The structure of lim 1 \displaystyle{\lim}^{1} -groups
Ioannis Emmanouil · 2026
Read The structure of lim 1 \displaystyle{\lim}^{1} -groups by Ioannis Emmanouil (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-08108.txt.
Introduction
Ioannis Emmanouil writes a short note on lim-one groups. These groups measure how far a module sits from its completion. For a shrinking chain of submodules, the lim-one group is the cokernel of the map into the completion. The note asks if every lim-one group arises in that way.
Emmanouil proves that it does. The form is canonical and functorial. Any inverse sequence of modules admits a map from a chain sequence that matches on lim-one. The proof uses Quillen's small object argument, as recast by Eklof and Trlifaj. It also uses a construction of Salce that gives enough projectives.
Catalog record
| CREATOR | Ioannis Emmanouil |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 34 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.08108
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The structure of lim 1 \displaystyle{\lim}^{1} -groups Thanks: Research funded by the Hellenic Foundation for Research and Innovation (H.F.R.I.) under the ”3rd Call for H.F.R.I. Research Projects to Support Faculty Members and Researchers”, project number 24921 Ioannis Emmanouil Abstract . If ( A n ) n (A_{n})_{n} is a decreasing filtration of a module A A and A ^ = lim n A / A n \widehat{A}=\mbox{$\displaystyle{\lim}_{n}\!\!$}\ A/A_{n} , then lim n 1 A n \mbox{$\displaystyle{\lim}_{n}^{1}\!$}\ A_{n} is identified with the cokernel of the canonical map A ⟶ A ^ A\longrightarrow\widehat{A} . In this note, we show that any lim 1 \displaystyle{\lim}^{1} -group is canonically of that form: For any inverse sequence of modules ( X n ) n (X_{n})_{n} there exists an inverse sequence ( A n ) n (A_{n})_{n} as above and a morphism ( A n ) n ⟶ ( X n ) n (A_{n})_{n}\longrightarrow(X_{n})_{n} , depending functorially on ( X n ) n (X_{n})_{n} , that induces an isomorphism on lim 1 \displaystyle{\lim}^{1} . The proof is based on Quillen’s small object argument, as formulated by Eklof and Trlifaj in their investigation of the existence of enough injective objects in certain cotorsion pairs, and also uses a construction by Salce that provides enough projective objects therein. 0. Introduction The right derived functors of the limit functor in module categories were introduced by Milnor [ 8 ] and Yeh [ 13 ] , in order to compute such topological invariants as the homotopy groups of the limit of a sequence of fibrations or the cohomology of an ascending union of spaces. Roos showed in [ 9 ] that the higher limits vanish in degrees \geq 2 , if the indexing category is the ordered set of natural numbers. (This assumption will hold throughout this note.) The most manageable criterion that implies the vanishing of \displaystyle{\lim}^{1} is the Mittag-Leffler condition, that was introduced by Grothendieck in [ 6 ] . This condition turns out to be equivalent to the vanishing of \displaystyle{\lim}^{1} for inverse sequences of countable abelian groups [ 5 ] , but the relation between the two conditions is slightly more complicated in general; cf. [ 1 , Theorem 1.3] or [ 3 , Corollary 6] . It is shown in [ 2 ] that the vanishing of \displaystyle{\lim}^{1} is equivalent to a certain transfinite version of the Mittag-Leffler condition. Another class of examples over which the computation of \displaystyle{\lim}^{1} is well-understood is provided by decreasing filtrations: If (A_{n})_{n} is a decreasing sequence of submodules of a module A=A_{0} and \widehat{A}=\mbox{$\displaystyle{\lim}_{n}\!\!$}\ A/A_{n} is the Hausdorff completion of A with respect to the linear topology that is induced by the filtration, then \mbox{$\displaystyle{\lim}_{n}^{1}\!$}\ A_{n} is identified with the cokernel of the canonical map A\longrightarrow\widehat{A} . Therefore, assuming that the linear topology on A is already Hausdorff, the group \mbox{$\displaystyle{\lim}_{n}^{1}\!$}\ A_{n} measures the obstruction to completeness of A . The reader may consult [ 12 , \lx@sectionsign 3.5] for a proof of this result. Our goal in this paper is to show that any \displaystyle{\lim}^{1} -group is of that form, i.e. it coincides with the value of \displaystyle{\lim}^{1} on an inverse sequence that is associated with a suitable decreasing filtration. More precisely, we prove the following result. Theorem. For any inverse sequence of modules (X_{n})_{n} there is an another inverse sequence of modules (A_{n})_{n} and a morphism of inverse sequences f:(A_{n})_{n}\longrightarrow(X_{n})_{n} , which depends functorially on (X_{n})_{n} , such that: (i) the inverse sequence (A_{n})_{n} represents a decreasing filtration of A=A_{0} by submodules, (ii) the modules A_{n} are free for all n , (iii) the successive quotients A_{n}/A_{n+1} are free modules for all n and (iv) f induces an isomorphism \mbox{$\displaystyle{\lim}_{n}^{1}\!$}\ A_{n}\stackrel{{\scriptstyle\sim}}{{\longrightarrow}}\mbox{$\displaystyle{\lim}_{n}^{1}\!$}\ X_{n} . The proof proceeds by working in the abelian category {\mathcal{C}} of inverse sequences of modules. We express \displaystyle{\lim}^{1} as the Ext 1 -functor on {\mathcal{C}} that is associated with a certain inverse sequence and consider the cotorsion pair in {\mathcal{C}} that is cogenerated by that inverse sequence. The reader is referred to [ 4 ] for the definition and basic properties of cotorsion pairs. We use the version of Quillen’s small object that was formulated by Eklof and Trlifaj [ 4 , 11 ] and a construction by Salce [ 10 ] to obtain the completeness of that cotorsion pair. It will turn out that the result stated above is an immediate consequence of this completeness. Notations and terminology. We fix a unital associative ring R and consider only left modules over R . As mentioned above, with the proof of the Theorem in mind, we only consider inverse sequences of modules in this paper and refrain from considering diagrams of modules indexed by more general directed sets or other small categories. 1. An Ext-description of \displaystyle{\lim}^{1} We consider the abelian category {\mathcal{C}} of inverse sequences of modules; it is precisely the category of contravariant functors from the ordered set \omega to the category of modules. In this section, we elaborate on the discussion in [ 2 , \lx@sectionsign 2] and identify the \displaystyle{\lim}^{1} -functor to the Ext 1 -functor on {\mathcal{C}} induced by a suitable inverse sequence, a result that is certainly known to the experts. Let X=(X_{n})_{n} be an inverse sequence of modules with structure maps (s_{n}:X_{n}\longrightarrow X_{n-1})_{n} . We consider the linear map 1-S:{\displaystyle{\prod}}_{n}X_{n}\longrightarrow{\displaystyle{\prod}}_{n}X_{n}, whose composition with the projection onto X_{n} coincides with the projection from the product to X_{n}\oplus X_{n+1} , followed by [1,-s_{n+1}]:X_{n}\oplus X_{n+1}\longrightarrow X_{n} , for all n . Then, \lim_{n}X_{n}=\mbox{ker}(1-S) and \mbox{$\displaystyle{\lim}_{n}^{1}\!$}\ X_{n}=\mbox{coker}(1-S) . The reader is referred to Weibel’s book [ 12 , \lx@sectionsign 3.5] for background and general properties of the \displaystyle{\lim}^{1} -functor. If M is a module and n a non-negative integer, then we may consider the inverse sequence M(n) , consisting of M in degrees \leq n and 0 in degrees >n , with all structure maps M\longrightarrow M given by the identity map of M . It is easily seen that for any inverse sequence X=(X_{i})_{i} there is an isomorphism of abelian groups \phi_{n}:\mbox{Hom}_{\mathcal{C}}(M(n),X)\stackrel{{\scriptstyle\sim}}{{\longrightarrow}}\mbox{Hom}_{R}(M,X_{n}), which is natural in both M and X . Since exactness in {\mathcal{C}} is defined degreewise, we conclude that M(n) is a projective object in {\mathcal{C}} for all n\geq 0 , if M is a projective module. We also note that there is an obvious monomorphism \iota_{n}:M(n)\longrightarrow M(n+1) in {\mathcal{C}} for all n\geq 0 , which is natural in M ; it is given by the identity map of M in degrees \leq n . Moreover, the following diagram is commutative for any inverse sequence X=(X_{i})_{i} (1) \begin{array}[]{ccc}\mbox{Hom}_{\mathcal{C}}(M(n+1),X)&\stackrel{{\scriptstyle\phi_{n+1}}}{{\longrightarrow}}&\mbox{Hom}_{R}(M,X_{n+1})\\ {\scriptstyle{\imath_{n}^{*}}}\downarrow&&\downarrow{\scriptstyle{(s_{n+1})_{*}}}\\ \mbox{Hom}_{\mathcal{C}}(M(n),X)&\stackrel{{\scriptstyle\phi_{n}}}{{\longrightarrow}}&\mbox{Hom}_{R}(M,X_{n})\end{array} Here, we denote by s_{n+1}:X_{n+1}\longrightarrow X_{n} the structure map of X ; the additive map (s_{n+1})_{*} is the corresponding structure map of the inverse sequence (\mbox{Hom}_{R}(M,X_{i}))_{i} . This diagram is natural in both M and (X_{i})_{i} . The colimit M(\infty)=\mbox{colim}_{n}M(n) is the inverse sequence consisting of M in each degree, with all structure maps given by the identity of M . There is a short exact sequence in {\mathcal{C}} (2) 0\longrightarrow{\textstyle{\bigoplus_{n}}}M(n)\stackrel{{\scriptstyle 1-I}}{{\longrightarrow}}{\textstyle{\bigoplus_{n}}}M(n)\longrightarrow M(\infty)\longrightarrow 0, which is natural in M . Here, 1-I is defined so that its restriction to M(n) is the composition of (1,-\iota_{n}):M(n)\longrightarrow M(n)\oplus M(n+1) , followed by the inclusion of M(n)\oplus M(n+1) into