Sir William Jones, a British judge in India, first defined the Indo-European language problem in one famous sentence in 1786. Jones had arrived in Calcutta in 1783 to establish the rule of British law over both the excesses of the English merchants and the rights of their Indian subjects, who obeyed an already functioning and very ancient system of Hindu law. To understand Hindu law, Jones had to learn Sanskrit. His teachers, outstanding Hindu scholars, taught him to read the Vedas, the ancient religious texts that lay at the root of Hindu religion. The Rig Veda, the oldest Vedic text, was known to be more than two thousand years old, but exactly how much older, no one knew. Three years after his arrival in Calcutta, Jones presented a lecture to the Asiatic Society of Bengal, in which he uttered the following oft-quoted words:
The Sanskrit language, whatever be its antiquity, is of a wonderful structure; more perfect than the Greek, more copious than the Latin, and more exquisitely refined than either; yet bearing to both of them a stronger affinity, both in the roots of verbs and in the forms of grammar, than could possibly have been produced by accident; so strong indeed, that no philologer could examine them all three, without believing them to have sprung from some common source, which, perhaps, no longer exists.
Jones concluded that Sanskrit had sprung from the same source as Greek and Latin, the classical languages of European civilization, and added that Celtic, Persian (Iranian), and German probably belonged to the same family. For Europeans the news was startling. The civilization of faraway India turned out to be a long-lost cousin. What was the parent language? Where had it been it spoken and by whom? What historical events made its daughter tongues the dominant languages from Scotland to India? Finally, just how big was the family?
These questions created a debate that has spanned two hundred years and has yet to be resolved. It has inspired episodes of genocide, dry academic discourses, and romantic fantasies. Scholars trying to solve this problem created the discipline of linguistics in the nineteenth century. Their principal interest was comparative grammar, sound systems, and syntax, which provided the basis for classifying languages, grouping them into types, and otherwise defining the relationships between the tongues of humanity, none of which had ever been attempted. They divided the Indo-European language family into twelve major branches, distinguished by innovations in phonology, or pronunciation, and in morphology, or word form, that appeared at the root of each branch and were maintained in all the languages of that branch. The branches included most of the languages of Europe (excluding Basque, Finnish, Estonian, and Magyar); the Persian language of Iran; Sanskrit and its many modern daughters (Hindi and Urdu); and numerous extinct languages, including Hittite and Tocharian. Modern English was assigned to the Germanic branch. The analytic methods invented by these philologists are used to describe, classify, and explain language variation all over the world.
In the 1780s the German Romantic philosopher Johann Gottfried Herder argued that language creates the categories and distinctions through which humans give meaning to the world. Each language therefore generates and is enmeshed in a closed social community, or "folk," that is meaningless to an outsider. After the 1859 publication of Charles Darwin’s Origin of Species, the Romantic conviction that language was a defining factor in identity was combined with new ideas about social evolution. Race, language, and culture were interpreted as a package that endowed some nations with a superior biological-spiritual-linguistic essence and consigned others to the back row. The policies that forced the Welsh (including Sir William Jones) to speak English and the Bretons to speak French were rooted partly in the search for a "pure" national heritage derived from a single heroic and superior race of Anglo-Saxons or Gauls.
The theoretical mother tongue that gave birth to all twelve branches is called Proto-Indo-European. The speakers of the mother tongue soon were molded to fit a national-racial stereotype. The name "Aryan" began to be applied to them because the authors of the oldest religious texts in Sanskrit and Iranian, the Rig Veda and Avesta, called themselves Aryans. The term "Aryan" should be confined only to this Indo-Iranian branch of the Indo-European family. The Vedas were a newly discovered source of mystical fascination in the nineteenth century, however, and in Victorian parlors the name "Aryan" soon spread beyond its proper linguistic confines.
The gap through which the name escaped from India was provided by the Rig Veda itself: the Vedic Aryans described themselves as invaders who had conquered their way into the Punjab. A feverish search for the "Aryan homeland" began. Researchers have placed it confidently in places ranging from India and Pakistan to Russia, Turkey, central Europe, and even the North Pole and Atlantis. Some homelands were proposed not for innocent reasons but to provide a historical precedent for nationalist or racist claims to privileges and territory. In the 1920s the German scholar Gustaf Kossinna attempted to demonstrate on archaeological grounds that the Aryan homeland lay in northern Europe, centered in Germany. Kossinna illustrated the prehistoric migrations of the Indo-Germanic Aryans with neat black arrows that swept east, west, and south from his presumed Germanic core. Nazi armies followed his pen twenty years later.
The fundamental errors that led an obscure linguistic mystery to erupt into racial genocide were the equation of race with language and the assignment of evolutionary superiority to certain language and race groups. Indo-European, the linguistic phenomenon, became the "Indo-Europeans," a racial-spiritual fantasy. Prominent linguists have always pleaded against these ideas. The Aryans themselves, according to their own texts, used Aryan as a religious-linguistic category. The Rig Veda was a ritual canon, not a racial manifesto. Making the proper sacrifices to the right gods, which required performing the traditional prayers in the traditional language, made a person an Aryan.
Any attempt to solve the Indo-European problem has to begin with the realization that the term Proto-Indo-European refers to a language community. Race, poorly and inconsistently defined, cannot be linked in any predictable way with language. Because definitions are cultural, scientists cannot provide a true boundary between races. Moreover, archaeologists have their own, quite different definitions of race, based on traits of the skull and teeth that often are invisible in a living person. However race is defined, languages are not normally sorted by race—all racial groups speak a variety of languages. Culture, however, often is associated with lan-guage—the language a person speaks can lead others to make assumptions about one’s character, religion, dietary preferences, and so on. These are stereotypes, of course, and people often confound them. How, then, do we connect language with culture in a reliable and predictable way?
LANGUAGE AND MATERIAL CULTURE
Many archaeologists think that it is impossible to identify a prehistoric language group, because language is not reflected in any consistent way in material culture. People who speak different languages might make houses or pots in the same way, and people who speak the same language often make pots or houses in different ways. Likewise, a language can spread without a corresponding change in material culture, and vice versa. Language and culture are correlated predictably under some circumstances, however. We have erred in trying to find a single class of material culture that correlates reliably with language; we should focus instead on frontiers.
Where we see a robust frontier represented in material culture—not just different pots but also different houses, graves, cemeteries, town patterns, ritual icons, diets, and dress designs—that persists in the same location for centuries or millennia, it tends to be a linguistic frontier as well. Persistent ethnol-inguistic frontiers seem to occur under relatively few conditions, principally at ecological boundaries and at the end point of certain kinds of migrations. There was, for instance, a persistent ethnolinguistic frontier between English-speaking immigrants and the indigenous Celtic Welsh in Wales. This divide separated populations that spoke distinct languages (Welsh/English), built particular kinds of churches (Celtic/Norman English), managed agriculture in varying ways with different tools, used disparate systems of land measurement, employed different standards of justice, and maintained a wide variety of distinctions in dress, food, and custom. In cases such as this, where a clear material culture frontier persists in the same place for hundreds of years, language tends to be correlated with the boundary. This insight permits us to identify at least a few probable linguistic frontiers on a map of purely archaeological cultures, a critical step in finding the Proto-Indo-European homeland.
HOW RECONSTRUCTION WORKS
Historical linguistics gave us not just static classifications but also the astounding ability to reconstruct at least parts of early languages for which no written evidence survives. The methods that make this possible rely on regularities in the way sounds change inside the human mouth. For example, the sound k, as in "kiss" (or any consonant made with the back of the tongue), followed by the sound e, as in "set" (or any other vowel made with the tip of the tongue), is likely to shift forward on the palate toward the front vowel—to ts- and then to s.
This happened when the Latin word centum (meaning "hundred" and pronounced kentum) became the old French cent (pronounced tsohnt) and then the modern French cent (pronounced sohnt). A shift in the other direction, backward on the palate from ce- to tse- to ke-, is quite unlikely. Given the terms centum and cent, and no other historical information about them, we could say that the sound of the Latin word makes it the older form, that the modern French form could have developed from it according to known rules of sound change, and that an intermediate pronunciation tsohnt probably existed before the modern form appeared. Both words are from the same Indo-European branch, Italic, which produced Latin and from Latin all the Romance languages, including French. Indo-European words for "hundred" from different branches of the language family can be compared in this way to see whether all can be derived from a single hypothetical ancestral word. The proof that Latin centum in the Italic branch and Lithuanian shimtas in the Baltic branch are related in this way, that they are cognates, is the construction of the ancestral root.
Root forms converge on one unique "root" sequence of sounds that could have evolved into all of them by known rules. The comparative method cannot force a regular reconstruction on an irregular set of sounds; for example, if terms in several branches have borrowed sounds from local languages, those borrowings might not fit the expected rules of regular sound change. For this reason, much of the Proto-Indo-European vocabulary, perhaps most of it, never will be reconstructed. Regular groups of cognates permit archaeologists to reconstruct a Proto-Indo-European root for the word "eye" but not for "eyebrow," for "snow" but not for "rain," and for "foot," but not for "hand." Proto-Indo-European certainly had such words, but we cannot safely reconstruct how they sounded.
Still, linguists have reconstructed the sounds of thousands of other words. A reconstruction based on cognates that survive in eight different Indo-European branches, like *k’mtom-, the Proto-Indo-European root for "hundred," is much more reliable and probably more true than one based on cognates in just two branches. The accuracy of reconstruction has been confirmed by archaeology. Three separate archaeological discoveries of ancient inscriptions have provided opportunities to test whether the sounds that had been suggested by linguists for ancient phases of three Indo-European languages—Hittite, Mycenaean Greek, and archaic German—actually appeared in the inscriptions. In all three cases the linguists were proved correct.
For example, linguists working on the development of Greek had proposed *kw (pronounced like the kw- in "queen") as the ancestral sound that developed into Greek t before a front vowel or p before a back vowel. The reconstruction remained hypothetical until the discovery and decipherment of the Mycenaean Linear B tablets, which showed that the earliest form of Greek, Mycenaean, had the predicted kw, where later Greek had t or p before front and back vowels. Such discoveries have confirmed that many reconstructed terms can be regarded as more than just abstractions.
The extent to which reconstructed terms can be thought of as real is the subject of debate. We should not imagine that reconstructed Proto-Indo-European was ever spoken anywhere. After all, it is fragmentary (and most of the language this reconstruction represents never will be known). The reconstructed language, which averages centuries of change, is homogenized, stripped of many of the peculiar sounds of its individual dialects. The same can be said of the English language as presented in Merriam-Webster’s Dictionary. This dictionary contains the word "ombre" (a card game popular in the seventeenth and eighteenth centuries) as well as "hard disk" (post-1978). Thus, its vocabulary brings together about four hundred years of English. No person has ever spoken this version of English. Nevertheless, many of us find the dictionary useful as a guide to spoken English. Reconstructed Proto-Indo-European is similar—it might not be a true language, but it certainly refers to one.
If a reconstruction is based on a large and diverse set of cognates from both Europe and Asia and includes a cognate from an ancient language, the only conclusion we can draw is that such a term existed in the parent language. Proto-Indo-European is a partial grammar and a partial set of pronunciation rules attached to the abundant fragments of a very ancient dictionary. To an archaeologist, that is more valuable than a roomful of potsherds. The reconstructed vocabulary of Proto-Indo-European is a guide to the thoughts, concerns, and material culture of actual people.
THE PRIZE: THE RECONSTRUCTED VOCABULARY
The reconstructed vocabulary includes word clusters that suggest that the speakers of Proto-Indo-European were farmers and stockbreeders: we can reconstruct words for "bull," "cow," "steer," "ram," "ewe," "lamb," "pig," and "piglet." There is a term for "butter" and perhaps one for "cheese."
When these people led their cattle and sheep out to the "field," they walked with a faithful "dog." They knew how to "shear wool," which they used to "weave" textiles. They tilled the earth with a scratch plow, or "ard," which was pulled by "oxen" wearing a "yoke." They turned their threshed grain into flour by "grinding" it with a hand "pestle," and cooked their food in clay "pots." They had "bees" and "honey."
They divided their possessions into two categories: items that could be moved and those that could not. In fact, the root for "movable wealth" (*peku-, the ancestor of such English words as "pecuniary") became the term for "herds" in general. Terms for male family members suggest that they inherited their rights and duties through the father’s bloodline (patrilineal descent). The absence of equivalent terms for the wife’s family indicates that wives lived with the husband’s family after marriage (patrilocal residence). "Chiefs" probably supervised political relations within their kin group, and there were formally instituted "warrior bands." A male sky deity ("sky father"), a thunder god, and a pair of sky twins were worshipped. Two senses of the sacred seem to have been recognized: "that which is imbued with holiness" and "that which is forbidden." Many of these practices and beliefs are simply unrecoverable through archaeology. The proto-vocabulary offers the hope of retrieving some of these details of ritual and custom. Reconstructed Proto-Indo-European is a long, fragmentary word list left by people who lived in a time and place unilluminated by any other kind of textual evidence. The list becomes useful, however, only if we can determine from where it came. To do that, we must locate the Proto-Indo-European homeland. First, however, we must know when Proto-Indo-European was spoken.
DATING PROTO-INDO-EUROPEAN: THE TERMINUS POSTQUEM A dictionary is dated easily by its most recent words. The terminus post quem, the date after which reconstructed Proto-Indo-European must be placed, can be established in much the same way, by the vocabulary. Words for things that were invented at a known date, such "wagons" and "wheels," can have existed in a language spoken only after that date. Proto-Indo-European began to split into different branches after the date indicated by these reconstructed words.
The most important words from this perspective are the reconstructed words for the basic tools (ard and pot) and products of agriculture (field, grain, cow, bull, calf, ram, ewe, lamb, milk, and cheese), for "wagons" and their parts, and for "wool." The agricultural vocabulary establishes that the speakers of Proto-Indo-European could not have been hunters and gatherers.
The term for "wool" provides a more precise date. The reconstruction is based on cognates in almost all branches from Welsh to Indic, so it certainly was in the vocabulary before the breakup into branches began. Wool sheep are mutants, bred to produce fleeces made entirely of the fine, curly fibers that wild sheep originally had just as an undercoat beneath their long, hairy coats. The best estimate is that wool sheep were bred in Mesopotamia about 4000 b.c. and then spread westward into Europe, eastward into Iran and India, and northward into the Caucasus Mountains and the Russian/ Ukrainian steppes. From the wool perspective, Proto-Indo-European was spoken after 4000 b.c.
The vocabulary for wagons provides stronger guidance. At least five terms can be reconstructed with great confidence: two nouns for "wheel," another for "axle," a noun for "harness pole" (a "thill"), and a verb meaning to "go or convey in a vehicle." Cognates for these terms occur in all the major branches of Indo-European. Furthermore, all words but "thill" are based on recognizable Proto-Indo-European roots. For example, one reconstructed "wheel" root, *kwekwlos looks very much as if it was created from another root, *kwel-, a verb that means "to turn." Thus, *kwekwlos is not just a random string of phonemes; it is "the thing that turns." This kind of cross-referencing within a reconstructed vocabulary increases confidence in both reconstructions. Finally, most of the reconstructed "wagon" terms turn out to have a kind of vowel structure called an "o-stem" that generally is thought to identify a late stage in the development of Proto-Indo-European, so the vocabulary is internally consistent in its phonology. The speakers of Proto-Indo-European had wagons and talked about them using words of their own invention.
Based on archaeological evidence it is fairly certain that the wheel-and-axle principle was invented after 4000 b.c., probably after 3500 b.c. A track preserved under a barrow grave at Flintbek in northern Germany might have been made by wheels and might be as old as 3600 b.c. All other evidence for wheeled vehicles—written signs, artistic images, three-dimensional clay models, and wheels them-selves—first appears in the archaeological record between 3500 and 3000 b.c. Thus, late Proto-Indo-European must have been spoken after 4000 and possibly after 3500 b.c. Before then, no language had words for "wagons" or "axles."
DATING PROTO-INDO-EUROPEAN: THE TERMINUS ANTE QUEM
Proto-Indo-European has been created on the basis of systematic comparisons of all of the known Indo-European daughter languages. The terminal date for the reconstructed language—the date after which our reconstructed form becomes an anachronism—should be related in some way to the separation of its oldest independent branches. If Proto-Indo-European is defined as the language that was ancestral to all of the Indo-European daughters, then it is the oldest reconstructable form. The later daughters did not evolve directly from Proto-Indo-European but from some intermediate, evolved set of late Indo-European languages that preserved aspects of the mother tongue and passed them along.
Internal evidence—the appearance within a branch of phonological archaisms and innovations not shared with other branches—helps identify the oldest branches. All of the branches cannot be placed with confidence in a sequence, but most linguists agree that Anatolian was the first branch to separate. It appears in the oldest known inscriptions in any Indo-European language, dated 1920-1820 b.c., at Karum Kanesh II in Turkey. Anatolian is so archaic and idiosyncratic that it must represent a very early stage in Proto-Indo-European. Italic and Celtic also seem archaic and should be included within the next set of branches to form, although their earliest inscriptions are much later, about 600500 b.c. Reconstructed Proto-Indo-European becomes increasingly anachronistic after the set of separations that includes Italic and Celtic. Greek, documented in Linear B by 1450 b.c., probably split off from a more evolved set of Indo-European dialects and languages centuries after the dialects that led to Italic and Celtic. The sound changes that identify Indo-Iranian emerged after the separation of the Greek branch. Old Indic Sanskrit had emerged from Indo-Iranian by 1450 b.c., the date of the oldest Sanskrit inscriptions in the Mitanni texts. Common Indo-Iranian must be older than 1450 b.c., at least as old as 1700 b.c.
The older separations—Greek, Italic and Celtic, and Anatolian—form a sequence that must predate 1700 b.c. Although their exact place in the sequence is debated, Germanic and Tocharian certainly also split away before Indo-Iranian. The latest possible date for Proto-Indo-European can be set at about 2700 b.c., leaving just a millennium—almost certainly not enough time—for the evolution ofAn-atolian, Italic, Celtic, Mycenaean Greek, Germanic, Tocharian, and Indo-Iranian. Long before 1700 b.c., the language that has been reconstructed as Proto-Indo-European had evolved into something else or, more accurately, into a variety of late dialects that continued to diverge in various ways in different places. By at least 2000 b.c., and probably long before, what we know as Proto-Indo-European was a dead language.
LOCATING THE HOMELAND
It has been proposed that Proto-Indo-European was spoken in Anatolia in about 7000-6500 b.c. and then spread through Europe and eastward across the Eurasian steppes with agriculture and animal herding. This idea is appealing, but it cannot be correct. It requires a breakup into daughter branches in about 6500 b.c., when the first pioneer Anatolian farmers migrated to Greece, with subsequent branch formations and separations as the farming economy was carried northward into temperate Europe between 6000 and 3000 b.c. By 4000-3500 b.c. the Indo-European language family should have been quite diverse, according to this proposal.
For this chronology to be correct, we would have to assume that the wool and wagon vocabularies were created long after the breakup of the Indo-European branches and then were borrowed into each daughter branch. Linguists generally have rejected this accommodation, however, because the vocabulary does not exhibit phonological traits indicating that it was created within a later Indo-European language. No other technical vocabulary is known to have been borrowed in a standardized form into the Indo-European daughter languages after they were scattered from Scotland to India— for example, the vocabulary for iron technology is quite diverse in the daughter languages. The linguistic evidence is against the Anatolia solution.
If not Anatolia, then where? Linguists have long tried to find animal or plant names in the reconstructed vocabulary that refer to species that lived in just one part of the world. The reconstructed term for "salmon," *lok*s, was once famous as a proof that the Aryan homeland lay in northern Europe. Animal and tree names seem to narrow and broaden in meaning easily, however. They are even reused and recycled when people move to a new environment. The most specific meaning that linguists would now feel comfortable ascribing to the reconstructed term *lok*s- is "trout-like fish." Most linguists agree that the fauna and flora designated by the reconstructed vocabulary are temperate-zone types (bear, otter, beaver, lynx, and horse), not Mediterranean (cypress, olive, and laurel) or tropical (monkey, elephant, palm, and papyrus).
"Bee" and "honey," however, are very strong reconstructions. The term for "honey," *medhu-, also was used for an intoxicating drink that played a prominent role in Proto-Indo-European rituals. Honeybees are not found in northern Eurasia east of the Ural Mountains, across Siberia, because the hardwood trees (lime and oak, particularly) that honeybees prefer as nesting sites become rare east of the Urals. That removes all of Siberia and much of northeastern Eurasia from contention, including the Central Asian steppes of Kazakhstan.
The horse, *ek*wo-, is solidly reconstructed and seems also to have been a potent symbol of divine power for the speakers of Proto-Indo-European. Although horses lived in small, isolated pockets throughout prehistoric Europe, they were rare or absent in the Near East, Iran, and the Indian subcontinent and were numerous and economically important in the daily meat diet only in the Eurasian steppes. The term for "horse" removes the Near East, Iran, and the Indian subcontinent from serious contention, and it encourages us to look closely at the Eurasian steppes.
Finally, we can use the information that the speakers of Proto-Indo-European were familiar with agriculture and herding. In the northern forest zone of Russia and the Baltic, economies based on fishing, hunting, and gathering were retained until after 2500-2000 b.c. The switch to herding and farming happened after 2000 b.c. in the Siberian forest zone east of the Urals. That, too, eliminates Siberia and Kazakhstan and casts doubt on the Russian-Baltic northern forest zone. We are left with temperate Europe and the western steppes and perhaps the temperate parts of the Caucasus Mountains and Anatolia.
WHO WERE THE NEIGHBORS?
The neighbors of the speakers of Proto-Indo-European can be identified through words and forms borrowed between Proto-Indo-European and other language families. Proto-Indo-European shows strong links with Proto-Uralic, a key ancient language of the northern Russian forests, and weaker links with a language ancestral to Proto-Kartvelian, spoken in the Caucasus Mountains. Proto-Indo-European and Proto-Uralic shared two kinds of linkages. One, revealed in shared pronouns, noun endings, and basic vocabulary could be ancestral: the two proto-languages seem to have shared an ancient common ancestor, perhaps spoken by Ice Age hunters east of the Carpathians. The relationship is so remote, however, that it can barely be detected.
The other link between the two languages seems cultural: some Proto-Indo-European words (to wash, water, to give, merchandise, to fear) were borrowed by the speakers of Proto-Uralic, perhaps through a shared trade jargon. The fact that the reconstructed roots are similar in phonological form and meaning indicates that they were loans rather than inheritances.
These two kinds of linguistic relationship—a possible common ancestral origin and interlangu-age borrowings—suggest that the Proto-Indo-European homeland was situated near the homeland of Proto-Uralic. Uralic is a broad language family, like Indo-European. Its daughter languages are spoken across the northern forests of Eurasia from Finland to the Siberian Pacific. The Proto-Uralic homeland is thought to have been in the southern part of the forest zone near the Ural Mountains. Many researchers believe that the best case can be made for a homeland west of the Urals, and some argue for the east side. Almost all agree that Proto-Uralic was spoken in the forests between the Oka River on the west and the Irtysh River on the east, probably before the adoption of a herding economy (2500-2000 b.c.). This leaves a possible contact zone south or southwest of the Ural Mountains.
Coincidentally, this is the direction in which we find the second neighbor. Proto-Indo-European interacted with the languages of the Caucasus Mountains, primarily those that are classed as southern Caucasian or Kartvelian, the family that produced modern Georgian. Many terms have been proposed as loanwords to Proto-Indo-European from Proto-Kartvelian (and even Semitic). The few such loanwords that are widely accepted (such as those for "silver" and "bull") might be words that were carried along trade and migration routes far from the Semites’ Near Eastern homeland. The phonology of the loans suggests that none of these language contacts was direct—all of the loanwords passed through unknown intermediaries between the known three. One intermediary is required by chronology, since Proto-Kartvelian generally is thought to have existed after Proto-Indo-European and Proto-Semitic.
Who, then, were the neighbors? Proto-Indo-European exhibits strong links with Proto-Uralic and weaker links with a language ancestral to Proto-Kartvelian. The speakers of Proto-Indo-European lived between the Caucasus and Ural Mountains but had deeper linguistic relationships with the people who lived around the Urals. The region between the Caucasus and the Urals is the Russian and Ukrainian steppe—a place long identified as a strong candidate for the Indo-European homeland. Does contemporary archaeology support this solution?
THE ARCHAEOLOGY OF THE PROTO-INDO-EUROPEAN HOMELAND
In the North Pontic region, north of the Black Sea, the first farmers were Cri§ culture pioneers who migrated from southeastern Europe and the Lower Danube Valley. Their arrival created a cultural frontier northwest of the Black Sea in modern Ukraine, between the Dnieper and Dniester Rivers, that persisted for 2,500 years, from about 5800 to 3500 b.c. Two distinct cultural systems existed side by side, east and west of the Dnieper-Dniester frontier. Substantial differences in material culture distinguished the immigrants and their cultural descendants (Cri§, Linear Pottery, and Tripolye) from the indigenous societies and their cultural descendants (Dnieper-Donets, Mariupol, Sredny Stog, and Yamnaya). The two traditions differed in house forms; settlement types; economy; ceramic style, decoration, and technology; stone tool types; mortuary rituals; the presence or absence of female figurines; and metallurgical techniques—in other words, they maintained distinctions in almost every aspect of material culture for millennia.
Another persistant cultural frontier coincided with an important ecological frontier. It separated the foragers of the northern forest zone west of the Ural Mountains, the probable Proto-Uralic homeland, from the cattle herders and sheepherders to the south, in the Pontic-Caspian steppes. This economic-ecological frontier, too, persisted for 2,500 years, from about 5000 to about 2500 b.c. A bundle of cultural distinctions defined the forest/steppe frontier, including variations in house forms, pottery types, stone tools, and burial practices.
Finally, on the eastern edge of the Pontic-Caspian steppes there was yet a third clear and persistent cultural frontier, a north-south line extending from the southern slopes of the Ural Mountains to the deserts north of the Caspian Sea. Long after herding was adopted in the Pontic-Caspian steppes (about 5000 b.c.), the societies of the Kazakh steppes to the east remained foragers—such groups as the Atbasar, Surtanda, and Tersek-Botai. They made quite different kinds of pots and stone tools, did not use cemeteries, and had distinctive house forms. Like the first two frontiers, this one persisted for at least 2,500 years, until about 2500-2000 b.c. In all three cases it is clear from published archaeological reports that the cultures on either side of the frontiers knew and interacted with each other, but rather than assimilating, they remained distinct for millennia.
The historic cases cited earlier suggest that these material-culture frontiers almost certainly were linguistic frontiers as well. They surrounded and enclosed precisely the region identified in other ways as the probable Proto-Indo-European homeland. While we should not make the mistake of assuming that all of the people of the Pontic-Caspian steppes were Indo-European-speakers, we can safely suppose that Proto-Indo-European was spoken somewhere in the Pontic-Caspian region between 4000 and 2000 b.c.
Archaeology thus reveals a set of cultures in this region at this time that fits all of the requirements of the reconstructed vocabulary: they sacrificed horses, cattle, and sheep; cultivated grain at least occasionally; drove wagons; and expressed institutionalized status distinctions in their funeral rituals. They occupied a part of the world, the steppes, where the sky is by far the most striking part of the landscape, a fitting environment for people who believed that their most important deities lived in the sky. Archaeological evidence for migrations from this region into neighboring regions is well established. The sequence and direction of these movements match those suggested by linguistics. Two movements toward the west could represent the detachment of the pre-Anatolian (Cernavoda I into the eastern Balkans) and then the pre-Italic/pre-Celtic dialects (Yamnaya into eastern Hungary), followed by a third movement toward the east (Sin-tashta-Petrovka) that could represent the detachment of Indo-Iranian. The archaeology of the region provides a new window onto the lives of the people who spoke Proto-Indo-European and the process by which it became established and began to spread.
Warfare has been defined in both broad and narrow terms. In the broad view, warfare is armed conflict between any social or political units. In this view, societies as diverse as bands of Mesolithic hunter-gatherers, Neolithic farming tribes, Celtic high chiefdoms, petty states, and the Roman Empire can conduct war. The narrow definition confines war to state-level societies—those with the hierarchical organization to centrally direct armies that are led by, if not consisting wholly of, full-time military specialists. This constricted view is historically misleading and anthropologically absurd. Roman legionnaires routed and killed by warriors of a Celtic hill tribe were just as vanquished as those beaten by a Persian army. Indeed, it took the Romans more time and manpower to conquer the small Celtiberian tribes of northern Spain (four to six legions and two hundred years of continuous fighting) than it took them to subdue Macedonia and Greece (two to four legions and, in total, about twenty years of intermittent combat). Under the narrow definition, the very terms "prehistoric warfare" and "tribal warfare" are oxymorons, which means that recent tribes such as the Apache, Maori, and Taureg never made war. For many reasons, then, the broad definition is preferable and is used here.
Archaeological evidence for warfare is recovered in four categories: human remains; fortifications; weapons and armor; and artistic representations. Only when classical authors begin to describe warfare of their societies with the so-called barbarians of more northerly Europe were there written accounts to supplement the physical evidence revealed by archaeology.
HUMAN REMAINS
Human remains often bear witness to the traumas caused by weapons. These include sword cuts, the indentations made by stone axes and adzes, and depressed fractures made by maces or other blunt-force weapons. The most common type of weapon traumas found on victims of early warfare are embedded stone or bone projectile points. Any of these types of traumas can be considered the cause of death, especially when there are no signs that the wound healed.
Archaeological evidence for warfare can also be seen in the treatment of the body after death. Bodies of war victims were often left where they fell or dumped into mass graves. Bodies that were not buried soon after death often suffered mutilation by animal scavengers. War victims were also mutilated in the course of hostilities. One common type of peri-mortem (i.e., about the time of death) mutilation is known as "overkill," which involves striking the victim with numerous blows or multiple projectiles— any one of which would have been fatal. Another kind of mutilation involves the taking of war trophies—heads, hands, or other body parts—leading to burials with either too few or too many body parts for the individual interred. There is also sometimes evidence for cannibalization of the victims. These types of mutilation suggest that the victors wanted to either humiliate their victims or to acquire the victim’s spiritual power.
