For me, only really great poetry fulfills the function of poetry at all. The rest is just to keep the institution going, to have poetry at all, because you couldn't have it without having a critical mass of it, quantitatively, for it to exist.
I don't feel this way about film, for example. Since I don't have high expectations of novels I think all novels fulfill the function of novel-reading, if I can stand to read the novel in question at all.
But then again, some poetry does fulfill that function by coming close to being great, with that shock of recognition. "I bought a dishmop / having no daughter." The second caveat: I can feel that great poetry comes in several forms, like the seemingly minor poetry of Ron Padgett. Thirdly, you can agree with me about this point, that only great poetry fulfills the function of poetry at all, but disagree about what poems and poets are actually great.
Of course, you can disagree with me that only great poetry fulfills the poetic function at all, but then I would probably disagree with you about what that function really is. I would define it tautologically as the function that only great poetry fulfills! I don't want to define it, otherwise, because any definition would feel exterior to poetry itself.
Poetry in translation, for me, rarely fulfills that particular function. For enough to come through, in translation, is exceedingly rare. The kind of translation that makes excuses for itself. The translation must enjoy a high degree of autonomy in relation to the original; it must be a translation of an already great poem, etc...
There's a version of a Lorca poem by Spicer that is superior than the original, for example. Lorca's poem is not great, in this case.
Borges points out logically that there is no a priori reason why a translation cannot be better than the original (or just as good). Say there are two poets of equal ability. One writes poem. The other writes a poem in another language that also happens to be a translation of the first poem. Logically, there is no reason to think that poem 1 will always be better than poem 2. This seems obvious, unless we think of the original as a sacred object.
But with a really great poem, then you get a situation where this poem has a one in a million chance of coming into existence. Then the chances that any particular translation of this poem will also be one in a million are, well, one in a million. If we think it is harder to write a poem when trying to translate than while writing your "own" poem, then the chances are less than one in a million. If you think that it is easier to be great when translating from a great poem, then the odds go the other way. Maybe every thousandth translation of the one in a million poem will be great.
I too dislike it. I cannot defend my feelings intellectually, but I cannot stop having them either.