If the over-riding purpose of this conference is to explore how historical analysis of global processes affects our understanding of regions, then a comparative treatment of borderlands is relevant to our purposes, since we normally conceive of a borderland as a particular kind of region. To address a subject as diverse as borderlands effectively, however, we must consider at the outset the problem of taxonomy, for we have to classify and define the phenomenon in question with sufficient precision to permit the identification of those shared attributes that can be productively compared and to insure that we are in fact examining comparable case studies. We should therefore begin by asking ourselves what we mean when we talk about a region. How, in other words, do we identify, describe, and demarcate regions as meaningful units of study?

            In taking this initial step toward a definition of borderlands, the recognizably uniform physical characteristics of a given geographical area‹its location, relief features, climate, soil, natural resources, flora, and fauna‹obviously demand our attention. More fundamental for our purposes, however, are the distinctive patterns of human interaction and the correlation we find between human interrelations and a distinctive geographical place. The territorial extent of that correlation between the physical and the human determines the approximate boundaries or outer zones of a historically significant region, which can be viewed as something relatively small‹the Vendee in France, for example‹or seen, with different purposes in mind, as very large— the Amazon basin being an obvious instance. We thus imagine a region as a constantly changing physical-cultural entity with imprecise borders that ebb and flow over the centuries.  In any historical epoch, a region does not necessarily coincide with political or administrative boundaries. It can sometimes transcend the jurisdiction of two or more modern nation-states. A region often exhibits considerable internal diversity, despite the widespread physical features and networks of socio-economic relations that define it. Hence, for analytical purposes, we normally feel compelled to divide a larger region into smaller sub-regions.¹

            A rich body of recent scholarship demonstrates the potential of applying regional approaches to the study of borderlands. Two markedly different books illustrate this point. The first example is a volume of essays on region-wide interaction in the American Southwest in which various authors explore large-scale themes in various periods of prehistory.² Among other topics, these authors describe the nature of human interaction within archaeologically delineated spatial divisions as well as patterns that operated across regional boundaries. Juxtaposing the tremendous diversity and recurring similarities prevalent throughout the Southwest, they consciously avoid the use of formal models such as world-system theory, arguing that even a regional-system concept can mask important socio-cultural variation.  Nonetheless, their treatment of warfare, land use, economic exchange, religious practices, migration, and cultural diffusion demonstrate the value of thinking in terms of regional interaction.

            In a similar vein, Karen Wigen�s book on Japan�s lower Ina valley shows us how the modern transformation of a clearly definable region must be understood on the basis of Japan�s changing role in East Asia and in strictly national terms as well.³ Her study, tying history to geography, assumes that Tokugawa Japan and early modern Europe share a number of traits critical to a satisfactory explanation of how and why the Ina valley became a periphery of a Tokyo-centered national economy in the period from 1750 to 1920. To complete her task, Professor Wigen examines human-land interactions and employs spatial analysis in conjunction with her efforts to delineate a region in functional terms, i. e. as an integrated economic unit, and to demonstrate how it was constructed and reconstructed over time. Placing this study of a regional metamorphosis shaped by global, national, and local forces along side of recent scholarship on regional interaction in the prehistoric American Southwest, we gain a reasonably clear sense of the rich conceptual and methodological resources we can draw upon to develop regional approaches appropriate for a comparative and cross-cultural investigation of borderlands.

            But before we proceed, we must still decide what basic traits mark off a region as a borderland.  A recent AHR Forum reminds us of how difficult addressing this issue invariably turns out to be. Jeremy Adelman and Stephen Aron, treating the North American experience from the seventeenth through the early nineteenth century, described a frontier as øa meeting place of peoples in which geographic and cultural borders were not clearly defined,Ó while øreserving the designation of borderlands for the contested boundaries between colonial domains,Ó and arguing that the øshift from inter-imperial struggle to international coexistence turned borderlands into bordered lands.Ó⁴ Yet Evan Haefeli immediately criticized them øfor failing to maintain a consistent distinction between frontier and borderland phenomena.Ó Pointing out the unusual fluidity and instability of North American frontiers as opposed to those on other continents that tended to coalesce along ecological transition zones and endure for extended periods of time, he suggested that a borderland and a frontier must be carefully differentiated and that a borderland should be understood øas a place where autonomous peoples of different cultures are bound together by a greater multi-imperial [or more recently multi-national] context.Ó⁵

            Meanwhile, Michiel Baud and Willem van Schendel have made a systematic effort to create a conceptual framework for the comparative study of borderlands. Pointing out that the problem of borderlands became worldwide only with the consolidation of modern nation-states in the eighteenth and nineteenth century, they reject a purely state-centered approach and adopt øa cross-border perspective, in which the region on both sides of a border is taken as the unit of analysis.Ó⁶ Borders may be political divides, but they nonetheless forge unique transnational patterns of interaction shaped in part by complicated relationships between regional elites, the common folk, and the two state governments whose authority meets at the political divide. Distinctive political, social, economic, and cultural patterns impart a measure of unity to a borderland and help to determine its geographical extent, which may expand or contract over time. Baud and van Schendel regard the circumstances in which several frontiers come together, whether in peaceful or conflicted ways, as an embryonic borderland, the nineteenth-century history of  northern Mexico and the American Southwest offering a classic instance of a borderland that emerged out of a frontier situation involving two modern states.⁷

            Baud and van Schendel also suggest how we might examine borderlands as particular kinds of closely knit yet permeable regions of interaction where human relationships have been shaped by the sharp juxtaposition of  strongly contrasting cultures. They show how cross-border perspectives can be used to explore the divisions and accommodations between separate peoples within a shared meeting ground. They emphasize the distinctive conditions that create the social space and identity of a borderland, and the historical consequences provoked by the resulting nexuses of distinctive interests.  But Baud and van Schendel direct their attention primarily toward recent experience, leaving questions about how to handle the existence of borderlands prior to the emergence of modern state borders largely unanswered. The clear delineation of political boundaries is a comparatively recent practice, and even in Europe national governments did not always survey their borders until after the Napoleonic wars.⁸ Yet  borderlands as we are discussing them here can be traced back at least as far as the crystallization of the earliest known complex societies.  By roughly 3500 B.C.E., the ancient Sumarians, having asserted their dominance over the alluvial lands of the Tigris and Euphrates, maintained persistent contacts, sometimes peaceful and sometimes violent, with pastoral nomads to their west, where savannas merged into the immense Arabian desert, and to the north and east, where the plains abutted up against the wooded foothills of the Taurus and Zagros mountains. The consequent interaction between settled lowland peoples and neighboring upland herders so obviously different in their lifeways produced the very first borderlands we can describe in any detail. While Baud and van Schendel recognize that borderlands do have a long history and exhibit substantial differences from one region of the world or one historical period to the next, they have relatively little to say about the problem of how meaningful comparisons might actually be made.⁹

            Despite its challenges, comparative history gives us many advantages. We are drawn into a systematic mode of inquiry that imposes a measure of control over our project and invariably compels us to be analytical. At the most elementary level, we line up two or more case studies, asking a number of questions about each instance. While our difficulties tend to mount as the distances of time and space increase, the objects of our investigation can be either widely separated or tightly clustered in time and space as long as they all fit a workable definition of the phenomenon under consideration and offer a range of relevant common features for us to examine. A rigorously implemented comparison of parallel types grounded in exacting empirical research can explain the presence of a given phenomenon in a variety of unique circumstances. It can highlight clear-cut differences as well as persistent similarities from one case to another. It helps us to distinguish purely local from more broadly applicable factors, pointing us toward generalizations that extend beyond a particular historical context. It reveals the larger significance of a given experience. At its most sophisticated level, comparative history helps us formulate, test, and refine hypotheses capable of explaining perceived relationships between two or more specified variables.¹⁰

            For purposes of comparative analysis, students of borderlands have an incredible range of possible combinations from which to choose.¹¹ However, even scholars from outside the United States find the American Southwest/northern Mexico borderlands difficult to ignore. One reason may be its similarity with other experiences elsewhere in the world, particularly with regard to its extended history.¹² A prehistory in the Southwest shaped by ecological circumstances was followed by a persistent Spanish presence that stretched across the entire continent. European imperial rivalries in the eighteenth century led to the loss of Spanish territorial holdings and claims prior to Mexican independence in 1821. Less than three decades later the United States had gained control over the immense northern reaches of what had been New Spain.  Thereafter, the actions of the U. S. and Mexican governments and the relationships between the two nations quickly øturned borderlands into bordered lands,Ó¹³ in the words of Jeremy Adelman and Stephen Aron, who view North American borderlands essentially as contested regions separating European colonial territories, and who rightly insist that native Americans must be included in the treatment of conflicts and accommodations that did so much to dictate the direction of this long history.

            Recent concerns over the inability of the United States to control its southern border effectively and over the impact of NAFTA have intensified interest  in the American Southwest/ northern Mexico borderlands. Comparing it with the borderlands between the United States and Canada dramatizes just how important historical, cultural, and ethnic differences tend to be in the making of the phenomenon. The United States and Canada share a boundary line that stretches across the entire continent.  Disputes over that boundary line, focusing at different times on the Oregon territory, the upper reaches of the Louisiana Purchase, the Great Lake region, and the northern boundary of Maine, punctuated diplomatic relations between the United States and various European powers, above all Great Britain, during the early decades of the nineteenth century. Canada long served as a sanctuary for run-away slaves and occasionally for native American tribes, such as the bands following Sitting Bull following the Battle of the Little Bighorn.  At the outset of the twenty-first century, it is a vital trading partner for the United States. Patterns of interaction between peoples living in the Detroit-Windsor area resemble those between peoples in the environs of El Paso-Ciudad Juarez in many respects. Yet scholars in the United States have paid comparatively little attention to the borderlands separating the two countries, a reminder of the extent to which human perceptions go into the making of the phenomenon.

            Whatever other case studies might be selected, no student of world-historical borderlands can afford  to ignore China�s northern frontier with the pastoral nomads of Inner Asia. Here we discover an immense and complicated interaction zone stretching from Manchuria through Mongolia and northern China to Turkestan. In the heart of this huge area, characterized by immense geographical and cultural variation, the intensive agriculture practiced on the easily tilled loess soils of the Yellow River�s middle stretches have historically given way to rainfall agriculture and herding in the drier areas beyond the river valleys, and imperceptibly to pastoral nomadism in the desiccated steppe lands of Mongolia. Interaction between nomadic herders to the north and more sedentary communities on the outskirts of the agricultural areas, sometimes conflicted and violent, often genuinely symbiotic, exercised powerful influences on the development of Chinese civilization, as historians have readily recognized.¹⁴ Building on the work of Owen Lattimore, experts have increasingly emphasized the concomitant affects of these connections on steppe societies, especially after pastoralists on the great grasslands of  Eurasia had mastered the skills of horseback riding after 900 B. C. E.¹⁵

            Viewed in cross-cultural terms, nomadic states, typically assuming the form of imperial confederacies, seem to have developed and disintegrated with a timing that matched the rhythms of Chinese dynastic politics. The spread of horseback riding into Mongolia shortly before 400 B. C. E. provoked the organization of Chinese armies that combined infantry with cavalry, setting off a cyclical pattern of interactions that played itself out repeatedly over the next several millennia. Three periods of imperial consolidation alternated with corresponding interludes of intensified struggle and chaos, the conquest of China by the Mongol Yuan dynasty constituting the major aberration in this cyclical pattern.  The Han and Hsiung-nu empire dominated the first period. During the second, the Sui, T�ang, and Sung dynasties ruled China while the central steppe fell under sway of two Turkish states and then the Uighurs.  In the third, the Ming had to deal with the power of Oirats and Eastern Mongols, while the Ch�ing confronted the Zunghars, who forged the last of the extensive steppe empires. By the middle of the  eighteenth century, when the autonomous existence of nomadic and semi-nomadic peoples throughout the world had started to succumb to the expanding power of agriculturally based, city centered societies, an on-going confrontation between Russia and China had replaced the interactions between the sedentary Chinese and the pastoralists to the north as the dominant factor shaping conditions in the borderlands of Inner Asia.

            Running through the middle of the transition zone where the lifeways of farmers and pastoralists intermingled and where the interactions between the Chinese and their nomadic neighbors conditioned historical developments was the Great Wall, whose oldest segments had been built by the Ch�in dynasty in the late third century B. C. E. to delineate the outer edges of their territorial domains and to keep the people considered their own separated from the decidedly different and always menacing outsiders dwelling beyond the range of their effective power. For China�s rulers the Great Wall served political as well as military purposes. Ever anxious about their ability to control peripheral regions, they were usually reluctant to encourage trade across their northern borderlands lest economic exchange bind the local Chinese too closely to steppe cultures. Above all else, they conceived of the Great Wall less as a clear, continuous line and more as a cordon sanitaire, a barrier restricting the movement of people and goods over what they regarded as the approximate boundary of their state. Only during the era of Ming rule did a dynasty make a sustained attempt to construct a permanent line of defense designed to shield their northern frontier from invasion.¹⁶

            Nomadic confederacies, by contrast, normally depended on opportunities to be found south of the Great Wall, since exploitable resources available on the open steppe seldom proved sufficient to build and maintain elaborate political structures. When nomadic leaders conducted destructive raids into China, they were often trying to extract diplomatic and trade agreements out of Chinese authorities. Their interests seemed best served by the preservation of a stable Chinese regime capable of offering them annual subsidies, marriage alliances, formal recognition, and other valuable concessions. Given the limitations of their military systems and administrative methods, they were seldom tempted to initiate campaigns conquest in lands beyond the Great Wall that they could not possibly retain for very long.

When both imperial China and one or more nomadic confederacies grew strong, they divided the borderlands region between them. A dominant steppe power, such as the Hsiung-nu state, would then conduct some form of what Thomas Barfield has called an øouter frontier strategy.Ó Officials at the Chinese court typically viewed appeasement as a means of preventing disruptive incursions and weakening potentially dangerous outsiders. When, by contrast, China succumbed to chaos, discord between competing tribes usually prevailed on the steppelands. Rival nomadic leaders, struggling to build a new empire, found themselves ill-served by a weak and disunited China. Initially vulnerable but ultimately triumphant tribal chieftains tended to adopt an øinner frontier strategyÓ aimed at the eliciting of Chinese assistance without loss of autonomy. The Chinese, hoping to encourage internecine tribal struggles and lay a foundation for amicable relations with the ultimate victor, would willingly provide the assistance even if their cooperation contributed in the long run to the reappearance of a dominant‹and menacing‹steppe empire.¹⁷

            An examination of the persistent patterns of interaction characteristic of China�s northern borderlands invariably evokes comparisons with the various frontiers of Imperial Rome, a case study of special interest, given the diversity of instances that it provides as well as the rapidly changing nature of borderland experiences amidst an extraordinary rise to greatness and an equally haunting decline. Events at any given time in northern Britain, North Africa, along the Rhine and the Danube, and in Southwest Asia, where the Romans confronted the power first of the Parthian Empire and then of the Sassanid (Persian ) Empire, suggest a series of striking parallels along with the anticipated differences. Historians and archaeologists alike have recently been inclined to reject the concept of a frontier world neatly divided into clear-cut halves‹Roman and non-Roman or barbarian‹and to adopt C. R. Whittaker�s view of Roman frontiers as spheres of complicated interaction between diverse frontier peoples.¹⁸ Experts have demonstrated that the Romans themselves disagreed on the exact limits of their empire, and had difficulty discerning any meaningful border separating them from outsiders. Along their various frontiers, peoples with different backgrounds intermingled, eventually blurring cultural distinctions. Romans and those they called øbarbariansÓ overcame their differences sufficiently to intermingle, live and work together, and exchange artifacts, ideas, and practices. Individuals dwelling within a borderlands region were readily transformed into cultural hybrids, neither Romanized barbarians nor barbarized Romans. Trade goods carried across the Rhine and the Danube reached the very heart of the empire on the one side and as far afield as Scandinavia on the other.¹⁹ Just as scholars investigating the American borderlands are learning how to delineate a history shared by the northern-most states of Mexico and what we describe as the Southwest, and experts focusing on China�s northern borderlands are linking events in Inner Asia with what was simultaneously occurring to the immediate south of the Great Wall, so specialists concentrating on ancient Rome are forging a broader understanding of its imperial frontiers, one that more fully integrates the rise and fall of a great empire with the unfolding of the diverse cultures surrounding its core provinces.

            C. R. Whittaker�s contributions to recent conceptual advances rest on a rejection of the idea that Roman power was confined within set boundaries delineated by rivers and walls. Despite the Roman understanding that their civilization was in fact limited in extent, those limits remained vague, in part because they never sensed that the empire had reached its ultimate ends. Even in the Middle East, where power relations remained in rough balance for extended periods of time, a fixed, linear frontier failed to emerge. The Romans refused to accept the Euphrates as a natural boundary. Like the Rhine and the Danube, the Euphrates ran through a band of territory that tended to draw together peoples who were ethnically and culturally diverse through the development of local socio-economic ties. Lands on either side of these rivers had more in common with each other than they did with neighboring regions. The Romans saw the great rivers as vehicles of easy transport and communication for their armies. In a manner that resembled their Parthian or Germanic neighbors, they sought to extend their control beyond the rivers whenever such advances proved feasible. Similarly, neither Hadrian�s Wall in northern Britain nor the African limes in what is now Tunisia and Libya were erected over a cultural dividing line. In both cases, the Romans sought to monitor trade and the movement of people rather than establish a defensive barrier against intruders. Along all of the empire�s borderlands non-Roman elites shared lifeways with Romans dwelling in nearby towns or encampments while rural folk, much less affected by Roman culture traits, had little in common with their social betters. Frontier societies, built to a substantial degree on peaceful accommodation, permitted regular interchange back and forth over extended periods, and when Roman authorities felt pressure from outsiders, it frequently came in the form of sporadic raids or the movement of small bands rather than massive assaults by invading hordes. Given the different circumstances faced by the Romans from one border region to the next, they never could commit themselves to a grand strategy covering the defense of the entire empire.²⁰

            Another fascinating borderland, indicative of just how diverse the phenomenon has been over the centuries, can be found in a region that William H. McNeill has called øEurope�s steppe frontier,Ó²¹ the far western end of Eurasia�s grasslands running from Hungary�s Great Alfold  through the Ukraine‹a name itself derived from a word meaning borderland‹all the way to the area of the North Caucacus. From time immemorial the flat, open plain of central Hungary had served as an arena of contact and confrontation between Europe and Asia. Long before the arrival of the Magyars at the end of the ninth century C. E., Scythians, Sarmations, Huns, Bulgars, and Avars had conquered the Great Alfold, often leaving the extensions of this vast plain laying beyond the Danube River�s right bank under the control of European peoples. Hence it was, as C. A. Macartney once observed, that øthese two elements‹Europe and Asia‹[repeatedly] strove for mastery, [although] neither ever achieved it quite completely.Ó²² As a consequence, the great basin of the middle Danube became part of Europe�s eastern frontier, a segment of that broad geographical transition zone where the culture of the West blended gradually into the culture of Asia.²³

            What made Europe�s steppe frontier distinctive among the other case studies we have thus far considered was its more heavily militarized character, which by the mid-sixteenth century set it apart from conditions in the relatively secure areas adjacent to the contested zone. The Ottoman Turks, already firmly established in the Balkans and the southern Ukraine by the end of the fifteenth century, successfully occupied central Hungary in 1543, extending their sway to within eighty miles of Vienna and making Hungary a part of the borderlands between Christendom and Islam.²⁴ For several centuries thereafter, they confronted a line of often hostile Christian states‹the Venetian Republic, the Habsburg monarchy, the Kingdom of Poland Lithuania, and imperial Russia‹whose rulers and subject populations tended to view the Moslem invaders as an alien horde and a direct threat to their own way of life.²⁵  Protracted hostilities were most intense between the Turks and the Austrians, who maintained a more or less organized military border between themselves and the Ottoman Empire until 1881, and between the Cossack and Tartar communities in the southern Ukraine. Within the disputed frontier zone, usually several hundred miles in width, Transylvania, the two Rumanian principalities of Moldavia and Wallachia, and Christian rulers of Georgia as well as the Cossack war bands and the Tartar khanate of Crimea exploited the chronic rivalries between the Turks and their enemies to retain a measure of autonomy.²⁶ Recurring wars fought to advance mutually incompatible interests in the borderlands region did not preclude extended interludes of formal peace. Nor did they prevent normal exchanges between ostensibly alien peoples. The city-republic of Dubrovnik (Ragusa), a virtually independent part of the Ottoman Empire, continued to serve as a conduit for the passage of Western trade goods and cultural influences into the Balkans, while Krakow seems to have benefited at least initially from the Turkish onslaught by becoming an intermediary between Habsburg lands and the Ottoman portions of Hungary.  From one end of Europe�s steppe frontier to the other, a mosaic of constantly shifting relationships between local elites and the governments of the intruding states, and between local elites and a given area�s ordinary people exercised as much influence on the social and political history of the region as conditions within the various competing powers or the more thoroughly studied interactions between them.²⁷

            Just how far a state�s authority extended into this transition zone remained open to debate. The Vienna court, for example, always claimed the whole of its Hungarian inheritance. The sultan, on the other hand, regarded any ruler who paid tribute to Constantinople‹and that included the Habsburg emperor‹as his vassel and therefore subordinate to his will.²⁸ Such conditions would not long survive the consolidation of what McNeill has termed øbureaucratic empires.Ó By the middle of the eighteenth century, the unincorporated borderlands between Austria, Russia, and Turkey had been substantially reduced.  Following the lead of Austrian and Russian armed forces, agricultural settlements spread rapidly onto open ground. By 1800 the entire frontier zone had been absorbed by the three surrounding empires, although the imprint of what had happened in the region since the Turks first conquered the central plains of Hungary still remained essentially unaltered.²⁹

            The four case studies that we have briefly examined hardly exhausts the range of possibilities for the comparative study of borderlands. They suggest, however, the potential embedded in this approach.  Each formed closely-knit yet permeable geographical areas where human interrelationships were shaped by a sharp juxtaposition of strongly contrasting cultures. Despite obvious differences, each in its own way revealed the structural similarities of borderland dynamics that have  prevailed in widely separated historical periods and different places around the world. These four examples are sufficient to demonstrate how much we miss when we pursue our study of the past exclusively within a framework of distinct states and separate cultures. They certainly dramatize the inadequacies of our conventional maps. By revealing unexpected historical connections to be found in apparently unrelated human experiences, they remind us of the costs of scholarly fragmentation as well as the value specialization. In particular, because of the various ways they can assist us in our efforts to reconfigure concepts of  øareaÓ and østate,Ó comparative and cross-cultural approaches to the phenomenon of borderlands are highly germane to the objectives of this conference on interactions.