Diachronic Frontiers: Landscape Archaeology in Highland Albania*

Modern frontier studies began over a century ago with Frederick Jackson Turner’s presentation of “Th e Signifi cance of the Frontier in American History,” delivered in 1893. In this now iconic paper, Turner not only outlined a general concept of the frontier, but also called attention to the variability of cultural encounters in frontier zones and their bidirectional transformative power. Subsequent scholarship focused more on the former aspect of Turner’s thesis: the frontier as “the outer edge of the wave—the meeting point between savagery and civilization,” and less on the latter: “the wilderness masters the colonist” (Turner 1920 [1893]: 34). More recently, frontier researchers have reoriented their perspective on frontiers from ‘edges of advancement,’ to ‘zones of contact and interaction’ (e.g.,


introduction
M odern frontier studies began over a century ago with Frederick Jackson Turner's presentation of "Th e Signifi cance of the Frontier in American History," delivered in 1893.In this now iconic paper, Turner not only outlined a general concept of the frontier, but also called attention to the variability of cultural encounters in frontier zones and their bidirectional transformative power.Subsequent scholarship focused more on the former aspect of Turner's thesis: the frontier as "the outer edge of the wave-the meeting point between savagery and civilization," and less on the latter: "the wilderness masters the colonist" (Turner 1920(Turner [1893]]: 34).
More recently, frontier researchers have reoriented their perspective on frontiers from 'edges of advancement,' to 'zones of contact and interaction' (e.g., The modern practice of archaeological survey-regional, intensive, diachronic, and interdisciplinary-is well-suited to the study of frontiers.In this paper we provide the example of the Shala Valley Project, which studies the northern Albanian mountain valley of Shala, home to the Shala tribe.Northern Albania is the only place in Europe where tribal societies survived into the 20t century.We attribute their survival to the frontier position of northern Albania, wherein tribal chiefs controlled access to and through valley systems.Shala provides a classic example of a "refuge" society, perched within a strongly contested peripheral zone.The tribe actively and creatively resisted state incorporation during both the Ottoman (Early Modern) and Modern periods.The northern Albanian frontier may have formed much earlier, though, perhaps as early as the Bronze Age.We bring a broad array of evidence to bear on this question, drawn from the ethno-historical, excavation, and of course, survey-archaeological records.

frontiers and archaeological survey
Th e goal of this paper is to demonstrate how regional survey data can be employed to elucidate aspects of culture contact in frontier zones.We set out to achieve this goal by fi rst briefl y presenting the aspects of regional survey in general that can directly inform the study of culture contact.And second, by presenting as a case study the preliminary results of our own fi eldwork in highland Albania.
Archaeological survey is an ideal technique with which to study contact zones.In the fi rst place, it is regional in scale-as are frontiers.A survey project can construct its sampling universe to encompass an entire zone where contact occurred, rather than focusing on isolated sites within that broader spatial arena.Secondly, archaeological survey is diachronic.Frontier zones in one period are often frontier zones in other periods as well.Th e diachronic study of a single region with multiple episodes of culture contact can be highly illuminating.Th irdly, regional survey can be interdisciplinary in approach, including not just archaeologists, but historians, ethnographers, and geo-scientists as well.
Fourth, and fi nally, surveys collect data about settlement patterns, landscape change, and patterns of land use that pertain to the individuals most aff ected by frontiers: those who lived in them, so-called 'people without history' to paraphrase Wolf (1997).
In this light, our project, the Shala Valley Project (SVP),¹ parallels Lightfoot's approach at Fort Ross of a "holistic, diachronic, and broadly comparative" examination of culture contact (Lightfoot 1995: 202).Th e SVP combines all of these factors in order to study one small frontier zone: it is regional in scope, diachronic and interdisciplinary in approach, and is gathering multiple lines of evidence to demonstrate how people living in a high-mountain valley successfully resisted incorporation by multiple external powers.

the shala valley project
Albania is a small country located along the Adriatic coast of the Balkan Peninsula between Greece to the south and the former Yugoslavia to the east and north (Figure 1).Th e northern high mountains are typically described as being extremely remote.For example: [Northern Albania is] among the wildest and most inaccessible [areas] of the Balkan Peninsula and peopled for the most part by savage and fanatical mountaineers.(Sir Arthur Evans 1885, quoted in Hammond 1976: 35) Th e degree to which the mountaineers of northern Albania have or have not been isolated from the outside world is open to question.Our initial research indicates that materials and people moved in and out of the mountains with relative ease.Th at said, if the people of Shala had wanted to isolate themselves, they certainly had the means to do so; the mountains that surround the valley approach 3000 meters and can be crossed only via a small number of high-altitude passes.Th e southern entrance to the valley is at a point where the Shala River cuts through the mountains between spectacular cliff s, the 'Gates of Shala.' Sealing the valley would have been a relatively easy matter and in fact was done (with tree trunks) in the early 20t century when the Montenegrins invaded Albania and took Shkodër (Durham 1914: 15, 27, 35).Extreme isolation may help to explain the origins and persistence in northern Albania, Shala included, of so-called 'tribal' societies (see Boehm 1983Boehm , 1984aBoehm , 1984b regarding similar societies in Montenegro).However, we also suspect that their position in a frontier zone might help to explain the formation and preservation of the northern Albanian tribes.Gjeçov (1989; see also Hasluck 1954).Today, aspects of the tribal system continue to operate in places like Shala, in matters of kinship, such as marriage and inheritance for example (Odile 1989;Whittaker 1968Whittaker , 1976)).Large households (shpia) organized into neighborhoods (mehalla) share patrilineal descent from a common apical ancestor thereby forming exogamous segmented clans (fi si).Several neighborhoods and fi si together compose a single village.Political power is vested in the person of the family patriarch (zot i shpi).Family heads are appointed or elected to a village council (kuvend) that makes decisions of importance to the whole community.A single council member is elected 'headman' or kryeplak.In Ottoman times, several villages and fi si might be politically joined in a bajrak (a 'banner') led by a bajraktar (a 'banner chief ').Bajraks formed loose tribal confederations; e.g.those of the Shala 'tribe' joined Shosh, Shala's nearest neighbor to the south, and several other tribes, to form the Dukagjin 'confederacy' (farë), one of ten tribal confederations in northern Albania.
Life in Shala is diffi cult.Th e climate is Continental to Alpine and the winters are long and hard.Th e economy is currently built around sedentary agropastoralism (i.e.'mixed' village farming; see Halstead 1990) and functions at or just below subsistence levels; government aid or remittances from overseas relatives bridge the gap.Settled agro-pastoralism and tribal socio-political organization together have had a profound eff ect on Shala's landscape and built environment (Plate 1).Large stone houses, some of them fortifi ed (kulla, smallwindowed towers that provide protection and sanctuary to wanted men), dot elaborately terraced foothills.Fields are irrigated and run-off is controlled by a complex system of small dams and canals.Property boundaries are marked by stone walls and fences, and deeply entrenched, clearly very old paths link fi elds and homes.
During the summers of 2005 and 2006, the SVP conducted research in the village of Th eth (which is divided into nine neighborhoods), located at the upper end of the Shala Valley, and in lower Shala, in the neighborhoods of Nderlysaj, Gak, Lekaj Musha, Gimaj, and Nen Mavriq (Figures 2 and 3).We intensively surveyed all of the cultivated or cleared land in both areas (circa 4 sq km in 683 tracts, 15-m walker spacing).Many of the forested areas around neighborhoods were surveyed extensively, as were several of the high-altitude pastures.In Th eth, all visible architecture (460 structures) was located and recorded.All structures were mapped and photographed, and many were drawn.At the time of archaeological and/or architectural survey, preliminary interviews were results Evidence for prehistoric settlement in the valley is confi ned to the Middle Paleolithic period (Figure 3; SVP Site 001) and perhaps to the Bronze and Iron Ages (Figures 2 and 3; SVP Site 006, and perhaps Sites 002, 005, and 008).Th ere are two major historical periods during which we know that our study region was a frontier: the Ottoman period and the period during which the Albanian ².Th e architectural survey will recommence in lower Shala in .Ethnographic interviews have taken place primarily in Th eth, but will be extended south in .Our goal is to survey archaeologically the whole Shala tribal territory, from the headwaters of the river in Okol to the Gates of Shala.About two-thirds of the territory has been surveyed thus far.nation state formed (fi rst under King Zog and subsequently under the totalitarian Communist dictatorship of Enver Hoxha).Th e situation in Shala during the earlier historical periods (e.g., during the Roman and Byzantine periods) is unclear, but it may be that the valley was abandoned during this time, or only used on a temporary, seasonal basis.³In both the Ottoman period and during the early 20t century imperial powers encroached upon the territory of local, tribal groups with the goal of incorporating them into their bureaucratic systems.Th e relationships between tribes were mitigated by local politics, but also by outside forces.When an external group impinged on their territory, local diff erences were put aside to react to the greater threat.Th us, the activity of core states within the northern Albanian frontier had strong and lasting eff ects on the region's socio-political systems.

Prehistoric Shala
Humans (in this case Homo sapiens neanderthalensis) fi rst entered Shala during the Middle Paleolithic period, probably during the last interglacial (circa 131,000-114,000 years ago) when the mountains would have been free from permanent ice and snow cover.In 2004 and 2005 we collected stone tools of classic Mousterian type at SVP Site 001 (Figure 4).Th e valley would not have been inhabitable in Upper Paleolithic times, and we have found no evidence for Mesolithic or Neolithic occupation.
At the tail end of the 2005 fi eld season, we identifi ed a small structure at the southernmost tip of the neighborhood of Grunas in Th eth (Figure 2, Plates 2 and 3).Th e architectural survey indicated that the ruined building was unlike any other known structure in Shala.As a result the site (dubbed SVP Site 006) was subjected to test excavations in 2006.⁴ Preliminary results indicate that Site 006, which actually is composed of fi ve diff erent structures and is associated with various terraces and walls, is probably Late Bronze Age and/or Early Iron Age in date (circa 1000 bc; Figures 5, 6,  and 7).In four test units and eleven shovel tests, we found many pieces of dark  red, burnished pottery that is similar to prehistoric pottery found elsewhere in northern Albania.At the very top of the cultural horizon we found a few small lumps of iron.Stone tools, including one fl ake of tan chert that may have been imported, and fragments of bone were found throughout.Th ere also was a substantial amount of charcoal, and we have submitted samples for radiocarbon dating. 5One interesting aspect of Site 006 is that it is located in a defensible, strategic position at a natural choke point in the valley.Th e site is protected on the west by the Th eth River, which passes through a deep gorge at this point, and steep cliff s.To the north of the site is a hill which is spanned by the remains of large, rubble walls.Th e eastern edge of the site appears to have been fortifi ed (Figures 5 and 6), though we cannot yet be certain that the walls and terraces at the site are prehistoric.⁶Th e spur of land upon which the site is situated comes to a point at its southern end, which looks out over the lower part of the valley.
Our working hypothesis is that Site 006 was a late prehistoric stronghold that controlled access to the northern reaches of Shala.It may be that the pastures at Th eth were intrinsically valuable and worth controlling.Or it may be that as early as the Bronze Age the Shala Valley constituted an important transportation route for people-shepherds, traders, warriors-moving back and forth between the valley of the Drin River and points north, in Montenegro.⁷Th us, it is possible that already during the late periods of prehistory and just prior to Roman conquest (the fi rst Roman incursions in northern Albania occurred in 229 bc), Shala fi lled a frontier position, perched between competing 'Illyrian' power centers to the south and west along the coast between Shkodër and Dalmatia and to the north and east in interior Montenegro and Kosova.Whatever the case, the valley appears to have been abandoned in Roman and ⁵.Th e pottery from Site  is similar to two pieces of pottery found in  near the rock shelters at Okol in Th eth (Site ).Site  is a ring of fi ve large stones embedded into the ground around a central stone.It is near Site  and so may be prehistoric.Site  is complex of large, overgrown walls located in the neighborhood of Gimaj, possibly prehistoric.Sites  and  will be shovel-tested in .
⁶.We plan to study and perhaps date the walls and terraces at Site  using various geophysical methods, including magnetic susceptibility and radiocarbon dating.Th is will take place in .
⁷. Th e Drin River valley is located to the south of Shala and in ancient times was the main east-west corridor between the Adriatic and the Balkan interior, Kosova in particular (Baçe ; Hoxha ; Palavestra ; Përzhita ; Përzhita and Hoxha ).

Ottoman
According to tradition, the ancestors of Shala's modern occupants arrived in Late Medieval times (sometime before ad 1500), having fl ed Ottoman persecution elsewhere.As a result, Shala may constitute a so-called 'region of refuge' (Beltran 1979;cf. Boehm 1983cf. Boehm , 1984;;Hall 2000: 241).Many of the oldest houses in Shala (some in ruins, many bearing elaborate carvings, a few associated with Late Medieval and Early Modern pottery) mark the valley's fi rst neighborhoods (parts of Gak, Gimaj, and Lekaj in lower Shala, and perhaps slightly later, Okol, lower Gjelaj, and Kolaj/Ulaj in Th eth) and architectural survey indicates their gradual expansion as households grew and split.
About the time the fi rst villages were established a fortress was built atop a prominent peak in the neighborhood of Dakaj (Figure 3 and Plate 4).In 2006 we carefully mapped the entire site (SVP Site 007; Figure 8).Our survey produced a large collection of pottery, all of which seems to date to the Late Medieval period (ad 1200-1500).Much of the pottery is glazed and decorated, some with incised lines and/or combing, while some is heavily-tempered, coarse cooking ware.We also found ceramic wasters and pieces of iron slag, which may indicate on-site production of pottery and iron tools.Dakaj, it seems, was not simply a fortress or refuge site but it had an industrial function as well.
Th e site retains some of its circuit walls (Figure 8), but local landowners indicate that wall stones had been carted away to build nearby houses.Th ere are also the remains of at least two, perhaps more, large building complexes.
Th ese were likely residential (given the presence of cooking wares and industry), but according to local tradition, a church once existed at Dakaj.Careful surveys of the fi elds that ring Dakaj (Figure 3) produced no artifacts, so it is not clear whether Medieval houses or a village existed somewhere in the direct vicinity of the site.Certainly the site is well situated for defense and monitoring of the valley.Th ere are excellent views in all directions.It is unclear, however, who controlled Dakaj: local elite, representatives of Venetian interests, or church offi cials are all possibilities.Continued work at Dakaj, as well as archival research, may help answer this question.In any case, the Late Medieval settlement system is dominated by one large site, Dakaj, repeating a pattern established in the Bronze Age at Grunas.It may be therefore that Dakaj and Grunas served a similar function: control of movement into and through the valley.Th e ability to isolate and protect the valley took on even greater significance in the coming centuries as the Ottomans put increasing pressure on the northern tribes.Ottoman encroachment triggered several interesting responses on the part of Shala's inhabitants in the areas of economy, land use, and social organization.Given northern Albania's geography, climate, and environment, we might predict some form of long-distance, seasonal transhumance of the type that exists, for example, amongst the Koutsovlachs of the high Pindos range of Greece (Chang 1992;Chang and Tourtelotte 1993), but Shala's inhabitants are fully sedentary village farmers.One important question, then, is why yearround village agro-pastoralism developed in Shala in Early Modern times.It may be that Ottoman attempts to incorporate northern Albania and Albanians, through forced conversion to Islam for example (which began in earnest in the 16t century; Pollo and Puto 1981: 90), encouraged a retreat to high valleys, such as Th eth.Th is is the story villagers tell.But fl ight from Ottoman persecution cannot have been the only factor driving migration to and year-round settlement in northern Albania.Another plausible explanation is that changes in settlement and land use elsewhere in Albania and nearby regions spurred indirectly socio-economic change in Shala.
For example, in the 15t century, the Ottomans extended the timar system (whereby Ottoman soldiers [i.e.sipahis] managed tracts of land for the Sultan) from south and central to north Albania causing population displacements and a change in systems of land tenure (Pollo and Puto 1981: 64, 66, 88-89;Winnifrith 1992).As a result, people may have moved deeper into the mountains not only to escape forced conversion, but also to avoid becoming landless serfs tethered to foreign landlords.With increased immigration, population centers in the mountains may have become larger and access to resources, such as good land, circumscribed.In Th eth, as in northern Albania generally, the response was agricultural intensifi cation, perhaps through terracing and irrigation.Investments in the built environment created the landscape visible today.As the landscape became more structured so too did the socio-political system.It was at this time that the tribal system as recorded in the Kanuni i Lekë Dukagjinit probably evolved, a dynamic, social response to the pressures and possibilities of life in a frontier zone.
One major diff erence between seasonally-transhumant Koutsovlachs and settled, high-altitude Albanian agro-pastoralists is that the Vlachs do not pos-  Th is may be due in part to population pressure and greater resource scarcity in the Dukagjin as compared to the Pindos, requiring a more complex system of land management, but there are equally important historical reasons for the existence and persistence of Albanian tribes.Unlike northern Albania, Vlach kinship relationships and political systems are only very loosely defi ned.In Ottoman times Vlach herders were well-integrated into the larger regional economy; greater freedom of movement allowed a fl exible Vlach economy that responded to risk through regional exchange.Conversely, northern Albanians typically have resisted economic and political integration.In fact, there is good evidence that the Ottomans deliberately discouraged integration by exploiting tribal rivalries.Economic and political competition reduced the possibility of tribal confederation and coordinated, mass uprisings (as had happened under Skanderbeg, the leader of the 15t-century Albanian resistance movement).Ottoman offi cials actually created the bajrak system (in the late 18t century; the fi rst written reference to bajraks is from 1783; Ulqini 1991: 167) so that 'banner chiefs' might be played one against the other (Plate 5).In this social atmosphere-population growth and/or movement, shifts to sedentary village agro-pastoralism, scarce resources, increased investment in landscape management, increasingly circumscribed grazing territories, Ottoman political meddling-the kinship and tribal systems were not abandoned, as might have happened with full incorporation into the world-system (Galaty 2002); rather the customary laws that defi ned social and political relationships were codifi ed in the Kanun, which was then strictly enforced, which is exactly what Hall (2000: 241) predicts for so-called 'regions of refuge.'It may be that as confl icts between political segments intensifi ed a ' juridical' solution was sought.

Formation of the Nation State
Th e northern tribes were a political force in post-independence Albania (Vickers 1999).King Zog (neé Zogu), who ascended to the throne in 1928, was a chief of the powerful Mati tribe.In the interest of regional stability, it was the goal of various western governments to aid in the establishment of a stable Albanian state and the northern Albanian frontier played a key role in the national and international power struggles that accompanied the establishment of a centralized government in Tirana.Tribal politics were primarily local in scale, however tribal leaders were well aware of the ramifi cations of impending Albanian statehood.As the Tirana Government (as it was called) took form it sought to incorporate the northern highlands into its political and administrative structure.Th e northern tribes saw the Tirana Government as another external body attempting to impinge on their territory and curtail their traditional lifestyle (Lane 1923).In order to succeed, the central government knew it would have to appease and subdue the country's major social classes: the northern highland tribes (consisting primarily of Catholic and Sunni Muslim pastoralists), the urban Orthodox and Catholic middle class, and the lowland Shiite peasants-remnants of the Ottoman feudal system.In this regard, the perpetually feuding tribes had common interests that fostered political unity.Ultimately some degree of integration was achieved through a combination of negotiation and violence.
In 1921, Rose Wilder Lane, daughter of Laura Ingalls Wilder and someone well familiar with frontiers, visited the Shala Valley of northern Albania with a pair of Red Cross co-workers who hoped to establish a school in the mountains.Th ey were accompanied by Rrok Perolli, an agent of the interior ministry, Rexh Meta, a 12-year-old Muslim orphan (and head of his household) whom Rose eventually put through Cambridge University, and a pair of well-armed gendarmes.In her memoir of the trip, Th e Peaks of Shala (1923), Lane vividly describes the customs and beliefs of the members of the Shala tribe she encountered and recounts a number of discussions she witnessed concerning the place of tribal society within the nascent nation state.
Th roughout her account, the Shala and neighboring Shosh tribes are in the midst of a blood feud.She relates one story in which a Shala man abducted a Shosh woman so that his son might marry her (Lane 1923: 30-31).Not an unusual event, except the woman turned out to be married and her husband took off ense.To preserve his honor, the husband shot and killed the son.According to the Kanun, the murder was unjustifi ed, since it was the father who should have been shot not his son.Th e result was a series of retribution killings (sanctioned by the law code) that ultimately settled the dispute.Such feuds were Diachronic Frontiers  endemic to tribal life and leave their material traces in the form of kulla.
Punctuating periods of blood feud are besas, local peace treaties during which all killings cease.A besa may be performed to negotiate an end to a feud, to give men an opportunity to participate in the harvest, or to cope with an outside threat.Such was the case in the mid 1920's.
On December 17, 1924, after a period of exile, former prime minister and future king, Ahmet Zogu, returned to Albania across its northern frontier with well-armed Yugoslavian and Russian troops.A week later he installed as dictator in Tirana.Unlike his previous attempts at rule, which included negotiation and appeasement, Zogu treated his rivals with ruthless violence.Among his goals was the disarmament of the mountain tribes-needless to say, not a popular policy.In November 1926, partly funded by the Italian and Yugoslav governments (the very governments that had supported Zogu's return), the northern Gheg tribesmen united and launched a revolt.It was crushed within two weeks.In the ensuing year, men who participated in the revolt were detained and executed.
A plaque, shown to us in Th eth by a descendant of Kol Marku (whose name is second on the list), memorializes the men from both the Shala and

Plate 5 -1920s-Era Photograph of the Bajraktar of Th eth Symbolically Surrendering His Gun to Representatives of the Tirana Government
Note: Used with the permission of the Phototekë Marubi in Shkodër, Albania.
Shoshi tribes who participated in the revolt and were executed together in 1927 (Plate 6).Just a few years earlier, these same men would have eagerly killed one another in the midst of the above-mentioned blood feud.Th e inhabitants of the Shala Valley, while seemingly isolated local actors, living what their contemporaries considered to be an antiquated lifestyle, were nevertheless aware of their role in the global arena.Th ey knew that national-in fact international-laws, such as where to draw borders, would strongly aff ect them and their families; they had no choice but to declare besa and revolt.Th is time, though, historical forces were working against them and their traditional defense mechanism, isolation, failed.
Th e pacifi cation of the northern tribes is strongly refl ected in the results of our fi eldwork.We have identifi ed a major shift in building style that according to oral testimony dates to the Zogist period.In the late '20s to early '30s, houses became signifi cantly less fortifi ed: small windows-frengji-were closed up and/or replaced by large, glass-paned windows; traditional access to secondstory living quarters, by ladder or wood stairs that could be pushed away easily during an attack, was replaced by fi rst-fl oor doors and internal staircases; animals were moved out of houses and into barns; and most dramatically, large families split up-the need for communal defense no longer existed.Feud was the engine that powered the Kanun.Without feud and the Kanun, and the frontier situation that necessitated them, the tribal system was doomed.

world-systems theory and the northern albanian tribes
In the foregoing discussion we have repeatedly (and rather generally) referenced world-systems theory in an attempt to explain the patterns of settlement and settlement change documented thus far in Shala.We have referred to Shala as a 'frontier,' but also as a 'region of refuge.'In light of the results discussed above, it is now possible to describe more precisely a world-systems model for Shala.
Taking a world-systems approach, we might defi ne northern Albania, Shala in particular, as 'marginal,' located within a wide, peripheral frontier zone that stretched across the mid-section of the Balkan peninsula (see Sherratt 1993 for the 'nuclear-margin' concept specifi cally).Th e nature of this frontier shifted through time depending on the strength and goals of regional imperial expansion.Various empires touched northern Albania, from several directions, sometimes simultaneously (at which times the region was a 'contested periphery,' see Allen 1997;Chase-Dunn and Hall 1997;Cline 2000) but none it seems managed to fully conquer and/or incorporate the region.It appears to have operated largely outside the interest or control of the imperial politicaleconomies that surrounded it.Th e question is why northern Albania escaped conquest, particularly when so many surrounding regions were 'incorporated' into expanding empires.Processes of incorporation appear to have operated along a continuum that runs from the complete economic domination of peripheral communities by core states (so-called 'eff ective' or 'formal' incorporation) to very informal interactions between core states and isolated, independent regions (so-called 'contact' peripheries) (see Hall 1986Hall : 391-392, 1998Hall : 256, 2001: 242): 242).Most episodes of incorporation appear to fall somewhere in between these two idealized extremes and in reality expanding states used a variety of tactics depending on the situation.'Regions of refuge' are typically thought to fall somewhere in the middle range, and refuge societies often manage to exploit their marginal, frontier position in order to avoid formal incorporation.Kardulias (1999) refers to this situation as 'negotiated peripherality.' In late prehistoric times, Shala may have been valuable as prime grazing territory, as a transportation route, or both, and thus worth controlling.We hypothesize that this was the prime function of SVP Site 006: to monitor movement through the mid-section of the valley.It may be that those who lived at Site 006 stayed there through the early historical periods, but the valley seems to have been abandoned or at least very lightly used in Roman and Byzantine times.Th is is quite surprising given the evidence for these periods in regions surrounding Shala, along the Drin River road in particular.Extensive surveys in tribal territories between Shala and the Drin, planned for 2007, may help resolve this mystery.If evidence for early historical settlement is found there, then the negative evidence from Shala will become more meaningful.
Shala appears to have been re-populated on a permanent basis in Medieval times, as documented at the site of Dakaj (SVP Site 007).As described above, Dakaj's role as a fortress is beyond doubt, but it is not yet clear who lived there and whether they dominated the newly established villages or protected them, or both.Local inhabitants of Shala claim that their ancestors arrived as refugees, seeking to escape Ottoman conquest and forced religious conversion elsewhere.We have no reason to doubt these origin stories, but as described above, we believe the foundation and evolution of northern Albania's tribal societies were subject to more complex processes of cause and eff ect.Shala was isolated, and its inhabitants may have negotiated their peripherality in order to avoid formal incorporation by the Ottoman state, but still they were aff ected by changes in the outside, imperial world.We have argued, for example, that immigration to the mountains was likely driven also by changes in land tenure on the plains.Increases in population and population pressure were met with an intensifi cation of agricultural production, including terracing and irrigation.It is also no accident that New World crops such as maize, beans, and squash were intro-duced to Albania sometime during this period, allowing much larger crop yields (Andrews 1993).Population growth stimulated formation of the tribal system as well as systems of oral customary law and blood feud.
Th ere are several key assumptions in the above sketch that deserve further discussion and are subject to investigation by archaeological survey, as we have defi ned it.Our model assumes that populations in Shala grew and that population pressure was the result.Beginning with Malthus (1803Malthus ( , 1830)), most scholars who study mountain eco-systems have asserted that their carrying capacity is rather low and that mountain economies are particularly prone to collapse (e.g.McNeill 1992: 2-7).Consequently, many mountain societies carefully regulate marriage, birth, and immigration rates.Emigration from mountain to plain may also serve to relieve population pressure.Whereas these strategies have been well documented for the Medieval-Modern period in the European Alps (e.g., Viazzo 1989), Mediterranean mountain regions have received far less attention (McNeill 1992 is a noteworthy exception).Most scholars working in the Alps argue further that mountain societies depended on isolation as a means of buff ering and limiting population growth (e.g.Netting 1981).Alpine anthropologists applied to the Alpine situation the concept of the 'closed corporate community' : small, endogamous villages of small, economically and politically independent nuclear families that closely regulated rules regarding marriage, land tenure, and inheritance.Immigration and emigration were limited.Th is view of Alpine communities was challenged by Cole andWolf (1999 [1974]), who argued that in fact, historically most Alpine communities have been 'open' communities, the frontier reserve from which low-altitude urban imperial cores drew surplus people and products.Th us, the real key to understanding mountain demography was not to track the degree to which mountain communities were isolated and isolationist, but rather to understand how the shifting frontier situation aff ected the movements of individuals in and out of the mountains.Cole and Wolf (1999) therefore view migration as a primary release valve limiting over-population in mountain valleys.Recent scholarship shows that migration was one possible strategy for limiting population in the Alps, but that the migration rate varied through time from village to village (Viazzo 1989).
Mediterranean mountain societies, northern Albania included, diff er from the Alps in several important ways.Northern Albanian family structure is dominated by the so-called 'root' or 'complex, joint' family (as opposed to the 'stem' family common in the Alps) (see Gruber and Pichler 2002), in which several generations live together in one large house and share all proceeds of their work (the so-called 'zadruga' ; Burns 1976).Th e land is owned by the patriarch (zot i shpi) and is divided between sons upon his death or retirement.Th ere are no restrictions on marriage and birth and all men are allowed and encouraged to marry and have as many children as possible.Women of the family marry out, and wives are brought in from other, non-related clans.Several of the main population checks employed in the Alps do not exist therefore in northern Albania, and there are two possible means whereby population was controlled: emigration and blood feud.
Blood feud must have had a tremendous eff ect on northern Albanian populations during some periods.Catholic Church documents indicate that in the period 1901-1905, the male death rate from feud in Shala stood at 26 (from Nopsca 1925cited in Coon 1950: 27;additional statistics in Whitaker 1968: 272-274).Feud was also the primary cause for emigration.Men would have fl ed the mountains not to marry, as was the case in the Alps, but to escape feuds.Th us, we need archaeological means of tracking not just population change, but feuding and emigration as well.
Christopher Boehm (1983Boehm ( , 1984aBoehm ( , 1984b) ) came to many similar conclusions about the Montenegrin tribal zone, which is in the same mountainous region as the northern Albanian tribal zone.As mentioned above, our application of the 'refuge area' hypothesis to Albania is drawn from Boehm's work.Boehm (1984b: 26) implies that Montenegrin and Albanian tribal systems are similar due to parallel cultural evolution.We fi nd this argument very diffi cult to accept given that the two cultures are nearly identical in all respects.Rather, we would argue that the two are similar because they share a common ancestral culture; whether Slavic or Albano-Illyrian is unclear.At some point the region was divided along a linguistic/ethnic/religious frontier (the Montenegrins are Orthodox Christian).Given that the Slavic migration began in the 6t century ad, this frontier was perhaps the fi rst to directly aff ect Shala and may have resulted from eff orts on the part of the Orthodox and Catholic churches to control the region, one of the earliest bastions of Christianity in Europe.We might also implicate the Byzantine and Venetian states, though the record of their impact in the tribal zone is less evident than that of the churches.
Th is 'ethnic frontier' has continued to aff ect the northern Albanians, Shala in particular, down to the present day.Th is was especially true after the formation of the Montenegrin state in the early to mid-19t century.Th e fi rst thing the 19t-century 'vladikas' did when they assumed control of the country was to end blood feuds and raiding, thereby forming a tribal coalition that could eff ectively resist the Turks (Boehm 1984a: 12-14).⁸Independence was achieved in 1878, but by then the tribal system had been completely disassembled.So, state formation in Montenegro happened at the expense of the Montenegrin tribes, and since state formation did not begin in Albania until 1912 (and was not really achieved until 1945 with installation of the Communist dictatorship), the Albanian tribal system survived into the 20t century.
As a consequence of events in Montenegro, in the 19t century northern Albanian tribes found themselves situated fi rmly between two predatory, enemy ⁸.Th e Montenegrin vladikas were paramount chiefs/bishops elected by a council of tribesmen and ecclesiastical leaders; in  the position of vladika was secularized and changed to that of an hereditary monarch.

Robert Schon & Michael L. Galaty 256
Diachronic Frontiers  states, where before there had been only one.Th e ethnic frontier was transformed into a political frontier, and the Montenegrins became a much more serious threat.Th e northern Albanian tribes appear to have maintained their independence from the Ottomans over several centuries through a wide variety of strategies (all of which fi t nicely the expectations of 'negotiated peripherality' in a 'refuge area'): they occasionally paid token taxes; sometimes agreed to fi ght alongside Turks, in particular if the enemy was Serb; and revolted when necessary, which typically entailed declaring a general besa.But with Montenegrin state formation and independence, Albanian tribes more often agreed to Ottoman demands, for money, troops, and peace.During this period bajraktars gained power at the expense of local chieftains and councils, throwing the tribal political system into disarray.Th e result was a spasm of violence (recorded by the Catholic church as a spike in death from feuding; see above), and a general weakening of the tribal defenses, at precisely the time the Albanian nation-state began to form.Th us, by the time of Zog, even truces between sworn enemy tribes, such as Shala and Shosh, could not save the tribal system from its dismemberment at the hands of the Tirana Government.It was the frontier that created the northern Albanian tribesman, and the transformation of the frontier that in the end sealed his fate.In order to test the above model it is necessary to deploy an archaeological survey methodology that can collect data relating to population, feud, emigration, ethnicity, and religion.Fortunately, our practice of archaeological survey is regional, interdisciplinary (combining traditional archaeological approaches, with ethnography and history), diachronic, and focused on the landscape and land use.Our methods are designed to locate artifacts, but also to identify and record architecture and oral and archival history.In two years of work, we have documented periods of village growth in Shala (primarily changes in the number of houses) that indicate increases in population at key moments of transition, always in response to changes in the frontier context.⁹Th ere also may have been expansions in the terrace and irrigation systems at times of population growth, though this conclusion awaits scientifi c testing.We have also identifi ed changes in house architecture (mentioned above) that correspond to decreases in blood feud beginning in the Zogist period.Emigration is much more diffi cult to view in the material record, but we do have some evidence for extra-valley contacts, particularly in the form of imported, exotic pottery.Th is at least indicates that Shala was not completely closed to the outside and that products, as well as ⁹.A detailed description of the timing of these village expansions, written by Wayne Lee, is available in the fi nal report of the SVP' s  fi eld season.http://www.millsaps.edu/svp/SVP%202005%20Final%20Report.pdf.
people, moved into and presumably out of the valley, possibly helping to blur ethnic and religious lines.
We also think that the movement of people, whether immigration or emigration, directly aff ected Shala because the valley served as an attractive, alternative transportation route, in particular for individuals who preferred to remain hidden from imperial eyes, such as raiders, traders, priests, shepherds, and smugglers.Th is may have begun as early as the Bronze Age and continued into the Medieval period, but was certainly the case during Ottoman and later times.Th e Shala route was used into recent times by herders and traders heading north to the market at Gusinjë, making Th eth somewhat of a crossroads, the place to stop and rest before tackling the Qafe e Pejës, the pass over the mountains.Th e route north was technically closed after wwi when Gusinjë was awarded to Montenegro, and completely closed after wwii.According to our elderly informants, the closing of the border in 1945 marked the most memorable event of their lives, more memorable even than the beginning or end of Communism.

conclusion: frontier life and archaeology
While our fi eldwork is still in its early stages, a few conclusions may nevertheless be drawn concerning the utility of our approach.Much of the previous narrative relies on ethnohistory, with material remains supplementing and fi netuning the picture oral and written sources provide.A pottery scatter, a wood plaque, or a fortifi ed hideout are isolated fi nds that add color to a preexisting picture, but say little on their own.Th is may be expected as they all appear within the bounds of living villages-for which oral and written history are preserved.
Th e strength of archaeological survey, however, lies in its ability to discern patterns in the longue durée (Braudel 1972).To meet this goal we focus on changes in land use and settlement patterns from prehistory to the present over a wide area.As of yet, we have not sampled a suffi ciently large universe nor discovered an adequate number of settlements to form concrete inferences concerning long-term patterns of change in these realms.As we expand into adjacent valley systems, we anticipate that meaningful, comparative examples of land use and settlement will emerge.Such patterns will not only complement the image we get from historical sources, but will also add elements about which history is mute.
Figure 2 -Map of Tracts and Structures Surveyed by the Shala Valley Project in 2005

Figure 3 -
Figure 3 -Map of Tracts and Structures Surveyed by the Shala Valley Project in 2006

Figure 4 -
Figure 4 -Middle Paleolithic Stone Tools from SVP Site 001red, burnished pottery that is similar to prehistoric pottery found elsewhere in northern Albania.At the very top of the cultural horizon we found a few small lumps of iron.Stone tools, including one fl ake of tan chert that may have been imported, and fragments of bone were found throughout.Th ere also was a substantial amount of charcoal, and we have submitted samples for radiocarbon dating. 5One interesting aspect of Site 006 is that it is located in a defensible, strategic position at a natural choke point in the valley.Th e site is protected on the west by the Th eth River, which passes through a deep gorge at this point, and steep cliff s.To the north of the site is a hill which is spanned by the remains of large, rubble walls.Th e eastern edge of the site appears to have been fortifi ed (Figures5 and 6), though we cannot yet be certain that the walls and terraces at the site are prehistoric.⁶Th e spur of land upon which the site is situated comes to a point at its southern end, which looks out over the lower part of the valley.Our working hypothesis is that Site 006 was a late prehistoric stronghold that controlled access to the northern reaches of Shala.It may be that the pastures at Th eth were intrinsically valuable and worth controlling.Or it may be that as early as the Bronze Age the Shala Valley constituted an important transportation route for people-shepherds, traders, warriors-moving back and forth between the valley of the Drin River and points north, in Montenegro.⁷Th us, it is possible that already during the late periods of prehistory and just prior to Roman conquest (the fi rst Roman incursions in northern Albania occurred in 229 bc), Shala fi lled a frontier position, perched between competing 'Illyrian' power centers to the south and west along the coast between Shkodër and Dalmatia and to the north and east in interior Montenegro and Kosova.Whatever the case, the valley appears to have been abandoned in Roman and View of SVP Site 006 from Above and to the EastByzantine times, fully so by the beginning of the Slavic migrations in the 6t century ad.

Figure 5 -
Figure 5 -Map of the Hinterland of SVP Site 006

Figure 6 -
Figure 6 -Map of the Immediate Vicinity of SVP Site 006 Figure 7 -Map of Structures 001-003 at SVP Site 006 Showing Location of Excavation Units and Shovel Tests Figure 8 -Map of Dakaj

Figure 1 :Figure 2 :Figure 3 :Figure 4 :Figure 5 :Figure 6 :Figure 7 :Figure 8 :
Figure 1: Map of Albania showing location of the Shala Valley (M.L. Galaty) Figure 2: Map of tracts and structures surveyed by the Shala Valley Project in 2005 (M.L. Galaty) Figure 3: Map of tracts surveyed by the Shala Valley Project in 2006 (M.L. Galaty) Figure 4: Middle Paleolithic stone tools from SVP Site 001 (A.Bushati) Figure 5: Map of the hinterland of SVP Site 006 (M.L. Galaty and R. Schon) Figure 6: Map of the immediate vicinity of SVP Site 006 (M.L. Galaty and R. Schon) Figure 7: Map of structures 001-003 at SVP Site 006 showing location of excavation units and shovel tests (M.L. Galaty and R. Schon) Figure 8: Map of Dakaj (M.L. Galaty)

Map of Albania Showing Location of the Shala Valley
In the Dukagjin region of northern Albania, oral customary laws regarding kinship relations and tribal political organization were codifi ed by a 15t century chief named Lekë Dukagjini.Th e Kanuni i Lekë Dukagjinit was not transcribed in full until the early 20t century by a Franciscan priest, Father Shtjefën ¹. See http://www.millsaps.edu/svp.
. SVP Sites  and  are Early Modern scatters of pottery associated with existing and ruined house compounds.SVP Site  is the Late Medieval fortress of Dakaj, discussed below.See Figures  and .
³⁴.A longer report on the excavations at SVP Site , with maps and photos of artifacts, is available at the project website: http://www.millsaps.edu/svp/SVP%20final%20report%202006.pdf Diachronic Frontiers 