The stumbling block, frequently mentioned, to identifying Countisbury with arx Cynuit appears to be the phonology. In a Hillfort Study Group report of 1989, the archaeologist Aileen Fox describes the uncertainty: ‘The difficulty is that the place name is entered in Domesday Book as ‘Contesberia’, and subsequently as ‘Cuntesbea’ [sic, recte Cuntesberia] (Devon Place Names I, p 62) and the experts are divided as to whether it is linguistically possible to be derived from Cynuit … .’
A second stumbling-block is the lack of any reference to Countisbury predating Domesday, which might have given a clue as to the derivation and dispelled all doubt.
Among the Gaulish place names studied, Conte in Jura is specially interesting, see below: an earlier attested form is Condate (E. Nègre, Toponymie, 1084 (DR), where 1084 seems to be the date and I think DR is a source because TP, in the item above it, is Tabula Peutingeriana).
So Gaulish/Celtic Cóndate > Conte seems possible; which gives support to Cóndate + Saxon gen. s + burh > Contesberie (plus the presence of a river confluence in both cases). Many French towns called Condé certainly derive from Condate, so something like Condé + s + burh would be phonetically possible, dependent on the chronology. The d/t alternation is not a problem (cf. a 13th-c form Cundeburye).
Although there are close to a hundred existing place names in France which are probably/possibly derived from Condate, they seem to have no close relationship with the Celtic god Condatis, only one inscription, in Allonnes, Sarthe, having been found. Condat(e), then, merely designated the confluence, not the presence of a cult. There are, however, inscriptions in Rennes (Condate Redonum) indicating the cult of a god Mullo, also identified with the Roman Mars, as is Condatis.
In Britain, only Northwich (I think) in Cheshire had an attested name Condate (Antonine Itinerary), but there are a handful of inscriptions and the remains of three altars in north east England mentioning Condati(s), and one inscription recently found in Scotland, proving the existence of a cult.
Camden’s sketch of the altar fragment (below), discovered near heavily Romanised Piercebridge, co Durham, but now lost, has been interpreted as:
men(sor) eṿọc(atus) imp(eratum)
ex ius(su) sol(vit) l(ibens) a(nimo) ”
which is beyond me to comment on. I do know that it was actually found at a place called Coniscliffe, which may look an interesting name – but no: Mills has produced a pre-Domesday (1040) form Cingcesclife, which points to ‘king’s cliff’ as the derivation; so useless to produce forms like Cunis-bere, a 16th-c. version of Countisbury, as resembling Conis-cliffe; which is, however, by the confluence of the river Tees and Piercebridge Beck. But as confluences go, it isn’t much to write home about …
There is one reasonable objection: Countisbury is not actually by the confluence noted, Watersmeet, where the East Lyn meets Hoaroak Water. However, Condres, Lozère, for example, was Condate in the 4th-c. Tabula Peutingeriana – see the illustration above – (for the Gallic insertion of r, cf Lond-inium > Lond-res). The present day town is now, according to Nègre, 2 km away from the confluence. Countisbury village is about 1 km from the confluence; but there was an Iron Age settlement actually overlooking the confluence, and not built for defensive purposes. People (presumably Iron Age Dumnonii, or ‘Ancient Britons’) lived there – unlike on Wind Hill where no signs of a settlement have been found.
So I would suggest that this settlement was the original ‘Cóndate’ of the Iron Age Celts, and that it later moved closer to the defences of the mighty rampart of Wind Hill, perhaps when inter-tribal warfare became more common, or perhaps as a response to the arrival of the Romans.
And what of this peaceful early settlement by the confluence? More of that later.