1 A Little History of Mediterranean Island Prehistory
Abstract
Before the mid-1970s, a distinctive subfield of ‘Mediterranean island archaeology’ cannot be said to have existed; there were only archaeologies of individual islands or island groups. So far as the Mediterranean is concerned, an interest in studying islands in a collective and comparative framework, and trying to understand what impact the quality of insularity may have on material culture and human behaviors, can be traced directly to two influential articles by J.D. Evans in the 1970s. This chapter addresses the development of Mediterranean island prehistory from Childe to Evans’s watershed papers, and charts the emergence of a comparative and explicitly quantitative island archaeology, heavily informed by biogeography, in the 1980s and 1990s. It then moves on to the critique of the ‘new’ Mediterranean island archaeology that emerged in the early twenty-first century, and highlights how it has opened up new avenues of inquiry in insular prehistory, not least by emphasizing connectivity, island identities, and the formation of distinct island communities. Using data from the period between the later Upper Palaeolithic and the Late Bronze Age, it seeks to draw out the practical and heuristic consequences of different paradigms, and to suggest future areas of development in Mediterranean island prehistory.
Introduction
The distinctiveness of the Mediterranean, an inland sea as its name indicates, surely depends to a significant extent on the multitude of islands it contains. Indeed, for many of the sea- and sun-seeking holidaymakers who seasonally double the population of the region (Inglis 2000), ‘the Med’ is virtually synonymous with the island-based tourist destinations – Mallorca, Corfu, Mykonos, Cyprus – to which they throng. Other inland seas enclose islands, of course, but not at all like the Mediterranean in their size and distribution: for instance, the Caspian Sea, one-sixth its size, has numerous islands, but all small, very close to the coast, and mostly uninhabited. In the Mediterranean, by contrast, some 150 islands are larger than 10 sq km, 50 larger than 100 sq km, and nine surpass 1000 sq km. An exact count of all its islands is unrealistic, however, because there also exists a myriad of tiny islets, some barely cresting the surface – certainly more than 5000 islands in total (Figure 1.1).
Figure 1.1. Satellite image of the Aegean and its closely crowded islandscape. (NASA, The Visible Earth – http://visibleearth.nasa.gov/).
This, moreover, is merely ‘our’ Mediterranean, now. In reality, the islands we study today are remnants of a drowned landscape, much more of which was above water at the glacial lowstand of the sea (some 120 m below current levels), about 17,000 years ago. Sea-level rise since the glacial maximum, to −25 m around 8000 BP and −7 m around 5000 BP, created a dynamic succession of island configurations dramatically different from the present (van Andel and Shackleton 1982; Shackleton et al. 1984; Lambeck 1996; Lambeck and Chapell 2001; Lambeck and Purcell 2005), to which prehistorians interested in the post-Pleistocene occupation of the islands must attend. Yet even a glance at the map of today’s Mediterranean, which closely approximates the way things have been for the past several thousand years, reveals some major contrasts in size and spatial arrangement of islands within the several separate zones into which the basin can be divided (east Mediterranean, Aegean, Illyrian-Ionian, Sicilian Channel, Tyrrhenian, and west Mediterranean). This immediately makes the Mediterranean ‘something of a unique case’ (Broodbank 2000: 38) when compared with other island theaters such as southwest Oceania, Polynesia, Island Southeast Asia, the Caribbean, or the Aleutians. Furthermore, as we now know from decades of research in various circum-Mediterranean countries, there has been human occupation for tens or hundreds of thousands of years in areas in sight and within easy reach of some of the islands; and yet (leaving aside the very new and controversial discoveries at Plakias on Crete; Strasser et al. 2010; 2011), the first settlement on almost all Mediterranean islands is a development no earlier than the Mesolithic, and in the majority of cases the Neolithic or later (Dawson 2008).
These patterns, both geographical and archaeological, might have been expected to encourage a comparative island archaeology of the Mediterranean, both between different parts of the region and between the Mediterranean and other island-rich parts of the globe. In this respect, the Mediterranean enjoys some advantages, since the exploration of prehistoric sites on some of its islands began almost as early as in some of the surrounding mainland regions, and generally much earlier than it did in, for example, most of the Caribbean or Polynesia. Yet before the 1970s, a distinctive subfield of ‘Mediterranean island archaeology’ cannot be said to have existed: there were only archaeologies of individual islands or island groups.
So far as the Mediterranean is concerned, an interest in studying islands in a collective and comparative framework, and trying to understand what impact the quality of insularity may have had on material culture signatures, can be traced directly to a couple of influential articles by John Evans (1973; 1977). Much of the subsequent work they inspired tended to focus on patterns and processes of earliest island settlement, to fixate on islands as supposedly isolated places, to be heavily influenced by island biogeography, and to emphasize alleged oddities of island cultures. Meanwhile, the head start enjoyed by Pacific island archaeology was also being eroded by developments in the Caribbean theater, as well as a growing number of studies of islands in peri-Arctic waters. In all of these areas, including the Mediterranean, the past 30 years have seen a steadily growing adherence to more theoretically informed perspectives, and the subfield’s distinctiveness has been recognized by the creation of the Journal of Island and Coastal Archaeology, Island Studies Journal, Shima: The International Journal of Research into Island Cultures, and so on. Recent years have seen an expectable reaction to these developments. Certain scholars have critiqued the alleged Eurocentric notions of what it means to be an island/islander, doubting that few islands have ever been truly isolated, and proposing new approaches that revolve around concepts such as ‘islandscapes’ or ‘seascapes.’ More recently yet, interest has shifted to an emphasis on connectivity, island identities, and self-perceptions, and the formation of distinct island communities.
This chapter, then, aims to discuss these changing intellectual and disciplinary underpinnings of island archaeology as a component of Mediterranean prehistory. We begin by asking why island archaeology in the Mediterranean was such a late development, relative to major island theaters elsewhere. We then outline the distinctive Mediterranean island archaeologies that developed in the 1980s and 1990s, the critiques of them posed in the 1990s and 2000s, and some of the themes that are central in current work. Our discussion seeks to draw out the practical and heuristic consequences of different paradigms, and to suggest areas of current development in Mediterranean island prehistory.
Mediterranean Island Prehistory prior to the 1970s
Before the appearance of John Evans’s seminal work (1973; 1977) on the utility of islands as ‘laboratories for the study of culture change,’ there existed very little literature in Mediterranean prehistory that attempted an explicitly ‘island’ approach. The failure of an insular archaeology to develop in Mediterranean prehistory until the 1970s places it a couple of decades, at least, behind developments in other major maritime theaters (e.g., Sahlins 1955, and his ideas about ‘esoteric efflorescence’ in the Pacific). Why was this the case? The lag, we believe, can be attributed largely to the dominance of a Childean diffusionist paradigm in prehistoric Mediterranean studies; and it was only as a result of the increasingly untenable nature of this framework in the 1960s that there emerged the possibility for a truly insular archaeology to develop.
Childe (1930; 1957) conceived of maritime Europe in the Neolithic and Bronze Age as an essentially contiguous zone, through which cultural and social innovations could flow with relative ease. A central component to this thought was the perceived chronological primacy of the development of ‘civilization’ in southwest Asia (Childe 1936). Considering the ancient Near East to be the core of Old World social complexity, Childe envisaged a gradual spread outward from this core, either via demic processes or by adoption and acculturation. In this scheme, the Bronze Age cultures of the Aegean derived the trappings of social complexity from the pristine societies of southwest Asia and the Nile Valley. In turn, these sociocultural, symbolic, and technological suites were passed on, via macro-diffusionist processes, to the megalithic cultures of the central and western Mediterranean (Childe 1957: 16–18, 48–56, 252–64). Thus, the parallels both within the Bronze Age of the western Mediterranean, and also beyond it, in Iberia, Brittany, and ultimately Wessex, could be explained by the gradual spread of cultural forms flowing along maritime routes, all with an ultimate origin in the east. In Childe’s conception, the advent of ‘civilization’ became increasingly recent from southeast to northwest, in a majestic cline reaching from Sumer to Wessex (Figure 1.2). While noting the sometimes highly esoteric aspects of the insular cultures of the western Mediterranean, he understood these to represent the peculiar expression of eastern cultural forms in western contexts (Childe 1930: 193–94, 197; 1957: 261).
Figure 1.2. Map of the Mediterranean and Europe showing the chronological links used by Gordon Childe to date prehistoric cultures, with reference to the historical calendars of ancient Egypt and Mesopotamia (Renfrew 1973: fig. 6).
Concurrent with Childe’s synthetic works of the 1920s and 1930s, a diffusionist framework, itself rather insular in nature if not (consciously) in subject matter, was being developed at the inception of eastern Mediterranean maritime prehistory. The excavation and publication of Knossos by Sir Arthur Evans established Cretan (‘Minoan’) archaeology as a coherent discipline, but more specifically provided evidence of trans-pelagic influences on an insular material culture that demanded explanation (Evans 1921; 1927). Noting the striking cultural changes between the Neolithic and Bronze Age phases of the tell’s occupation, and the apparent debt owed by the latter to the Nile Valley, Evans supposed Early Minoan I to be a product of profound Egyptian influence in the form of outright migration, which transformed ‘…the rude island culture’ (Evans 1921: 16, 66; 1927: 25–59; Pendlebury 1939: 37). Yet he was equally comfortable in looking to the west, noting parallels between Middle Minoan II motifs and the spirals of Maltese Tarxien.
Like Childe, Evans saw the Mediterranean as a great connector, facilitating cultural osmosis from east to west, positing cultural continuities over large expanses of space and time, and the gradualized spread of these continuities from the south to the north and west (cf. Mosso 1907: 329). In Evans’s view, this acculturation (for Crete, at least) was not passive: ‘Insular, but not isolated, it [Crete] was thus able to develop a civilization of its own on native lines and to accept suggestions from the Egyptian or Asiatic side without itself being dominated by foreign conventionalism’ (Evans 1921: 25). Nonetheless, increasingly large data sets, now from the Cyclades as well as Crete, seemed to confirm the understanding of the insular eastern Mediterranean as enabling the flow of culture from southwest Asia to points west. The mounting evidence for ‘Minoanizing’ processes – taken to indicate a Cretan thalassocracy (Atkinson et al. 1904) – further hinted at the degree to which the maritime Mediterranean had been interconnected in prehistory. It was these models, and those like them, that Childe integrated into an explanatory mechanism on a much larger scale, to account for apparent continuities at the continental, as opposed to purely regional level (1930; 1957).
Childe’s immense erudition, and his ability to articulate coherently vast and diverse data, rendered his narrative in The Dawn of European Civilisation (1957) the most powerful vision of prehistoric circum-Mediterranean cultural development during his lifetime, and even after his death. Seminal studies of the prehistoric archaeology of the major western groups relied upon a Childean diffusionist mechanism to explain rapid change or technological innovation. Bernabò Brea (1957) and Lilliu (1962), for Sicily and Sardinia respectively, employed Childe’s conception of the dynamics of culture change, invoking invasion and migration as the causal factors lying behind the significant cultural watersheds of the Neolithic and Bronze Age (e.g., Bernabò Brea 1957: 38, 41, 136). Again, the model was pan-Mediterranean: technological advances reached the shores of the western Mediterranean islands carried by Aegean longboats (Bernabò Brea 1957: 41; cf. Broodbank 2000). The lone dissenting voice during this period was John Evans who, in his work on the Maltese sequence (Evans 1959), was keen to emphasize the enigmatic aspects of (what was then understood to be) the Maltese Bronze Age. While seeking to situate the Maltese archipelago within the wider western Mediterranean context, Evans (1959: 65, 160) nonetheless conceived of the ‘esoteric’ Maltese temple architecture, and the apparent lack of extensive use of metals, as indicative of the development of idiosyncrasies resulting from relative cultural isolation (1959: 30, 133). Nonetheless, Evans still adhered in large part to the general Childean framework and its chronological component.
The dominance of Childe’s compelling master narrative appears to be a primary reason for the lack of any consideration of the unique properties of islands as bounded environmental units. In Childean terms, they were of course hardly bounded at all: with the Mediterranean seen as a connector rather than divider, islands were unusually highly exposed to maritime-borne cultural innovation and diffusion. Under this scheme, the palatial societies of the Aegean and the (admittedly esoteric) insular cultures of the western Mediterranean were merely parts of a far wider region that was relentlessly subject to the impacts of Near Eastern innovations that provided the mechanisms of cultural change.
If the dominance of Childe’s legacy into the 1960s explains the failure of an explicitly insular Mediterranean archaeology to emerge, then the breakdown of the diffusionist paradigm likewise played a decisive role in its development. As became clear in the late 1960s and 1970s (Renfrew 1967; 1968; 1973; but cf. Trump 1980), the long-accepted view that many aspects of prehistoric culture had their origins in the Near East and only later, via the Mediterranean, diffused into ‘barbarian’ Europe became increasingly difficult to sustain. Renfrew’s book Before Civilization (1973) brilliantly demonstrated the revolutionary effects of the tree-ring-based re-calibration of radiocarbon dates in snapping many of the basic links on which the traditional chronologies had been grounded. For example, the earliest cultural phases on Malta became earlier by almost a millennium, with consequences for the entire Neolithic and Bronze Age chronology developed by Evans in the 1950s (Renfrew 1973: 147–66; Evans 1977). With the Maltese Tarxien period now ending at ca. 2200 BC or even earlier (Renfrew 1973: 152), the explanatory connection between the megalithic architecture of the western Mediterranean and its alleged Aegean or eastern Mediterranean origins became untenable.
This dislocation had two consequences for the archaeology of the western basin. First, it suggested that the trajectories of cultural development there were more temporally heterogeneous than had been realized, with the Maltese Neolithic separated from the Sardinian Nuraghic and the Balearic Talayotic; consequently, they could no longer plausibly be regarded as outcomes of the same process. More importantly, the severing of Childe’s temporal equivalencies removed the causal mechanisms for the development of nascent social complexity throughout the western Mediterranean and northwestern Europe in general. The kicking away of the exogenous explanatory crutch necessitated the development of new modes of thinking about the development of novel cultural forms in island contexts.
Matters were different in the eastern Mediterranean, where the advent of calibrated radiocarbon dates did not entirely dislocate the correlation of the Aegean Bronze Age with established Egyptian chronologies, as first proposed by Sir Arthur Evans (1921). Renfrew (1973: figs 20–21) depicted this as an arc or ‘fault line’ running around the Aegean and eastern Mediterranean, within which dates in the third millennium BC were not much altered by the re-calibration. Nevertheless, the rethinking of chronologies, cultural dynamics, and the explanation of cultural change made necessary in the west by the ‘radiocarbon revolution’ also prompted the development of alternative models for the study of Aegean islands. John Evans, who had already stressed the development of cultural idiosyncrasy in the Maltese island group, published the Neolithic-period excavations at Saliagos in the Cyclades and at Knossos on Crete in a manner that placed a new emphasis on the importance of the operation of local cultural processes, constrained by insularity (Evans and Renfrew 1968: 91; Warren et al. 1968: 275).
These various developments were foundational for two publications that – despite varying greatly in scope, length, and aim – turned out to be of huge significance in making the argument for the role of endogenous processes in the development of social complexity in insular contexts. Evans’s article ‘Islands as laboratories of culture change’ (1973) – essentially a manifesto, and extraordinarily influential, despite its extreme brevity – arose from an awareness that new modes of explanation were required to account for the cultural developments witnessed on Malta and the other western islands in prehistory. Given his earlier conviction that local and insular processes played a significant role in the emergence of idiosyncratic forms (Evans 1959: 30, 133), it is unsurprising that he situated local process at the heart of his generalizing claim for the utility of islands in the study of culture change. Renfrew’s massive and magisterial study The Emergence of Civilisation (1972) – based on data from the prehistoric Cyclades, though ranging much further afield, and also adopting an explicitly systems-theory approach – likewise provided support for the insight that there could exist a functional relationship between insular environments and specific trajectories of cultural evolution. Both Evans and Renfrew, significantly, made reference to Pacific island data, the first time that Mediterranean island archaeology had looked beyond its own narrow purview. These publications enabled the development of the application of an insular archaeology at a pan-Mediterranean level.
Developments in the 1980s and 1990s
The collaboration between Evans and Renfrew in 1964–65 in the excavations on the tiny islet of Saliagos was certainly important in providing, for the first time, detailed evidence of later Neolithic settlement in the Cyclades, but the resulting report (Evans and Renfrew 1968) was largely traditional in outlook. Fieldwork undertaken on another Cycladic island a few years later indicates how fast the conceptualization of island prehistory was changing. Renfrew’s excavations at the Bronze Age town of Phylakopi on Melos (1974–77) took place within the context of an island-wide survey directed by John Cherry, arguably the first example in an insular setting of the types of systematic, intensive surveys that were shortly to proliferate in many parts of the Mediterranean (Cherry 2003; Alcock and Cherry 2004).
The monograph on the Melos project, An Island Polity (Renfrew and Wagstaff 1982), brought together the researches of an interdisciplinary group of prehistorians, classicists, geologists, geomorphologists, and geographers, all focused on a theoretical framework set out very explicitly in its opening pages. It emphasized a diachronic systems approach, characteristic of its time, with the island ecosystem of Melos serving as the primary focus, even though set within a regional system (the Cyclades), an areal system (the Aegean), and a larger world-system (the east Mediterranean and beyond). A generally biogeographic outlook was placed front and center, and indeed one of the mantras of island biogeography – ‘In the science of biogeography the island is the first unit that the mind can pick out and begin to comprehend’ (MacArthur and Wilson 1967: 3) – appeared on the very first page, along with references to other island theorists of the time (Vayda and Rappaport 1963; Fosberg 1963). Whatever the verdict may be on this project three decades later, there is an undeniable sense of excitement emerging from the chapters of this monograph that the Mediterranean world had excellent data and original theoretical perspectives to contribute to the developing field of global island archaeology.
This was the context of a paper by Cherry (1981), often cited as foundational for comparative Mediterranean island prehistory. It was prompted, in part, by the realization that clear evidence existed, from the Franchthi Cave in the Argolid, that obsidian had been acquired on Melos as early as the eleventh millennium BC (Perlès 1979), and yet survey on Melos, including the obsidian quarry sites themselves (Torrence 1986), had not revealed clear evidence of pre-Neolithic activities. This encouraged a wider examination of the pattern and process of Mediterranean island colonization, set within a generally biogeographical framework, used as a heuristic tool. The intention was not to assert that islands had been colonized according to the same principles as those of plants and animals – a brilliant paper by Broodbank and Strasser (1991) showed how wrong such an assumption would be in the case of the deliberate initial colonization of Crete, for example – but rather to evaluate how the size of islands and their relative accessibility might correlate with the observed patterns of colonization (cf. Diamond 1977).
This broad-brush study yielded some provocative empirical generalizations. One was that – contrary to a number of overly optimistic claims based on weak data – colonization of Mediterranean islands prior to the Neolithic was vanishingly rare, despite proxy evidence (such as the distribution of Melian obsidian) for earlier knowledge of, and movement among, some of the islands. Cherry’s review also indicated that, as in most other parts of the world, the settlement of islands is a relatively late phenomenon, at least compared to surrounding mainland regions: with certain notable exceptions, this did not begin until the Neolithic, took off only in the earlier stages of the Bronze Age (i.e., between ca. seven and four millennia ago), and followed a sequence that, to some extent, reflected island size and remoteness (Malone 1999; Dawson 2008). Moreover, the proportion of islands occupied by a given stage appeared to be higher in the west Mediterranean than in the east, at least until the late third millennium BC, an observation perhaps explained by disparities in rank-size distributions in either area.
Unsurprisingly, given the huge increase in fieldwork results and the major growth of interest in an island archaeology of the Mediterranean over the subsequent 30 years, some of these conclusions have fared better than others. The accumulation of new data, for example, has essentially erased the east–west distinction just noted (Figure 1.3; Dawson 2008). Likewise, not only has skepticism about significant Neolithic settlement in the Aegean islands been shown to be unfounded (Broodbank 2000: 117–41), but there now also exists a number of instances, well documented by excavation and radiocarbon dating, of Mesolithic island occupations in this area (Galanidou 2011: fig. 1).
Figure 1.3. Plots of the cumulative percentage of east and west Mediterranean islands settled, by millennium: (a) based on data as they existed before 1980, and (b) based on data up to 2005 (Cherry 1981: fig. 5; Dawson 2008: fig. 7.4).
One immediate, and positive, consequence of the paper by Cherry (1981), and several subsequent papers by the same author, was to draw the Mediterranean into conversation, on an equal footing, with discussions about comparative colonization processes in other island theaters (e.g., most prominently, Keegan and Diamond 1987). Another was the development of more nuanced applications of biogeographical approaches, informed by better understandings of island palaeogeography (e.g., Held 1989a; 1989b; 1993). Colonization of particular islands may be affected, for example, by the existence of ‘stepping-stone’ islands, or geographic configurations conducive to serving as seagoing ‘nurseries.’ The very term ‘colonization’ is too simplistic, and requires tricky distinctions to be made between seasonal visits, resource exploitation, first colonization, permanent settlement, and even later abandonment (Cherry 1981: 48–49; Vigne 1989: 41; Dawson 2008). The theory of island biogeography (MacArthur and Wilson 1967) rightly emphasized the predictive importance of ‘area effects’ and ‘distance effects,’ but in a cultural setting these are factors that need considerable modification and elaboration. Thus, distance refers to the widest water barrier between an island and its nearest landfall. That may have varied throughout the last deglaciation, and the relative difficulty of making a crossing depends on winds and currents, as well as the knowledge, skill, and the endurance of seafarers. Knowledge of a distant island may depend on sensory clues (clouds, seabirds, smells), but also on its ‘target area,’ i.e. the angle it subtends on the horizon when viewed from the staging area (Keegan and Diamond 1987; see Figure 1.4 for an illustration from the southern Aegean). Held’s work in the 1980s (1989a; 1989b: 78–104) utilized the concept of a simple ‘target/distance’ (T/D) ratio as an approximate measure of the difficulty (or probability) of reaching an island, and from it flowed many useful insights both within the Mediterranean and beyond. It underscored the relative isolation (in T/D terms) of Cyprus and some of the Balearics (thus revealing the precociously early occupation of the former, and helping explain the late settlement of the latter), and demonstrated that, worldwide, all islands with known Pleistocene occupation have high T/D scores.
Figure 1.4. Illustration of the ‘target/distance ratio’ in the case of the configuration of islands in the southern Aegean, seen from the perspective of the southern tip of the Knidos peninsula (Turkey) as an embarkation point. Crete, which presents a target of 39 degrees, divided by its distance from the next neighbouring island, has a lower probability of contact (expressed as a target/distance ratio) than, for example, Rhodes (when winds and currents are taken into account) (Patton 1996: fig. 3.5).
Another important development during this period of research on Mediterranean island prehistory was an enhanced understanding, based on numerous excavations at early sites, of the chronological sequence of Holocene animal extinctions and introductions. Because the late Pleistocene fauna of many Mediterranean islands is well studied, it has become increasingly possible to evaluate people’s far-reaching impact on island ecosystems, particularly in contexts of overlap between humans, endemic fauna, and humanly introduced species. On Corsica, for instance, the entire mammalian fauna derives from human introductions over the past nine millennia, and all Pleistocene animals have become extinct, directly or indirectly through human action. Yet, in the case of smaller species, there was also a seven-millennium-long period of co-existence (see summary in Cherry 1990: 194–97, fig. 9). While humans seem to be the major catalyst for animal extinctions on Mediterranean islands, it is often unclear whether this occurred as a result of overkill, the introduction of new species, or human modifications of the natural environment. Understanding the overlap (if any) of humans and island pygmy faunas, and the possible role of the former in exterminating the latter, have been major, still unresolved, areas of contention in the case of the Late Epipalaeolithic on Cyprus at the Aetokremnos site (Simmons 1999; Knapp 2010), and of human–Myotragus relationships in the Balearics (Ramis et al. 2002).
Most of the research in this period, summarized above – and much other work not specifically cited here (for references, consult the exhaustive bibliographies in Cherry [1990] and Broodbank [2000]) – was conducted in terms characteristic of its time: a generally processual outlook; a strong emphasis on model testing, quantification, and generalization; and considerable weight placed on ecological and environmental factors. The methods and achievements of this phase of work on island prehistory was usefully synthesized, for the whole Mediterranean, in Patton’s book Islands in Time (1996). While it is an unreliable source with many errors, and is itself now seriously out-of-date, it nonetheless served as a pointer to the styles of research that were soon to become dominant – as intimated by its subtitle Island Sociogeography and Mediterranean Prehistory. Where were prehistoric people, communities, and social processes on islands? Biogeography, islands as laboratories, modeling initial colonization, studying faunal extinctions, and so on were certainly fascinating and productive avenues of research, but they left a very great deal of the lives and identities of prehistoric Mediterranean islanders unexplored and unexplained.
Mediterranean Island Prehistory into the Twenty-First Century
The biogeographic paradigm and approaches derived from it (e.g., Held 1993) largely continued to dominate Mediterranean island archaeology until the late 1990s, due in part to their apparent success in accounting for broad patterns seen in the initial colonization of the basin (Cherry 1990). Yet this continued focus on the earliest stages of island prehistory came at the expense of a lack of attention to many aspects of subsequent insular dynamics (Patton 1996; Broodbank 1999: 237) – human impacts on island ecosystems, differing trajectories of sociocultural and political development, the significance of networks of inter-island and island–mainland interactions, and the emergence of distinctive island identities and materialities. There was also a continued interest in the esoteric and the introverted in island contexts (e.g., Stoddart et al. 1993; Malone and Stoddart 2004; Kolb 2005), despite reminders that this is not an exclusive outcome of island living (Rainbird 1999: 227).
The development of the discipline within the Mediterranean, as a result, became increasingly divergent from that within other theaters. Particularly notable was a growing call in the Pacific for the recognition of continued interconnectivity in post-colonization phases as the cause of large-scale homogenous phenomena such as the Lapita complex (Kirch 1988). Strong challenges were mounted against the bounded nature of the prehistoric island and limitations in environmental diversity as a condition of island life. Evidence was presented to indicate the reticulate nature of prehistoric (primarily Melanesian) societies, in contrast to the prevailing dendritic model of insular cultural development throughout the wider Pacific (Gosden and Pavlides 1994; Hunt and Fitzhugh 1997; Terrell et al. 1997). Mediterranean-based studies remained, by comparison, relatively unconnected to the larger world of prehistoric island archaeology.
Major advances came from the work of Cyprian Broodbank (1993; 2000) and Paul Rainbird (1999; 2007: 26–45), both – significantly – drawing inspiration from sources outside circum-Mediterranean archaeology. Although approaching the topic from starkly different perspectives and dealing with very diverse bodies of material, the work of these scholars in conjunction nonetheless challenged the basic assumptions of the then-dominant paradigm in Mediterranean island archaeology. In particular (and echoing the central theme of debates in the Pacific), both suggested that insularity, as a condition obtaining upon island societies, was best understood as constituted contingently as well as necessarily – that is, prehistoric island isolation might be an outcome of the operation of cultural systems, rather than a premise of this operation.
Rainbird (1999; 2007) both developed a critique of island archaeology as informed by biogeographic principles and also provided an alternative that sought to de-center the island as the focus of study. For him, the category itself was misplaced, an aspect of a largely literary western meta-narrative that has sought to constitute the insular as an object to the non-insular subject (Rainbird 1999: 217–25; 2007: 4–11, 26–39; cf. Baldacchino 2008), thereby empowering the colonizing ‘Cartesian’ westerner. Island archaeology in general, and Mediterranean island prehistory in particular, by utilizing biogeographic principles and seeking generalizations, had failed to provide meaningful information about the nature of human colonization and occupation of individual islands (Rainbird 2007: 36; cf. Broodbank 2006: 218–19). Rather than understanding islands as entities bounded as a consequence of their environmental properties, Rainbird conceived of the sea as the great connector, uniting dispersed and depauperate island communities via a uniform medium, thereby facilitating movement and rendering ‘the island’ meaningless as a discrete unit of analysis. The sea was something traversed by prehistoric peoples, whose very travels helped constitute it as a liquid plain, just as knowable as the land. Using a phenomenological framework akin to that of Tilley (1994), Rainbird constructed a manifesto for a landscape archaeology of what he called the seascape, a term intended to draw attention away from islands and toward the experiential aspects of prehistoric maritime lifeways (1999: 229–32; 2007: 45; cf. Boomert and Bright 2007). Broodbank (2000: 34), at just this time, was also stressing the need for an archaeology of the sea to match that of the land.
Rainbird’s general approach, however, has not garnered many adherents (see the commentaries in the Journal of Mediterranean Archaeology 12 [1999] 235–60). Certainly, the introduction of agency, by emphasizing the behavior and perceptions of people on islands, rather than islands themselves, is a welcome development; and yet a proposed archaeology that shifts its focus away from the material and environment (as Rainbird 2007: 45) seems oddly counterintuitive. His central point, however, is well taken, and recalls the strengths of the diffusionist conception of the Mediterranean as a briny continental highway: it is undeniable, in cases ranging from the Lapita expansion, to the spread of the Neolithic into and across the Mediterranean, to the broad distributions of the obsidian of Monte Arci on Sardinia or of Melos, that some prehistoric societies were comfortable with and demonstrably capable of traversing significant distances by sea. Moreover, while Tarxien Temple-phase Malta may seem isolated, Ghar Dalam Malta, or Tarxien Cemetery-phase Malta, do not (Stoddart et al. 1993; Stoddart 1999; Robb 2001); and since the geometric properties of the island did not change over the centuries, variables other than distance, size, and visibility must have been implicated in the apparent waxing and waning of relative insularity. Geographic isolation, then, at least in the Maltese example, may well have been a necessary condition for the development of esoteric insular cultural dynamics of the sort identified by John Evans – but it was not a sufficient condition. In other words, cultural isolation and esoteric efflorescence, when it occurs, may in part be functions of contingent social processes. Evans’s (1973) notion of islands as laboratories for the study of culture change may have been ill-founded, yet certain islands as a result of specific historical processes have chosen, as it were, to become laboratories.
The work of Cyprian Broodbank (1989; 1993; 2000) to some extent anticipated elements of Rainbird’s critique of island archaeology, and it has certainly opened up a range of innovative ways of interrogating and thinking about the available data. His main object of study, like Renfrew’s (1972), was the Early Bronze Age Cyclades. This regional rather than insular focus, perhaps more readily defensible than elsewhere in the Mediterranean, militated against any approach seeking to deal with each island in isolation. As Broodbank noted (2000: 175–210), from colonization until the second millennium BC the archipelago seems to have functioned as a relatively coherent cultural unit, with many elements of material culture and other aspects of island life shared from Andros to Amorgos, even if a self-aware sense of ‘Cycladic’ identity was not to emerge until very much later. The group demands to be treated as a whole, because insularity, whether cultural or environmental, was highly dependent on context and degree of incorporation within various shifting intra-Cycladic networks. These islands were neither self-sufficient isolates (which, despite Rainbird’s [1999: 288] characterization, was never the position adopted by biogeographical island archaeology), but nor were they locked in world-systemic integration. Depending on the temporal and geographic scale in question, an individual island or island group might have been more or less connected according to the time of year, the productivity of the harvest, the need for a bride for a grown son, the strength of the meltemi, the size of the annual tunny shoals, or the proximity of the next island.
The key recognition is that a range of variables contributes to how insular or how connected a given community is able to render itself (Broodbank 2000: 92–96, 175–210; 2008; Fitzpatrick et al. 2007; Fitzpatrick and Anderson 2008). Islanders exercise their agency within a highly circumscribed environment, yet the boundary and degree of that circumscription fluctuate at a series of temporal scales, presenting and foreclosing on novel capacities and environments in a manner that is infrequently experienced by mainlanders. Broodbank (2000: 21–26) captured the mutability of the parameters and horizons of island life, as well as its liminality, with the notion of islandscape. Defining the various spheres of insular living for a particular community, the islandscape also consequently serves as an appropriate unit of analysis. It has the particular advantage of recognizing the local and specific nature of island experience without abandoning the changing, yet highly circumscribing, conditions of island living, thereby negotiating an epistemological path between island-as-laboratory and the phenomenological seascape.
This understanding of the island world as mutable and semi-permeable resonates with Robb’s (2001) conception of the insularity of Malta in the fourth and third millennia BC. For Robb too, cultural process and geometric reality combine to create conditions in which connectivity, or lack of it, satisfies cultural needs both despite and because of environmental factors. Robb also makes the important point that just as insularity or connectivity as conditions are not absolute, neither are their social and material outcomes. Malta in the fourth and early third millennia BC was isolated, to be sure, but only relative to its participation in wider central Mediterranean networks in the preceding and succeeding centuries. Mediterranean isolation is only such, then, when viewed through a local lens. Little in the way of Mediterranean insular experience is comparable to the exceptionally remote example of Rapa Nui or even Hawaii; conversely, Ibiza and Formentera in the Balearic islands appear to have been settled many millennia after Cyprus (Gómez Bellard 1995), far longer than it took Pacific islanders to journey from island southeast Asia to New Zealand and Rapa Nui. This illustrates nothing save to reinforce that local processes and perceptions are inherently relative.
A further major contribution of Broodbank’s work was to move beyond speculation about relative connectivity and to explore the social and material consequences of insular distance and its crossing. The former was achieved in part by drawing on graph theoretical techniques, as earlier applied in the Pacific, to develop a subtle application of Proximal Point Analysis for simulating patterns of growth in interaction and connectivity among Cycladic communities, as a consequence of colonization, increase in settlement, and intensification of seafaring (Broodbank 2000: 180–210) (Figure 1.5). An important insight was his suggestion that the calorific and labor investment necessary to paddle the ‘canoes’ depicted in Keros-Syros art (Broodbank 1989) within and beyond the archipelago was enough to make any non-local journey a significant social undertaking, being an inherently high-risk (albeit presumably also high-return) strategy (Broodbank 1993; 2000: 276–319). As a consequence, he proposed that the capacity to traverse open water, visit exotic shores, and successfully return home having bartered or traded for goods might generate social capital. In this understanding, accretion of capital by such means was implicated in the emergence of a ‘voyaging ethos’ in the third-millennium BC Aegean, in which successful navigation and its associated behaviors was a dominant feature in intra-communal social relations (as, for instance, in the ethnographically attested Melanesian exchange systems). This in turn partly helps to account for the apparent prominence of ‘sea-going nurseries’ and ‘voyaging centers’ in the third-millennium Aegean (Broodbank 2000: 212–21), and the widespread circulation of pottery, lithics, and other goods within and beyond the archipelago. Indeed, prehistoric seafaring and voyaging have attracted increased attention in recent years, both in the Mediterranean (Broodbank 2006; Knapp 2010) and globally (Anderson et al. 2010).
Figure 1.5. A sequence of Proximal Point Analyses simulating patterns of local interaction in the Cyclades under conditions of demographic growth in the Early Bronze Age (Broodbank 2000: fig. 53).
Broodbank’s book (2000) deservedly won prizes, and it has served as a powerful catalyst for pushing Mediterranean island prehistory, and even global island archaeology, in fruitful new directions entering the twenty-first century. A rich set of research questions were laid out for consideration, many yet to be fully explored. For instance, it may be necessary to reconsider how insularity is manifested at the level of the community: if some members were more ‘connected’ (via status, access to seacraft, absence of sailing taboo, etc.) while others were more ‘isolated,’ this forces a reconsideration of blanket definitions of insularity. Parallels between patterns of resource acquisition and consumption, and indices of increased or decreased insularity, might contribute to this discussion. As another example, more research on the social dimension of voyaging, beyond the context of the Early Bronze Age Aegean discussed by Broodbank, may well be illuminating. This is particularly so in cases in which it is reasonable to think that maritime activity was not a casual, routine activity without social connotations – for instance, in the spread of the maritime Neolithic from Thessaly to points west during the sixth and fifth millennia BC (Zeder 2008; Zilhão 2001).
Whither Mediterranean Island Prehistory?
With the collapse of faith in the biogeographic notion of islands as bounded, naturally defined areas for investigation, and the move beyond a somewhat myopic focus on patterns and processes of colonization, Mediterranean island prehistory has both matured and diversified in the face of an increasingly rich data set and a willingness to adopt new approaches. It can reasonably be claimed that, in being exposed to the discussions surrounding insularity versus connectivity, it has been drawn into the wider field of global island prehistoric archaeology and anthropology, especially concerning the methodological debates ongoing in the Pacific and Caribbean; this exposure, moreover, has had effects outside the purely anglophone academy (e.g., Gili et al. 2006).
Some trends are very promising. Mediterranean island prehistory, a latecomer to comparative and theoretically informed discussions of insularity, finally has a seat at the table. It is now included in global overviews of island archaeology and seafaring (e.g., Cherry 2004; Anderson et al. 2010). Mediterranean prehistorians have themselves organized conferences and edited books that concern comparative island archaeologies on a global scale (e.g., Waldren and Ensenyat 2002; Conolly and Campbell 2008), and in one instance have even shifted the focus of their current field endeavors to the Caribbean (Cherry et al. 2012). At the same time, it must be recognized that, if comparisons between island contexts within the same basin need careful explication, merely superficial comparisons between far-flung island theaters may turn out to be only moderately illuminating. What do the Balearics or the Cyclades have to do with Micronesia or the Aleutians? Context-specific dynamics should not be forgotten in the rush to compare and contrast, lest the outcome merely reproduce inherent differences.
Nevertheless, there is a new willingness to engage with processes whose study might benefit from careful comparison. The work of Helen Dawson (2006; 2008; 2010; 2014) exploring the complexities of Mediterranean island colonization – a line of research that some had imagined was past its use-by date – in the light of increasingly rich chronometric data and methods applied to other colonization episodes, is a particularly good example. Scholars of island cultures in areas otherwise mired in local empiricism have begun to appreciate the benefits of wider horizons (e.g., Berg 2010; Copat et al. 2010). In a different direction, network-based techniques that aim to map and quantify connectivities, and the wider effects of their outcomes, represent another research initiative in which prehistoric Mediterranean island archaeology has much material and many insights to contribute (e.g., Knappett et al. 2008; Knappett 2011).
A striking feature of Mediterranean island prehistory early in the second decade of the twenty-first century is the sheer quantity of information now available. Chance finds, rescue digs, research excavations, and regional projects have generated new data almost everywhere. A major factor has been the large number of landscape survey projects undertaken since the 1980s that have now reached publication. Gkiasta’s recent book The Historiography of Landscape Research on Crete (2008), for example, labors to synthesize results from more than 30 projects on a single island; inevitably, it encountered the difficulty of distinguishing between disparate regional sequences and incongruities among data sets collected with different methodologies (Alcock and Cherry 2004). The major research effort required to synthesize the earliest prehistory of 115 Mediterranean islands (Cherry 1981) was rendered out-of-date in a number of important essentials within a decade (Cherry 1990) and, for the Aegean, at least, required a further major overhaul by the time of Broodbank’s study (2000). Large-scale comparison – the very stuff of island archaeology – has become more, not less, difficult. At the same time, the torrent of fresh discoveries has encouraged the writing of authoritative new interpretative syntheses of the prehistories of individual islands, whether traditionally culture-historical in outlook (e.g., Leighton 1999, for Sicily), or informed by theoretical perspectives unimaginable a generation ago (e.g., Knapp 2013, for Cyprus; Skeates 2010, for Malta).
This chapter has not aimed to summarize any empirical facts about Mediterranean island prehistories, but rather to highlight trends in the development of distinctive approaches to the comparative study of islands within this basin. Nonetheless, it may be appropriate to conclude with a few examples of the extent to which new facts confront what we thought we knew, forcing constant revision of our models. On Cyprus, where until relatively recently the earliest evidence was that of the aceramic Neolithic Khirokitia Culture, new excavations and surveys, informed by high-resolution radiocarbon dating, have pushed back the prehistory of the island by several millennia and necessitated a wholescale re-evaluation of interactions among Late Epipalaeolithic and Early Aceramic Neolithic seafarers, foragers, and agriculturalists (see Knapp 2010; 2013: 48–119 for a recent summary). Cretan prehistory has been shaken up by the report of numerous Mesolithic sites in the Plakias area of the south coast (themselves several millennia earlier than the hitherto accepted earliest evidence from aceramic Neolithic Stratum X at Knossos), but also by the claims – not yet fully supported or widely accepted – of Lower Palaeolithic tools, between 130,000 and 700,000 years old, in the same area (Strasser et al. 2010; 2011). In the Cyclades, excavations at Kavos on the island of Keros have brought to light a massive deposit of fragmentary marble vessels and figurines, many of the latter evidently deliberately broken prior to their transport to the island for deposition, implying the existence of some kind of unique Cycladic-wide ‘sanctuary’ or ritual center (Renfrew et al. 2009). Lastly, to cite a negative example, earlier assessments of the antiquity of settlement in the Balearics have been shown to be overoptimistic, with a colonization stage now no earlier than the late third millennium BC, and full habitation with an indigenous culture only in the mid-second millennium Cal BC (Ramis et al. 2002; Ramis 2010: 75–76; see also Ramis, Chapter 3).
Essential to the development of maturity within Mediterranean island prehistory has been the recognition that many causal factors must be combined, in order to account for the development of island lifeways. Broadly environmental factors are coming to be recognized as just as significant as the social and symbolic aspects of insularity. Recent years have recognized the need to view the terminal Pleistocene and early Holocene Mediterranean as relatively dynamic with regard to both climate and sea level. Thus, Broodbank, in exploring the development of initial Mediterranean seafaring activity, highlights how changing ecological parameters alter human perceptions of productive and unproductive landscapes, thereby potentially transforming islands from inhospitable to relatively welcoming environments over the course of just a few centuries (Broodbank 2006: 208–11; 2008: 75). This integration of cultural and environmental factors offers a convincing account of the initial causation of Mediterranean colonization, and suggests that the motivation and stimuli (rather than simply the capacity) to establish and maintain island communities must be addressed in regional research. This also necessitates accepting that demographic growth exerts very real pressures, especially in liminal and insular societies (Shennan 2009). A correlation between demographic growth, preference for certain soils and ecological niches, and the speed and rate of colonization may illuminate aspects of the spread of island and coastal lifeways through the early Holocene Mediterranean.
Finally, the creation of meaningful landscapes in pristine or semi-pristine colonized environments is a promising avenue of research. Knapp (2008) has attempted to reconcile debates surrounding Cypriot Bronze Age insularity with a sophisticated discussion of ethnicity and identity. In moving toward an island archaeology that balances social concerns with the environmental constraints of insular life, the construction of island identities, their retention, and indeed their abandonment (Dawson 2008; 2014) will feature large in understanding why and how islanders made the material and cultural choices they did. In this respect, the book Material Connections in the Ancient Mediterranean: Mobility, Materiality and Identity (van Dommelen and Knapp 2010), with its focus on the intersection of movement, insularity, and senses of self and place, can be taken as epitomizing the current state of theory in Mediterranean island archaeology at the end of the first decade of the third millennium.
What trends are now emerging in its second decade? Two recent publications illustrate the degree to which diversity now characterizes Mediterranean island prehistory. The micro-focus of Bevan and Conolly’s (2013) innovative survey of tiny Antikythera reveals the minutiae of diachronic island life, while also hinting at patterning in the struggle to eke out a living on islands out of the stream. At the other end of the spectrum, the grand sweep of Broodbank’s (2013) vision of Mediterranean prehistory contextualizes islanders as inhabitants of an interlinked Middle Sea that is far greater than the sum of its parts. The common ground is the recognition that ebbs and flows have equally bound islands to, and severed them from, wider social worlds, and that these processes of engagement and withdrawal are at least in part rooted in ecodynamics operating at various scales. New studies located between these two poles continue to emerge, giving the lie to the demise of island archaeology (as Rainbird 1999 suggested); a renewed interest in human arrival at, subsistence on, and abandonment of islands is noteworthy (e.g., Phoca-Cosmetatou 2011; Dawson 2014; Simmons 2014). Mediterranean island prehistory is, in the final analysis, in rude health: a new generation of island prehistorians, bringing to bear new approaches to old (and new) data and revitalizing questions first posed in the 1970s, promises to keep it so.