Settlement Patterns in Mediaeval and Post-Mediaeval Sphakia: Issues from the Archaeological and Historical Evidence
L. Nixon, S. Price, O. Rackham, and J. Moody, as submitted originally in 1999 for Proceeding of the Mediaeval and Post-Mediaeval Conference, Corfu, May 1998. A revised version appeared in 2009 in John Bintliff and Hanna Stöger (eds), Medieval and Post-Medieval Greece: The Corfu Papers, BAR International Series 2023. Oxford: Archaeopress, 43-54.
The Sphakia Survey is an interdisciplinary archaeological field project studying the eparchy of Sphakia (470 km2), which had not previously received systematic investigation. Our objective is to reconstruct the sequence of human interaction with the environment in southwestern Crete, from the time that people arrived (ca 3000 B.C.) up to the end of the Turkish period (A.D. 1900). Sphakia, a rugged and remote area, includes most of the White Mountains, and ten major gorges, notably Samaria. Thanks to a generous permit obtained from the Greek Archaeological Service through the Canadian Archaeological Institute at Athens, researchers began working in Sphakia in 1987. Fieldwork was conducted in 1987-90, 1992; final site revisiting 1996; study seasons 1992-95, 1997-99.
Our survey has a serious general commitment to the synthesis of all available evidence--environmental, archaeological, and documentary--for all three major periods of the survey (Prehistoric = PH, Graeco-Roman = GR, and Byzantine-Venetian-Turkish = BVT). We have located some 312 sites, about 50% of which have a mediaeval or post-mediaeval phase. For the purposes of this paper, Byzantine means 962 - 1204; Venetian, 1204 - 1669; and Turkish means 1669 - 1900. We are most grateful to the specialists who have helped us with this material: for pottery, Margrete Hahn (Odense) and Pamela Armstrong (Oxford); for mediaeval Greek, Rosemary Bancroft-Marcus (Oxford); for Ottoman documents, Machiel Kiel (Utrecht). We would like to emphasize how much the synthetic work of this (and other) survey projects depends on the expertise of many different specialists, and we thank all of them.
We have preliminary figures for site numbers in relevant periods. There is some evidence for the period between ca 800 and the 14th c. but not much -- three sites with 8th-9th c. sherds (the Late Roman to Early Byzantine period in Sphakia), and four with 11th-12th c. sherds -- whereas eight sites have sherds from the 13th-14th c., and 21 have 15th c. sherds. We suspect that the post-antique settlement pattern was established not long after the turn of the 1st millennium A.D., although we cannot show this with complete certainty. We are reasonably sure that it was established by the middle of the Venetian period, because we have datable churches in villages throughout the eparchy, five of which are at least 14th c. in date. For the later mediaeval and post-mediaeval periods, this settlement pattern, as far as we can tell, remains relatively stable.
In this paper we will discuss three issues arising from the material.
1. Dating Simple Churches and Vernacular Architecture
Until recently, dating simple churches and vernacular buildings had not been a high priority. Thus many of the churches in Sphakia, which, as elsewhere, are architecturally simple, are currently undated by experts in the period unless they have datable frescoes; BVT houses too are difficult to date. But diachronic survey projects like the Sphakia Survey have begun to ask questions about landscape use, which could be answered if dates were available for these structures.
In Sphakia, the tiny church of Agios Pavlos, on the coastal route between Agia Roumeli and Loutro, is a good example of how a secure date can help with the interpretation of a local landscape. Agios Pavlos is an isolated church (xoklisi) dated to the 10th - 11th centuries A.D. because of its identification with a church said to have been built by Ioannes o Xenos in his will (Tomadakis 1948; Gallas, Wessel, Borboudakis 1983: 256-257). Its position is key: it shows the location of fresh water, and it marks the junction of an important path up to the village of Agios Ioannis. Knowing the date of Agios Pavlos therefore suggests that the settlement pattern of traditional modern Sphakia was already emerging at the turn of the first millennium. Current work in Sphakia is showing that xoklisia like this usually mark significant features of the landscape, so their dates are definitely important.
A second example will confirm the importance of dates for xoklisia. The simple rectangular church of Agios Niketas lies near the eponymous fort at Frangokastello, on the coast of the Frangokastello Plain. Built over an earlier, larger 6th c. basilica with a mosaic, it was constructed before 1371. We know this because of a graffito, not because of any architectural features (Spanakis n.d. vol. II: 387; Sanders 1982: 123, 165). Nearby to the north in the Frangokastello Plain, in the locality called Lakkos is another simple mediaeval church, Agios Ioannis. Marble spolia, presumably early Christian, are built into it, but these do not give us a construction date. A date for Agios Ioannis might clarify the specific function of both churches within the local landscape, and the stages of BVT use of the plain; and these are only two of the xoklisia in this area.
The same problem occurs with vernacular architecture. The villages of Patsianos and Kapsodasos at the northern end of the Frangokastello Plain definitely belong to the (B)VT period, but we do not know exactly when the houses within them were built. Again, these construction dates could be correlated with those of the churches, in order to see how the Frangokastello Plain developed after the end of the GR period. Similarly, the Skordyli House in the Anopolis neighbourhood of Gyro was dated to the Venetian period by Gerola, but this is not a precise date, given the length of the Venetian period (Gerola 1905-32: IV.266, fig. 166). And what about all the other houses in Gyro, not to mention those in the other nine old neighbourhoods of Anopolis?
Another important category of vernacular architecture in Sphakia comprises the shepherds' huts (mitata) in the summer pastures of the White Mountains. These structures, though obviously BVT in date, cannot be given a precise date without excavation, and the same applies to the mitata elswhere in Crete such as Mt Dikte in Lasithi, and Psiloreiti/Mount Ida (Syrmakezis 1988, which focusses on the technique of corbelling). The pottery associated with the Sphakiote mitata does give us some idea of when, and how, these structures and the areas around them were used; we have B and BV, VT and T pottery, but no pure V. To the pottery we can add documentary evidence for cheese production and distribution. Cretan cheese generally is well known in the V period as a major export, and some of this may well be from the White Mountains. But the first explicit testimony is from the very early T period (Randolph 1687; cf. Praktikidis 1900: 72 [of 1818]; Raulin 1869: 220; Olivier 1801: II. 225-226; Sonnini 1801: I. 432; Triantaphyllidou-Baladié 1988: 202-203). Thus the pottery and the available texts mentioning the White Mountains 'agree' on the late V/early T period. But there is no doubt that reliable construction dates of individual mitata, however they were obtained, would clarify the development of transhumant pastoralism in the Venetian, as well as the Turkish, period in Sphakia.
2. Comparability of Major Epochs
It is difficult to compare major epochs in our Survey, a problem which applies to most other survey projects as well; luckily, the problems of comparison in themselves are informative. One of our goals is to assess, accurately, the size, function, and phases of settlements dating to the Prehistoric, Graeco-Roman, Byzantine-Venetian-Turkish epochs. How is it possible to make a meaningful comparison between a late prehistoric site known mainly from sherd scatters, an Iron Age site with pottery and possibly structural remains, and a BVT site with buildings that are currently occupied, where earlier sherds are still 'locked in', only to be released some 50 years after the settlement is deserted? To give specific examples, consider the summer village of Asphendou (with its currently undatable houses, as already discussed, some of which are still in use) ; Loutro in the Late Roman period, with standing remains of basilicas, tombs, and other structures; and a Minoan site in the Frangokastello Plain, marked only by lines of walling, stone piles and sherds.
One way of doing this is to construct settlement hierarchies for each epoch (and for the periods within them where possible). The B period is still problematic for various reasons, but our knowledge of the VT period is more secure. There are very few deserted VT settlements; once the traditional modern settlement pattern was fully established, it lasted almost until the end of the 2nd millennium A.D. All VT settlements (as opposed to other, smaller loci of human activity marked only by pottery, as in other epochs) have plannable structural remains, if not actual houses. They cannot be dated precisely, as we have seen, but they definitely belong to this long period. The presence or absence of VT pottery is thus less relevant in terms of estimating site size (though of course it might, like good architectural dates, make finer chronological phasing possible). We can therefore produce figures for settlement sizes in a consistent way for all VT settlements in this period, and then establish whether they vary in any significant way.
Furthermore, within the VT period we can say that there was relatively little variation in settlement size. There are big villages in Sphakia, but they are usually big in aggregate because they are made up of separate, not contiguous, neighbourhoods. There are two possible exceptions, Agios Ioannis with its two distinguished and datable churches (Panagia and Ag. Ioannis, both 14th c.; Gallas et al. 1983: 257), and Khora Sphakion with its multitude of small, presumably family chapels (13 listed in 1637, more than for any other settlement in Sphakia; Khaireti 1968). Why larger Sphakiote villages tend to be 'aggregated' (separate neighbourhoods, as in Anopolis) rather than nucleated is a question that we cannot answer here. What is worth noting is that Sphakia in this period has only the lower end of the total Cretan hierarchy: VT Sphakia has villages and smaller loci of human activity, but it has neither cities nor towns.
Indeed the history of settlement in Sphakia since the earliest period for which we have evidence (Final Neolithic/Early Minoan) is one of gradual attrition/depletion, with occasional hiccoughs e.g. in the Late Roman period, in terms of settlement hierarchy. In the palatial periods, Sphakia has neither palaces nor 'villas' but it does have two or three large settlements, which might come next in a generalized Minoan hierarchy. During the Iron Age, Sphakia does have a sharply differentiated settlement hierarchy, with larger sites (some of which were known as poleis, at least in the Hellenistic period; see Moody et al. 1998). Loutro in the Late Roman period was clearly a substantial settlement, with no fewer than five basilicas. But in the VT period, it is clear that we have only the lower end of the settlement hierarchy, as stated above. On the whole, the upper levels of the settlement hierarchy in Sphakia are topsliced as the scale and complexity of life on Crete (and beyond) increases.
In the end it may be more significant to note that (B)VT settlements tend to be rather similar in size, and therefore lack any real hierarchy, in contrast to the other two epochs of the Sphakia Survey, PH and GR most certainly do have significant differences in size, and therefore in settlement hierarchy. What has emerged from the discussion above is a useful question: what proportion of the overall settlement hierarchy for Crete (and possibly beyond) might Sphakia have, for each of the three major epochs, and the periods within them? It is important to have a sense of what the answer to this question might be, because it helps us to see what is actually available for comparison in each epoch.
3. Using Archaeology and Texts
Here we look at the question of how to make best use of archaeology and texts in an area like Sphakia, which has two kinds of written evidence: particular documents that illuminate general issues (e.g. pasturage treaty of 1435), and systematic documents for specific times and places in Sphakia (e.g. Ottoman records for different taxation areas of Sphakia). The question of using historical plans and pictures of Sphakia is also discussed. These issues apply to other diachronic survey projects as well.
Let us begin with an example of the complementary relationship between texts and archaeology. A document of 1435 gives the treaty between two families of Khora Sphakion, the Skordyli and the Valerianoi, who had been feuding over pasturage rights (Vourdoubakis 1939). It gives the disputed boundary between their two territories which was to be resolved by a marriage, which starts from Khora Sphakion and goes more or less north, naming reference points and toponyms on the way. Many of these can be indentified today, such as the church of Agioi Apostoloi, which has a complex triconch plan (Gerola 1905-32: II. 245-246, figs. 297-298; Lassithiotakis 1961-62: 188 fig.14, 191-192) Though Gerola dates the church to the 16th c., the text of 1435 suggests that there was already a church here in the 15th c. Another is Ergasteria; the name for the locality survives, though its date would be hard otherwise to establish as there are no structures here apart from a cave mandra and a bee enclosure (which may or maot be the one mentioned in the document). No BVT pottery was found at Ergasteria. To the north, towards the Sphakiano Gorge, lies another church, Agios Pavlos. It could scarcely be more architecurally simple, consisting of one blocky rectangular room; the document establishes that there was a church here in 1435, and this structure may well be it. Without the document we would have had no idea of the age of any church in this location.
Archaeological and ethnographic work throughout the eparchy has shown that the landscape of Sphakia is used not in discrete environmental units, but in broad north-south slices. Within each slice three altitudes can usually be distinguished: down, always on the coast; middle; and up. These different altitudes are exploited in different ways in different periods. In this area, Khora Sphakion and Ergastiria are down (0-300 m above sea level; Agios Pavlos (790m) and the summer village of Kaloi Lakkoi (1140m) are both part of the middle altitude; and up is the level of the mitata in the summer pastures above it (1500m). During most of the VT period all three altitudes in this area were used, in order to maximise the benefits of transhumant pastoralism specialised in cheese production; we cannot be certain about the B period.
Thus the document shows the precise relationship between some of the particular settlements in this area. In other words we had worked out the overall pattern archaeologically, and the document helped us with details which we would not otherwise have been able to obtain. Further fieldwork was crucial in enabling us to confirm whether individual buildings and toponyms were still in use, and whether they were where the document said they were.
Another example reinforces the need for reconnaissance. We have three 17th c. pictures of Khora Sphakion: Monanni 1622; Basilicata 1630; and Boschini 1645. We know that the Venetians built a watchtower at Khora, probably in the 15th c.; in the early 16th c. it was turned into a fort and served as living quarters for the local Provveditore. The fort was expanded by the Turks, and destroyed in the uprising of 1878 (Khatzidakis 1881: 83). The location of the (enlarged) Venetian fort is clear in all three pictures, and so is the placement of the houses -- most of them are not on the coast, but concentrated mainly in one area above it, which Boschini labels 'Mesocorio'. There is an architecturally complex church at the extreme eastern edge of this area visible in all three pictures, almost certainly Agioi Apostoloi. The much smaller groups of buildings on the shore and to the east and west are not named.
Field visits have added crucial information to our own picture of the development of Khora Sphakion over the last three centuries. The fort, studied also by Gerola (1905-32: I. 256-260), is indeed ruined, today overgrown by pines, and -- disconcertingly -- more or less invisible. We have learned that the village has four neighbourhoods: Mesokhori, Tholos (to the east), Georgitsi (to the west), and Bros Gialos on the shore. It is easy to see why people originally preferred to live well above the hot and humid coast, because at a slightly higher altitude, summer breezes make the heat more bearable, yet the land is not too exposed in winter. Georgitsi, now completely deserted, was much larger than the small group of houses shown in the 17th c. pictures. Are they accurate enough for us to say that this neighbourhood grew after the mid 17th c.? The houses and churches here are, as usual, undated. It is not difficult to explain why people left Georgitsi in the later 20th c. The abandonment of Mouri after World War II (Georgitisi was on the main mule track leading north to this village); the construction of the modern car road to Anopolis well below it, along the coast, in the 1960s; touristic preferences for sea and sand since the 1970s; all three have contributed to the desertion of Georgitsi and the enlargement of the coastal neighbourhoods. Tholos and Mesokhori have fared somewhat better, but there too, many houses are no longer occupied. Most buildings still in use in Khora are near the sea, in Bros Gialos.
The three 17th c. pictures give us information about Khora Sphakion towards the end of the Venetian period, which can be checked against later accounts like Gerola's in the early 20th c., and must be checked against the actual village as it has survived into the 20th. Khora Sphakion is one of several places in Sphakia where the village name has continued, but the patterns of occupation within it have changed. We shall see another example below in the case of Agia Roumeli.
Detailed Venetian records were not compiled for all of Sphakia, but there are important Ottoman records. Records of poll-tax (haraç) payers in 1659 for the Rethymnon region include two villages that were then in the eparchy of Ag. Vasilios, but now fall within Sphakia: Skaloti and Argoule (Stavrinidis 1970). In addition there are the systematic Ottoman records for what was then Sphakia. Because Sphakia was assigned to the vakif (charitable foundation) of Mecca and Medina, the relevant records are held not with those for the rest of Crete, which was administered directly by the Ottoman Empire, but with the records of the vakif.
For example, for we have very detailed information for the village of Agia Roumeli for one year in the middle of the 17th c. (Istanbul 1649/50: 98). We have names of the adult males, the number of households, and the amount of taxes and tithes paid for various commodities (wheat, barley, oats lentils, cotton, grape must, 'unlawful innovations' i.e. pigs); landuse (grazing, beehives, gardens); and processing equipment (e.g. watermills, which could be used for grinding grain or sawing wood); and we can compare these figures with those for other villages. Thus on the basis of the Ottoman records for 1649/50 we can suggest that people in Agia Roumeli produced less olive oil, and paid more tax on beehives, than other places recorded that year in Sphakia.
But information from documents always needs to be checked on the ground. In this case, an important factor in understanding and using the information in the Ottoman records is knowing the location of the village. 'Agia Roumeli' is the name applied today to the modern touristic village on the coast, with its hotels and restaurants serving people who hike the Samaria Gorge. Before we had access to these written records, our field visits had already shown that the only candidate for a village of the Ottoman and earlier periods is the mostly deserted old settlement which lies inland, and is invisible from the sea. Here, then, was the place to look for some of the material evidence so carefully tabulated in the document.
The houses in the old village are 'old', that is, they cannot be dated precisely, but their construction is of stone rather than cement, and their form is traditional. In number they correspond more or less to the number of households listed in the text. Almost no structures on the coast are older than this century, indeed most of them are hotels and rent rooms built in or after the 1970s, apart from the V church of the Panagia (Gerola 1905-32: II. 181-182 and figs 118-119, plan and photo). Near the old village there are indeed several watermills; when exactly they were constructed, and whether they were used for grain or wood, are questions still to be answered, but the mills do exist (Cf. Valianos 1985). Similarly, the area around the old village is extremely arid, therefore good for thyme and bees, and there are bee enclosures nearby (also currently undatable). Of course people in the Ottoman period made use of the coast; but they did not actually live in a village there, as our reconnaissance has shown definitively.
In Khora Sphakion we saw that the village had remained in more or less the same location, but that the pattern of use had changed. In the case of Agia Roumeli, the name has migrated, and only a site visit will reveal the location of the older village. Again, we know from our archaeological work throughout the eparchy of Spakia that people often alternated between a tendency to live inland and upland, and a tendency to live on the coast. It is therefore crucial to be sure that we know precisely where settlements were, and when. Documents alone do not always give us this information.
Mediaeval and post-mediaeval Greece used to be studied principally by art historians in terested in icons and frescoes, and by historians using only texts. We hope to have shown how archaeological survey, combining diachronic questions and synthesising both material and documentary evidence, can contribute to future work in this area.
Lucia Nixon, Magdalen College, Oxford
1. Sphakia Survey Publications
1988 'Archaeological Survey in Sphakia, Crete', L. Nixon, J. Moody,
and O. Rackham, Echos du monde classique/Classical Views (EMC/CV) 32 n.s.
2. Other Publications
Boschini, M. 1645. Il regno tutto di Candia delineato a parte, e intagliato,
The Sphakia Survey: Internet Edition, University of Oxford 2000
Designed and Developed by ACDT at University of Oxford
This page last updated