The transmission of pottery technology among prehistoric European hunter-gatherers

Abstract

Human history has been shaped by global dispersals of technologies, although understanding of what enabled these processes is limited. Here, we explore the behavioural mechanisms that led to the emergence of pottery among hunter-gatherer communities in Europe during the mid-Holocene. Through radiocarbon dating, we propose this dispersal occurred at a far faster rate than previously thought. Chemical characterization of organic residues shows that European hunter-gatherer pottery had a function structured around regional culinary practices rather than environmental factors. Analysis of the forms, decoration and technological choices suggests that knowledge of pottery spread through a process of cultural transmission. We demonstrate a correlation between the physical properties of pots and how they were used, reflecting social traditions inherited by successive generations of hunter-gatherers. Taken together the evidence supports kinship-driven, super-regional communication networks that existed long before other major innovations such as agriculture, writing, urbanism or metallurgy.

Fig. 1. Study area, site locations and examples of reconstructed forms for the pottery styles included in this study.

Illustrated are reconstructions from the (1) Eastern Baltic, (2) Western Baltic, (3) Upper Dnieper, (4) Bug-Dniester, (5) Middle Don, (6) Lower Don, (7) Northern Caspian, (8) Lower Volga, (9) Middle Volga and (10) Upper Volga regions. Map based on the ASTER Global DEM v.3 with ecotones based on generalized mid-Holocene estimates from ref. ; it should be noted that the boundary between steppe and forest is likely to have been highly diffuse.

Fig. 2. A spatiotemporal model for the spread of pottery technology.

a, Predicted start date for pottery vessels in the region based on spatial–temporal regression models. b, The modelled posterior probability distribution for selected locations based on radiocarbon evidence compared to the predicted dates. For details on how the models are constructed, see Supplementary Methods.

Fig. 3. Lipid residue analysis of hunter-gatherer ceramics.

a,b, δ¹³C_16:0 and δ¹³C_18:0 values of sample with (a) and without (b) aquatic biomarkers and statistical ellipses (1σ) of modern animal fats from Eastern Europe (provided online; see Data Availability Statement). c, Relative frequency of identified food commodities by region following the criteria described in the Methods. d, A surface model interpolating Δ¹³C values (for error maps and other results, see Supplementary Fig. 12).

Fig. 4. Biplots of correspondence analysis scores for traits (site-wise) recorded through inspection of the archaeological ceramics and the organic residue analysis.

Axes orientation chosen to best illustrate the recapitulation of geographic coordinates.

Fig. 5. Correlation coefficients for pairwise Mantel tests for Jaccard and geographical distance matrices.

The partial Mantel tests indicate the strength of the correlation between organic residues and pottery characteristics when holding a great-circle distance constant. The Mantel correlation of distance matrices tests a null hypothesis that there is no relationship between the cultural, biomolecular and geographical ‘distance’. For cases in which the null hypothesis was rejected, the Pearson correlation coefficients r produced by these Mantel tests are illustrated, with P values (two-sided, null hypothesis r = 0) and 95% CIs contained in Supplementary Table 7.

Fig. 6. Mantel correlograms showing the scale of spatial autocorrelation.

Significant similarity (Mantel r > 0) or dissimilarity (Mantel r < 0) is indicated by filled circles. Error bars indicate bootstrapped 95% CIs.

Fig. 7. The spatial, temporal and network proximity of selected sites in the study.

a, Locations for sites with isochrones representing a spatio-temporal model of diffusion. b, Posterior probability distributions of the date of the start of pottery use at each site. c, NeighborNet networks for ceramic and biomolecular dissimilarity data.