Bayesian chronological modelling of the Early Iron Age in Southern Jutland, Denmark

The dissertation investigates temporal processes of change in Early Iron Age material culture from Southern Jutland, Denmark (c.500-200 BC). The chronological framework of the period is mainly based on typo-chronological analyses of metalwork, and to a lesser degree of pottery, from large urnfield cemeteries. The chronological framework is unconstrained by scientific dating, which makes it difficult to correlate metalwork and pottery that appear disproportionate in funerary and settlement contexts. The dissertation continues a long tradition of chronological research using artefact assemblages from the urnfields, but it also presents the first large-scale 14C dataset in support of an absolute chronological framework. The majority of datable material from the urnfields is cremated bone and because 14C dating cremated bone is a relatively new method, it is necessary to investigate methodological aspects relating to laboratory techniques. A laboratory intercomparison demonstrates that differences in pretreatment do not affect the radiocarbon results and that results are reproducible between laboratories (Rose et al. 2019). Another important aspect when dating cremated bone is that it will be affected by wood-age offsets through carbon substitution, effectively causing cause calendar date offsets between the real cremation events and the radiocarbon dates. The dissertation explores statistical modelling of offsets through additional case studies (Rose et al. 2018; Kristiansen et al. forthcoming), before proposing a new statistical model aimed at wood-age offsets in cremated bone (Rose 2020). The new statistical model is based on the empirical distribution of wood-age offset in an experimental dataset and on archaeological combinations of cremated bone and archaeobotanical remains. The statistical model is applied to all dates on cremated bone from 95 burials from three urnfields, Aarupgaard, Aarre and Søhale, in order to take account of their inherent wood-age offsets. The corrected dates provide indirect dates on the associated artefacts of metalwork and pottery and together with prior information on artefact typology and site formation processes (e.g. how the urnfields developed spatially over time) this is combined in a Bayesian framework aimed at modelling temporal processes of change in the material culture. 24 artefact types are modelled and out of these, nine metalwork and five pottery currencies have sufficient dates to be modelled independently. The models support an existing typological sequence of metalwork, but it also suggests that the ceramic sequence needs to be revised. The models suggest discrepancies in existing ceramic sequences (Jensen 2005), and in contrast to earlier works (Becker 1961; Jensen 2005), it is proposed that pins with circular head is treated as a single type regardless of head size. It is also proposed to divide the different types of belt equipment into an early group and a late group to minimize misclassification. The currency models provide a necessary correlation of metalwork and pottery, but they also offer independent evidence of the temporal processes of change in material culture by demonstrating periods with particularly rapid and slower change of the Early Iron Age material culture. The currency models have inhomogeneous temporal structures, with most metalwork being in use for shorter periods than pottery, but how quickly or slowly a type was introduced, and later abandoned, is very individual. Overall metalwork is found to be chronological sensitive, whereas pottery appears to be more conservative in relation to change. The dissertation evaluates the chronological framework based on the new 14C evidence and the transitional from Bronze to Iron Age in Southern Jutland is estimated as a period starting in the early 7th century BC. The Bronze-Iron Age transformation is traditionally considered to occur c.500 BC and to coincide with the introduction of the urnfield phenomenon, but the urnfields are now estimated to start already in the early 6th century BC. The urnfields are abandoned in the last half of the 3rd century BC, which coincides with the traditional transformation from Early to Late Pre-Roman Iron Age c.250-200 BC