On a problem of Amitsur and Small

Elad Paran · 2024

Read On a problem of Amitsur and Small by Elad Paran (2024) 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-2412-06230.txt.

Introduction

Elad Paran, writing alone in 2024, settles a ring theory question that Amitsur and Small posed in 1978. The setting is a polynomial ring in two variables over a division ring, a number system where division works but multiplication need not commute. The question concerns maximal left ideals, special substructures that are as large as possible without being the whole ring.

Paran builds a concrete example. In his ring, a maximal left ideal of the two-variable ring meets the one-variable ring in a set that is not maximal there. That single construction answers the 1978 problem in the negative. Results like this matter because polynomial rings over division rings model many systems in algebra. Knowing where a natural guess fails keeps later theory honest.

Catalog record

CREATORElad Paran
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE17 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-06230

Browse the full library · About the Tower of Babel

Full text

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

On a problem of Amitsur and Small Elad Paran Abstract . We construct an example of a division ring D D and a maximal left ideal M M in the polynomial ring D ⁡ [ x , y ] D[x,y] in two central variables over D D , such that the intersection M ∩ D ⁡ [ x ] M\cap D[x] is not a maximal left ideal in D ⁡ [ x ] D[x] . This resolves a ring-theoretic problem of Amitsur and Small raised in 1978. 1. Introduction Let D be a division ring and let M be a maximal left ideal in the ring D[x_{1},\ldots,x_{n}] of polynomials in n central variables over D . In [ 3 ] , Amitsur and Small prove a form of Nullstellensatz for such ideals: The left quotient module D[x_{1},\ldots,x_{n}]/M is finite over D . A key lemma in their proof states that the intersection of M and D[x_{k}] is non-zero, for any 1\leq k\leq n . In the case where D is a field, one can of course say more: The intersection M\cap D[x_{k}] is a maximal ideal. This is a classical fact, essential to dimension theory. However, in the case where D is an arbitrary division ring, Amitsur and Small write “We remark that we are unable to show that maximal left ideals in D[x_{1},\ldots,x_{n}] intersect D[x_{1},\ldots,x_{k}] , k<n , in maximal or even semi-maximal left ideals” [ 3 , p. 356] . The second part of this problem was solved by Small and Robson in [ 14 , Proposition 5.3] , who showed that if M is a semi-maximal left ideal in D[x_{1},\ldots,x_{n}] then M\cap D[x_{1},\ldots,x_{k}] is semi-maximal for any 1\leq k\leq n (see also [ 11 , Theorem 6.8, p. 360] ). However, the first part of the problem (whether maximal ideals contract to maximal ideals) remained unresolved by the mention works. Some related results are given by Resco in [ 13 ] , who notes that “It remains unknown, however, whether a maximal right ideal need contract to a maximal right ideal” [ 13 , p. 70] 1 1 1 Some of the mentioned papers work with left ideals, and some with right ideals, but of course, the two notions are interchangeable. . Years later, in a paper of Rowen from 1995, he mentions [ 15 , p. 2272] that the solution to this question is negative “as shown recently by Amitsur and Small”. However, no reference is given by Rowen, and there does not seem to be a paper from that period containing a solution. The author had written to Small, who does not recall that he or Amitsur had solved this problem, and to Rowen, who also does not recall what the mentioned sentence in his paper refers to. (Notably, Amitsur had died at September 1994, the same month when Rowen’s paper was submitted for publication.) In this note we resolve the problem of Amitsur and Small (or re-discover its solution, had Amitsur and Small already found it), by providing an example of a division ring D and a maximal left ideal M in D[x,y] , such that M\cap D[x] is not maximal. This example is described in § 2 below, where a simple recipe for constructing such examples is given by Proposition 2.3 . Then in § 3.1 we discuss some connections of this problem to recent works, leading to several natural follow-up problems. 2. Maximal left ideals that contract to non-maximal left ideals Throughout this note, given a division ring D , we denote by D[x_{1},\ldots,x_{n}] the ring of polynomials in n central variables over D , and in the case of one or two variables, we shall write D[x] or D[x,y] , respectively. Given a polynomial f=f_{n}x^{n}+f_{n-1}x^{n-1}+\ldots+f_{1}x+f_{0}\in D[x] and an element a\in D , we denote by f(a) the usual substitution f(a)=f_{n}a^{n}+f_{n-1}a^{n-1}+\ldots+f_{1}a+f_{0} . It is well-known that the substitution map f\mapsto f(a) is generally not a homomorphism from D[x] to D , however it satisfies the following “product formula”: If f,g\in D[x] then (fg)(a)=0 if g(a)=0 , and (fg)(a)=f\big(a^{g(a)}\big)g(a) if g(a)\neq 0 , where a^{g(a)} denotes the conjugation g(a)ag(a)^{-1} (see [ 9 , Proposition 16.3, p. 263] ). We say that a\in D is a zero of f\in D[x] if f(a)=0 . This is equivalent to x-a being a right-hand factor of f [ 9 , Proposition 16.2, p. 262] . We shall need the following lemma: Lemma 2.1 . Let a,b,c be elements of a division ring D such that (ab)c=c(ab) and (a+b)c=c(a+b) . Then the left ideal in D[x,y] generated by (x-a)(x-b),y-c is a proper left ideal. Proof. Suppose to the contrary that there exist f,g\in D[x,y] such that f\cdot(x-a)(x-b)+g\cdot(y-c)=1. The polynomial ring D[x] is an Ore domain, hence admits a unique quotient division ring, which we denote by D(x) . Let us view the above presented equality as taking place in the ring D(x)[y] . Consider the substitution map y\mapsto c from D(x)[y] to D(x) . Applying this map to the presented equality, we get, using the substitution formula mentioned above, that f(c^{(x-a)(x-b)})\cdot(x-a)(x-b)=1. (Let us emphasize that in this equality (x-a)(x-b) serves as a “constant” – an element of the division ring D(x) .) Now, by our assumptions, c commutes with a+b and with ab , hence c commutes with (x-a)(x-b)=x^{2}-(a+b)x+ab . Thus in the ring D(x) , we have c^{(x-a)(x-b)}=c . But this means that f(c^{(x-a)(x-b)}) , which a priori is an element of D(x) , equals f(c) , which belongs to D[x] . This contradicts the last presented equality, since then f(c^{(x-a)(x-b)})\cdot(x-a)(x-b)=f(c)\cdot(x-a)(x-b)\in D[x] is either 0 or of degree at least 2 in x . ∎ Given a division ring D and n\geq 1 , we shall denote by D_{c}^{n} the subspace D_{c}=\{(a_{1},\ldots,a_{n})\in D^{n}{|}a_{i}a_{j}=a_{j}a_{i}\hbox{ for all }1\leq i,j\leq n\} of D^{n} . Lemma 2.2 . Let a_{1},\ldots,a_{n} be elements of a division ring D . Then the left ideal in D[x_{1},\ldots,x_{n}] generated by x_{1}-a_{1},\ldots,x_{n}-a_{n} is a proper left ideal if and only if (a_{1},\ldots,a_{n})\in D_{c}^{n} . Moreover, if this left ideal is proper, then it is maximal. Proof. The first claim is given by in [ 1 , Lemma 2.1] , and the second claim is given by [ 1 , Proposition 2.2] – both claims are formulated there for the case where D is the real quaternion algebra, but their proofs hold verbatim for any division ring. ∎ We remark that by the above lemma, the points of D_{c}^{n} are precisely the points for which substitution in polynomials in D[x_{1},\ldots,x_{n}] is well-defined, as discussed in [ 1 , §2] . Next, we have: Proposition 2.3 . Let D be a division ring and let a,b,c be elements of D satisfying the following conditions: a) abc=cab . b) c(a+b)=(a+b)c . c) Every zero in D of the polynomial (x-a)(x-b) does not commute with c . Then the left ideal M generated by (x-a)(x-b),y-c in D[x,y] is a maximal left ideal, whose intersection with D[x] is not a maximal left ideal in D[x] . Proof. By the preceding lemma M is a proper left ideal in D[x,y] . To show that it is maximal, suppose that f\in D[x,y] does not belong to M . We must show that M+D[x,y]f=D[x,y]

Read the complete text