(detailed programme with abstracts below)
August 24th 2017
|10-10.15am||Registration and coffee|
|10.15-10.30am||Welcome: Tom Brughmans, Anna Collar, Fiona Coward|
|Session 1: Ego-networks and individuals|
|10.30-11am||Matthew Hammond & Cornell Jackson||Analysis of Ego-Networks of Witnesses to Medieval Charters: Opportunities and Challenges|
|11-11.30am||Lena Tambs||Mixing Emic and Etic Perspectives for Studying Small-Scale Communities of the Past: The Case of Pathyris (c. 165-88 BCE)|
|11.30-12pm||Esther Lewis||Was Bristol a ‘hotbed of heresy’?: a discussion of the use of networks in the pursuit of medieval heresy.|
|Session 2: Accounting for temporal change|
|1-1.30pm||Ray Rivers & Tim Evans||Is machine time historical time?|
|1.30-2pm||Elisa Grandi||Social networks and path dependency in the international economic advising. The World Bank first mission in Colombia and its aftermaths (1949-1954)|
|2-2.30pm||Cécile Rivals||The modeling of urban spatial dynamics in long time spans: The use of graph theory in Saint-Antonin- Noble-Val (Tarn-et-Garonne, France) from the 14th to the 19th centuries|
|Session 3: Networks and Geographical Space|
|3-3.30pm||Denis Hakszer||Networks vs. Ideology: Network transformation in Caria and the Dodecanese during the Classical Period|
|3.30-4pm||Katherine Crawford||Movement within the Cityscape: the potential of applying network methods|
|4-4.30pm||Mark Groenhuijzen & Philip Verhagen||(Finding the Limits of) Network Approaches to Local Transport in the Dutch Part of the Roman Limes|
|4.30-5.30pm||Keynote: Eleftheria Paliou|
August 25th 2017
|Session 4: Categorisations and Material Similarity Networks|
|9-9.30am||Beatrijs De Groot||Networks of ceramic assemblage similarity and the Neolithisation of South-East Europe|
|9.30-10am||Clara Filet, Fabrice Rossi & Aurélia Feugnet||Co-presence Analysis and Significance Scores:
when network studies highlight choices of Iron Age elites
|10-10.30am||Fiona Coward||‘All things being equal?’ Material networks of the early Neolithic in the Near East|
|Session 5: Methodological Decision making and the consequences of research design|
|11-11.30am||Christine Johnston||Choice and Consequence in Research Design: Ceramic Networks and the Study of Political Economies|
|11.30-12pm||Lennart Linde||Networks, Agents and Interactions: How to use Agent Based Models to Investigate Archaeological Network Formation|
|Session 6: Modelling movement and transportation|
|1.30-2pm||Ignacio Morer, Luce Prignano, Francesca Fulminante, Sergi Lozano & Albert Diaz-Guilera
|Modelling transportation networks in protohistoric scenarios: assessing the impact of incomplete datasets|
|2-2.30pm||Christina Williamson & Onno van Nijf||Whose network? The complexities of multiscalar festival networks in the Graeco-Roman world|
|2.30-3pm||Carrie A. Fulton||Adding Shipwrecks into Maritime Networks: Shipping Meat in the Roman Mediterranean|
24th August 2017
10-10.15: Registration and coffee
Session 1: Ego Networks and Individuals
Analysis of Ego-Networks of Witnesses to Medieval Charters: Opportunities and Challenges
Dr Matthew Hammond (University of Glasgow, soon to be Kings College London) and Dr Cornell Jackson (University of Greenwich)
A Leverhulme Trust funded project running from 2013 to 2016 enabled extensive social network analysis of the People of Medieval Scotland 1093-1286 database, the most comprehensive extant prosopographical database of a medieval kingdom, covering over 6000 persons and institutions drawn from over 15,000 charters. SNA specialist Jackson and historian Hammond explored a number of methods for analysing the material, many of which are laid out in an extensive free e-book and a number of interactive online graphs (http://www.poms.ac.uk/social-network-analysis/). The results of most studies are fully available as excel spreadsheets. Approaches taken included a) familial, employment, and lordship relationships, b) grantors and beneficiaries, c) a broad range of studies examining people who witnessed documents together, and an experimental agent-based modelling study. Our presentation at this conference would focus on the ego-networks created from the broadest study of co-witnesses, with an eye to the challenges presented by the uneven and incomplete nature of the medieval evidence and some possible interpretative approaches to allow us to make use of the some of the potentially promising results, as laid out in Chapter 7 of the e-book (http://www.poms.ac.uk/e-books/social-network-analysis-and-the-people-of-medieval-scotland-1093-1286-poms-database/). Actors identified by various means as the most influential are shown to have very low ego-network densities; this significance of this is considered in light of great variations in ego-net size. This opens up a possible method for identifting brokers with high levels of bridging capital. The significance of this, however, is tied closely to an understanding of the documents which the actors has witnessed.
Mixing Emic and Etic Perspectives for Studying Small-Scale Communities of the Past: The Case of Pathyris (c. 165-88 BCE)
Lena Tambs, University of Cologne
Some 30 km South of Thebes (modern Luxor) in Egypt, the subsidiary military camp of Pathyris (modern Gebelein) was established sometime between 165 and 161 BCE. For about 75 years – or three generations – an unusually rich body of surviving written sources pertaining to the lives and affairs of the Pathyrites allows for the small-scale community they formed to be studied in detail.
The research project here presented, aims to study social and economic life in Ptolemaic Pathyris through a distinct network approach. It centres on its individual members and their interrelations as revealed by c. 450 texts associated with 21 ancient archives from the site. Particularly relevant here, is Social Network Analysis (SNA). By means of mapping a significant number of specific socio-economic relations in a systematic bottom-up approach, SNA can tackle the diversity of the source material and enable the relational data contained in the texts to be studied on an inter-archival, as well as an archive-by-archive basis.
SNA offers new perspectives and methodological tools for organizing, analyzing and interpreting attributed relational data. In order to make sense of evolving patterns and observations, their meaning must though be read in relation to the larger picture. The talk argues for a balanced relation between emic and etic perspectives by highlighting (1) the project’s applied method of attributing node and edge entries with categorizing labels and (2) subsequent contextualisation of the networks in time and place. In both cases, pros and cons of emic and etic approaches are critically discussed.
Was Bristol a ‘hotbed of heresy’?: a discussion of the use of networks in the pursuit of medieval heresy.
Esther Lewis, University of Nottingham
The debate surrounding heresy and lollardy in pre-Reformation Bristol is unresolved. Historians remain in a deadlock, with neither side being able to conclusively prove their own point or disprove the others’. Late medieval Bristol was divided between the dioceses of Bath and Wells and of Worcester. This split in jurisdiction is probably the reason that there is little evidence for heresy in Bristol despite its reputation among medieval contemporaries as a centre of dissent.
I have used social network theory and Gephi to take a new approach to this issue. I have used the relationships with religious institutions present in last wills and testaments to reconstruct the laity’s pious networks across the medieval town in the first quarter of the fifteenth century. The graphs show mainstream Catholic networks that bear a resemblance to the networks of dissenting activity glimpsed in the heresy trials. A closer analysis of the networks within the affected parishes shows the relationship between seemingly ‘orthodox’ parishioners and their dissenting neighbours.
This paper will discuss how networks have been used in relation to this specific historical debate, and go on to discuss this approach in relation to medieval history more widely. It will discuss the advantages and draw backs of using this methodology, and emphasise the need for a qualitive analysis alongside the use of networks.
Session 2: Accounting for Temporal Change
Is machine time historical time?
Ray Rivers and Tim Evans (Imperial College London)
Historical networks evolve for endogenous and exogenous reasons that are difficult to distinguish. To clarify the issue we pose the simpler question which concerns the way in which we code for network models: to what extent does machine time – the way in which the code is implemented chronologically – reflect the historical time evolution of the system under study? The relevance of this is that it forces us to consider the extent to which we encode, or evade, ideas of historical determinism, contingency, and even chaos.
At one (ontic) extreme the code assumes initial conditions appropriate to a specific historic past and takes evolution to a ‘good enough’ network as a proxy for the historical evolution from that past state. From this viewpoint endogenous algorithmic iteration in machine time is history in the making.
The other (epistemic) extreme says that the computer is doing no more than showing the relaxation in machine time from an arbitrarily chosen initial condition to the ‘more likely’ required answer and nothing societal is to be attributed to the nature of the path of relaxation, only the final state. That is, history is an attempt to maintain ‘good enough’ functionality under (manageable) global stress whereas, from the previous viewpoint, history is an attempt to achieve ‘good’ functionality from a non-optimal beginning. The complementary nature of the time-scales of these approaches and the extent to which they can be resolved, as exemplified by city-state formation, is the key ingredient of this talk.
Social networks and path dependency in the international economic advising. The World Bank first mission in Colombia and its aftermaths (1949-1954).
This paper applies social networks (SNA) analysis to the history of World Bank’s early development programs and focuses on World Bank General Survey mission in Colombia (1949), the first the Bank organized in the country. SNA is applied to analyze the interaction between international experts and Colombian business and political elites over time, to uncover the evolving strategies that oriented the adoptions of specific development programs by the World Bank, as well as by local actors.
This paper, part of a PhD thesis recently defended, assumes that the relational pattern among the international advisors and the local staff strongly influenced the way in which international economic advising evolved. Therefore, the main methodological challenge concerns how to analyse social networks through time and connect them with the evolution of the loans the World Bank disbursed in Colombia after this first mission. To observe the way in which networks evolve the different steps of the mission, we applied both the notion of relational chain, proposed by Michel Grossetti and the ego-centered networks analysis of some main actors involved in the mission. Indeed, ego-centered networks constitute a very powerful tool to catch the multiplexity of a relational network and they can be fruitfully applied to enlighten the complexity of the actors’ role in the networks. Nonetheless, it is difficult to transpose static ego-centered network to the dynamics of a process. This paper proposes to adopt the notion of chains relational chain to overcome this difficulty.
The modeling of urban spatial dynamics in long time spans: The use of graph theory in Saint-Antonin- Noble-Val (Tarn-et- Garonne, France) from the 14th to the 19th centuries
Cécile Rivals, Université Toulouse
The scarcity and heterogeneity of planimetric documents before the systematic realization of the Napoleonic Cadastre in the 19 th century is a handicap for the study of spatial dynamics. Fiscal sources, such as terrier and compoix, allow us to overcome these difficulties. These documents, considered ancestors to the Napoleonic Cadastre, appeared in France during the 13th century. Used until the 18th century, there are tens of thousands in municipal and departmental archives. They identify the lands of a territory in order to levy taxes. Parcels are described and placed in space by the use of microtoponyms and surrounding features. The lack of map reference complicates the utilization of terrier and compoix. The necessity of an approach which oversteps these difficulties led to an application of graph theory in order to modelling spatial information. The contiguity of plots is considered as links of a network. Thus, it is possible to use the topological attributes of the documents for modelling landscape with the help of graph theory. Each fiscal register and each plots plan on an area are modelled as a graph. The comparison of each graph, which reflects different states of the landscape, is based on links between parcels, to reproduce the dynamism of plots. This approach leads us to understand the evolution of urban or rural landscape and of the society which created it. The Connected Past 2017 could be an opportunity to present the treatment of spatial information contained in fiscal sources of a small medieval town in south of France.
Session 3: Networks and Geographical Space
Networks vs. Ideology: Network transformation in Caria and the Dodecanese during Classical Period
Denis Hakszer, Charles University, Prague
This paper considers networks analysis (using Cytoscape) as a way to explore the political and economic interactions in the area of Dodecanese archipelago and the coastal part of Caria during the Classical period (5. and 4. centuries BC). It builds on the close study of rich epigraphic and material evidence and seeks to illustrate how much and in which way did the political and exchange networks transform over time during rather significant shifts of ideology when individual city-states adhered to policies imposed by powers such as the Delian League or Persian administration. This diachronic analysis of network structure is based on spatial-temporal modelling within a GIS framework. To what extent did exchange networks follow patterns established by political networks which were primarily influenced by ideology? What additional insights could an approach to interaction such as this yield? What are its limitations?
Religious Movement within the Cityscape: the potential of applying network methods
Katherine Crawford, University of Southampton
Evidence of ritual movement within the ancient city is rarely discernible within the archaeological record. Transient events, like religious processions, were primarily held in the memories of those that attended, heard, or read about the ritual. While occasionally marked with commemorative architecture, these sources of evidence to do not provide insight into how processions were integrated within the ancient cityscape. Previous scholarship has relied on literary accounts and corresponding urban topography which has resulted in a connect-the-dots approach to determining possible processional routes. This study develops a new framework that questions how urban architecture and social activity structured ritual movement within the city of Ostia Antica, Rome’s ancient port. Urban space functionalities are used as a proxy for the impact on ritual movement; betweenness centrality is used to find specific zones of engagement within Ostia to denote areas were processional movement likely passed. Rather than trying to map specific routes, this approach considers processions as dynamic events that engaged with both the people and built environment of a cityscape. By examining the intra-site relationships between space and ritual activity, the ways in which religion was disseminated across the city can be addressed.
(Finding the Limits of) Network Approaches to Local Transport in the Dutch Part of the Roman Limes
Mark Groenhuijzen & Philip Verhagen, Vrije Universiteit Amsterdam
The aim of the ‘Finding the limits of the limes’ project is to study the cultural landscape of the Dutch limes zone, and in particular the spatial and economic relations between the local and Roman military populations. An important facet that structured this relationship is the assumed increasing frequency of local transport movements, brought about by the large requirements of the Roman army for food, wood and fuel, and the increasing potential of the local population to produce a surplus to at least partly provision the military population.
However, due to the lack of archaeological evidence, questions remain pertaining the organisation of local scale transport. For example, how were goods moved from the local to the military population? Was there a direct connection from source to sink, or a more intricate socio-economic structure wherein intermediary places were involved?
Since there is little archaeological evidence to rely on, the aim of this paper is to approach the aforementioned question through computational archaeological techniques, including spatial analysis and network science. More specifically, some methodological issues that arise are: what is a suitable network construction technique to model transport networks; what is the role of uncertainty in the archaeological site dataset; what is the significance of network analysis results? By examining hypotheses on the organisation of local scale transport in this way, we also aim to give an overview of our exploration on the possibilities, and the limits, of applying computational approaches to our archaeological questions.
4.30-5.30pm: KEYNOTE Eleftheria Paliou
5.30pm Wine reception
25th AUGUST 2017
Session 4: Categorisations and Material Similarity Networks
Networks of ceramic assemblage similarity and the Neolithisation of South-East Europe
Beatrijs de Groot, UCL, Institute of Archaeology
This presentation will discuss the results of a systematic study of ceramic assemblage similarities between Early Neolithic sites in western Anatolia and south eastern Europe. The aim of this study was to use network analyses to visualise intercultural relationships between the regions surrounding the Aegean Sea, during the period between ca. 6600-5600 BC, and through these means, identify the location and timing of social interactions and/or migration through this region.
Ceramic assemblage similarities were calculated on the basis of the number of shared ceramic attributes per site-phase, as derived from site-reports and publications. This presentation will focus on the methods used and the problems encountered during this study. How, for example, can we adequately make use of previously published materials and account for variation in their quality and research foci? Furthermore, how does the sample size of the ceramic assemblages influence the diversity of their attributes, and how does this diversity affect the relationships observed?
Finally, this presentation will discuss what the networks and relationships this analysis produced actually represent; are the network patterns indeed representative of ancestral relationships between sites, or can different mechanisms (copying, innovation) or the biases discussed cause similar network patterns? In conclusion, this presentation will showcase how a detailed consideration and comparison of ceramic assemblage similarities can help shed light on the processes of interaction and migration that underpin the spread of farming from Anatolia to Europe.
Co-presence Analysis and Significance Scores: when network studies highlight choices of Iron Age elites.
Clara Filet, Fabrice Rossi & Aurélia Feugnet, Université Paris Panthéon-Sorbonne
During the last decades, the discovery of thousands of items imported from the Mediterranean area to Celtic Europe testifies a major economic and political upheaval that occurred at the end of the Iron Age (IIIth – Ist c. BC). The inventory 57 735 of them is used to identify and characterise distribution structures (regional selections, main axes, etc.).
From a 2-mode network formalisation of the data table (objects – sites), we use network analysis to detect patterns in the association of imported artefacts that are significantly often found together. For this purpose, the significance of those recurrent associations is evaluated via a specific statistical score, used to identify the part of random situations and highlight the part which may result from intentional associations. The objectives are to highlight groups of imports that may have circulated together and to emphasize regional selections by local populations. At this stage, two main systems of imports have been highlighted, respectively, in West and Central Europe, materializing, a socio-cultural frontier inside the Celtic world. Within those two systems, the heterogeneity of associations systems of artefacts from a region to another emphasizes important regional particularisms: some areas received/selected certain types of goods that their neighbours did not received or wanted to receive.
‘All things being equal?’ Material networks of the early Neolithic in the Near East
Fiona Coward, Bournemouth University
Archaeological network research typically relies on material culture as a proxy for past social networks. In many cases, a range of different types of material culture are subsumed into reconstructed connections between sites/populations/nodes. However, not all kinds of material culture are equivalent, and different types of objects may be caught up in rather different forms of social relationship and thus circulate among individuals/groups in different ways. For example, ethnographically ‘personal’ items such as jewellery may perhaps have more social and cultural significance and thus be traded, gifted or exchanged with more care than commodified goods such as functional tools such as awls or scrapers. Thus analyses which include many different types of material culture may in fact be conflating a range of different kinds of social relationship and patterns of connection. However, in many societies the dividing line – if there is one – between significant and functional is very fluid, making it impossible to categorise items definitively without understanding the specific cultural context. Using data from the early Neolithic of the Near East this paper will investigate whether and how different forms of material culture produce different kinds of social network and the potential ramifications of this both for archaeological network research, and for understanding past use of material culture.
Session 5: Methodological Decision making and the consequences of research design
Choice and Consequence in Research Design: Ceramic Networks and the Study of Political Economies
Christine Johnston, University of California, Los Angeles
This paper explores the use of network analysis for the quantitative assessment of trade systems and the profiling of institutional structures in incorporated polities. This study focuses on trade in the Eastern Mediterranean during the Late Bronze Age (ca. 1700 to 1200 BCE), and includes a network analysis of Cypriot and Mycenaean pottery circulated throughout Egypt, Cyprus, and the Levant. The ceramic trade networks of Egypt and the Levant are contrasted to examine the correlation between network measures and the presumed political structures extant in these two regions during this period. These profiles then form the control groups for evaluating import networks in Cyprus to elucidate Late Cypriote political organization, which remains highly contested.
The analysis of ceramic distribution demonstrates a high degree of variability across the regions of study, including differentiation in the relative popularity of open dining vessels and closed containers, the diffusion of vessel circulation, the concentration of large assemblages and rare forms at higher-order sites, and the centralization of distributional systems around political centers. These contrasting network attributes signify regional differences in demand and consumption, while overall network centralization and density measures indicate diverging mechanisms for import circulation. Various network iterations are produced according to different taxonomic systems—chronological, ware type, and form function—in order to explore the impact of analytical categorization on network measures. Similarly, the impact of regional boundary setting in research design is examined. Results of this analysis demonstrate the efficacy of network methods for the examination of trade and political economies.
Networks, Agents and Interactions: How to use Agent Based Models to Investigate Archaeological Network Formation
Lennart Linde, Goethe-University Frankfurt
During recent years the terms “Network Analysis” and “Agent Based Model” (ABM) have both become huge buzzwords within the archaeological community. But beyond their individual hype cycle both approaches to archaeological data display a vital potential. This presentation will try to illustrate how they can complement each other in the ongoing quest to make sense of our data.
Archaeological networks are reconstructed from broken links and dysfunctional nodes through various material proxies. While we came up with more and more ways to deal with the emergent uncertainties of this approach, our interpretation of the resulting graphs is still strictly narrative. Wedenominate certain interactions to be traceable through our chosen proxies. But are they?
This is where ABMs can become a promising addition to the network analysts toolbox. ABMs allow us to generate virtual networks based on our assumptions about the character of interactions we try to trace. They can serve as computational laboratories through which we can explore the formation of our archaeological networks under certain assumptions and preconditions.
This talk will give a (very) short introduction to the basic structures of ABMs and how they can be utilized to monitor our implications on the genesis of our networks. The second part will give a small case study on exploratory analysis of network formation.
(To be confirmed)
Session 6: Modelling movement and transportation
Modelling transportation networks in protohistoric scenarios: assessing the impact of incomplete datasets
Ignacio Morer1, Luce Prignano1, Francesca Fulminante2, Sergi Lozano3 & Albert Diaz-Guilera1
1 Universitat de Barcelona
2 University of Cambridge
3 Institut Catalá de Paleoecologia Humana i Evolució Social – IPHES
The period between the Late Bronze Age and the Archaic Age is a time of change and development in the Italian Peninsula, which led to the creation of regional ethnic and political groups and to the formation of the first city-states. In previous works [1,2] we studied the transportation networks of two neighboring regions within a network modeling framework, aiming to explore the dynamics between city-states: We designed network models corresponding to three competing hypotheses about the dominant mechanism underlying the creation of new connections, and then compared model outcomes and empirical networks by looking at global topological measures.
However, these cases represent the rather unique situation in archaeology in which a complete dataset is available. Fragmentation is a typical trait of archaeological records, and it is crucial to determine whether our approach is useful when facing a more common scenario of partial information. Specifically, we must assess its impact on model validation, which critically depends on the completeness of the record.
Thus, to explore the potential limitations of this methodology, we artificially damaged our empirical networks to resemble incomplete systems. Our main goal is to understand how the type and the amount of damage affects the process of network modelling, with special attention to the key phase of model validation.
 Prignano, L., Morer, I., Fulminante, F., and Lozano, S. (2016). Modelling terrestrial route networks to understand inter-polity interactions. A case-study from Southern Etruria. arXiv:1612.09321 [physics.soc-ph]
 Fulminante, F., Prignano, L., Morer, I., and Lozano, S. (2017). Coordinated decisions and unbalanced power. How latin cities shaped their terrestrial transportation network. Front. Digit. Humanit. 4:4. doi:10.3389/fdigh.2017.00004
Whose network? The complexities of multiscalar festival networks in the Graeco-Roman world
Christina Williamson, University of Groningen
Panhellenic games in antiquity brought the ancient Hellenic population together; major festivals as at Olympia and Delphi not only provided a means for rivalry, but also the performance of the global community and the reiteration of shared values. In the Hellenistic period however cities across the Greek world began organizing festivals at their own sanctuaries with Panhellenic aspirations. This became a common means of self-promotion, but also one of the driving factors behind an increasingly globalizing network. By the imperial period there was a virtual ‘agonistic explosion’, as Louis Robert observed, and we believe this to be one of the fundamental ties in the network that facilitated the rule of Rome in the Greek east.
In our project, Connected Contests, we investigate this phenomenon at various scales, interpreting network dynamics at global, regional and local levels. In the first place our focus lies on prosopographic data drawn from epigraphy, but successive phases will include spatial and material culture as well. Clearly these festivals created networks at different resolutions, as we see in the social, civic, interstate, and oikoumene and imperial spheres. But how can we go beyond ‘network thinking’ and integrate inherently different types of sources, of varying quality, to fully analyze the complexities of multiscalar festival networks? This paper discusses some of the issues that we have encountered in our project with tentative suggestions for ways forward.
Adding Shipwrecks into Maritime Networks: Shipping Meat in the Roman Mediterranean
Carrie A. Fulton, University of Toronto
This paper examines the challenges that arise from using mobile assemblages within formal network analyses. That is a shipwreck as a networked assemblage. In discussions of maritime networks, shipwrecks are often omitted altogether (e.g. papers in Ducruet et al 2016)
However, I introduce several elements.
1) There is a chronological component to interpreting objects within the shipwreck that I argue can best be understood by examining the chaîne opératoire of the items onboard, and more broadly stated the chaîne opératoire of the cargo.
2) How do we integrate an actual physical geographical element into the description of maritime networks?
3) For maritime networks can we use the emphasis on the physical mobility of objects to examine the cultural or social significance of mobility?
4) Are shipwrecks the nodes in our system or are they physical evidence for ties?
This paper thus moves past the divide of trying to describe cargoes as either homogenous or heterogeneous and also the pitfall of trying to signal out one object on board the ship as the main way to compare ships. Instead, through network analysis, I present a way of moving beyond this distinction and analyzing the commonalities and differences in assemblages. Such an approach also moves beyond helping us to classify certain items as either cargo or personal items.