I gave a talk on early Coptic books a few weeks ago in which I mentioned the results of some radiocarbon analyses of Coptic codices. Now I learn that some of what I said is already out of date! A recent e-mail from Mike Holmes prompted me to check in on the OxCal website, where I saw that a new version of the OxCal calibration program is up and running using new data (IntCal20). The new calibration curve appears to have a noticeable effect on the dates of some early Christian manuscripts.
The results of radiocarbon analysis yield ages in terms of “radiocarbon years before present (BP).” Because levels of carbon-14 in the atmosphere fluctuate over time, these results need to be converted into calendar dates through a process of calibration. Objects with known dates, usually tree rings, are subjected to radiocarbon analysis, and these results are compiled and used to translate radiocarbon years BP into calendar dates for objects of unknown dates. Every so often, the accumulated and refined data for calibration is used to generate a new calibration curve.
The appearance of a new calibration curve means that older analyses can be re-calibrated using the new data. For this reason, authors who publish the results of radiocarbon analyses should report not only calendar ages but also radiocarbon years BP so that the calendar dates can be easily recalculated when new calibration data becomes available (less conscientious authors will omit BP dates). In some cases, the new calibration yields very little change to the resulting calendar dates, but in other cases, the change is meaningful.
I’ll illustrate with a couple of the examples that I mentioned in my talk. The first is the Manichean Synaxeis codex at the Chester Beatty Library. Samples of papyrus from a leaf of this codex were subjected to AMS analysis in 2013. The weighted mean for the samples was 1630 ± 22 radiocarbon years BP. The calibration (using IntCal13) yielded the following outcome:
The resulting span of calendar years was 350-534 CE (95.4% probability). The authors who published the analysis quite reasonably interpreted these results in the following way: “70 to 72% of the probability density function spans the period 380 CE to 435 CE. All things being equal, it is most probable that the Medinet Madi codices date to somewhere within the last quarter of the 4th to the early decades of the 5th centuries AD.”
Calibrating this radiocarbon years BP date with the new program noticeably adjusts the calendar dates.
The calendar range is now 405-538 CE (95.4% probability). Using the new calibration curve, then, the possibility of a fourth century date for the papyrus appears diminished. The probability of a date in the first part of the fifth century remains relatively high, but a date in the sixth century is no longer quite so easily excluded.
We see less change in the case of another Coptic book, the Glazier codex. In 1994, a piece of the leather wrapping band of the codex was tested and yielded a result of 1565 ±45 BP. When this data was published in 1999, the calibrated age was reported as 420-598 CE (95% probability). Using those numbers with the new curve, we get the following results:
The range of calendar dates is now 418-590 CE (95.4% probability). So, we see in this instance very little change.
From a general comparison of the IntCal13 and IntCal20 curves over the period of 100-700 CE, it seems that IntCal20 will shift calendar dates slightly later in most cases or keep them roughly the same. One of the frustrating things about the fluctuations of carbon-14 levels is reflected in the wiggle in the calibration curve that occurs during the third and fourth centuries CE. This dip has become more pronounced in IntCal20 with a higher peak and a lower trough, which means that the usefulness of AMS analysis remains very limited for distinguishing between material dating to the third century CE and material dating to the fourth century CE.