Dotty Dot-Dots and Late Roman Grots

Coiney dot-distribution map (Richard Henry)

Richard Henry writes on Twitter:

" I am now starting to compile Roman coin profiles for the 10,500 parishes in England for my PhD. Late Roman coins recorded with @findsorguk and coin hoards (AD 364-378) from the south west here highlight some interesting distributions, particularly around Taunton. [...] Here are the parishes with above average coin loss in the south west. The data includes @findsorguk, the Roman Rural Settlement, the work of Philippa Walton and other sources. In total I have recorded just under 100,000 coins from the south west".

On being asked by Joseph Lewis "what are the soft/hard constraints [on this distribution pattern]?" the researcher replies: "They are constraints to metal detecting either where it is banned (hard constraints) or less likely (soft constraints). It explains some of the major gaps you see in PAS distributions such as Salisbury Plain". In fact what the pattern represents is not only collecting affordances, but also preferences, but above all reporting. With an estimated 27000 detectorists and 10500 parishes (and just 100 000 coins from a group of them), it's not hard to see that for an activity that has been going on now for fifty years (!) if there were consistent records, there should be a super abundance of information on their coin finds in that period. In fact all that is happened is that sites have been emptied into pockets and a sample of unknown representativeness has been recorded. Brodie talks of why archaeologists love the PAS and get "entangled" with it - because it gives them material on a plate for this kind of reseach. So it is interesting to see the next comment:

Cat Lodge @CatLodge1 11 g.
W odpowiedzi do @richardhenryfsa @Durotrigesdig and @findsorguk
Plenty more on the North Somerset HER, if you're interested!

She later explained that these consist of "lots of antiquarian finds rather than detected", but then these too have their own recording biases and distortions.

I have mixed feelings here. First it is great to see a text on coins that treats findspot as information. Most coineys like to study the pictures and writing on a pile of decontextualised coins tipped out on a tabletop and call it "numismatic research" (see the ACCG in the US). Here is quite clear that knowing what is from where (and found in relation to what) is important information.  

Secondly though this seems to be a repeat of patters that were being established for just this region in the 1970s (Peter Fowler comes to mind), hillforts with Late Roman pottery (including African Redslip Ware as I vaguely recall), villas with late occupation (late mosaics too). I wonder whether just studying the coins as a statistical pattern from chance reports is going to add a lot to this pattern? Also readers will know, I am not a great fan of any interpretations that involve merely making a dot distribution map  and talking about the broad pattern (though must admit this one is rather endearing). 

I'll be interested to see what this research produces in the end and whether all that site trashing by artefact hunters and collectors of which this is a minor byproduct can be justified.