Two generalizations of the nonabelian tensor product. (Q1946102)

The new idea of the present paper is a generalization of the notion of nonabelian tensor product, largely studied in a series of contributions of R. Brown, G. Ellis, D. Johnson, J.-L. Loday and E. Robertson almost 30 years ago. There is a continuous interest in the topic, since it has deep connections with various branches of mathematics such as algebraic topology, K-theory and low dimensional homology. The ``box tensor product'' appears in Definition 1.2 (of the paper under review) and is based essentially on a variation of the notion of ``compatible actions'', which plays a fundamental role in the theory of the nonabelian tensor products. Some illuminating examples show that the box tensor product is in general more complicated than the nonabelian tensor product and, unfortunately, the examples are quite technical to describe and often the help of GAP is required. A classical theorem of Ellis on the finiteness of the nonabelian tensor product \(G\otimes H\) of the groups \(G\) and \(H\) is generalized here for the box tensor product (see Theorems 4.8, 6.3 and 6.4). One of the main ingredients is an application of the famous Dietzmann Lemma (see for instance Lemma 4.1) to the context of the box tensor product. On the other hand, the literature on the nonabelian tensor products has also an alternative approach, due to some contributions of N. Rocco and I. Nakaoka, who noted that it is possible to embed the nonabelian tensor product \(G\otimes H\) of the groups \(G\) and \(H\) in an overgroup \(\eta(G,H)\), defined in terms of generators and relations. This allows us to place \(G\otimes H\) in a bigger group, of which we may have more information. Section 5 of the paper under review shows that a similar overgroup can be constructed for the box tensor product (see Theorem 5.2) and its properties are discussed. A series of interesting consequences and finiteness conditions are contained in the remaining sections.