That is quite correct. If you look up the mathematical definition of associativity, the behaviour of the "<" does not fit it. In fact associativity is only a pr

That is quite correct. If you look up the mathematical definition of associativity, the behaviour of the "<" does not fit it. In fact associativity is only a property that can be possessed by functions or operations that operate on a particular set. Thus, both the inputs and the result of the function or operation have to belong to the set. For the "<" operation, the inputs are numbers and the result is boolean. A boolean value cannot be an input for the "<" operation (if you take the classical definition of "<"). Therefore, "<" does not operate on a particular set. Therefore "<" cannot be associative. The behaviour of "<" the author refers to is not associativity. Therefore, the languages he criticises do not really get associativity wrong, he just needs to brush up on his math. Now you can say "well, ok this behaviour is not associativity but it is a well know behaviour of "<" and ">". Why can't the languages get it right?" I think the answer is that this (a<b<c<d) is one of those notations that look ok in algebra class but are really problematic if you actually try to define and use a logically consistent function notation, so it is not a problem for me, if some languages do not perform (a<b<c<d) as expected in order to keep their semantics clean.