AttestedLevioxava at position 222 in the Ravenna Cosmography.

Where:  Perth, on the river Tay.  The Cosmography lists Levioxava after Poreoclassis (Carpow) and before Cermium (Strageath).  Perth otherwise has no Roman name.

Name origin:  The Lev- part came from PIE *leb- ‘lip’, referring to a river squeezed between higher ground, as discussed at length here and here.  Also on the river Tay, a little way upstream, was Leviodanum.  If the name is segmented Levi-oxava, a meaning is required for oxava, perhaps like the Oxus, a great river of ancient Asia.  If it is segmented Levio-xava the second part could be a variant of *sava, well known as an ancient river name, as discussed under Sabrina.  However, the river Tay clearly had an ancient name like Tava, and it would make sense for Perth to be at the ‘lips of the Tay’.  Possible reasons why a letter X might occupy the slot of a letter T include the infamous Tau Gallicum, pronounced like TZ and written like a crossed D or S, or miscopying of the name off an original Roman map: see the little piece of one Cosmography manuscript here, in which the X in Levioxava can be contrasted with the T in Victorie.
[Levio mss part]

Notes:  Fitting the Cosmographys's sequence of names to known Roman sites is like a game of musical chairs, with Levioxava ending up without a seat when the music stops.  Perth is now a clear winner, but previously a range of other possibilities have been discussed.  The Roman camp near Forteviot would be a strong candidate if Marcotaxon had not already claimed it.  A previous best guess focussed on the river Earn, which flows into the river Tay in much the same configuration as the river Sava flows through former Yugoslavia into the Danube.  Near Bridge of Earn, a canoeist reports “a weir about 150m below the road bridge ... which washes out at high tide”.  At about NO107194 there is a curious figure-8 structure in the river Earn, around Kirkton Pouch and an oxbow lake, at about the modern limit of potential flooding from an exceptionally high tide, where conceivably some kind of Roman weir or causeway remains to be discovered, which would have served to deepen the river upstream so that boats could carry supplies to the fort at Strageath.  Ekwall (1928:308) suggested that the river names Okement, Ogmore, Ogwr, and Ogwen had a first element related to Greek ωκυς ‘swift’ or Latin ocior ‘swifter’.  Both the Ock in Surrey and the Ock in Oxfordshire (whose alternative name may be Charn, from Cearn) run fast upstream where they drain hills, but lower down they tend to spread out and cause floods, much as the river Earn does; see here for one map of its potential flooding.  Celtic scholars have been slow to recognise the significance of PIE *oku- ‘swift’ because its descendants are a bit hidden, such as Welsh diog ‘lazy’, eog ‘salmon’, and ebol ‘foal’ (from equus ‘horse’).  There is confusion from other roots meaning ‘sharp’, ‘eye’, ‘upper’, and ‘eight’. Delamarre (2017,65-70) realised that Octo in ancient names need not always mean number 8, but went off into a flight of fancy about frost fairs, and missed the fact that *Octodurus was on the upper Rhone which “est un fleuve capricieux” that periodically causes devastatating floods.  Also, in that area, Selkirk (1995) “would like to bet ... that Inchaffray Abbey hides a Roman site”.

You may copy this text freely, provided you acknowledge its source as www.romaneranames.uk, recognise that it is liable to human error, and try to offer suggestions for improvement.
Last edited 29 October 2022     To main Menu