The Reconstruction of Settlement Patterns on Thera in Relation to the Cyclades

Basic information is lacking on several characteristics and cannot be interfered from the rest of the Cyclades because of similar deficiencies in the record. Prolegomena for a reconstruction of settlement patterns in pre-eruption Thera are offered, and the paper concludes with a plea for the collection of more relevant data.

SETTLEMENT PATTERNS: GENERAL

Settlements may be considered to be, fundamentally, the "facilities men build in the process of occupying an area" (Kohn 1954). They are centers of accumulation, concentration, dispersal and organization, but consist of the structures within which people live and perform activities as different as peparing food, smelting copper, sheltering animals or worshipping their gods. Within settlements there are various networks which allow them to function, on one level as socio-economic systems (e.g. kinship, "neighborhoods") and, on another, as physical entities (e.g. streets, drains). Settlement patterns may be recognized wherever several such organisms co-exist in space. Although a settlement pattern might be described in terms of all the characteristics outlined, in practice various levels of abstraction are often used. Settlements may be reduced to dimensionless points and patterns discussed simply in terms of the geometrical arrangements. Settlements are located on different types of sites, as well as in various situations with regard to such things as resources, and these patterns can be analysed. Similarly, the hierarchies of size, function and status involve spatial patterns, as do variations between settlements in their overall physiognomies and internal structures.

The ability to reconstruct settlement patterns in at least some of this detail depends upon the available evidence. For very early times, it consists mainly of the physical remains or traces of the settlements themselves, with no descriptions or maps to give additional insights. Inevitably, the data are incomplete, particularly on such basic items as the number of settlements co-existing in a chosen study region at a given time. These problems are particularly acute for Thera in the period before the great eruption of c. 1500 B.C.

SETTLEMENT PATTERNS: THERA

The available information on settlements in Thera has been ably summarized by Sperling, and I am pleased to acknowledge my considerable debt to his publication (Sperling 1973).

Eleven sites are known for the pre-eruption period (fig. 1), but more probably existed. Six of the known sites were found in one comparatively small area where gullying had exposed walls and other objects. Contemporary sites elsewhere on the present land mass of Thera and Therasia are probably buried under thick deposits of ash and pumice, whilst it is possible that one or two settlements were blown literally sky high by the explosion which formed the caldera.

The available data are thus a sample, probably biased in its spatial distribution, and drawn from a total population of unknown size. Accordingly, a very familiar and fundamental archaeological problem, is raised in trenchant form in Thera, namely the amount of inference that can be drawn legitimately from the characteristics displayed by the components of the sample (Binford 1964). The problem may be ignored on the grounds that nothing can be done about it or that it does not matter too much if research is concerned primarily with, say, the stylistic development of pottery or wall painting. But the problem will not go away. In the discussion of settlement patterns, an incomplete population and an unrepresentative sample create real difficulties.

The most basic difficulty is of reconstructing a point pattern of settlements for the island for the pre-eruption period. Without that skeleton, other types of settlement pattern can be neither fully reconstructed nor adequately assessed. Eleven possible settlement sites might be thought to represent a relatively large sample, at least if one supposes that the island has never supported more than 17 settlements, as today ⁽¹⁾. But with more than half of the original island lost when the caldera was formed, one might argue that there should have been space in the pre-eruption period for at least half as many again, giving a theoretical maximum of 25 or 26. On an isomorphic surface these might be expected to have shared the island more or less equally between them, allowing 6 or 7 kms to each. The theoretical spacing would be 2.96 kms. But, of course, the hypothetical regularity of the pattern would have been disturbed by variations in terrain and the distribution of cultivable land. Differences in the size and functions of settlements might also be expected to have disturbed the hypothetical regularity of the point pattern. In fact, the clustering of known sites in the south-west of the island, as well as the mean spacing between known sites of 2.06 kms, does seem to suggest a degree of concentration in the pre-eruption settlement pattern. Its extent and significance, however, can not be judged. Was the pattern of center and satellites seen here repeated elsewhere in the island or was it unique to a particular area? Were the other settlements more or less regularly spaced or were they randomly arranged? There is no way of satisfactorily answering these questions from the Theran data, and little help is forthcoming from the comparative study of the Cycladic group of island in the same period since this aspect of settlement study appears to have been somewhat neglected in favor of the excavation of large sites.

A second difficulty is that of discovering the types of location - sites and situations - preferred by settlements in Thera during the Late Bronze Age. The problem, however, is not just uncertainty about the representativeness of the sample. Ash and pumice have buried the sites so completely that their topographical configurations cannot now be discovered, at least without very extensive area stripping ⁽²⁾. Burial of the ancient land surface also means, taken with the formation of the caldera and alterations in the coastline of the outer rim of the island, that settlement situations cannot be appraised realistically. The relationships to cultivable land or water resources, for example, cannot now be established with any certainty. Almost all that can be said is that, assuming the present outer rim of the island represents something like the original coastline (Sperling 1973), then no settlement in pre-eruption Thera could have been more than 8 or 9 kms from the sea. On the other hand, the amount of imported goods discovered at the main site near Akrotiri might indicate that one or more settlements were situated on the coast, though the mere fact of importation does not require port settlement or any size or elaboration beyond these of modern skales.

More reliable conclusions perhaps can be drawn about settlement hierarchy and structure, partly as a result of Marinatos' excavations near Akrotiri and partly because some parallels can be drawn from elsewhere in the Cyclades, though the usefulness of these is undetermined by two unfortunate traits. The first is the failure of regional surveys to record significant details of settlements in standardized form (Hope Simpson 1965; Scholes 1956) ⁽³⁾. The second trait is the apparently ad hoc basis on which area sampling has been carried out in excavations. This weakens any interference about the form and structure of the whole settlement.

On Thera, excavations at the main site near Akrotiri have exposed densely-packed buildings related to well-defined, narrow streets (fig. 2). These characteristics were not repeated on other Theran sites (Page, 1970, 24 - 28; Sperling 1973). The implication might be that the Akrotiri excavations have uncovered part of a town, as Marinatos suggested (Marinatos 1972). The other sites may be those of looser, smaller and probably less important settlements. Such a simple hierarchy has been suggested for most other Cycladic island during the period (Renfrew 1972).

The structures of the apparently dominant settlements may also be similar. Buildings generally consist of rectangular units ⁽⁴⁾ grouped into cellular blocks bounded by narrow streets. At Phylakopi in Melos, where a large area was uncovered in the 1890s, the degree of regularity is striking (fig. 3) (Atkinson 1904; Hutchinson 1953). Regularity seems clear at Agia Irini in Keos (Caskey 1971) and may be apparent elsewhere (e.g. Naxos). These characteristics, perhaps indicative of comprehensive settlement planning (Hutchinson 1953), allow a positive interpretation of the hints of regularity revealed by the various excavations near Akrotiri in Thera.

Here, as elsewhere in the Cyclades, a complete settlement plan has not been uncovered. Consequently, the size and shape of the settlement are far from clear, and its articulation as well as its internal networks, remain mysteries. The absence of a defining or defensive wall may be significant, but we shall not know until its non-existence has definitely been proved. Nonetheless, the temptation is to follow Marinatos in regarding the building density and defined street system at Akrotiri as urban attributes, and to imagine that the evident sophistication of its art and culture was that of a settlement not only performing recognizable urban functions for the whole island but also of a significant node in the spatial system of the southern Aegean. In fact, the data are not suitable to allow any significant differentiation between settlements on the island, particularly when it is recalled that most of them come from Akrotiri anyway. Comparative study across the Cycladic islands, though limited by the scarcity of excavated sites, indicates that perhaps the Akrotiri settlement was not the unique center misleadingly suggested by its incredible state of preservation. Each island may, have possessed at least one nucleated, comparatively populous and sophisticated centre.

CONCLUSION

If the conclusion of this paper seems somewhat negative, they reflect the present state of knowledge about settlements in the Cyclades during the Late Bronze Age as much as the inherent difficulties of Thera.

They should be seen as challenge to the archaeologists who, after all, are the collectors of the primary data on which all of us seek to interpret. The biggest need is for systematic surveys of all the islands to pin point occupation and activity sites. Linked with this is the requirement that data relevant to settlement study per se should be collected. Particular attention should be given to the size and form of identified sites, for such information will tell us about point patterns, sites and situations and possibly hierarchies and functional differentiation. Beyond meeting such basic requirements, the need is for more - and possibly larger-scale - excavations, based around careful sampling designs, to reveal the internal articulation of settlements. Comparative study across the islands may then allow the in-filling of some of the lacunae presented in the unpromissing but fascinating case of pre-eruption Thera.

- (1). Sperling (1973) supposed a possible total of 12 settlements in the Roman period.

- (2). The pre-eruption surface is exposed on a very small scale at Akrotiri, Balos and in the Phera Quarry (Sperling, 1973).

- (3). For example, Hope Simpson (1965) did not note the prominent "fortification" wall at Phylakopi on Melos. The survey by Scholes (1956) confines itself largely to pottery types.

- (4). The mean width to length ratio of 26 units measured on the small-scale published plans of Akrotiri is 1 : 1.7.

---------------------------------------

------------------------------------