Infinite Versions of Hilbert’s Nullstellensatz

A. Bernhard Zeidler · 2025

Read Infinite Versions of Hilbert’s Nullstellensatz by A. Bernhard Zeidler (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-2502-10566.txt.

Introduction

A Bernhard Zeidler compiles infinite-dimensional versions of a classic algebra theorem, Hilbert's Nullstellensatz, in a solo 2025 paper. The classic result links systems of polynomial equations to the ideals that vanish on their solutions, and it anchors much of algebraic geometry.

Zeidler gathers a long list of forms of the theorem that stay valid when the number of variables grows without bound. He shows the forms match each other. He also proves that the strong form of the theorem persists in very large polynomial rings. A short paper of this kind serves specialists as a reference map for a deep and scattered topic.

Catalog record

CREATORA. Bernhard Zeidler
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE26 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-10566

Browse the full library · About the Tower of Babel

Full text

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

Infinite Versions of Hilbert’s Nullstellensatz A. Bernhard Zeidler Abstract . We compile a long list of equivalent formulations of Hilbert’s Nullstellensatz in infinite dimensions, and prove a persistence result for the strong Nullstellensatz in large polynomial rings. 2020 Mathematics Subject Classification 13A15,14A04 1. Presentation of the Results This note is devoted to Hilbert’s Nullstellensatz for polynomial rings of infinite Krull dimension. We collect classical and recent equivalences (see Theorem 1.1 ) and provide complete, elementary proofs (see Section 2 ) of which some are new or generalize existing ones. Moreover, given an ideal in a polynomial ring of dimension lower than the cardinality of the base field, we show that its extensions to larger polynomial rings satisfy the strong Nullstellensatz, even if their cardinality exceeds that of the base field (see Corollary 1.5 ). Let us fix the setting and notation: All rings and algebras are assumed to be commutative, non-zero, with unit and their homomorphisms are assumed to preserve the unit element. Given a ring F and a (not necessarily finite) set I , let A:=F[t_{i}\mid i\in I] be the polynomial algebra over F . The set of maximal ideals of A is denoted by \mbox{\rm{Smax}}(A) . Any ideal {\scalebox{1.10}{$\mathfrak{a}$}}\unlhd A defines the zero set \mathds{V}({\scalebox{1.10}{$\mathfrak{a}$}})\subseteq\,F^{I} and any X\subseteq\,F^{I} defines the vanishing ideal \mathds{I}(X)\unlhd A . The vanishing ideal of a single element x\in F^{I} is denoted by {\scalebox{1.10}{$\mathfrak{m}$}}_{x} . Theorem 1.1 . Let F be a non-zero ring, I be a non-empty index set and the polynomial algebra A=F[t_{i}\mid i\in I] . Then the following statements are equivalent: (a) F is an algebraically closed field and |I|<|F| holds for the cardinalities. (b) Any proper ideal {\scalebox{1.10}{$\mathfrak{a}$}}\unlhd A admits an F -algebra homomorphism A/{\scalebox{1.10}{$\mathfrak{a}$}}\to F . (c) Any maximal ideal {\scalebox{1.10}{$\mathfrak{m}$}}\unlhd A admits an F -algebra homomorphism A/{\scalebox{1.10}{$\mathfrak{m}$}}\to F . (d) Any maximal ideal {\scalebox{1.10}{$\mathfrak{m}$}}\unlhd A is of the form {\scalebox{1.10}{$\mathfrak{m}$}}={\scalebox{1.10}{$\mathfrak{m}$}}_{x} for some x\in F^{I} . (e) The assignment x\mapsto{\scalebox{1.10}{$\mathfrak{m}$}}_{x} defines a bijection F^{I}\to\mbox{\rm{Smax}}(A) . (f) F is a field and any {\scalebox{1.10}{$\mathfrak{m}$}}\in\mbox{\rm{Smax}}(A) and i_{1},\dots,i_{n}\in I admit x_{i_{k}}\in F with {\scalebox{1.10}{$\mathfrak{m}$}}\cap F[t_{i_{1}},\ldots,t_{i_{n}}]\ =\ \langle t_{i_{1}}-x_{i_{1}},\,\dots,\,t_{i_{n}}-x_{i_{n}}\rangle. (g) F is a field and for any ideal {\scalebox{1.10}{$\mathfrak{a}$}}\unlhd A we get \mathds{V}({\scalebox{1.10}{$\mathfrak{a}$}})=\emptyset\Leftrightarrow{\scalebox{1.10}{$\mathfrak{a}$}}=A . (h) F is a field and any field extension F\subseteq\,K=F[u_{i}\mid i\in I] is trivial. (i) For any ideal {\scalebox{1.10}{$\mathfrak{a}$}}\unlhd A we have \sqrt{{\scalebox{1.10}{$\mathfrak{a}$}}}=\mathds{I}\mathds{V}({\scalebox{1.10}{$\mathfrak{a}$}}) . The presumably most popular equivalence, (a) \Leftrightarrow (d), is attributed to Kaplansky, apparently without concrete reference (compare [ 6 , p. 407, footnote 1] ). We provide a direct proof here. Statement (b) is the chromatic version [ 3 , Thm. 6.1] by Burklund, Schlank and Yuan. We present a slight modification of their proof, to match our formulation. Statement (h) is taken from [ 6 ] and (i) is gained from (d) classically via Rabinowitsch’s trick . We neither saw (c) nor (f) before, but they’re elementary. The rest is common folklore. Remark 1.2 . It is clear, that the strong Nullstellensatz 1.1 .(i) gives rise to a one-to-one correspondence between radical ideals of A=F[t_{i}\mid i\in I] and algebraic subsets of F^{I} . All we need is F being an algebraically closed field and |I|<|F| . Remark 1.3 . Note that we assumed F to be a field in (a), (f), (g) and (h) only. In the other statements, this is implied. In fact it is possible to drop this assumption in (f) as well, by allowing n=0 . In this case we have {\scalebox{1.10}{$\mathfrak{m}$}}\cap F=0 for any {\scalebox{1.10}{$\mathfrak{m}$}}\in\mbox{\rm{Smax}}(A) and this is equivalent to F being a field. Remark 1.4 . In 1952 S. Lang proved a similar equivalence: If F is an algebraically closed field and P\subseteq\,F is the prime field of F , then property (d) holds true if and only if I is finite or |I|<\mbox{\rm{trdeg}}(F:P) . The latter is much the same as (a), since the transcendence degree satisfies |F|\ =\ \mbox{\rm{max}}\hskip-1.70709pt\left\{\,|\mathds{N}|,\,\mbox{\rm{trdeg}}(F:P)\,\right\} Corollary 1.5 . Let F be an algebraically closed field and I\subseteq\,J be index sets, such that I is non-empty and |I|<|F| . If now {\scalebox{1.10}{$\mathfrak{u}$}}\unlhd F[t_{j}\mid j\in J] is an ideal of the form {\scalebox{1.10}{$\mathfrak{u}$}}={\scalebox{1.10}{$\mathfrak{a}$}}\,F[t_{j}\mid j\in J] for some ideal {\scalebox{1.10}{$\mathfrak{a}$}}\unlhd F[t_{i}\mid i\in I] , then we still get the identity \mathds{I}\,\mathds{V}({\scalebox{1.10}{$\mathfrak{u}$}})\ =\ \sqrt{{\scalebox{1.10}{$\mathfrak{u}$}}} 2. Proof of Theorem 1.1 and Corollary 1.5 It is well-known, that: If F is a field and I is a non-empty index set, then the ideal {\scalebox{1.10}{$\mathfrak{m}$}}_{x}=\mathds{V}(x)=\langle t_{i}-x_{i}\mid i\in I\rangle is a maximal ideal of A=F[t_{i}\mid i\in I] and the following map is well-defined and injective F^{I}\,\,\hookrightarrow\,\,\mbox{\rm{Smax}}(A)\;:\;x\,\mapsto\,{\scalebox{1.10}{$\mathfrak{m}$}}_{x} Definition 2.1 . We say that A is geometric , iff this map is surjective. That is iff every maximal ideal {\scalebox{1.10}{$\mathfrak{m}$}}\in\mbox{\rm{Smax}}(A) of A is of the form {\scalebox{1.10}{$\mathfrak{m}$}}={\scalebox{1.10}{$\mathfrak{m}$}}_{x} for some x\in F^{I} . Proposition 2.2 . Let F be a field, I a non-empty index set and A=F[t_{i}\mid i\in I] the F -algebra of polynomials. Then we get the following statements (i) If I is finite and F is an algebraically closed field, then A is geometric. (ii) If I is infinite, then A is geometric if and only if A[s] is geometric. In this proposition (i) is the classic (weak) Nullstellensatz, of which numerous proofs exist: A short, elegant proof has been given by Zariski in [ 8 ] using the Noether Normalisation theorem. Krull and Goldman introduced the theory of Jacobson resp. Hilbert rings, which can be seen as a generalization of this, e.g. refer to [ 4 ] . In 1999 Munshi reduced the techniques of Jacobson rings to a minimum - refer to [ 7 ] , in 2006 Arrondo gave an astoundingly simple proof in [ 1 ] . And in 2018 a proof using Gröbner bases has been provided by Glebsky and Rubio-Barrios [ 5 ] . Statement (ii) is quite obvious, we took it from [ 6 ] . Let us split the proof of Theorem 1.1 into parts. The first insight is due to [ 3 ] : Claim 1: Under assumption (b), F is an algebraically closed field. Proof of claim 1: Let us first prove, that F is a field: Consider any a\in F with a\not\in F^{\ast} and define the ideal {\scalebox{1.10}{$\mathfrak{a}$}}:=\langle t_{i}\mid i\in I\rangle+aA . Then the homomorphism A\,\,\twoheadrightarrow\,\,\raisebox{3.0pt}{$F$}\hskip-1.0pt\raisebox{-1.0pt}{\scalebox{1.5}{/}}\hskip 1.07pt\hskip-2.0pt\raisebox{-3.0pt}{$aF$}\;:\;f\,\mapsto\,f(0)+aF is surjective and has kernel \mathfrak{a} , which is a proper ideal, since a\not\in F^{\ast} . Therefore A/{\scalebox{1.10}{$\mathfrak{a}$}} and F/aF are isomorphic as F -algebras and by (b) there is a homomorphism \varphi of the form \varphi:A/{\scalebox{1.10}{$\mathfrak{a}$}}\to F . This gives rise to a homomorphism of F -algebras \raisebox{3.0pt}{$F$}\hskip-1.0pt\raisebox{-1.0pt}{\scalebox{1.5}{/}}\hskip 1.07pt\hskip-2.0pt\raisebox{-3.0pt}{$aF$}\,\,\stackrel{{\scriptstyle\raisebox{-1.50694pt}[0.0pt]{$\sim$}}}{{\longrightarrow}}\,\,\raisebox{3.0pt}{$A$}\hskip-1.0pt\raisebox{-1.0pt}{\scalebox{1.5}{/}}\hskip 1.07pt\hskip-2.0pt\raisebox{-3.0pt}{${\scalebox{1.10}{$\mathfrak{a}$}}$}\,\to\,F Explicitly this is F/aF\to F:b+aF\mapsto b1_{A}+{\scalebox{1.10}{$\mathfrak{a}$}}\mapsto\varphi(b1_{A}+{\scalebox{1.10}{$\mathfrak{a}$}})=b . But from a+aR\mapsto a and a+aR\mapsto 0 we find a=0 and therefore F is a field. Next we have to show, that F even is algebraically closed: Fix any j\in I , let H:=I\setminus\left\{\,j\,\right\} and consider an arbitrary, non-constant polynomial f\in F[t_{j}]\subseteq\,A . Now define the ideals {\scalebox{1.10}{$\mathfrak{h}$}}:=\langle t_{i}\mid i\in H\rangle and {\scalebox{1.10}{$\mathfrak{a}$}}:=fA+{\scalebox{1.10}{$\mathfrak{h}$}} . Then it is clear, that F[t_{j}]\,\stackrel{{\scriptstyle\raisebox{-1.50694pt}[0.0pt]{$\sim$}}}{{\longrightarrow}}\,A/{\scalebox{1.10}{$\mathfrak{h}$}}:g\mapsto g+{\scalebox{1.10}{$\mathfrak{h}$}} is an isomorphism of R -algebras, that maps fF[t_{j}] to {\scalebox{1.10}{$\mathfrak{a}$}}/{\scalebox{1.10}{$\mathfrak{h}$}} . And therefore we find the isomorphism \raisebox{3.0pt}{$F[t_{j}]$}\hskip-1.0pt\raisebox{-1.0pt}{\scalebox{1.5}{/}}\hskip 1.07pt\hskip-2.0pt\raisebox{-3.0pt}{$fF[t_{j}]$}\;\,\stackrel{{\scriptstyle\raisebox{-1.50694pt}[0.0pt]{$\sim$}}}{{\longrightarrow}}\,\;\raisebox{3.0pt}{$A/{\scalebox{1.10}{$\mathfrak{h}$}}$}\hskip-1.0pt\raisebox{-1.0pt}{\scalebox{1.5}{/}}\hskip 1.07pt\hskip-2.0pt\raisebox{-3.0pt}{${\scalebox{1.10}{$\mathfrak{a}$}}/{\scalebox{1.10}{$\mathfrak{h}$}}$}\;\,\stackrel{{\scriptstyle\raisebox{-1.50694pt}[0.0pt]{$\sim$}}}{{\longrightarrow}}\,\;\raisebox{3.0pt}{$A$}\hskip-1.0pt\raisebox{-1.0pt}{\scalebox{1.5}{/}}\hskip 1.07pt\hskip-2.0pt\raisebox{-3.0pt}{${\scalebox{1.10}{$\mathfrak{a}$}}$}\;:\;g+fF[t_{j}]\,\mapsto\,g+{\scalebox{1.10}{$\mathfrak{a}$}} Now F[t_{j}]/fF[t_{j}] is a \deg(f) -dimensional F -vector space and as f is non-constant, we have \deg(f)\geq 1 . In particular F[t_{j}]/fF[t_{j}]\not=0 and therefore A/{\scalebox{1.10}{$\mathfrak{a}$}}\not=0 . That is \mathfrak{a} is a proper

Read the complete text