1911 Encyclopædia Britannica/Indo-European Languages
INDO-EUROPEAN LANGUAGES. The Indo-European (I.E.) languages are a family of kindred dialects spread over a large part of Europe, and of Asia as far as India.
The main branches so far identified fall easily into two groups of four. These groups are distinguished from one another by the treatment of certain original guttural sounds, k(c), g, kh, gh, which one group shows as consonants, while the other converts them into sibilants. The variation is well shown in the word for “hundred”: Gr. ἑ-κατόν, Lat. centum, Old Irish cēt; Sanskrit śatam, Zend satəm, Lithuanian szim̄tas, Old Bulgarian (Old ecclesiastical Slavonic) sŭto. In the first three the consonant is a hard guttural (the Romans said kentum, not sentum), in the others it is a sibilant (the Lithuanian sz is the English sh).
The first group (generally known as the centum-group) is the Western and entirely European group, the second (generally known as the satem-group) with one exception lies to the east of the centum-group and much its largest part is situated in Asia. To the centum-group belong (1) Greek; (2) the Italic languages, including Latin, Oscan, Umbrian and various minor dialects of ancient Italy; (3) Celtic, including (a) the Q-Celtic languages, Irish, Manx and Scotch Gaelic, (b) the P-Celtic, including the language of ancient Gaul, Welsh, Cornish and Breton: the differentiation, which exists also in the Italic languages, turning upon the treatment of original kw sounds, which all the Italic languages save Latin and the little-known Faliscan and the (b) group of the Celtic languages change to p. With these go (4) the Germanic or Teutonic languages, including (a) Gothic, (b) the Scandinavian languages, Swedish, Danish, Norwegian, Icelandic—differentiated in historical times out of a single language, Old Norse,—(c) West Germanic, Including English and Frisian, Low Frankish (from which spring modern Dutch and Flemish), Low and High German.
To the satem-group belongs (1) Aryan or Indo-Iranian, including (a) Sanskrit, with its descendants, (b) Zend, and (c) Old Persian, from which is ultimately descended Modern Persian, largely modified, however, by Arabic words. This group is often divided into two sub-groups, Indo-Aryan, including the languages of India, and Iranian, used as a general title for Zend and Old Persian as the languages of ancient Iran. Although the sounds of Indo-Aryan and Iranian differ considerably, phrases of the earliest form of the one can be transliterated into the other without change in vocabulary or syntax. (2) To the west of these lies Armenian, which is so full of borrowed Iranian words that only in 1875 was it successfully differentiated by Hübschmann as an independent language. It is probably related to, or the descendant of, the ancient Phrygian, which spread into Asia from Thrace by the migration of tribes across the Hellespont. Of ancient Thracian unfortunately we know very little. (3) North of the Black Sea, and widening its borders in all directions, comes the great Balto-Slavonic group. In this there are two branches somewhat resembling the division between Indo-Aryan and Iranian. Here three small dialects on the south-east coast of the Baltic form the first group, Lithuanian, Lettish and Old Prussian, the last being extinct since the 17th century. The Slavonic languages proper themselves fall into two groups: (a) an Eastern and Southern group, including Old Bulgarian, the ecclesiastical language first known from the latter part of the 9th century A.D.; Russian in its varieties of Great Russian, White Russian and Little Russian or Ruthenian; and Servian and Slovene, which extend to the Adriatic. (b) The western group includes Polish with minor dialects, Czech or Bohemian, also with minor languages in the group, and Sorb. In the satem division is also included (4) Albanian, which like Armenian is much mixed with foreign elements—Latin, Greek, Turkish and Slavonic. The relation between it and the ancient Illyrian is not clear.
Besides the languages mentioned there are many others now extinct or of which little is known—e.g. Venetic, found in clearly written inscriptions with a distinctive alphabet in north-eastern Italy; Messapian, in the heel of Italy, which is supposed to have been connected with the ancient Illyrian; and possibly also the unknown tongue which has been found recently on several inscriptions in Crete and seems to have been the language of the pre-Hellenic population, the finds apparently confirming the statement of Herodotus (vii. 170) that the earlier population survived in later times only at Praesos and Polichne. Names of deities worshipped by the Aryan branch are reported to have been discovered in the German excavations at Boghaz-Keui (anc. Pteria, q.v.) in Cappadocia; names of kings appear in widely separated areas elsewhere in Asia, and a language not hitherto known has recently been found in excavations in Turkestan and christened by its first investigators Tocharish. So far as yet ascertained, Tocharish seems to be a mongrel dialect produced by an intermixture of peoples speaking respectively an I.E. language and a language of an entirely different origin. The stems of the words are clearly in many cases I.E., but the terminations are no less clearly alien to this family of languages. It is remarkable that some of its words, like ku, “dog,” have a hard k, while the other languages of this stock in Asia, so far as at present known, belong to the satem-group, and have in such words replaced the k by a sibilant.
Till the latter part of the 18th century it was the universal practice to refer all languages ultimately to a Hebrew origin, because Hebrew, being the language of the Bible, was assumed, with reference to the early chapters of Genesis, to be the original language. Even on these premises the argument was unsound, for the same authority also recorded a confusion of tongues at Babel, so that it was unreasonable to expect that languages thus violently metamorphosed could be referred so easily at a later period to the same original. The first person to indicate very briefly the existence of the Indo-European family, though he gave it no distinctive name, was Sir William Jones in his address to the Bengal Oriental Society in 1786. Being a skilled linguist, he recognized that Sanskrit must be of the same origin as Greek, Latin, Teutonic (Germanic) and possibly Celtic (Asiatic Researches, i. p. 422; Works of Sir W. Jones, i. p. 26, London, 1799). Unfortunately Sir William Jones’s views as to the relationship of the languages were not adopted for many years by later investigators. He had said quite definitely, “No philologer could examine them all three (Sanskrit, Greek and Latin) without believing them to have sprung from some common source, which perhaps no longer exists.” Friedrich Schlegel, who learnt Sanskrit from Alexander Hamilton in Paris nearly twenty years later, started the view that Sanskrit, instead of being the sister, was the mother of the other languages, a mistake which, though long since refuted in all philological works, has been most persistent.
Curiously enough the history of the names given to the family is obscure. The earliest known occurrence of the word “Indo-European” is in an article in the Quarterly Review for 1813 by Dr Thomas Young. The term has been in use in English and in French almost continuously since that date. But a glance at Dr Young’s article will show that he included under Indo-European many languages like Basque, Etruscan and Arabian (his term for Semitic), which certainly do not belong to this family of languages at all; and if the term is taken to mean, as it would seem to imply, all the languages spoken in India and Europe, it is undoubtedly a misnomer. There are many languages in India, as those of the Dravidians in Southern India and those of Northern Assam, which do not belong to this family. On the other hand there are many languages belonging to the family which exist outside both India and Europe—Zend, Old Persian, Armenian, Phrygian, to say nothing of languages recently discovered. The term most commonly used in Germany is “Indo-Germanic.” This was employed by Klaproth as early as 1823. It is said not to have been invented by him, but by whom and when it was invented is not quite ascertained. It is an attempt to name the family by its most easterly and most westerly links. At the time when it was invented it had not yet been settled whether Celtic was or was not a member of this family. But in any case the term would not have been wrong, for members of the Germanic stock have been settled for above a thousand years in Iceland, the most westerly land of Europe, and for the last four centuries have increasingly dominated the continent of America. As has been pointed out by Professor Buck of Chicago (Classical Review, xviii. p. 400), owing to the German method of pronouncing eu as oi, the word “Indo-Germanic” is easier for a German to pronounce than “Indo-European.” Attempts to discover a more accurate and less ponderous term, such as “Indo-Celtic” or “Celtindic,” have not met with popular favour. Aryan (q.v.) is conveniently brief, but is wanted as the proper term for the most easterly branch of the family. What is wanted is a term which does not confuse ethnological and linguistic ideas. Not all speakers of any given language are necessarily of the same stock. In ancient Rome Latin must have been spoken by many slaves or sons of slaves who had no Latin blood in their bodies, though a slave if manumitted by his master might be the father or grandfather of a Roman citizen with full rights. Plautus and Terence were both aliens, the one an Umbrian, the other an African. The speakers of modern English are even a more multifarious body. A possible name for the family, implying only the speaking of a language of the stock without any reference to racial or national characteristics, could be obtained from the name for man, so widely though perhaps not altogether universally diffused throughout the family—Sanskrit vīras, Lithuanian wyras, Lat. vir, Irish fer, Gothic waír, &c. If the speakers of these languages were called collectively Wiros, no confusion with ethnological theories need arise.
It is customary to talk of the roots, stems and suffixes of words in the Indo-European languages. These languages are distinguished from languages like Chinese by the fact that in the great majority of words suffixes can be separated from roots. But the distinction between them and the so-called agglutinative languages is one of degree rather than of kind. In the agglutinative languages, or at any rate in some of them, some of the post-fixed elements have still an independent value. In the Indo-Germanic languages no one can say what the meaning of the earliest suffixes was. Suffixes which have developed in individual languages or individual sections of this family of languages can often be traced, e.g. the often quoted -hood in English words like “manhood,” or the English -ly in “manly,” which has gradually extended till it is actually attached to its own parent like in “likely.” But all recent investigation goes to show that before the Indo-European languages separated they possessed words with all the characteristics which we recognize in substantives like the Latin dominus or verbs like the Greek δείκνυται. Or, to put the same fact in another way, by the comparative method it is impossible to reach a period when the speakers of Indo-European languages spoke in roots. A “root” is only a convenient philological abstraction; it is merely the remnant which is left when all the elements that can be analysed are taken away; it is therefore only a kind of greatest common measure for a greater or smaller body of words expressing modifications of the same idea. Thus, though by no means the earliest form of the word, the English man might be taken as the “root” from which are derived by various suffixes manhood, manly, mannish, manful, manned (past tense), manned (participle), unman, mannikin, &c. How far the suffixes which can be traced back to Indo-European times (i.e. to a time before the separation of the languages) had existence as separate entities it is impossible to say. From what we see of the later history of the languages it is much more probable that both forms and signification were very largely the result of analogy. For in the making of new words analogy plays a much larger part than any reference to general principles of formation or composition. New words are to a large extent, even in modern times, the invention of persons unskilled in the history of language.
The first to point out that the term Indo-European (or Indo-Germanic) was not used uniformly in one sense was Professor Kretschmer in his Einleitung in die Geschichte der griechischen Sprache (Göttingen, 1896), pp. 9 ff. It is in fact used in three senses. (1) Indo-European is treated as preceding and different from all its descendants, a single uniform speech without dialects. But, strictly, no such language can exist, for even individual members of the same family differ from one another in pronunciation, vocabulary, sentence formation, etc. Thus it appears impossible to ascertain what the Indo-European term for the numeral 1 was, since different languages show at least four words for this, three of them presenting the same root with different suffixes: (a) Sanskrit eka (= *οι-qṷo-); (b) Zend aeva, Old Persian aiva, Greek οι-(ϝ)ο-ς (= *οι-ṷo-); (c) Greek οἰνή, “ace,” Latin unus (older oenus), Old Irish oen, Gothic ains, Lithuanian vénas (where the initial v has no more etymological signification than the w which now begins the pronunciation of the English one), Old Bulgarian inŭ; (d) Greek εἷς, ἔν (= *sem-s). But the Indo-European community must have had a word for the numeral since the various languages agree in forms for the numerals 2 to 10, and the original Indo-European people seem to have been able to count at least as far as 100. On the other hand, if the Indo-European language must have had dialects, the line of differentiation between it and its descendants becomes obliterated. (2) But even when a word is found very widely diffused over the area of the Indo-European languages, it is not justifiable to conclude that therefore the word must have belonged to the original language. The dispersion of the Indo-European people over the areas they now inhabit, or inhabited in the earliest times known to history, must have been gradual, and commerce or communication between different branches must have always existed to some extent; the word might thus have been transmitted from one community to another. When a word is found in two branches which are geographically remote from one another and is not found in the intermediate area, the probability that the word is original is somewhat stronger. But even in this case the originality of the word is by no means certain, for (a) the intervening branch or branches which do not possess the word may merely have dropped it and replaced it by another; (b) the geographical position which the branches occupy in historical times may not be their original position; the branches which do not possess the word may have forced themselves into the area they now occupy after they had dropped the word; (c) if the linguistic communities which possess the word have a seaboard and the intervening communities have not, the possibility of its transmission in connexion with early sea-borne commerce must be considered. At the dawn of European history the Phoenicians and the Etruscans are great seafarers; at a later time the Varangians of the North penetrated to the Mediterranean and as far as Constantinople; in modern times sea-borne commerce brought to Europe words from the Caribbean Indians like potato and tobacco, and gave English a new word for man-eating savages—cannibal. Thus with Kretschmer we must distinguish between what is common Indo-European and what is original Indo-European in language. (3) A word may exist in several of the languages, and may have existed in them for a very long time, and yet not be Indo-European. Hehn (Das Salz, ed. 2, 1901) rejects salt as an Indo-European word because it is not found in the Aryan group, though in this case he is probably wrong, (a) because, as has been shown by Professor Johannes Schmidt, its irregular declension (sal-d, genitive sal-nes) possesses characteristics of the oldest Indo-European words; (b) because the great plains of Iran are characterized by their great saltness, so that the Aryan branch did not pass through a country where salt was unknown, although, according to Herodotus (i. 133), the Persian did not use salt to season his food. Since Kretschmer wrote, this argument has been used very extensively by Professor A. Meillet of Paris in his Dialectes indo-européens (Paris, 1908). In this treatise he brings forward arguments from a great variety of facts to show that in the original Indo-European language there were dialects, the Aryan, Armenian, Balto-Slavonic and Albanian, as we have seen, forming an oriental group with novel characteristics developed in common, although in various other characteristics they do not agree. Similarly Italic, Celtic and Germanic form a Western group, while Greek agrees now with the one group now with the other, at some points being more intimately connected with Italic than with any other branch, at others inclining more towards the Aryan. This grouping, however, is by no means exclusive, members of either group having characteristics in common with individuals of the other group which they do not share with the other languages of their own group (Meillet, p. 131 ff.).
From all this it is clear that in many cases it must be extremely uncertain what is original Indo-European and what is not. Some general characteristics can, however, be predicated from what is handed down to us in the earliest forms of all or nearly all the existing languages. (1) The noun had certainly a large number of distinct cases in the singular: nominative, accusative, genitive, ablative, locative, instrumental, dative. In the plural, however, there was less variety, the forms for dative and ablative being from the earliest times identical. In the dual, the oblique cases cannot be restored with certainty, so little agreement is there between the languages. In the locative-singular the ending -i seems to have been of the nature of a post-position, because in various languages (notably in Sanskrit) forms appear without any suffix. In the locative plural also the difference between the -su of Sanskrit and early Lithuanian (Slavonic -chu) on the one hand, and of -σι in Greek on the other, seems to be best explained by supposing that the -u and -i are postpositions, a conclusion which is strengthened by the Greek rule that -σ- between vowels disappears. In the instrumental singular and plural it is noticeable that there are two suffixes—one, represented in Germanic and Balto-Slavonic only, beginning with the sound -m, the other, surviving in most of the other languages for the plural, going back to an Indo-European form beginning with -bh. Professor Hirt of Leipzig has argued (Idg. Forschungen, v. pp. 251 ff.) that -bh- originally belonged to the instrumental plural (cf. the Lat. filiabus, omnibus, &c.), and the forms with -m- to the dative and ablative. But this is merely a conjecture, which has no linguistic facts in its favour, for the -bi of the Latin dative tibi, which has parallel forms in many other languages, belongs to the pronouns, which show in their declension many differences from the declension of the noun (cf. also Brugmann, Grundriss (ed. 2), ii. 2, p. 120). (2) The adjective agrees with its noun in gender, number and case, thus introducing a superfluous element of agreement which is not found, e.g. in most of the agglutinative languages. Thus in phrases like the Greek ἡ καλὴ κόρη or the Latin illa pulchra puella the feminine gender is expressed three times, with no advantage, so far as can be detected, over the modern English, that fair maid, where it is not obviously expressed at all. In this respect and also in the employment of the same case endings for the plural as well as the singular, in the plural after a syllable expressing plurality, the agglutinative languages have a distinct superiority over the Indo-European languages in their earliest forms. Some languages, like English and Modern Persian, have practically got rid of inflexion altogether and the present difficulty with it; others, like modern German, as the result of phonetic and analogical changes have even intensified the difficulty. (3) In the personal pronouns, especially those of the first and second persons, there is widely spread agreement, but more in the singular than in the plural. Forms corresponding to the English I and thou, the Latin ego and tu, are practically universal. On the other hand the demonstrative pronouns vary very considerably. (4) The system of numerals (subject to slight discrepancies, as that regarding 1 mentioned above) is the same, at least up to 100. (5) In the verb there were at first two voices, the active and the middle, and three moods, the indicative, the subjunctive and the optative. It has been suggested by Professors Oertel and Morris in Harvard Studies, xvi. (p. 101, n. 3) that the similarity which exists between the earliest Greek and the earliest Aryan in the moods is the result of a longer common life between those two branches. But of this there is no proof, and the great difference in the treatment of the sounds by these two branches (see below) militates very strongly against the supposition. The tense forms indicated originally not relations in time but different kinds of action. The distinctive forms are the present, the perfect, and the aorist. The present indicated that an action was in progress or continuous, the aorist on the other hand regarded the action as a whole and, as it were, summed it up. The aorist has sometimes been said to express instantaneous action, and so it does. But this is not the essence of the aorist; the aorist may be used also of a long continued action when it is regarded as a whole. Greek shows this very clearly. In Athenian official inscriptions it was usual to fix the date of the record by stating at the commencement who was the chief magistrate (archon) of the year. This was expressed by the imperfect (ἠρχε). But when reference was made to a past archonship, that was expressed by the aorist (ἦρξε). The same characteristic is evident also in prohibitions; thus, in Plato’s Apology of Socrates, μὴ θορυβήσητε is “Do not begin to make a disturbance,” μὴ θορυβεῖτε is “Do not keep on making a disturbance.” These points are most easily illustrated from Greek, because Greek, better than the other languages, has kept the distinctive usages of both moods and tenses. The perfect as distinguished from the other forms expresses either repetition of the action, emphasis, or the state which results from the action expressed by the verb. Different languages regard this last in different ways. Sometimes the state resulting from the action is so characteristic that the perfect is almost an independent verb. Thus in Greek κτάομαι is “I acquire,” but κέκτημαι (the perfect) is “I possess,” the result of the action of acquiring. On the other hand the perfect may mean that the action has come to an end. This is specially common in Latin, as in Cicero’s famous announcement of the execution of the Catilinarian conspirators,—Vixerunt (“They have lived” = “They are no more”). But it is by no means confined to Latin. The pluperfect, the past of the perfect, is a late development and can hardly be reckoned Indo-European. In Greek the forms clearly arise from adding aorist endings to a perfect stem. The forms of Latin are not yet completely explained—but it is certain that the specially Latin meaning expressing something that was past at a time already past (relative time) is a late growth. When Homeric Greek wishes to express this meaning it uses most frequently the aorist, but also the imperfect as well as the pluperfect, the notion of relative time being derived from the context. In the earliest Latin the pluperfect is not uncommonly used with the value of the aorist perfect. As regards the future it is difficult to say how far it was an original form. Some languages, like Germanic, preserve no original form for the future. When the present is found not to be distinctive enough, periphrastic forms come in. In other languages, like Latin and Greek, there is constant confusion between subjunctive and future forms. It is impossible to distinguish by their form between δείξω (future) and δείξω (subjunctive), between regam (future) and regam (subjunctive). A special future with a suffix -sḭo- (syo) is found only with certainty in the Aryan group and the Baltic languages. The future perfect is, strictly speaking, only a future made from a perfect stem; in the Latin sense it is certainly a late development, and even in early Latin, videro has occasionally no different meaning from videbo. The imperative, which was originally an exclamatory form to the verb, of the same kind as the vocative was to the noun, and which consisted simply of the verb stem without further suffixes, developed, partly on the analogy of the present and partly with the help of adverbs, a complete paradigm. The infinitives of all the languages are noun cases, generally stereotyped in form and no longer in touch with a noun system, though this, e.g. in early Sanskrit, is not always true. The participles differ only from other adjectives in governing the same case as their verb; and this is not an early distinction, for in the earliest Sanskrit all verbal nouns may govern the same case as their verb.
The system here sketched in the barest outline tended steadily to fall into decay. The case system was not extensive enough to express even the commonest relations. Thus there was no means of distinguishing by the cases between starting from outside and starting from inside, ideas which, e.g. Finnish regards as requiring separate cases; without a preposition it was impossible to distinguish between on and in, though to the person concerned there is much difference, for example between being on a river and in a river. There are other difficulties of the same kind. These had to be got over by the use of adverbs. But no sooner had the adverbs become well established for the purpose of defining these local relations than the meaning was felt to exist more in the adverb than in the case ending. For this syntactical reason, as well as for mechanical reasons arising from accent (q.v.), the case system in some languages fell more and more into desuetude. In Sanskrit it has been kept entire, in Balto-Slavonic the only loss has been the disappearance of the original genitive and its replacement by the ablative. In Latin the locative has been confused with the genitive and the ablative, and the instrumental with the ablative. The loss of the locative as an independent case had not long preceded historical times, because it survives in Oscan, the kindred dialect of the neighbouring Campania. Greek has confused ablative with genitive, except for one small relic recently discovered on an inscription at Delphi; in the consonant stems it has replaced the dative by the locative form and confused in it dative, locative and instrumental meanings. In some other members of the family, e.g. Germanic, the confusion has gone still farther.
The fate of the verb is similar, though the two paradigms do not necessarily decay at the same rate. Thus Latin has modified its verb system much more than its noun system, and Greek, while reducing seriously its noun forms, shows a very elaborate verb system, which has no parallel except in the Aryan group. From the syntactical point of view, however, the Greek system is much superior to the Aryan, which has converted its perfect into a past tense in classical Sanskrit, and to a large extent lost grip of the moods. The decay in Aryan may be largely attributed to the power, which this group developed beyond any other, of making compounds which in practice took the place of subordinate sentences to a large extent. The causes for the modifications which the Latin verb system has undergone are more obscure, but they are shared not only by its immediate neighbours the other Italic dialects, but also to a great degree by the more remote Celtic dialects.
The origin and spread of the Indo-European languages has long been, and remains, a vexed question. No sooner had Bopp laid the foundation of Comparative Philology in his great work, the first edition of which appeared in 1833-1835, than this question began to be seriously considered. The earlier writers agreed in regarding Asia as the original home of the speakers of these languages. For this belief there were various grounds,—statements in the Biblical record, the greater originality (according to Schlegel) of Sanskrit, the absurd belief that the migrations of mankind always proceeded towards the west. The view propounded by an English philologist, Dr R. G. Latham, that the original home was in Europe, was scouted by one of the most eminent writers on the subject—Victor Hehn—as lunacy possible only to one who lived in a country of cranks. Latham’s view was first put forward in 1851, and in half a century opinion had almost universally come over to his side. Max Müller indeed to the last held to the view that the home was “somewhere in Asia,” and Professor Johannes Schmidt of Berlin, in a paper read before the Oriental Congress at Stockholm in 1889, argued for a close contact between early Indo-European and Assyrian civilization, from the borrowing of one or two words and the existence of duodecimal elements in the Indo-European numeral system side by side with the prevalent decimal system—the dozen, the gross, the long hundred (120), &c. At 60 the systems crossed, and 60 was a very characteristic element in Assyrian numeration, whence come our minutes and seconds and many other units.
Even before Latham a Belgian geologist, d’Omalius d’Halloy, in 1848 had raised objections to the theory of the Asiatic origin of the Indo-Europeans, but his views remained unheeded. In 1864 he brought three questions before the Société d’anthropologie of Paris: (1) What are the proofs of the Asiatic origin of Europeans? (2) Have not inflectional languages passed from Europe to Asia rather than from Asia to Europe? (3) Are not the speakers of Celtic languages the descendants of the autochthonous peoples of Western Europe? (Reinach, op. cit. p. 38). Broca in replying to d’Omalius emphasized the fact which has been too often forgotten in this controversy, that race and language are not necessarily identical. In 1868 Professor Benfey of Göttingen argued for the south-east of Europe as the original home, while Ludwig Geiger in 1871 placed it in Germany, a view which in later times has had not a few supporters.
Truth to tell, however, we are not yet ready to fix the site of the original home. Before this can be done, many factors as yet imperfectly known must be more completely ascertained. The prehistoric conditions of Northern, Western, Central and South-eastern Europe have been carefully investigated, but important new discoveries are still continually being made. Investigation of other parts of Europe is less complete, and prehistoric conditions in Asia are at present very imperfectly known. In Western Europe two prehistoric races are known, the palaeolithic and the neolithic. The former, distinguished by their great skill in drawing figures of animals, especially the horse, the reindeer, and the mammoth, preceded the period of the Great Ice Age which rendered Northern Europe to the latitude of London and Berlin uninhabitable for a period, the length of which, as of all geological ages, cannot definitely be ascertained. For the present purpose, however, this is of less importance, because it is not claimed that the Indo-European stock is of so great antiquity. But when the ice again retreated it must have been long before Northern Europe could have maintained a population of human beings. The disappearance of the surface ice must have been followed by a long period when ice still remained underground, and the surface was occupied by swamps and barren tundras, as Northern Siberia is now. When a human population once more occupied Northern Europe it is impossible to estimate in years.
The problem may be attacked from the opposite direction. How long would it have taken for the Indo-European stock to spread from its original home to its modern areas of occupation? Some recent writers say that it is unnecessary to carry the stock back farther than 2500 B.C.—a period when the civilizations of Egypt and Mesopotamia were already ancient. Wherever the original home was situated, this date is probably fixed too low. The discussion, moreover, is in danger not only of moving in one vicious circle but in two. (a) The term “Indo-European stock” necessarily implies race, but why might not the language have been from the earliest times at which we can trace it the language of a mixed race? (b) It is usual to assume that the Indo-European stock was tall and blond, in fact much as the classical writers describe the early Germans. But the truth of this hypothesis is much more difficult to demonstrate. In most countries known to the ancients where blond hair prevailed, at the present day dark or brown hair is much more in evidence. Moreover the colour of fair hair often varies from childhood to middle life, and the flaxen hair of youth is very frequently replaced by a much darker shade in the adult. It has been often pointed out that many of Homer’s heroes are xanthoi, and it is frequently argued that ξανθός means blond. This, however, is anything but certain, even when Vacher de Lapouge has collected all the passages in ancient writers which bear upon the subject. When Diodorus (v. 32) wishes to describe the children of the Galatae, by whom apparently he means the Germans, he says that their hair as children is generally white, but as they grow up it is assimilated to the colour of their fathers. The ethnological argument as to long-headed and short-headed races (dolichocephalic and brachycephalic) seems untrustworthy, because in countries described as dolichocephalic short skulls abound and vice versa. Moreover this classification, to which much more attention has been devoted than its inventor Retzius ever intended, is in itself unsatisfactory. The relation between the length and breadth of the head without consideration of the total size is clearly an unsatisfactory criterion. It is true that to the mathematician ¾ or 6⁄8 or 9⁄12 are of identical value, but, if it be also generally true that mental and physical energy are dependent on the size and weight of the brain, then the mere mathematical relation between length and breadth is of less importance than the size of the quantities. Anthropologists appear now to recognize this themselves.
The argument from physical geography seems more important. But here also no certain answer can be obtained till more is known of the conditions, in early times, of the eastern part of the area. According to Ratzel the Caspian was once very much larger than it is now, and to the north of it there extended a great area of swamp, which made it practically impossible for the Indo-European race to have crossed north of the Caspian from either continent to the other. At an early period the Caspian and Black Sea were connected, and the Sea of Marmora and the Dardanelles were represented by a river which entered the Aegean at a point near the island of Andros. While the northern Aegean was still land divided only by a river, it is clear that migration from south-eastern Europe to Asia Minor, or reversely, might have taken place with ease. Even in much later times the Dardanelles have formed no serious barrier to migration in either direction. At the dawn of history, Thracian tribes crossed it and founded, it seems, the Phrygian and Armenian stock in Asia Minor; the Gauls at a later time followed the same road, as did Alexander the Great a generation earlier. At the end of the middle ages, Asia sent by way of the Dardanelles the invading Turks into Europe. The Greeks, a nation of seafarers, on the other hand reached Asia directly across the Aegean, using the islands, as it were, as stepping-stones.
Though much more attention has been devoted to the subject by recent writers than was earlier the practice, it is doubtful whether migration by sea has even now been assigned its full importance. The most mysterious people of antiquity, the Pelasgians, do not seem to be in all cases the same stock, as their name appears merely to mean “the people of the sea,” Πελασγοἰ representing an earlier πελαγς-κοι, where πελαγς is the weak form of the stem of πέλαγος, “sea,” and -κοι the ending so frequent in the names of peoples. A parallel to the sound changes may be seen in μίσγω, for *μίγ-σκω, by the side of μίγ-νυμι. As time goes on, evidence seems more and more to tend to confirm the truth of the great migrations by sea, recorded by Herodotus, of Lydians to Etruria, of Eteocretans both to east and west. An argument in favour of the original Indo-Europeans being seated in north-western Germany has been developed by G. Kossinna (Zeitschrift für Ethnologie, 1902, pp. 161-222) from the forms and ornamentation of ancient pottery. It has certainly not been generally received with favour, and as Kossinna himself affirms that the classification of prehistoric pottery is still an undeveloped science, his theory is clearly at present unequal to the weight of such a superstructure as he would build upon it. As the allied sciences are not prepared with an answer, it is necessary to fall back upon the Indo-European languages themselves. The attempt has often been made to ascertain both the position of the original home and the stage of civilization which the original community had reached from a consideration of the vocabulary for plants and animals common to the various languages of the Indo-European family. But the experience of recent centuries warns us to be wary in the application of this argument. If we cut off all past history and regard the language of the present day as we have perforce to regard our earliest records, two of the words most widely disseminated amongst the Indo-European people of Europe are tobacco and potato. Without historical records it would be impossible for us to discover that these words in their earliest European form had been borrowed from the Caribbean Indians. Most languages tend to adopt with an imported product the name given to it by its producers, though frequently misunderstanding arises, as in the case of the two words mentioned, the potato being properly the yam, and tobacco being properly the pipe, while petum or petun (cp. petunia) was the plant.
The first treatise in which an attempt was made to work out the primitive Indo-European civilisation in detail was Adolphe Pictet’s Les Origines indo-européennes ou les Aryas primitifs (1859-1863). The idyllic conditions in which, according to Pictet, early Indo-European man subsisted were accepted and extended by many enthusiastic successors. The father, the protector of the family (pater from pā, protect), and the mother (mater from mā, to produce) were surrounded by their children (Skt. putra), whose name implied that they kept everything clean and neat. The daughter was the milkmaid (Skt. duhitā from duh, milk), while the brother (Skt. bhrātār), derived from the root of ferre, “bear,” was the natural protector of his sister, whose name, with some hesitation, is decided to mean “she who dwells with her brother,” the notion of brother and sister marriage being, however, summarily rejected (ii. p. 365). The uncle and aunt are a second father and mother to the family, and for this reason nepos, Skt. napāt, is both nephew and grandson. The life of such families was pastoral but not nomad; there was a farmstead where the women were busied with housewifery and butter-making, while the men drove their flocks afield. The ox, the horse, the sheep, the goat and the pig were domesticated as well as the dog and the farmyard fowls, but it was in oxen that their chief wealth consisted. Hence a cow was offered to an honoured guest, cows were the object of armed raids upon their neighbours, and when a member of the family died, a cow was killed to accompany him in the next world. Even the phenomena of nature to their naive imaginations could be represented by cows: the clouds of heaven were cows whose milk nourished the earth, the stars were a herd with the sun as the bull amongst them, the earth was a cow yielding her increase. Before the original community, which extended over a wide area with Bactria for its centre, had broken up, agriculture had begun, and barley, if not other cereals, and various leguminous plants were cultivated. Oxen drew the plough and the wagon. Industry also had developed with the introduction of agriculture; the carpenter with a variety of tools appears to construct farm implements, buildings and furniture, and the smith is no less busy. Implements had begun with stone, but by this time were made of bronze if not of iron, for the metals gold, silver, copper, tin were certainly known. Spinning and weaving had also begun; pottery was well developed. The flocks and herds and agriculture supplied food with plenty of variety; fermented liquors, mead, probably wine and perhaps beer, were used, not always in moderation. A great variety of military weapons had been invented, but geographical reasons prevented navigation from developing in Bactria. Towns existed and fortified places. The people were organized in clans, the clans in tribes. At the head of all, though not in the most primitive epoch, was the king, who reigned not by hereditary right, but by election. Though money had not yet been invented, exchange and barter flourished; there were borrowers and lenders, and property passed from father to son. Though we have no definite information as to their laws, justice was administered; murder, theft and fraud were punished with death, imprisonment or fine (Résumé général at end of vol. ii.).
Further investigation, however, did not confirm this ideally happy form of primitive civilization. Many of Pictet’s etymologies were erroneous, many of his deductions based on very uncertain evidence. No recent writer adopts Pictet’s views of the Indo-European family. But his list of domesticated animals is approximately correct, if domestication is used loosely simply of animals that might be kept by the Indo-European man about his homestead. Even at the present day domestication means different things in the case of different animals. A pig is not domesticated as a dog is; in areas like the Hebrides or western Ireland, where cattle and human beings share the two ends of the same building, domestication means something very different from the treatment of large herds on a farm extending to many hundreds of acres. In other respects the height of the civilization was vastly exaggerated. That the Indo-European people were agricultural as well as pastoral seems highly probable. But as Heraclides says of the Athamanes (Fragmenta hist. Graec. ii. 219), the women were the agriculturists, while the men were shepherds. Agriculture begins on a very small scale with the dibbling by means of a pointed stick of a few seeds of some plant which the women recognize as useful either for food or medicine, and is possible only when the people have ceased to be absolutely nomad and have fixed settlements for continuous periods of some length. The pastoral habit is broken down in men only by starvation, if the pasture-lands become too cramped through an excessive increase of population or are seized by a conqueror. As has been well said, “of all the ordinary means of gaining a livelihood—with the exception perhaps of mining—agriculture is the most laborious, and is never voluntarily adopted by men who have not been accustomed to it from their childhood” (Mackenzie Wallace, Russia, new ed. i. p. 266, in relating the conversion of the Bashkir Tatars to agriculture). Even the plough, in the primitive form of a tree stump with two branches, one forming the handle, the other the pole, was developed, and to this period may belong the representations in rock carvings in Sweden and the Alps of a pair of oxen in the plough (S. Müller, Nordische Altertumskunde, i. 205; Dechelette, Manuel d’archéologie, ii. pp. 492 ff.). The Indo-European civilization in its beginnings apparently belongs to the chalcolithic period (sometimes described by the barbarous term of Italian origin eneolithic) when copper, if not bronze had come in, but the use of stone for many purposes had not yet gone out. While primitive Indo-European man apparently knew, as has been said, the horse, ox, sheep, goat, pig and dog, it is to be observed that in their wild state at least these animals do not all affect the same kind of area. The horse is an animal of the open plain; the foal always accompanies the mother, for at first its neck is too short to allow it to graze, and the mare, unlike the cow, has no large udder in which to carry a great supply of milk. The cow, on the other hand, hides her calf in a brake when she goes to graze, and is more a woodland animal. The pig’s natural habitat is the forest where beech mast, acorns, or chestnuts are plentiful. The goat is a climber and affects the heights, while the sheep also prefers short grass to the richer pastures suited to kine. To collect and tame all those animals implies control of an extensive and varied area.
What of the trees known to primitive Indo-European man? On this the greater part of the arguments regarding the original home have turned. The name for the beech extends through a considerable number of Indo-European languages, and it has generally been assumed that the beech must have been known from the first and therefore must have been a tree which flourished in the original home. Now the habitat of the beech is to the west of a line drawn from Königsberg to the Crimea. The argument assumes that its distribution was always the same. But nothing is more certain than that in different ages different trees succeed one another on the same soil. In the peat mosses of north-east Scotland are found the trunks of vast oaks which have no parallel among the trees which grow in the same district now, where the oak has a hard struggle to live at all, and where experience teaches the planter that coniferous trees will be more successful. On the coast of Denmark in the same way the conifer has replaced the beech since the days of the “kitchen middens,” from which so much information as to the primitive inhabitants of that area has been obtained. But with regard to the names of trees there are two serious pitfalls which it is difficult to avoid. (a) It is common to give a tree the name of another which in habit it resembles. In England the oriental plane does not grow freely north of the Trent; accordingly, farther north the sycamore, which has a leaf that a casual observer might think similar, has usurped the name of the plane. (b) In the case of the beech (Lat. fagus), the corresponding Greek word φηγός does not mean beech but oak, or possibly, if one may judge from the magnificent trees of north-west Greece, the chestnut. It has been suggested that the word is connected with the verb φαγεῖν to eat, so that it was originally the tree with edible fruit and could thus be specialized in different senses in different areas. If, however, Bartholomae’s connexion of the Kurd būz, “elm” (Idg. Forschungen, ix. 271) be correct, there can be no relation between φαγεῖν and φηγός, but the latter comes from a root *bhāuĝ, in which the g would become z among the satem languages. The birch is a more widely spread tree than the beech, growing as luxuriantly in the Himalayas as in western Europe, but notwithstanding, the Latin fraxinus, which is almost certainly of the same origin, means not birch but ash, while the word akin to ash (Gr. ὀξύη) appears in Latin without the k suffix as os- in Latin ornus, “mountain ash,” for an earlier *osinos, cp. Old Bulgarian jasenŭ (the j has no etymological value), Welsh and Cornish onnen, from an original Celtic *onna from *os-nā. One of the most widely spread tree names is the word tree itself, which appears in a variety of forms, Gr. δρῦς, Goth triu; Skt. dāru, δόρυ, &c., which is sometimes as in Greek specially limited to the oak, while the Indian deodar (deva-dāru) is a conifer. O. Schrader, who in his remarkable book, Sprachvergleichung und Urgeschichte (1883, 3rd ed., 1906-1907), locates the original home in southern Russia, would allow the original community (ii. p. 178) to be partly within, partly without the beech line. The only other tree the name of which is widely spread is the willow: the English with, withy, Lat. vitex, Gr. ἰτέα for ϝιτέα, Lithuanian wýtis, Zend vaêti. Otherwise the words for trees are limited to a small number of languages, and the meaning in different languages is widely different, as Gr. ἐλάτη, “pine,” Old High German linta, “linden,” with which go the Latin linter, “boat,” and Lithuanian lentà, “board.” The lime tree and the birch do not exist in Greece, and the Latin betula is a borrowing from Gaulish (Irish bethe), the native word fraxinus, as we have seen, being used for the ash. The equation of the Latin taxus, “yew,” with Gr. τόξον, “bow,” is no doubt correct; Schrader’s equation of Skt. dhanvan, “bow,” with the German tanne, “fir,” must, if correct, show at least a change of material, for no wood is less well adapted for a bow than fir. The only conclusion that can be drawn with apparent certainty from the names of trees is that the original settlements were not in the southern peninsulas of Europe.
Some of the names for cultivated plants are widely spread, but like the names of trees do not always indicate the same thing. This is not surprising if we consider that the word corn, within the Teutonic languages alone, means wheat in England, oats in Scotland, rye in Germany, barley in Sweden, maize in the United States of America. Thus the Skt. yáva means corn or barley, in Zend corn (modern Persian jav, barley, but in the language of the Ossetes yeu, yau is millet), the Gk. ζεά is spelt, the Lithuanian jawaĩ corn, the Irish éorna barley (Schrader, Sprachvergleichung3 ii. p. 188). The word bere or barley itself is widely spread in Europe—Latin far, spelt, Goth, barizeins, “of barley,” Old Norse barr, Old Slav, bŭrŭ, a kind of millet (ibid.). But the original habitat of the cultivated grain plants has not yet been clearly established, and circumstances of many kinds may occasion a change in the kind of grain cultivated, provided another can be found suitable to the climate. In early England it is clear that the prevalent crop was barley, for barn is the bere-ern or barley-house.
The earliest tree-fruits found in Europe are apparently those discovered by Edouard Piette as Mas d’Azil in a stratum which he places between palaeolithic and neolithic. They included nuts, plums, birdcherry, sloe, &c., and along with them was a little heap of grains of wheat. If Piette’s observations are correct, this find must go back to a date long preceding the fruits found by Heer in the pile-dwellings of Switzerland. Here also cherry-stones were found, though the modern cherry is said to have been imported first by Lucullus in the first century B.C. from Cerasus in Pontus, whence its name. In the pile-dwellings a considerable number of apples were found. They were generally cut up into two or three pieces, apparently to be dried for winter use. In all probability they were wild apples of the variety Pirus silvatica, which is found across the whole of Central Europe from north to south (Buschan, Vorgeschichtliche Botanik, p. 166). The original habitat of the apple is uncertain, but it is supposed to be indigenous at any rate south of the Black Sea (Schrader, Reallexikon, s.v. Apfelbaum). The history of the name is obscure; it is often connected with the Campanian town Abella, which Virgil (Aeneid, vii. 740) calls malifera, “apple-bearing.” Here also the material for fixing the site of the original habitat is untrustworthy.
The attempt has been made to limit the possible area by a consideration of three animals which are said not to occur in certain parts of it—(a) the eel, which is said not to be found in the Black Sea; (b) the honey bee, which is not found in that part of Central Asia drained by the Oxus and Jaxartes; (c) the tortoise, which is not found in northern areas. From evidence collected by Schrader from a specialist at Bucharest (Sprachvergleichung,3 ii. p. 147) eels are found in the Black Sea. The argument, therefore, for excluding the area which drains into the Black Sea from the possible habitat of the primitive Indo-European community falls to the ground. Honey was certainly familiar at an early age, as is shown by the occurrence of the word *medhu, Skt. mádhu, Gr. μέθυ (here the meaning has shifted from mead to wine), Irish mid, English mead, Old Slav, medŭ, Lithuanian medùs honey, midùs mead. Schrader, who is the first to utilize the name of the tortoise in this argument, points out (op. cit. p. 148) that forms from the same root occur in both a centum and a satem language—Gr. χελύς, χελώνη, Old Slav. žĭly, želŭvĭ—but that while it reaches far north in eastern Europe, it does not pass the 46th parallel of latitude in western Europe. This argument would make not only the German site for the original home which is supported by Kossinna and Hirt impossible, but also that of Scandinavia contended for by Penka.
From the foregoing it will be seen that the arguments for any given area are not conclusive. In the great plain which extends across Europe north of the Alps and Carpathians and across Asia north of the Hindu Kush there are few geographical obstacles to prevent the rapid spread of peoples from any part of its area to any other, and, as we have seen, the Celts and the Hungarians, &c., have, in the historical period, demonstrated the rapidity with which such migrations could be made. Such migration may possibly account for the appearance of a people using a centum language so far east as Turkestan. But our information as to Tocharish is still too fragmentary to decide the question. It is impossible here to discuss at any length the relations between the separate Indo-European languages, a subject which has formed, from somewhat different points of view, the subject of Kretschmer’s Einleitung in die Geschichte der griechischen Sprache and Meillet’s Les Dialectes indo-européennes.
Bibliography.—Besides the articles on the separate languages in this Encyclopaedia the following works are the most important for consultation: K. Brugmann (phonology and morphology) and B. Delbrück (syntax), Grundriss der vergleichenden Grammatik der indogermanischen Sprachen (1886-1900), ed. 2, vol. i. (1897); of vol. ii. two large parts, including the stem formation and inflexion of the noun, the pronoun and the numerals, have been published in 1906 and 1909. A shorter work by Brugmann, Kurze vergleichende Grammatik der indogermanischen Sprachen, dealing mainly with Sanskrit, Greek, Latin, Germanic and Slavonic, appeared in three parts in 1902-1903. A good but less elaborate work is A. Meillet, Introduction à l’étude comparative des langues indo-européennes (1903, 2nd ed. 1908). For the ethnological argument: W. Z. Ripley, The Races of Europe (1900); G. Sergi, The Mediterranean Race (English edition, 1901). Other works, now largely superseded, which deal with this argument are K. Penka, Origines Ariacae (1883), and Die Herkunft der Arier (1886), and I. Taylor, The Origin of the Aryans, N.D. (1890). The ethnologists are no more in agreement than the philologists. For the arguments mainly from the linguistic side see especially O. Schrader, Sprachvergleichung und Urgeschichte (3rd ed., 2 vols., 1906-1907)—the second edition was translated into English by Dr F. B. Jevons under the title Prehistoric Antiquities of the Aryan Peoples (1890); Reallexikon der indogermanischen Altertumskunde (1901); M. Much, Die Heimat der Indogermanen (1902, 2nd ed. 1904); E. de Michelis, L’Origine degli Indo-europei (1903); H. Hirt, Die Indogermanen (2 vols., 1905-1907); S. Feist, Europa im Lichte der Vorgeschichte und die Ergebnisse der vergleichenden indogermanischen Sprachwissenschaft, 1910, in W. Sieglin’s Quellen und Forschungen zur alten Geschichte und Geographie. Important for special sections of this question are S. Müller, Nordische Altertumskunde (2 vols., 1897-1898), and Urgeschichte Europas (1905); V. Hehn, Kulturpflanzen und Haustiere (1870), 7th ed. edited by O. Schrader, with contributions on botany by A. Engler (1902); J. Hoops, Waldbäume und Kulturpflanzen im germanischen Altertum (1905). Delbrück has devoted a special monograph to the I.-E. names of relationships, from which he shows that the I.-E. family was patriarchal, not matriarchal (Die idg. Verwandtschaftsnamen, 1889). E. Meyer, from Tocharish being a centum language, has revived with reserve the hypothesis of the Asiatic origin (Geschichte des Altertums,2 I. 2, p. 801).
- (P. Gi.)
- E. Meyer, Sitzungsberichte der Berliner Akademie (1908, pp. 14 ff.), and more fully in Kuhn’s Zeitschrift (xlii. pp. 17 ff.); also Geschichte des Altertums (i. 2, 2nd ed. pp. 807 ff.).
- Sieg und Siegling, “Tocharisch, die Sprache der Indoskythen” (Sitzb. d. Berl. Ak. 1908, pp. 915 ff.).
- No. xix. p. 255, “Another ancient and extensive class of languages, united by a greater number of resemblances than can well be altogether accidental, may be denominated the Indo-European, comprehending the Indian, the West Asiatic, and almost all the European languages.”
- Leo Meyer, “Über den Ursprung der Namen Indogermanen, Semiten und Ugrofinner,” in the Göttinger gelehrte Nachrichten, philologisch-historische Klasse, 1901, pp. 454 ff.
- The vocative is not strictly speaking a case at all, for it stands outside the syntax of the sentence. It was originally an exclamatory form consisting of the bare stem without case suffix. In the plural the nominative is used to supply the lacking vocative form.
- For the history of the controversy see the excellent summary in Salomon Reinach’s L’Origine des Aryens: Historie d’une controverse (1892). Max Müller’s latest views are contained in his Biographies of Words and the Home of the Aryas (1888). See Schmidt’s Die Urheimat der Indogermanen und das europäische Zahlsystem (1890).
- “Geographische Prüfung der Tatsachen über den Ursprung der Völker Europas” (Berichte der k. sächsischen Ges. d. Wissenschaften, 1900, pp. 34 ff.).
- See the essay on “Evolution and the Science of Language,” in Darwin and Modern Science (1909), p. 524 f.