Thanks are due to Carl Knappett and the journal editors for the invitation to participate in the volume. The ideas presented here owe a significant debt to productive exchange with various colleagues over the past few years, especially Elizabeth S. Greene, Carl Knappett, Walter Scheidel, and Thomas Tartaron.
- 1 E.g. network metaphors among the various contributions to “A Forum on Trade” in Scheidel 2012.
1Despite a scholarly fondness for general network metaphors,1 few of the broad theoretical constructs or specific methodological tools of network analysis have been employed in studying the archaeological evidence for Roman maritime connectivity. Yet there is a growing awareness of the role played by webs of personal and community relationships in structuring exchange systems across the breadth of the ancient world, with the result that social network analysis may increasingly help to guide our approaches to the historical and material evidence (Brughmans 2013; Knappett 2013). Taco Terpstra’s (2013) recent treatment of ethnic communities of foreign traders, for example, suggests that the social bonds within these communities served to underpin long-distance trade in the Roman Mediterranean.
- 2 E.g. Arnaud 2005; see also the ORBIS project, described by Scheidel & Meeks (2012) and available o (...)
2Most historically based studies of Roman exchange networks have been developed on a macro scale, focusing on long-distance movement of individual goods and empire-wide geographies of transport.2 This scale certainly represents a logical outcome from the compilation of widely ranging and uneven historical, epigraphic, and archaeological evidence as well as an extensive spatial and chronological breadth. Such an expansive view facilitates inquiry into a central concern of research on in the ancient economy over the past few decades: the structure and institutional functioning of markets for maritime exchange in the Roman Mediterranean (e.g. Bang 2008; Temin 2012). Discussions have long focused of how integrated—or ‘modern’—ancient markets were, and to what degree goods and information flowed in reasonably steady and reliable correlations. It is no surprise, then, that the interdependence of networks and markets is an area of growing concern for socioeconomic studies, and that social network analysis can productively inform our approaches to the historical and material evidence of economic phenomena (e.g. Rauch & Casella 2001).
3Yet while the common lenses for studying Roman connectivity are Mediterranean-wide, it is generally accepted that the most routine economic exchanges in the ancient world—at least in terms of volume preserved in the archaeological record—were overwhelmingly local. Bryan Ward-Perkins (2001: 169) summed up this evaluation of short-haul and long-distance maritime movement of goods for late antiquity: “For every amphora that travelled across the Mediterranean, dozens almost certainly circulated only within a regional economy, and yet more only entirely locally.” The social relationships that structured networks governing day-to-day exchange played out primarily on small scales, extending perhaps no farther than individual communities or neighboring settlements. The movement of products within such circles can be difficult to detect and may not have required permanent containers or professional middlemen. In the case of seaborne exchange, long-distance transport is most easily identified through either massive harbor facilities like Portus near Rome or shipwrecks carrying thousands of transport amphoras. Despite this archaeological visibility, the majority of goods surely traveled in small ships, traversing short seas along known coastlines and between socially connected communities.
4Against this incongruity between the most celebrated connections of the Roman Empire as a whole and the routine of local maritime exchange, Thomas Tartaron’s recent analysis of Mycenaean networks offers a welcome typology as a conceptual tool for appreciating and analyzing precisely such issues of scale (Tartaron 2013: 186 tbl. 6.1). Developing at the most local level are what Tartaron terms “coastscapes”, which include the coastal zone and adjacent visual seascape. His “maritime small world” is comprised of many coastscapes connected intimately and routinely through the frequent round trips—of just a couple of days—by fishing vessels and small coastal traders. Together, he argues, these complementary scales of maritime network activity played the most fundamental and under-recognized socioeconomic role in the everyday life of Mycenaean coastal communities. Though developed explicitly for an Aegean Bronze Age context, Tartaron’s approach nonetheless provides a useful framework that should also challenge archaeologists and historians to parse carefully the scales of maritime activity that characterized the spatially and quantitatively extensive communication during the Roman era. Distinctions centering on range of operation, size of ventures, and actors involved in exchange has to some extent featured in the famous model of connectivity developed by Peregrine Horden and Nicholas Purcell (2000) who stress the importance of micro-regional variation within a broader Mediterranean fabric. An approach emphasizing regional scales brings to the forefront critical issues regarding the relative roles and variable integration of modes of network activity. What did the different maritime scales look like in the Roman world? To what extent were they integrated into one reasonably functional market for goods shipped across the empire? Did regional and inter-regional systems of exchange grow directly from the merchants and mechanisms of smaller socioeconomic networks, or did they function on a largely independent scale, with agents and actors creating their own discrete systems?
5Even at the height of ancient Mediterranean connectivity, the notion of one pan-Mediterranean market for goods presents something of a problem. Views of Roman transport—in the Edict of Diocletian or along the stationes of the Piazzale delle Corporazioni—highlight the hubs and links through which large quantities of goods moved (Arnaud 2005, 2007; Scheidel 2013), but the human relationships and knowledge upon which exchange and markets depended remained constrained by logistics and technology. The inquiry I undertake briefly here centers on detecting a regional scale of network activity—between short-haul commerce and longer-distance flows of goods—and identifying how this regional phenomenon may or may not have interacted with other maritime network scales. Given the scope of the topic and state of the evidence, the discussion here is limited to the broader theory of social network analysis rather than specific formal modeling, although more detailed analytical approaches should eventually be possible. My case study comes from the late Roman period on Cyprus in the northeast corner of the Mediterranean, for which a sufficiently detailed body of evidence is available to distinguish the activity of multiple socioeconomic networks (Fig. 1).
Map of the northeast corner of the Mediterranean region showing sites of interest.
6I argue that a distinctly regional scale of seaborne connectivity played a fundamental role in structuring maritime interaction and market formation in the northeast Mediterranean. This network was based on small cargos moving over distances of a few days up to a week or so around and between the island and adjacent mainland. This regionalism of small-scale but routine connectivity involved small ships with minimal crews, cargos of just a few tons, and the unadorned beaches and anchorages that served as occasional ports and markets. The unpredictable coastal circuits of cabotage of the sort proposed by Horden & Purcell (2000: 365) as “the basic modality for all movements of goods and peoples in the Mediterranean before the age of steam” surely played a role in this regional network, but so too did direct exchange. For example, merchants who traveled between small beach-side coves in Cyprus and those in Cilicia created a dense network of ties throughout the area and integrated its economic participants into a shared regional market. While individual harbors were surely not entirely independent of larger economic currents that connected the empire, the region appears to have been the most critical scale of network interaction for many communities in this part of the Roman east during late antiquity. What follows outlines in turn the ports, ships, and distances that constituted the nodes and links of this regional network, followed by a brief discussion of the extent to which this scale of activity may have been integrated alongside longer-distance modes of shipping.
7Unlike state-driven international exchange involving massive ships and built harbors, regional exchange centered on ports that were little more than coastal beaches where goods could be loaded and unloaded. Often these were coves protected from prevailing winds, but any low-lying shore that was accessible from both land and sea could be suitable, even in the absence of natural or manmade shelter. Lacking the built infrastructure of harbor walls, these sites are far less conspicuous than the built ports of the larger cities and towns along the coast, and yet they are arguably more numerous. These sites often exhibit scatters of pottery, but the identification of economic activity in this setting can be difficult to determine as surface survey rarely provides substantial evidence for coins, weights, or other such distinctive indicators of exchange. On Cyprus, underwater and coastal surveys have brought to light several locations that probably served in this capacity, particularly during late antiquity. On the basis of ceramics (underwater and on shore), anchors, and topographic suitability, opportunistic ports have been identified at Avdimou Bay (Fig. 2), near Cape Zevgari along west Akrotiri, and at Zygi-Petrini (Leidwanger 2013c). Although the routine exchange carried out at these locations required little in the way of infrastructure and no permanent port architecture, these sites nonetheless provided a ready outlet for agricultural produce from the immediate hinterlands and a convenient space for the import of goods. Comparisons between the material records of consumption at settlements served by these sites suggest that the ports were strongly connected not only to other parts of the island, but also to the adjacent mainland; very few imports probably arrived from beyond this circumscribed region.
View of the beach and anchorage at Avdimou Bay along the south coast of Cyprus
8The utilization of these nearly ephemeral ports around Cyprus appears to have been the business of small-capacity ships. Only a few Roman and late Roman wrecks around the island have been surveyed in sufficient detail to provide estimates of cargo size, and the shallowness of most near-shore sites complicates the reliability of cargo estimates. Detailed survey of the 2nd century Fig Tree Bay shipwreck off the southeast coast brought to light evidence for 4-5 tons of agricultural goods carried in fewer than 150 amphoras (Fig. 3), the overwhelming majority of which were produced on the neighboring mainland of Cilicia and northwest Syria (Leidwanger 2013a). Another 2nd century shipwreck off the Akamas peninsula comprised just 50 or more amphoras drawn most likely from the west coast of Cyprus or the opposite shore around Anemurium (Leonard 1995: 144-145). One late Roman shipwreck explored at Cape Zevgari off Cyprus’ south coast provides details of what might be a common mode of exchange around the island during the period of interest. This vessel foundered while carrying its agricultural load in perhaps 150 or so jars of probable Cilician manufacture, amounting to no more than 5 tons (Leidwanger 2007: 308-311). Little research has been conducted to date off the island’s northern shores, but two late Roman shipwrecks discovered four decades ago at Cape Andreas suggest a similar model: small groups of regional late Roman amphoras most likely from Cilicia and Cyprus, along with some less common products of the area, including in one case bulk terracotta sarcophagi produced on this part of the mainland (Green 1973). Provenience and distribution study of this sarcophagus type attests to the development of a thriving regional market for such a specialty product centered on this corner of the northeast Mediterranean (Parks & Neff 2002).
Remains of the small cargo from the Roman shipwreck at Fig Tree Bay off the southeast coast of Cyprus
- 3 All sailing times are derived from a GIS-driven method for estimating average voyage durations dev (...)
- 4 For shipwrecks off the Syrian coast, see Kampbell forthcoming. The work of Nicholas Rauh and his c (...)
9Sailing distances—marked by the short sea lengths along the coasts and between the Cyprus and the mainland—provide another measure of the scale of network activity. Regional links here need not have included extensive open-water voyages or long durations at sea. Prevailing winds complicated the matter significantly, in general assisting outbound journeys from Cyprus to the mainland and slowing the return trip. In average winds, the short crossing could have been accomplished in a day or two (Fig. 4 and 5). A round trip from nearly anywhere along the island to this coast could have been accomplished easily within a span as little as two or three days up to a week or slightly more. For local coastal traders making stops while hugging the shore, short hauls of no more than about 10 km between ports meant just a few hours of sailing at a time and probably no overnights away from shelter.3 Realistically, journey times were probably somewhat shorter than these figures, at least for certain slower estimates; unlike sailors headed off on long journeys, a small-time mariner traversing these distances probably had more freedom to wait out an unhelpful wind or ominous sky. Taking advantage of the diurnal cycle of on- and offshore winds would sometimes have allowed more progress against the prevailing winds. The opposite shores of Cilicia and Syria have not been explored in detail—at least not with a primary maritime focus—but the few hints of maritime network development suggest again a heavily regional pattern in the circulation of goods.4
10While small cargos and short hauls have been seen as the norm for maritime transport in antiquity, the tonnages and distances evident from shipwrecks and opportunistic ports off the coast of Cyprus represent an even smaller size and more intensely regional scale of merchant activity than generally acknowledged. In terms of cargo capacity, the vessels appear an order of magnitude lower than the 70-80 tons Lionel Casson (1995: 171) argued for typical small Roman merchantmen. Casson’s numbers reflect a perspective derived primarily from historical accounts of the smallest scale of state-sponsored trade, or exchange associated with wider imperial networks. But a picture of even smaller ships, distances of one to several days, and rather limited ranges of operation between Cyprus and the mainland adds a more nuanced economic geography to the flat and scale-free uniformity often portrayed in short-haul exchange. The benefits of merchant ships with limited size and range could have been substantial. They may have been handled by just a few sailors. Their shallow draft would have permitted them safe passage among the shallow coastal waters as well as easy access and quick loading at simple beachside anchorages. Traveling such short distances, in familiar waters and along well-known coasts would have been a key advantage in terms of safety and security. Given the sailing times above, they could have made multiple trips in a week; for the shortest hops, they may have returned in the same day or the next morning.
Suggested sailing times based on an origin at Seleucia, represented as rings of sailing days using a GIS-driven method described in Leidwanger 2012b
Suggested sailing times based on an origin at Paphos, represented as rings of sailing days using a GIS-driven method described in Leidwanger 2012b.
11The ports, ships and distances described above all point to the operation of a network system of dense connectivity on a geographical scale that was at once larger than local but also distinct from Mediterranean-wide exchange and limited to this corner of the Roman world. For those living along and near the coasts of Cyprus, Cilicia and Syria, this pattern of connectivity would have been a fundamental feature of the rhythm of regional economic life, fostering the type of reliable and persistent commerce that formed a background cadence for prosperity in maritime trade. With common trips, and with parties certainly well known to each other, reliable information could flow around the region quite quickly and a considerable level of integration could be achieved on this geographical scale.
12Determining the social impact of such integration is another matter. Patterns of consumption provide one possible view, but few assemblages in the late Roman countryside of Cyprus have been quantified sufficiently and in such a way as to shed light on this question. An exception comes from the site of Kalavasos-Kopetra, a market town of perhaps 600 inhabitants located up the Vasilikos Valley just inland from the probable opportunistic port at Zygi-Petrini noted above. Here, Marcus Rautman has traced the record of imports during late antiquity, revealing that imported amphoras outnumber local Cypriot jars nearly four to one. This proportion is startling for a site that lay outside the island’s major commercial ports and should be described as minor at best. The high percentage can be better understood through the clarification that nearly all of the imports are drawn from a restricted area along the neighboring mainland, mirroring the economic geography presented by the network picture above (Rautman 2003: 168-175, 169 tbl. 5.5; Rautman et al. 1999; Rautman 2013). Cyprus’ larger port cities record substantial numbers of imports as well—including Amathus, the nearest major harbor to Zygi-Petrini—yet assemblages there exhibit a range of materials drawn more widely from not only the island but the whole of the Mediterranean (Kaldeli 2008: 153-158, 238-241 and 500 tbl. 13 and fig. 23). The material evidence points to two distinct spatial networks of maritime exchange centered on different ports, ships, agents, and commodities.
13Such a picture presents no small difficulty in determining the level of social integration that strictly regional economic connections might have fostered, but it stands to reason that reliable and routine transport was a feature of the maritime landscapes that marked daily life on the island. These small cargos shipped from minor ports might reflect not only the small-scale end of a permanent and distinct class of merchant mariners but also the occasional ventures of fishermen and other socioeconomic groups firmly embedded within the extra-urban Cypriot community. Viewed from this perspective, at least some part of regional seaborne exchange might represent a maritime extension of the agricultural economy rather than a specialized group of professional merchant sailors. Through this network, second-tier and smaller agricultural communities were interconnected across the island and throughout this corner of the Mediterranean, laying the foundation for one type of regional economic market.
14I would contend that these ports and mariners served as something other than a subsidiary level of the same networks that linked the larger urban coastal hubs, whether on Cyprus or across the broader Mediterranean. The operation of a regional maritime economy was dependent on particular economic agents, communities, vessels, goods and mechanisms, many of which may have been appropriate for only this scale of maritime circulation. The same logic and evidence suggest that regional networks also need not reflect a natural outgrowth of the most local Cypriot coasting patterns—a “scaling up” of the individuals and infrastructure that worked on a local level. Yet some of the simple beachside anchorages discussed above may have also served this most local of exchange, raising the possibilities for some level of integration—or at least intersection—of economic networks on local and regional scales. If we are to appreciate and analyze the complexity of the Roman maritime economy, then a network-based approach must emphasize this diversity of scales and allow exploration of their integration.
15The pan-Mediterranean scale of network analysis of the Roman maritime connectivity is critical for investigating issues of integration and information flows within the context of one of the ancient world’s most massive economic systems. Viewed through the lens of economic regionalism, the Roman Mediterranean can be characterized not only by the distinct geographies of network activity but by the different models and mechanisms particular to each network scale. The interplay of these scales serves as an index for investigating a particularly problematic issue of the ancient Mediterranean economy: the possible role of markets and market development across the ancient world. Rather than revealing a single integrated “trade network” in the Roman Mediterranean, this approach suggests multiple intersecting regional and inter-regional networks centered on distinct products, ships, agents, communities, and mechanisms.