In a 1980 article, Colin Renfrew lamented how, in the decades following the 1950s radiocarbon revolution a “Great Divide” had formed between anthropological archaeology and classical archaeology. The former, since the 1950s, had been moving towards quantitative and statistical methodologies alongside explicit theory to generate new ideas about the processes of human development. Classical archaeology, however, remained rooted in the classification and publication of immense datasets from the “Big Digs” run primarily by foreign schools in Athens since the second half of the 19th century. Renfrew emphasized what classical archaeology could offer to a more integrated discipline, namely its long tradition of meticulous documentation stemming from these Big Digs, an enormous potential for processualist archaeologists.[1] He presciently suggested that “while we need not be mathematicians ourselves, can we ignore the potential of mathematical thinking in a discipline such as ours which inevitably requires the handling of large quantities of data, which have to be ordered and sorted into a pattern so that we can hear their ‘Voice of Silence’?” (1980: 294).
We are now in the swing of Kristian Kristansen’s Third Science Revolution (Kristiansen 2014),[2] where the main driver is “Big Data” and computational techniques used to analyze and visualize this data. Michael Loy’s book is a product of this new era, which brings Renfrew’s insights to fruition (although Loy does not specifically cite Renfrew’s article or Kristiansen’s). Indeed, the first pages evoke the era of the Big Digs, noting, like Renfrew, the focus of these projects on generating painstaking catalogues of data—a focus that, for Loy, raises a broader question: “why investigate at a ‘big’ scale if not to ask ‘big’ questions?” (4). Loy proposes to use this legacy data to ask bigger questions about behavioural patterns reflected in these archaeological remains, specifically: “How formally organized were the political communities of early Greece, and what role within these communities was played by elite peers?” and “To what extent can methods of modern economics be appropriately used to analyse the ancient economy?” (28). The datasets for this analysis involve the material remains (sculpture, pottery, coins, and inscriptions) of communities around the Aegean between 700 and 500 BCE. They are drawn from a range of sources, including original excavation reports from these Big Digs, secondary catalogues, and online databases.
The Introduction (Chapter 1) lays out the rationale for answering these questions. The historical questions above, on political formations and economic structures, are less foregrounded than the quantification and analysis applied to the material remains. Introductory discussions include consideration of the methodological problems of employing datasets from Big Digs, including variations in recording, terminology, and completeness, among others, that inevitably plague synthetic studies. Loy also introduces the main framework, Social Network Analysis (SNA), used here “as a heuristic for sorting groups of similar materials together according to units of place around the Aegean basin” (13). Material affiliations between sites are thus understood as proxies for shared activities: “it is those similar practices undertaken in the same way at the same places that are understood to represent the historically meaningful building blocks of wider networks. Nothing more and nothing less.” (13) Throughout this work, Loy combines SNA with other methods, particularly spatial analyses. In this sense, the methodology is roughly in line with previous archaeological studies that combine spatial and social network analyses (e.g., Mills et al. 2013). The methodology is thus not novel (the SNA is employed primarily for identifying inter-community affiliations), but the amalgamation of legacy data from decades of fieldwork using SNA visualizations combined with spatial and other exploratory data analyses offers exciting potential for asking new questions of old data. The basic unit of analysis within these frameworks is communities (primarily city-states), and particularly the elite actors that drove the formation of these networks.
The body of the book consists of five chapters and a conclusion. Readers will find a real benefit in reading these chapters in succession, as each one builds on preceding results. Each chapter uses a combination of models to lay out the patterning and distribution of material data, and concludes with a discussion of the political and/or economic behaviours that might best account for such patterns. Chapter 2 leads with an analysis of the transportation patterns of marble sculpture used to make kouroi and korai statues. Loy aggregates all known data points of complete and fragmentary marble statues (305 kouroi and 174 korai, although this dataset is later reduced to 207 pieces with secure provenance) using Richter’s catalogues of kouroi (1942) and korai (1968), updated with more recent studies. Major data points include estimating the tonnage of each marble piece to arrive at the total volume of marble used to sculpt the surviving corpus,[3] along with the origin points of the marble and the ultimate resting places of the statues to relate them to major shipping routes across the Aegean in the Archaic period. Employing Proximal Point Analysis (PPA) combined with the origins and end points of these marble statues, Loy identifies four main corridors of movement for this marble, modeled as a map visualization of routes travelled between marble quarry and destination coloured by the relative weights of marble moving along them. The product is not the typical node-link graph of SNA, but the combination with PPA is ultimately more realistic, as node-link graphs elide the varied transport routes an object can take from source to destination, while the combination of source node and target node with PPA allows at least a hypothetical evaluation of how raw material moved across the Aegean. Loy concludes by emphasizing an elite-driven, socially embedded economy operating at a significant size and scale through the movement of marble freight (83).
Chapter 3 models the distribution of fineware ceramics around the Aegean in 50-year time slices of 32 sites, with a representative sample of pottery drawn from 50 publications and the Miletos Archive, yielding a database of 25 735 pieces. This modeling includes distribution maps yielding the relative percentages of Attic, Corinthian, Ionian, and local pottery across these 32 sites as well as Principal Component Analysis (PCA) to visualize the similarity of all sites to one another based on the co-variance of their ceramic assemblages. Loy concludes overall that knowledge networks and infrastructure were driving these patterns, positing that networks began to shape themselves around specialized production over the 7th and 6th centuries. Using PPA to reconstruct weighted shipping routes, Loy argues that the shipping of finewares paved the way for the rise of heavy marble freight in the 6th century (Chapter 2) and the entangled networks of luxuries, semi-luxuries, and commodities demonstrate a mixture of embedded and market-driven economic behaviour. These two chapters thus nicely build on one another to illustrate the emergence of entangled economic networks.
Loy elaborates on these entangled networks in Chapter 4, which attempts to explain how coinage spread quickly across the Aegean in the 6th century BCE. The main analytical component is the weight standards in use, and the networks of knowledge and interaction that can be reconstructed from their distribution. 631 distinct coin series from 60 sites are visualized through network models produced in Gephi in geographic and non-geographic (“force atlas”) configurations. These patterns are compared to patterns of amphora distribution from the same periods. Both datasets show how networks emerged as small clusters of interacting communities before taking on Aegean-wide linkages, albeit at different times (the late 7th century for amphoras and mid-6th century for coinage). The key issue, for Loy, is that these economic networks were entangled with one another: “The contacts that communities made by interacting in one way could be used to form new types of connections, and they allowed for the rapid development of new economic networks” (183).
Loy takes on epichoric alphabets as measures of collective identities and political affiliations between communities in Chapter 5. This chapter opens with an overview of theories of identity in the material record, arguing that writing styles should also be included in material processes of identity construction (192–193). Loy employs similarity matrices to model the co-variance of five types of letter groups, and test earlier regional groupings of alphabet styles devised by Adolf Krichhoff and Anne Jeffery. The results are again presented as network visualizations produced in Gephi, with sites as nodes and edge strength represented by the number of shared letter shapes. Loy uses a modularity tool in Gephi that sorts nodes into clusters or “communities”, based upon the number of shared edges between nodes as well as the strengths of those edges. Network graphs with community clusters are presented in geographic and non-geographic formats to compare the formation of these clusters to their spatial proximities. These results present a more dynamic reality of clustering variations across the 7th and 6th centuries than the static regional groupings represented in earlier studies. These clusters are not necessarily driven by geography, but possibly by the different contexts of writing, from small portable objects to large stone inscriptions (237–241). But Loy concludes that these dynamic patterns were ultimately due to shifting identities and political affiliations (246). The rationale for tying letter shape choices to political identity ultimately needs more demonstration, as it is not clear why these choices would explicitly point to an expression of political identity as opposed to other types of communities of practice (or even more unconscious cultural patterning).
Chapter 6 summarizes the results from previous chapters to situate them within a broader historical framework to test the role of historically attested political networks (using the Kalaureion amphiktyony and the Ionian League as case studies) in determining the patterns thus far observed (247). Loy downplays anachronistic representations of political alliances in later literary sources and argues instead for a bottom-up approach offered by social networks constructed from material datasets. Such an approach allows a more fine-grained, dynamic view of the complex intersections between identity expression, political alliances, and economies. The conclusion further considers how the data-driven results in this work inform us on questions that historians have typically asked of Archaic Greece, including the structure of the ancient economy and the “rise of the polis”. Loy suggests that the models presented resonate with Bresson’s (2008) view of an economy that exhibited a combination of market behaviour and social embeddedness. The data patterns reveal a more gradual process of peer polity interaction and competition along emerging intertwined economic and knowledge networks. Loy emphasizes the community (rather than the larger regions that archaeologists typically study) as the driver of these networks that enabled the formation of larger political configurations over time (287).
As noted above, Loy’s work in many ways successfully bridges classical archaeology’s “traditional” past with techniques of quantification and computational analysis emerging in more recent decades. The last paragraph makes a plug for returning to the legacy data of Big Digs with new eyes and new techniques — a worthy exhortation. The visuals produced from this data are, for the most part, excellent, and links to higher-resolution network models on the author’s website are provided in figure captions. Readers will also find appendices of all the datasets on this site. Some imaging could be improved, in particular the pale and overlapping labelling on PCA scatter plots. A few methodological issues are obscured, for instance the methodology for fitting broad date ranges for some inscriptions into 50-year time slices. The author mentions more information will be provided below, but this is not forthcoming; it turns out the methodology is outlined in the caption for Table 5.2.
A question raised by configuring the data within such models is who or what becomes lost. One thing that struck me was the focus on the mobilities of artifacts (and the social connections these mobilities represented), with little focus on the mobilities of people. The question of the adoption of coinage is a case in point in which network models might be obscuring other mobilities. For instance, Panhellenic sanctuaries undoubtedly brought elites together in venues of competition, where various citizens might have learned about coinage and brought this knowledge back to their home city-states, a phenomenon that maps and network models charting where specific weight standards were used would miss.[4] Importantly, Loy eventually does come to recognize sanctuaries, particularly Panhellenic ones, as venues for structuring political identities (230–232), and highlights the unique roles of sanctuaries as small geographical units that had big effects on political and economic patterning (286) — perhaps a venue for future investigation.
There are a few further areas that might be expanded in Loy’s work: the economic side of Loy’s overall argument is better articulated than the political, in general. In addition, theoretical and mathematical concepts of SNA might be fruitfully integrated to further explicate the adoption of material markers of identity across the Aegean (e.g., strong and weak ties; different centrality measures). But by and large these criticisms expressed are simply ways this energizing study prompted this reader to think about the data and its patterns. Loy has done admirable and dogged work in aggregating an immense amount of legacy data into new forms that can reinvigorate old debates and introduce new questions about the development of communities and economies in the Archaic Aegean. Loy’s conclusions, based on the aggregation and modeling of this data, in some cases confirm what we apprehended through vaguer, qualitative comparisons, and at other times introduce novel findings, particularly demonstrating the dynamism of these entangled economic networks over time and the differing roles of urban and sanctuary sites in these networks. Too long have the catalogues of the original Big Digs sat in disparate tomes, and this study will be an inspiration to future researchers, another bridge over Renfrew’s “Great Divide” worth building upon.
Works Cited
Bresson, A. 2008. L’économie de la Grèce des cités (fin VIe-Ier siècle a. C.). Paris: A. Colin.
Kristiansen, K. 2014. “Towards a New Paradigm? The Third Science Revolution and Its Possible Consequences in Archaeology.” Current Swedish Archaeology 22: 11–34.
Mills, B. et al. 2013. “Transformation of social networks in the late pre-Hispanic US Southwest.” PNAS 110.15 5785–5790.
Mooring, J. 2022. “Weak and strong ties in the diffusion of coinage during the Greek Archaic period.” In Networks and the Spread of Ideas in the Past: Strong Ties, Innovation and Knowledge Exchange, edited by A. Collar, 79–98. London: Routledge.
Renfrew, C. 1980. “The Great Tradition versus the Great Divide: Archaeology as Anthropology?” AJA 84.3: 287–298.
Notes
[1] Renfrew was specifically addressing the Archaeological Institute of America on its centenary; certainly, there were classical and anthropological archaeologists prior to this address and in the following decades who were building bridges between these disciplines.
[2] The First and Second Science Revolutions being, respectively, the discovery of deep time and the advent of scientific dating in archaeology.
[3] Loy uses estimates of the original height of statues to then estimate the weight of the block of marble used to carve each. To account for inevitable variations, he ran a number of possible input values through a series of calculations and took the median average (62).
[4] See, for instance, Mooring 2022 (this reference is included for the readers’ information; the reviewer recognizes that the author would not have been able to incorporate this reference given the time frame of publication).