Continuing the logic of the expanding geographical and cultural scope, the inter-regional/intercultural maritime interaction sphere involves interactions and networks that extend beyond the Mycenaean maritime culture area. As outlined in Chapter 2, Mycenaean extra-Aegean connections are often deduced by plotting the distribution of Mycenaean objects found beyond the Aegean and the nonAegean objects found in Mycenaean contexts (Burns 2010: 36—40; Cline 1994; Lambrou-Phillipson 1990; van Wijngaarden 2002). The historical sketch and the discussion of the problems of interpreting the very partial evidence presented there need not be repeated, but a few key points might be emphasized (see Table 6.1).
We do not know how often, how far, or from where, Mycenaean ships might have ventured beyond familiar waters. Sporadic visits of Mycenaeans to far-flung lands beyond the Aegean seem assured for Cyprus and the northern Levantine coast in the East, as well as the shores of southern Italy and Sicily in the West. The catalogue of imported objects, which we might have expected to travel home with Mycenaean merchants, is neither impressive quantitatively nor widely distributed geographically, however. A dominant role for ships operated by intermediaries based at Cypriot or Syro-Canaanite entrepots (represented by the Uluburun and Gelidonya wrecks) in maintaining longdistance networks cannot be discounted. It is likely that a series of intermediate nodes was interposed between the East and the Mycenaean heartland, with down-the-line and freelance trading the rule and direct voyages or diplomatic missions the exception. Around the rim of the Aegean, Crete, with its long history of engagement with the East, and Miletos or Rhodes, enjoying proximity and close contact with the Hittites, were well positioned as ports of call and transshipment points. Whether ships were controlled by palaces or independent merchants, there was ample opportunity to mix private enterprise with official business. Long-distance voyaging was the realm of seagoing ships manned by master navigators and seasoned sailors with knowledge of open-sea sailing and experience with sea lanes and hazards en route to distant places. Traversing intercultural space, these ships entered ultimately into the realms of intracultural regions and small worlds, instantiating the intersection and overlap of the nested scales of interaction described in this chapter. Once inside a small world, the xenoi relied on their accumulated knowledge and the accommodation offered by locals for safe landfall at one or more anchorages. This would also have been the case for ships traveling long distances within the Aegean world, as when a ship from the Argolic Gulf made landfall in a Cretan small world at the Gulf of Mirabello. The case for Mycenaean permanent presence in enclaves or colonies beyond Aegean and Ionian shorelines is equivocal at best and dubious at worst.
In closing this section, it needs to be stressed that geographically defined spheres of interaction are always contingent to time and place. I have chosen to define four maritime interaction spheres based on geographical scale, frequency of interaction, and cultural identity. They are meant to fit what we know about the political organization and maritime technology of the LBA, as well as the moderate distances and dense distribution of islands that characterize the Aegean region. It would make little sense to apply these spheres without modification to the widely dispersed islands of Oceania, or to the enormously different conditions of political organization and maritime technology of Greece during the Roman period. Even in the Bronze Age, we can observe coastscapes emerge, thrive, decline, and disappear; small worlds alternate between cohesion and fragmentation; and long-distance connectivity of Aegean polities with the eastern and central Mediterranean fluctuate over time with the political and economic conditions of the day. For this reason, our frameworks must be flexible enough to accommodate a diachronic history of adjustment and reconfiguration. In the following chapter, I will attempt to show how such histories can be written within the maritime cultural landscape framework.
Connectivity and Social Network Theory
Following Fernand Braudel (1972), a central theme of Horden and Purcell's The Corrupting Sea is that the extreme fragmentation of Mediterranean coastlands and islands encourages intensive local interactions by sea, while relatively easy maritime communication allows these to be expanded to form larger networks as environmental, economic, and political conditions permit. The emphasis on microregions as the fundamental building block for networks of all sizes is no accident: “The short hops and unpredictable experiences of cabotage are. . . the basic modality for all movements of goods and peoples in the Mediterranean before the age of steam" (Horden and Purcell 2000: 365). Malkin et al. (2007: 1) describe the symbiosis of two tiny, neighboring Aegean islands, Herakleia and
Schinoussa. Herakleia's harbor is protected from the southeast, Schinoussa's from the northwest. As winds and weather change, fishermen and yachtsmen rush to move their boats from one island to the other — a relatively common occurrence that illustrates another element of Horden and Purcell's paradigm, the instability and unpredictability of the Mediterranean climate. The coastal world (or “mental horizon") of these islanders consists first of their village; second of their linked harbors; and third of the Aegean and Mediterranean beyond. This conceptualization of the maritime world conforms well to the coastscapes, small worlds, and larger spheres of interaction proposed in this chapter. Malkin and colleagues (2007: 1) call the relationship between Herakleia and Schinoussa a “fractal" of Mediterranean networks: operating at a small, local scale, but exhibiting dynamics that can extend to the whole of the Mediterranean. This analogy from the natural world is an intriguing one, but debatable — is the architecture of a local maritime network repeatable and remarkably similar at any scale? Because of the quantum leap in knowledge and professionalism, as well as the different kinds of exchanges that characterize the transition from small worlds to larger spheres of interaction, I am skeptical that this should be so.
Yet these scholars rightly question the notion that autarchy or true isolation existed in the Aegean, insisting instead that microregions and microecologies, while distinctive, cannot be separated from the wider networks of which they form a part. Balanced against the certainty that in aggregate, small-scale interaction moved far more goods and people than did long-haul traffic, as well as the likelihood that rather few Bronze Age coastal dwellers ever ventured far from home, must be the realization that the ways that microregions interact and form clusters (i. e., connectivity) are as important as their internal features. As Horden and Purcell (2000: 465) put it, “The wider historical context is as potent a factor in the workings of the microecology as is the local physical environment or the human responses to it." Still, I would defend the separation (however fuzzy) of the different scales of interaction for analytical purposes, in agreement with Michael Galaty and colleagues (Galaty, Parkinson et al. 2009: 43), who caution that they possess contrasting dynamics that cannot be revealed when conflated.
The ways that connectivity works highlight both the corporate responses of (to us, faceless) societies to opportunities and constraints of the physical and social environment, and the decisions and actions of intrepid individuals and small groups. Horden and Purcell's (2000) “four definite places" and the hundreds of other examples they offer to illustrate historical connectivity draw attention to the insights that emerge from written documents and the limits of our knowledge in the absence of them. They offer a kind of template for the establishment of maritime relations that accords agency to individual coastal settlements and their inhabitants: two ports, located a short voyage apart, have goods to exchange and people whose mutual interests and understanding of each
Other promote friendly relations. They make a pact to encourage exchange, and set about improving their port facilities.6 Local agency is a common feature of social networks, which are often understood as self-organizing systems “ . . . due to the local decisions made by the individual vertices" (Barabasi and Albert 1999: 512). Once established, such relations can become durable, carrying on in spite of the vagaries of environment, politics, and economics in the wider world, and even if the rationality and expediency of the connection itself is lost (Horden and Purcell 2000: 128). A reason for this may be the deep social ties that have developed over time. In periods of growth, the information, resources, services, and people flowing along these paths stimulate the creation of new vertices and paths, resulting in clusters that coalesce at larger and larger scales; thus, the formation of world systems from microregions.
Network Models and the Aegean Bronze Age
Two important attempts have been made at spatial modeling of maritime networks in the Aegean Bronze Age: Broodbank's (2000) for the Cycladic Islands in the EBA, and that of Knappett and colleagues (2008) more broadly for the Aegean, with a focus on the MBA. Network analysis has not been explicitly applied to the Mycenaean period.
Broodbank's network model consists of a Proximal Point Analysis (PPA) that simulates interaction networks given certain assumptions about the number and location of interacting nodes (in this case, settlements).7 PPA predicts patterns of connections between points distributed in space, conventionally by connecting each point with the three closest to it. The webs formed by these connections generate network clusters, as some points accumulate more links by virtue of their proximity to a larger number of other points. The denser clusters hypothetically mark out interaction “centers" where communication ought to flow most easily. The actual placement of points in the model is a problem given the fragmentary nature of the archaeological record. Broodbank addresses this by placing known sites on the map and then proceeding to add points to simulate the growth of population over time. The number and location of these points is determined by varying the amount of land area required to place a point on an island. He creates four different models (PPA 1—4) by adding a point for every 150, 100, 75, and 50 square kilometers, respectively (Fig. 6.2), and then compares the results to the apparent settlement patterns of the Neolithic to EBA Cyclades. PPA 1, with only 19 points, is taken to resemble the Neolithic—EB I pattern of low-density networks in which larger settlements and longer-distance connections are vital for survival. By contrast, the more heavily populated and highly connected EB II is simulated best by PPA 4, in which 54 points yield more localized, high-density networks and maximal small-island participation.
Broodbank's PPA is, like any other model, laden with assumptions, simplifications, and choices that affect its utility in representing reality (Knappett et al. 2008: 1010). In this case, it is assumed that communities interact most intensely with their nearest neighbors, since longer journeys are riskier and more time consuming. Sea travel is taken to be uniform in all directions. Sites are deemed to be of roughly equal size and distributed evenly in space among the islands. The links between them are similarly undifferentiated: one node can connect with any other directly or through a series of short hops as constrained by the available propulsion technologies of rowing and paddling. The only variables that can be adjusted to induce change are the number of sites in the system or the number of links each site can form. While this set of rules and assumptions obviously oversimplifies and distorts the reality of these networks, it is important to recognize that Broodbank's PPA was designed for the limited geographical world of the Cyclades and for a time in which boats were propelled by human power. Settlements were smaller and less hierarchically organized than in the later Minoan and Mycenaean palatial worlds. Within these bounded circumstances, the model can claim some success in explaining the presence or absence of certain centers around the Cyclades. For example, no such center is known on the large island of Andros, which in spite of possessing fresh water and arable land, is shown to be out of the mainstream of near-neighbor connectivity; conversely the settlement of Daskaleio-Kavos flourished on tiny, resource-poor Keros in the Erimonisia group by virtue of its position in a dense web of crisscrossing links. The fit, in terms of prominence or insignificance of sites, between the model and the archaeological record is not perfect, but Broodbank's analysis does demonstrate that network centrality was an important factor in the role that specific settlements played in maritime connectivity and in the way that clusters of small worlds were constructed. Yet PPA is limited in its application — it is not likely to simulate eras with large travel ranges well, or translate easily to greater geographical scales.
Knappett et al. (2008) sought to build on Broodbank's beginnings to devise a more sophisticated network model with wider applicability. Their model, which they characterize as “imperfect optimisation," uses a detailed mathematical equation to express the notion that participants in a network tend to strike a reasonable, though never perfectly optimal, balance between the costs and benefits of maintaining maritime connections. The model is far more flexible because it incorporates more of the variables that influence connectivity and allows the weight of each variable to be modified, either experimentally or to reflect current understandings of the archaeological record.
To assess the likelihood of connection between two sites, or the connective potential of any single site in a network, each site is coded with several variables, including a kind of estimate of importance based on site size, population, and available resources. Where the connectivity between any two sites is concerned,
6.2 Broodbank's PPA versions 1-4, based on different initial and growth conditions. Brood-bank 2000: 184, fig. 53. Courtesy of Cambridge University Press.
The energy required to maintain contacts is a combination of the physical distance between them and the fraction of effort each devotes to the interaction. These values lead to an equation:
H = - KR - XE + jP + jxT,
Where H yields a quantitative representation of the energy balance between the costs of supporting the local population (P) versus the costs of maintaining links (T), and the benefits of exploiting local resources (R) versus the benefits of maintaining links (E). The characters k, X, j, and jx are constants that record the relative importance or weight accorded to each factor. The flexibility of this model resides in accommodating a broader number of variables that can influence connectivity, as well as the ability to alter variables as population, settlement patterns, or technologies (e. g., seafaring or mining) evolve. At the same time, the ratios of the constants can be modified to reflect different cost-benefit relationships between local affairs and nonlocal interaction.
Even with this more sophisticated and encompassing approach, Knappett's model carries its own assumptions and simplifications. A central assumption is
The network centrality of large sites, like Knossos, which are more connected than smaller sites and attract new connections preferentially because of their greater ability to acquire and control the resources needed to sustain and benefit from overseas contacts. Further, large communities tend to target each other, creating longer-distance connections and network hierarchies. This “gravitational pull" can aid in linking distant settlements and holding large-scale networks together. The disproportionate role of well-positioned nodes finds strong support in network theory. Albert-Laszlo Barabasi and Reka Albert (1999) describe two common properties of networks: continuous growth by the addition of new vertices (i. e., nodes or points), and preferential attachment by which new vertices attach disproportionately to sites that are already well connected. These patterns can be observed in many kinds of social networks, such as the linking of documents in the World Wide Web or the preferential citation of certain articles in scientific literature (Barabasi and Albert 1999: 510). These dynamics have several implications for Aegean Bronze Age maritime networks. One is that a node that acquires more connections than another one will accumulate them at an increasing rate, causing the difference in connectivity between the two to multiply as the network grows (Barabasi and Albert 1999: 511). Such conditions may help explain the emergence of Mycenae during the Shaft Grave Era, with its wide access to exotic goods, or the rise of Knossos to an unparalleled position on Crete. Furthermore, the rapid growth in disparity between well - and less well-connected sites may result in the kind of explosive growth to prominence that seems to describe Mycenae in the Shaft Grave Era.
With the advance of seafaring technology and the emergence of large centers in protopalatial Crete, conditions were set for Aegean-scale networks to grow, requiring a model of greater scope and variability than Broodbank's PPA. Knappett and colleagues published a few variations on the imperfect optimisation model, including one in which the benefits of trade were incrementally increased (k = 1.0, 2.0, 3.0, and 4.0: Knappett et al. 2008: 1015—1016, fig. 4). At each increment, the links between geographically distant areas of the Aegean — the mainland, Cycladic Islands, Crete, the Dodecanese and Asia Minor — strengthened and particularly well-positioned sites such as Akrotiri on Thera became crucial “intermediate" nodes in holding the larger network together, in spite of their modest size. Removing these nodes, as when Thera was destroyed in a volcanic cataclysm in the middle second millennium BC, can (and did) cause major disruptions in the broader network (Knappett et al. 2011).
When a PPA analysis was performed on the same data set according to Brood-bank's rules, however, Crete failed to link to the Cyclades, and the Dodecanese Islands were entirely isolated from all other Aegean networks (Knappett et al. 2008: 1019, fig. 7). This is hardly surprising as it is difficult to generate large-scale networks when a node can only connect to its three nearest neighbors. Social network theorY8 Predicts the constant addition of new vertices and the creation of shortcuts between well-connected nodes, linking local clusters into “small worlds" (Watts and Strogatz 1998) and further to large-scale networks that may feature direct connections between powerful centers, or emporia that attract connections from the entire sailing universe.
Knappett's model can be manipulated to simulate admirably enough the kinds of networks that plausibly existed in the MBA Aegean, but it too has limitations, mainly the ambiguity of the quantitative value of the variables and constants that the mathematical equations use. How exactly, we might ask, are numerical values calculated for abstract concepts or data fields for which there is only fragmentary information? A few examples will illustrate this problem. The means of quantifying the “fraction of effort" that one site puts into its relationship with another is not explained; it seems perhaps to be equated with the fraction of trade, but how this is derived is unknown (Knappett et al. 2008: 1014, fig. 2). Part of the effort of maintaining the connection involved the ease or difficulty of intersite travel, but in this model distance and travel remain simplistic. While Knappett and colleagues recognize the need to arrive at travel times (“daily transport distance") rather than simple linear distances, the model does not incorporate any calibrations and we are left with a uniform, essentially friction-free sea. In a dissertation on Roman maritime trade in the eastern Mediterranean, Leidwanger (2011: 90—121) takes initial steps toward a textured seascape by constructing a GIS model using wind speed and directional data, as well as historical sources and sea-trial data from the replica ship Kyrenia II, to arrive at a friction factor, which is finally converted into a buffered map of sailing days from a given location. These maps of “cost weighted path distance" can be used to estimate the differences in travel for outbound and return voyages. For instance, applying the friction factor, the sailing time from lalysos on Rhodes in the southeastern Aegean to Salamis on eastern Cyprus should be around four days, whereas the voyage in the other direction could take more than eleven (Leidwanger 2011: figs. 2.7, 2.10; Fig. 6.3). This kind of information would be valuable as an input for Knappett's distance and effort variables. Even these advances leave room for development, since one could envision friction factors incorporating currents as well as a stochastic element for storms and other hazards.
Other ambiguities arise as an inescapable result of the fragmentary nature of the archaeological record. Population (let alone population density) and carrying capacity figures are simply not available for most LBA settlements, at least not without wide margins of error. Lacking reliable quantitative data for these and other categories, calculations of site importance or cost-benefit for local and long-distance interaction are necessarily open to challenge. It is the double quandary of acquiring robust numerical values for structural features and then translating them through mathematical equations into social behavior that has led many historians and archaeologists to adopt a cautious attitude
(a)
6.3 Maps of cost-weighted path distance for eastbound and westbound journeys in the eastern Mediterranean. (a) Sailing days from Ialysos, Rhodes; (b) Sailing days from Salamis, Cyprus. Courtesy of Justin Leidwanger.
(Malkin et al. 2007: 6). One could continue the discussion along these lines, but it is sufficient to conclude at this point that social network analyses hold promise, as yet not realized. With further development and robust data, they have the potential to reconstruct maritime social and economic networks for prehistory, and to identify the dominant engines driving their formation and change over time. The purpose of the present critique is not to advance a skeptical position, or to deliver a new, improved network model. Instead, I wish to expose the potential pitfalls of working with quantitative models that, if not invested with robust empirical data, give results that may seem authoritative but are in fact illusory. These are concerns of which Broodbank and Knappett were quite aware; hence their cautionary statements. I also want to alert readers that in the case studies to follow in the next chapter, data are often not amenable to quantification. In such cases, I follow Malkin and colleagues in adopting more qualitative characterizations of the maritime networks that I seek to analyze.
World History
