Finding the Limits of the Limes

Posted on March 10, 2017


It was heart breaking to find out the the ‘Limes’ was not the fruit but the defensive boundary of parts of the Roman Empire. But I quickly got over that when I found out that the conference was on modelling the past, an interest of mine. These are the videos from the conference for the ‘Finding the Limits of the Limes’ project. You can find out more about project here- -and the videos of the presentations are here:

Setting the scene: characterising Batavian society at the edge of empire in the Dutch river area

Nico Roymans (Vrije Universiteit Amsterdam)

Modelling the cultural landscape of the Dutch Roman Limes: approach, results and prospects Philip Verhagen (Vrije Universiteit Amsterdam)

What did the Romans ever do for us?

Wim Jongman, University of Groningen

‘What did the Romans ever do for us?’ Few have nailed the fundamental question of Roman provincial history and archaeology with greater precision than the Monty Python team in their Life of Brian. For decades, the answer has been a pretty damning one, with a florid rhetoric of post-colonial scepticism. Similarly, even for its imperial core the success of the Roman economy has been seriously questioned by the overwhelming intellectual stature of its most perceptive analyst, the late Sir Moses Finley.
In my paper intend to argue that such scepticism has been unwarranted, once we take the vast quantities of new archaeological data into account. For a few centuries before and after the beginning of the modern era, the Roman Empire witnessed a remarkable growth of both population and standard of living, unprecedented levels of urbanization and truly urban life, more trade and connectivity than ever before and for a long time after, and favourable institutions.
I also want to argue that this prosperity was not the product of exploitation of the periphery by the imperial core, but was a genuine economic growth that extended to the provinces and benefiting large numbers of people.

Simulating the Roman farm: Investigating the limits and possibilities of the agricultural economy in the Roman Dutch limes zone via agent-based modelling

Jamie Joyce, Vrije Universiteit Amsterdam
This paper presents the final results of an agent-based modelling approach to investigate the limitations on the agricultural economy of rural settlements in the Dutch limes zone between the late Iron Age and Middle Roman periods.
This research was undertaken under the remit of Finding the limits of the limes project specifically to understand the possibilities of agricultural production under the natural and cultural conditions present in the region in the Roman period. Specifically, we have used a spatial dynamic approach to reveal how the availability of land and labour affected agricultural productivity of farmers and possibilities of surplus production.
We have developed an agent-based model using NetLogo to analyse the principle determinants and limitations on subsistence and surplus farming in the region. The model simulates in unison the major tasks within a mixed agricultural economy: arable farming, animal husbandry and collection of wood for fuel and construction. Initial experiments were conducted to analyse the demand placed on land and labour by settlements undertaking these tasks within a subsistence-based economy. Further scenarios were simulated in which diverse agricultural strategies were undertaken to evaluate the feasibility of these different approaches in producing surplus.
This paper presents a detailed description of the simulation model including assumptions, behavioural rules and primary characteristics before presenting the results of key scenarios simulated, including both scenarios used as heuristic devices and those based upon data derived from archaeological data. Subsequently, the principle findings of this ABM approach to modelling the ancient rural economy of the Dutch limes zone will be shown and, in particular, how these findings impact on the current state of knowledge. Finally, the usefulness and limitations of our approach to the investigation of a complex system such as agriculture will be offered.


The economy of Roman wine. Productive landscapes, archaeological data, quantification and modelling. Case Study Research: “Regio Laeetana-Hispania Citerior Tarraconensis” (1st century BC-3th century AD) Antoni Martin Oliveras, University of Barcelona
The study of the Roman viticulture economy has multiple fields of knowledge and expertise with enormous possibilities for research. Most studies have in common to use the archaeological information and the written sources as a complementary support to confirm the absolute chronology of a settlement, a socio-economic phenomenon or an exact location of a wine production centre or a pottery activity in a specific territory.
Regional variability is a key point for understanding the changing patterns of rural settlement and its evolution as the specific interaction between intra-regional and extra-regional economic networks. These studies can be conducted by geospatial and geoeconomic analyses in different territorial levels: macroespacial (regio) mesoespacial (territorium) and microespacial (torcularium atque figlina).
The level of dependence of the rural population in the regional market, respect for local urban centers and their subsequent screening in foreign markets (in our case study research: Western Europe, Italian Peninsula and Rome itself), responding to a series of socioeconomic patterns and behaviors that may be modeled and studied by different economic and econometrical ways.
The extensive use of mathematical models, statistical and linear programming to analyze, interpret and make predictions/regressions and reconstructions on the evolution of those complex systems, regarding, inter alia, different variables as the potential production of a region or territory, the regional consumption, the surplus production that could be traded in foreign markets, and other variables such as the sale prices, the market reactions, the production and transport costs, and the trends of consumption, is an increasingly widespread reality.
This paper present a PhD Research Project that try to analyse in four scenarios, the answers to this questions and the evolution of this complex economic system, related with the production processes, the long-distance trade and the consumption of Laeetanian wine in the Roman period, between the 1st century BC and the 3th century AD.

Simulating land-use for the Lower Rhine-Meuse delta in the Roman period

Maurice de Kleijn, Vrije Universiteit Amsterdam
This paper presents a methodological framework for understanding and simulating past land-use developments. As a case study, the framework is applied for the Roman period in the Lower Rhine-Meuse delta region. The arrival of Roman soldiers in the Lower Rhine delta in the first century AD brought great change for the local inhabitants and their surrounding landscape. First, Roman armies were (partially) reliant on the local population to produce cereals and meat for them. Second, the fortification structures required timber for construction and maintenance. Additionally, wood was required as fuel. This research simulates the impact of these changes and tests various scenarios on the demand for various land-use types by applying a quantitative modelling approach.
Farmers and Forts in Moesia Inferior: Modelling agricultural strategies on the Lower Danubian Frontier Eli Weaverdyck, University of California, Berkeley

This project analyzes changing rural settlement patterns in the Lower Danubian Plain from the Late Iron Age through Late Antiquity in order to elucidate the role played by garrison settlements in the economic strategies of peasants living near the Roman frontier. Using GIS, univariate statistical analysis and multivariate equation-based modeling, I interrogate rural landscapes for evidence of agricultural intensification and ancient marketing systems.
Settlements are the most durable traces left behind by rural, non-elites and the landscapes around settlements fundamentally shaped the opportunities and constraints these people faced. By identifying the features that attracted and repelled ancient settlers we can begin to understand the conditions under which settlement decisions were made. Using data collected from published and semi-published archaeological investigations in northern Bulgaria, I compare the productive potential of settlement territories to randomly distributed territories of the same size. I investigate the rural marketing through the concept of market potential – a measure of the total accessibility of marketing opportunities from any given location. I create market potential variables that include forts, exclude them, and include them as a negative influence. I then create a series of logistic regression models, varying only the market potential variables used. By comparing the goodness of fit of these models with one that excludes market potential altogether I can determine whether marketing opportunities influenced rural settlement and, if they did, which variable best approximates the ancient marketing system.
I come to the following conclusions: 1) the security provided by the Roman frontier allowed rural agriculturalists to settle in the most productive locations within the Danubian plain while the collapse of Imperial control at the end of the fourth century drove settlers to occupy less productive, but more defensible landscapes; and 2) Roman peasants sold their produce in small towns and cities, but not military bases.


From population dynamics to settlement patterns. Linking archaeological data to demographic models of the Dutch limes

Philip Verhagen, Vrije Universiteit Amsterdam

Even while the Dutch river area is among one of the most intensively researched regions in the world, when it comes to estimating its population in the Roman period, the available evidence is difficult to interpret. Previous approaches have departed from either estimates on the basis of settlement densities, or on idealized models of army recruitment and the size of the occupational forces and urban settlement. The demographic processes themselves however have hardly been the subject of study, leading to a largely unexplained model of substantial population increase in the late 1st and early 2nd century, followed by decline in the 3rd century AD.
We have tried to tackle this from two different angles: firstly, we have created a generalized simulation model of household demographics, taking into account historical evidence on mortality and assumed army recruitment rates; and secondly, we have analyzed the archaeological data for chronological accuracy, in order to better understand the population development trends. The simulation models point to various scenarios that can influence long- and short-term population dynamics, in which the biological factors of mortality and fertility are not the only drivers. On the other hand, the archaeological data suggests that the earlier postulated patterns of population growth and decline are real, at least as far as it concerns the observed presence of settlement. Some features in the data may obscure more subtle patterns, such as the abundance of chronologically undistinctive ‘native’ pottery, the lack of well-dated metal finds in many surveys and the scarcity of Late Roman finds that may not just be related to a lower number of people living in the area, but also to reduced material wealth.
Even so, it seems that the observed increases and decreases in settlement density are difficult to explain by natural demographic processes, and hence simulation modelling can be used to further explore possible scenarios.

Modelling evidence densities: past population variation or modern structuring affordances Chris Green, University of Oxford
This paper is based upon the work of the English Landscapes and Identities Project (EngLaId) at the University of Oxford (2011-2016; ERC funded), covering material recently published as Green et al. 2016. EngLaId was a project that took a long-term approach to understanding the archaeology of the English landscape from the Middle Bronze Age (1500 BC) to the Domesday survey of AD 1086. It was a legacy data project that collated a massive amount of material from existing repositories and then used that as the basis for modelling continuities and differences across England across c. 2,500 years.
In particular, this paper will discuss differences in relative densities of evidence between (the modern country of) England in the Roman period (AD 43 – 410/1) as against the preceding Iron Age (c. 800 BC – AD 42) and succeeding early medieval periods (AD 411 – 1065). In particular, it will question the degree to which these differences (and similarities) are due to differences in past population density or due to differences in both modern recovery practices and past usage of archaeologically durable material culture. Inevitably, the answer is a little bit of both.
Green, C., C. Gosden, A. Cooper, T. Franconi, L. ten Harkel, Z. Kamash & A. Lowerre. 2016. “Understanding the Spatial Patterning of English Archaeology: Modelling Mass Data, 1500 BC to AD 1086.” Archaeological Journal [online first]
A different vision of ancient settlement dynamics: creation and application of a model of evolution of the Antique habitat of the Plateau Lorrain Antonin Nüsslein, École Pratique des Hautes Études, Paris

Research concerning settlement dynamics is mainly based on data from archaeological field survey. This method of investigation provides researchers with a lot of information that can help to identify trends and to model the evolution of settlement structure at different scales.
Nonetheless, field survey data, even when dated, is incomplete and only shows a snapshot of the settlements. This static information lacks a certain number of parameters which are essential to perceive the inherent evolution of the habitats and therefore to visualize their own evolution within the dynamics of the settlement trajectory networks. On the other hand, data from archaeological excavations enable us to detect those phenomena.
This paper aims to propose a methodological approach to try to resolve this lack of parameters : the creation of an evolutionary model of the habitats from the information collected during excavations. Applied to the sites discovered by prospection, and combined with other analytical tools, the model allows for a better understanding of the diversification phenomena of the habitat and the modalities of spatial development of the settlement.
For now, this tool that offers solutions to balance the static approach provided by surveys data, was designed for the study of Antique settlement of the Plateau Lorrain, but the concept can be applied to other periods and to other regions as well.

Integrating detection and modelling of ancient pathways: towards a spatio-temporal, multiscalar approach

Philip Verhagen (Vrije Universiteit Amsterdam) & Laure Nuninger (Chrono-Environnement UMR 6249, Université de Bourgogne-Franche Comté)
Pathway detection and modelling operate from different perspectives that seem to have little in common apart from using the same base data. Material traces of pathways can be detected from LiDAR and fieldwork, but they are usually fragmented and are difficult to date, interpret and connect to other elements in the landscape. Path modelling on the other hand is usually hindered by a severe lack of material evidence to test models of ancient movement, and therefore often remains speculative.
In order to better understand past movement and the development of ancient pathways, it is necessary to consider all the relevant spatial and historical dimensions of movement, the connectivity of the resulting networks, and the material traces left behind. We suggest that such an analysis should start by considering the full historical trajectory of pathways, from their construction and associated practices up to their development during their use and after their abandonment. Such a spatio-temporal, multi-scalar perspective on ancient pathways is crucial to enhance our ability to detect and interpret structures, and to link the detected elements to their modelled counterparts.
In this paper, we will present a state of the art and develop some ideas to build a framework for such an approach.

Diverse movement in a dynamic environment: modelling local transport in the Dutch part of the Roman limes Mark Groenhuijzen, Vrije Universiteit Amsterdam
This paper presents part of the research on local transport in the Dutch part of the Roman limes that was carried out in the context of the ‘Finding the limits of the limes’ project. Research on transport has traditionally focussed on long-distance connections, often related to historical and archaeological evidence. We know very little of local scale transport, perhaps precisely because of the lack of material to work with. Least-cost path models have often been applied in archaeology to bridge this gap, for example to investigate possible routes or to look into the factors that have governed movement in the past.
Reconstructing transport movements through least-cost path models is not straightforward: there are many dynamic parameters that change the way in which local transport materialises. The role of the natural environment must firstly be considered, particularly in the dynamic river landscape of the Dutch Roman limes. Secondly, transport can utilise multiple modes such as regular walking, ox-cart transport and water-based transport, and a single movement can potentially use more than one mode of transportation. Finally, there are many different functions available in path modelling as well as a distinction between energy-based and time-based models, each having its own applications and assumptions. How these aspects are considered and integrated in our research will be the core of this paper. The next step is then to move from reconstructed transport paths to transport networks, which will be briefly alluded to but be dealt with in greater detail in another paper.

Postdicting Roman Roads in the NW Iberian Peninsula César Parcero-Oubiña , Alejandro Güimil-Fariña, João Fonte (INCIPIT – CSIC, Santiago de Compostela) and José Manuel Costa García (University of Santiago de Compostela)

GIS-based digital modelling tools, such as the well-known least cost paths (LCP), have been widely used in archaeology in the recent years as ways to approach the forms of mobility in the past. Quite frequently they have been used as tools to reconstruct (predict) the routes of ancient paths and roads, especially in prehistoric contexts. However they can also be used within a “postdictive” perspective where the aim is not to reconstruct lost parts of evidence (the missing route of a road) but rather to find a logic in the pieces of evidence we actually know about (the archaeological record). This approach is especially useful when we have a significant amount of evidence available, in highly formal archaeological contexts.
This is usually the case of the Roman roads, as in the NW Iberian Peninsula. Here, in the Roman provice of Gallaecia, a number of scholars have contributed extensively to the reconstruction of the likely routes of the main Roman roads in the last decades. In this paper we will show a complementary approach that, based on a LCP analysis, tries to tackle the question of why the roads follow those particular routes. By doing so, we will try to understand the influence of topographic factors in the road network, and the role of some particular sites in the social, administrative and political organization of the territory.

Walking Between Gods and Mortals: reconsidering the movement of Roman religious processions Katherine Crawford, University of Southampton
Movement was a central component of Roman religious processions. While scholars acknowledge the regularity of processional activity, studies have traditionally focused on major processions such as the Roman triumph rather than considering the wider occurrence of religious processions. The nuances of how a procession traversed a city’s streets and its urban impact are not easily revealed due to the paucity of information from ancient literary sources. As a transient event with their occurrence largely held in the memory of those who attended or heard about them, processions leave little standing trace within the archaeological record. This paper applies a network analysis and agent-based modelling approach to consider new ways of studying the movement patterns of religious processions at Ostia, Rome’s ancient port. Focus is transitioned from trying to reconstruct routes based off of a fragmented archaeological and historical record to questioning the ways in which a cityscape affected a procession’s direction. This approach considers processions as dynamic events that engaged with both the people and built environment of a cityscape. By re-examining approaches to studying processional movement, we can gain new insight into how religious movement shaped social and religious interactions within the ancient city.

Network science in Roman studies: the potential and challenges Tom Brughmans, University of Oxford
As far as the application of network science in Roman studies is concerned, the low-hanging fruit has almost all been picked. We have created networks of the Antonine itineraries and of Cicero’s letters, and many more networks of large datasets of Roman material culture. Network science has recently become more commonly applied in archaeology and Roman studies in particular have seen a diverse and creative use of network science techniques, thanks to the availability of both written and material sources as well as long-running debates on theoretical models. Roman studies now face the challenge, not of copying network science techniques from sociology and physics, but to develop original network science models and methods that explicitly address its key research themes. Such work should take the unique strength of network science as its starting point: the formal representation and study of (theories about) relational phenomena and data. In this presentation I will review the use of network science in Roman studies and make some suggestions of how the potential of network science can be explored in the future. The review reveals theoretical and methodological challenges that Roman studies share with archaeology and history more broadly in their application of network science. How to deal with material data in network approaches? How to explore long-term change in social systems with network methods that have traditionally been applied to study more short-term changes? How to address sampling issues and missing data? Addressing these challenges and thinking through the relational phenomena underlying key research themes in Roman studies will reveal ways forward.

Possibilities and challenges in the use of networks to study socio-economic relations in the Dutch part of the Roman limes Mark Groenhuijzen, Vrije Universiteit Amsterdam

Network theory and network analysis have become integral components in our research on the cultural landscape of the Dutch part of the Roman limes, particularly concerning local-scale transport. The application of network analysis on reconstructed transport networks has provided avenues for approaching both old and new archaeological questions, yet there are many considerations and reservations that have to be made in the process. Both sides of the coin will feature in this paper.
The first part will present the application of formal network analysis on reconstructed transport networks, and what we can interpret from the results, for example related to existing archaeological hypotheses on the socio-economic structure of the region. Using the extensive site dataset and chronological information available, networks of various modes of transport can be compared, the development of transport networks through time can be studied, and the role of individual sites within the network, ranging from rural settlements to Roman castella, can be examined. The second part of the paper will present some challenges that were encountered and considered during our research using network analysis to study transport networks. Firstly, a number of network construction techniques and their merits for building transport networks will be discussed. Secondly, the stability of networks and the robustness of network analysis results under the influence of uncertain information will be presented, showing that a significant part of the results are unreliable, and how that can be used in the interpretation of the network. Finally, the application of an alternative approach to chronological uncertainty in the archaeological site dataset and how that changes the networks is shown.
Network analysis to model and analyse the Roman transport and mobility Pau de Soto, Universidade Nova de Lisboa

The analysis of Roman infrastructures to understand the transport costs and the commercial routes and processes is an indispensable way to know the benefits and shortcomings of the transportation system created in Roman times. It is well known that the Roman Empire built the first big transport network in Western Europe and also in parts of Eastern Europe, the Middle East and Northern Africa. This overwhelming task included not only the construction of roads with their correspondent bridges, but also the building of river ports and maritime harbours. Such a huge effort aimed to create an integrated economy covering all the Roman provinces on the Mediterranean Sea and Atlantic Ocean.
In the last years, we have attempted to reconstruct the Roman transport conditions by modelling travel costs and times with the help of GIS and Network Analysis applications. The main geographical focus of this project was the NE of Hispania. It was necessary devote a significant effort to the gathering, documentation, analysis and digitisation of Roman communications with high precision. With the aim of using these methodology in a much broader geographic frame, the entire Iberian Peninsula and other European territories were analysed with less detailed transport networks. It allowed us to discover very interesting patterns. The results of such applications provide us with new information to understand the distribution of commodities, product competition and problems of stagnation in ancient economies such as that of Ancient Rome.
The ability to see graphically and numeral those costs values which until now they could only be guessed, can open new perspectives and justifications to the speeches made on the work done until today. In fact, the comparison between these results and the analysis of archaeological and historical interpretations should complement each other, clarifying and offering more elements for a global vision.
At the same time, it has been possible to observe how the construction of a complex communication network, especially based on the creation of land routes, meant an important element for the integration of new territories to the Roman provincial model. To understand the morphology of these networks, we have applied some Social Network Analyses to understand the configuration and performance of the Roman mobility in these territories.
The same methodology can be used to analyse other historical periods and therefore to elaborate comparisons of the same territories along the centuries. This is the main focus of our current project, named Mercator-e and will allow us to explore the transport network evolution of the Iberian peninsula from Roman times to the XIXth Century.

Multi-scalar approach to long-term dynamics, spatial relations and economic networks of the Roman secondary settlements in Italy: towards a model? Angelo Castrorao Barba (University of Palermo), Stefano Bertoldi (University of Pisa), Gabriele Castiglia (Pontifical Institute of Christian Archaeology)
In Roman landscapes particular sites defined as secondary settlements, (known also as vici/villages, minor centers, agglomération secondaire, stationes/mansiones) have played an ‘intermediary’ role between the cities and the other rural structures (villae/farms), linked to the medium- and long-distance economic and commercial trajectories. The aim of this paper is to apply a multi-scalar approach to modelling the long-term spatial relations and the connectivity with the Mediterranean exchange network of the Roman secondary settlements. On a macro-scale we have analyzed a sample of 219 edited sites to understand the diachronic trends and the spatial dynamics of attraction / proximity to the significant elements of the landscape as towns, roads, rivers and coastline. The Ombrone Valley (Tuscany – Italy) represented a micro-scale case study of a complex systems in which the imported pottery (amphorae, ARS, ingobbiata di rosso) found in the vicus/mansio of Santa Cristina in Caio, the Roman villa of La Befa and the town of Siena (Saena Iulia) held diagnostic ‘macro-economic’ perspectives. The results showed how the secondary settlements played a nodal position in the Roman landscape in terms of resilience (long period of occupation until the Early Middle Ages), spatial organization with a close relationship to natural and anthropic infrastructures and trade functions linked to the Mediterranean routes.


Posted in: Uncategorized