It is a good question even if it seems to be inviting speculations rather than historical references. The obvious answer is that we cannot tell how Aristotle would proceed but there is a good exemple of what could be the central issue as it generated the s.c. linguitics wars. (and further a book with that title by Randy A. Harris, Oxf. 1995).
In 1958 Emile Benveniste published Categories de pensee et categories de langue (Jstor), a remarkable analysis how Aristotle took the grammar of his native language to be of metaphysical significance. The mistake was not irreparable as people already in the middle ages (e.g. R.Bacon) understood that it was just an instance of a more general universal grammar. The idea has been reworked in early modern times and it became rather popular with Chomsky's work, who was the commander in chief of one of the warring parties.
The other side was promoting generative semantics but at that time there was already a modern reworking of Aristotle by Joseph Greenberg in his Language Universals (1966). A less known contribution was made Polish linguists and Anna Wierzbicka, who emigrated to Australia, achieved some notoriety with her work on semantic primes or primitives. As the name suggest such entities cannot be defined in a non-circular way in language - but they are understood as they embody human experience. (The best known example seems to be "time" all attempts at whose definition turn out to be circular.)
So what would be Aristotle's view depends essentially on what would be the conception that distinguishes syntax from semantics and the arguments with which it is implemented.