Dating ancient Mediterranean civilisations

Posted June 27th, 2010 by Andrea Vianello

Since the earliest day of archaeological research there has been one problem: determine the chronology of artefacts and cultures. All human cultures record time in some way, and one of the most ancient such methods is the listing of kings or chiefs, as it happened in Egypt, the ancient Near East, and Rome. The question of the origin of time or the age of Earth has interested many people and it can be recognised in the Near Eastern lists, which start with mythological characters. A turning point in the matter, at least in western cultures, happened in the mid-seventeenth century, when Archbishop James Ussher determined that Earth had been created on the 23rd of October 4004 BC, according to a study of the Bible. Ever since that moment, western scholars have attempted to proof or undermine that claim, especially through scientific methods as science was just beginning to revolutionise the western way of thinking. At first geologists started to look at the age of Earth, and then astronomers expanded the question to the whole Universe. Eventually archaeology came around and focussed on the thornier problem of determining when humans appeared on Earth and establish the succession of human cultures. Several timelines of human cultures have been produced, and you can find a very good one on the Internet, courtesy of the Metropolitan Museum. Archaeology however split for convenience the chronological sequences between the approximate absolute chronology that linked human cultures to geological and astronomical periods as well as known dates and the relative chronologies that provided a name to archaeological temporal contexts and linked many of them on a temporal scale.

Chronology has been an essential aspect of archaeological research. In 1960 Professor Willard F. Libby obtained the only Nobel Prize so far awarded for research of great relevance to archaeology, thanks to his newly devised method of radiocarbon dating, or C14 dating. Since then, the method has been improved (mostly by adding calibrations) and other dating methods have been entered the regular use of archaeologists. Anyone that is familiar with radiocarbon dates will know that any precise date is followed by the indication of probability, i.e. how many years off the date could be. It is rare however to see recent dates with approximations of hundreds of years as it was the case at the beginning of radiocarbon dating. Dating the Egyptian and Aegean (pre- and proto-Greek) cultures should therefore have been achieved in 2010, 50 years since radiocarbon dating began to be used, because those are some of the best known cultures with the longest timespan of archaeological research, historical lists of kings for Egypt, and high interest that never abated among public and scholars. Instead, this year we have to come to the conclusion that we still do not know, despite the best efforts in the field.

A study published by Christopher Bronk Ramsey et alii in Science in June 2010 reveals that whilst there are periods for which the different chronological methods agree, there are also differences, some of which can be attributed to seasonal differences producing imbalances in the regional calibration of isotopes, but there are also differences that we cannot explain yet. A huge problem is the scientific dating of the Santorini (Thera) eruption, which simply disagrees with established relative chronologies and their synchronisation with absolute dates and the relative chronologies of other cultures. The most precise range of dates for the eruption is 1627 to 1600 BC according to Sturt W. Manning, which disagrees with relative chronologies. I have encountered the problem myself, and for my 2005 book I have produced a table of concordance of several chronologies in the Mediterranean (especially Aegean to central Mediterranean), which shows the Helladic chronology to be off by 125 years at least (figure 1).


Figure 1: chronologies after Vianello 2005, 104. The vertical red line (Thera eruption range) should fall within the LH I B period.

I should state that I support the scientific chronology that would stretch the Mycenaean period (recognisable by the “LH” prefix in the “Helladic” row) to a better length. According to my own studies on Aegean-type (including Mycenaean-type) products in the Mediterranean, west of the Aegean, such an early date would agree with the length of time and local relative chronologies there and better explain the sudden appearance and wealth of the Mycenaeans, as witnessed by Circle A and B burials at Mycenae.

Bronk Ramsey in his study reports that the time of the eruption, which can be associated to artefacts of the beginning of the Egyptian New Kingdom, should be 1570 to 1544 BC (Ahmose’s reign, p. 1556), also using radiocarbon dating, which is at least 30 years off what has been proposed by Manning. The problems do not end here, since the dating of a key archaeological site, Tell el-Dabca, disagrees with  all dating. Using historical archaeology, Manfred Bietak suggests that the eruption would have happened between 1530 and 1480 BC, but after radiocarbon dating and calibration the date is 1720 to 1640 BC, at least 13 years before Manning’s proposal and 70 years before Bonk Ramsay’s chronology. The chronologies proposed by Bronk Ramsey and Bietak differ between 70 and 170 years (figure 2).


Figure 2: Chronology after Bruins, Science 328(5985), 1490. It shows some incongruences in the dating between Bronk Ramsey (yellow), Bietak (blue) and Manning (red column to the left).

In figure 2, the eruption should fall within the end of Late Minoan I A or better the very beginning of Late Minoan I B, and the beginning of the New Kingdom. It would be easy to suggest that Bietak is wrong, because most of the radiocarbon dates from Tell el-Dabca are systematically 100 to 200 years older, but there is nothing wrong on how such dates have been obtained and the site has one of the best archaeological sequence (i.e. relative chronology) in Egypt for the period being debated.

The result is that there is no secure dating for the Egyptian chronology, to a degree that the older Middle Kingdom presents far fewer problems than the New Kingdom. The Aegean and Egyptian chronologies remain detached, despite their geographical proximity (figure 3).


Figure 3: map showing the location of Santorini, Tell el-Daba and Egypt. Copyright by Bing Maps, Microsoft.

The solution may be in more radiocarbon dates, as suggested Hendrik J. Bruins in the same issue of Science reporting Bronk Ramsay’s study. It is also possible that the Santorini eruption (or particular environmental conditions) may have affected the calibration of radiocarbon dating in certain areas in ways that we cannot decipher yet. If this is true, then there will be important periods in history when dating will be very loosely determined, and because environmental changes and cultural changes may be quite intertwined (see Jared Diamond’s Collapse or head to Edge 114 or even his TED talk for some of his ideas) and their correlation may be critical to understand the past. For instance, the period of the Santorini eruption marks the beginning of the New Kingdom in Egypt, after the end of the Hyksos, and the beginning of the Mycenaeans in the Aegean, after the collapse of the Minoans. It was hardly an uneventful time.

It is a sobering thought that radiocarbon dating has yet to fulfil its promise of precise chronologies. One of the central issues in archaeology remains just that. According to the most recent studies, the Santorini eruption may have occurred at any time between 1720 and 1550 BC, and that is a period 170 years long. The best compromise among the chronologies would see Bietak’s 1640 BC date being closer to Manning’s 1627 BC than to Bronk Ramsey’s 1570 BC and therefore the eruption would have happened in the second half of the seventeenth century BC. If that chronology is correct, the Egyptian chronology would have to be rewritten from Nubia to the delta of the Nile. If Bronk Ramsey is correct, the chronology of the Aegean and the delta of the Nile would be at least 50 years older if we consider the traditional one, or at least 30 years  more recent if we consider the radiocarbon dating, which also matches the dendrochronology obtained in northern Italy. Since changes of at least 30 to 170 years need to be made, it is obvious that all absolute dates suggested for kings reigning across the Mediterranean are imprecise.

To conclude, all recent radiocarbon studies are welcome, but none of them as yet allows anyone to move away from established chronologies. Dating the volcanic eruption securely by a variety of means would provide a precisely dated reference point in time. Since the ashes from Santorini have been found dispersed across many lands, dating the time of deposition of the ashes would automatically date many contexts and provide help for calibrating regional isotopes as well for synchronising different regional relative chronologies. Until then, it is safer to stick to relative chronologies. It is my unprofessional (I am not involved in radiocarbon dating) opinion that the date provided by Manning may be the most precise because that research focused on just one being very aware of the possible chemical interference caused by the volcanic eruption, whereas all other chronologies ignore that aspect. For instance, I note in Bronk Ramsay’s study (fig. 2, p. 1555) that “there is also a scatter of outliers (…) that show no systematic pattern and have no single explanation”. These outliers are concentrated in the period 1600 to 1100 BC and may suggest that some further uncalibrated interference occurred during this period and are missing, at least in that concentration, from the calibration curves of the Old and Middle Kingdom, even if it should be considered that less samples were available for older periods. A seasonal effect has been detected and a local offset of 20 +- 5 years has been applied to the dates of the New Kingdom, and it may be that such seasonal effect, which may be correct for periods more distant from the eruptive event, was far more significant at that time. A robustness test as carried out by Bronk Ramsey would detect a prolonged effect, but would probably not detect a time-constrained, exceptional anomaly. Its effect would be diluted in the longer periods considered. There is therefore in my opinion a greater chance that the New Kingdom may have begun earlier, around 1600 BC, with radiocarbon dates pointing to a date as early as 1570 BC. In that case, both Aegean and Egyptian chronologies would be earlier of at least 50 years than previously thought, and that difference may be accounted for by calibration problems due to the volcanic eruption for what concerns the radiocarbon dating and by problems in recording time at a stressful time for the ancient societies.