Symmetric Bessmertnyĭ Realizations and Field Extension Problems in Characteristic 2 - A Differential Algebra Approach
Soumya Sinha Babu, Aaron Welters · 2026
Read Symmetric Bessmertnyĭ Realizations and Field Extension Problems in Characteristic 2 - A Differential Algebra Approach by Soumya Sinha Babu, Aaron Welters (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-2605-04910.txt.
Introduction
Soumya Sinha Babu and Aaron Welters give a short algebraic proof of a realization theorem for systems theory in the special setting of characteristic two. Bessmertnyĭ realizations write certain matrix-valued functions as corners of larger symmetric linear systems. A recent proof of the symmetric case in characteristic two was algorithmic in style.
Using differential algebra, the authors define formal derivatives for rational functions over such fields and study their constants. That reduces the matrix problem to checks on diagonal entries alone. As a bonus, they prove a new result on the field extension problem and, in the scalar case, count how abundant counterexamples in characteristic two can be.
Catalog record
| CREATOR | Soumya Sinha Babu, Aaron Welters |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 51 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.04910
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Symmetric Bessmertnyĭ Realizations and Field Extension Problems in Characteristic 2 - A Differential Algebra Approach Soumya Sinha Babu, Aaron Welters Abstract We present a short, purely algebraic proof of the Symmetric Bessmertnyĭ Realization Theorem in the characteristic 2 2 case recently proved in [ 5 ] . Symmetric Bessmertnyĭ realizations are Schur complements of affine linear symmetric matrix pencils, and they arise naturally as state-space representations in linear systems theory. In contrast with the algorithmic approach in [ 5 ] , we use differential algebra: by defining formal partial derivatives on multivariate rational functions over fields of positive characteristic and considering their corresponding field of constants, we obtain scalar criteria for symmetric and homogeneous symmetric realizability in characteristic 2 2 , effectively reducing the matrix-valued problem to its diagonal entries. As a consequence, we prove a new theorem on the field extension problem for symmetric and homogeneous symmetric Bessmertnyĭ realizations. Finally, in the scalar case, we identify realizable rational functions with vector spaces over appropriate fields of constants and quantify the abundance of counterexamples in characteristic 2 2 . Keywords: symmetric Bessmertnyĭ realizations; multivariate rational matrix functions; linear matrix pencils; characteristic 2 ; formal partial derivatives; field of constants; field extension problem MSC2020: Primary 15A54; Secondary 15A22, 15B57, 12H05, 13N15, 16W10, 93B25, 93C35, 93B15, 12F20 1 Introduction Let \mathbb{F} be a field with characteristic denoted by \operatorname{char}(\mathbb{F}) and set z=(z_{1},\ldots,z_{n}) , where z_{1},\ldots,z_{n} are n indeterminates. Then \mathbb{F}[z] will denote the integral domain of all polynomials in z_{1},\ldots,z_{n} with coefficients in \mathbb{F} . Next, we will denote the field of rational functions as \mathbb{F}(z) and the subset of all homogeneous rational functions of degree d as \mathbb{F}(z)_{d} . The set of all m\times k matrices with entries in \mathbb{F} will be denoted by \mathbb{F}^{m\times k} , and if A=[a_{ij}]\in\mathbb{F}^{m\times k} is a matrix then A^{T}=[a_{ji}]\in\mathbb{F}^{k\times m} will denote its transpose {and, similarly for \mathbb{F}[z] , \mathbb{F}(z) , or \mathbb{F}(z)_{d} }. We now introduce the following important definitions. Definition 1 . An (affine) linear matrix pencil (LP) is an element A(z)\in\mathbb{F}[z]^{m\times m} of the form A(z)=A_{0}+z_{1}A_{1}+\cdots+z_{n}A_{n}, where A_{0},A_{1},\ldots,A_{n}\in\mathbb{F}^{m\times m} . If A_{j} is a symmetric matrix for each j=0,1,\ldots,n , then A(z) is called a linear symmetric matrix pencil (sLP). If A_{0}=0 , then A(z) is called a linear homogeneous matrix pencil (hLP) and if, in addition, it is a sLP then it is called a linear homogeneous symmetric matrix pencil (hsLP). Definition 2 . An element F(z)\in\mathbb{F}(z)^{k\times k} is Bessmertnyĭ realizable , i.e., has a Bessmertnyĭ realization ( BR ) [over the field \mathbb{F}(z) ], if there exists a linear matrix pencil A(z)=[A_{ij}(z)]_{i,j=1,2}\in\mathbb{F}[z]^{m\times m} partitioned into a 2\times 2 block matrix form such that F(z) is the Schur complement of A(z) with respect to A_{22}(z) , that is, \displaystyle F(z)=A(z)/A_{22}(z)=A_{11}(z)-A_{12}(z)[A_{22}(z)]^{-1}A_{21}(z), in which case we call A(z) the Bessmertnyĭ realizer of F(z) . If A(z) is a sLP then F(z) is said to have a symmetric Bessmertnyĭ realization ( SBR ) . If A(z) is a hLP then F(z) is said to have a homogeneous Bessmertnyĭ realization ( hBR ) . If A(z) is a hsLP then F(z) is said to have a homogeneous symmetric Bessmertnyĭ realization ( hSBR ) . In [ 5 ] , the following theorem was proved; we will call it the Bessmertnyĭ-Elsinger-Orzel-Welters Theorem, or the Bessmertnyĭ Realization Theorem for short. Theorem 1 (Elsinger-Orzel–Welters (2026)) . Let F(z)\in\mathbb{F}(z)^{k\times k} . Then: ( a ) F(z) has a BR. ( b ) F(z) has a hBR if and only if F(z)\in\mathbb{F}(z)_{1}^{k\times k} . ( c ) F(z) has a SBR if and only if F(z)^{T}=F(z) and any one of the following three conditions holds: ( i ) n=1; ( ii ) \operatorname{char}(\mathbb{F})\not=2; ( iii ) n\geq 2 , \operatorname{char}(\mathbb{F})=2, and the diagonal entries of F(z) are in the subspace \displaystyle\operatorname{span}\{q_{0}^{2}+z_{1}q_{1}^{2}+\cdots+z_{n}q_{n}^{2}:q_{i}\in\mathbb{F}(z)\}, (1) where the span is over the field \mathbb{F} . ( d ) F(z) has a hSBR if and only if F(z)\in\mathbb{F}(z)_{1}^{k\times k} , F(z)^{T}=F(z) , and any one of the following three conditions holds: ( i ) n=1,2; ( ii ) \operatorname{char}(\mathbb{F})\not=2; ( iii ) n\geq 3 , \operatorname{char}(\mathbb{F})=2, and the diagonal entries of F(z) are in the subspace \displaystyle\operatorname{span}\{z_{1}q_{1}^{2}+\cdots+z_{n}q_{n}^{2}:q_{i}\in\mathbb{F}(z)_{0}\}, (2) where the span is over the field \mathbb{F} . The original proof of this theorem in [ 5 ] was algorithmic and rather long (about 26 pages) with most of the effort and pages dedicated to proving statement (c)(iii) in the case k=1 . Thus, the main goal of our paper is to give an alternative, purely algebraic proof of this important theorem that simplifies and shortens the one from [ 5 ] and to also provide more clarity and insight on why the statement is true. As we shall see we can achieve this by proving the following theorem using a new approach based on differential algebra. Theorem 2 . Let F(z)\in\mathbb{F}(z)^{1\times 1} and suppose \operatorname{char}(\mathbb{F})=2 . Then the following statements are true: (i) If F(z) has a SBR then \displaystyle F(z)\in\operatorname{span}\{q_{0}^{2}+z_{1}q_{1}^{2}+\cdots+z_{n}q_{n}^{2}:q_{i}\in\mathbb{F}(z)\}, where the span is over the field \mathbb{F} . (ii) If F(z) has a hSBR then \displaystyle F(z)\in\operatorname{span}\{z_{1}q_{1}^{2}+\cdots+z_{n}q_{n}^{2}:q_{i}\in\mathbb{F}(z)_{0}\}, where the span is over the field \mathbb{F} . One way to see how differential algebra is involved when \operatorname{char}(\mathbb{F})=2 is to note the following two lemmas that we use to prove Theorem 2 . The first lemma is well-known, but for the reader’s convenience we will prove it in Sec. 2 (see Proposition 1 ). The second lemma we will prove in Sec. 3 before giving the proof of Theorem 2 . Lemma 1 . If \operatorname{char}(\mathbb{F})=p and \partial/\partial z_{1},\ldots,\partial/\partial z_{n} are the formal partial derivatives for the field \mathbb{F}(z) then \displaystyle\mathbb{F}(z^{p})=\left\{r(z)\in\mathbb{F}(z):\frac{\partial}{\partial z_{i}}r(z)=0,\text{ for each }i=1,\ldots,n\right\}, (3) where \mathbb{F}(z^{p}) is the rational field in z^{p}=(z_{1}^{p},\ldots,z_{n}^{p}) . Lemma 2 . If \operatorname{char}(\mathbb{F})=2 then \displaystyle\operatorname{span}\{q^{2}:q\in\mathbb{F}(z)\}=\mathbb{F}(z^{2}),\;\operatorname{span}\{q^{2}:q\in\mathbb{F}(z)_{0}\}=\mathbb{F}(z^{2})_{0}, (4) \displaystyle\operatorname{span}\{q_{0}^{2}+z_{1}q_{1}^{2}+\cdots+z_{n}q_{n}^{2}:q_{i}\in\mathbb{F}(z)\}=\mathbb{F}(z^{2})+z_{1}\mathbb{F}(z^{2})+\cdots+z_{n}\mathbb{F}(z^{2}), (5) \displaystyle\operatorname{span}\{z_{1}q_{1}^{2}+\cdots+z_{n}q_{n}^{2}:q_{i}\in\mathbb{F}(z)_{0}\}=z_{1}\mathbb{F}(z^{2})_{0}+\cdots+z_{n}\mathbb{F}(z^{2})_{0}, (6) where \mathbb{F}(z^{2}) is the rational field in z^{2}=(z_{1}^{2},\ldots,z_{n}^{2}) with \mathbb{F}(z^{2})_{0}:=\mathbb{F}(z^{2})\cap\mathbb{F}(z)_{0} the subfield of \mathbb{F}(z^{2}) of homogeneous degree-zero elements. There are two unexpected applications of this approach. The first is that we know when \operatorname{char}(\mathbb{F})=2 there are many more non-SBRs than SBRs and we can quantify this as follows in the case k=1 . First, \mathbb{F}(z) is a 2^{n} -dimensional vector space over \mathbb{F}(z^{2}) (this is well-known, for instance, see [ 3 , Prop. 2.2] , but for the benefit of the reader we prove it in Sec. Appendix: Auxiliary Proofs and Discussions ) and ( 5 ) is an (n+1) -dimensional subspace of \mathbb{F}(z) over \mathbb{F}(z^{2}) (by Corollary 2 ). From which it follows by Lemma 2 and Theorem 1 that there are many more non-SBR elements (within the space of all rational functions over the field \mathbb{F} ) since the “density” of non-SBR elements is 1 as \lim_{n\rightarrow\infty}\frac{n+1}{2^{n}}=0 . Similarly, it can be shown that \mathbb{F}(z)_{1} is a 2^{n-1} -dimensional vector space over \mathbb{F}(z^{2})_{0} (as proven in Sec. Appendix: Auxiliary Proofs and Discussions ) and ( 6 ) is an n -dimensional subspace of \mathbb{F}(z)_{1} over \mathbb{F}(z^{2})_{0} (by Corollary 2 ) from which it follows by Lemma 2 and Theorem 1 that there are many more non-hSBR elements (within the space of all homogeneous degree-one rational functions over the field \mathbb{F} ) since the “density” of non-hSBR elements is 1 as \lim_{n\rightarrow\infty}\frac{n}{2^{n-1}}=0 . Second, we