynamics. InTopics in symbolic dynamics and applications (Temuco, 1997), volume 279 ofLondon Math. Soc. Lecture Note Ser., pages 57–88. Cambridge Univ. Press, Cambridge, 2000

work page 1997

[7] [7]

Flow equivalence and isotopy for subshifts.Dyn

Mike Boyle, Toke Meier Carlsen, and Søren Eilers. Flow equivalence and isotopy for subshifts.Dyn. Syst., 32(3):305–325, 2017. Corrigendum: ibid, page (ii)

work page 2017

[8] [8]

Poset block equivalence of integral matrices.Trans

Mike Boyle and Danrun Huang. Poset block equivalence of integral matrices.Trans. Amer. Math. Soc., 355(10):3861–3886, 2003

work page 2003