>>How secure is the derivation? I would expect something like Latin *horbis
/ *horbs,
>The loss of h- is in itself not problematic: there are more examples of
>early loss of h-. As to the vocalism, that's another issue.
"Obscure, perhaps dialectal, is the patternless change of o > Latin u
before r + consonant in some words, as furnus "oven" (fornus rare) beside
fornax "furnace" (furnax rare); also in loanwords like purpura "purple dye"
, Greek porphura. This has the nature of a regular sound law in Sabellian,
but in Latin for the most part or + consonant remains unaffected, and this
is assumed to be the proper development for urban Latin." (Sihler section
45)
Meiser has something similar, but also says
"Occasionally we meet Latin CurC as a reflex of *CrC (syllabic r), when w or
a labiovelar precedes the resonant r ..... No conditions can be adduced for
curro, scurro .... urbs (*wrdhi-, as Umbrian uerfale "a bordered area" <
*werdh ?). In part ur < *or appears as an inner-Latin (not city-Roman)
development..."
So the vocalism, although unexplained, has parallels, and is not a problem.
Peter