Wednesday, December 3, 2008

The Santorini Eruption and the limits of knowledge

For at least a decade now there has been a major discrepancy in the dating of the eruption of the volcano at Thera/Santorini, which some people think destroyed the Minoan civilization. The eruption is mentioned in Egyptian records that can be tied into the Egyptian king lists, which extend unbroken down to Roman times, and the Egyptian records give the date of 1530 BC. But archaeologists have found numerous pieces of charred wood from the volcanic ash layer, and the radiocarbon dates they get are significantly older. The latest date comes from a complete olive branch buried in the eruption, and it comes out as 1613 BC, plus or minus 10 years.

Now nobody really expects radiocarbon dates to be exactly right all the time, so this discrepancy has been swept under the rug for years. But as the dates pile up it is really starting to seem that either the calibration curve we use for converting radiocarbon years to calendar years is out of whack, or there is something seriously wrong with the Egyptian king lists, which have been treated as a reliable source for centuries. One of the two pillars we use for dating events in the ancient Mediterranean is wrong.

I mention this because I have long been and remain a skeptic about our ability to have precise, accurate knowledge about the distant past. The further we look back in time, the more our vision is blurred, and the more we are dealing with probabilities rather than certainties. Radiocarbon dating is a wonderful tool, but it is subject to many limitations -- the changing level of C14 in the atmosphere (which is why we need the calibration curve), the possibility of local variations in the C14 level created by things like the ocean reservoir effect (the ocean is full of old carbon) or the outgassing of ancient carbon by volcanoes, the re-use of old wood, and the movement of charcoal through the soil because of worms, groundhogs, tree roots, and the like. I once submitted three charcoal samples from a single large pit and got back dates that translate to AD 1700, AD 650 (which matched the artifacts), and 2000 BC.

No comments: