Principles

When a sample is dated using the carbon-14 dating method, the result obtained by the laboratory (the so-called "conventional date", which consists of a mean value and its margin of error, the so-called "standard deviation") must be corrected so that it can be converted by the archaeologist into a real calendar age.

Expressed in years "before present" (the present being, by convention, fixed at the year 1950, the year in which the method was invented), the measurements carried out in the laboratory are based on the assumption that the level of carbon-14 in the atmosphere (and therefore present in all living beings) has been constant throughout time. However, we now know that this rate has varied considerably, giving rise to distortions of sometimes significant magnitude between calculated age and real age, with the latter generally being underestimated.

In order to correct for these distortions, a curve known as the "calibration;" curve has been established based on the carbon-14 dating of control samples whose age was previously known, having been measured using separate methods (dendrochronology, uranium-thorium method, etc.). This curve, which is regularly supplemented, clarified or corrected by geochronologists, accounts for the variations in the carbon-14 content of the Earth's atmosphere over the last 50,000 years.

Thus, thanks to the equivalences defined by the cross-referencing of dating methods, it is possible to determine the real age of the samples analysed. For example, for a "conventional" age of 15,000 ± 100 BP obtained in the laboratory, the "IntCal20" calibration curve indicates a significantly older chronological interval, between 18.6 and 18.1 thousand years before present (so-called "calibrated" dating).

The use of "calibrated" ages, expressed in the form of time intervals, should therefore be prioritised when discussing the age of an object, charcoal or painting that has been determined using carbon-14 dating. And this is all the more crucial when we want to place dated anthropological events back into their climatic and environmental contexts, the latter of these generally being defined based on chronological data obtained using methods that do not require this type of correction.

What are the consequences for Lascaux?

For the Recent Palaeolithic, the systematic application of calibration is a relatively recent phenomenon. Until the 2000s, ages obtained using the carbon-14 method were generally expressed as "conventional" dates, i.e. without correction. In the 1980s, for example, press articles headlined "Lascaux, 17,000 years ago", should today, after correction, read as "Lascaux, around 20,500 years ago".

Without going into the details of the controversies surrounding the age of the cave, we must bear in mind that it is important not to mix "conventional" and "calibrated" dating in the same discourse, lest it become impossible to understand anything!

To simplify matters, and based on the results of the three main dating programmes carried out to date (2020), two sets of chronological benchmarks are now acceptable for the Lascaux cave:

  • Set expressed in "conventional" dates (uncalibrated) 
    1970s: around 17,000 years before present 
    1990s/2000: around 19,000 years before present 
    2020: around 17,500 years before present;
  • Series expressed in calibrated dates (recommended use) 
    1970s: between 21 and 20.5 thousand years before present 
    1990s/2000s: between 23 and 22.5 thousand years before present 
    2020: between 21.5 and 21 thousand years before present.

See also: Why is Lascaux both 18,000 and 21,000 years old? [in French]