Final Report Summary - EUROFARM (Transmission of innovations: comparison and modelling of early farming and associated technologies in Europe)
The EUROFARM project aims at understanding the factors structuring the spread of early farming practices across Europe, with a focus on the western Balkans. To this purpose, the project works through a combination of fieldwork (northern Bosnia, southern Montenegro), analysis of existing archaeological collections (Bosnia and Herzegovina, Croatia, Montenegro, Serbia), meta-analysis of the literature and computational modelling. From a methodological point of view, the project is inherently comparative, tackling different categories of evidence (e.g. plant and animal remains, lithics, pottery, radiocarbon), and focusing on the two main corridors through which early farming spread (Adriatic basin and Danube catchment).
While earlier work generally considered the role of the human agents involved (i.e. late local foragers vs. incoming farmers), the EUROFARM project focuses on the description, analysis and simulation of the archaeological record, in order to understand how its variability for this period and area informs us about the diffusion of agriculture, stock-breeding, and associated technological innovations (e.g. pottery, settlement pattern). The earlier stages of the project were dedicated to surveying the – very fragmented – literature, and then supplementing this record by analysis of existing museum collections (radiocarbon, zooarchaeology, flotation of archaeobotanical samples, lithics, pottery technology), and targeted fieldwork in areas largely devoid of previous archaeological investigations. The next stage corresponds to the analysis of the resulting datasets, through an original combination of qualitative and quantitave methods. Qualitative assessment of the data is a key – yet often overlooked – stage in order to identify and characterise biases existing through different categories of information (e.g. spatial discrepancy in the distribution of 14C dates, impact of sieving pratices on zooarchaeological assemblages, extend of sampling and recording of plant macro-remains), and to factor them in subsequent analyses. A variety of statistical tools were used (e.g. summed radiocarbon probability distributions, correspondence analysis, range of diversity indexes borrowed from numerical ecology). Furthermore, agent-based models were developed in order to simulate, within a formal mathematical framework, several hypotheses regarding the respective role of spatial, ecological and demographic factors in shaping the archaeological record. These simulations were compared to archaeological data using similar statistical tools in order to test the validity of the aforementioned hypotheses and to evaluate the most likely scenarii.
The final results indicate a limited role of the last foragers in the diffusion of early farming, as their population density seems to have been very low, especially when compared to that of the incoming new farming groups. Both qualitative and quantitative analyses demonstrate clear differences in the archaeological record of the earlier farming communities spreading, on the one hand, across the Adriatic basin and, on the other hand, across the Danube catchment. Although these differences occur in two ecologically-distinct corridors, they cannot be accounted for solely in ecological terms. Taking a long-term perspective helps to solve this apparent paradox. Both diffusion streams, whilst retaining original features, follow the same temporal trajectory characterised by initial low diversity during the spread, followed by a gradual rise in diversity throughout the rest of the Neolithic sequence. Computational models indicate that such two-step trajectory can be explained by a process of founder-effect, through which an expanding population undergoes a drop in diversity, whilst the later increase is related to continuous interaction between groups located at the back and at the front of the expansion wave. This process, though described in a single case-study, is systemic and therefore of application for the rest of the spread of farming across Europe, and, generally speaking, to any human dispersal driven by demographic pressure.
While earlier work generally considered the role of the human agents involved (i.e. late local foragers vs. incoming farmers), the EUROFARM project focuses on the description, analysis and simulation of the archaeological record, in order to understand how its variability for this period and area informs us about the diffusion of agriculture, stock-breeding, and associated technological innovations (e.g. pottery, settlement pattern). The earlier stages of the project were dedicated to surveying the – very fragmented – literature, and then supplementing this record by analysis of existing museum collections (radiocarbon, zooarchaeology, flotation of archaeobotanical samples, lithics, pottery technology), and targeted fieldwork in areas largely devoid of previous archaeological investigations. The next stage corresponds to the analysis of the resulting datasets, through an original combination of qualitative and quantitave methods. Qualitative assessment of the data is a key – yet often overlooked – stage in order to identify and characterise biases existing through different categories of information (e.g. spatial discrepancy in the distribution of 14C dates, impact of sieving pratices on zooarchaeological assemblages, extend of sampling and recording of plant macro-remains), and to factor them in subsequent analyses. A variety of statistical tools were used (e.g. summed radiocarbon probability distributions, correspondence analysis, range of diversity indexes borrowed from numerical ecology). Furthermore, agent-based models were developed in order to simulate, within a formal mathematical framework, several hypotheses regarding the respective role of spatial, ecological and demographic factors in shaping the archaeological record. These simulations were compared to archaeological data using similar statistical tools in order to test the validity of the aforementioned hypotheses and to evaluate the most likely scenarii.
The final results indicate a limited role of the last foragers in the diffusion of early farming, as their population density seems to have been very low, especially when compared to that of the incoming new farming groups. Both qualitative and quantitative analyses demonstrate clear differences in the archaeological record of the earlier farming communities spreading, on the one hand, across the Adriatic basin and, on the other hand, across the Danube catchment. Although these differences occur in two ecologically-distinct corridors, they cannot be accounted for solely in ecological terms. Taking a long-term perspective helps to solve this apparent paradox. Both diffusion streams, whilst retaining original features, follow the same temporal trajectory characterised by initial low diversity during the spread, followed by a gradual rise in diversity throughout the rest of the Neolithic sequence. Computational models indicate that such two-step trajectory can be explained by a process of founder-effect, through which an expanding population undergoes a drop in diversity, whilst the later increase is related to continuous interaction between groups located at the back and at the front of the expansion wave. This process, though described in a single case-study, is systemic and therefore of application for the rest of the spread of farming across Europe, and, generally speaking, to any human dispersal driven by demographic pressure.