If not "univocal," then what?

Carl Conrad notes that a lingering problem that he perceives with DE is the
assumption of a univocal relationship between original text and translation. 
He also notes in defense of some GNB passages that the GNB "quite often
expresses the idea intelligibly and accurately in phrasing that I would not
call paraphrasing but translation in the best sense."  In this same vein,
Mari Broman Olsen offers as a maxim: "ALL translation is paraphrase."
So, some questions:
1.  If not a univocal relationship, then what?
2.  What remains of the difference between paraphrasing and "translation in the
best sense," or, of any kind of differentiation between meaning in textual
authorship/readership and meaning in "translation"?
3.  If Mari's observation is vindicated, then what makes for good or less good
and, 4. If the RSV/NRSV leave a "hard to understand" greek passage hard to
understand in translation, what were they missing which would have resulted in
an easy to understand translation?  Is the NIV, with its propensity to render
every possible difficulty in as "vanilla" a translation as possible, a better
treatment of hard to understand passages in greek?
There, I think that will do for some opening queries!