Dyer groups have the falsification by fellow-traveller property
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.
Ada: I'm Ada.
Emmy: And I'm Emmy, here with Kolya.
Ada: We're looking at the paper: "Dyer groups have the falsification by fellow-traveller property".
Emmy: Here's the gist of the paper. Dyer groups have been shown to possess the falsification by fellow-traveller property with respect to a specific generating set, which implies that they form a Cannon pair and have finitely many cone types The main result of this paper is that every Dyer group admits a finite generating set with respect to which it satisfies the FFTP (Theorem 3.2) and hence, forms a Cannon pair (Corollary 3.9).
Ada: Introduction to Dyer Groups and Related Concepts Dyer groups form a class of groups that unify the study of Coxeter groups and graph products of cyclic groups, defined by vertex- and/or edge-labelled graphs They are instances of periagroups
16, 17: and share fundamental properties such as the rationality of their growth series
30: and a common solution to the word problem
29: .
Emmy: The notion of cone types is introduced by Cannon in the 80s, where it corresponds to the set of elements which extend it geodesically.
Ada: The Falsification by Fellow-Traveller Property (FFTP) The falsification by fellow-traveller property (FFTP) was initially introduced in the 90s to approach the problem of determining which pairs of groups together with a finite generating set admit a regular language of geodesics.
Paper summary: Ada: We just touched on how this paper shows Dyer groups have a finite generating set where they satisfy the falsification by fellow-traveller property >
Emmy: It's important to remember what this property actually means in the context of group theory. It’s about whether there is a regular language of geodesics on the Cayley graph, and this property is strictly stronger than just having some kind of regularity >
Kolya: So, if they prove that Dyer groups have this FFTP with respect to a generating set X S, what does that actually tell us about the group's growth or its geometry? Is it just a technicality, or does it constrain the shape of the group structure itself >
Ada: It constrains it. Because as we saw, this property implies that Dyer groups have finitely many cone types >
Emmy: Which is a big deal because cone types relate to how elements extend geodesically in the Cayley graph, and having finitely many of them gives us a very controlled geometric structure >
Kolya: So for someone interested in the probabilistic side, what does that mean about random walks on these groups? Does having finite cone types put an upper bound on the spectral radius of those walks >
Ada: The paper mentions computing the spectral radius of the simple random walk on its Cayley graph eighteen twenty-two twenty-four twenty-five, and this property is used in that context > <ref:2602.18220#pg2,the spectral radius of the simple random walk on its Cayley graph>
Emmy: It suggests that for these groups, the upper and lower estimates for the spectral radius of those random walks are more tightly controlled than we might expect from just looking at their general growth series nineteen twenty-two > <ref:2602.18220#pg2,upper and lower estimates for the spectral radius of>
Kolya: So what's the catch here? The paper constructs this specific set X S by modifying a standard generating set and interpolating between it and these so-called syllables; how does that modification affect the resulting bounds on the spectral radius >
Ada: Well, they include elements like x alpha v where f(v) is finite, along with x, plus one minus one v where the function of v is infinite >
Conclusion: Ada: So to wrap up, this paper by Megan Howarth is showing that Dyer groups, despite their complexity arising from being graph products of cyclic groups and Coxeter systems, still have a very nice geometric structure when you choose the right generating set >
Emmy: The main point is that finding this specific finite generating set makes the group behave much more nicely geometrically because it guarantees finitely many cone types >
Kolya: So for someone listening who doesn't know much about group theory, what does this mean practically? It just means that if you are studying a Dyer group, you can use tools from geometric group theory with a finite set of building blocks instead of dealing with the full complexity of the infinite presentation directly >
Ada: Exactly. It gives us a concrete way to study them that has strong geometric guarantees, and it connects this whole class of groups back to well-studied examples like Coxeter groups >
Emmy: The paper is important because it provides a specific framework—the falsification by fellow-traveller property—that lets us connect the abstract algebraic definition of these groups to concrete, quantifiable geometric properties >
University of Geneva
math.GR
Submitted: 2026-02-20
Updated: 2026-11-05
Comments: 33 pages, 2 figures. Comments are welcome!
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 83/100
The gist: Dyer groups have been shown to possess the falsification by fellow-traveller property with respect to a specific generating set…
Key concepts
- Dyer Groups
- These are groups formed by vertex- and/or edge-labelled graphs. They serve as a unifying structure for studying Coxeter groups and graph products of cyclic groups, possessing properties like rational growth series and a common solution to the word problem.
- Falsification by Fellow-Traveller Property (FFTP)
- This property relates to whether a group has a regular language of geodesics. Satisfying FFTP is much stronger than having finitely many cone types and implies that the group has only finitely many cone types with respect to a suitable generating set.
- Cone Types
- Introduced by Cannon, cone types represent the set of elements that extend a given geodesic path indefinitely. Having finitely many cone types means the group has a finite structure related to its geometric properties.
Terminology
Summary
Dyer groups have been shown to possess the falsification by fellow-traveller property with respect to a specific generating set, which implies that they form a Cannon pair and have finitely many cone types The main result of this paper is that every Dyer group admits a finite generating set with respect to which it satisfies the FFTP (Theorem 3. p. 12) and hence, forms a Cannon pair (Corollary 3. p. 19).
Introduction to Dyer Groups and Related Concepts
Dyer groups form a class of groups that unify the study of Coxeter groups and graph products of cyclic groups, defined by vertex- and/or edge-labelled graphs They are instances of periagroups [16, 17] and share fundamental properties such as the rationality of their growth series [30] and a common solution to the word problem [29]. The notion of cone types is introduced by Cannon in the 80s, where it corresponds to the set of elements which extend it geodesically. p. 0
The Falsification by Fellow-Traveller Property (FFTP)
The falsification by fellow-traveller property (FFTP) was initially introduced in the 90s to approach the problem of determining which pairs of groups together with a finite generating set admit a regular language of geodesics. The FFTP implies the finiteness of the cone types and is strictly stronger [12, 27]. p. 1 Among groups known to satisfy this property with respect to a suitable generating set are geometrically finite hyperbolic groups, virtually abelian groups, Coxeter groups, Artin groups of large type, dihedral Artin groups and Garside groups. p. 1
Construction of the Finite Generating Set XS
The paper constructs a specific finite generating set XS to demonstrate the FFTP for Dyer systems. The set XS is obtained by modifying the standard one and interpolates between it and the so-called set of syllables (Definition 2. p. 13). This set XS includes elements of the form x α v where f(v) < ∞, along with x ±1 v where f(v) = ∞. p. 0
Proof of FFTP via Mediangle Graphs
The proof establishes that Cay(D, XS) is a locally finite mediangle graph (Proposition 3.1). The verification of the four axioms of Definition 2. p. 66 is shown through inductive arguments based on the structure of Cay(D, S(X)) and comparisons between Cay(D, XS) and Cay(D, S(X)). p. 8 The proof also details how path transformations (T1, T2, T3) shorten non-geodesic paths such that they become k-fellow travelling with a constant related to M = max of m(e). p. 5
Conclusion and Implications
The combination of the lemmas leads to Theorem 3.2, proving that (D, XS) has the falsification by fellow-traveller property with constant 2M. p. 5 This result implies that (D, XS) has finitely many cone types and thus is a Cannon pair. p. 4 The paper concludes that Dyer groups have finitely many cone types with respect to the set XS. p. 0 It is noted that the FFTP is sensitive to the choice of generating set, and it is an open problem whether Dyer groups have finitely many cone types with respect to all generating sets. p. 4 The authors also suggest that one can build on XS to exhibit infinitely many generating sets for which a Dyer group has the FFTP. p. 0
References
[1] Y. Antol´ın, Counting subgraphs in FFTP graphs with symmetry, Math. p. 6 Proc. Cambridge Philos. Soc., 170(2):327–353, 2021
[2] Y. Antol´ın and L. p. 6 Ciobanu, Finite generating sets of relatively hyperbolic groups and applications to geodesic languages, Trans. p. 0 Amer. Math. Soc., 368(11):7965–8010, 2016
[3] B. Brink and R. B. Howlett, A finiteness property and an automatic structure for Coxeter groups, Math. p. 0 Ann., 296(1):179–190, 1993
[4] J. W. Cannon, The growth of the closed surface groups and the compact hyperbolic Coxeter groups, Geom. p. 0 Dedicata, 16(2):123–148, 1984
[5] J. W. Cannon, The combinatorial structure of cocompact discrete hyperbolic groups, Geom. p. 0 Dedicata, 16(2):123–148, 1984
[6] J. W. Cannon and P. p. 0 Wagreich, Growth functions of surface groups, Mathematische Annalen, 293(1):239–257, 1992
[7] R. Charney and J. Meier, The language of geodesics for Garside groups, Math. p. 0 Z., 248(3):495–509, 2004
[8] L. Ciobanu and G. Crowe, Conjugacy geodesics and growth in dihedral Artin groups, 2024
[9] S. Cleary, M. Elder, and J. p. 5 Taback, Cone types and geodesic languages for lamplighter groups and Thompson’s group F, J. p. 4 Algebra, 303(2):476–500, 2006
[10] M. W. Davis, The Geometry and Topology of Coxeter Groups, Springer Monographs in Mathematics. p. 0 Springer Cham, 2025
[11] M. Dyer, Reflection subgroups of Coxeter systems, Journal of Algebra, 135(1):57–73, 1990
[12] M. Elder, Regular geodesic languages and the falsification by fellow traveler property. p. 5 Algebr. Geom. Topol., 5:129–134, 2005
[13] M. J. Elder, Finiteness and the falsification by fellow traveler property, In Proceedings of the Conference on Geometric and Combinatorial Group Theory, Part II (Haifa, 2000), volume 95, pages 103–113, 2002
[14] D. B. A. Epstein et al., Word processing in groups, Jones and Bartlett Publishers, Boston, MA, 1992
[15] A. Genevois, Groups acting on quasi-median graphs, An introduction, S´eminaire de th´eorie spectrale et g´eom´etrie, 35:43–68, 2021
[16] A. Genevois, Rotation groups, mediangle graphs, and periagroups: a unified point of view on Coxeter groups and graph products of groups, 2022
[17] A. Genevois, Rotation groups virtually embed into right-angled rotation groups, Canadian Journal of Mathematics, pages 1–25, 2025
[18] S. Gouezel, A Numerical Lower Bound for the Spectral Radius of Random Walks on Surface Groups, Combinatorics, Probability and Computing, 24(6):838–856, 2015
[19] R. Grigorchuk and C. Kravaris, On the growth of the wallpaper groups, Acta Crystallographica Section A Foundations and Advances, 78(4):371–383, 2022
[20] D. F. Holt, Garside groups have the falsification by fellow-traveller property, Groups Geom. p. 0 Dyn., 4(4):777–784, 2010
[21] D. F. Holt and S. Rees, Artin groups of large type are shortlex automatic with regular geodesics, Proceedings of the London Mathematical Society, 104(3):486–512, 2012
[22] M. Howarth and T. p. 0 Nagnibeda, Cone types and spectral radius of hyperbolic triangle groups and hyperbolic tessellations, Groups Geom. p. 0 Dyn., 19(2):477–493, 2025
[23] J. Loeffler, J. Meier, and J. Worthington, Graph products and Cannon pairs, Internat. p. 0 J. Algebra Comput., 12(6):747–754, 2002
[24] T. Nagnibeda, An estimate from above of spectral radii of random walks on surface groups, Journal of Mathematical Sciences, 96(5):3542–3549, 1999
[25] T. Nagnibeda, Random walks, spectral radii and Ramanujan graphs, In Random walks and geometry, pages 487–500. p. 1 Walter de Gruyter, Berlin, 2004
[26] W. D. Neumann and M. p. 1 p. 1 Shapiro, Automatic structures, rational growth, and geometrically finite hyperbolic groups, Invent. p. 1 Math., 120(2):259–287, 1995
[27] W. D. Neumann and M. Shapiro, A short course in geometric group theory, Notes for the ANU Workshop, January/February 1996
[28] G. A. Noskov, Bounded shortening in Coxeter complexes and buildings, In Mathematical structures and modeling, No. p. 0 8 (Russian), pages 10–14. Omsk. Gos. Univ., Omsk, 2001
[29] L. Paris and M. Soergel, Word problem and parabolic subgroups in Dyer groups, Bull. p. 0 Lond. Math Soc., 55(6):2928–2947, 2023
[30] L. Paris and O. Varghese, The spherical growth series of Dyer groups, Proc. p. 0 Edinb. Math. Soc., (2), 67(1):168–187, 2024
[31] J. Parkinson and Y. Yau, Cone types, automata, and regular partitions in Coxeter groups, Adv. p. 0 Math., 398:Paper No. p. 0 108146, 66, 2022
How the training signal is built
The proof of Theorem 3.2 proceeds by combining Lemmas 3.6, 3.7 and 3.8 to obtain the result. p. 1 This involves analyzing how non-geodesic paths can be shortened using three operations: (T1) Removing a backtrack, (T2) Shortening a triangle, and (T3) Applying a flip. Lemma 3.4 shows that these transformations reduce the fellow-travelling constant for non-geodesic paths. p. 5
Where the gains come from
The core of the proof lies in showing that minimal non-geodesic paths can be shortened by a path α such that γ and α M-fellow travel with constant ≤ 2. This is achieved by analyzing minimal paths based on whether their endpoints are separated by distance n-1 or n-2. Lemma 3.7 and Lemma 3. p. 48 provide the necessary machinery to construct these shorter paths using convex even cycles and 3-cycles. p. 3 The final result is that (D, XS) has finitely many cone types and thus is a Cannon pair. p. 4 It is also noted that one can build on XS to exhibit infinitely many generating sets for which a Dyer group has the FFTP. p. 0
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Ability to determine if a given group is a Dyer group by analyzing its presentation data (G, m, f). This involves checking for specific relation structures derived from vertex and edge labeling maps as defined in Definition 2.1.
-
Capability to construct a finite generating set XS for any given Dyer group D using the algorithm described in Definition 2.3, enabling the application of powerful geometric tools like mediangle graphs to analyze its Cayley graph Cay(D, XS).
-
Functionality to verify if a pair of groups and a finite generating set admit a regular language of geodesics by checking if they form a Cannon pair, leveraging the proof that (D, XS) has this property (Corollary 3.9).
-
Implementation of path shortening algorithms using transformations T1, T2, and T3 to find shorter paths for non-geodesic paths in the Cayley graph Cay(D, XS), as detailed in Definition 3.3 and Lemma 3.4.
-
Detection of whether a non-geodesic path in Cay(D, XS) can be M-fellow traveled by a shorter path with constant M, based on the analysis derived from Lemma 3.6 and Lemma 3.7 concerning hyperplanes H and cycle conditions.
Sources
- Conjugacy geodesics and growth in dihedral Artin groups
- Rotation groups, mediangle graphs, and periagroups: a unified point of view on Coxeter groups and graph products of groups
Related papers
- Idempotence criteria for lazy cellular automata
- From algebraic orthogonality to RAAG embedding obstructions in hierarchically hyperbolic groups
- A finitely generated Linear group without property (T) that is neither inner amenable nor properly proximal
- From almost tight p-group enumeration to 2-group domination among finite groups
- Character Nonvanishing under Coprime Actions
- Uncountably many 2-spherical groups of Kac-Moody type of rank 3 over F 2