Celtic swash-S motif and the structure of evidence

Swash-S motif on button-and-loop fastener
Wild's Class VIa, 2nd century AD
Portable Antiquities Scheme,  HAMP3302 

This post is a continuation of yesterday's post and must, boringly, start with the same picture. H. E. Kilbride-Jones, in Celtic Craftsmanship In Bronze shows this same type and records that example as coming from Cilurnum, Northumberland (Chesters Roman Fort). Elsewhere in his book, he refers to a swash-N motif, but does not apply that designation to this one. In his chapter on dress-fasteners, he says that most of them "have been found in the former territories of the Votadini and Brigantes". both tribes had important metal-working sites, but it is the Traprain Law hill fort in Votadini territory that seems more likely to have produced this type than the workshops in the Brigantes territory. We cannot be sure of this, however, without clear evidence like comparative metal or enamel analyses, or the discovery of moulds. With so few and scattered find spots that we can associate with this motif, and the knowledge that all of these workshops were making "mass-produced" small objects for trade that seem to have been taken far and wide by Roman soldiers, we cannot make very confident claims about its source. It is possible that this rare type came from a small workshop yet to be discovered. This might even be in a different tribal territory --we just do not know. The scales are tipped slightly in favor of Traprain Law because there is another example of the motif on a driving-harness snaffle bit (Jope, 2000, Pl 279 a-b, page 303, Birrenswark, Dumfriesshire). Its "official" museum photograph seems to be deliberately useless, and the National Museum of Scotland, unlike the British Museum, wants no truck with its images plastered all over the Internet. The quoted find spot "Burnswark" is an alternative spelling for Birrenswark. As I cannot freely snatch images from Jope, I found one that was labelled "out of copyright", so I eliminated the yellowing from the image of that page, and enhanced it to make it closer to the original drawing:

Bronze driving-harness snaffle-bit decorated with red and yellow enamel
Birrenswark, Dumfriesshire, Last half of the 1st cent AD (click to enlarge)

The information on that page is also more useful than its museum catalogue worksheet. I especially found it interesting that Roman coins from Nero to Trajan had been found in the neighborhood. While not proving the date of the bit, it does add weight to Jope's dating of the type. This was clearly not part of any farm cart rig. I am sourcing it as from a driving-harness because of the extra decoration on the left which has the swash-S motif at its centre. This is part of a chariot rig -- with one for each pony, the more elaborate decoration would not be wasted on the inside. You don't necessarily need a bit on a driving harness, it depends on the horse or pony. No harnesses are known of that time, so we cannot be sure of their configuration or use. I can't be of much help in techniques, having spent very little time behind a horse. I know next to nothing of the commonest English style of horse-riding, either. It seems a bit primitive to me. I mean, really, a rein in each hand? What does one do when you need to rope a steer? I did have to gee-haw once, on a badly trained, nasty and skittish pinto stallion from an Indian reservation here. It was bad enough to start with, but when a truck sped by me and the country road gravel showered the horse, he bolted and there was nothing I could do about it. After pulling his head up toward the sky to no avail. I decided to ride it out. That was fine until he decided to cross a ditch and enter a wood, still at full gallop. A low tree branch hit me in the chest, back I went, jamming one foot in the stirrup and being dragged through the wood before my foot came free. My memory of the incident is spotty. I had several injuries and might have been knocked out in the wood. I woke up in field, realizing that I had also broken my arm. Then, I had to catch the damned horse which at that point was quietly grazing as if nothing had happened. I have had an aversion to pintos ever since. Even the cars named after them had a habit of catching fire when rear-ended. But, as Shakespeare wrote, "What's in a name?" A pinto horse is just a colour variation, anyway. I should have known better when I learned that the horse had not been trained to neck rein and had been sold off. I don't mind a bit of excitement -- I even rode a cutting horse, once -- talk about turning on a dime!

That the snaffle-bit had a lot of wear and was repaired a few times is interesting. The Torrs Farm pony-cap had also been repaired, apparently in Ireland. That is interesting because Dumfriesshire is adjacent to Kirkcudbrightshire, where Torrs farm is located. The bit is not of an Irish style, though, and no bits of its type have been found in Ireland. Iron age sites are not thick on the ground in Dumfriesshire, the only one I know of has not been excavated at all.  It seems most likely that the snaffle-bit had not travelled that far -- Roman soldiers seemed just to go for the trinkets, so Traprain Law is a slightly better possibility for the button-and- loop fastener. Traprain Law seems to mainly have objects with multi-coloured enamel within small cells, though and the fastener does not quite fit the profile.

I am defining the Swash-S motif as an implied S-shape with a keel at each of the four corners. Keeling does have an earlier history, but we have to take all of the elements of the motif into consideration. As Vincent Megaw says, "Similar is not the same as same".

A related motif, of course, is the Swash-N. While the Stitchill, Roxburghshire collar has a wildly elaborated form of this motif which includes two keels, and the Swash-N is a response to a large rectangular (Roman), space, I think a closer relative to the S-form on the fastener is the armlet which was found at Plunton Castle, Borgue, Kirkcudbrightshire.

Plunton Castle, Kirkcudbrightshire, armlet
late 1st cent AD. (click to enlarge)

Although dress fasteners are concentrated in Votadini and Brigantes territories, I think we can rule out the Brigantes for the Swash-S motif.  I think it possible, though, that the motif was mainly on objects made by and/or for the Selgovae, a large and fairly well fortified tribe who left their largest hillfort when the Romans invaded in 79 AD. The date strongly suggests a possibility of artisans moving to another area where, later, they were engaged in making these dress-fasteners for the Romans. The snaffle-bit is most certainly a piece of warrior finery but it, too, might have had quite a history before it arrived where it was excavated. We cannot, however, rule out the possibility that Selgovae artisans stayed in their own territory and also made a few dress-fasteners for Roman soldiers along Hadrian's Wall.

Find spots do not give us a full picture, but at least it is better than imagining archaeologically excavated sites full of context and dateable associated finds. This is not the common pattern for Britain for this sort of material as a whole, but in this region, such is pure fantasy. Again, we return to one of the subjects of yesterday's post and realize that when political factors come into play, especially nationalistic archaeology and archaeologists who criticize the value of stray finds and public involvement from detectorists and collectors alike, even to the point of criticizing art-historical methodologies used on unprovenanced objects -- even if first published on Ebay! Then, we have to take into consideration the man as well as what he writes. It is the postmodern thing to do. Neglecting comparative typology means that you will be able to write very little about the subject, and what you do write will not be very good. This could even lead condemnations of the subject itself! Before I go onto the next topic, here are few more Swash-S links that you might find useful:

http://finds.org.uk/database/artefacts/record/id/86014
http://finds.org.uk/database/artefacts/record/id/55925
http://finds.org.uk/database/artefacts/record/id/148466
http://finds.org.uk/database/artefacts/record/id/275719

Printed literature can make tracking motifs and elements very difficult. Those who do have the sort of intellectual powers that are not very common. Others who lack such abilities might start applying The Fox and the Grapes phenomenon to the task. The databases we see are often not wide enough in their scope (apart from those like the Celtic Coin Index which includes archaeological excavated, stray find, collection and trade examples). Databases also suffer from designs too slanted to certain types of interests, mix and match terminology, and human inadequacies of all sorts. If you enter a term like "bronze", the interface should ask you if an analysis has defined the object as such and if not, should list all such variations as "copper or copper alloy". It might even ask you about brass. You cannot have identical things scattered all over the place. Clean the data before you build the database and base its design on clean data.

My  idea for the database of the future, has at its core, three qualities:

Object; Attributes; Subsidiary Objects

These are all "nested": an attribute can become an object in its own right, and have its own attributes, and subsidiary objects, and so can a subsidiary object. I will give you a simple example: A Ford Pinto is the object. Its steering wheel is a subsidiary object -- if someone has removed the steering wheel, it is still a subsidiary object that is designated as "missing". Cars are also an object, and in that case, the Ford Pinto is an attribute: cars can, but do not have to always include Ford Pintos. A parking lot is an object and cars are one of its attributes, as are marked parking spaces. Parking lots might also contain other objects that are usually found in such places: trucks, vans, even tanks and troop carriers.

All of these are attributes -- we are not that surprised to find them there and we might even expect to do so (depending on what sort of parking lot we are looking at). A box of oranges on the back seat of a Ford Pinto is a subsidiary object, not an attribute. If you see a Ford Pinto, you do not think of boxes of oranges. If the Ford Pinto is engulfed in flames, then the fire is a subsidiary object. However if there are enough Ford Pintos recorded as being in flames, then the subsidiary object will automatically be promoted to an attribute!

The box of oranges on the back seat of our burning Pinto also has sets of attributes and subsidiary objects --which might include flames,  too, in this case, but will likely include oranges, even if it is an empty orange box. The oranges have skins, or they might have been peeled, they might have seeds and so on -- everything is nested.

They connect via "fluid nodes" This is a very crude brain model, nowhere near as good as the one in our skulls, but one that can be programmed to rearrange the data both as it is entered into the database and as the database is queried at certain times. While behaving a bit like a search engine in some cases, it would be far more sophisticated as it is constantly altering our perceptions of reality. Each database would be used for specific sorts of tasks -- It might be a museum database, or it might be a police database. It could be used for any interest. Each database would be linked through networks to similar databases used for other things, but it would not connect the data unless it "spotted" certain associative patterns. For example, a stolen museum object on a police database would alert its original museum automatically. Or a stolen object might later show up in a museum, and it would alert the police database. Data is being sent back and forth all the time. We do not yet have the technology to build and operate such things. Until that time, we must muddle along as best we can, and always be on the lookout for things in places we do not think about, or like.

Have a good weekend!