Population Games and Rorschach Tests September 6, 2011Posted by Beachcombing in : Ancient, Medieval , trackback
Beachcombing had some fun the other day writing about ancient history and population estimates. Last night reading in the ‘wee hours’ he came across another lovely example of this: the insane modern debate about the population of Roman Britain.
Now post-war estimates for the population of Roman Britain have gone as low as 200,000 and as high as 5,000,000. Estimates for the Anglo-Saxon invaders who were invited or/and fought their way into Britain c. 400 have ranged, meanwhile, from as few as 10,000 to as many as 1,500,000.
The casual reader may be struck or even surprised at the range of numbers offered here and it is a bewilderment that Beachcombing shares particularly given that there are no census returns and virtually no documentation from the period. But then attempts have been made to ‘scientise’ these numbers by comparing the numbers involved to tribes in modern Yemen (really) or even, Beach loves this, by counting bodies in Romano-British cemeteries.
Forget the statistical voodoo though. An eleventh-century work, Domesday Book remains our first real evidence for fifth-century population: and this is perhaps another way of saying that we have no evidence at all.
Domesday is, in itself, a sobering reminder of the absurdity of scholars arguing over this or that number for fifth-century Britain. Historians of Norman England using the Domesday census, even with written records before them, still argue about numbers: two or four million represent the extremes.
Population estimates for fifth-century Britain, for either the British or the Anglo-Saxons are, instead, a matter of archaeological guess-work. Admittedly the temptation to put some kind of a number down is great. But when this temptation comes one need only think of the Romano-British city of Silchester where population estimates have ranged over the years from 300-7500. Silchester (Calleva to friends) is our best excavated Romano-British city, and if numbers for a delimited unit like this range over 1000% is it not optimistic for historians to talk of getting within 200% for a national estimate?
And yet the ritual population game continues. A young archaeologist on the make discovers that population density was higher or lower in the fifth century in a Yorkshire valley (or at least in the ten square metres of that valley he excavates), and so the numbers are bumped up or bumped down by twenty percent.
However, as the original numbers are sheer fantasy this exercise proves nothing. It would be more meaningful to say ‘Smith estimated H in 1890, this was adapted by Freeman to M after the Great War and then Darke suggested, on the basis of his work on the Downs, Q: personally I’d like to scale our estimates back to G’.
Classicists don’t, like disturbed children, do Rorschach tests, they guess at the population of Roman provinces. And their estimates reveal their dirtiest secrets. Looking over the numbers, for example, given for Anglo-Saxon invaders it is striking that those who believe the Anglo-Saxons were largely peaceful ‘long-haired tourists who just occasionally beat up the locals’ keep the numbers of that rabid rabble (Beachcombing’s ancestors) as low as possible.
Here’s hoping that Genetics will come riding in on a white charger.
Any other mad historical estimates from historians, Beachcombing needs to know! drbeachcombing AT yahoo DOT com