Pith. sign in

REVIEW 2 major objections 3 minor 27 references

There is only one Farey map

T0 review · 2 major / 3 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read For every dimension $n$, up to relabeling and permuting vertices, exactly three simplex-splitting two-map iterated function systems are contractive, and only one of them—the Farey–Mönkemeyer map—has a continuous coding map.

desk verdict A likely correct and important classification of contractive two-symbol IFS; the proof's endgame relies on unchecked 'direct inspection' for arbitrary n and needs tightening. read the letter →

arxiv 2506.02984 v1 pith:JUXIJHAL submitted 2025-06-03 math.DS math.NT

Giovanni Panti This is my paper
classification math.DSmath.NT MSC 11J7037M25
keywords continuedfractionsmultidimensionaliteratedfunctionsystemssimplex-splittingFareymapGauss-typetopologicalcontractivitynonnegativematrices
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 a classification result for two-symbol continued fraction algorithms in every dimension. It considers all ways of splitting the standard $n$-dimensional simplex $\Delta$ into two smaller unimodular simplexes using nonnegative integer matrices, and asks when the resulting iterated function system is contractive: for every infinite sequence of choices, the nested images shrink to a single point. The answer, up to relabeling the two pieces and permuting the vertices, is that exactly three such systems exist for every $n=1,2,\dots$: the Mönkemeyer algorithm and its two orientation variants. The point of the result is that contractive systems are exactly the algorithms that assign distinct expansions to distinct points, so the theorem says there are precisely three such two-symbol continued fraction algorithms in any dimension. Only one of the three, the Farey–Mönkemeyer map, has a continuous coding map.

What carries the argument

The load-bearing object is the incomplete incidence graph $G^-$ of the pair $(A_0,A_1)$: its nodes are the vertices of the simplex, a directed edge $j \leftarrow i$ records that the matrix sends vertex $e_i$ to $e_j$, and the two vertices that map to the shared Farey midpoint $(e_0+e_1)/2$ are left without a full edge. Lemma 3.3 shows that contractivity forbids two loops for the same word at distinct nodes and two loops for incomparable words at the same node; these two forbidden patterns, together with the Protasov–Voynov theorem used in Lemma 3.2, force the pair of graphs into exactly three chain shapes. Those shapes determine the matrices, yielding the three contractive pairs.

What would settle it

Run a computer search over all orbit representatives of pairs $(P_0,P_1)\in S_{n+1}^2$ for $n=4$ (there are 1283 orbits according to the paper's count) and check contractivity directly by computing whether every intersection $\bigcap_{t\ge0} A_{a\upharpoonright t}[\Delta]$ is a singleton; any contractive orbit not equivalent to $(M_0,M_1)$, $(M_0,M_1F)$, or $(M_0F,M_1)$ would refute Theorem 2.1.

Watch

Extended reading notes

Core claim

The central claim is that, after the natural action of the symmetric group on the vertices of $\Delta$, the only contractive simplex-splitting pairs $(A_0,A_1)$ with $A_0,A_1 \in GL(n+1,\mathbb{Z})$ nonnegative are the three pairs $(M_0,M_1)$, $(M_0,M_1F)$, and $(M_0F,M_1)$. Contractivity here means that for every infinite word $a$, the intersection $\bigcap_{t\ge0} A_{a\upharpoonright t}[\Delta]$ is a singleton. The paper proves this by showing that the incomplete incidence graphs of the two matrices must avoid two forbidden patterns—two loops for the same word at distinct nodes, or two loops for incomparable words at the same node—and that these exclusions force the graphs into one of three chain configurations. The three configurations correspond exactly to the three pairs above, and all three are contractive because their time-$t$ partitions coincide with the Mönkemeyer partition. Exactly one of the three, the Mönkemeyer pair, gives a continuous Gauss-type map $G:\Delta\to\Delta$.

Load-bearing premise

The proof depends on an external theorem saying that any semigroup of nonnegative matrices whose elements all have spectral radius 1 can be simultaneously block-triangularized with finite irreducible diagonal blocks; if that theorem does not apply to the monoid generated by a contractive simplex-splitting pair, then Lemma 3.2, the graph restrictions, and the whole classification collapse.

Editorial extensions

If this is right

  • In every dimension $n$, any simplex-splitting two-symbol IFS that is contractive must be equivalent to one of the three pairs $(M_0,M_1)$, $(M_0,M_1F)$, $(M_0F,M_1)$; no other such algorithm exists.
  • For all three contractive pairs, the time-$t$ partitions $\{\Delta_w : |w|=t\}$ are identical, so the geometric coding of points is the same; the algorithms differ only in which branch is called $0$ and which is called $1$, and in the orientation of the pieces.
  • The Gauss-type map $G$ is continuous in exactly one of the three cases, namely the Mönkemeyer (Farey) algorithm; in the other two cases the two branches disagree across the common face.
  • Because contractive simplex-splitting systems give every point a coding and distinct points distinct codings, the theorem identifies, in every dimension, the complete list of two-symbol continued fraction algorithms with unique expansions.

Reading between the lines

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

  • What the author leaves implicit is that, since the three algorithms share the same cylinders, any observable that depends only on the cylinder partition—cylinder diameters, symbolic complexity, the combinatorics of the coding tree—is identical across all three; only the labeling or orientation of the two branches changes.
  • The same forbidden-pattern technique could be pushed to $m$-symbol simplex-splitting IFS for $m>2$; the paper's Rauzy-gasket example suggests non-contractive behavior is typical, so one would expect a short finite list of contractive systems in each dimension.
  • A direct test of the classification: because the continuous case is unique, one could try to prove, without the full graph machinery, that any two-symbol algorithm with a continuous Gauss-type map must be the Farey–Mönkemeyer map by checking the orientation-coherence condition used in Section 3.
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.

Referee Report

2 major / 3 minor

Summary. The paper classifies, for every dimension n, all pairs (A0,A1) of nonnegative unimodular integer matrices that split the standard n-simplex into two subsimplexes and whose associated projective iterated function system is topologically contractive. Working up to the symmetric group action, it claims that there are exactly three such pairs: the Mönkemeyer pair (M0,M1) and the two variants (M0,M1F) and (M0F,M1). It further shows that only the Mönkemeyer pair gives a continuous Gauss-type map. The proof reduces the classification to a graph-combinatorial statement (Claim 3.1) and uses a theorem of Protasov and Voynov on matrix semigroups with constant spectral radius.

Significance. If correct, the result is a striking uniqueness theorem: in every dimension there are only three two-symbol continued fraction algorithms with unique expansions, and only one of them is continuous. The reduction of the problem to graph data is elegant, and the use of the Protasov–Voynov theorem is natural. The three candidate families are known to be contractive from earlier work, so the paper's contribution is the exclusion of all other possibilities. The main structural lemmas (Lemmas 3.2 and 3.3) are well motivated and appear sound. However, the proof as written relies on two 'direct inspection' assertions for arbitrary n that are load-bearing and are not fully demonstrated.

major comments (2)
  1. [§3, final paragraph] The exclusion of a single short backtrack i→i+1 is dismissed by 'direct inspection shows that ... the incidence graph of A10 would then be disconnected with an isolated loop at i'. This step is load-bearing: it is what yields ki=i for all i and hence identifies the parabolic and hyperbolic cases with M1 and M1F. The check is finite but n-dependent, and no general-n argument is given. Please replace this with an explicit description of the incidence graph of A10 for arbitrary n and a proof of the asserted disconnectedness, or otherwise supply a rigorous argument. As written, the proof leaves open the possibility that an unclassified contractive pair exists for some n.
  2. [§3, Lemma 3.5, last paragraph] In the (0h),(1p) subcase with k2=2, the contradiction uses the assertion that 'by direct inspection' the incidence graph of A01 has a loop of length n shortened to one of length n/2, and consequently A01^{n/2} has a singleton and several 2×2 diagonal blocks, spectral radius >1, and a fixed vertex e0. This is a general-n assertion that is not proved. Please provide an explicit construction of the graph for arbitrary even n, or a proof by induction, since this step is essential for eliminating the (0h) possibility.
minor comments (3)
  1. [§3, Lemma 3.5] The sentence 'If k2 ≠ 2, then for certain r, s both k20r0 and 21sk2 occur' is hard to parse; please write the words with explicit exponents and parentheses, and state explicitly which incomparable loops result.
  2. [§3, Lemma 3.2] The construction of the positive vector x in the proof of Lemma 3.2 is sketched in a parenthetical remark; since it is an essential step, it would be clearer to present it as a numbered sub-argument.
  3. [§3, Example 3.4] The diagrams for n=4 are helpful, but the 'obvious tail' for larger n is not immediately verifiable; a short formal description of the edge sets for the three pairs (M0,M1), (M0,M1F), and (M0F,M1) would make the example more rigorous.

Circularity Check

0 steps flagged · score 2.0 of 10

No substantive circularity: the uniqueness classification is derived from graph-spectral constraints, not from its own conclusion; self-citations are auxiliary and the unexpanded 'direct inspection' steps are proof gaps, not circular reductions.

full rationale

The classification proof is self-contained against the claimed conclusion. Claim 3.1 starts from an arbitrary contractive simplex-splitting pair (A0,A1)=(NP0,LP1) and, via Lemma 3.2 (which imports the external Protasov-Voynov theorem [24]) and Lemma 3.3, forces the incomplete incidence graphs into one of three shapes. None of these inputs contains the conclusion that the Monkemeyer pair and its two variants are the only contractive pairs. Contractivity of the three exhibited pairs is imported from Monkemeyer [20] and Schweiger [25], both external to the author, and extended to (M0,M1F) and (M0F,M1) by the explicit identity of their time-t partitions with those of (M0,M1); this is independent support, not circularity. The self-citations [22] and [23] occur only in Remark 1.2 and in historical remarks; they are not used in the proof of Theorem 2.1 or Claim 3.1. The only soft spots are the two 'direct inspection' assertions in the proof of Lemma 3.5 and at the end of Claim 3.1, concerning the incidence graphs of A01 and A10 for arbitrary n. These are unexpanded n-dependent graph checks and are correctly described by the skeptic as proof gaps, but they are not circular: they do not assume the conclusion or reduce it to an input by construction. A possible error there would be a correctness flaw, not a circularity.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The proof relies on standard nonnegative matrix theory and one specialized external theorem (Protasov-Voynov) as a black box. No free parameters are fitted, and no new entities are postulated.

assumptions (4)
  • standard math Protasov-Voynov theorem: a semigroup of matrices with all spectral radii equal to 1 can be block-triangularized with finite irreducible diagonal blocks
    Invoked in Lemma 3.2 to derive a contradiction from a monoid with spectral radius 1; located in Section 3, proof of Lemma 3.2.
  • standard math Perron-Frobenius theorem for nonnegative matrices
    Used to assert that spectral radius is an eigenvalue with a nonnegative eigenvector in Lemma 3.2.
  • standard math Monkemeyer's theorem that the Farey-Monkemeyer IFS is contractive in every dimension
    Quoted from [20] and [25] to prove that the three listed IFS are contractive; Section 2, before Theorem 2.1.
  • standard math Hutchinson operator fixed-point theory for contractive IFS
    Background used in Remark 1.2(6) and Example 1.3; not load-bearing for the main classification.

how reviews work

0 comments
Cite this review

Pith. "Pith review of There is only one Farey map." pith.science (2026). https://pith.science/paper/JUXIJHAL

@misc{pith2026250602984,
  author       = {Pith},
  title        = {Pith review of: There is only one Farey map},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JUXIJHAL}},
  note         = {Machine review of arXiv:2506.02984}
}
read the original abstract

Let A_0, A_1 be nonnegative matrices in GL(n+1,Z) such that the subsimplexes A_0[Delta], A_1[Delta] split the standard unit n-dimensional simplex Delta in two. We prove that, for every n=1,2,... and up to the natural action of the symmetric group by conjugation, there are precisely three choices for the pair (A_0, A_1) such that the resulting projective Iterated Function System is topologically contractive. In equivalent terms, in every dimension there exist precisely three continued fraction algorithms that assign distinct two-symbol expansions to distinct points. These expansions are induced by the Gauss-type map G: Delta --> Delta with branches A_0^{-1}, A_1^{-1}, which is continuous in exactly one of these three cases, namely when it equals the Farey-Monkemeyer map.

Figures

Figures reproduced from arXiv: 2506.02984 by the authors.

Figure 1
Figure 1. The time-5 partition and the least fixed point R points in projective 2-space of the proximal elements in the monoid generated by A0, A1, A2 [4, §3.1]. It is easy to show that R is the image of the Rauzy gasket [1], [18] under an appropriate projective map. There are uncountably many incompa￾rable fixed points between ∆ and R; for example, for every x in the boundary of ∆ the sequence At{x} converges to the closure … view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

27 extracted references · 27 canonical work pages

  1. [1]

    Avila, P

    A. Avila, P. Hubert, and A. Skripchenko. On the Hausdorff dimension of the Rauzy gasket.Bull. Soc. Math. France, 144(3):539–568, 2016

    work page 2016

  2. [2]

    P. R. Baldwin. A multidimensional continued fraction and some of its statistical properties. J. Statist. Phys., 66(5-6):1463–1505, 1992

    work page 1992

  3. [3]

    Banakh, W

    T. Banakh, W. Kubi ´s, N. Novosad, M. Nowak, and F. Strobin. Contractive function systems, their attractors and metrization. Topol. Methods Nonlinear Anal., 46(2):1029–1066, 2015

    work page 2015

  4. [4]

    Benoist and J.-F

    Y. Benoist and J.-F. Quint. Random walks on reductive groups , volume 62 of Ergebnisse der Mathematik und ihrer Grenzgebiete. Springer, 2016

    work page 2016

  5. [5]

    Berman and R

    A. Berman and R. J. Plemmons. Nonnegative matrices in the mathematical sciences , volume 9 of Classics in Applied Mathematics . Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA, 1994. Revised reprint of the 1979 original

    work page 1994

  6. [6]

    C. J. Bishop and Y. Peres. Fractals in probability and analysis , volume 162 of Cambridge Studies in Advanced Mathematics. Cambridge University Press, 2017

    work page 2017

  7. [7]

    Borsuk and S

    K. Borsuk and S. Ulam. On symmetric products of topological spaces. Bull. Amer. Math. Soc., 37(12):875–882, 1931

    work page 1931

  8. [8]

    Bruin and S

    H. Bruin and S. Troubetzkoy. The Gauss map on a class of interval translation mappings.Israel J. Math., 137:125–148, 2003

    work page 2003

Show all 27 references
  1. [9]

    de la Harpe

    P. de la Harpe. Topics in geometric group theory. University of Chicago Press, 2000. 12 G. PANTI

    2000

  2. [10]

    H. Federer. Geometric measure theory , volume 153 of Die Grundlehren der mathematischen Wissenschaften. Springer, 1969

    1969

  3. [11]

    N. P. Fogg and C. Noˆ us. Symbolic coding of linear complexity for generic translations on the torus, using continued fractions. J. Mod. Dyn., 20:527–596, 2024

    2024

  4. [12]

    Fougeron and A

    C. Fougeron and A. Skripchenko. Simplicity of spectra for certain multidimensional continued fraction algorithms. Monatsh. Math., 194(4):767–787, 2021

    2021

  5. [13]

    F. R. Gantmacher. Applications of the theory of matrices . Interscience Publishers, Inc., New York, 1959

    1959

  6. [14]

    T Garrity and O. V. Osterman. On the linear complexity associated with a family of multidi- mentional continued fraction algorithms. https://arxiv.org/abs/2410.02032, 2024

    2024 arXiv

  7. [15]

    Heersink

    B. Heersink. An effective estimate for the Lebesgue measure of preimages of iterates of the Farey map. Adv. Math., 291:621–634, 2016

    2016

  8. [16]

    S. Isola. From infinite ergodic theory to number theory (and possibly back). Chaos Solitons Fractals, 44(7):467–479, 2011

    2011

  9. [17]

    S. Ito. Algorithms with mediant convergents and their metrical theory. Osaka J. Math. , 26(3):557–578, 1989

    1989

  10. [18]

    N. Jurga. Hausdorff dimension of the Rauzy gasket. https://arxiv.org/abs/2312.04999, 2023

    2023 arXiv

  11. [19]

    Kesseb ¨ohmer and B

    M. Kesseb ¨ohmer and B. O. Stratmann. A multifractal analysis for Stern-Brocot intervals, con- tinued fractions and Diophantine growth rates. J. Reine Angew. Math., 605:133–163, 2007

    2007

  12. [20]

    M ¨onkemeyer

    R. M ¨onkemeyer. ¨Uber Fareynetze in n Dimensionen. Math. Nachr., 11:321–344, 1954

    1954

  13. [21]

    Nogueira

    A. Nogueira. The three-dimensional Poincar ´e continued fraction algorithm. Israel J. Math. , 90(1-3):373–401, 1995

    1995

  14. [22]

    G. Panti. Multidimensional continued fractions and a Minkowski function. Monatshefte f ¨ur Mathematik, 154:247–264, 2008

    2008

  15. [23]

    G. Panti. Purely periodic continued fractions and graph-directed iterated function systems. Ramanujan J., 65(1):447–475, 2024

    2024

  16. [24]

    V. Yu. Protasov and A. S. Voynov. Matrix semigroups with constant spectral radius. Linear Algebra Appl., 513:376–408, 2017

    2017

  17. [25]

    Schweiger

    F. Schweiger. Multidimensional continued fractions. Oxford University Press, 2000

    2000

  18. [26]

    E. S. Selmer. On the irreducibility of certain trinomials. Math. Scand., 4:287–302, 1956

    1956

  19. [27]

    E. S. Selmer. Continued fractions in several dimensions. Nordisk Nat. Tidskr. , 9:37–43, 95, 1961. Department of Mathematics, Computer Science and Physics, University of Udine, via delle Scienze 206, 33100 Udine, Italy Email address: giovanni.panti@uniud.it

    1961

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.