Read Languages In the World Online
Authors: Julie Tetel Andresen,Phillip M. Carter
Languages In the World (31 page)
No homeland has received more attention than that of the Proto-Indo-Europeans. Early philologists based their hypotheses solely on linguistic evidence, in the kinds of
flora and fauna that could be reconstructed to belong to the source language as a way to narrow down the geographic range. The Proto-Indo-Europeans knew bears, for instance. They also knew wolves, rabbits, beavers, mice, weasels, deer, geese, ducks, snakes, sheep, goats, pigs, dogs, and horses, among other animals; as for plants, they encountered oak, beech, pine, birch, and willow trees. In addition, the kinds of animals and plants they did not encounter are also significant: no anciently common Indo-European words can be reconstructed for elephant, camel, lion, tiger, monkey, crocodile, rice, bamboo, or palm. Because there are common words for snow and freezing cold, more or less widely spread over the Indo-European territory, the conclusion is that the homeland had to be in a place that was fairly cold.
An early group of philologists put special consideration on two words, in particular:
bee
and
beech
. The ability to reconstruct a PIE word for âbee' is of interest because the honeybee is native to Europe but not to the locations in Asia ever considered as possible PIE homelands; and the honeybee is relevant because PIE also has a reconstructed word for âhoney.' The beech tree was at one time considered a particularly promising way to pinpoint the homeland because the range of the common beech is relatively limited: it is more or less confined to central Europe and is not native east of Poland and the Ukraine. Unfortunately, no one can be sure the PIE word that developed in various IE languages and came into English as
beech
referred to the tree we know as the beech tree. The word and the thing correspond in Latin and the Germanic languages. However, the word means âoak' in Greek, and in other languages it came to designate âelder' and âelm.' There is the further problem of determining whether the current native range of the beech tree is the same as its range 6 kya.
The point here is that the lexical testimony for the homeland is obviously important, but it also leaves wriggle room. In addition, evidence from two now-extinct branches of the Indo-European stock, Hittite and Tocharian, can and should come into play, even if there are not enough records of them to weigh in on the bee/beech issue. The Hittites had an empire about 1600 years ago on the Anatolian plateau of what is present-day Turkey. The Tocharians lived around 2000 years ago in Xinjiang, the northernmost and westernmost province of China. Clearly, these two branches expand the scope of the geographic search.
At present, there are at least seven competing hypotheses for the PIE homeland, among which we will consider only three:
The first is archeologically based, and it is the Kurgan theory. It was first proposed in the 1970s by Lithuanian-American archeologist Marjia Gimbutas, who placed the PIE homeland in the Russian steppes, perhaps present-day Ukraine, north and a bit east of the Black Sea, around the fifth millennium BCE. The name
kurgan
comes from a Turkic loanword in Russian for the type of burial mounds that are found at archeological sites in the region. The story here revolves around the reconstructed words for âhorse,' âwheel,' and âaxle,' and the question of whether or not the Indo-Europeans had domesticated the horse and had wheeled conveyances; and these technological innovations are seen as the cause of the Indo-European expansion. In a further effort
to harmonize our terminology with our time-depths, we reserve the term
expansion
to refer to the movements of proto-speakers away from a homeland, that is, movements that have taken place in near prehistory. In the older version of the PIE expansion, the Proto-Indo-Europeans were represented as proud horsemen, noble warriors absorbing or eliminating the pastoralists/peasants they encountered as they moved into new territories. Gimbutas reinterpreted this earlier representation in a more violent light and argued that these later PIE warrior-like Indo-Europeans replaced an earlier more peaceful, female-centered society.
The second hypothesized site for the PIE homeland is the Anatolian steppes of present-day Turkey around the seventh millennium BCE. The Anatolian hypothesis was first promoted by British archeologist Colin Renfrew, and it is based on a larger hypothesis that in the past 10,000 years, most of the language families that now exist derive from populations who developed or acquired systematic forms of food production. When these populations took up agriculture and/or animal husbandry, they dispersed and displaced the hunter-gatherers into whose territories the farmers expanded. Population densities of farmers are 10â100 times higher than those of hunter-gatherers, within the same environment. This is because farmers, with their livestock, grains, vegetables, and orchards, have much more food available to them. Hunter-gatherers, in contrast, have to spread out over large areas to hunt the game and to find the edible wild plants.
11
Major agricultural-origin regions include China, Mesoamerica, sub-Saharan Africa, as well as Southeast Asia, and these regions are hypothesized to be the homelands of many of the major language stocks of today, as we will see below. The Middle East is also such a region, and so a population moving from there into the Anatolian plains is not difficult to imagine. The acquisition of agricultural practices not only supported larger populations than those of the hunter-gatherers
12
but also prompted language dispersal. Thus, archeologist Peter Bellwood believes that Indo-European languages spread not by language shift on the part of populations the Indo-Europeans encountered â in his opinion, not necessarily sedentary hunter-gatherers â but by the movement of a significant number of Indo-European speakers themselves (2001:33).
The third candidate for the PIE homeland is farther east, in fact east of the Caspian Sea, specifically in the vicinity of Bactria-Sogdiana, which is around the present-day TajikistanâUzbekistan border. The homeland is proposed at a time-depth of 6â7 kya. This is Johanna Nichols's position (Nichols 1999). She has long maintained that while many historical linguists assume that Indo-European is a normative language family, it is in fact the odd man out. For one thing, the usual amount of branches language families around the world show at this time-depth is up to three; Indo-European has a dozen or more. In addition, PIE shows a typological diversity that, say, the Austronesian, Turkic, and Mongolic families do not. Third, even before PIE broke up, an
internal phonetic division can be identified, and it split along geographic lines.
13
From what is known from archeological and textual evidence of a mythology and a poetic tradition, along with what can be reconstructed socially, culturally, archeologically, linguistically, and even ecologically, only a very particular homeland could produce a language stock with such unique characteristics.
Nichols identifies Bactria-Sogdiana as the most likely epicenter from which the Indo-Europeans spread, since this location puts them at the frontier of Near Eastern civilization and gives them access to the Eurasian steppe across which they would eventually fan out. This particular locus reconciles the conflicting lexical evidence about the PIE homeland, which suggests both urbanized and nomadic practices as well as knowledge of both dry grasslands and forests. As opposed to Gimbutas, Nichols sees the Proto-Indo-Europeans on the move not as warriors but rather as nomadic pastoralists and herders who introduced their language to the nonmoving populations they encountered in their expansion and induced language shift in those populations.
We, the authors, have not brought forth the arguments of these three accounts in order to decide among them. Rather, we mention them to point out that even for an extremely well-studied language stock, not all researchers agree about how the available evidence for the homeland lines up or even about what the available evidence is or should be. We also note that population genetics, once again, may eventually help decide the issue. Already in the 1970s, there were genetic clues that the first farmers living in Europe had links to ancient farming cultures in the Near East. More recently, several large-scale studies of bones show that:
- only 30% of modern Europeans show genetic linkages to the earliest farmers coming into Europe about 8500 years ago;
- hunter-gatherers and farmers lived side by side for a very long time in Europe;
- groups moved around quite a bit; that is, it is not the case that when a group settled in a particular place in Europe 8, 6, or even 4 kya, they necessarily stayed there;
- the Neolithic transition from Paleolithic hunting and gathering to farming was prolonged and nonlinear, with successive waves of farmers coming into Europe who learned to coexist with the populations already there (Balter 2013a, 2013b, Brandt et al. 2013).
Our tentative conclusion is that the very first farmers entering Europe from the Anatolian plain were not speaking a language from the Indo-European stock.
With the exception of the Aboriginal Australians and perhaps the Melanesians of Papua New Guinea and the Khoisan of southwest Africa, the settlements of the homelands of the various established families have occurred in relatively recent times, namely within the last 10,000 years or so. These settlements coincide with the shift from the Old Stone Age to the New Stone Age, along with the introduction of agriculture and
metalworking. In terms of geological epochs, these settlements occurred during the shift from the Pleistocene, which extends back 2,588,000 years and covers the period of the earth's glaciations, to the Holocene, which began 11,700 years ago at the end of the last Ice Age and is associated with warmer periods.
14
As we have just seen, the identification of the homeland for a well-studied stock such as Indo-European is far from settled. Quite a number of other issues in historical linguistics also remain unsettled, beginning with the terminology. Historical linguist Lyle Campbell distinguishes between the terms
family
and
stock
on the basis of evidence. For Campbell (1998:187), a family is a grouping whose evidence of relationship is strong, while a stock is a grouping whose evidence of relationship is weak. Indeed, Campbell even cautions against using terms such as
stock
,
phylum
, and
macrofamily
because they cause confusion. However, in Chapter 3, we introduced Johanna Nichols's definitions, which are based not on evidence but on time. For her, a family has a time-depth of 2500â4000 years and a stock a time-depth of 5000â10,000 years, while a phylum refers to a group of languages that may have more or less evidence of being related at a time-depth greater than 10,000 years. We continue to follow Nichols's usages here.
Going beyond Indo-European, we offer a historical and geographic overview for the rest of the stocks (with representative families and/or languages noted in parentheses) in the following subsections.
The San may have been one of the first populations to differentiate from the East Africans who eventually peopled the world, given that the San Y-haplogroup suggests a separation at least 60 kya. Instead of migrating north and east as did the East Africans, the ancestors of the Khoisan migrated south. Because the time-depth puts us well out of the range of 10,000 years that we are saying is the norm to confidently propose a stock, Khoisan may be only a convenient geographical cover term.
The homeland is in West Africa. Although the Sahara Desert is now one of the driest places on earth, between 11 and 5 kya it was a green region containing savannah grasslands and humid tropical forests. The increasingly arid conditions began in the west 5 kya. A Bantu expansion began, propelled by agricultural practices, about 3 kya in modern Cameroon and eastern Nigeria such that now the lower half of Africa is dominated by speakers from this branch, called NigerâCongo B. Western sub-Saharan Africa is peopled by speakers of NigerâCongo A, such as Wolof, spoken primarily in Senegal, and Yoruba, spoken throughout West Africa. These languages are known for vowel harmony, tone systems of three or more levels, vowel nasalization, SOV word order, and a comparative using âexceed' as the comparative marker.
The eastern Sudan is the hypothesized homeland for Nilo-Saharan probably 8 kya. Languages in this stock are dotted all over Eastern, Central, and West Africa, lacing among Afro-Asiatic languages to the north and NigerâCongo
languages to the south. The structural features of the languages are variable. Some are tone languages, and some not. Most have SVO word order, but there are those with SOV as well as some with VSO order. Their speakers are equally mixed in lifestyle, with some seminomadic pastoralists and others who are agriculturalists. The ethnogroup known as Fur were thrust onto the world stage in 2003, when they were caught in the midst of a geopolitical conflict known as the War in Darfur, which will be discussed in Chapter 9.
Europe was home to human habitation well before the Indo-Europeans arrived beginning about 4 kya. We know this from the cave paintings in France and Spain, some of which date back to 30 and 40 kya, and from the peat bogs in Northern Europe, whose excavations show that people lived there as long ago as 10 kya. Two language isolates in Europe, Basque, and now-extinct Etruscan, are a puzzle. Not enough is known about Etruscan, which was spoken on the Italian peninsula before the arrival of the Indo-Europeans, to speculate about its origins. Basque is well studied, known for ergativity, and clearly unrelated to Indo-European. However, it is not known whether Basque is a continuation of one of the ancient populations in Europe or is a more recent arrival.
As for the present, in addition to Indo-European languages in Europe are found Uralic languages, also sometimes called Finno-Ugric. Their homeland is probably just north of the eastern Eurasian steppe. The story of how Hungarian came to be separated geographically from its sister languages will be told in Chapter 8. These languages all have vowel harmony and agglutinative morphology, lack grammatical gender, and are remarkable for their varying numbers of cases, with certain varieties of one of the languages, Komi, having nearly 30. Finnish and Hungarian have 15 and 21, respectively, depending on how one counts. These case endings correspond to how English uses prepositions. For example, in Finnish, the inessive case (
-ssa
) is used to express situation in time or place
Suome-
ssa
âIn Finland,' the allative case (
-lle
) expresses direction or beneficiary of an action
Pane-n kirja-n pöyda-
lle
âI will put the book on the table,' the illative case (
-on
) expresses movement into something
Mene-n talo-
on
âI am going into the house,' while the translative case (
-ksi
) indicates a noun having undergone a change of state
Lumi muttu-I vede-
ksi
âThe snow turned into water.'
Caucasian has the oldest reconstructable protolanguage with a time-depth of 8000 years. The Caucasus Mountains are considered to be the homeland of this family. This area is what Johanna Nichols calls a
residual zone
, an area difficult for outsiders to enter and conquer, meaning that the indigenous languages there can survive and differentiate over long periods of time, often with unusual trajectories. The Georgian consonantal system, for instance, is notoriously difficult for nonnative speakers to acquire. It includes ejectives, which are voiceless consonants that include a simultaneous glottal closure. Thus, the voiceless aspirated
stop [p
h
] contrasts with the voiceless ejective stop [p']. The Caucasus has also been a region of recent conflict rooted in issues of language, as we will see in Chapter 9.
Proto-Afro-Asiatic is likely 10,000 years old. Two proposals have existed for the original homeland, one in the Middle East and the other in northeast Africa, generally the Horn of Africa. The matter has been decided in favor of Africa and is given credence by also being the place of origin of a Y-haplogroup that spread through the region. Thus, it seems that the northeastern part of the grassland that is now the Sahara Desert was the central homeland, with an eastward advance of the Semitic group into the Near East and the westward expansion of the Berber and Chadic groups. Another western expansion moved out from the center and met up with the western NigerâCongo languages, and this incursion prompted an eastward and southern expansion of the Bantu family. Because of the drying conditions in the Saharan region, Neolithic Afro-Asiatic farmers moved into river valleys and developed irrigation systems.
The people occupying the Nile developed the great Egyptian civilization. The most salient feature of the Semitic branch of the Afro-Asiatic languages is its word morphology based on the trilateral root, as mentioned in Chapter 5. The canonical example involved the root k-t-b âto write.' By means of internal vowel alternations, the following words can be generated:
| a kt u b | âI write' |
| k a t a b tu | âI wrote' |
| k a t a b a | âhe wrote' |
| k a t a b at | âshe wrote' |
| k i t Ä b | âbook' |
| ma kt u b | âoffice, desk' |
| ma kt a b at | âlibrary, bookstore' |
and so forth, nearly endlessly. Note that here the prefix
ma-
means âplace (where some activity involving writing happens).'
The Mongolic, Turkic, and Tungusic stocks all seem to have originated in Central Asia. Because these three groups of languages share striking similarities in syntactic structure and the pronoun system, they were originally considered to be in one stock, but now a possible phylum named Altaic is proposed for them. The Altai Mountains are in the western and southwestern part of Mongolia where Russia, China, and Kazakhstan all come together, and they are so named from the Mongolian word
alt
âgold' because of the gold deposits there.
This phylum has inspired endless debates over whether Japanese and/or Korean belong to it. In any case, the proto-speakers of Turkic, for example, being nomadic pastoralists, moved south and as far west as Turkey, along the route established as the Silk Road several thousand years later. In other words, modern-day Turkey, despite its name, is not the homeland of the Turkic stock and was originally occupied by
the Indo-European Hittites. Their structural features include: case systems, little head marking, accusative alignment, simple syllable structure, front rounded vowels, no genders, simple prosodic structure (no tones, fixed stress), agglutinative morphology, and very regular verbal word formation (Nichols 2011:188).
The proposed homeland for this stock is the Indus Valley in the northwest region of the Indian subcontinent, and the Harappan language and civilization identified with this region may or may not be Proto-Dravidian. This stock is another whose languages spread with the expansion of agricultural societies moving into areas occupied by hunter-gatherers. Dravidian languages are currently found now only in the southern portion of India, and the usual explanation is that speakers of Dravidian languages formerly occupied a wider area but were pushed to the south and east by the Indo-Europeans. Dravidian languages are well known for having retroflex consonants, where the tongue curves or curls itself on a point between the alveolar ridge and the hard palate. These became an areal feature of the South Asia and have been borrowed into languages of the Indic branch of the Indo-European stock.
The language stock/family map becomes particularly colorful, which is to say complicated, in this part of the world, perhaps because the land is so fertile and because there was a lot of movement and thus mixing of peoples, cultures, and languages over long periods of time. The controversial stock/phylum Austric is far from settled, and if clearer evidence for it could be found, it might include three major divisions: Austronesian, Austroasiatic, and Tai. As for the more settled categories, we have already addressed Austronesian.
The homeland for these languages seems to be the Mekong River at the place where modern-day Cambodia, Thailand, and Laos intersect. The family may well be only about 2000 years old.
Languages in this family are distributed throughout southern China and across Southeast Asia. The family name is fraught and has been known as Tai-Kadai and as Daic; its status as an independent language family is unclear and has been linked both to the Sino-Tibetan languages and to the Austronesian. One likely homeland of this family is the present-day Guizhou province of southern China. One proposal is that proto-Daic speakers migrated back to the mainland from Formosa/Taiwan at the beginning of the Austronesian breakup. This language family, like Austroasiatic, appears to be rather recent.
In addition to Austronesian, already discussed, Indo-Pacific languages are found in Papua New Guinea, Melanesia, and perhaps Tasmania. Indo-Pacific may well be an areal classification, because constructing a
protolanguage for these languages has proven difficult. On the other hand, population genetics now suggests that the reason for this difficulty might be that the language family is ancient and goes back perhaps as far as 40 kya. New Guinea, with its rugged terrain, counts as a residual zone, like the Caucasus Mountains. There are easily 1000 unique languages spoken in an area somewhat larger than the state of Texas.
Today's Aboriginal Australians are descendants of the earliest humans to occupy Australia and likely represent one of the oldest continuous populations outside Africa with a time-depth of 50 kya. Because of the time-depth, there is no way to know whether all of today's 250 aboriginal languages
15
descended from one protolanguage or whether the migrating humans who settled Australia spoke languages from different families. Once again, there is much complexity and controversy surrounding how to sort out the relationships among the Australian languages. There has been so much contact between and among the languages that many shared/similar features might not have been inherited from a protolanguage but rather have diffused through them. Furthermore, it is not necessarily the case that there was only one group to move into Australia; and it is highly likely that there were repeated movements from the west and the north before sea levels began to rise some 16 kya, isolating the continent. Multiple groups moving in and mixing in complicates an already-complicated picture.
For the sake of convenience, one specialist in these languages, R.M.W. Dixon, has organized the languages into 50 groups, only some of which qualify as being low-level genetic subgroups. There is no prehistoric time of breakup as such, if one considers the continent of Australia as a homeland, which, before the arrival of the Europeans in 1788, was occupied solely by the aboriginals.