You seem to be puzzled by the fact that Newtonian mechanics was retained in a successive theory and that this somehow precludes incommensurability. It seems that you would think of incommensurable theories as "not relating to each other" somehow and that retaining a former theory is incompatible with this requirement.
If this is your view, the point to be clarified here is that the possibility of including a former theory within a successive theory does not preclude incommensurability between those theories.
Both Kuhn and Feyerabend - who put forward the concept of incommensurability between scientific paradigms (Kuhn) viz. universal theories (Feyerabend) - were trained physicists, so they knew very well that you can derive Newtonian mechanics as a special case of STR.
So, what was Kuhn's point? He writes:
… the physical referents of these Einsteinian concepts are by no means identical with those of the Newtonian concepts that bear the same name. (Newtonian mass is conserved; Einsteinian is convertible with energy. Only at low relative velocities may the two be measured in the same way, and even then they must not be conceived to be the same.) (SSR, p. 102)
His point is that key concepts in both theories, while retaining the same name - like "mass" - have not only different meaning, but that the meanings of Newtonian mass and Einsteinian mass exclude each other. According to Kuhn, there is no concept of mass, in this case, which can consistently unite both meanings - thus they are "incommensurable" concepts, i.e. concepts "without a common measure". (How this gets us to incommensurable theories and paradigms is a lot more complicated, but you get the picture.)
Kuhn uses this point against what is known today as convergent realism, the view that science shows improving approximation to the truth. Why? Because, according to Kuhn, the concept of mass as devised by Einstein does not extend, but replaces, the concept of Newtonian mass. So, if we want to understand in which way science progresses through these shifts, Kuhn argues, the model of successive extension or piecemeal revision of a concept is not a good candidate.
Did Kuhn argue from this - as it is sometimes assumed - that one couldn't derive Newtonian mechanics as a special case of STR? Certainly not. He argued, however, that the special case derived is not actually Newtonian mechanics, but a substitute of it, a numerical approximation within STR. This qualification is of no interest to the working physicist, but it is of interest to someone - philosopher or physicist alike - who wants to argue that the progress of science consists in (old-style) convergent realism.