betaFeedback

WorksMathematics

Full tt-Linkage, Cancellation, and Almost Krull Domains

Abstract

Motivated by open questions of Kabbaj and Mimouni, we determine the position of full tt-linkage among the standard classes of integral domains and its relationship with almost Krullness. Full linkage for a star operation is equivalent to cancellation in the monoid of all nonzero fractional star ideals. For the tt-operation, this full cancellation law, denoted by TCL\mathrm{TCL}, lies strictly between the tt-almost Dedekind and completely integrally closed classes; on the PvMD\mathrm{PvMD} locus, we recover the known tt-almost Dedekind characterization. We prove that TCL\mathrm{TCL} and almost Krullness are incomparable. A monoid algebra arising from Reinhart's example satisfies TCL\mathrm{TCL} but is neither almost Krull nor a PvMD\mathrm{PvMD}. In the opposite direction, we construct a locally factorial almost Krull domain with a maximal ideal 0≠M=Mt=(M2)t⊊D0\ne M=M_t=(M^2)_t\subsetneq D. For this domain the canonical locally divisorial operation κκ is cancellative and κf=tκ_f=t, but cancellation does not descend to tt. In contrast, finite valuation patching proves that the Heinzer–Ohm and Arnold–Matsuda almost Krull non-PvMD\mathrm{PvMD}s satisfy TCL\mathrm{TCL} with t<κt<κ; in the latter example κ=vκ=v. Class-monoid, residual, and support criteria identify the obstruction to descent. Thus the missing tt-linkage node is governed by cancellation in the entire fractional tt-ideal monoid, rather than by local Krullness alone.

Introduction

This paper answers the open questions of Kabbaj and Mimouni concerning the position of the tt-linkage property in the hierarchy of integral domains and its relationship with almost Krull domains [20 Figure 1 and p. 11]. Their questions bring together two structural viewpoints: linkage expressed through colon ideals and cancellation in multiplicative monoids of fractional ideals. The aim is to identify what full tt-linkage measures when the local descriptions available for Prüfer vv-multiplication domains no longer apply.

Our work is motivated by a series of characterizations of domains in terms of linkage properties. Kabbaj and Mimouni showed that finite ordinary linkage characterizes Prüfer domains, whereas ordinary linkage for all nonzero ideals characterizes almost Dedekind domains [19]. In Linkage of ideals under star operations, they extended this framework to arbitrary star operations. They proved that finite tt-linkage and finite vv-linkage each characterize vv-domains [20 Theorem 3.1], while full vv-linkage characterizes completely integrally closed domains [20 Theorem 3.2]. They also established that finite ww-linkage characterizes Prüfer vv-multiplication domains (PvMDs), whereas full ww-linkage characterizes tt-almost Dedekind domains [20 Theorem 3.3]. Here and throughout, full refers to their linkage property for all nonzero ideals; the adjective serves only to distinguish this property from its finitely generated version.

Related work appears in the manuscript of Huang and Wang [18], who introduced a ww-analogue of linkage and studied its finite version in weakly Matlis domains. Recall that a domain is weakly Matlis if each nonzero element belongs to only finitely many maximal tt-ideals and no two distinct maximal tt-ideals contain a common nonzero prime ideal. Their Theorem 3.9 characterizes weakly Matlis PvMDs in terms of finite ww-linkage. In the star-operation framework adopted here, the corresponding characterization of PvMDs obtained by Kabbaj and Mimouni requires no weakly Matlis hypothesis [20 Theorem 3.3].

Since t=wt=w on every PvMD, these results also characterize full tt-linkage within this class: a PvMD has full tt-linkage if and only if it is tt-almost Dedekind [20 Corollary 3.4]. Outside the class of PvMDs, however, these results do not settle the relationship between full tt-linkage and the almost Krull property. Following Pirtle, a domain DD is called almost Krull if DND_N is a Krull domain for every maximal ideal NN of DD [28, 29, 30].

The present paper investigates this relationship, which is also left unresolved in Figure 1 of [20]. More specifically, our investigation is guided by the following questions posed by Kabbaj and Mimouni at the end of their article [20 p. 11, Open Questions].

Figure 1. The position of the tt-linkage node from [20 Figure 1], as determined by Theorem 6.7. Solid arrows are implications. Within a box, full tt-linkage and TCL are equivalent, as are the two conditions in each of the tt-almost Dedekind and CIC boxes. The lower-left box records properties of almost Krull domains, not a converse characterization. No implication connects AK with TCL, or PvMD with TCL.

Question 1.1.

(i) Where should the tt-linkage property be placed in the diagram?

(ii) What is its precise relationship with almost Krull domains?

(iii) Or is it naturally aligned with another class shown in the figure?

Cancellation provides a natural approach to these questions. Gilmer proved that cancellation by all nonzero finitely generated ideals characterizes Prüfer domains, whereas cancellation by all nonzero ideals characterizes almost Dedekind domains [10 Corollary 1 and Theorem 3]. Anderson and Anderson developed the corresponding residual criteria [1]. The parallel with the linkage characterizations suggests that the missing node should be sought in the multiplicative structure of fractional tt-ideals rather than by replacing ww with tt in a local criterion.

The first theorem identifies full ⋆\star-linkage with full ⋆\star-cancellation for every star operation ⋆\star (Theorem 2.3). For ⋆=t\star=t, this identifies full linkage with the class TCL of domains whose monoid of all nonzero fractional tt-ideals, under (I,J)↦(IJ)t(I,J)\mapsto(IJ)_t, is cancellative. Writing tt-AD, AK, and CIC for the tt-almost Dedekind, almost Krull, and completely integrally closed classes, respectively, we obtain

t-AD⊊TCL⊊CIC⊆v-domain,TCL∩PvMD=t-AD.(1)t\text{-AD} \subsetneq\mathrm{TCL} \subsetneq\mathrm{CIC} \subseteq v\text{-domain}, \qquad\mathrm{TCL}\cap\mathrm{PvMD}=t\text{-AD}. \tag*{(1)}

The intersection identity recovers Kabbaj–Mimouni’s Corollary 3.4 in cancellation language; the strict containments locate the full tt-node beyond that previously determined intersection.

Two distinctions are essential. First, Chang’s theorem identifies ww-cancellation of an ideal with tt-local principality; it does not identify it with tt-invertibility without an additional tt-finiteness hypothesis [5]. Second, full cancellation tests arbitrary fractional ideals in all three positions, whereas the a.b. and e.a.b. conditions require the factor being cancelled to be finitely generated. Although these latter conditions agree for finite-type operations, neither is equivalent to full cancellation [8, 24]. We show that full tt-cancellation is strictly stronger than the a.b. property of tt, but does not force tt-invertibility of finitely generated ideals.

The almost Krull part of Question 1.1 has a particularly concrete interpretation. If every maximal localization of DD is Krull, then the canonical locally divisorial operation Iκ=⋂N∈Max⁡(D)(IDN)vNI_\kappa=\bigcap_{N\in\operatorname{Max}(D)}(ID_N)_{v_N} is cancellative and satisfies κf=t\kappa_f=t [6]. Thus the question whether almost Krullness implies full tt-linkage asks whether cancellation survives passage from this operation to its finite-type associate. Equality κ=t\kappa=t holds exactly on the almost Krull PvMD locus, but equality of operations is stronger than preservation of cancellation. We establish both possibilities when t<κt<\kappa.

Reinhart’s cancellative non-tt-Prüfer monoid, together with the monomial-ideal description of Kim and Park, yields a TCL-domain outside both the PvMD and almost Krull classes [22, 31]; see Theorem 3.3. A different finite-patching argument proves that the Heinzer–Ohm essential domain and the Arnold–Matsuda domain both belong to AK∩TCL∖PvMD\mathrm{AK} \cap\mathrm{TCL} \setminus\mathrm{PvMD}. Their local descriptions come from [3, 15, 16]; the full cancellation statements are proved here. In the Arnold–Matsuda example, t=κf<κ=vt = \kappa_f < \kappa= v, and a height-two non-well-behaved maximal tt-ideal has tt-closure smaller than its κ\kappa-closure.

For the opposite direction, we introduce infinitely many positive weights tending to zero into a Heinzer–Ohm-type construction. Every maximal localization of the resulting domain is a UFD, but one maximal ideal satisfies

0≠M=Mt=(M2)t⊊D,Mκ=D.(2)0 \ne M = M_t = (M^2)_t \subsetneq D,\qquad M_\kappa= D. \tag*{(2)}

Thus (M2)t=(MD)t(M^2)_t = (MD)_t while Mt≠DtM_t \ne D_t. The construction proves that almost Krullness does not imply TCL, even when all maximal localizations are factorial and satisfy their own full tt-cancellation laws (Theorem 5.3 and Corollary 5.4). Together with the monoid example, it gives the incomparability of AK and TCL, answering the second question in Question 1.1. It also sharpens the positive examples to

AK∩PvMD⊊AK∩TCL⊊AK.(3)\mathrm{AK} \cap\mathrm{PvMD} \subsetneq\mathrm{AK} \cap\mathrm{TCL} \subsetneq\mathrm{AK}. \tag*{(3)}

Theorem 6.7 collects the answers to all three questions, and Figure 1 records the resulting position of the missing node.

The paper has six sections. § 2 develops the residual language of linkage and cancellation. § 3 gives the monoid example and the Heinzer–Ohm patching argument. § 4 studies κ\kappa and proves full cancellation for the Arnold–Matsuda example. § 5 constructs the almost Krull domain failing TCL. § 6 describes the descent defect and records the final implication diagram. Throughout, DD is an integral domain with quotient field KK, and all fractional ideals are nonzero. Fields are included in the tt-almost Dedekind class; equivalently, the defining DVR condition is imposed on nonzero maximal tt-ideals. The abbreviation AK always means locally Krull, not an almost-factorial condition.

For any undefined terminology and notation, the reader is referred to [32].

Linkage, cancellation, and the full tt-node

This section places linkage and cancellation in a common notation and then uses the equivalence to locate full tt-linkage among familiar Prüfer-type conditions.

Star operations and residual linkage

Let F(D)\mathcal{F}(D) be the set of nonzero fractional ideals of DD, and let f(D)\mathfrak{f}(D) be its subset of finitely generated fractional ideals. A star operation on DD is a map I↦I⋆I \mapsto I_\star on F(D)\mathcal{F}(D) such that, for 0≠x∈K0 \ne x \in K and I,J∈F(D)I,J \in\mathcal{F}(D), (xI)⋆=xI⋆(xI)_\star= xI_\star, I⊆J⇒I⋆⊆J⋆I \subseteq J \Rightarrow I_\star\subseteq J_\star, I⊆I⋆I \subseteq I_\star, and I⋆⋆=I⋆I_{\star\star} = I_\star, and D⋆=DD_\star= D. It is of finite type if I⋆=⋃{F⋆:F∈f(D), F⊆I}I_\star= \bigcup\{F_\star: F \in\mathfrak{f}(D),\ F \subseteq I\} for every II. These definitions go back to Krull; standard references are [11 Section 32] and [14]. The more flexible semistar formulation was introduced by Okabe and Matsuda [26].

For I∈F(D)I \in\mathcal{F}(D), put I−1=(D:I)I^{-1} = (D : I), Iv=(I−1)−1I_v = (I^{-1})^{-1}, and It=⋃{Fv:F∈f(D), F⊆I}I_t = \bigcup\{F_v : F \in\mathfrak{f}(D),\ F \subseteq I\}. The stable finite-type operation associated to vv is Iw=⋂M∈t-Max⁡(D)IDMI_w = \bigcap_{M \in t\text{-}\operatorname{Max}(D)} ID_M; see [7, 33]. Thus w≤t≤vw \leq t \leq v. A fractional ideal is a ⋆\star-ideal if I=I⋆I = I_\star, and it is ⋆\star-invertible if (II−1)⋆=D(II^{-1})_\star= D. Here t-Max⁡(D)t\text{-}\operatorname{Max}(D) denotes the set of ideals maximal among the proper integral tt-ideals; these ideals are prime. A tt-ideal is tt-finite if it equals Ft=FvF_t = F_v for some finitely generated fractional ideal FF.

For fractional ideals AA and BB, write (A:B)={x∈K:xB⊆A}(A : B) = \{x \in K : xB \subseteq A\}. If AA and BB are integral ideals, the integral colon is (A:DB)=(A:B)∩D(A :_D B) = (A : B) \cap D.

Following [20 Definition 2.1], call two nonzero integral ideals II and JJ ⋆\star-linked if there is a nonzero ideal A⊆I∩JA \subseteq I \cap J such that I⋆=((A⋆:DJ))⋆I_{\star}=((A_{\star}:_{D}J))_{\star} and J⋆=((A⋆:DI))⋆J_{\star}=((A_{\star}:_{D}I))_{\star}. The domain has full ⋆\star-linkage if every such pair is ⋆\star-linked, and finite ⋆\star-linkage if this is required only for pairs of nonzero finitely generated ideals. The adjective full is omitted in [20].

The defining ideal can always be replaced by the product. This is [20 Lemma 2.2]; we include the proof to fix the residual notation.

Lemma 2.1. Two nonzero integral ideals II and JJ are ⋆\star-linked if and only if

I⋆=((IJ)⋆:DJ)⋆,J⋆=((IJ)⋆:DI)⋆.(4)I_{\star} = ((IJ)_{\star}:_{D}J)_{\star}, \qquad J_{\star} = ((IJ)_{\star}:_{D}I)_{\star}. \tag*{(4)}

Proof. If II and JJ are linked through AA, then I⊆(A⋆:DJ)I \subseteq(A_{\star}:_{D}J). Indeed, the latter colon is a ⋆\star-ideal: it is the nonzero intersection of DD and the ⋆\star-ideals j−1A⋆j^{-1}A_{\star} for 0≠j∈J0 \ne j \in J. Consequently, IJ⊆A⋆IJ \subseteq A_{\star} and (IJ)⋆⊆A⋆(IJ)_{\star} \subseteq A_{\star}. The inclusions I⊆((IJ)⋆:DJ)⊆(A⋆:DJ)I \subseteq((IJ)_{\star}:_{D}J) \subseteq(A_{\star}:_{D}J) yield the first identity after taking ⋆\star-closures. The second follows by symmetry. Conversely, if the displayed identities hold, choose A=IJA = IJ, which is a nonzero subideal of I∩JI \cap J. □

A fractional ideal AA is a ⋆\star-cancellation ideal if (AB)⋆=(AC)⋆(AB)_{\star}=(AC)_{\star} implies B⋆=C⋆B_{\star}=C_{\star} for all B,C∈F(D)B,C \in\mathcal{F}(D). The full ⋆\star-cancellation law holds if every member of F(D)\mathcal{F}(D) is a ⋆\star-cancellation ideal. This extends the classical notions studied by Gilmer and by Anderson–Anderson [1, 10].

For the identity operation dd, Anderson and Anderson call AA a quasi-cancellation ideal when (AF:A)=F(AF:A)=F for every finitely generated fractional ideal FF. Their residual and inclusion criteria are the dd-counterparts of the lemma below [1 Lemma 1]. If AA is finitely generated, quasi-cancellation, cancellation, and invertibility are equivalent [1 Corollary 1 and Theorem 1]. At the level of the whole domain, every nonzero finitely generated ideal is cancellative exactly when DD is Prüfer, whereas every nonzero ideal is cancellative exactly when DD is almost Dedekind [1 Theorem 6]; compare [10 Corollary 1 and Theorem 3]. By contrast, every nonzero ideal is merely quasi-cancellative exactly when DD is a completely integrally closed Prüfer domain [1 Theorem 7]. Thus the distinction between finite residual tests and unrestricted cancellation is present before any nontrivial star operation is introduced.

The following lemma is the star-operation residual form of these classical equivalences and converts equality-cancellation into the inclusion form needed later.

References

  1. [1]D. D. Anderson and D. F. Anderson, Some remarks on cancellation ideals, Math. Japon. 29 (1984), 879–886.
  2. [2]D. D. Anderson, D. F. Anderson, M. Fontana, and M. Zafrullah, On v-domains and star operations, Comm. Algebra 37 (2009), no. 9, 3018–3043.arxiv.org/abs/0809.2947
  3. [3]J. T. Arnold and R. Matsuda, An almost Krull domain with divisorial height one primes, Canad. Math. Bull. 29 (1986), no. 1, 50–53. doi:10.4153/CMB-1986-009-6.DOI
  4. [4]N. Bourbaki, Commutative Algebra, Springer-Verlag, Berlin, 1989.
  5. [5]G. W. Chang, Characterizations of ⋆-cancellation ideals of an integral domain, Comm. Algebra 37 (2009), 3309–3320. doi:10.1080/00927870802502795.DOI
  6. [6]S. El Baghdadi, L. Izq uie, and A. Tamoussit, Almost Krull domains and their rings of integer-valued polynomials, J. Pure Appl. Algebra 224 (2020), no. 6, Article 106269.
  7. [7]M. Fontana, J. A. Huckaba, and I. J. Papick, Prüfer Domains, Monographs and Textbooks in Pure and Applied Mathematics, vol. 203, Marcel Dekker, New York, 1997.
  8. [8]M. Fontana and K. A. Loper, Cancellation properties in ideal systems: a classification of e.a.b. semistar operations, J. Pure Appl. Algebra 213 (2009), 2095–2103.
  9. [9]M. Fontana, K. A. Loper, and R. Matsuda, Cancellation properties in ideal systems: an e.a.b. not a.b. star operation, Arab. J. Sci. Eng. 35 (2010), no. 1D, 45–49.arxiv.org/abs/1002.2388
  10. [10]R. W. Gilmer, Jr., The cancellation law for ideals in a commutative ring, Canad. J. Math. 17 (1965), 281–287.DOI
  11. [11]R. Gilmer, Multiplicative Ideal Theory, Marcel Dekker, New York, 1972.
  12. [12]R. Gilmer, Completely Semigroup Rings, University of Chicago Press, Chicago, 1984.
  13. [13]M. Griffin, Some results on v-multiplication rings, Canad. J. Math. 19 (1967), 710–722.DOI
  14. [14]F. Halter-Koch, Ideal Systems: An Introduction to Multiplicative Ideal Theory, Marcel Dekker, New York, 1998.
  15. [15]W. Heinzer and J. Ohm, An essential ring which is not a v-multiplication ring, Canad. J. Math. 25 (1973), no. 4, 856–861. doi:10.4153/CJM-1973-088-5.DOI
  16. [16]E. Houston, S.-E. Kabbaj, and A. Mimouni, ⋆-reductions of ideals and Prüfer v-multiplication domains, J. Commut. Algebra 9 (2017), no. 4, 491–505. doi:10.1216/JCA-2017-9-4-491.DOI
  17. [17]E. Houston and M. Zafrullah, Well-behaved prime t-ideals and almost Krull domains, Comm. Algebra 51 (2023), no. 1, 276–287. doi:10.1080/00927872.2022.2098308.DOI
  18. [18]K. Huang and F. Wang, A short note on domains with w-linkage property, preprint, version dated October 16, 2021.
  19. [19]S. Kabbaj and A. Mimouni, Linkage of ideals in integral domains, Algebr. Represent. Theory 24 (2021), 799–809. doi:10.1007/s10468-020-09969-7.DOI
  20. [20]S. Kabbaj and A. Mimouni, Linkage of ideals under star operations, Rend. Circ. Mat. Palermo (2) 75 (2026), Article 184, 13 pp. doi:10.1007/s12215-026-01503-w.DOI
  21. [21]B. G. Kang, Pruifer v-multiplication domains and the ring R[X]_{N_v}, J. Algebra 123 (1989), 151–170.DOI
  22. [22]H. Kim and Y. S. Park, The t-class group of a semigroup ring, Algebra Colloq. 8 (2001), no. 4, 419–426.
  23. [23]R. Matsuda, Cancellation properties in ideal systems of monoids, Int. Electron. J. Algebra 9 (2011), 61–68.
  24. [24]R. Matsuda, Note on cancellation properties in ideal systems, Comm. Algebra 43 (2015), no. 1, 23–28.DOI
  25. [25]R. Matsuda and T. Sugatani, Cancellation ideals in pseudo-valuation domains, Comm. Algebra 23 (1995), no. 11, 3983–3991.DOI
  26. [26]A. Okabe and R. Matsuda, Semistar operations on integral domains, Math. J. Toyama Univ. 17 (1994), 1–21.
  27. [27]B. Olberding and A. Reinhart, Radical factorization in finitary ideal systems, Comm. Algebra 48 (2020), no. 1, 228–253. doi:10.1080/00927872.2019.1640237.DOI
  28. [28]E. M. Pirtle, Integral domains which are almost Krull, J. Sci. Hiroshima Univ. Ser. A-I Math. 32 (1968), 441–447.DOI
  29. [29]E. M. Pirtle, Families of valuations and semigroups of fractionary ideal classes, Trans. Amer. Math. Soc. 144 (1969), 427–439.DOI
  30. [30]E. M. Pirtle, On a generalization of Krull domains, J. Algebra 14 (1970), 485–492.DOI
  31. [31]A. Reinhart, On monoids and domains whose monadic submonoids are Krull, in M. Fontana, S. Frisch, and S. Glaz (eds.), Commutative Algebra: Recent Advances in Commutative Rings, Integer-Valued Polynomials, and Polynomial Functions, Springer, New York, 2014, pp. 307–330.DOI
  32. [32]F. G. Wang and H. Kim, Foundations of Commutative Rings and Their Modules, 2nd ed., Singapore: Springer Nature, 2024.
  33. [33]F. Wang and R. L. McCasland, On w-modules over strong Mori domains, Comm. Algebra 25 (1997), 1285–1306.

Paper details

vlnc:W_kzsxxth5
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 3

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents