As a clue to look for evidence, I'd suggest muscular and innervation cues that support prehensile upper lips, as Jura indicates referencing Breda (2008). More mobile tissue means more extensive nerves/blood vessels, which should reflect a large maxillary branch of the trigeminal and anterior branches of the facial nerve, which exit through the narial opening. Extensive facial innervation (see, tapirs, rhinos, etc) likely results in less "rhinaric" tissue of the snout. Stiffened upper lips and all that. A broader osteological permission might be the relative width of the snout and associated nostril-adjacent tissues: prehensile tissues caused for a retracted, open snout. But, as noted, sometimes that's not so clear: giraffids have narrow snouts, good for high browsing and selective feeding. Presence of absence of an incisiform ("premaxilla") and teeth upon it is also no good indicator, given the variation seen in artiodactyls and perissodactyls.
As with many things, osteological correlates are an inference to or from a given condition, rather than a hard-fast rule of inclusion/exclusion of it. So we should take these things with a handful of salt.