Afternoon tea with my friends Åke and Petra inspired me to re-run this post from March 2010. Professional dendrochronology is still almost entirely a black-box in-house endeavour, that is, it is still not great science. Field archaeologists: when you saw your wood samples for dendro, get two samples and send one to the amateur community! They practice open data sharing. Check out Åke & Petra’s web site!
How long ago was the time of Emperor Augustus? Most educated people, including professional historians and archaeologists, will reply “about 2000 years” if you ask them. But a considerable number of amateur dendrochronologists say “about 1800 years”. And because of an unfortunate peculiarity in how professional dendrochronologists work, it is very hard to convince these dissident amateurs that they are wrong. Because they’re actually thinking straight given the data available to them.
If you look at published dendro curves for the transalpine provinces of the Empire, you find that they contain two main blocks of information covering the past 2500 years or so. There’s one that extends solidly from today and back to about AD 400, consisting of many tightly interlinked samples. And then there’s a Roman-era block that is also quite solid internally. But between the two blocks is a period of about 200 years when there are very few samples. It appears to be hard to find preserved timber that grew in the 3rd and 4th centuries AD. There are enough samples to satisfy professionals that they actually have the whole 2000 years covered, but the sample overlaps for the gap between the blocks are few and rather weak.
Professional dendrochronologists explain this lack of finds by reference to the cessation of Roman building projects and to deforestation during the Imperial centuries. According to the accepted model, the reason that there are so few samples covering the gap between the blocks is that few trees were of a suitable age for construction timber during that time and even fewer were used to build anything that has been preserved. Trees that fell into bogs and rivers at the time would have too few rings to be of much use to dendrochronology.
Dissident amateurs instead think that the Roman block and the recent block have been joined incorrectly, and that there shouldn’t be any gap at all between them in the diagrams. According to them, the professionals have been fooled by the early historians Dionysius Exiguus and Beda Venerabilis into thinking that the Western Empire fell 1600 years ago, using this as an axiom in their work with the dendro curves, when in fact it happened only 1400 years ago. A common idea about why this should be so is that the Church of Rome added a couple of centuries to its age to gain legitimacy: in other words, a conspiracy of early historians.
I mentioned published dendro curves. The rub here is that most dendro data are never published. They are kept as in-house secrets in dendro labs in order for these to be able to sell their services to archaeologists. So when the amateurs challenge the professionals’ opinion, all the latter can reply is “We know we’re right but we can’t show you how we know”. And that is of course an unscientific approach to the issue. The amateurs rarely get access to 1st Millennium wood samples, and basically have to work with the past 1000 years in their own studies. And so they cultivate a dissident opinion that could swiftly be laid to rest — or be accepted as fact — if wood samples and measurement databases were only made public.
My guess, though, is that any Roman archaeologist could solve the controversy quite easily, perhaps even using published radiocarbon dates. All you need are a couple of well-sourced dates for contexts known to be from about the time of the first emperors, such as Pompeii. (But if you know that a context is from that time, then you have very little reason to pay for radiocarbon dating.) Because although the calibration curve for radiocarbon depends on dendrochronology, several of the available datasets are not from European wood samples. And there is of course no inherent bias about where on the diagram the fall of Rome should be in North American dendrochronology, for instance.