Patterns in Time and the Tempo of Change: A North Atlantic Perspective on the Evolution of Complex Societies. more

Patterns in Time and the Tempo of Change: A North Atlantic Perspective on the Evolution of Complex Societies Kevin P. Smith Abstrøcl Between AD II75 and 1250 medieval lceland transformed itselffrom a network of decentralized simple chiefdoms into a unified proto-state. Uniquely, a vast corpus of vernacular writing-much wrilten by the chieftains themselves-describes actors' ideologies, histories, motivations, and understandings of the processes involved. Archaeological data provide alternative perspectives, highlighting processes thal extend over temporal scales beyond actors' abilities to observe or manage. How robust can our explanatory frameworks be if the changes we seek to expløin occur too rapidly to be monitored by most archaeological methods? Do archaeological perspectives provide valuable or illusory insights on the processes involved? Issues ofScale: Actors, Processes, Sources, and Change in Early Medieval lceland In operation and evolution of Icelandic chiefdoms, as interpreted and recorded by the actors involved, over a span of more than 150 years. The available texts include law codes, annals, poetry, histories, and sagas-extensive literary compositions retelling the accepted histories of regions, families, and individuals. The best-known among the latter are the Family Sagas,literaryhistorical works set several centuries before the dates oftheir composition (Clover and Lindow 1985; Kristjánsson 1997). However, the genre also includes the so-called Contempo- recent years, many archaeologists studying political change in complex societies have shifted course from a concern with processes operating on whole social groups or societies to perspectives that emphasize the motivations and dynamic interactions of the people who participated in these systems over long periods of time. The context of debate has expanded from one that viewed the individual as somewhat irrelevant to one that includes individuals as actors with motivating power who constitute important, if hard to recognize, creators of the archaeological record. I do not intend to review the various sides ofthis debate nor to describe the contributions of those who have addressed it so well (e.g. rary Sagas-less familiar, sweeping narratives which their composition (McGrew 1970; McGrew and de- scribe events that took place within the lifetimes of their authors or within the memories of those living at the time of Thomas 1974; Thorsson 1988). These latter sagas describe processes Bourdieu 1977;Brumfiel1992; Brumf,rel and Fox 1994; Dobres and Robb 2000; Giddens 1979; Pauketat 1994,2000). Rather, in keeping with the theme of this volume, my concern is with the very different temporal, spatial, social, and material-analytical scales these different approaches to perceiving the past require us to monitor and understand. To explore these issues, I focus on a relatively unique, important, and under-studied case of complex society transformations that took place in Iceland cluring the l2'l'-13th centuries AD. My goal in introducing this material is to examine how different the processes involved in the transformations would look when viewed from the perspective of the analyst confronting archaeological data and from the very differently scaled vantage points of actors who participated in the system and observed its operation. The archaeological and documentary records of early medieval Iceland (circa AD 1000-1264) should be important to anyone studying complex societies as this is one of the only cases, world-wide, in which chiefdom consolidation and state formation were described in written form, nearly contemporaneously, by participants directly involved in the processes involved. As in most chiefly societies, dytastic history and "singing the praise" of ruling elites were important forms of entertainrnent and legitimation in Icelancljc chieftains' households. Both orality (the poetic construction of history) and literacy became hallmarks of' elite culture, serving as potent symbols of rank and as vehiclcs for transmitting the esoteric knowledge maintained, used, and created by Norse chieftains and their peers (e.g. Helms 1979). One result of this linkage between literacy and power was the creation of a suite oftexts that describe in detail the internal of clrange in Iceland circq AD 1120-1280. The authors of many of these works are known, as are their sources. For example, Sturla Þórðarson (1214-1284), the most prolific author, was a paramount chieftain who lived in westem lceland. Through his contacts and alliances, he had access to the perspectives of all his peers and rivals. He and his family members were directly involved in the processes of centralization and state formation that he described. His father and uncles consolidated complex chiefdoms in westem and northem lceland. Sturla Þórðarson himself ruled one of them and his cousin unified Iceland in the 1240s, creating the frrst Icelandic proto-state. Yet, Sturla Þórðarson's family were among those who lost the most in 1264 when lceland was annexed by the Norwegian state. The narratives he, and other similar authors, created were, therefore, informed by personal knowledge and same-generation sources, yet were constrained from unbridled partisan advocacy by their creation for audiences that included other actors with first-hand knowledge of the same events, V/hile these sources are cer- tainly not fully objective, they are nonetheless remarkable for their detail, the comparability of events recorded in different accounts, and their even-handedness in recounting processes of trir"rmph and defeat. As indigenous sources of information on the intemal operation ancl cfynamic processes of ohange in chiefcloms, they are unparalleled. Early Medieval Iceland: Background Describing a complex historical reality at its broadest temporal and social scales, the story told by the medieval Icelandic authors is this: Iceland was discovered by Norwegian Vi83 î ltxplortng the Kole oÍ Analytlcat ùcar(j ln ¡\ItJllaculogrual llll.çrIJtçtauurr kings in the iate 9tl' century AD and w¿s settled primalily by elitõ refugees and their supporters fleeing the process ofstate formation in Norway. In Iceland they created a society led by chieftains and organized by a commonly accepted set of laws that established ground-rules for regulating conflict among chieftains and their supporters. These ground-rules seem to have been intended explicitly to prevent any one chicftain or other words, to protect the status quo of autonomous chiefly authority that had been violated by the creation of a state in Norway. householder had to declare his alliance with one of the chief- tains, but not necessarily the one living nearest to hilr. Householders could theoretically change their allegiancc, þy public declaration at an assembly, to other chieftains if theí felt that their interests were not being rnet, that their chieÈ tains were becoming too avaricious, or that they could not expect potent support on their behalf. The creation ofassemblies for resolving disputes, the establishment of chieftaincy faction from establishing hegemony over the others-in as a decentralízed, non-regional power structure, and the provisions for shifting alliances were all clearly intended to prevellt any one chieftain from consolidating power, while supporting the legitimacy of the chieftains' rule. Yet, despite the existence of leveling mechanisms to prevent the consolidation of authority into fewer hands, most of Iceland's districts were being transformecl by the late l2'l' cen- Icelandic chieftains (goðar) were recruited by birth from a relatively closed social class that viewed itself as distinct from the non-elite, to whom they were, however, related through more distant levels of kinship. They inherited titles to offices (goðorð) that existed independently of the individuals holding them. They lived in much larger houses, ate better food, drank more beer and wine, and canied more expensive weapons than the non-elite. They mobilized supporters for offensive and defensive action, held feasts, and had the right to intetpret law and execute justice through lethal means. They distributed prestige goods to their peers and supporters to maintain alliances and support networks. Most of these prestige goods were obtained ttu'ough trade with Europe and were secured for elite use through laws that gave chieftains the rights to purchase hrst from foreign merchants and to set prices on those merchants' remaining goods (Gelsinger 198 1). Chieftains and their kinsmen traveled overseas to learn and gain status in the courts of European kings, to reinforce their authorify through knowledge of distant places and ways, and to symbolize it by the gifts and styles they carried back from abroad (Bagge 1991). Through literacy and poetry they transmitted esoteric lore, including the law codes that dehned and supported their rights. tury into tenitorially bounded polities ruled by chieftains who had subordinated or eliminated their rivals (Sigurðsson 1989, 1999). By AD 1220 power in Iceland was held by six kindreds engaged in nearly constant civil war. Their leaders controlled complex chiefdoms that had absorbed formerly independent chieftaincies as subordinates, and in 1246, one of them, Þóröur kakali Sighvatsson, unified all of Iceland. He then established a court rnodeled on European royal exam- In AD 930, a coalition of these chieftains established an annual, national-level assernbly-the Alþing-where major disputes were settled and the laws were reviewed, as well as a series of regional assemblies for resolving minor disputes at the local level. By AD 965, thirty-nine decentralized chief- ples, appointed regional governors, and raised an armed bodyguard from the non-elite to enforce his interests, punish his enemies, and ensure tax collection. Leaders among the non-elite, who had held consensual authority among their peers in earlier centuries, were promoted to higher positions of authority within this system, while Þórður assumed sole authority tojudge disputes, approve localjudges, and enforce judgements-thereby subordinating the authority of the Alþing and establishing an official executive power for the first time in Iceland's history. Before this time, disputes could be judged according to the law but the responsibility to enforce judgements or take action against offenses was the duty of the aggrieved party, generally with the support of his chieftain, and was accomplished tluough blood-feud (Miller re90). Þórður kakali's polify endured while he occupied its highest seat, but fell apart when he traveled to Norway and never retumed. His potential heirs and enemies fought over its parts, briefly reconsolidating elements of his realm, until the entire island came under the control of the Norwegian state in 1264. What is important, however, is that once Þórður established the mechanisms, structures, and lexicons of state control in lceland, they endured among the elite. Each of his potential successors tried to re-establish his realm, or a realm modeled on it, rather than returning to earlier power structures. Iceland's non-elites, in contrast, resisted this process and sought to re-establish a model of less-centralized govelnance with greater local self-control. The incorporation of Iccland into the Norwegian state followed the rejection of Icelandic elites' olaims to succession by non-elites in two northem districts during the 1250s-those same districts ccntralization resulted ir'r the crcation of regional polities (ciricfdoms and complex chiefdoms) of cver-iucleasing size. Ultirnately, I suspect tlrat many, or most, chiefdoms known archaeologically went through phases of territorial integration matching the tladitional view of chiefdorns and tbrough periods ofpersonal or non-territorial integration (chieftaincies) as a process parallel to, or linked with, the "cycling" described by Wright (1984) and elaborated by Anderson (l 994). taincies were recognized in thirteen districts distributed around the country (Figure 1). Chieftaincies were defrned by law as constituting authorify over men (mannaþrrað), rather than rule over ten'itories.l Accordingly, each independent ' Tre tenns chiefdom, chieftaincy, and chieftain are not used sy'nonymously in this papcr. Chiefdont is used to reference the antlrropological construct first articulated by Oberg ( I 955), and since expanded and modified by many others (arnong them Canreiro l98l; Earle 1997; Service l97l; Wright 1984) to describe pre-state polities integrating many local cornmunities under one hereditary ruler's authority. The tem chiefdom caries, from its inception, assumptions that I question in the body of this paper about the territorial integrity of the polity under the chieftain's control. Chieftaíncy therefore represents an altemative, yet linked, perspective referring tìot only to the oflrce (in Icelandic gorlorð) hekl by a chieftain but also to a politically corporate, suprâ-regional polity comprised ol indìviduals allied, in a non- territorial way, to fhe chieftain. As an altentate fonn of regional polity linked to the chiefdor¡ concel)t, the chiel'taincy diflèrs subtly in its nontenitorial yet still supra-local irn¡rlications for power-holding and political :-^^--^-.^¡i- rr fncrÌr¡orâtiotì. lhe term chicftaitt (in iceiandicgoJi,pl. goöar) i<jontifios thc holder of such an oftice, whether his/her power base is ãrganized as a tenitorial polity (a chiefdom in the ..standar<ì', sense) or non-territorially, yet regionally, into an alliance network built around the leader (a chieftaincy in this paper's sense). ln lceland, chiefdonts a.i,d chieftaincles appeàr to liave co-existed at tirnes before AD I 150, employing the same lexiion and legitimating chartcrs for power-holding. After thai rirne processes of regiõnal 84 -*r$: fn. F ,:, werc arnong the first to pledge tribute to the king of Norway in return for his promises of internal peace. )y * scale, and consequences of conflict in this case of state for_ mation. I turn to this subject now in order to reflect on the motives, ,y f- It 1- ï $ å È Conceptualizing Change: Actors' Motivations vs. Archaeologists' Expectations neo-evolutionary terms, Icelandic society transformed itself between AD 1175 and 1220 from a confederation of simple chiefdoms-with one or two levels of decisionrnaking hierarchy above the household level-to a competitive network of complex chiefdoms. From that, it developed indigenously into a state with four levels of decision-making hierarchy and the elements of a bureaucracy. The transformation fi'om simple to complex chiefdoms took less than two generations; the evolution from complex chiefdoms to a unified state took less than one - in fact, Þórður kakali's unification and transformation of the system took less time than it takes to eam an undergraduate degree. The entire process occupied a temporal span half the duration of a common archaeological phase and less than the typical 1-sigma error of a standard radiocarbon date. It is unlikely that as archaeologists we could monitor the actual processes involved in the rise of complex chiefdoms or the state with changes occurring at this speed, yet there is nothing to suggest that Iceland's trajectory was unusual. The emergence of the Mississippian polity of Cahokia, for example, appears to have taken place within one political generation (Pauketat 1994) and is matched by similarly rapid trajectories in other parts of the In Negotiating in a World of Hurt: Conflict as process in Medieval lceland Tlre remarkable detail contained ín the Contemporary Sagas frequently allows the growth and development of episodes of conflict among chieftains across the island to be exarnined at temporal scales of a year or less and at social scales fìne enough to identify most of the major actors and their motivations. Figure 2 is a graphic representation of one such escalating conflict between two chiefly lines in westem Iceland. Each horizontal line represents one of the 36 original Icelandic chieftaincies (see Figure 1). Black circles and squares indicate aggressors in each incident of dispute; white circles or squares indicate the group attacked. Circles imply lethal conflicts, while squares designate episodes of legal dispute or non-lethal combat. Vertical lines link parties directly involved in the aggression - who actually attacked whom while diamonds identify chiefly allies who came to each party's defense. I want to draw attention to two characteristics of the temporal and social structure of aggressíon: feud chains and alliance cascades. The concept of afeud chain, coined by Jesse Byock (1982), reflects the social fact that scores are rarely evened in interpersonal conflict and that one event, lethal or legal, often leads to another. Such feud chains are seen in the enduring and cumulative nature of the conflicts represented in Figure 2. The incidents here include clashes between chieftains as well as hostile interactions among their non-elite suppotters into which elites were drawn by their obligations to provide support for their followers. This is an important point that I will return to shortly. In feud chains, small incidents built up cumulatively. Some were settled by legal means (primarily the exchange of valuables) and others were resolved by lethal combat; yet the score was never really evencd. Defeats in legal disputes eroded chieftains' reputations for being able to defend their followers and required honor to be avenged, while wounds and deaths called out for revenge. Although the number of people involved in each incident could be iy IE .o r le Ì ç rt ttIS r11 Í -X rS v te .e t,d h re ir IS Le )e Ist re ie rf f)f it ir "S Mississippi valley (Brain 1989; Knight 1997; Knight and Steponaitis 1998; Steponaitis 1998; Pauketat 2000). Similarly, where tight chronological control over the processes can be obtained, the rise and dl.namic cycling of primary states in Mesoamerica (Flannery and Marcus 1983; Joyce 2000; Kowaleski 1990; Marcus 1998; Schele 1990), Egypt (Kemp 1989; Lehner 2000), and elsewhere (Spencer 1990) appears to have been similarly rapid. V/hat ls unusual about the Icelandic case of state fonnation is that the available documentation allows us to understand the internal dynamics at very much flrner social and chronological scales than are available in other parts of the world where states arose. .e re lr .e .s n iS rf rf o .s ; ,h d Although many different theories have been proposed to account for developmental processes in complex societies, including the rise of states, conflict theories (e.g. Cameiro 1970,1977,1981) loom large and would seem to be appropriate in considering the Icelandic case. Conflict was obviously a central concern of the Icelandic elite. The processes of aentralízation and political consolidation were to a iarge degree promoted and sustained by the twin powers of the sword (conflict) and the feast (alliance). In sorne cases, the sword enforced what the feast could not, while at other times the feast consolidated and rectified what the sword had accornplished. The literature of medieval Iceland is a literature of conflict, written by elites for elites, and about elite conoerns. Af times the theme is conflict as a form of resolution, at other times the theme is the resolution of unchecked aggression; but conflict forms tho core of the nanatives whether sagas or law cocles are considered (Byook 1982; Miller 1990). What motivated conl'lict, how it was prosecuted, how it developed, and what it achieved can be examined in detail in these sources, since this was what interested the sagas' authors most. Although the Icelandic data allow other theoretical perspectives to be examined in equal detail, 85 small-fewer than l0 people on a side in most of these events-their significance loomed larger as incidents accumulated. One of the cumulative effects of feud chains was the unleashing of alliance cascqdes. Vy'hen local disputes between chieftains and their factions intensified, allies secured over the years by friendship, gifts of prestige goods, and marriages were calied upon for support. As each chicftain tried to gain advantage or build enough support to subordinate or annihilate his rivals, the number of allies called into the fray increased. When allies lost men or accumulal.ed wounds, enmity deepened and spread ovet'wider areas of the countly, tuming what was onÇe a local dispute into a regional conflict or wicler teusion zone. The feud chain seen itr F-igure 2 starts with a casc of petry thelt involving two chieftains' followers and ends 38 years later with the involvement of chieftains drawn fi'om all parts of the island, the absorption of onc chieftaincy by another, and the formation of the core of a regional cornplex chieftlom that endured long after *,-r------o -- of chieftains, which drew the noose tighter each time they were mobilized and eventually threatened to tum every local conflict into a regional conflagration. A concrete example from western Iceland may, perhaps, provide a better sense of the mechanisms involved in these processes, the scales at which they operated, and their consequences at both the local and the regional levels. Wood, Wounds, and Death: Actorst Perspectives on Social Change in Saga (McGrew 1970; McGrew and Thomas 1974; Sturlttnga ThorssJn t 918) during the late- i 2'r' tluough early-1 3tl' centuries (Figure 3). Allies called upon for support expected support in return. Consequently, alliance cascades had the poteniial to create polarrzed factions or competing confederacies when compared to the wider-scale picture of concurt'ent cou- this dispute. As cornplex as this sequence looks, it pales flicts among chicftains and their suppolters described lì'orn the enemies of the other coalition's allies. As a result, each coalition member's obligations to provide lethal support was balanced against the fear of acting rashly, knowing that reprisals would ensue in their own districts for wounds give¡ in this dispute. Ultimately, the chieftains played political chess so well that the legal cases deadlocked at the Alþing. Both sides had balanced tlueats against obligations so skillfully that no one was able to gain supremacy and no one was going to get satisfaction. Rather than the affair being settled at the assembly, Þórð the Red's frustration grew until he at- tacked Hámund's chieftain, Þórð Sturluson, with an axe. Fighting ensued between the factions at the assembly. Four men were killed and many were wounded. Two of the dead came from each faction, but three of them were actually suppofiers of chieftains from distant districts who had come in as allies of the principals and who had not in any way been implicated in the original feud (Thorsson 1988:184-185). V/hile the local dispute was eventually resolved and compensation was obtained for the four deaths, enmity between the coalitions grew. What started as a small-scale, local conflict became, through the process of alliance cascading, a tactical struggle between members of two regionally organized strategic blocks that were growing and becoming stronger through repeated examples of such small-scale feuds. These, in turn, stafied when non-elite householders needed the support of elite backers to prosecute their local disputes. They became extended when the chieftains used these local disputes to demonstrate their abilities to defend their supporters and to advance their own competitive agendas against their rivals. Finally they expanded across the island as the allies of allies and enernies of enemies were drawn into what began as local affairs through the cascading obligations imposed by alliances. In AD 1195, two farmers in westem Iceland (Þórð the Red and Hámund Gilsson) argued about who was cutting more brushwood from a hillside they owned in common. It led to no violence, but soured feelings between the neighbors. Some time later, a young man took a walk with a young woman. Her father objected and ran after them. The young man, possibly unaware that her father was behind, threw his axe over his shoulder and gouged her father's eye out. The young man was a friend of Þórð the Red, and the father was Hámund's kinsman. Bittemess grew when the young man's family refused to compensate the girl's father for his lost eye. Some months passed and one day the young man and two of his friends (one of them being Þórð the Red's son) rode down the valley to visit a ship berthed nearby. As they passed Hámund's farrn, I{ámund, one of his workmen, and the wounded father attacked them. In the ensuing frght, Hámund, Þórð the Red's son, and two others were badly wounded. As this feud chain grew, both sides knew that more violence was certain and sought support from their chieftains. Hámund called on Þórð Sturluson-a member of a powerful western Icelandic chiefly family-to give him aid. Þórð Sturluson sent his brother with a contingent of men to stay at Hámund's home all winter. No further violence occurred that season. However, Þórð the Red was a suppofier (þingmaður) of a different chieftain, Kolbein the Young, one of lceland's most powerful paramounts, whose power base was in northetn lceland. In the summer, both chiefly factions gathered allies and took their cases to the Alþing, hoping to use nonlethal means to end the dispute. Þórð Sturluson exploited his kinship ties to build a coalition of six paramounts and chieftains from northeln, southem, and western lceland. While I{ámund sought support from Þórð Sturluson to obtain compensation for the wounds he and his fr.iends had suffered, Þórð Sturluson hoped to use this case to counteract Kolbein the Young's growing support network in westem lceland. Kolbein, for his part, callcd in sup¡rort from his own allies, all paramount chicflains fiom nofihern and sot¡thern Iceland, in order to defend ltis þingntaöur anci cxtencl iris sphere of influence. What were the outcomes of this conflict? Men certainly died and goods definitely changed hands in compensation, troops were mustered and tensions escalated both within factions and across regions-but at the end of the contest no chiefdom expanded at the expense of another, no one gained greater access to high-value resources, and no one controlled a greater share of the system's prestige goods. Contrary to many archaeological rnodels, the goals of the conflict were never to capture resources and its cause was neither environmental nor social circumscription. This conflict, like most others described in medieval Icelandic sources, had its roots in interpersonal problerns that are cofilmon in isolated rural settings: accusations of theft, seduction, loss of family honor, minor boundary disputes, and unintended accidents, mixed with personality conflicts, family hatreds, obstinacy, and a willingness to use violence to redress old affronts. For the non-elites, the affair met their goals: vengeance was obtained, rights were upheld, and justice was served. The chieftains involved, in tum, maintained their status and gained authority among their followers by providing potent suppod and fulfilling the obligations that were the flip side of commoners' agl'eements to support thern. As this classic case of an alliance cascade developing from a feud chain expancled, each side strategically called in favors 86 i use the tcrm aglcement intcntionaily because the non-eiite were always far lnore nulrrerous than the chieftains and their families. Lacking overwheltningly better annaments, it was always possible, at least in theory, to overthrow or ignore eiite claims to rule. However, through their wealth, accepted authority, distant kinship corulections, and wider networks of support, Icelandic elites cleally had something concrete to offer-the ability to gather enough force to uphold their followers' rights against their neighbors and to maintain the pcace when peace could not be regained at the local level. This was a concrete ancl indisputably valuable reason for accepting and supporting the asymmetries of the system and ultimately formed the basis for sustained elite power in lce- tackecl one after another ofthe rival centers ofchiefly power, driving out their'leaders, killing their supp<lrters, and bring- ing their territories under his control. He then began to re_ make the society's ruling apparatus in the image of a Scanclinavian royal court, establishing the first Icelandic proto-state. Modeling Change: Data, Processes, and Issues of Scale The Icelandic documentary data suggest that the process and progress of state formation and chiefdom consolidation proceeded aI temporal scales faster than those we can normally monitor through archaeological means. It also suggests that the factors motivating social change and considered important to the actors were very different from those that are central to most of the explanatory models we employ. Many of these processes also took place at social scales too finegrained or too broadly spread for us to see easily. In fact, how much of this would we see archaeologically? How accurate would organizational models be that we might deduce from analyses of this society's archaeological record? What would we be likely to suggest were important variables generating social change? Would they coincide with the factors that actually motivated Icelandic actors' behavior? Archaeology in Iceland is still relatively young and this period has been relatively neglected, but by combining archaeological and historical data we can get a good sense of regional settlement patteming, comparable to regional surveys and suitable for modeling. Other data in the archaeological record suggest long-term patterns worthy of consideration, but would they have been perceived as significant by the actors land (and perhaps elsewhere)-not knowledge of distant places, esoteric rituals about the unfathomable, control of exotic goods, or similar symbols of elite rule. Most of these latter social phenomena were also present in lceland, and served as inter-elite currencies to establish and rnaintain alliances or to strengthen elites' claims for legitimate succession to office, but they appear to have been mechanisms for maintaining po'wer rather than causative factors for the acceptance of sustained power differentials or for transformative changes in the socio-political system. Ultimately, this minor case-one of hundreds described in the Contemporary Sagas-started with a non-lethal dispute among neighbors over brushwood, developed into a lethal local feud, and led, after it expanded to draw in people from all parts of Iceland, to a marriage between two elite kindreds on one side of the dispute and the formation of tightly knit coalitions among distant chiefly houses on both sides. Such coalitions, forged through mutual needs for support and bound ever more strongly together by accumulating shared resentments, guided the trajectories along which conflict and regional political consolidation took place over the next 50 years. As chieftaincies declined in number and more ofhces whose lives were trends? far shorter than the duration of these were controlled directly or indirectly by a small number of emerging paramounts, intermarriage among their houses tended to bring alliance cascades into play faster, over wider regions of the island, and with higher stakes. As disputes alnong these paramounts strengthened in intensity and bitterness, more often than ever before the only possible resolution for breaking the chains of feud and averting the cascading effects that turned local disputes into regional conflicts was to attack one's rival directly with the intent to kill or to drive him from the country. In so doing, a sphere of influence or a polity was left without a leader. Yet, despite the expectations we might have that such situations would provide opportunities for the victorious polity to expand, chieftains more often retreated without annexing these vacant areas or held them very briefly. It is clear that one reason for this instability and inability to expand polities was that the victorious chieftains' titles, genealogies, and past alliances provided them with no accepted basis for receiving the supporf of locál non-elites. Further, the territories in question were often too distant to be controlled easily or directly from their estates. Not infrequently, the void would be fillcd by a kinsman of the recently deposed leader and the cycle of rising tension, feuding, and alliance furmation would start again. The economic basis of medieval Icelandic households was a combination of northem European pastoralism coupled with fishing and wild resource acquisition (Amorosi 1992; McGovern 1990; McGovem et al. 2001; Smith 1995;Zutter 1992). Excavations, surveys, and historical research concur that the structure of settlement in medieval lceland consisted of dispersed, unfortif,red farmsteads with enclosed hay-fields and outbuildings, similar to the present-day settlement pattem in rural Iceland (Smith and Parsons 1989; Vésteinsson 1998). These farmsteads, privately owned as alienable property (e.g. Earle 2000) and normally separated by half a kilometer or more, were occupied by single or joint households, often with a few servants (Miller 1990). Archaeologicai data suggest that median household size doubled from the time Iceland was settied to the l3th century (Smith nd). A bi- or tri-partite division of household sizes in the 12th-13'h century data can reasonably be associated with the residences of common farmers, estates of local non-elite leaders, and elite compounds. In 1242, Þórður kakali Sighvatsson set out explicitly to break this cycle, to avenge his father and brothers who were victims of its violence, and to reclairn the chieftaincy that was his birthright. In a sweeping set of campaigns spanning just four years he gathered the supporters of his deceased family members as well as disaffected subordinate elites and at87 No villages, towns, cities, or other nucleated population centers existed in Iceland until the 18tl'cenfury' With one r¡ndated exception, regional refuges or hillforts are absent. Neither elito nor non-elite residential complexes were iocatcd in clearly defensible looations. Doculentary sources indicate that sornc elite residential compounds were tcmporarily fortified with turf walis, 3-4 meters tall. These were erected in times of hostility and were removed when the conflict ended' some of these fortifications stood for In the 13tl' ""nirry, sites were apparently never fortified' decades, yet other elite Explor.ing thc Role of Analytical Scale in Archaeological lnterpretatton non-elite leaders, and even on some small farms' of varying sizes were found in chiefcomplexes' on the estates of rnany local tains' residential Churches or chapels The distribution of residential sites in the Borgarfiörður district of westem lceland, circa AD 1180, is shown in Figure 4. This reconstruction is based on a combination of medieval ,our.", and archaeological research'2 Although many more non-elite fatms were tikely to have been occupied at this time than are shown in this view, Figure 4 provides an accurate view of the distribution of elite sites and most of the larger non-elite sites across Borgarfiörður 20-30 years before the regional consolidation of chiefly authority' Ur"rfortunately, this reconstruction bears littlc, if any, reser¡blance to the documented organization of the regional system in the late 12'h century. As Figure 6 shows, thc district was split by one ofthe four boundaries that separated lceland into four Quarters. Until circa 1150, ohieftains and their follow- ers living northwest of this boundary met each spring at an assembly site located on the norlhern edge of the district's Iowlands, while chieftains and their supporters living south of this boundary met each year near Reykjavík, far to the south. The assembly site for the two chieftains residing at the In Figure 5 the information presented in Figure 4 has been recast to examine patterns of regional political control that we rnight reasonably infer for this region employing simple models based on central place theory that have been used to arølyze settlement pattern data in complex societies from many paÍs of the world (e.g. Arnold 1997; Peebles and Kus 1977; Renfrew 1974,1975; Renfrew and Level 1979; Scull 1999), including Viking Age Scandinavia (Thurston 1996). Thiessen polygons (Hodder and Orton 1976) were drawn around this region's seven contemporary elite residential complexes to predict polities from political centers. This exercise suggests a relatively even spacing among chiefly centers, comparably sized territories, and roughly similar numbers of non-elite farmsteads in each one. Each chiefly center is located approximately one day's journey on foot or with pack horses from the next, as would be predicted in models that integrate economic proccsses with settlement structure in chiefdoms (Peebles and Kus 1977). Each pte' dicted territory incorporates some areas of habitable lowland pasture, upland grazing ranges, river segments, and fi'eshwater lakes. Each territory's settlement pattenì includes sites of three sizes, the two largest of which contain sacred structures, suggesting a one- or two-tier control hierarchy above the level of the household (Wright 1984). From these patterns, it might well be concluded that the region contained a network of competing simple chiefdoms or perhaps emergent complex chiefdoms. Among these, one territory (A) contains all three of the assembly sites known to have existed in the westernmost edge of the district was located in Breiðafiöröur, the next fiord system to the north. Theoretically, the Quarter system both freed and constrained alliances between elites and non-elites. Householders were required to ally themselves with one of the 9-12 chieftains living in their Quarter but were not required to ally themselves with the chieftain living nearest to them: chieftaincy gave authority over men, not control over territories. However, as the dispute over brushwood showed, by the late 12th century some farmers established alliances with chieftains from more distant regions, beyond their Quarters. Þórð the Red [1], for example, was a þingmqður of Kolbein the Young, from lceIand's North Quarter. Hámund's [2] chieftain, Þórð Sturluson, lived in the West Quarter. Both Hámund and Þórð the Red lived in the South Quarter, Neither Hámund nor Þórð the Red were allied to the chieftain in whose Thiessen polygon-defined territory their homes were located, because political leadership itself was not teritorially defined. In this caso, where regional patterns of alliance and power are extremely well-documented, it is clear that modeling tools such as Thiessen polygons could easily lead us to reach spurious conclusions conceming both the economic and political foundations of elite control and also non-elite community organization. district, as well as its main harbor, to which foreign merchants came each summer carrying prestige and domestic goods. A second territory (B) controls access to this harbor from the sca and commands th¡ee subsidiary landing sites. We might well predict that one of these two polìties would emerge as a paramountcy through its control over access to exotic goods, its incorporation of the most extensive areas of Iowland pasture, its access to the diverse resources of the coastal/ terrestrial interface, its incorporation ofthe district's initial elite settlement (in B), or the presence in its territory of assembly sites, which were symbolically, ritually, and politically-charged locations withjn the lanclscape. 2 The distribution of substantial non-elite farms with attached churches might, for example, suggest a subordinate level of administrative control within each modeled chiefly territory. However, circa 7180, the distribution of such farms actually reflects the division of the countryside into local communities (hreppar) of non-elite farmers established for mutual support and the coordination of seasonal activities. These communities' leaders were elected by their non-elite peers and frequently met at the church-farms' Although chieftains may have exetted considerable control over the hreppar in which their estates were locatecl (Sigurðsson 1999), the hreppur system was, in theory at least, independent of elite control prior to the period of regional consolidation (SteinWilkshuis 1982). As a lower-level, egalitarian political structure existing side-by-side with the hierarchical networks connecting farmers to chieftains, the hreppur system demonstrates the simultaneous operation of heterarchical and hier- I)ata used to gencrate Figure 4 include the names ofall farms recorded iu medieval clocuments for this region, combined with ìnformation gathered lrorn archaeologícal research conducted rvit'hin the mapped area (for this, see especially Olafsson 1996). Although lìo area as largc as this--,the figure oovers an area of roughly 6,500 km'?-has yet beou surveyed in lceland, the general pattenìing of sites in the densest areas shown in Figure 4 is comparable to settlement densities and patterning reconstructed from postrnedieval censuses and archaeological surveys tlìroughout the country (e.g. Gestsson and Briem 1954; Lirusson 1967; ólaßson 1996; Rafnsson 1990; I 99 Sveinbjamardóttir 1992; Vésteiusson g). archical principles within this society, as has been documented in many other prehistoric chiefdoms and states (e.g. Earle 1977,1978; Ehrenreich et al' i995; Joyce and Henclon 2000; Mehrer 2000). However, it is unlikely that the specific interlocking dirneusions of either structuring principle could be accurately identif,red in this setting r¡'ithout the documetltary clues. The assumption that hierarchy means territorial controi (implicit in our use of terms like "centers" and the application of modeling tools such as Thiessen polygons to our data) would, in this case, reduce the actual complexity of this society-with its non-tenitorial hierarchies and territori88 g w fi $ Ë ally defined non-elite communities-to a rnap that is admittedly satisfying but nonetheless grossly inaccurate in representing the society's organization and dynamic structure' I I the church farms in this district were directly controlled by Reykholt and much of the revenue they collected as tithes may have escaped its control. However, while this map of Borgar{örður (Figure 8) suggests only one level of elite contlol over the district's many households, it only tells part of the story. Reykholt ín 1240 was the center of a complex chiefdom, but it was one that extended over a much wider region, in which Borgarfiörður was just a part (Figure 9). In addition to the six titles he held in Borgarfiörður, Snorri Sturluson controlled the titles to at after regional political consolidation had taken place' AII of the elite residential complexes present in 1180 were still occupied and the time elapsed between Figures 6 and 7 is less than the use-histories of known elite structures in Iceland. It is therefore likely that these sites would have maintained largely the same form and layout in 1240 as they had 1180, Figure 7 shows the same district 60 years later, around 1240, even though the political status of their occupants had changed and they were no longer the residential bases of independent chieftains. However, given that these sites would still look, archaeologically, like chiefly centers, the resulting regional pattem of very large paramount chieftains' compounds, somewhat smaller elite estates, intermediatesized church-farms, and smaller, non-elite farmsteads would probably suggest that the region was divided between two complex chiefdoms, each with at least three levels of administrative control above the household. The complex chiefdom centered on Reykholt would appear, in this reconstruction, to have had within its territory a population of perhaps 8,00011,000 people distributed among one paramount's compound, five subordinate elite centers, at least 16 churchfarms, and a much larger number of non-elite farms' The territories dehned by the Thiessen polygon boundaty drawn between the centers of the region's two apparent complex chiefdoms include harbors, assembly sites (one for the westem polity is located just north of the boundary of Figure 7), and extensive areas ofboth lowland and upland zones' From this information, it might be suggested that control over basic agricultural resources, assembly sites, and trade were essential to the emergence of Icelandic complex societies. While this is likely, to some degree, it is worth noting that the centers which actually did gain regional power by 1240 (Figure 5, C and G) were not those (Figure 5, A and B) that had the district's harbors, assembly sites, and most-extensive gtazing lands within their Thiessen polygon-defrned territories, circa 1180. This suggests that factors other than exchange, proximity to symbolically charged political sites, or agricultural potential actually guided the processes by which chieftains gained and maintained political power, which accords with the Icelandic actors' perspectives, described above. least seven other chieftaincies in northern, western, and southem Iceland. Other chieftains, themselves holders of regional power and generally more than one chieftaincy, served him as subordinate allies, although they remained theoretically independent. ln less than a generation, Reykholt's paramount had gained hegemony over most of western Iceland, an area occupied by perhaps 30,000 people, not just the district suruounding his residential complex. A few years later, Reykholt's domain was incorporated into a still larger paramountcy and ten years after that the site had become the seat of a local govemor in a newly unified Icelandic state. Although the site remained an elite residence throughout this process, the positions of its occupants in regional and islandwide political hierarchies had changed faster than the tlpical construction histories responsible for producing two of the most reliably measurable and most frequently employed ar- chaeological correlates and residence sizes. of hierarchical relationships-site Would Snorri's vast cornplex chiefdotn, or these processes of change, be archaeologically visible as such? The continued occupation of standing elite structures at each of the five subordinate chiefly centers within his own home district might encourage archaeologists to suggest models of shared power, competition among equals, or peer-polity interaction (sensø Renfrew and Cherry 1986) within the Borgarfiörður district. The presence of comparable site hierarchies in each of the teruitories he controlled would make it hard to recognize them archaeologically as parts of his domain. Archaeo- logical reconstructions incapable of monitoring processes Overall, the reconstructed regional pattern in Figure 7 is intuitively satisfying. Again, however, the contrast with the historically recorded system is critically important (l-igure 8). Rather than being secondary elite centers, each of the that played out over such vast regions and at temporal scales more rapid and more f,rne-grained than either architectural changes within sites or the resolution of standard dating tools might well see no differences in regional settiement pattems during the 60 year period covered by Figures 6-9. 'I'he Shape of Time: Archaeological Realities and Cultural Dynamics former chieftains' estates within Reykholt's territory had actually been subsumed within its home territory and had been demoted to the status of local non-elite church-fanns after the chieftain residing at Reykholt, Snori Sturluson (AD 1119-1241), consolidated power in the district, circa 120310. Snor¡i appears to have held all of thesc titles himself or managed them on behalf of those who had previously held them tluough hereditary right (lngvarsson 1987). The earlier division of the district bctween two Quarters was abolished and the assembiy site for the West Quarter, which had been moved to the center of the district around 1150, became the regional assembly site for Snorri's polity (Jóhanltesson 1974:77). Reykholt itself owned many outlying farrns directly, managing them as components of its estate, but few of 89 I have asked throughout this exercise whether we, as archaeologists, could recognize, tkough settlement pattern analyses, the correlates of social changes that took place over large areas and at temporal scales measuring 60 years or less' \\¡hat just palt of one? Would we sec any of this compiexity or interpret it accurately? In medieval lceland, given the long use-lives of the buildings involved and the shorter temporal scales at which political proçesses spun themselves out, we would probably miss phases 75-100 ycars, or longer, in duration? What would we see if our st¡rvey domains took in only one valley system, or if we had an archaeological record builf of typical D^VLvLlL'b turf defenmost of what we seek to understand' If tclnporary be locatecl or had bee' removed when sìlr" *utts could 'ot dung"rou, conditions disappeared, we--would.probably only o".äiu" a pattern of larger and smaller undefended färm.t"ads, uo.yìng in scale through a continuum, with some at We each level it-ing ritual facilities (churches ancl chapels)' tugg"tt this was a minimally ranked society with might well juJ otre levei of control (farms with churches) above the household. Differences in the sizes of these farms might well be used to infer differences in wealth, but perhaps not evidence for differential political control. f the fofiifications were apparent and if each elite household was fortiflted, we could pèrhaps identify a two-tier hierarchy with political dimensions, as well as wealth-based ranking, but we would probably not be able to monitor the changes that transformed ihe regional system since the structures elites built were likely to endure through more than one phase, even if the elite status ofthe site's occupants did not. rnental degradation due to human activities (Dugmore and Buckland 1991). Each of these patterned sets of relationships resonates well with argurnents that archaeologists have used to explain trajectories of evolutionary change in cornplex societies. They include the carrots of expanding labor forces and increased access to exotic goods, as well as the sticks of climatic change, ecological degradation, and population increases in circumscribed areas. However, what is interesting to note is that none of these factors were identified by the Icelandic actors/authors as having had any direct, perceived, or important roles as causes for conflict, political consolidation, or systernic change. When historically reported years of good and bad harvests, famines, or other stresses are plotted against the annual record of conflict and consolidation from Iceland, it is clear that conflict happened in both good years and bad. There is not a single episode in lhe Contemporary Sagas where a chieftain or his followers set off intentionally to defeat a rival and annex his land in order to offset local resource shortages. Simiiarly, chieftains had the right to make first purchases and to set prices at markets in their districts, and clearly they used those advantages to supply themselves with prestige goods (Gelsinger 1981). Yet there are no instances in which the expressed goal of a raid or campaign was to establish exclusive control over a harbor. On the other hand, there are examples in which chieftains' efforts to exert such control led foreign merchants to pack up and sail away to other districts with less arrogant leaders. Would we correctly understand the role of conflict in this society? With no regional mustering places, no central fortresses, isolated fatmsteads set in non-defensible locations, historic sources that indicate no campaigns of regional destruction, and only temporary walls of turf thrown up around household compounds too small to shelter regional populations, it would be hard to conclude that conflict was an active force leading to regional consolidation, as the documentary soul'ces imply. The image of a relatively peaceful, barely hierarchical society that we would be likely to derive from the archaeological record would clearly be at odds with our expectations for a system of complex chiefdoms developing into a state, especially one in which conflict was an important causal element of that process. Yet, that is what the indigenous sources tell us happened. lceland's archaeological record, can we be so sure that we are really seeing what we hope to be understanding in other archaeological cases where no written records exist to help us refine our expectations? Is it possible that sotne of our models, designed to monitor structure and change on temporal and social scales so different from those in which actors operated (even when our models are actor-centered), could lead us away from understanding the patterning of the past? What long-term pattems do we see in the archaeological record of Sturlung Age Iceland? Data drawn from excavations and suweys undertakcn across the country suggest a slow and incremental growth in household size and also an increasing range of household sizes (Smith nd). Surveys indicate more farms were built than abandoned in the upland zones, implying an increasing population and dynamically fluctuating settlement areas. We see a shift in herding strategies towards a greater emphasis on sheep and less concern with cattle (Amorosi 1991). We see increases tluough time in the frequency of imported luxury and don-iestic goods in household assemblages, with some fluctuations in access to foreign goods around the tirne of chiefdom consolidation, There ale suggestions that rnore petmanenl and corr¡rlex stnrctures were built in at least one harbor after I 150 (Herrnanns-Auðardóttir 1986, 1999). We see a slow deterioration in regional clirnatic conditions, but not yet the onset of the Little Ice Age (Ogilvie 1991), plus indicators ofincreased erosion and environ- If it would be so easy to misinterpret The one-way street we often paint in theoretical models seems more often to have run two ways when viewed from an actor-scaled, rather than a system-level, perspective. Chieftains, even those who could claim hegemony over 30,000 people and nearly one-quarter of the country' were constrained by the realily that actors at almost every level of the system had different goals than theirs. Snorri himself was assassinated by a coalition of his former sons-in-law who had many axes to grind-and did; while around 1255 nonelite farmers in nofihem Iceland walked away from chieftains and rcjected chieftaincy altogether. The system itseli a network of competing and allied interests, worked at very different scales depending on from whose perspective it was seen, yet activities undefiaken at each level of scale had the ability to impact every other scaled subsystem according to their intersecting logic. This implies, as Flannery (1999) suggests, that neither actor-based nor processual/systemic models are likely to have sufficient explanatory power alone to illuminate complex plocesses of social change in the past, but are more profitably seen as complementary approaches to building complex exPlanations. Seen fi'om the actors' perspective, processes of political change in rnedieval Iceland seem to confirm the American politician 'fip O'Neill's femous dictum that all politics are ultirnately local. The nextts of interactions and mutual obligations betweeu chieftains and their far more nutnerous llonelite supporters-that is, what chieflains actually did-was mole irnportant for perpctuating the system and rnotivating social change than what they ate, exchanged, built, or had accompanying them in death. The daily logic of the system' characterized by the intersecting and often contradictory short-term goals of interacting and variously obligated ac90 tors, was a rnore potent vehicle for change than transgenerational, incremental "systemic" adaptations or slowly developing trends in phenomena external to the systern of social interaction itself, even if those trends and adaptations helped to set the stage on which change occurred. Unfortunately for us, that relegates most of what we can recover, count, and analyze to the category of contextual rather than causal information, suggests that archaeological research complex chiefdoms in others, and complete transfonnations of the system at still other tirnes. to explain why sirnilar actions undertaken at different points in time led to quite different results: cycling of simpleìhief_ doms during some per.iods, the appearance and cóllapse of more often provides us with tools to monitor the consequences of change than to understand its causes, and implies that in many cases we give objects and other things that we €xcavate far more importance as motivating forces than they deserve in the processes we hope to understand. The social, spatial, and temporal scales at which we build our models directly influence the ways we expect the past to have been, color the answers we get back fiom our studies, and lead us to accept some explanations for why change happens rather than seeking alternatives. Models such as those that posit conflict as a prime mover in the evolution of complex societies generally presume longer than subgenerational temporal scales for processes to play out and describe abstract social groups, rather than individuats (at Conclusions What can we conclude from this examination of political change in medieval Iceland? First, the Icelandic documentary material implies that actor-centered perspectives on social change are ultimately likely to be the most appropriate tools for understanding why, how, and when major social transformations took place, and for interpreting ways in which people adapted to and used those conditions of change to create the social landscapes in which they operated. However, currently we can only rarely resolve chronologies or material patterning to the extremely fine-grained resolution necessary for monitoring actors archaeologically at temporal scales similar to those in which critical actions took placeperiods of years or decades at the longest. Consequently, most actor-based theories are probably unverifiable and unfalsifiable using current archaeological data. As a result, one has least non-elite individuals), as the agents thlough which change occurs (Cameiro 1970, 1981). Chieftains represented as social types rather than individual people become strategists engaged in trans-generational campaigns to capture resources, with the assumption that their authority was felt in bounded polities into which non-elites were slotted as passive pawns. The Icelandic data, on the other hand, suggest that the social scale at which political actions took place was very personal and fast-paced. Regional crises could tum on the obligations that existed between a chieftain and any one of his followers. And, most critically, the Icelandic situation implies that non-elites' interests and activities were as essential to processes of political change as were the elites' (cf. Pauketat 2000). Successful modeling of political change may therefore require multi-centric approaihes for understanding the past, rather than a choice of "top-down" or "bottom-up" models. to question whether such approaches emphasize current theorists' roles as actors more than they reflect the interests ofthe actors being theorized about.3 Conversely, the perspectives gained from observing longterm or system-level patterning in the archaeological record appear to highlight processes that happened too subtly, too slowly, over periods of time too long, and across areas too extensive to have been perceived as important by most actors in past societies. Subtle changes through time in the scale and organization ofhouseholds, variations in the frequencies with which exotic goods passed through the system, shifts in the proportions of cattle or sheep herded, or fluctuations in the condition of regional resource bases appear to have occurred at temporal scales too much longer than individual lives to have been observed by participants in the system or to have been motivations for political action by real actors. 'While these archaeologically defined processes and patterns are important, the Icelandíc data suggest that actions leading to feud chains, alliance cascades, and transformative processes of political change found their origins and motivations in conflicts and events that would have happened and did happen regardless of these external factors. This suggests that much of what we cun'ently meaEure as archaeological data may record the contextual fabric of life during periods of change, without being directly related to the caøs¿s of change. It may be, however, that this changing context helps 3 Re-envisioning the concept of the chiefdom from one that necessarily equates chieftains with control over territories to one that sees them occupying positions of negotiated and potentially decentralized authority over their supporters may throw into question many established models, but fits well with ethnographic and ethnohistoric accounts of life in chiefdoms and small states. Reorienting our conception of chiefdoms to allow for deçentralized and negotiated political processes, rather than assuming centralized decision-making in bounded polities, provides a social basis for understanding why chieftains were endured, asking what useful services they provided, examining how they were able to manipulate the system, and considering how the dynamics of those interactions rray have shaped trajectories ofsocial change. I am not arguing that all chiefdoms were organized like those in Iceland, yet the emergence of stateJike formations here, within the context of a network of interacting and competing chieftaincies, is essentialiy similar to social frameworks inferred behind other cases of state formation around the world (Collis 1994; Flannery 1999; Renfrew and Cheny 1986; Wright 1984) anct meshes well with ethnohistorically described cases of political cycling in chiefdoms worldwide (e.g. Anderson 1994; Kristiansen 1991; Johnson 1999; Persson 1983). These considerations alone suggest that loeland's detailed docuureutary sources and emerging archacological record are worthy of investigation by archaeologists and can be used for testing and developing explanatory theories that However, see Wright et al. (1989) for an encouraging examplc ofhighresolution archaeological analysis, Schele (1990) and others for advances made in understanding dynastic changes in the Maya world, and Kohler and Gurnerman (2000) for a range of prornising approaches ernploying multiscalar modeling. incorporate actors' motivations and rapid, on-the-ground dynamic processes into the longer-tetm perspcctives pro91 vided by the archaeological record itself. In fact, several anthropologists and arclraeologists have already explored the potential value of using lceland's records as a tnodel for uuderstanding the "Germanic mode of production" in European prehistory (Gihnan 1995), comparing the econotnic sttucture of European and Pollmesian chiefdoms (Earle 1997:123-124; 2000:47), and examining the dynamics of stratified societies without state-like integrative mechanisms (Byock 1982, 1988,2001; Durrenberger 1992; Smith and Parsons 1989)' The Icelandic case also suggests that the numerical superiority of non-elite actors must always be considered a potent References Cited force when modeling the dynamics of complex societies, especially in contexts where the weapon-ry everyone held was basically the same. Vy'here no evidence exists to indicate a coercive monopoly of force by the elite-or when the point G. 1994 The Savannah River Chiefdoms: Potilical Change in the Late Prehístoric Southeast University of Alabama Press (Tuscaloosa, AL). Amorosi, Thomas 1991 'Icelandic archaeofauna: a preliminary review' Acta Archaeologica 6l:272-284. Amorosi, Thomas 1992'Climate impact and human response in northeast lceland: archaeological investigations at Svalbarð, 1986-1988' in C.D. Morris and D.J. Rackham (eds) Norse and Later Settlement and Subsistence in the North Atlantic University of Glasgow (Glasgow): 103-1 38. Arnold, CJ. 1997 An Archaeologt of the Early Anglo-Saxon Anderson, David Kingdoms 2"d ed. Routledge (London). Bagge, Sverre 1991 Society and Politics in Snoni Sturluson's Heimsbringla Uníversity of California Press (Berkeley, CA). Bourdieu, Piene 1977 Outline of a Theory of Practice (translated by R. Nice) Cambridge University Press (Cambridge). Brain, Jeffrey P. 1989 Winterville: Late Prehistoric Culture Contacl in the Lower Mississippi Valley Mississippi Department of Archives and History, Archaeological Report No.23 (Jackson, MS). by the non-elite could rarely, if ever, be taken for granted' Their mutual obligations and needs were quite possibly critical elements of any complex system's true dynamics' As archaeologists, this means that we must put far more time into understanding past societies at the scales of the household and community, attempting to understand the divergent interests of actors occupying different positions within social, political, and economic instifutions (Brumfiel 1992; Brumflrel and Fox 1994), and refining our ch-ronological control to scales adequate for monitoring the interaction of actors with one another and with their sunoundings. Acknowledgments of interest is how an elite obtains a coercive monopoly on force-it must be considered that the acceptance of the elite Brumfiel, Elizabeth M. 1992 'Distinguished lecture in archaeology. Breaking and entering the ecosystem: gender, class, and faction steal the show' American Anthropologist 9aQ):551-s67. Fieldwork in lceland, on which many of the thoughts expressed here are based, has been done in coordin(ltion wilh the Department of Archaeology of the National Museum of Iceland, with funding generously provided by rhe National Geographic Society, the American-Scandinavian Foundø' tion, Norðurá\, Eimskip, Sparisjöður Myrasi,slu, the Bulfalo Museum of Science, Brown University, and the University of Michigøn. A large number of individuals and lcelandic community organizations, too many to name here, have provided material, financial, emotional, and intellectual support 'ttaage of over the years. Special thanks are due to Sér Geir teins s on, all Reykholt, Guðmundur O la¡tton, Sigurður B ergs of my crew members, my doctoral committee members, the patient editors of thß volume, Michèle lfayeur Smith, who 'cares for and eicourages me, and Jessica and Emilie, who Ìceep me smiling. ,4ny errors remaining in this presentation, despite their well¿neant commenÍs and warnings, remain the sole responsibility ofthe author. Brumfiel, Elizabeth M. and J.V/. Fox (eds) 1994 Factional Competition and Political Development in the New World Cambridge University Press (Cambridge). Byock, Jesse L. 1982 Feud in the lcelandic Saga University of California Press (Berkeley). Byock, Jesse L. 1988 Medieval lceland: Society, Sagas, and Power Hisarlik Press (Enfield Lock, UK). Byock, Jesse L. 2001 Viking Age lcelønd Penguin Books (Harmondsworth, UK). Carneiro, Robert L. 1970 'A theory of the origin of the state' Science 169:733-738. Carneiro, Robert L. 1977 'Political expansion as an expression of the principle of competitive exclusion' in R. Cohen and E. Service (eds) Origins of the State: The Anthropology of Political Evolution Institute for the Study of Human Issues (Phi la delphia) :20 5 -223 . Carneiro, Roberl L. 1981 'The chiefdom: precursor to the state' in G.D. Jones and R.R. Kautz (eds) The Transition to Statehood in the New World Cambridge University Press (Cambridge):37-79. Clover, Carol J. and John Lindow 1985 Old Norse-Icelandic Literature: A Critical Guide (Islandica XLI) Cornell University Press (Ithaca). Collis, John 1995 'States without centers? The Middle La Tène period in temperate Europe' in B. Arnold and D.B. Biographical Sketch Kevin P. Smith has conducted archaeological excavations and surveys in Iceland since 1985, working primarily with the National Museurn of lceland. He received his B.A. in Antlrropology from Haverford College (Haverford, PA) and his M.A. in Anthropology from the University of Michigan (Ann Arbor, MI). ile is cuireirtiy ABD iir the Deparlment of Anthropology at the University of Michigan and Deputy Director/Chief Curator for the Haffenreffer Museurn of Anthropology at Brown University (Providence and Bristol, Rhode Island, USA). Gibson (eds) Celtic Chiefdom, Celtic State Cambridge University Press (Cambridge):75-80, Dobres, Marcia-Anne and John Robb 2000 Agency in Archaeology Routledge (London). I)ugmore, Andrew J. and Paul C. Iluckland 199 I 'Tephrochronology and Late Holocene soil e¡osion in South lceIand' in J.K. Maizels and C. Casteldine (eds) Environmental Change in lceland, Past and Present Kluwer Acadcmic Press (Dordrecht, Netherlands):147 -160. 92 Durlenberger, E. Paul 1992 The Dynamics of Medieval lceland; Political Economy & Lilerature University of Iowa Press (Iowa City, IA). Earle, Timothy 1977 'A reappraisal of redistribution: complex Hawaiian chiefdoms' in T' Earle and J. Ericson (eds) Exchange Systems in Prehistory Academic Press (New J. Robb (eds) Agency in Archaeology Routledge (London):71-9L Joyce, Rosemary and Julia A. Hendon 2000 'IJeterarchy, history, and rnaterial reality: "communities" in Latc Classic Honduras' in M.A. Canuto and J. Yaeger (eds) The Archaeology of Communities: A New World Perspective Routledge (London ): I 43-l 60. Kemp, Barry J. 1989 Ancient Egypt: Anatomy of a Civilizallon Routledge (London). Knight, Vernon 1., Jr. 1997 'Some developmental parallels between Cahokia and Moundville' in T.R. Pauketat and T.E. Emerson (eds) Cahokia: Domination and ldeology in York):213-232. Earle, Tirnothy 1978 Economic and Social Organization of a Complex Chiefdom, the Halelea District, Kaua'i, ÍIawaii University of Michigan Museum of Anthropology, Anthropological Papers 63 (Ann Arbor, MI)' Earle, Timothy 1997 How Chiefs Come to Power: The Political Economy in Prehisrory Stanford University Press the Mississippian þI/orld University of Nebraska Press (Stanford, CA). Earle, Timothy 2000 'Archaeology, property, and prehistory' Annual Review of Anthropology 29:39'60Elu'enreich, Robert M., Carol L. Crumley, and Janet E. Lely (eds) 1995 Heterarchy qnd the Analysis of Complex Societies Archaeological Papers of the American Anthropological Association No. 6 (Alexandria, VA). Flannery, Kent V. 1999 'Process and agency in early state formation' Cambridge Archaeologic al Jour nal 9(1):3 -21' Flannery, Kent V. and Joyce Marcus 1983 'The origins of the state in Oaxaca: editors' introduction' in K.V. Flannery and J. Marcus (eds) The Cloud People: Divergent (Linco ln, NE) :229 -247 . Knight, Verron J., Jr. and Vincas P. Steponaitis 1998 'A new history of Moundville' in V.J. Knight and V.P. Steponaitis (eds) Archaeology of the Moundville Chiefdom Smith- sonian Institution Press (Washington, DC):1-25. Kohler, Timothy A. and George J. Gumerman 2000 Dynamics in lfuman and Primate Societies: Agent-Based Model- ing of Social and Spatial Processes Oxford University Press (New York). Kowaleski, Stephen A. 1990 'The evolution of complexify in the Valley of Oaxaca' Annual Review of Anthropologt 19:39-5 8. Evolution of the Zapotec and Mixtec Civilizations Aca- demic Press (New York):79-83' Gelsinger, Bruce E. l98l lcelandic Enterprise: Commerce Carolina Press (Columbia, SC). Kristiansen, Kristian 1991 'Chiefdoms, states, and systems of social evolution' in T' Earle (ed) Chiefdoms: Power, Economy, and ldeologt Cambridge University Press and Economy in the Middle Ages University of South Gestsson, Gisli and Jóhann Briem 1954 'Byggðarleifar í Þjórsárdal' Árbók hins islenzka Fornteifafelags I 9 5 4 :5 -22. Giddens, A. 1979 Central Problems in Social Theory: Action, Structure, and Contradiction in Socíal Analysis University of Califomia Press (Berkeley, CA). Gilman, Antonio 1995 'Prehistoric European chiefdoms: rethinking "Germanic" societies' in T.D. Price and G. Feinman (eds) Foundations of Social Inequality Plenum Helms, Mrary 1979 Ancient Panamq: Chiefs in Search of Power University of Texas Press (Austin, TX). Hermanns-Auðardóttir, Margrét 1986'Fornleifarannsóknir að Gásum og víðar í Eyjafrrði árið 1986' Sulur2T:3-39. Hermanns-Auöardóttir, Margrét 1999'Arkeologiska undersökningar av handels-platsen vid Gásir' in A. Christophersen and A. Dybdahl (eds) Gáslr - En Internasional ÍIandelsplas s i Nord-Atlanteren T apir Forlag (Trondheim, Norway):9-36. Press (New York):235-25 1. (Cambridge):16-43. Kristjánsson, Jónas 1997 Eddas and Sagas: Iceland's Medieval Literature (translated by P. Foote) Hið islenska Bókmenntafélag (Reykj avík)' Lárusson, Björn 1967 The Old lcelandic Land Registers C.W.K. Gleerup (Lund). Lehner, Mark 2000 'The fractal house of Pharaoh: Ancient Egypt as a complex adaptive system, a trial formulation' in T.A. Kohler and G.J, Gumerman (eds) Dynamics in Human and Primate Societies: Agent-Based Modeling of Social and Spøtial Processes Oxford University Press Q'{ew York):275-354. Marcus, Joyce 1998 'The peaks and valleys of ancient states: an extension of the dynamic model' in G'M. Feinman and J. Marcus (eds) Archaic States School of American Research Press (Santa Fe):59-94. McGovem, Thomas H. 1990 'The archaeology of the Norse Nor1h Atlantic' Annual Review of Anthropology 19:331' 351. Hodder, Ian and Clive Orton 1976 Spatial Analysis in Archaeology Carnbridge University Press (Cambridge). Ingvarsson, Lúðvík 1987 Goðorð og Goðorðsmenn (Egllsstaðir, Iceland). Jóhannesson, Jón 1974 A History of the Old lcelandic Com' monwealth (istendinga saga, franslated by Haraldur Bessason) University of Manitoba lcelandic Series Vol. lI, University of Manitoba Press (Winnipeg, Manitoba). Johnson, Eric S. 1999 'Cornmunity and confederation: a political geography of Contact Period southern New England' in M.A. Leviue, K.E. Sassarnan, and M.S' Nassaney (eds) The Archaeological Northeasl Bergin & Garvey (Westport, CT): I 55- I 70. Joyce, Artltur A. 2000 'The founding of Monte Albán: sacred propositions and social practices' in M-4. Dobres and 93 McGovern, Thomas H., Sophia Perdikaris, and Ciay'ton Tinsley 2001 'The economy of landnám: the evidence of zooarchaeology' in A. Wawn and Þ. Sigurðardóttir (eds) Approaches to Vínland Sigurður Nordal Institute Studies 4 UòCrLw, Julia (trans.) 1970 Sturlunga Saga, Volume L' The Saga of Hvatnm-Sturla and The Saga oJ' the lcelanders American-scaudinavian Foundation (f{ew York). McGrew, Julia and R. George Thornas (trans') 1974 Sturlunga Saga, Volune 2" Shorter Sagas oJ'the lcelanders (Reykjavík):1 54-165. rwavnå Publishers and The twãyne Publishers and The Americarr-Scandinavian Foundation (New York). Mell'er, Mark 2000 'Heterarchy and hierarchy: the community plan as institution in Cahokia's polity' irr M'A' Canuto røiii"r, wittiu,r, i. rsso Blood-taking and Peacemaking" ' '^i:;"d, of ChiLaw, and Society in Saga lceland University of social structure among the o;;ti il;iltì" 195i ìã.ír"r¿ tribes of South and central America' American Iceland A'D' ogirui", Àrtrù B.c. 1991 'Climatic changes in -'c.865 1598' Acta Archaeologica 61:223-251' to í Borgar' Ólafsson, Guömundur 1996 Friðlystar Fornleifur AnthroPologist 57: 472-487 ' of and J. Yaeger (eds) The Archaeology -ie. (London) Wo, t lp ersp ec iv e Routledge t Communities: A :44-5 7' Snrith, Kevin P. nd The Economic Structure of Medieval Icelandic Chiefdorns" Political Economy in the Slurlung Period, AD II50-1264 Doctoral dissertation, in prep', Department of Antll'opology, University of Michigan, An¡ Arbor. Smith, Kevin P. an<J Jeffrey R. Parsons 1989 'Regional archaeological research in lceland: potentials and possibilities' in E.P. Durrenberger and G' Pálsson (eds) The Anthropology of Iceland University of Iowa Press (Iowa City, Spencer, Charles S. 1990 'On the tempo and mode cago Press (Chicago)' ''Tvpes Ilt):179-202. of state i,orðaísí,slu (Rit hins islenska Fornleifafélags og Þjóðmin'íorofnrTstands 2) H\ó Íslenska Fornleifafélag og Þjóðminlutufo ÍttunAs (ReYkjavík) Cahokia Pauketat, Timothy R. 1994 The Ascent of Chiefs: formation: neo-evolutionism reconsidered' Journal of An- thropological Archaeology 9(1): 1-30. Stein-Witkèshuis, Martina 1982 'The right to social welfare 8:343-352. Steponaitis, Vincas in early medieval lceland' Journal of Medieval History P. 1998 'Population trends at Mound- àrd U¡ts¡tsippian Politics in Native North America Uni- versity of Alabama Press (Tuscaloosa, AT ì'Pauketai, Timothy R. 2000 'The tragedy of the commoners' in M-A. Dobres and J. Robb (eds) Agency in Archaeology Routledge (London): I 13-129. Peebles, Christopher S. and Susan M' Kus 1977 'Some archaeological correlates of ranked societies' American AnPersson, Johnny 1983 'Cyclical change and circular exchange: a re-examination of the Kula ring' Oceania Rafnsson, Sveinbjörn 1990 Byggðaleifur í Hrafnkelsdal og á (eds) '4rSmithsonian Instichaeologt of the Moundville Chiefdom tution Press (Washington, DC):26-44. Sveinbjarnardóttir, Guðrún 1992 Farm Abandonment in Me- uill"' in V.J. Knight, Jr. and V.P' Steponaitis tiquity 421421-448' dievât and Post-Medieval lceland: An Interdisciplinary Study Oxbow Monograph 17, Oxbow Books (Oxford)' Thorsjon, Örnólfur (ed) 1983 Sturlunga ^Saga Svart á Hvítu (Reykjavík)' Thursion, Tina 1996 'Landscapes of power and landscapes of conflict; geographical approaches to culture change in saQ):32-77. Brúardölum ¡ntt nins islenska Fornleifafélags 1) Hið Íslenska Fomleifafélag (Reykjavík). Iron Age Denmark' in D.A. Meyer, P'C' Dawson, and Renfrew, Colin 1974 'Space, time, and polity' in M' Rowlands and J. Friedman (eds) The EvoluÍion of Social Sys/ems Duckworth (London):89-1 14. Renfrew, Colin 1975 'Trade as action at a distance' in J'A' Sabloff and C.C. Lamberg-Karlovsky (eds) Ancient Civilization and Trade lJniversity of New Mexico Press (Albuquerque):3-59. D.T. Hãnna (eds) Debatíng Complexity, Proceedings of the 2dh Annual Chacmool Conference The Archaeological Association of the University of Calgary (Calgary, Alberta):428-442. Vésteinsson, Orri 1998 'Pattems of settlement in lceland: a study in pre-history' Saga-Book of the Viking Society 25:1- Wright, Henry Renfrew, Colin and John Cherry 7986 Peer Polity InteracPress (Cambridge). tion and Socio-Political Change Cambridge University Renfrew, Colin and E.V. Level 1979 'Exploring dominance: predicting polities from centres' in C' Renflew and K'L' -Cooke (eds) Transþrmations; Mathematical Approaches to Culture Change Academic Press (New York):145-168' Schele, Linda M. 1990 Forest of Kings: The Untold Story of the Ancient MayaMorrow (New York). 1984 'Prestate political formations' in W' Sa=nders, H. Wright, and R.M. Adams (eds) On the Evolution of Complex Societies: Essays in Honor of Harry Hoij er tJ ndena Pub lications (Malibu, C lt') :4 l -7 7' Wright, Henry T., Richard Redding, and Susan Pollock 1989 'lùonitoring interannual variability: an example fi'om the period of eãrly state development in southwestern lran' in Þ. Halstead and J. O'Shea (eds) Bad Year Economics: Cultural Responses to Risk qnd (Jncertainty Cambtidge Uni- 29. T. Scull, Christopher 1999 'social archaeology and AngloSaxon kingdom origins' iIl T. Dickinson and D. Griffiths (eds) The Making of Kingdams Anglo-Saxon Studies iir Archaeology and History 10, Oxford University Commil tee for Archaeology (Oxford): I 7-24. Service, Elman R. 1971 Primitive Social Organization: An Zttter, Cynthia M' 1992 'Icelandic plant and land-use patterns: àrchaeobotanical analysis of the Svalbarð midden (6706-60), northeastern lceland' in C. D' Morris and D' J' Rackham (eds) Norse and Later Settletnent and Subsistence in fh,e North Atlantic University of Glasgow (Glasgow):139-148. versify Press (Cambridge): I 06-1 1 3. Evolutionary Perspective 2nd ed. Random House Qt{ew YorÐ. Sigurðsson, Jón Viðar 1989 Fra Goðorðum til Ríkiø: Þróun 13. Otd (Sagnfræðirannsóknir 10, Sagnfræðistofnun Háskóla islands) Bókaútgáfa MemrinGoðavalds á 12. og garsjóðs (Reykjavík). Sigurðsson, Jón Viðar \999 Chieftains and Power in the Icelandic Contmanwealth (translated by Jcan LundskærNielsen) Oclense University Press (Odense). Smith, Kevin P. 1995 'Landnárn: the settlement of lceland in archaeological and historical perspective' llorld Archaeology 26:319-347. 94 Figure 1: According to medieval lcelandic sources, thirty-six chieftaincies (goöorð) were recognized by law in AD 930; three more were added in 965. The actual number of chieftaincies may have varied through time but these offices endured in name, at least, throughout the Commonwealth period (AD 930-1264) even if some were later absorbed into larger polities. Dark triangles locate the areas where the 36 "full and ancient" goöorð had their seats; white triangles identify the three chieftaincies added after 965. Numbers carry through to Figures 2 and 3. 36 t¡-tJt 35 34 33 32 3l 30 2C West 26 25 24 23 22 2t 20 -l¡ Eoutft t8 t7 16 t5 14 13 12 1l I I 1 ó Eâst 5 4 3 1 't 1t11222 5ô789012 00000000 Figure 2: Graphic representation of the feud chains and alliance cascades in The Saga of Hvamm-Sturla (McGrew 1190 with 1gZ0). This dispute between two adjacent western lcelandic chieftaincies began in AD 1147 and ended circa the entanglement of chiefly houses in three Quarters of lceland, and the the a'bsorption of one chieftaincy by another, in Figure initiation of another feud chain in the North Quarter. Numbers on the y-axis refer to the chieftaincies mapped 1; symbols are as described in the text (half black/half white iquares identify mediators called to arbitrate settlements). 9s llllllll 1i :, I .i. å I i i i 36 35 34 *""¡ :i 30 29 28 27 26 25 24 29 22 I I I ì I V/est I 2t i ì ì 20 -t9 t? so$th t8 '16 i I t5 t4 t3 _-_---__--1 l2 I ll i T -19 Esst I s 7 6 fi 5 4 3 a '| fi Ë ï g l¡|lIll|lll|llllllIlll|llllllIll|llll|l fr 111t12222 567890123 000000000 Figure 3: Graphic representation of intersecting and concurrent feud chains and alliance cascades across lceland, circa AD 1147-1235, according to íslendinga Saga (McGrew 1970). Note the ever-decreasing number of independent chiefly lines and the increasing tempo and spatial extent of hostilities as time progresses. Numbers on the y-axis refer to the chieftaincies mapped iñ figure 1; symbols are as described in the text (half black/half white squares identify mediators called to arbitrate settlements). lllllllll å ä ø ü $ & fr g ü il g g & n ffi # r 6 g $ H Ë ffi ft g x ffi ß Atlantic Ocean H ffi nd E m ù Chieftain's center o Non-elite church farm Non-elite farm Harbour Minor lantling place Figure 4: The Borgarfiörður region of western lceland, circa AD 1180, according to medieval texts and archaeological research. While the number and location of chieftains' centers, non-élite church farms, harbours, and landing places are probably accurate, non-élite farms are certainly under-represented. Dark gray shading identifies land over 200 meters AMSL;white areas are glaciers. 96 å t, It rr Figure 5: The Borgarf,örður region of western lceland, circa AD 1180, with Thiessen polygons drawn around seven chieftains' centers located within the district and one (H) in the adjacent Breiðafjöröur district. Dark gray shading identifies land over 200 meters AMSL; white areas are glaciers. v al )S l0 center of the map is the boundary between the South and West Quarters. Dark lines link chieftains' centers to the Assembly Sites where they and their followers met each spring. [1] and [2] identify the farms of Þorð the Red and Hámund Gilsson, respectively, with gray lines linking them to their chieftains. Figure 6: The Borgarfiörður region of western lceland, circa AD 1180, showing political boundaries and alliances known from historical sources and discussed in the text. The bold line following the line of the river Hvítá through the 9l