The fine structure of LD-equivalence (Q1590057)

An LD-system (or Left Distributive system) is a set provided with a binary operation which satisfies the so-called identity (LD): \(x(yz)=(xy)(xz)\). The free LD-system based on a set \(X\) is the quotient of the set \(T_\infty\) of all terms with variables in \(X\) under the least congruence that forces the identity (LD) to hold. Contrary to the structure of the free groups or of the free monoids, the combinatorial structure of the free LD-systems is quite complicated and badly understood. The paper under review follows several other ones by the same author whose program is to understand the LD-systems. Such a program is made more natural and attractive by the deep connection between LD-systems and braids. The first significant result of the paper concerns the word problem in the free LD-systems. The question is to determine an algorithm which decides whether two terms represent the same element in the free LD-system. The Polish algorithm is a quite simple method for solving the word problem. It has already been considered before, but its convergence is still an open problem. Here the author proves that this algorithm converges for a large family of pairs of terms. Afterwards, the author provides a new method for computing some normal forms in the free LD-systems. This method is much more useful than the others known before. With a free LD-system one can associate a group \(G_{\text{LD}}\) which contains the braid group \(B_\infty\), a monoid \(M_{\text{LD}}\) which contains the monoid \(B^+_\infty\) of positive braids, and a morphism \(f\colon M_{\text{LD}}\to G_{\text{LD}}\). This group plays for the LD-systems the same role as Thomson's group for associative systems, and its study is the hard core of the subject. The Embedding Conjecture says that \(f\) is injective. The third relevant result of this paper is that, for a large family of \(a\) in \(M_{\text{LD}}\), there is no \(b\in M_{\text{LD}}\), \(b\neq a\), such that \(f(b)=f(a)\). In particular, this extends the well-known result of Garside which says that \(B^+_\infty\) embeds in \(B_\infty\).