Kato-Milne cohomology group over rational function fields in characteristic 2, II

Ahmed Laghribi, Trisha Maiti · 2025

Read Kato-Milne cohomology group over rational function fields in characteristic 2, II by Ahmed Laghribi, Trisha Maiti (2025) 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-2504-15957.txt.

Introduction

Ahmed Laghribi and Trisha Maiti, two authors writing in 2025, extend an exact sequence in algebra. They work in a cohomology theory adapted to fields of characteristic two. Their goal is a sequence like a known one from the theory of quadratic forms. The result generalizes an earlier Milnor-style sequence in the same setting. The paper is the second part of a series on these groups over rational function fields.

Catalog record

CREATORAhmed Laghribi, Trisha Maiti
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE88 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-15957

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Kato-Milne cohomology group over rational function fields in characteristic 2, II Ahmed Laghribi, Trisha Maiti Abstract . Our aim in this paper is to prove in the setting of Kato-Milne cohomology in characteristic 2 2 an exact sequence which is analogue to the Milnor-Scharlau sequence [ 8 , Theorem 6.2] . This is an extension of the Milnor exact sequence proved in [ 6 ] . 1991 Mathematics Subject Classification 11E04, 11E81, 13N05 1. Introduction In [ 3 , Thoerem 6.2] Aravire and Jacob determined the Witt group of nonsingular quadratic forms W_{q}(\mathcal{F}(x)) , where \mathcal{F} is a field of characteristic 2 and x is an indeterminate over \mathcal{F} . To this end, they proved the analogue of the Milnor-Scharlau exact sequence [ 8 , Theorem 6.2] known in characteristic not 2 . Since nonsingular quadratic forms in characteristic 2 are related to differential forms by Kato’s isomorphism ( 1 ), it is natural to ask for an analogue of the Milnor-Scharlau exact sequence in the setting of Kato-Milne cohomology. This is what we will accomplish in this paper by following the same process developed in [ 3 ] . In [ 6 ] we already proved, in the setting of Kato-Milne cohomology, the analogue of the Milnor exact sequence [ 7 ] . So in this paper, we extend that short exact sequence using the transfers described in Definition 3.5 . To state our main result, we need to recall some backgrounds on quadratic forms, differential forms and Teichmüller lifting. Let F be a field of characteristic 2 . Up to isometry, any quadratic form \varphi over F can be written up to isometry as follows: \varphi\simeq[a_{1},b_{1}]\perp\cdots\perp[a_{r},b_{r}]\perp\langle c_{1}\rangle\perp\cdots\perp\langle c_{s}\rangle, where [a,b] and \langle c\rangle denote the quadratic forms aX^{2}+XY+bY^{2} and \langle c\rangle , respectively. The quadratic form \varphi is said to be nonsingular if s=0 . For a_{1},\cdots,a_{n}\in F^{*}=:F\setminus\{0\} , let \left<a_{1},\cdots,a_{n}\right>_{b} denote the diagonal bilinear form given as follows: ((x_{1},\cdots,x_{n}),(y_{1},\cdots,y_{n}))\mapsto\sum_{i=1}^{n}a_{i}x_{i}y_{i}. Let W(F) be the Witt ring of nondegenerate symmetric bilinear forms over F , and W_{q}(F) the Witt group of nonsingular quadratic forms over F . It is known that W_{q}(F) is endowed with a W(F) -module structure induced by tensor product [ 4 ] . Let I(F) be the fundamental ideal of W(F) . For m\in\mathbb{N}_{0} , let I^{m}(F) be the m -th power of I(F) . This ideal is additively generated by the m -fold bilinear Pfister forms \langle\langle a_{1},\ldots,a_{m}\rangle\rangle_{b}:=\langle 1,a_{1}\rangle_{b}\otimes\ldots\otimes\langle 1,a_{m}\rangle_{b} for a_{i}\in F^{*} (we take I^{0}(F)=W(F) ). Similarly, the subgroup I_{q}^{m+1}(F)=I^{m}(F)\otimes W_{q}(F) is additively generated by the (m+1) -fold quadratic Pfister forms \langle\langle a_{1},\ldots,a_{m};b]]:=\langle\langle a_{1},\ldots,a_{m}\rangle\rangle_{b}\otimes[1,a] , for a_{i}\in F^{*} and b\in F . Note that I_{q}^{1}(F)=W_{q}(F) . The notation \overline{I}_{q}^{m+1}(F) means the quotient I_{q}^{m+1}(F)/I_{q}^{m+2}(F) . For any integer m\geq 1 , let \Omega_{F}^{m} be the space of m -differential forms over F . We also take \Omega_{F}^{0}=F and \Omega_{F}^{m}=0 for m<0 . Let H_{2}^{m+1}(F) be the Kato-Milne cohomology group of degree m , defined by H_{2}^{m+1}(F)=\Omega_{F}^{m}/(\dd\Omega_{F}^{m-1}+\wp(\Omega_{F}^{m})), where \dd and \wp denote the differential and the Artin-Schreier operators, respectively. We refer to [ 6 , Introduction] for definitions. For each m\in\mathbb{N} , we have Kato’s isomorphism f(F)_{m+1}:H_{2}^{m+1}(F)\xrightarrow{}\overline{I}_{q}^{m+1}(F) defined by (1) f(F)_{m+1}\left(\overline{x\frac{\dd x_{1}}{x_{1}}\wedge\ldots\wedge\frac{\dd x_{m}}{x_{m}}}\right)=\overline{\langle\langle x_{1},\ldots,x_{m};x]]}. Let v_{p}:F\xrightarrow{}\mathbb{Z} be a discrete valuation of F with p as an uniformizer. We use \overline{F}_{p} and F_{p} to denote the residue field and the completion of F , respectively. Now we define the group W_{1}(H_{2}^{m+1}(F_{p})) as the cokernel \text{coker}(\alpha^{\prime}:H_{2}^{m+1}(\overline{F}_{p})\xrightarrow{}H_{2}^{m+1}(F_{p})), where \alpha^{\prime} is naturally induced by a Teichmüller lifting 1 1 1 We refer to [ 6 , Appendix] for an explicit description of Teichmüller lifting. \alpha:\overline{F}_{p}\xrightarrow{}F_{p} . Moreover, the group W_{1}(H_{2}^{m+1}(F_{p})) is independent from the choice of a Teichmüller lifting [ 6 , Corollary 2.6] . The inclusion map F\xrightarrow{}F_{p} induces a homomorphism \beta_{1}:H_{2}^{m+1}(F)\xrightarrow{}H_{2}^{m+1}(F_{p}) . Let \beta_{2}:H_{2}^{m+1}(F_{p})\xrightarrow{}W_{1}(H_{2}^{m+1}(F_{p})) be the homomorphism induced by projection. Then, we have a (residue) map \partial_{p}:H_{2}^{m+1}(F)\xrightarrow{}W_{1}(H_{2}^{m+1}(F_{p})) defined by \partial_{p}=\beta_{2}\circ\beta_{1}. Now consider F=\mathcal{F}(x) the rational function field in one variable x over a field \mathcal{F} of characteristic 2 . Let p\in\mathcal{F}[x] be a monic irreducible polynomial or p=\frac{1}{x} . Let F_{p} ( resp. \overline{F}_{p} ) denote the completion ( resp. the residue field) of F with respect to the p -adic valuation. By [ 6 , Theorem 4.2] , H_{2}^{m}(\overline{F_{p}}) is isomorphic to a quotient group of W_{1}(H_{2}^{m+1}(F_{p})) . Moreover, the Scharlau’s transfer s_{p}^{\prime}:W_{q}(\overline{F_{p}})\xrightarrow{}W_{q}(\mathcal{F}) , along with Kato’s isomorphism ( 1 ) induce the homomorphism s_{p}^{\prime\prime}:H_{2}^{m}(\overline{F_{p}})\xrightarrow{}H_{2}^{m}(\mathcal{F}) . Consequently, we have the transfer map s_{p}^{*}:W_{1}(H_{2}^{m+1}(F_{p}))\xrightarrow{}H_{2}^{m}(\mathcal{F}) , which is the composition map W_{1}(H_{2}^{m+1}(F_{p}))\xrightarrow{}H_{2}^{m}(\overline{F_{p}})\xrightarrow{}H_{2}^{m}(\mathcal{F}) (we refer to Definition 3.5 for details). Our main result is the following theorem: Theorem 1.1 . Let \mathcal{F} be a field of characteristic 2 and F=\mathcal{F}(x) the rational function field in one variable x over \mathcal{F} . Then, the following sequence in exact: 0\xrightarrow{}H_{2}^{m+1}(\mathcal{F})\xrightarrow{{\color[rgb]{0,0,0}i}}H_{2}^{m+1}(F)\xrightarrow{\bigoplus\partial_{p}}\bigoplus_{p}W_{1}(H_{2}^{m+1}(F_{p}))\xrightarrow{\bigoplus s^{*}_{p}}H_{2}^{m}(\mathcal{F})\xrightarrow{}0, where i is induced by the inclusion \mathcal{F}\xrightarrow{}F and p varies over \frac{1}{x} and all monic irreducible polynomials of \mathcal{F}[x] . Now we briefly summarize the proof of this theorem, for which we will adapt some arguments from [ 3 ] . By [ 6 , Theorems 4.9, 4.10] , it suffices to prove that {\rm Im}(\oplus\partial_{p})={\rm ker}(\oplus s^{*}_{p}) . In Section 2 , we prove a decomposition of W_{1}(H_{2}^{m+1}(F_{p})) which will be needed in Section 3 to describe the transfer maps s_{p}^{*} . In Section 4 , we prove that \oplus(s_{q}^{*}\circ\partial_{q}) vanishes on L_{0}+L_{0}\wedge\overline{\frac{\dd p}{p}} for any fixed monic irreducible polynomial p . In fact, the map \oplus(s_{q}^{*}\circ\partial_{q}) vanishes on H_{2}^{m+1}(\mathcal{F}(x)) (Theorem 5.1 ) , which proves that {\rm Im}(\oplus\partial_{q})\subset{\rm ker}(\oplus s^{*}_{q}) . Then, by some calculations we get the other inclusion, which completes the proof of the theorem. 2. A decomposition of W_{1}(H_{2}^{m+1}(F_{p})) Let \mathcal{F} be a field of characteristic 2 . A subset \{t_{i}\hskip 2.84526pt|\hskip 2.84526pti\in\mathcal{I}\}\subset\mathcal{F} is said to be a 2 -basis of \mathcal{F} if \{\prod_{i\in\mathcal{I}}t_{i}^{\epsilon_{i}}\hskip 2.84526pt|\hskip 2.84526pt\epsilon_{i}\in\{0,1\} and \epsilon_{i}=0 for almost all i\in\mathcal{I}\} is an \mathcal{F}^{2} -basis of \mathcal{F} , where \mathcal{I} is an ordered set. When F=\mathcal{F}(x) and p a monic irreducible polynomial, the group W_{1}(H_{2}^{m+1}(F_{p})) is given in [ 6 , Theorem 4.2] . In this section we prove another decomposition of W_{1}(H_{2}^{m+1}(F_{p})) , which will be needed in Section 3 to introduce the transfer maps s_{p}^{*} that we talked about before. Let L_{p} be a field of characteristic 2 , complete with respect to a discrete valuation v_{p}:L_{p}\xrightarrow{}\mathbb{Z} and p an uniformizer. Let \overline{L}_{p} denote the residue field of L_{p} , and \overline{A}_{p}=\{s_{i}\hskip 2.84526pt|\hskip 2.84526pti\in\mathcal{I}^{\prime}\} a 2 -basis of \overline{L}_{p} , where \mathcal{I}^{\prime} is an ordered set. Set \mathcal{I}=\mathcal{I}^{\prime}\cup\{\infty\} , where s_{\infty} denotes p . Now we extend the ordering of \mathcal{I}^{\prime} to an ordering of \mathcal{I} by taking i<\infty for all i\in\mathcal{I}^{\prime} . Since L_{p} is complete, we may consider \overline{L}_{p}\subset L_{p} and {A}_{p}=\{s_{i}\hskip 2.84526pt|\hskip 2.84526pti\in\mathcal{I}\} as a 2 -basis of L_{p} . We define two groups \overline{\mathcal{T}}_{p}=\bigoplus\limits_{|\mathcal{I}^{\prime}|}(\mathbb{Z}/{2\mathbb{Z}}) and {\mathcal{T}}_{p}=\bigoplus\limits_{|\mathcal{I}|}(\mathbb{Z}/{2\mathbb{Z}}) . Now \mathcal{T}_{p} has lexicographically ordering with respect to 0<1 . Namely, for any I=(I_{i})_{i\in\mathcal{I}} and J=(J_{i})_{i\in\mathcal{I}}\in\mathcal{T}_{p} we have I<J if there exists an element k\in\mathcal{I} such that I_{k}<J_{k} and I_{i}=J_{i} for all i>k (we consider the ordering from right to left). Since \overline{\mathcal{T}}_{p} can be considered as a subgroup of {\mathcal{T}}_{p} , the group \overline{\mathcal{T}}_{p} is also lexicographically ordered. Let L_{p}^{\prime}=\{l\in L_{p}\hskip 2.84526pt|\hskip 2.84526ptv_{p}(l)=0\} and for I=(I_{i})_{i\in\mathcal{I}}\in{{\mathcal{T}}_{p}} , let take

Read the complete text