Figure 5.2 A diagram of the sequence and approximate dates of splits in early Indo-European as proposed in this book, with the maximal window for Proto-Indo-European indicated by the dashed lines. The dates of splits are determined by archaeological events described in chapters 11 (Anatolian) through 16 (Iranian and Indic).
Before we step into the archaeology, however, we should pause and think for a moment about the gap we are stepping across, the void between linguistics and archaeology, a chasm most Western archaeologists feel cannot be crossed. Many would say that language and material culture are completely unrelated, or are related in such changeable and complicated ways that it is impossible to use material culture to identify language groups or boundaries. If that is true, then even if we can identify the place and time of the Indo-European homeland using the reconstructed vocabulary, the link to archaeology is impossible. We cannot expect any correlation with material culture. But is such pessimism warranted? Is there no predictable, regular link between language and material culture?
CHAPTER SIX
The Archaeology of Language
A language homeland implies a bounded space of some kind. How can we define those boundaries? Can ancient linguistic frontiers be identified through archaeology?
Let us first define our terms. It would be helpful if anthropologists used the same vocabulary used in geography. According to geographers, the word border is neutral—it has no special or restricted meaning. A frontier is a specific kind of border—a transitional zone with some depth, porous to cross-border movement, and very possibly dynamic and moving. A frontier can be cultural, like the Western frontier of European settlement in North America, or ecological. An ecotone is an ecological frontier. Some ecotones are very subtle and small-scale—there are dozens of tiny ecotones in any suburban yard—and others are very large-scale, like the border between steppe and forest running east-west across central Eurasia. Finally, a sharply defined border that limits movement in some way is a boundary; for example, the political borders of modern nations are boundaries. But nation-like political and linguistic boundaries were unknown in the Pontic-Caspian region between 4500 and 2500 BCE. The cultures we are interested in were tribal societies.1
Archaeologists’ interpretations of premodern tribal borders have changed in the last forty years. Most pre-state tribal borders are now thought to have been porous and dynamic—frontiers, not boundaries. More important, most are thought to have been ephemeral. The tribes Europeans encountered in their colonial ventures in Africa, South Asia, the Pacific, and the Americas were at first assumed to have existed for a long time. They often claimed antiquity for themselves. But many tribes are now believed to have been transient political communities of the historical moment. Like the Ojibwa, some might have crystallized only after contact with European agents who wanted to deal with bounded groups to facilitate the negotiation of territorial treaties. And the same critical attitude toward bounded tribal territories is applied to European history. Ancient European tribal identities—Celt, Scythian, Cimbri, Teuton, and Pict—are now frequently seen as convenient names for chameleon-like political alliances that had no true ethnic identity, or as brief ethnic phenomena that were unable to persist for any length of time, or even as entirely imaginary later inventions.2
Pre-state language borders are thought to have been equally fluid, characterized by intergrading local dialects rather than sharp boundaries. Where language and material culture styles (house type, town type, economy, dress, etc.) did coincide geographically to create a tribal ethnolinguistic frontier, we should expect it to have been short-lived. Language and material culture can change at different speeds for different reasons, and so are thought to grow apart easily. Historians and sociologists from Eric Hobsbawm to Anthony Giddens have proposed that there were no really distinct and stable ethnolinguistic borders in Europe until the late eighteenth century, when the French Revolution ushered in the era of nation-states. In this view of the past only the state is accorded both the need and the power to warp ethnolinguistic identity into a stable and persistent phenomenon, like the state itself. So how can we hope to identify ephemeral language frontiers in 3500 BCE? Did they even exist long enough to be visible archaeologically?3
Unfortunately this problem is compounded by the shortcomings of archaeological methods. Most archaeologists would agree that we do not really know how to recognize tribal ethnolinguistic frontiers, even if they were stable. Pottery styles were often assumed by pre—World War II archaeologists to be an indicator of social identity. But we now know that no simple connection exists between pottery types and ethnicity; as noted in chapter 1, every modern archaeology student knows that “pots are not people.” The same problem applies to other kinds of material culture. Arrow-point types did seem to correlate with language families among the San hunter-gatherers of South Africa; however, among the Contact-period Native Americans in the northeastern U.S., the “Madison”-type arrow point was used by both Iroquoian and Algonkian speakers—its distribution had no connection to language. Almost any object could have been used to signal linguistic identity, or not. Archaeologists have therefore rejected the possibility that language and material culture are correlated in any predictable or recognizable way.4
But it seems that language and material culture are related in at least two ways. One is that tribal languages are generally more numerous in any long-settled region than tribal material cultures. Silver and Miller noticed, in 1997, that most tribal regions had more languages than material cultures. The Washo and Shoshone in the Great Basin had very different languages, of distinct language families, but similar material cultures; the Pueblo Indians had more languages than material cultures; the California Indians had more languages than stylistic groups; and the Indians of the central Amazon are well known for their amazing linguistic variety and broadly similar material cultures. A Chicago Field Museum study of language and material culture in northern New Guineau, the most detailed of its type, confirmed that regions defined by material culture were crisscrossed with numerous materially invisible language borders.5 But the opposite pattern seems to be rare: a homogeneous tribal language is rarely separated into two very distinct bundles of material culture. This regularity seems discouraging, as it guarantees that many prehistoric language borders must be archaeologically invisible, but it does help to decide such questions as whether one language could have covered all the varied material culture groups of Copper Age Europe (probably not; see chapter 4).
The second regularity is more important: language is correlated with material culture at very long-lasting, distinct material-culture borders.
PERSISTENT FRONTIERS
Persistent cultural frontiers have been ignored, because, I believe, they were dismissed on theoretical grounds.6 They are not supposed to be there, since pre-state tribal borders are interpreted today as ephemeral and unstable. But archaeologists have documented a number of remarkably long-lasting, prehistoric, material-culture frontiers in settings that must have been tribal. A robust, persistent frontier separated Iroquoian and Algonkian speakers along the Hudson Valley, who displayed different styles of smoking pipes, subtle variations in ceramics, quite divergent house and settlement types, diverse economies, and very different languages for at least three centuries prior to European contact. Similarly the Linear Pottery/Lengyel farmers created a robust material-culture frontier between themselves and the indigenous foragers in northern Neolithic Europe, a moving border that persisted for at least a thousand years; the Criş/Tripolye cultures were utterly different from the Dnieper-Donets culture on a moving frontier between the Dniester and Dnieper Rivers in Ukraine for twenty-five hundred years during the Neolithic and Eneolithic; and the Jastorf and Halstatt cultures maintained distinct identities for centuries on either side of the lower Rhine in the Iron Age.7 In each of these cases cultural norms changed; house designs, decorative aesthetics, and religious rituals were not frozen in a single form on either side. It was the persistent oppositio
n of bundles of customs that defined the frontier rather than any one artifact type.
Persistent frontiers need not be stable geographically—they can move, as the Romano-Celt/Anglo-Saxon material-culture frontier moved across Britain between 400 and 700 CE, or the Linear Pottery/forager frontier moved across northern Europe between 5400 and 5000 BCE. Some material-culture frontiers, described in the next chapters, survived for millennia, in a pre-state social world governed just by tribal politics—no border guards, no national press. Particularly clear examples defined the edges of the Pontic-Caspian steppes on the west (Tripolye/Dnieper), on the north (Russian forest forager/steppe herder), and on the east (Volga-Ural steppe herder/Kazakh steppe forager). These were the borders of the region that probably was the homeland of Proto-Indo-European. If ancient ethnicities were ephemeral and the borders between them short-lived, how do we understand premodern tribal material-culture frontiers that persisted for thousands of years? And can language be connected to them?
I think the answer is yes. Language is strongly associated with persistent material-culture frontiers that are defined by bundles of opposed customs, what I will call robust frontiers.8 The migrations and frontier formation processes that followed the collapse of the Roman Empire in western Europe provide the best setting to examine this association, because documents and place-names establish the linguistic identity of the migrants, the locations of newly formed frontiers, and their persistence over many centuries in political contexts where centralized state governments were weak or nonexistent. For example, the cultural frontier between the Welsh (Celtic branch) and the English (Germanic branch) has persisted since the Anglo-Saxon conquest of Romano-Celtic Britain during the sixth century. Additional conquests by Norman-English feudal barons after 1277 pushed the frontier back to the landsker, a named and overtly recognized ethnolinguistic frontier between Celtic Welsh-speaking and Germanic English-speaking populations that persisted to the present day. They spoke different languages (Welsh/English), built different kinds of churches (Celtic/Norman English), managed agriculture differently and with different tools, used diverse systems of land measurement, employed dissimilar standards of justice, and maintained a wide variety of distinctions in dress, food, and custom. For many centuries men rarely married across this border, maintaining a genetic difference between modern Welsh and English men (but not women) in traits located on the male Y chromosome.
Other post-Roman ethnolinguistic frontiers followed the same pattern. After the fall of Rome German speakers moved into the northern cantons of Switzerland, and the Gallic kingdom of Burgundy occupied what had been Gallo-Roman western Switzerland. The frontier between them still separates ecologically similar regions within a single modern state that differ in language (German-French), religion (Protestant-Catholic), architecture, the size and organization of landholdings, and the nature of the agricultural economy. Another post-Roman migration created the Breton/French frontier across the base of the peninsula of Brittany, after Romano-Celts migrated to Brittany from western Britain around 400–600 CE, fleeing the Anglo-Saxons. For more than fifteen hundred years the Celtic-speaking Bretons have remained distinct from their French-speaking neighbors in rituals, dress, music, and cuisine. Finally, migrations around 900–1000 CE brought German speakers into what is now northeastern Italy, where the persistent frontier between Germans and Romance speakers inside Italy was studied by Eric Wolf and John Cole in the 1960s. Although in this case both cultures were Catholic Christians, after a thousand years they still maintained different languages, house types, settlement organizations, land tenure and inheritance systems, attitudes toward authority and cooperation, and quite unfavorable stereotypes of each other. In all these cases documents and inscriptions show that the ethnolinguistic oppositions were not recent or invented but deeply historical and persistent.9
These examples suggest that most persistent, robust material-culture frontiers were ethnolinguistic. Robust, persistent, material-culture frontiers are not found everywhere, so only exceptional language frontiers can be identified. But that, of course, is better than nothing.
Population Movement across Persistent Frontiers
Unlike the men of Wales and England, most people moved back and forth across persistent frontiers easily. A most interesting fact about stable ethnolinguistic frontiers is that they were not necessarily biological; they persisted for an extraordinarily long time despite people regularly moving across them. As Warren DeBoer described in his study of native pottery styles in the western Amazon basin, “ethnic boundaries in the Ucayali basin are highly permeable with respect to bodies, but almost inviolable with respect to style.”10 The back-and-forth movement of people is indeed the principal focus of most contemporary borderland studies. The persistence of the borders themselves has remained understudied, probably because modern nation-states insist that all borders are permanent and inviolable, and many nation-states, in an attempt to naturalize their borders, have tried to argue that they have persisted from ancient times. Anthropologists and historians alike dismiss this as a fiction; the borders I have discussed frequently persist within modern nation-states rather than corresponding to their modern boundaries. But I think we have failed to recognize that we have internalized the modern nation-state’s basic premise by insisting that ethnic borders must be inviolable boundaries or they did not really exist.
If people move across an ethno-linguistic frontier freely, then the frontier is often described in anthropology as, in some sense, a fiction. Is this just because it was not a boundary like that of a modern nation? Eric Wolf used this very argument to assert that the North American Iroquois did not exist as a distinct tribe during the Colonial period; he called them a multiethnic trading company. Why? Because their communities were full of captured and adopted non-Iroquois. But if biology is independent of language and culture, then the simple movement of Delaware and Nanticoke bodies into Iroquoian towns should not imply a dilution of Iroquoian culture. What matters is how the immigrants acted. Iroquoian adoptees were required to behave as Iroquois or they might be killed. The Iroquoian cultural identity remained distinct, and it was long established and persistent. The idea that European nation-states created the Iroquois “nation” in their own European image is particularly ironic in view of the fact that the five nations or tribes of the pre-European Northern Iroquois can be traced back archaeologically in their traditional five tribal territories to 1300 CE, more than 250 years before European contact. An Iroquois might argue that the borders of the original five nations of the Northern Iroquois were demonstrably older than those of many European nation-states at the end of the sixteenth century.11
Language frontiers in Europe are not generally strongly correlated with genetic frontiers; people mated across them. But persistent ethnolinguistic frontiers probably did originate in places where relatively few people moved between neighboring mating and migration networks. Dialect borders usually are correlated with borders between socioeconomic “functional zones,” as linguists call a region marked by a strong network of intra-migration and socioeconomic interdependence. (Cities usually are divided into several distinct socioeconomic-linguistic functional zones.) Labov, for example, showed that dialect borders in central Pennsylvania correlated with reduced cross-border traffic flow densities at the borders of functional zones. In some places, like the Welsh/English border, the cross-border flow of people was low enough to appear genetically as a contrast in gene pools, but at other persistent frontiers there was enough cross-border movement to blur genetic differences. What, then, maintained the frontier itself, the persistent sense of difference?12
Persistent, robust premodern ethnolinguistic frontiers seem to have survived for long periods under one or both of two conditions: at large-scale ecotones (forest/steppe, desert/savannah, mountain/river bottom, mountain/coast) and at places where long-distance migrants stopped migrating and formed a cultural frontier (England/Wales, Britanny/France, German Swiss/French Swiss). Persistent identity depend
ed partly on the continuous confrontation with Others that was inherent in these kinds of borders, as Frederik Barth observed, but it also relied on a home culture behind the border, a font of imagined tradition that could continuously feed those contrasts, as Eric Wolf recognized in Italy.13 Let us briefly examine how these factors worked together to create and maintain persistent frontiers. We begin with borders created by long-distance migration.
MIGRATION AS A CAUSE OF PERSISTENT MATERIAL-CULTURE FRONTIERS
During the 1970s and 1980s the very idea of folk migrations was avoided by Western archaeologists. Folk migrations seemed to represent the boiled-down essence of the discredited idea that ethnicity, language, and material culture were packaged into neatly bounded societies that careened across the landscape like self-contained billiard balls, in a famously dismissive simile. Internal causes of social change—shifts in production and the means of production, in climate, in economy, in access to wealth and prestige, in political structure, and in spiritual beliefs—all got a good long look by archaeologists during these decades. While archaeologists were ignoring migration, modern demographers became very good at picking apart the various causes, recruiting patterns, flow dynamics, and targets of modern migration streams. Migration models moved far beyond the billiard ball analogy. The acceptance of modern migration models in the archaeology of the U.S. Southwest and in Iroquoian archaeology in the Northeast during the 1990s added new texture to the interpretation of Anasazi/Pueblo and Iroquoian societies, but in most other parts of the world the archaeological database was simply not detailed enough to test the very specific behavioral predictions of modern migration theories.14 History, on the other hand, contains a very detailed record of the past, and among modern historians migration is accepted as a cause of persistent cultural frontiers.
The Horse, the Wheel, and Language: How Bronze-Age Riders From the Eurasian Steppes Shaped the Modern World Page 13
