Prehistoric settlement of the Iberian Peninsula: Demographic dynamics versus C14 frequencies

Abstract

In this work we analyse 1,395 14C dates from the Iberian prehistory, belonging to the 13000-5000 BP chronological span. The only selection criterion considers their accuracy: a standard deviation of less than a century was required. It is the largest compilation published for this geographical area. Each date has been calibrated using OxCal program and IntCal13 calibration curve before the construction of the summed calibrated date probability distributions (SCDPD). We have followed the habitual methodology for the SCDPD, incorporating a mathematical routine into our process that avoids duplications of overlapping data from contexts repeatedly dated. At the same time, for a detailed evaluation, the exercise is also performed individually for different regions: Cantabrian region, Ebro basin, Central plateau, Catalonia, Valencia, Andalusia and Portugal. The aim of this work is to evaluate in which way this kind of compilations can be interpreted as demographic proxy, in this case as the Iberian population dynamics during the final Paleolithic and the first half of the Holocene. In this sense, we think that for interpreting the SCDPD it is firstly necessary to evaluate different factors that could distort the final distribution: factors related to the calibration curve itself, to taphonomic and research biases, or to the methodology employed. We also observe the individualised behaviors between the series of samples of long and short life that, contrary to the assumption, do not show striking dissonances. On the other hand, the participation of some very specific archaeological contexts, such as shell middens, can originate false demographic processes derived from their high archaeological visibility and their easy dating. For this reason, we justify the exclusion of certain archaeological sites, like the multiple megalithic burials for the end of our sequence, in these kind of exercises. They could create an overrepresentation. On the other hand, our interpretation is linked to the detailed knowledge of the archaeological contexts. We think that the understanding of a SCDPD only makes sense in the permanent dialogue between the radiocarbon data and the known archaeological record (and even the absent one). It is necessary to know the cultural (pre)historical development involved; the functionality of the sites; the archaeological research programs… The opposite is to make interpretations without a critical reading of the archaeological record. In this sense, as an example, the sharp drop in the date probability frequency after the start of the Neolithic, which in the logic of these kind of studies would attribute to a loss of population, could be related with changes in the strategies adopted by those societies. In conclusion, we think that the compilation of 14C dates can produce illusory effects, already warned critically by A. Bayliss, that they must be corrected by the archaeologist - the data by themselves do not do archeology-. The SCDPD cannot be taken as a proxy of prehistoric demography. The absence of data in temporary sections and geographic areas could refer to invisibility problems of the archaeological record, and the modifications in the probability frequencies are related with different human decisions about the exploitation of the territories, the number and characters of the deposits generated, or the inherent problems of the dating method and the exploitation of the results. The demographic patterns are hidden under these problems.