REVIEW 3 major objections 4 minor 29 references
Dynamics of continuous maps induced on the space of probability measures
T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read For interval maps with a periodic point of odd period, transitivity of the map and of its induced probability-measure map are equivalent.
desk verdict The interval transitivity criterion is a genuine sharp result and the non-autonomous counterexamples are worth having; the paper deserves serious refereeing despite a few proof-writing blemishes. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central machinery is the pushforward operator $\hat f$ on $\mathcal{M}(X)$, together with the Prohorov metric making $\mathcal{M}(X)$ compact. Most proofs run through two approximation lemmas: the set of finitely supported measures $\frac{1}{n}\sum_{i=1}^{n}\delta_{x_i}$ is dense in $\mathcal{M}(X)$, and any finite family of open subsets of $\mathcal{M}(X)$ can be entered by such measures whose atoms lie in prescribed open subsets of $X$. These lemmas allow mixing, weak mixing, exactness, chain properties, and sensitivity to be transferred back and forth between the base dynamics and the measure dynamics.
What would settle it
Exhibit a continuous interval map with a periodic point of odd period greater than 1 and a dense orbit whose induced map on $\mathcal{M}(I)$ is not topologically transitive; Theorem 2.3 says no such map exists.
Extended reading notes
Core claim
The paper's central claim is a sharp transitivity criterion for the induced system. For a continuous self-map $f$ of a compact interval $I$, write $\hat f$ for the map on the space $\mathcal{M}(I)$ of Borel probability measures defined by $\hat f(\mu)(A)=\mu(f^{-1}(A))$. Theorem 2.3 states that if $f$ has a periodic point of odd period different from 1, then $(I,f)$ is transitive if and only if $(\mathcal{M}(I),\hat f)$ is transitive; the theorem also exhibits a piecewise-linear transitive interval map with no such periodic point whose induced system is not transitive, so the condition cannot be relaxed. Around this criterion the paper establishes equivalences for total transitivity, mixing, mild mixing, exactness, and several sensitivity notions on intervals, along with a one-way inheritance of Li-Yorke and distributional chaos in the general compact metric setting.
Load-bearing premise
The proof of the main interval criterion relies on an imported classification theorem stating that, under the odd-period condition, transitivity of an interval map implies mixing; if that classification is wrong or inapplicable where it is quoted, the transitivity equivalence would need another argument.
Editorial extensions
If this is right
- On intervals, any transitive map with a periodic point of odd period greater than 1 has a transitive induced measure system, so the measure space is not merely a passive recorder of the base dynamics.
- Sensitive interval maps induce sensitive measure systems; the paper proves the stronger conclusion that cofinite sensitivity transfers, which also covers syndetic and ergodic sensitivity.
- Every transitive interval map forces infinite topological entropy on $\mathcal{M}(I)$, so finite entropy on the measure side is impossible for transitive autonomous interval systems.
- Chain transitivity of a compact system makes its induced measure system chain mixing, chain weakly mixing of all orders, chain exact, and chain transitive; surjectivity of every map in a non-autonomous sequence suffices for the same conclusion.
- If a surjective non-autonomous system is not mixing, its induced measure system does not have shadowing, so shadowing is not automatically inherited by the measure space.
Reading between the lines
- The odd-period criterion suggests a usable test in wider classes of spaces: failure of transitivity to lift to the measure space may be driven by absence of nontrivial odd periodic structure rather than by compactness or dimension alone.
- The zero-entropy non-autonomous example indicates that time-variation can erase the entropy amplification that autonomous transitive intervals display; a natural extension is to decide exactly which non-autonomous sequences restore infinite entropy on the measure space.
- The chain-mixing result for surjective sequences, paired with the two counterexamples, implies that non-surjectivity is the main obstruction to the converse; a sharper converse would probably characterize chain-mixing induced measure systems among non-surjective maps.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies the relations between the dynamics of a continuous map (or non-autonomous system) on a compact metric space X and the dynamics of the induced pushforward map on the space M(X) of Borel probability measures with the Prohorov metric. The principal result is Theorem 2.3: for an interval map (I,f) that has a periodic point of odd period different from 1, (I,f) is transitive if and only if (M(I),\hat f) is transitive, and a counterexample shows that the odd-period condition is sharp. The paper also proves equivalence of mixing, mild mixing, exactness, and several forms of sensitivity between the original and induced systems, studies topological entropy, chain mixing and chain transitivity, shadowing and specification, weak mixing of higher orders, Li-Yorke and distributional chaos, and equi-conjugacy. A number of results are extended to non-autonomous systems, with examples separating autonomous and non-autonomous behavior.
Significance. If the results hold, the paper gives a clean, sharp criterion for the transitivity equivalence on intervals and substantially extends the existing theory of induced measure systems to non-autonomous dynamics. The main strength is the combination of general transfer results with carefully constructed counterexamples, such as the ball-separation argument in Theorem 2.3(ii), the two-point examples for chain mixing and shadowing, and the entropy example in Theorem 2.6. The reliance on standard external classifications (Ruette's interval theorem, Balibrea-Oprocha's weak mixing result) is transparent and does not appear to be circular. The paper is a solid contribution to the induced-measure-space dynamics literature, provided the technical issues below are corrected.
major comments (3)
- [Section 2.1, proof of Theorem 2.3(i)] The proof states: "transitivity of (I,f) is equivalent to mixing of (I,f) by Theorem 2.20 in [23]." As written, this equivalence is false for the map constructed in part (ii) of the same theorem, which is transitive but not mixing. The intended statement is that, under the hypothesis that (I,f) has a periodic point of odd period different from 1, transitivity is equivalent to mixing. The proof should explicitly repeat the odd-period qualifier to avoid a logically false intermediate claim.
- [Section 4, Theorem 4.3, Eq. (4.11), and Theorem 4.4] The measure \hat\nu defined in (4.11) is written as (1/|A|)\sum_{j\in A}\delta_{y_j} + (1/(n_0-|A|))\sum_{j\notin A}\delta_{x_j}. This does not define a probability measure: the coefficients sum to 2, and the expression is undefined when |A|=n_0. The same problem occurs in the definition of \hat\nu_i in the proof of Theorem 4.4. The intended weights are almost certainly 1/n_0 (and 1/n_i respectively), and with that correction the subsequent inequalities are valid. As printed, the proofs are invalid and must be corrected.
- [Section 3.1, proof of Theorem 3.3] The proof invokes "Lemma 3.7" to assert that \hat f_0^k is surjective, but no Lemma 3.7 exists in the manuscript. The intended reference appears to be Lemma 3.2, applied iteratively to each surjective map f_i, which gives surjectivity of each \hat f_i and hence of the composition. This citation error should be fixed.
minor comments (4)
- [Section 2.3, proof of Theorem 2.12] The equality \hat f_0^{n_0}(\overline{B_{P_d}(\delta_x,r)}) = \hat f_0^{n_0}(D) is not literally correct; by continuity and compactness the left side equals the closure of \hat f_0^{n_0}(D). The intended conclusion about the distance from \delta_y to the image of the closed ball still follows, but the equality should be replaced by the appropriate closure statement.
- [Section 2.1, Lemma 2.5] The forward direction of the uniform convergence proof only writes the inequality \hat f_n(\mu)(A) \le \hat f(\mu)(A^\varepsilon)+\varepsilon. The symmetric inequality needed for the Prohorov metric should also be stated or explicitly derived from the same uniform convergence assumption.
- [Section 3.2, proof of Theorem 3.9] The proof applies the shadowing property to a finite \delta_0-chain, while the definition of shadowing concerns infinite pseudo-orbits. The argument should note that a finite chain can be extended to an infinite \delta_0-pseudo orbit, for instance by concatenating chains supplied by Theorem 3.3, before shadowing is applied.
- [General] There are several typographical slips, such as "exsits" in the definition of sensitivity in Section 4 and "Denfine" in the proof of Theorem 4.4. These should be corrected in the final version.
Circularity Check
No circularity found: all load-bearing results are either proved in-paper or imported from external independent sources.
full rationale
The central derivation chain of Theorem 2.3(i) is: transitivity of the interval map (I,f), combined with the hypothesis of a periodic point of odd period different from 1, implies mixing via the external interval classification theorem of Ruette ([23, Theorem 2.20]); mixing of (I,f) is then lifted to mixing of (M(I),\hat f) by the paper's own Theorem 2.11, which is proved independently in Section 2.3 from Lemmas 2.8 and 2.10 and does not invoke Theorem 2.3. Mixing of (M(I),\hat f) implies transitivity by the elementary measure-neighborhood argument of Proposition 2.2. No step in this chain assumes its conclusion or defines a quantity in terms of the predicted quantity. Theorem 2.3(ii) supplies an explicit counterexample with a direct ball-separation proof, so the odd-period hypothesis is not smuggled in through the conclusion. The non-autonomous weak-mixing example (Theorem 2.7) imports Theorem 6 of Balibrea--Oprocha [1], an external source, and the zero-entropy example (Theorem 2.6(ii)) imports Theorem 12 of [1] and Theorem E of [12], also external. The only self-citations, [24] and [25], appear in a general introductory sentence about applications of non-autonomous systems and carry no load in any proof. There is an expositional imprecision in the proof of Theorem 2.3(i), where the sentence 'transitivity of (I,f) is equivalent to mixing of (I,f) by Theorem 2.20 in [23]' does not repeat the odd-period qualifier, and a minor misreference to 'Lemma 3.7' in Theorem 3.3 (the intended statement is Lemma 3.2); these are proof-reading issues, not circular reductions. No fitted parameters are renamed as predictions, no uniqueness theorem is imported from the authors' own prior work, and no ansatz is smuggled in via self-citation. The derivation is therefore self-contained against external benchmarks and receives no circularity score.
Assumptions & free parameters
assumptions (7)
- domain assumption Standard compact metric space theory: M(X) with Prohorov metric is compact and \hat f_n is continuous whenever X is compact and f_n continuous.
- standard math Ruette [23, Theorem 2.20]: for continuous interval maps, transitivity together with a periodic point of odd period different from 1 implies mixing, and total transitivity is equivalent to mixing.
- standard math Kolyada-Snoha [12, Theorem E]: if a non-autonomous sequence f_n converges uniformly to f, then the topological entropy of the sequence is at most h(f).
- standard math Balibrea-Oprocha [1, Theorem 6]: the non-autonomous system on S^1 built from all finite compositions of the rotation R and the map T and their inverses, in alternating order, is weakly mixing of order 2.
- standard math Kurka [14, Proposition 2.6]: a chain transitive continuous self-map of a compact metric space is surjective.
- standard math Moothathu [19, Theorem 2]: sensitivity of an interval map implies cofinite sensitivity.
- standard math Walters [28, Theorem 6.1]: any Borel probability measure on a compact metric space is regular, so a Borel set of positive measure contains a closed subset of arbitrarily close measure.
Cite this review
Pith. "Pith review of Dynamics of continuous maps induced on the space of probability measures." pith.science (2026). https://pith.science/paper/XOSBUO6Z
@misc{pith2026190807676,
author = {Pith},
title = {Pith review of: Dynamics of continuous maps induced on the space of probability measures},
year = {2026},
howpublished = {\url{https://pith.science/paper/XOSBUO6Z}},
note = {Machine review of arXiv:1908.07676}
}
abstract
For a continuous self-map $f$ on a compact interval $I$ and the induced map $\hat f$ on the space $\mathcal{M}(I)$ of probability measures, we obtain a sharp condition to guarantee that $(I,f)$ is transitive if and only if $(\mathcal{M}(I),\hat f)$ is transitive. We also show that the sensitivity of $(I,f)$ is equivalent to that of $(\mathcal{M}(I),\hat f)$. We prove that $(\mathcal{M}(I),\hat f)$ must have infinite topological entropy for any transitive system $(I,f)$, while there exists a transitive non-autonomous system $(I,f_{0,\infty})$ such that $(\mathcal{M}(I),\hat f_{0,\infty})$ has zero topological entropy, where $f_{0,\infty}=\{f_n\}_{n=0}^\infty$ is a sequence of continuous self-maps on $I$. For a continuous self-map $f$ on a general compact metric space $X$, we show that chain transitivity of $(X, f)$ implies chain mixing of $(\mathcal{M}(X),\hat f)$, and we provide two counterexamples to demonstrate that the converse is not true. We confirm that shadowing of $(X,f)$ is not inherited by $(\mathcal{M}(X),\hat f)$ in general. For a non-autonomous system $(X,f_{0,\infty})$, we prove that if $(\mathcal{M}(X),\hat{f}_{0,\infty})$ is weak mixing of order $n$, then so is $(X,f_{0,\infty})$ for any $n\geq2$; while there exists $(X,f_{0,\infty})$ such that it is weak mixing of order $2$ but $(\mathcal{M} (X),\hat{f}_{0,\infty})$ is not. We then prove that Li-Yorke chaos (resp., distributional chaos) of $(X,f_{0,\infty})$ carries over to $(\mathcal{M}(X),\hat f_{0,\infty})$, and give an example to show that $(X,f)$ and $(\mathcal{M}(X),\hat f)$ may have no Li-Yorke pair simultaneously. We also prove that if $f_n$ is surjective for all $n\geq 0$, then chain mixing of $(\mathcal{M}(X),\hat f_{0,\infty})$ always holds true, and shadowing of $(\mathcal{M}(X),\hat f_{0,\infty})$ implies mixing of $(X, f_{0,\infty})$.
Figures
Reference graph
Works this paper leans on
-
[23]
Ruette, Chaos on the interval, University Lecture Se ries, 67
S. Ruette, Chaos on the interval, University Lecture Se ries, 67. American Mathematical Society, Provi- dence, RI, 2017. xii+215 pp
work page 2017
-
[1]
F. Balibrea, P. Oprocha, Weak mixing and chaos in nonauto nomous discrete systems, Appl. Math. Lett. 25 (2012), 1135–1141
work page 2012
- [2]
-
[3]
N. C. Bernardes Jr., U. B. Darji, Graph theoretic structu re of maps of the Cantor space, Adv. Math. 231 (2012), 1655–1680
work page 2012
-
[4]
N. C. Bernardes Jr., R. M. Vermersch, Hyperspace dynamic s of generic maps of the Cantor space, Canad. J. Math. 67 (2015), 330–349
work page 2015
-
[5]
N. C. Bernardes Jr., R. M. Vermersch, On the dynamics of in duced maps on the space of probability measures, Trans. Amer. Math. Soc. 368 (2016), 7703–7725
work page 2016
-
[6]
F. Blanchard, E. Glasner, S. Kolyada, A. Maass, On Li-Yor ke pairs, J. Reine Angew. Math. 547 (2002), 51–68
work page 2002
- [7]
Show all 29 references
-
[8]
Fern´ andez, C
L. Fern´ andez, C. Good, Shadowing for induced maps of hyp erspaces, Fund. Math. 235 (2016), 277–286
2016
-
[9]
Fern´ andez, C
L. Fern´ andez, C. Good, M. Puljiz, A. Ramirez, Chain tran sitivity in hyperspaces, Chaos Solitons Fractals 81 (2015), part A, 83–90
2015
-
[10]
Glasner, B
E. Glasner, B. Weiss, Quasi-factors of zero entropy sys tems, J. Amer. Math. Soc. 8 (1995), 665–686
1995
-
[11]
J. L. G. Guirao, D. Kwietniak, M. Lampart, P. Oprocha, A. Peris, Chaos on hyperspaces, Nonlinear Anal. 71 (2009), 1–8
2009
-
[12]
Kolyada, L
S. Kolyada, L. Snoha, Topological entropy of non-auton onous dynamical systems, Random Comp. Dyn. 4 (1996), 205–233
1996
-
[13]
Komuro, The pseudo orbit tracing properties on the sp ace of probability measures, Tokyo J
M. Komuro, The pseudo orbit tracing properties on the sp ace of probability measures, Tokyo J. Math. 7 (1984), 461–468
1984
-
[14]
Kurka, Topological and symbolic dynamics, Cours sp´ ecialis´ es [specialized courses], Soci´ et´ e math´ ematique de France, Paris, 11, 2013
P. Kurka, Topological and symbolic dynamics, Cours sp´ ecialis´ es [specialized courses], Soci´ et´ e math´ ematique de France, Paris, 11, 2013
2013
-
[15]
Kwietniak, P
D. Kwietniak, P. Oprocha, A note on the average shadowin g property for expansive maps, Topology Appl. 159 (2012), 19–27
2012
-
[16]
H. Liu, E. Shi, G. Liao, Sensitivity of set-valued discr ete systems, Nonlinear Anal. 71 (2009), 6122–6125. 26 HUA SHAO, HAO ZHU, AND GUANRONG CHEN
2009
-
[17]
J. Li, K. Yan, X. Ye, Recurrence properties and disjoint ness on the induced spaces, Discrete Contin. Dyn. Syst. 35 (2015), 1059–1073
2015
-
[18]
J. Li, P. Oprocha, X. Wu, Furstenberg families, sensiti vity and the space of probability measures, Non- linearity 30 (2017), 987–1005
2017
-
[19]
Moothathu, Stronger forms of sensitivity for dynami cal systems, Nonlinearity 20 (2007), 2115–2126
T. Moothathu, Stronger forms of sensitivity for dynami cal systems, Nonlinearity 20 (2007), 2115–2126
2007
-
[20]
K. P. Parthasarathy, Probability Measures On Metric Sp aces, Academic Press, New York and London, 1967
1967
-
[21]
Peris, Set-valued discrete chaos, Chaos Solitons Fr actals 26 (2005), 19–23
A. Peris, Set-valued discrete chaos, Chaos Solitons Fr actals 26 (2005), 19–23
2005
-
[22]
Rom´ an-Flores, A note on transitivity in set-valued discrete systems, Chaos Solitons Fractals 17 (2003), 99–104
H. Rom´ an-Flores, A note on transitivity in set-valued discrete systems, Chaos Solitons Fractals 17 (2003), 99–104
2003
-
[24]
H. Shao, H. Zhu, Chaos in non-autonomous discrete syste ms and their induced set-valued systems, Chaos 29 (2019), 033117
2019
-
[25]
Y. Shi, G. Chen, Chaos of time-varying discrete dynamic al systems, J. Differ. Equ. Appl. 15 (2009), 429–449
2009
-
[26]
ˇSotola, Relationship between Li-Yorke chaos and positive t opological sequence entropy in nonau- tonomous dynamical systems, Discrete Contin
J. ˇSotola, Relationship between Li-Yorke chaos and positive t opological sequence entropy in nonau- tonomous dynamical systems, Discrete Contin. Dyn. Syst. 38 (2018), 5119–5128
2018
-
[27]
Strassen, The existence of probability measures wit h given marginals, Ann
V. Strassen, The existence of probability measures wit h given marginals, Ann. Math. Statist. 36 (1965), 423–439
1965
-
[28]
Walters, An Introduction to Ergodic Theory, Springe r-Verlag, NewYork, 1982
P. Walters, An Introduction to Ergodic Theory, Springe r-Verlag, NewYork, 1982
1982
-
[29]
Wu, Chaos of transformations induced onto the space o f probability measures, Int
X. Wu, Chaos of transformations induced onto the space o f probability measures, Int. J. Bifurcat. Chaos 26 (2016), 1650227 (12 p.). Department of Mathematics, Nanjing University of Aeronaut ics and Astronautics, 211106, Nanjing, P. R. China E-mail address : huashao@nuaa.edu.c...
2016
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.