ω φιλε Μαρκε,
Amicus Marcus sed magis amica veritas.
I’ve resisted many earlier opportunities to take issue with your “semantic minimalism,” but it can’t be allowed to go unchallenged any longer. We agree, of course, that meaning is contingent: the meaning of a given utterance will vary according to the circumstances under which it's issued and received. J.L. Austin taught us that. Context is all: we have both said that.
But not literally all. There are the words themselves. “Call me Ishmael” could carry quite a variety of meanings in various contexts, but what you might be content to call its “basic” meaning is stable and definable in grammatical terms. Try going around saying it to people and noone will misunderstand what you have said, though they might not know quite what you mean by it.
Try going around saying “I speak true things” and people will think you’re mad. In other words the English sentences jaihare offered will not all communicate the same thing, regardless of the locutionary context. And nor will the Greek. And we should stick with the Greek, since English translations can be no better than approximate, and English analogies will likely be false.
50 out of 100 ancient Athenians, having chastized your impertinence, might have told you there’s no difference between το αληθες λεγω and τα αληθη λεγω. But you could have proved them wrong, by pointing out that the circumstances under which they would say the one were not always identical with the circumstances under which they would say the other. If the difference can’t well be rendered in English, that’s not to say there is
none. (I think that’s one of the traps you fall into.)
When it comes to θεος and θειος, no more than 1% of any ancient Greeks you polled would say they mean the same thing, and any who did would be wrong. The rest would tell you (if they had the patience: it would be wearing thin by now) that any number of things can be described as θειος, but nothing is a god but a god. And if you replied “Well, I’m a semantic minimalist, and I happen not to believe that. I think it is better to abandon the distinction between noun and adjective and to view both θεός as θεῖος as substantives which pretty much mean the same thing" they might (if they didn’t just box your ears) invite you to look at the tens of thousands of occurrences of θεος in their literature and challenge you to find one where θειος or θειον or το θειον could be substituted without affecting the meaning.
Adjectives can of course be “converted” into quasi-nouns by the article: ὁ χριστός, τὸ καλόν, “the poor.” But nouns can’t be converted into adjectives, or not without changing their morphology. (That's not true in English, but it is in Greek.)
In short, I submit that grammar and word choice make more semantic difference than you like to imagine. Examples previously touched on in other threads include aor.:impf. indicatives, aor.:pres. infinitives, and many more. Just because you can’t see a difference in any given instance doesn’t mean that there isn’t one. — You can of course retort that just because I can doesn’t mean there is.