FLO Explains UK's Medieval Coin Finds: Medieval Fairs

            155449310136 Collection of Medieval hammered coins found                           
metal detecting (eBayUK:  lotsauctions (28890*) Norwich)

• 
  

) 

Over on social media some brave PAS stalwarts attempt to do some 'outreach'. A brave attempt at generalisation followed the comments of one detector-sceptic archaeologist about the unrepresentatively-huge number of coin records in the PAS "database" (499,031 in the 1,057,185 PASD records) as opposed to what archaeological sites in England and Wales (stripped in the search for collectables) actually contain. He points out that "pottery turns up in large quantities and coins don't. Stratifed coin finds on medieval rural sites are like hen's teeth, yet detectorists seem to find loads in the plough-soil. Why is that? Where are they coming from?". The obvious answer is that UK detectorists do not collect representastive assemblages of the archaeological assemblages that they excploit, but "cherry pick" them for the bits they are interested in, and then they report a fragmenmt of that (selected on unclear criteria- nobody has studied this). The FLO (Dr Simon Maslin) however has another explanation:

A lot of medieval coin losses happen on sites of seasonal fairs and markets, which don't have structural remains. Also middening redistributes coins (along with everything else) on fields. Also: detectorists cover huge areas compared to excavations so there's a probability issue.

Hmm. So, leaving aside those 'fairs', if they cover a large area, they find more coins than archaeolgists. That's what he says. To clarify, let us take five classes of artefacts, A, B, C, D and E. Together they make up the bulk of the material culture of a particular group of communinities (like the one archaelogists conventiomally label 'cultures'). So wherever there is random activity of those communities, that material culture is in use. It follows then that these elements of material culture (the A, B, C etc artefacts) can in these places be lost, left, discarded or depositied. Like in a settlement and its assciated infrastructure. Occasionally the nature of the past activities of these communities - both at such places, or in other ones, will lead to a differential pattern of deposition of artefacts (let's say for example, artefacts C [coins] could be intentionally deposited in hoards or graves). But the PAS-big-data-apologists will say that the extent of detecting and its random, coverall, approach will level this out. So why does the PAS database show a skewed relationship between reported examples of artefact class 'C' [coins] in comparison with artefacts A, B, D, E? [everything else]? The effect is compounded even further in that many medieval coins found by detector-using artefact hunters are reported to other databases, like the ones in the Fitzwilliam Museum or the UKDFD without being in the PAS records. Also at this moment there are (checks) 4,200+ (lots of) hammered medieval British coins on sale by British sellers on eBayUK (some of those will be sold in ten days and replaced by others, in this way upwards of probably about 70 000 of them will pass through eBay each year - the PAS database has 82,970). 

We also have to bear in mind that the majority of English medieval pennies are very thin discs of hammered silver (of a few grams of metal) and comparatively difficult to find with a metal detector when randomly searching for single items over a wide area. Indeed their "tally" of such coins is used by UK detectorists as an index of their own, personal, skill.

So why do the so-called "big data" of the PAS not level up the effects of various deposition processes to produce the same A/B/C/D/E ratios as the excavated material culture of medieval societies? "Middening" is not the answer as Dr Maslin points out: "middening redistributes coins along with everything else on fields", so if everything is going on the midden (and why would "only coins" be going on the midden with the mucked-out cow poo?) you'd still get the same A, B, C, D, E ratios from metal detecting the fields.

This brings us to: "A lot of medieval coin losses happen on sites of seasonal fairs and markets, which don't have structural remains". This builds on the texts of numerous "how to" metal detecting manuals that suggest locating such sites (through written records, maps, field names and folklore) as potentially "productive"  sites to find stuff. It seems that the FLO is suggesting that there were other sites that are unknown from these sources, and that artefact hunters stumble across them and then hoover all the coins out. 

This should be easy to prove statistically by nearest neighbour analysis of surface findspots, a 'lost' market found by reporting metal detectorists exploiting it would be shown by the finding of discrete clusters of coin finds of similar dates and nature. I'd suggest they would be found at communication nodes, places where tracks converge for example, probably in a place where different zones of landuse are in contact. In fact professor Michael Lewis (Head of PAS) and Eljas Oksanen, got a lot of grant money to look at this issue (see here for example). I wrote about it here " Artefact collecting: creating or destroying the archaeological record?" Folia Praehistorica Posnaniensia 25 (2020) 39-91, I did not find the material available to me on this at all convincing (there is also reference in my text to recent work on what British "metal detecting" tells us about medieval coin circulation and use in general).