Reflection groups and invariant theory (Q5942989): Difference between revisions

The subject of the book are finite subgroups of \(\text{GL}(V)\) (\(V\) a finite-dimensional vector space) generated by linear bijections of \(V\) whose fixed spaces are hyperplanes, and the invariant theory associated with such groups. The first part of the book consists of the following chapters. Reflection groups, root systems, fundamental systems, length, parabolic subgroups, reflection groups and Coxeter systems, bilinear forms of Coxeter systems, classification of Coxeter systems and reflection groups, Weyl groups, the classification of crystallographic root systems, affine Weyl groups, subroot systems, formal identities (Mac Donald and Weyl identity). This is roughly the material of \textit{N. Bourbaki}, Groupes et algèbres de Lie, chapitres 4, 5 et 6 (1968; Zbl 0186.33001), when the material on Tits systems and BN-pairs is excluded. The second part of the book studies invariant theory. It leads to the Chevalley-Bourbaki-Shepard-Todd Theorem (1954): Let \(V\) be a finite-dimensional vector space over a field and \(G\) a finite subgroup of \(\text{GL}(V)\) such that the characteristic of the field does not divide \(|G|\). Then \(G\) is generated by simple mappings (`pseudo-reflections') if and only if its ring of invariants is a polynomial algebra. The third part deals with subgroups and conjugacy classes in finite Coxeter groups, eigenvalues and Coxeter elements. Also Carter's result that each element of a Weyl group is a product of two involutions is mentioned. A selection of nearly a dozen of books on reflection groups and invariant theory is available. Some of these books involve applications to Lie theory or crystallography. The present book does not study Lie algebras but many other non-standard items, e.g. the Cohen-Macauley-property of the ring of invariants, generalized invariants, and Hopf algebras in order to characterize pseudo-reflection groups in terms of harmonic elements. The book is intended to be a graduate text. However, it includes also very elementary arguments, e.g. a proof is given to the fact that a vector space over an infinite field is not a union of a finite number of proper subspaces. On the other hand, some proofs include phrases like `...now it is easy to deduce that...' where to me this seems not appropriate. The author gives (mainly in the basic parts of the book) extensive comments. Some readers may appreciate this, others may prefer a more rigid restriction to precise definitions and statements of facts and brief complete proofs. For example, a student who has never heard about free groups may want to read a precise definition of a free group with relations (page 11). The term `reflection' is used in several meanings. The definition on page 154 reads: a reflection on \(V\) (a finite-dimensional vector space) is a diagonalizable isomorphism \(V\to V\) which is not the identity map but leaves a hyperplane \(H\subseteq V\) pointwise invariant. Then it is claimed that the definition reduces in the case of the reals to `those previously considered for Euclidean space'. This sounds confusing. The traditional terms reflection and pseudo-reflection successfully defy terms used in literature on classical groups (a linear automorphism \(\varphi\) of a vector space is called simple if \(\dim V(1_V-\varphi)=1\); a simple mapping with \(V(1_V-\varphi)\subseteq\text{kernel}(1_V-\varphi)\) is called a transvection; a simple mapping in an orthogonal group is called a symmetry; and so on). Angles and angle-measures are among the most obscure concepts in elementary geometry. What precisely is meant by a `rotation through twice the angle between \(\alpha,\beta\)'? The proof of Lemma D on page 13 ends with the phrase `the following picture illustrates that \(\beta\) is moved (in this case, clockwise) through twice the angle between \(\alpha\) and \(\beta\)'. There are some errors; examples: On page 72 it is claimed that any two vectors of a root system \(\Delta\) fulfill: \(s_\alpha s_\beta\) has order \(m\) if and only if \((\alpha,\beta)=-\cos(\tfrac{\pi}{m})\) (\(s_\alpha\) denotes the symmetry with negative space \(\langle\alpha\rangle\)). In the definition of a Coxeter group (page 66) the requirement \(m_{ss'}=m_{s's}\) does not appear. The book is certainly a good source of information mainly in the more advanced areas of the subject. The references compile nearly a hundred articles and books.