'Supernatural' and 'immaterial' are not "broken concepts"
Todangst claims that terms like "immaterial" are meaningless. He wants us to believe that such terms are "broken concepts". Why should we believe this? Well, Todangst thinks we can’t positively denote anything with these terms. There’s nothing out there that these terms actually refer to. The terms, says Todangst, “are therefore meaningless, incoherent.” Thus, Todangst’s argument rests on the view that bits of language are meaningful when there is something to which they refer. (The problems that I am about to present for Todangst’s view will take us into some philosophy of language. This area of analytic philosophy may be unfamiliar to some readers. But do not let the jargon intimidate you. If there’s something you don’t understand, PM me and I’ll be happy to explain it for you.)
Todangst's argument is unsound. First, empirical linguistic evidence proves that many bits of language, whether they’re in the form of text or an acoustic blast from the mouth, are meaningful but do not refer to anything, e.g. “hello”, “and”, “on behalf of”, “alas”, etc. It is a fact that English speakers regularly use sentences with these non-referring expressions. Yet it would be absurd to say that none of these English speakers mean anything by what they say when they use such language.
Another, more technical, problem with Todangst's view is that it logically implies that there are no contingent statements that express identity between an object denoted by a proper name and a definite description. (I assume classic identity here: quantifying over any x and y, x = y only if, for any property or feature F, Fx if and only if Fy. This is an uncontroversial assumption. Moreover, a statement is contingent if and only if it expresses a proposition that is true in some possible worlds and false in others; again, uncontroversial.) Assume for reductio that Todangst’s view is true. Then statements like “Socrates = Plato’s teacher” become necessary truths, since “Socrates” and “Plato’s teacher” pick out the same object and therefore have the same meaning. But we know that it was only a contingent truth that Socrates was Plato’s teacher; so there are contingent identity statements. So we must reject Todangst’s view. Reductio complete. There are several other objections to the view. But I’ll leave them aside, since what I’ve said is sufficient to show that Todangst's argument is unsound.
To be clear, reference is important to language. But it’s not what constrains meaning. And nearly all contemporary analytic philosophers of language would reject views like Todangst’s. There are only a few very subtle versions of “referentialism” still around, e.g. theories that rely on intensional classes as the referents. (See e.g. certain entries at logicalsemanticism.wordpress.com) But these theories allow copulas and other apparently non-referring expressions like “sake” and “so on and so forth” to represent a sense or at any rate something that may not exist as a material object. Since they preserve the meaningfulness of terms like “immaterial”, Todangst would presumably not want to endorse them.
What’s worse, Todangst would still have to show that predicates like “immaterial” and “supernatural” don’t have a nonempty extension. I don’t think he’s given us any reason to believe this, but I’ll save this discussion for another time.
 Exact quotation: “There's literally nothing left over for these terms to refer to, so there's nothing left over for them to be. The terms are therefore meaningless, incoherent.”
Rude, offensive, irrational jackass.