Though I am uneasy about the term "substance" in this debate, it seems to me that Kant is perhaps the only truly consistent, plausible dualist.
He posits a unity of all things knowable, including the "physical" by any interpretation, and the existence of the noumenal,"unknowable" realms, which can never be reduced to the phenomenal. Though we may picture this as ghostly entities "outside" of conceptual grasp, it may be more feasible to think of it as a frequency or "scale" exceeding any possible measurement or a "speed" faster than light. Yet there are moral and other "reasons" within his system for positing such "unknowables." They can't just be Occamized.
In other respects, it is questionable whether even Descartes is a proper dualist, since he must posit some "pineal gland" or whatnot carrying out mind-matter conversions. (You are right, it appears to be largely a "semantic" problem, but I'm not sure that really says anything or washes out the complexity.) Since dualism suggests substances that can "never" interact "anywhere," we might suspect that "infinity" is involved, which cannot really sit within any physicalist system.
There are many candidates for things "outside" of any physical-causal sovereignty. Infinity, negentropy, possible worlds, utopia, god, entanglement. Noumena or Wittgenstein's "whereof we cannot speak." Somewhere over the rainbow. Are these irreducibly "dual" to the matter-energy continuum? Or are they unified in the consciousness that can conceive of them? If so, would that make schizophrenia the basic substance dualism?
Obviously no satisfactory "ontological proof" can apply to anything so construed. To me it remains, so far, a tangle. And Kant strikes me, for now, as the most coherent, workable expression of a fundamental dualism. So I cannot accept that it is "merely semantic." And anyone who could convincingly reduce such a swarm of problems to "semantics" would, I suppose, vault into philosophical prominence.