iterum sunt ciuitates in ipsa britania recto tramite una alteri connexae, ubi et ipsa britania plus angustissima de oceano in oceanum esse dinoscitur. id est:
|
Velunia |
Carriden |
10752 |
|
Volitanio |
Mumrills |
10752 |
|
Pexa |
Camelon? |
10753 = 1088 |
|
Begesse |
Rough Castle |
10753 |
|
Colanica |
Castlecary |
10754 |
|
Medionemeton |
Croy Hill |
10754 |
|
Subdobiadon |
Bar Hill |
10755 |
|
Litana |
Cadder |
10755 |
|
Cibra |
Balmuildy |
10756 |
|
Credigone |
Old Kilpatrick |
10756 |
Rivet and Smith (1979, 211) dismiss this list as containing one certain fort name and ‘not more than four’ others, preferring to emend Volitanio, <Pexa> and <Credigone> into the ethnic names Votadini, *Pecti (for Picti) and Creones. Most of the remainder they attach to places not on the wall. Whilst none of these emendations is impossible, the first name is not necessarily corrupt at all (it means ‘rather broad place’, ‘the expanse’: Richmond & Crawford 1949, 50. Rivet & Smith’s (1979, 509) dismissal of the name is unworthily disparaging). The second is positively startling. If the map source used by the Cosmographer were Flavian in date with Severan additions, as Rivet and Smith (1979, 196) conjecture, the mention of the Picts precedes any other by almost a century. This simply will not do: better to accept that the Cosmographer saw a line on his map listing the names of the forts on the Antonine Wall than to credit a Severan map with mention of a tribe ignored by the historians of the Caledonian wars that are hypothesised to have been the reason behind the map’s compilation.
Exactly how to identify the names with forts remains a problem, nevertheless. Velunia, the first name, is attested epigraphically at Carriden (Rivet & Smith 1979, 490), so we know that the Cosmographer is working from east to west, as he had done with Hadrian’s Wall. However, although nineteen forts have been identified on the Wall, a maximum of only ten is listed here. It has been suggested (Breeze & Dobson 1976, 142) that the solution might be to regard the Cosmographer’s list as giving the names of only those forts occupied during the brief Second Antonine occupation of Scotland, if this really exists (Hodgson 1995, 42). If we can still accept a modified later Antonine scheme, which is doubtful, this suggestion would have its merits, but only two forts (Bearsden and Rough Castle) can be shown conclusively to have remained unoccupied at this time. The remainder seem to have continuous occupation.
Richmond and Crawford (1949) attempted to use the topographical hints suggested the by the Celtic derivations of the names to locate them, but wrongly assumed that the list was working from west to east, and none of their identifications can be correct for this reason. Even if none can be accepted, though, their principle of using topographical references in the names is sound. On this basis, the second name, Volitanio, a ‘rather broad place’ (Ifor Williams in Richmond & Crawford 1949, 50), could well be Mumrills, overlooking the flat ground now occupied by Grangemouth. The form is a perfectly good Celtic name, despite Rivet and Smith’s (1979, 509) sniping comment that ‘it is acceptable up to a point’: if one were not looking for the ethnonym Votadini in the text, there would be no need whatsoever to emend Volitanio.
The next name, Pexa, which Richmond and Crawford (1949, 43) emended *Dexa, ‘south’ (literally ‘left’), is perhaps associated with a piece of ground regarded as being in some way southerly. The name recurs later in the Cosmographer’s list of places north of the Wall as Decha (at 1088). It could have been listed a second time if the place were close to but not actually on the line of the Wall and were in some way prominent on the map source. In this case, an identification with Camelon would be attractive, although it is difficult to see why Camelon might be regarded as southerly, unless it refers to its location south of the River Carron, a not especially convincing explanation.
<Begesse>, which may mean ‘ridge’ and perhaps ought to be emended *Begisso (see Rivet & Smith 1979, 266) would suit the fort at Castlecary, at the foot of a ridge running between Cumbernauld and the canal. However, Colanica, the following name, is clearly the same as the Κολανια of Ptolemy (Geography II.3,7) and the distances he implies fit an identification of this name as Castlecary, a fort which may have Agricolan antecedents (Richmond 1958a, 148; Frere 1987, 91); if this latter identification be accepted, *Begisso would then be the name of the fort at Rough Castle, in which case the hypothesis that this list refers to the apparently dubious second occupation of the wall cannot be upheld.
*Medionemetum does not suggest a topographical feature evident today, and the identification with Arthur’s O’on suggested by Rivet and Smith (1979, 417) is fanciful. The ‘middle sacred grove’ of the name is more suggestive of the worship of Silvanus, attested at a number of sites on the Antonine Wall, rather than a reference to a purely Classical religious structure.
Subdobiadonis extremely obscure as a name, although the ending suggests one ending with either *-iacum or *-iadunum (Rivet & Smith 1979, 463), and no topographic hints can be eked out of it. Litana, a ‘broad place’ (Richmond & Crawford 1949, 38), must surely be Balmuildy or Cadder. Dillemann (1979, 70) sees this name as a truncated doublet of *Litanomago, which occurs below at 1086. Rivet and Smith (1979, 245) prefer to emend the name to Alauna, which is quite acceptable; however, they identify it with Ardoch, the Αλαυνα of Ptolemy (II.3,7) to the north of the wall. This is not necessary, though, as Litana is a reasonable Celtic form. If we accept the identification with either Balmuildy or Cadder, the two previous names (Medionemetum and <*Subdobiacum>) could be Croy Hill and Bar Hill respectively.
Cibrais not necessarily to be identified with Ptolemy’s Κουρια (II.3,7) (as Rivet and Smith (1979, 317) have done), since *Ciura is a reasonable enough Celtic form, if of uncertain meaning (Williams in Richmond & Crawford 1949, 28). Nor is <Credigone> necessarily to be identified with the Creones, although it must be confessed that again the name is obscure. <Credigone> may contain a river-name ending with *-ona. This should be Old Kilpatrick, at the western terminus of the Wall, and *Ciura may be Balmuildy, in which case Litana would be Cadder.
The Antonine Wall remains a toponymic puzzle, but the extremely negative view of Rivet and Smith (1979, 211) is certainly too pessimistic. True, some of the names presented by the Cosmographer are hopelessly corrupt, but the tendentious emendation of names which are not, such as Volitanio and Litana, does a great disservice to the Cosmography as a whole. The approach proposed here accepts that the Cosmographer was a poor copyist, probably working with defective originals, and that he was not working in a manner which conforms to the logic of the modern western humanist tradition, but it does allow him a modicum of sense and plain speaking. If he saw a line of place symbols on a map, joined together in some way, there is little reason to announce the fact and then proceed to list tribal names and names of places far distant from the line as if they were in some way connected. The approach taken here allows us to make hypotheses regarding the names of some of the forts on the Antonine Wall; there can be little doubt that not all the identifications proposed or the suggested emendations are correct, but I firmly believe that the attempt is worthwhile, if only to stimulate further discussion.