Galois theory (Q5900395)

The first edition of this book appeared in 2006 and duly received a highly positive review by W. Kleinert in Zbl 1089.12001. In fact, that review is so comprehensive that it leaves very little to be added. The new edition contains several additions here and there (such as Newton's identities and linear disjointness) and a new chapter on transcendental extensions. This chapter establishes the existence and uniqueness of size of a transcendence basis, proves Lüroth's theorem and describes the group of automorphisms of a simple transcendental extension. It devotes a few pages to plane curves where an example of a non-rational curve is given, thus providing a transcendental extension of transcendence degree 1 that is not purely transcendental. The book is a valuable reference, covering many more topics than most of the standard books on the subject. These include, for example, infinite Galois extensions. However, an instructor planning to use the book for a one semester course must be selective. As the author suggests, such a course may cover Chapters 1 and 2, together with selections, from other chapters, dictated by the taste of the instructor and the limitations of time. The style is informal and at times unconventional, but always rigorous. One feels this from the very start when a field is defined, on p.~7, by as many as 14 axioms some of which are necessarily redundant. The key term [automorphism] seems to be never defined, but what harm can this possibly do? An [elementary progression] (on p.~x) must mean an arithmetic progression \(a, a+d, \dots\) with \(\gcd(a,d) = 1\); but this is not standard. The phrase [completely transcendental], coined by the author, serves its purpose perfectly, and the same holds for describing two extensions of \(F\) as [disjoint over \(F\)] when their intersection is \(F\). Usually, if a definition (or a theorem) has two parts (1) and (2), then one expects to be able to understand (2) without having to read the assumptions in (1); but this is not the case here as can be seen in Definition 2.5.1 (p.~22) and in Lemma 3.1.2 (p.~45). In several places, the equality sign stands for [is equal to] in a rather unconventional manner. For example, Lemma 3.4.1 states that [the field of symmetric functions \(F = D(s_1,\dots,s_d)\).] when it is more standard to state it as [the field \(F\) of symmetric functions is equal to \(D(s_1,\dots,s_d)\).]. Lemma B.1.2 (2) (p.~205) states that every prime ideal in a principal ideal domain is maximal; this oversight cannot be intentional, since it may be misleading to a beginner, and confusing to one who knows that the zero ideal is an exception. A splitting field of \(f(x) \in F[x]\) is defined, on p.~22, as [a field \(E \supseteq F\) such that \(f(x)\) splits over \(E\) but not over any proper subfield \(B\) of \(E\)]; many of this reviewer's students agreed that adding the phrase [containing \(F\)] at the very end of the definition would make it read better, but none went as far as disqualifying \(\mathbb{C}\) from being a splitting field of \(x^2+1\) over \(\mathbb{R}\) just because \(x^2+1\) splits over the proper subfield \(\mathbb{Q} \left(\sqrt{-1}\right)\). In his preface to this new edition, the author expresses his hope that in adding around 25 pages to the text, he has not introduced any new errors, and adds that that is probably too optimistic. This reviewer agrees: there are errors in the page numbers of at least 6 items of the index, and there is an extra ``s'' in [does not divides] in the very last paragraph of the book (p.~208). But there are also errors, usually minor and harmless, that persisted from the first edition. In Chapter 1, we find [different] for [difficult], [the] for [then], and a missing \(B\); on the first page of Chapter 2, the distributive law still reads [\(c(a+b) = ca+ba\)]; and so on. Every time this reviewer reviews a book, his strong belief that \textit{humans do err} gets stronger; so does his admiration for the degree of perfection that books printed in the first half of the past century enjoyed. It is true that careful proofreading can be rather costly, but books are not selling at remarkably low prices either.