CC BY-NC-ND
17
Sergey Kullanda Institute of Oriental Studies (Moscow)
Early Indo-European social organization and the Indo-European homeland
The homeland of a language family can be tentatively located with the aid of data on linguistic contacts, material culture, genetics, etc. Consequently, the current paper discusses the evidence that points to the specific traits of Proto-Indo-European social structure. It seems likely that the latter was based on age-sex stratification that had evolved into the varna system among the Indo-Iranians. Since age-sex stratification was particularly viable among cattle-breeding pastoralists, the search for the Indo-European homeland should focus on such mobile semi-sedentary cultures.
Keywords: kinship terms, age-sex stratification, cattle-breeding, varnas.
It is generally agreed that the homeland of a language family can be located with the aid of data on linguistic contacts, material culture, genetics, etc. In the light of this, I seek to evaluate the evidence pointing to the specific traits of Proto-Indo-European social structure.
Elsewhere, I have argued that early social organization of mankind had been based on age-sex stratification. Since I have already dwelt at length on this issue in (Кулланда 1998; Kullanda 2002), here I will confine myself to the broad outline of my arguments.
Analysis of Indo-European kinship terminology implies that Proto-Indo-European society was characterized by age-sex stratification (Hocart 1928; Idem 1937; Kullanda 2002). There are about twenty PIE etyma regarded, in accordance with the meaning of the majority of their reflexes, as kinship and relationship-by-marriage (affinal) terms. The reconstructed PIE kin-shipterm system is therefore usually divided into two subsystems, i. e., 1) consanguineous or blood-kinship terms and 2) affiliation or relationship-by-alliance terms. In doing so, however, one imposes on the society whose language one is trying to reconstruct one's own perception of kinship, notwithstanding conflicting evidence. Thus, PIE *bhrek2ter is traditionally included into the blood-kinship term subsystem with the meaning '(consanguineous) brother'. It is, however, at the same time traditionally considered a designation of any male member of the community/extended family belonging to the ego's generation, and rightly so (ИЭСОЯ II [1973]: 438-439; Трубачев 1959: 58 ff.; Benveniste 1969/1973, I: 213-214 [English version: 170171]; Szemerenyi 1977: 23-24; Гамкрелидзе и Иванов 1984/1995: 764 [English version: 666]). Nevertheless, the obvious discrepancy is tacitly ignored.
Moreover, by this approach morphologically uniform (in this case, formed with the aid of one and the same affix) terms are separated, while morphologically dissimilar ones are brought together. Meanwhile, 'the analysis of intrinsic subsystems discriminated on the basis of shared formal linguistic traits is far more effective than that of artificial subsystems distinguished on the basis of one's arbitrary perception of the outer world' (Поздняков 1989: 96).
Let us try to follow the above maxim and distinguish the intrinsic subsystems of PIE "kinship terms," if any. The first such subsystem to immediately attract one's attention is that of the social terms formed with the aid of the suffix *-ter, which Emile Benveniste regarded as a classifier for the lexical class of kinship terms (Benveniste 1969/1973, I: 212 [English version:
Journal of Language Relationship • Вопросы языкового родства • 9 (2013) • Pp. 137-144 • © Kullanda S., 2013
171])1. There are five such terms, namely: *phiter (its traditionally reconstructed meaning is 'father'), *mehiter ('mother'), *bhrehiter ('brother'), *dkughiter ('daughter') and *hiienhiter ('husband's brother's wife'). The reflexes of these terms and none other are invariably formed with the aid of the relevant suffix in all the subgroups of Indo-European language family.
We shall begin the analysis of these five terms with PIE *phiter. It is true that all of its recorded reflexes (not counting derivatives) mean 'father'. However, historical evidence does not allow one to reconstruct the same meaning for the prehistoric etymon, even with certain modifications.2 First and foremost is the polysemy of Latin pater. Even in such an idiom as pater familias this word cannot be interpreted as a mere kinship term, since familia either included freedmen and slaves or just meant 'slaves of one master' (Mommsen 1881: 60). Moreover, Latin pater also meant, without any amplifications, 'senator' (usually in the plural — patres). As one of the Latin etymological dictionaries puts it, the term 'does not indicate physical paternity which is rather implied by parens and genitor. Pater has a social significance.'3
Let us leave the term *phiter for the time being and turn to PIE *bkrehiter. As mentioned above, this etymon has been universally recognized as denoting not so much blood kinship as belonging to a large social group, usually assumed to be the extended family.
Let us turn now to the feminine terms. The survival of reflexes pertaining to either group membership or power and headship is less likely here, since men were invested with power and took more active part in social activities in historically recorded early Indo-European societies. Yet even in this case such reflexes can be traced.
Thus, Latin mater, a reflex of PIE *mehiter, "conveys, like pater, an idea of respect ... and is added to the name of a goddess. to honour her (Terra mater); the idea of maternity is not necessarily implied in the appellation: Vesta mater."
The majority of reflexes of PIE *dkughiter mean 'daughter.' At the same time, as Benveniste (1966: 34-50) has convincingly shown, Old Persian *duxgi, surviving as Elamite du-uk-si-is and meaning, like the cognate Middle Persian duxs, 'princess', eventually goes back, together with Armenian (borrowed from Middle Persian) dsxoy (queen), to PIE *dkughiter.4 In Oscan inscriptions the reflex of the etymon under consideration, fu(u)trei (dat. sg.), signifies a goddess of the circle of Ceres, probably Proserpine (Lejeune 1968: 72-73). This etymon also connotes authority/headship and thus cannot be treated as a mere kinship term.5
The last PIE etymon formed with the aid of the relevant suffix is PIE *hiienhiter. It is reconstructed on the basis of the following reflexes: Old Indian yatar-, Pushtu yor, Armenian ner, Greek (Homeric) plural eivaTepe^ (it must be noted that the Greek word was used almost exclusively in the plural; the singular evdrnp appears only in late inscriptions from Asia Minor — cf. Chantraine 1999: 323), Latin janitnces (re-interpreted after the feminine Nomina agentis of the genetrix type and recorded only in the plural), Lithuanian jente (gen. jenters), Russian
1 Moreover, according to Benveniste, it is "the suffix of kinship par excellence" (« le suffixe de parenté par excellence ») (1969/1973, I: 255 [English version: 205]).
2 Such as, for instance, Vater, Haupt der Grofifamilie, according to Julius Pokorny.
3 « Ce qu'indique pater, ce n'est pas la paternité physique, qui est plutôt indiquée par parens et par genitor. Pater a une valeur sociale » (Ernout et Meillet 2001: 487).
4 Szemerényi (1977: 20, note 67) refutes the derivation of Old and Middle Persian as well as Armenian words from PIE *dhugk2ter. However, in doing so he does not proceed from any linguistic considerations, but rather from his own conviction that such incompatible (from the point of view of the modern person) meanings simply cannot go back to the same etymon.
5 « Il [Ce mot]comporte, comme pater, une idée de respect... et s'ajoute au nom d'une déesse... pour l'honorer (Terra mater), et sans que l'idée de maternité soit nécessairement impliquée dans l'appellation: Vesta mater » (Ernout et Meillet 2001: 389).
jatrov'. All those words mean 'husband's brother's wife'. However, as Szemerenyi (1977: 92) rightly pointed out, the plural 'wives of brothers' may be reconstructed as a proto-language etymon as well.
Thus, what can be reconstructed here is a cluster of PIE words whose reflexes became mainly kinship terms and which were formed by means of an identical suffix. The relevance of the suffix in question can be inferred from the fact that it was attached in certain languages to those kinship terms that were originally formed otherwise6. Said cluster comprises four paired terms, which can be grouped on the basis of either gender (two masculine, i. e., *phiter and *bhrekiter, and two feminine, i. e., *mehiter and either *dkughiter or *hiienhiter) or relative age (two terms for seniors, i. e., *phiter and *mehiter, and two for juniors, i. e. *bkrehiter and either *dkughiter or *hiienhiter), as well as one feminine term that has no counterpart. Judging from the meaning of their reflexes, all these terms originally denoted membership of groups rather than individual positions. Moreover, certain reflexes of these terms pertain to authority and headship.
The reconstructed semantics would adequately describe a society that was, from the point of view of its members, regulated by age-sex stratification. «Competent» (that is, judging by ethnological parallels) married men are divided into juniors and seniors. The latter, judging by the reflexes meaning 'father' and the inevitable ethnological parallels, are those whose so-cial/classificatory children passed the initiation rites. Men's age grades have their exact female counterparts. It seems that, as is the case with men, women whose children passed the initiation rites are considered seniors. In addition, there also exists an age grade for girls who have reached puberty and are therefore entitled to marry (or have already married). It can be surmised that, due to the alleged dislocality of early marriage, both they and the wives of their classificatory brothers stay with their respective exogamous communities after getting married. In this case, members of the two groups of married men (juniors and seniors) and of the two groups of married women from the same exogamous community would have been not spouses, but siblings to each other, while their spouses would have lived in another community. Although the existence of an "odd" term for women seems to be indicative of such a situation, the latter assumption remains, of course, a mere conjecture.
Such a classification understandably includes only competent society members enjoying the socially active period of life. It leaves aside the non-initiated children, old men and women who have outlived the socially active life period, and initiated youths who have not yet won, through martial achievements, the right to get married. The markers for these groups were therefore unlikely to be formed by means of the suffix *-ter and consequently have not been discussed above. Let us see if later social institutions, such as the varna system, can be traced back to the alleged age-sex stratification.
To study the origin of the varna system, one has to look into Indo-Iranian and even Indo-European prehistory. Besides the famous Purusasukta story (RV X, 90, 11-12), Vedic texts provide very little information on the matter. They imply, however, that the kshatriya social class, charged with warring and political power, bears traces of the age class of young men. Thus, the Vedic Indra, a king and warrior par excellence, is depicted in the Rigveda as an unaging
6 One can cite in this respect Sanskrit jamatar, nanandar and naptar, which replaced the earlier form napat. Yet another example can be found in Persian where, as in French, post-tonic syllables were dropped. As a result of the loss of endings, Old Persian pitar-, matar-, bratar-, duxtar- became Middle Persian pit, mat, brat, duxt. A simulacrum of the old form had been, however, concocted by means of a new suffix added to the shortened form, hence Modern Persian pidar, madar, bradar, duxtar. By sheer analogy, the word pus 'son' (< puga) underwent similar suffixation and became pusar (Benveniste 1969: 256 [English version: 206]).
(ajuryâ) youth (yuvan) and mârya. One can, without stretching the point, trace these meanings to a protolanguage etymon marking membership in an age-sex group of young warriors. In this connection, the description of Indra and the Maruts in the Rigveda is of particular interest.
The Maruts in the Rigveda are youths (yuvanas and mâryas) mentioned exclusively as a group (sârdha, ganâ or vrata) without any individual characteristics. They were born simultaneously (sakâm jajnire [I, 164, 4]; sakâm jatah [V, 55, 3]), they are of the same nest (sânîlah [I, 165, 1; VII, 56, 1], neither senior nor junior and middle (téajyestha âkanisthasa udbhidô 'madhyamaso... [V, 59, 6]) (for more details see Wikander 1938: passim).
According to the Rigveda, the Maruts are samtapana [VII, 59, 9], that is, as Renou (1962: 46) puts it, born of the total burn, nés de la brûlure-totale. However, since the prefix sam- (here in the vrddhi form with a long vowel, sam-) implies the action of several agents, one is entitled to interpret the word samtapana as "those who passed through a common ordeal by fire", being an integral part of initiation rites.
A very interesting mention of Indra as a merchant (vanij) in the Atharvaveda (indram ahâm vanijam codayami, "I incite Indra the merchant" — XV 15, 1) that, at first sight, seems to be contradictory to his image of the war chief is in fact quite in keeping with it. Archaic trade was closely related to warring — suffice it to remember the Vikings who were both warriors and traders. This ambiguity was reflected in Indo-European languages where words meaning 'war booty' (Sanskrit lotra), 'gain, profit' (Latin lucrum) and 'pay' (German Lohn, etc.) go back to the same protolanguage etymon (Benveniste 1969: I, 166-169; Гамкрелидзе, Иванов 1984: 740).7 A classic case of this situation is the image of Wodan, the Germanic war god who was at the same time the protector of trade and in this capacity was identified with Mercury, the god of commerce in Roman mythology. As a result, the day of Mercury in Romance languages (French mercredi, etc.) became the day of Wodan (Dutch woensdag, English Wednesday).
Judging by ethnographic materials, a child becomes a young warrior at fifteen or sixteen. Certain, quite instructive, passages from early and traditional narrative sources can be cited in this connection. Thus, according to the Avestan Hom Yasht, during the golden age of Yima "looking fifteen-year-olds strolled about father and son." (panca.dasa fracaroiOe pita puOrasca raoSaësuua... — Y. IX, 5). 'Fifteen-year-old' here clearly means something like 'being in his prime.' There is a strikingly similar description of a band of adventurers in a Russian epic: "Volkh himself is fifteen // and all his men are fifteen-year-olds" (Кирша Данилов 2000: 73). Indian iconographic tracts also imply that the age of sixteen is the prime of life: "celestials are also to be represented as youths of sixteen summers" (dvirastavarsakarasca tatha karya divaukasah) (Citrasutra 38, 7 apud Chitrasutra 1978: 143, 173; see also Citralakshana III, 142)8.
I think that the presented materials enable us to trace the Kshatriya varna to the age-class of young warriors. Unfortunately, we have almost no information on the relation of the other two primary varnas to age-sex classification in Indian tradition; but in this case, we may be aided by Iranian tradition as well. Thus, according to a Scythian legend or myth recorded by Herodotus (IV 5-7), the Scythian progenitor, Targitaos, had three sons: Lipoxaïs, Arpoxaïs, and the youngest (vewTaroç) Colaxaïs. As the youngest brother succeeded in getting hold of sacred objects fallen from the sky, the elder brothers ceded the royal power to him. The story goes that "From Lipoxaïs originated those Scythians who are called the clan (yévoç) of Au-chatae; from Arpoxaïs, those called Katiari and Traspians; from the youngest the kings called Paralatae." (Ibid., 6.) (cf. Dumézil 1930; Benveniste 1938). Evidently, the social position of the
7 M. Mayrhofer, however, held that Sanskrit lotra was not cognate with the other above-mentioned words (EWAia 3: 449-450).
8 I am indebted to V. V. Vertogradova who drew my attention to these texts.
three brothers had nothing to do with seniority, otherwise the second son would have become the ancestor of kings. Neither can it be a mere error on the part of Herodotus, since in an independent Iranian tradition, that of the Middle Persian Bundahisn, it is the youngest son of Zar-dusht that becomes the ancestor of the warriors, resp. kings, while the eldest son is the ancestor of priests, and the second son is the ancestor of agriculturalists (Grantovskij 1960: 13-15). Moreover, parts of a system cannot be separated arbitrarily. If one of the social classes was related to an age class, the other two probably had the same origin. The next step in the interpretation of the Scythian genealogical legend, i. e. a correlation between its social interpretation and the hypothesis of W. Brandenstein [Brandenstein 1953: 183 sqq.] that Scythian yevea had originally meant age groups, was taken by E. Berzin. He held that among early societies "prevailed the division in three main age-classes corresponding to the essential social functions: 1) young warriors; 2) mature men, householders and producers of material goods; and 3) old men who performed cult duties... Thus age-classes became varnas." [Eep3MH 1986: 46).
Unfortunately, Brandenstein's and Berzin's works contained, besides reasonable guesses, a lot of far-fetched assumptions and outright errors,9 hence the sceptical attitude towards them. Nevertheless, the above-discussed data allow us to revisit the hypothesis of the age-sex origin of the varnas.
Thus, it seems likely that the Kshatriyas were originally warlike youths, the Vaishyas were mature men that ceased to be warriors par excellence and were charged mainly with farming and animal husbandry, and the Brahmans belonged to the age-class of elders performing priestly functions. It is all the more probable since there are mentions of age-sex classification in Vedic literature. Thus, the word vayas 'vigour, strength' in certain contexts means 'a stage of life.' One can cite such passages from the Rigveda as "passing through one life period (vayas) after another" (vayo-vayo vicarantah) [VIII, 55. 4], "[Soma] circulates through life stages" (pari... vayamsi... yati) (IX, 9. 1) or "when thou [i. e. Agni] growest old establishing one life stage after another, / thou goest circle-wise changing thy outlook" (vayo-vayo jarase yad dadhanah / pari tmana visurupo jigasi) (V, 15. 4), etc.
It is likely that Agni and Soma here personify the movement of human beings through life periods; when an age-group reaches the limit of the socially active age it leaves the system of age-classes and is substituted by a new group that assumes its name, that is, the circle-wise movement mentioned in the above-cited passages takes place. As to the age-grades in Vedic tradition, it is worth noting that in the Satapatha Brahmana (XII, 9, 1, 8) one encounters the word vayas meaning approximately 'the active life, vigorous age' and -vayasa 'life period' in such compounds as purvavayasa-, madhyamavayasa- and uttamavayasa- meaning the first, middle and last stage of life respectively. Let us cite this very instructive passage: "There are three sacrificial cakes, for this life (vayas) of man (purusa) consists of three parts. It is his life that [adhvaryu] wins for him by [means of] those [cakes]. The early life (purvavayasa) [he wins] by that of Indra, the middle [part of] life (madhymavayasa) by that of Savitar, and the last [part of] life (uttamavayasa) by that of Varuna. Having won the life from Death in the sequence [of its periods and] deities [yajamana] makes it immortal" (trayah purodasa bhavanti / tredhavihitam va idam purusasya vayah / vaya evasya taih sprnoti / purvavayasamevaindrena / madhymavayasam savitrena / uttamavayasam varunena / yatharupameva yathadevatam vayo mrtyoh sprtvamrtan kurute).
9 Thus, Brandenstein argued that Scythian yévea were related to the Tripolye farmers (the descendants of the elder brother), the bearers of the battle-axe culture (the descendants of the second son) and the Scythians proper (the descendants of the younger brother), which does not seem very plausible. Berzin posited that Indo-Europeans were mounted warriors while horseback riding was not widespread even in the Indo-Iranian epoch, etc.
All these and many other facts can only be explained as the result of the evolution of an age-stratified society, all the more so since linguistic analysis is indicative of the existence of age-sex stratification in the Proto-Indo-European community.
The obvious social-cultural implication of this conclusion is that age-sex stratification underlay the classificatory and individual kinship patterns and may therefore have been the first stage of social evolution. Yet while the remnants of age-sex classification had survived in a number of the recorded Indo-European traditions, it was among the Indo-Iranians that it evolved into the varna/pistra system. What circumstances, then, could have been favourable to the formation of such a system? Age-sex stratification has been particularly viable among pastoralists, and, more specifically, among those pastoralists who are not exactly nomadic but not quite settled cattle-breeders, such as the Maasai, Gikuyu, Oromo and other peoples of East Africa, the classic area of age-sex stratification. The same kind of economy was common to the Indo-Iranians, the only difference being the use of the horse. They also regarded cattle as their main wealth and raided their neighbours to drive their herds away: one may recall such Vedic notions as, for instance, gavyânt 'desirous of cattle/battle', gavyu 'id.' and gavya 'desire for cattle.'10 Cattle also played an important role in their spiritual life as evidenced by the Avestan notion of the soul of the cow, gdus uruuan-. Looking for such cattle-breeding cultures would hopefully narrow the focus of the search for the Indo-European and the subsequent Indo-Iranian homeland.
Literature
Берзин 1986 — Э. О. Берзин. Сивка-Бурка, вещая Каурка, или Древняя Европа в зеркале мифов и сказок // Знание — сила, № 11. [E. O. Berzin. Sivka-Burka, veschaya Kaurka, ili Drevnyaya Evropa v zerkale mifov i skazok // Znanie — sila, № 11.] Гамкрелидзе, Иванов 1984 — Т. В. Гамкрелидзе, Вяч. Вс. Иванов. Индоевропейский язык и индоевропейцы. Реконструкция и историко-типологический анализ праязыка и протокультуры. Тбилиси. [English version: Thomas. V. Gamkrelidze and Vjaceslav V. Ivanov. Indo-European and the Indo-Europeans. A Reconstruction and Historical Analysis of a Proto-Language and a Proto-Culture. Transl. by Johanna Nichols. Ed. by Werner Winter. Mouton de Gruyter. Berlin/New York 1995.] ИЭСОЯ — В. И. Абаев. Историко-этимологический словарь осетинского языка. Т. 1-5. Ленинград / Москва, 1958-1995. [V. I. Abaev. Istoriko-etimologicheskij slovar' osetinskogo yazyka. T. 1-5. Leningrad / Moskva, 19581995.]
Кирша Данилов 2000 — Древние российские стихотворения, собранные Киршею Даниловым (= Полное собрание русских былин, т. 1). СПб.: Тропа Троянова. [Drevnie rossijskie stikhotvoreniya, sobrannye Kirsheyu Danilovym (= Polnoe sobranie russkikh bylin, t. 1). SPb.: Tropa Troyanova.] Кулланда 1995 — С. В. Кулланда. Царь богов Индра: юноша — воин — вождь // Ранние формы политической организации: от первобытности к государственности. М. [S. V. Kullanda. Car' bogov Indra: yunosha — voin — vozhd' // Rannie formy politicheskoj organizacii: ot pervobytnosti k gosudarstvennosti. M.] Кулланда 1998 — С. В. Кулланда. Системы терминов родства и праязыковые реконструкции. // Алгебра родства, вып. 2, СПб. [S. V. Kullanda. Sistemy terminov rodstva i prayazykovye rekonstrukcii. // Algebra rod-stva, vyp. 2, SPb.]
10 See, e. g., the Vedic cliché: gavyata mánasa 'with the cattle-desirous mind' used in warlike context (RV III 31, 9; IV 1, 15; VI 46, 10), and the following passages from the Rigveda: 'Through cattle-desire for the Trtsu he set off to do battle against the men' (gavya tftsubhyo ajagan yudha nfn — VII 18, 7); "The chariot... goes desiring cattle, desiring horses" (rátho... gavyúr asvayúr lyate — IV 31, 14); "Let your cattle-desiring vajra be with you, O Indra" (sám te vájro vartatam indra gavyúh — VI 41, 2).
Поздняков 1989 — К. И. Поздняков. Лексика и грамматика в реконструкции элементов культуры // Лингвистическая реконструкция и древнейшая история Востока. Тезисы докладов Международной конференции, Москва, 29 мая — 2 июня 1989 г. Ч. 3. Москва: Наука, Главная редакция восточной литературы. С. 96100. [K. I. Pozdnyakov. Leksika i grammatika v rekonstrukcii elementov kul'tury // Lingvisticheskaya rekon-strukciya i drevnejshaya istoriya Vostoka. Tezisy dokladov Mezhdunarodnoj konferencii, Moskva, 29 maya — 2 iyunya 1989 g. Ch. 3. Moskva: Nauka, Glavnaya redakciya vostochnoj literatury. S. 96-100.]
Трубачев 1959 — О. Н. Трубачев. История славянских терминов родства и некоторых древнейших терминов общественного строя. Москва: Издательство Академии наук СССР. [O. N. Trubachev. Istoriya slavyanskikh terminov rodstva i nekotorykh drevnejshikh terminov obschestvennogo stroya. Moskva: Izdatel'stvo Akademii nauk SSSR.]
Фрейденберг 1978 — О. М. Фрейденберг. Миф и литература древности. М. [O. M. Frejdenberg. Mif i literatura drevnosti. M.]
Benveniste 1938 — Emile Benveniste. Traditions indo-iraniennes sur les classes sociales // Journal Asiatique, 230. P. 529-549.
Benveniste 1966 — Emile Benveniste. Titres et noms propres en iranien ancien. Paris: Librairie C. Klincksieck.
Benveniste 1969 — Emile Benveniste. Le vocabulaire des institutions indo-européennes. Vol. 1-2, Paris. [English version: Indo-European Language and Society. Transl. by Elisabeth Palmer. London: Faber and Faber Limited, 1973.]
Brandenstein 1953 — W. Brandenstein. Die Abstammungssagen der Scythen // Wiener Zeitschrift für die Kunde des Morgenlandes, Bd. 52, 1-2.
Chantraine 1999 — Pierre Chantraine. Dictionnaire étymologique de la langue grecque. Histoire des mots. Nouvelle édition mise à jour. Paris: Klincksieck.
Chitrasùtra 1978 — Chitrasütra of the Vishnudharmottara. C. Sivaramamurti (ed. and transl.). New Delhi: Kanak Publications.
Dumézil 1930 — Georges Dumézil. La préhistoire indo-iranienne des castes // Journal Asiatique, 219. P. 109-130.
Dumézil 1961 — Georges Dumézil. Les « trois fonctions » dans le Rgveda et les dieux indiens de Mitani // Académie royale de Belgique. Bulletin de la classe des lettres et des sciences morales et politiques. 5e Série, T. 47. P. 265298. Bruxelles.
Ernout et Meillet 2001 — Alfred Ernout et Antoine Meillet. Dictionnaire étymologique de la langue latine. Histoire des mots. Retirage de la 4e édition augmentée d'additions et de corrections par Jacques André. Paris: Klincksieck.
EWAia — Manfred Mayrhofer. Etymologisches Wörterbuch des Altindoarischen. Bd. 1-3. Heidelberg: C. Winter, Universitätverlag, 1992-2001.
Grantovskij 1960 — E. Grantovskij. Indoiranische Kastengliederung bei den Skythen. Moskau: Verlag für orientalische Literatur.
Hocart 1928 — A. M. Hocart. The Indo-European Kinship System // Ceylon Journal of Science, sec. G, vol. 2, pt. 4, pp. 179-204 (reprinted in: Imagination and Proof. Selected Essays of A. M. Hocart. Edited and with an introduction by Rodney Needham. Tucson: The University of Arizona Press, 1987, pp. 61-85).
Hocart 1937 — A. M. Hocart. Kinship Systems // Anthropos 32, pp. 545-551.
Kullanda 2002 — Sergey Kullanda. Indo-European 'Kinship Terms' Revisited // Current Anthropology, Vol. 43, No. 1.
Lejeune 1968 — M. Lejeune. « Fils » et « fille » dans les langues de l'Italie ancienne // Bulletin de la Société de Linguistique de Paris 62 (1): 67-86.
Mommsen 1881 — T. Mommsen. Römische Geschichte, 7e Aufl., Bd. 1: Berlin: Weidmann.
Pokorny 1959/69 — Julius Pokorny. Indogermanisches etymologisches Wörterbuch. Bd. 1-2. Bern/München: Francke Verlag.
Renou 1962 — Louis Renou. Études védiques et paninéennes X. Paris.
Szemerényi 1977 — Oswald Szemerényi. Studies in the Kinship Terminology of the Indo-European Languages. With Special Reference to Indian, Iranian, Greek, and Latin (= Acta Iranica 16). Téhéran-Liège: Edition Bibliothèque Pahlavi. Diffusion E.J. Brill, Leiden.
Wikander 1938 — Stig Wikander. Der arische Männerbund. Studien zur Indo-Iranischen Sprach- und Religionsgeschichte. Lund.
С. В. Кулланда. Социальная организация ранних индоевропейцев и индоевропейская прародина.
Прародину языковой семьи обычно ищут по данным языковых контактов, материальной культуры, генетики и т. п. Автор предлагает использовать для этой цели свидетельства об особенностях социальной структуры праиндоевропейцев. Есть основания предполагать, что эта система основывалась на половозрастной стратификации, эволюционировавшей у индоиранцев в систему варн. Поскольку половозрастное деление общества прежде всего характерно для подвижных полуоседлых скотоводов (оромо, масаи и др.), культура носителей праиндоевропейского, видимо, обладала теми же чертами.
Ключевые слова: термины родства, половозрастная стратификация, скотоводство, варны.
