WorksMathematics
The Auslander–Reiten Conjecture for Igusa–Todorov Algebras
Abstract
We prove that every Igusa–Todorov Artin algebra satisfies the Auslander–Reiten conjecture. More precisely, let be an -Igusa–Todorov witness, and let be the number of isomorphism classes of nonprojective indecomposable summands of . If a finitely generated module satisfies \Ext_A^i(M,A)=0 for every and \Ext_A^q(M,M)=0 for , then is projective. As applications, algebras of representation dimension at most three and algebras satisfying for which has finite representation type satisfy the Auslander–Reiten conjecture.
Introduction
Let be an Artin algebra over a commutative Artinian ring . We write -mod for the category of finitely generated left -modules, and all modules considered below belong to this category. For a module , let add denote the full subcategory of direct summands of finite direct sums of copies of . The Auslander–Reiten conjecture (ARC), formulated by Auslander and Reiten [4], asserts that
A module is self-orthogonal if for every $i>0. It is a generator if . Thus (1) is equivalent to the statement that every self-orthogonal generator is projective. The generalized Nakayama conjecture asserts that every indecomposable injective left -module occurs as a direct summand of some term in a minimal injective resolution of the left regular module . Auslander and Reiten [4] proved that the ARC holds for all Artin algebras if and only if the generalized Nakayama conjecture does. But for a given algebra, there is no such implication. The finitistic dimension conjecture, formulated by Bass [6], states that the supremum of the finite projective dimensions of modules in -mod is finite. Chen, Hu, Qin and Wang [9, Introduction] recall that its validity for all Artin algebras would imply the ARC for all Artin algebras. But this global implication does not give the corresponding implication for a fixed algebra. For further information on the interrelations among homological conjectures, see [13].
The ARC has a long history of investigation. In the representation theory setting, Auslander and Reiten [4] proved the conjecture for algebras of finite type and algebras with radical square zero. Hoshino [14] established the conjecture for symmetric Artin algebras with radical cube zero. The conjecture was further established for local Artin algebras with radical cube zero [24], special biserial algebras with radical cube zero [25], and a class of left Gorenstein algebras [5]. Moreover, the conjecture was shown to be invariant under derived equivalences [22]. For recent developments concerning the ARC for Artin algebras, we refer to Chen and Xi [10], Xia [23], and Zhang and Zhou [26]. In the commutative setting, Auslander, Ding and Solberg [3] extended the problem to commutative Noetherian rings and established its validity for complete intersection local rings. Further contributions were made by Huneke, Şega and Vraciu [16], as well as Huneke and Leuschke [15], who obtained significant vanishing results over Cohen–Macaulay and Gorenstein rings. Araya [1] showed that, for commutative Gorenstein rings, the validity of the conjecture can be reduced to the codimension-one case. For more details of the ARC for Noetherian rings, we refer to [11].
On October 6, 2026, OpenAI [18] released a preprint claiming to provide a counterexample to the ARC, in which the present paper is cited among the previously established positive results. This development further highlights the importance of identifying classes of algebras for which the ARC holds. Throughout the paper, we focus on the ARC for Artin algebras.
On the other hand, Igusa and Todorov [17] introduced their functions to study the little finitistic dimension. Wei [20 Theorem 2.3] subsequently introduced Igusa–Todorov algebras and proved that they have finite finitistic dimension. Their defining property is the existence of short presentations of syzygies by modules from one fixed additive subcategory. This class contains syzygy-finite algebras, as discussed by Wei [20] and Wei [21 Section 5]. Wei [19 Lemma 2.1] also showed that an algebra of representation dimension at most three is 0-Igusa–Todorov. In Proposition 4.2, we construct a 1-Igusa–Todorov witness under a condition on radical powers. In particular, this construction covers algebras with radical cube zero.
Further developments include the broader class of Lat-Igusa–Todorov algebras introduced by Bravo, Lanzilotta, Mendoza and Vivero [8]. Barrios, Lanzilotta and Mata [7] survey the corresponding homological functions and related classes. These results do not deduce ARC from the Igusa–Todorov property alone. In particular, finite finitistic dimension for one algebra does not imply the ARC for the algebra by any known general result.
The nonprojective support of a module is the set of isomorphism classes of its nonprojective indecomposable direct summands. We write for the cardinality of this set. Thus multiplicities do not enter . We also write
Recall that is -Igusa–Todorov if there is a fixed module such that each admits an exact sequence with , see Definition 2.1. The module is called a witness. Our first result is a finite test for projectivity.
Theorem I. (see Theorem 3.2) Let be an -Igusa–Todorov Artin algebra with witness , and put . If and
then is projective. Consequently, every Igusa–Todorov Artin algebra satisfies the Auslander–Reiten conjecture.
To the best of our knowledge, Theorem I is the first result that proves ARC for every Igusa–Todorov Artin algebra. It simultaneously covers syzygy-finite algebras and algebras of representation dimension at most three. Our argument uses the witness presentations themselves and gives a finite bound for the degree of a nonzero self-extension. Remark 3.5 explains why the proof does not extend directly to Lat-Igusa–Todorov algebras. The bound depends only on the nonprojective support of the chosen witness and not on . No optimality is asserted. The hypothesis still requires vanishing in every positive degree. The finite test concerns self-extensions only.
In Proposition 3.1, we prove the underlying count for a finite segment of syzygies. We pull back each projective cover extension along a presentation in . Vanishing pullbacks force the next syzygies into . Orthogonality in stable Hom then makes their nonprojective supports disjoint. Nonvanishing pullbacks give a triangular matrix of Ext lengths with positive diagonal. Its factorization through bounds its rank by . Each of the two classes therefore contains at most indices, contradicting the existence of such extensions. The matrix rank argument is related to that of Zhang and Zhou [26, Lemma 3.8]. The entries used here are differences of Ext lengths attached to witness presentations.
By [19 Lemma 2.1], an algebra of representation dimension at most three is 0-Igusa–Todorov. Theorem I supplies the new step from this presentation property to ARC and gives the following consequence.
Theorem II. (see Proposition 4.1) Every Artin algebra of representation dimension at most three satisfies the Auslander–Reiten conjecture.
The second application concerns a condition on powers of the radical. We construct an explicit 1-Igusa–Todorov witness under this condition.
Theorem III. (see Proposition 4.2 and Corollary 4.3) Let be an Artin algebra with radical . If and has finite representation type for some , then satisfies the Auslander–Reiten conjecture. In particular, every Artin algebra with radical cube zero satisfies the conjecture.
The same condition on radical powers in Theorem III appeared in the work of Dräxler and Happel [12] on the generalized Nakayama conjecture. Here we obtain ARC for arbitrary Artin algebras satisfying that condition. For split algebras of finite dimension with radical cube zero, Zhang and Zhou [26, Theorem 1.1] obtain a bound of , where is the number of simple modules. That bound uses a different parameter from and is not uniformly smaller for every choice of witness.
The paper is structured as follows. In § 2, we recall the definition of Igusa–Todorov algebras and establish the syzygy identities and a lifting result used in the proof. In § 3, we prove the estimate for a finite segment, deduce Theorem I and give a version restricted to self-orthogonal modules. In § 4, we prove Theorems II and III.
Throughout the paper, we assume that all algebras are Artin -algebras, where is a commutative Artin ring. We also assume that all modules are finitely generated left modules. For an algebra , denote by the category of finitely generated left -modules. For a module , choose a projective cover and define by the exact sequence . Higher syzygies are defined inductively, with . We write for modulo morphisms that factor through projective modules. If is an -module of finite length, then denotes its composition length. In particular, all Ext groups appearing in the paper have finite -length.
Preliminaries
In this section, we recall some preliminaries on Igusa–Todorov algebras and syzygies of modules.
We first recall the definition of Igusa–Todorov algebras given by Wei [20], see also [21 Section 5].
Definition 2.1. Let be an integer. The algebra is -Igusa–Todorov if there is a module such that every module admits an exact sequence
We call an -Igusa–Todorov witness. An algebra is Igusa–Todorov if it is -Igusa–Todorov for some .
Adjoining to a witness preserves the Igusa–Todorov property and does not change . We also need the following three lemmas on basic homological properties of syzygy modules in . For completeness, we provide brief proofs.
Lemma 2.2. Let . Then we have the following conclusions.
for every .
For and ,
For and every module ,
If , then is projective.
Proof. The first assertion follows by dimension shifting. Let
be a projective cover sequence. Every finitely generated projective module belongs to . Hence gives . Dimension shifting in the two variables now gives, for ,
Iteration proves (2).
For (3), the long exact sequence associated to the projective cover of identifies with the quotient of by the morphisms that extend to that projective cover. These morphisms factor through a projective module. Conversely, if a morphism factors as with projective, its first factor extends to the cover because
The quotient is therefore stable Hom, and dimension shifting proves (3).
For (4), suppose that . Then is projective and
The projective cover sequence of splits, making projective, contrary to the choice of . □
:::
Lemma 2.3. If , then has no nonzero projective direct summand.
Proof. Write $0 \to \Omega X \xrightarrow{u} P \to X \to 0$ for its projective cover sequence. If a projective module is a direct summand of , the projection extends to , because . The composite of the inclusion with is $1_Q$. Thus the image of is a direct summand of contained in . If with this copy of , then , and hence . Nakayama’s lemma gives . □
Lemma 2.4. Let , let be an integer, and suppose that for . Fix an integer and set for . Then
for ;
if ;
if , , and .
Proof. Part (1) is Theorem 2.2(2). For , part (3) of that lemma gives
For and , dimension shifting gives
These identities prove the remaining assertions.
The following property of pullback is very essential in the paper.
Lemma 2.5. Consider exact sequences
where is projective. If the image of in under pullback along is zero, then
In particular, if and , then .
Proof. The pullback splits if and only if lifts to a morphism with . There is then a morphism such that . The sequence
is exact. Surjectivity follows by first lifting the image in through . The description of the kernel follows from and . It splits because is projective. The last assertion follows since is closed under direct summands.
The finite self-extension test
In this section, we show the ARC for Igusa–Todorov algebras by giving a finite self-extension test.
The following estimate contains the counting argument. It requires presentations only for the specified syzygies.
Proposition 3.1 (A bound for a finite segment of syzygies). Let be nonprojective, let be a module, and let be integers. Suppose that
Set for . If, for every , there is an exact sequence
then .
Proof. Replace by and then by a basic module with the same additive closure. Write . Adding introduces only projective indecomposable summands, so this replacement preserves . It also preserves the assumed presentations. Since is nonprojective, Theorem 2.2(4)
gives . Every is consequently nonzero. Moreover, ensures that each is a positive syzygy of . It follows from Theorems 2.2 and 2.3 that and that has no nonzero projective direct summand. In particular, none of the is projective.
For , write the projective cover extension as
Since is not projective, . Denote the pullback map by
Partition the index set as
The indices in . For , Theorem 2.5 gives
If lie in and and have a common indecomposable summand , the composite
has nonzero stable class. Indeed, a factorization of this composite through a projective module would, after composing with the corresponding inclusion and projection, factor through a projective module. This would make projective, a contradiction. But Theorem 2.4(2) gives
Thus the nonprojective supports of the modules , , are nonempty and pairwise disjoint. By (4), they lie among the nonprojective summands of . Therefore .
The indices in . Write . If , then is immediate. So we assume that . Define the integer matrix by
For later use, applying to (2) gives the exact segment
We divide the proof of into three steps:
(1) We show that if in .
If , then . For , we have
Indeed, . Dimension shifting and Theorem 2.2(2) give
The middle map in (5), with and , is therefore an isomorphism. This proves .
(2) We show that for any in .
For the diagonal, apply to (3). Since , we obtain
Consequently, (5) gives the short exact sequence
Additivity of composition length yields
where strict positivity follows from . Thus is upper triangular with positive diagonal, so
(3) We show that .
Let be representatives of the nonprojective indecomposable summands of . By Krull–Schmidt, we may write
where and are projective. Since first Ext vanishes when its first argument is projective,
Hence , where
have sizes and , respectively. It follows that . Together with (6), this gives . The same conclusion covers , when the factorization forces .
Combining the two counts proves
as required. □
We now apply the estimate to an Igusa–Todorov witness and get the main result in this section.
Theorem 3.2 (Finite self-extension test). Let be an -Igusa–Todorov Artin algebra with witness , and put . If and
then is projective. Consequently, every Igusa–Todorov Artin algebra satisfies ARC.
Proof. Suppose that is nonprojective and set . For each , apply the -Igusa–Todorov property to . This gives a presentation in of . Thus Theorem 3.1 applies with and yields , a contradiction. The case is included, since then . Finally, the vanishing in (1) implies the hypotheses of the finite test. □
As an immediate result, we have the following corollary.
Corollary 3.3. Let be a self-injective Igusa–Todorov Artin algebra with witness . Every nonprojective module has a nonzero self-extension in some degree between and . In particular, every self-orthogonal -module is projective.
Proof. Since is injective as a left module over itself, . Apply Theorem 3.2. □
The estimate for a finite segment also gives a criterion restricted to self-orthogonal modules.
Proposition 3.4 (A criterion on self-orthogonal modules). Suppose that there are an integer and a module such that every module satisfying for all admits an exact sequence
Then satisfies ARC.
Proof. Put . If a nonprojective module satisfied the vanishing in (1), then every syzygy of would satisfy the same vanishing by Theorem 2.2. In particular, the assumed presentation is available for , . Applying Theorem 3.1 with and gives the contradiction . □
We end this section with a remark on Lat-Igusa–Todorov algebras.
Remark 3.5. The finite number of indecomposable isomorphism classes in is essential to both counts in the proof. Bravo, Lanzilotta, Mendoza and Vivero [8 Definition 5.1] introduced Lat-Igusa–Todorov algebras, for which the additional summands may range over a subcategory with infinitely many indecomposable isomorphism classes. The argument therefore does not establish the ARC for that larger class. In particular, Theorem 3.3 retains the Igusa–Todorov hypothesis.
Applications
In this section, we give applications to algebras of representation dimension at most three and algebras satisfying for which has finite representation type. Here denotes the Jacobson radical of .
Representation dimension at most three
A module is a generator-cogenerator if every indecomposable projective module and every indecomposable injective module belongs to . Auslander [2] defined the representation dimension of as the infimum of over all generator-cogenerators . The following known consequence will be used.
Proposition 4.1. If , then is 0-Igusa–Todorov. Consequently, satisfies ARC.
Proof. Choose a generator-cogenerator such that . Apply the characterization of Wei [19 Lemma 2.1] with parameter 1. It gives an exact sequence
for every , where . Hence is a 0-Igusa–Todorov witness. The second assertion follows from Theorem 3.2. □
A criterion using radical powers
An Artin algebra has finite representation type if has only finitely many isomorphism classes of indecomposable modules. A module is an additive generator of if .
Proposition 4.2. Let and let . Assume that and has finite representation type. If is an additive generator of , viewed as an -module, then
is a 1-Igusa–Todorov witness for .
Proof. Let and set . If is its projective cover, then . Consequently,
Put and . Both and are annihilated by , so they are -modules and hence belong to .
Form the pullback of a projective cover along . Its middle module fits into exact sequences
The second sequence splits, so the first becomes
Since , there is a module and an integer such that . Taking projective covers gives
Thus . The two left terms therefore lie in , as required. □
Immediately, we have the following corollaries.
Corollary 4.3. Under the hypotheses of Theorem 4.2, let . If and for , then is projective. In particular, satisfies .
Proof. Apply Theorem 3.2 to the witness in Theorem 4.2. □
Corollary 4.4. Every Artin algebra with radical cube zero satisfies .
Proof. Take in Theorem 4.3. The algebra is semisimple and therefore has finite representation type. □
These results prove Theorems II and III. For , one may take to be the direct sum of representatives of the simple modules. The support of the witness also involves the radicals of indecomposable projective modules. This explains why differs from the parameter counting simple modules in the work of Zhang and Zhou [26].
Remark 4.5. The generalized Nakayama conjecture and are equivalent as assertions over all Artin algebras. This does not give a pointwise implication from the former to the latter, as explained by Chen, Hu, Qin and Wang [9 Introduction]. The application concerning radical powers instead follows directly from the explicit Igusa–Todorov witness.
Although Igusa–Todorov algebras form a broad class of algebras satisfying the Auslander–Reiten conjecture, there also exist non-Igusa–Todorov algebras satisfying the conjecture. We end the paper with the following example on non-Igusa–Todorov algebras satisfying the Auslander–Reiten conjecture.
Example 4.6. Let , and let be the quiver with one vertex and three loops , and . Let be the ideal of generated by
and set . Then
is an 8-dimensional commutative local algebra, but not an Igusa–Todorov algebra.
(2) satisfies the Auslander–Reiten conjecture. In fact, every finitely generated -module with is projective.
Disclosure of computational assistance. Chatgpt was used for language editing and improving the clarity of the manuscript. All mathematical results, arguments, and proofs were developed and verified by the authors.
Acknowledgements. Xiaojin Zhang is supported by the National Natural Science Foundation of China (Grant Nos. 12171207 and 12371038). Panyue Zhou is supported by the National Natural Science Foundation of China (Grant No. 12371034).
Data Availability Data sharing is not applicable to this article, as no datasets were generated or analysed during the current study.
Conflict of Interests The authors declare that they have no conflicts of interest.
References
Xiaojin Zhang
School of Mathematics and Statistics, Jiangsu Normal University, Xuzhou 221116, Jiangsu, P. R. China
E-mail: [email protected]
Panyue Zhou
School of Mathematics and Statistics, Changsha University of Science and Technology, Changsha 410114, Hunan, P. R. China
E-mail: [email protected].
References
- [1]T. Araya, The Auslander–Reiten conjecture for Gorenstein rings, Proc. Amer. Math. Soc. 137 (2009), 1941–1944.arxiv.org/abs/0805.0210
- [2]M. Auslander, Representation dimension of Artin algebras, Queen Mary College Mathematics Notes, Queen Mary College, London, 1971.
- [3]M. Auslander, S. Ding and Ø. Solberg, Liftings and weak liftings of modules, J. Algebra 156 (1993), 273–317.DOI
- [4]M. Auslander and I. Reiten, On a generalized version of the Nakayama conjecture, Proc. Amer. Math. Soc. 52 (1975), 69–74.DOI
- [5]A. Bahlkeheh, A. Mahin Fallah and Sh. Salari an, On the Auslander-Reiten conjecture for algebras, J. Algebra 427 (2015), 252–263.
- [6]H. Bass, Finitistic dimension and a homological generalization of semi-primary rings, Trans. Amer. Math. Soc. 95 (1960), 466–488.DOI
- [7]M. Barrios, M. Lanzilotta and G. Mata, A survey on Igusa–Todorov functions, Rev. R. Acad. Cienc. Exactas Fís. Nat. Ser. A Mat. 118 (2024), article 146.arxiv.org/abs/2310.09283
- [8]D. Bravo, M. Lanzilotta, O. Mendoza and J. Vivero, Generalised Igusa–Todorov functions and Lat-Igusa–Todorov algebras, J. Algebra 580 (2021), 63–83.arxiv.org/abs/2002.07866
- [9]Y. Chen, W. Hu, Y. Qin and R. Wang, Singular equivalences and Auslander–Reiten conjecture, J. Algebra 623 (2023), 42–63.arxiv.org/abs/2011.02729
- [10]H. Chen and C. Xi, Homological theory of self-orthogonal modules, Trans. Amer. Math. Soc. 378 (2025), no. 10, 7287–7335.
- [11]L. W. Christensen and H. Holm, Algebras that satisfy Auslander’s condition on vanishing of cohomology, Math. Z. 265 (2010), 21–40.arxiv.org/abs/0801.0401
- [12]P. Dräxler and D. Happel, A proof of the generalized Nakayama conjecture for algebras with J²ℓ+1 = 0 and A/Jℓ representation finite, J. Pure Appl. Algebra 78 (1992), no. 2, 161–164.DOI
- [13]H. Enomoto and R. Marczinzik, On the homological conjectures for Artin algebras, arXiv:2609.19172.
- [14]M. Hoshino, Modules without self-extensions and Nakayama’s conjecture, Arch. Math. 43 (1984), 493–500.DOI
- [15]C. Huneke and G. J. Leuschke, On a conjecture of Auslander and Reiten, J. Algebra 275 (2004), 781–790.DOI
- [16]C. Huneke, L. M. Şega and A. N. Vraciu, Vanishing of Ext and Tor over some Cohen-Macaulay local rings, Illinois J. Math. 48 (2004), 295–317.
- [17]K. Igusa and G. Todorov, On the finitistic global dimension conjecture for Artin algebras, in: Representations of Algebras and Related Topics, Fields Inst. Commun. 45, Amer. Math. Soc., Providence, RI, 2005, 201–204.DOI
- [18]OpenAI An explicit counterexample to the Auslander–Reiten conjecture, preprint, 2026.
- [19]J. Wei, Finitistic and representation dimensions, preprint (2008), arXiv: 0803.3364v3.arxiv.org/abs/0803.3364
- [20]J. Wei, Finitistic dimension and Igusa–Todorov algebras, Adv. Math. 222 (2009), no. 6, 2215–2226.DOI
- [21]J. Wei, Derived categories and syzygies, preprint (2011), arXiv: 1109.6226.arxiv.org/abs/1109.6226
- [22]J. Wei, Tilting complexes and Auslander–Reiten conjecture, Math. Z. 272 (2012), 431–441.DOI
- [23]W. Xia The Auslander–Reiten conjecture for quantum complete intersections, preprint (2026), arXiv: 2609.24007V1.arxiv.org/abs/2609.24007
- [24]D. Xu, A Note on the Auslander-Reiten Conjecture, Acta Mathematica Sinica, English Series 29 (2013), 1993–1996.
- [25]D. Xu, Auslander-Reiten conjecture and special biserial algebras, Arch. Math. 105 (2015), 13–22.DOI
- [26]X. Zhang and P. Zhou, The Auslander-Reiten conjecture for algebras with radical cube zero, preprint (2026), arXiv: 2609.08679.arxiv.org/abs/2609.08679