I kind of get that, I think. But that's where I'm having trouble - if I'm getting it, why did THEY disagree?
But I don't think Ricoeur, who was writing on the philosophy of metaphor in science, OR Bachelard, who was writing explicitly about science (and so, obviously, about the possibility of "real" reference) would claim that metaphors ever really refer to anything concrete. They're still Hegelian and idealists in that sense, not epistemological realists.
I don't think they'd say that it's not a messy process, either. Obviously things inject, interfere, etc. That's the fluidity, even the chaotic nature of poetics, isn't it? That's part of its power to radically redescribe reality.
But what does Derrida mean when he says that "meaning" is not possible in language at all? Does he solely mean "intent that cleanly translates?" What is meaning, if not the thing that allows us to stably and approximately communicate with one another? That's what I don't understand. As murky as it might be, as difficult as it might be to pin down exactly, I can still be pretty goddamn sure that "this German historian cared about German identity" when I read his 500-page book trying to show Germany's continuity in history.
I tend to subsume Ricoeur into late Wittgenstein here, where communication isn't one-to-one either, it doesn't idealise linguistic "essences" (which I don't think good Heideggerians did either, obviously). "Meaning," and therefore "communication," is a matter of "seeing" - language that Ricoeur takes from Wittgenstein and uses a lot.
To have successfully learned a meaning is to say "ah yes, I see what you mean, there." The test of successful communication, of understanding, isn't whether you have internalised the "idealised essence" of that meaning (which is a nonsensical concept), but whether YOU CAN COMMUNICATE. If the person says "you see what I mean, do you? Show me, then," and you say something back to him that proves to him you understand, you have correctly learned the "rule" of how to apply that concept -- it seems to him (i.e., he "sees") that you followed the rule correctly.
But this, again, Ricoeur wouldn't deny. Each use of meaning is hermeneutic, interpretative, novel, creative, original. That doesn't mean there is NO stability whatsoever. That's what I'm confused about here. Is Derrida saying no stability is possible? Is it possible that he's presuming an implicitly metaphysics of meaning (of idealised essences of meaning) that doesn't exist?