This book frees your thinking on variables, constants and names. It starts with a semantics for variables that is completely free of syntax and philosophically the most pleasing solution ever. But even here, in the first chapter, you may ask yourself how the substitution [value] theorem could be expressed. Then even variables of the same name are admitted to refer to different things. From the fact that names in natural language may refer to different things then comes the suggestion to let the same constant denote different things. I think that this is not to be recommended in logic: Instead of letting Poincare denote different persons in the modeling

Met(Poincare,Poincare) of the sentence "Poincare met Poincare" (say the mathematician met the politician), we can model this as

Met(x,y) with choice-condition

C(x) = epsilon x'.name(x')="Poincare" and

C(y) = epsilon y'.name(y')="Poincare" using an indefinite version of Hilbert's epsilon.

The general defense of the referentialist standpoint against the Fregean one is heroic and breathtaking but not always convincing.

All in all, this is a book is a fascinating reading that frees your mind on logic.