Generalized $E$-algebras via $\lambda$-calculus I

Rüdiger Göbel¹ ; Saharon Shelah²

¹ Fachbereich Mathematik Universität Duisburg-Essen D-45117 Essen, Germany
² Institute of Mathematics Hebrew University Jerusalem, Israel and} Rutgers University New Brunswick, NJ 08903, U.S.A.

Fundamenta Mathematicae, Tome 192 (2006) no. 2, pp. 155-181.

Résumé

An $R$-algebra $A$ is called an $E(R)$-algebra if the canonical homomorphism from $A$ to the endomorphism algebra $\mathop{\rm End}_RA$ of the $R$-module ${}_RA$, taking any $a \in A$ to the right multiplication $a_r\in \mathop{\rm End}_RA$ by $a$, is an isomorphism of algebras. In this case ${}_RA$ is called an $E(R)$-module. There is a proper class of examples constructed in \cite{DMV}. $E(R)$-algebras arise naturally in various topics of algebra. So it is not surprising that they were investigated thoroughly in the last decades; see \cite{DG2,DV,Fa,FHR,GG, GSS, GSS1,GSt,Pi,PV}. Despite some efforts (\cite{GSS1, DV}) it remained an open question whether proper generalized $E(R)$-algebras exist. These are $R$-algebras $A$ isomorphic to $\mathop{\rm End}_RA$ but not under the above canonical isomorphism, so not $E(R)$-algebras. This question was raised about 30 years ago (for $R=\mathbb Z$) by Schultz \cite{Sch} (see also Vinsonhaler \cite{Vi}). It originates from Problem 45 in Fuchs \cite{Fu0}, that asks for a characterization of the rings $A$ for which $A\cong \mathop{\rm End}_\mathbb Z A$ (as rings). We answer Schultz's question, thus contributing a large class of rings for Fuchs' Problem 45 which are not $E$-rings. Let $R$ be a commutative ring with an element $p\in R$ such that the additive group $R^+$ is $p$-torsion-free and $p$-reduced (equivalently $p$ is not a zero-divisor and $\bigcap_{n\in\omega} p^nR=0$). As explained in the introduction we assume that either $|R| 2^{\aleph_0}$ or $R^+$ is free (see Definition 1.1).The main tool is an interesting connection between $\lambda$-calculus (used in theoretical computer science) and algebra. It seems reasonable to divide the work into two parts; in this paper we work in $ V= L$ (Gödel's universe) where stronger combinatorial methods make the final arguments more transparent. The proof based entirely on ordinary set theory (the axioms of ZFC) will appear in a subsequent paper \cite{GS}. However the general strategy will be the same, but the combinatorial arguments will utilize a prediction principle that holds under ZFC.

Détail

DOI : 10.4064/fm192-2-5

Keywords: r algebra called algebra canonical homomorphism endomorphism algebra mathop end r module taking right multiplication mathop end isomorphism algebras called module there proper class examples constructed cite dmv algebras arise naturally various topics algebra surprising investigated thoroughly decades see cite fhr gss gss gst despite efforts cite gss remained question whether proper generalized algebras exist these r algebras isomorphic mathop end under above canonical isomorphism algebras question raised about years ago mathbb schultz cite sch see vinsonhaler cite originates problem fuchs cite asks characterization rings which cong mathop end mathbb rings answer schultzs question contributing large class rings fuchs problem which e rings commutative ring element additive group p torsion free p reduced equivalently zero divisor bigcap omega explained introduction assume either aleph see definition main tool interesting connection between lambda calculus theoretical computer science algebra seems reasonable divide work parts paper work dels universe where stronger combinatorial methods make final arguments transparent proof based entirely ordinary set theory axioms zfc appear subsequent paper cite however general strategy combinatorial arguments utilize prediction principle holds under zfc