Daley, Andrew. 1998. Dictionary of Languages. London: Bloomsbury
Indo-European languages are spoken on every continent and by members of every racial group. English, Spanish, French and Russian have some official status in dozens of countries worldwide. Hindi, Bengali, Portuguese and German are just as essential in many multinational contexts. Latin, Greek, Sanskrit, Pali and Avestan are classical languages of religion, philosophy and culture. Greek, Armenian, Yiddish and Romani are the languages of worldwide diasporas. There are many other Indo-European languages, from Icelandic to Italian and Persian, whose literatures enrich humanity.
How is it known that all these languages are related to one another?
In the 17th and 18th centuries a European scholar, having begun to master Latin and Greek and one or two modem European languages, might well go on to learn Hebrew - the language of the Old Testament - and perhaps Arabic. These were readily seen to be wholly different in their structure from the ancient and modem languages of Europe. Then European interest in India began to grow, and there was the opportunity to study Sanskrit, the classical language of India, under the guidance of Indian teachers. It was realised that Sanskrit - quite unlike Hebrew and Arabic - showed pervasive similarities with Latin and Greek and other languages of Europe. How did this come about? The speakers of these languages lived thousands of miles apart, and history told of no early contact between them. It was a puzzle that Indo-European scholars, ever since, have continued to explore.
It was clear that the known Indo-European languages must have gradually grown apart through changes in sounds: hence the difference in the initial consonants of English brother, Latin frater, Greek phrater, Sanskrit bhratr, words that were clearly related both in sound and meaning. The essential breakthrough in Indo-European research was the gradual realisation that these sound changes were regular. In the development of each particular language or dialect they took place invariably, whenever a certain sound occurred in defined surroundings; if there appeared to be exceptions, the exceptions should in principle have explanations. In this way Indo-European studies helped to galvanise historical work on other languages too, for the same principles can be applied in all language families.
Once the regularity of sound change was accepted, it was possible to set up hypotheses about the sounds and the words of the parent language, 'proto-Indo-European': to make a formula (for example, *bhrater for 'brother') to which word forms in the known languages could be traced back.
The rebuilding of proto-Indo European
One line of this research led to the gradual reconstruction of proto-Indo-European as an apparently natural language, with a complete sound system and a vocabulary to match. This is not a simple task. Some of the pitfalls become obvious as one studies the result.
The formulae or reconstructed forms are, in general, the most economical possible: such that the simplest possible series of changes will link them with known language forms. 'I his makes them efficient as formulae, but it does not prove that they really existed in a parent language. The sound system of proto-Indo-European, variously reconstructed by generations of scholars, and sometimes given as many as four 'laryngeal consonants' (see below), differs noticeably from the sound systems of natural languages. The real proto-Indo-European can never have been a unitary form of speech but always a bundle of varying speech forms, 'sociolects' and dialects, as all real languages are.
The reconstructed vocabulary of proto-Indo-European is a collection of words that continued to be used, for some thousands of years, in at least two of the later branches of Indo-European. In the case of words that continued to be used in only one branch, there can usually be no proof of their ultimate origin, so they have to be discounted. Vocabulary is never static, especially as speakers migrate and as their economy and their society changes. Words are forgotten, new words are invented, and words are given new meanings. It is not surprising, after all this, that there are far fewer words in the proto-Indo-European dictionary than there are in any natural language.
Finally, proto-Indo-European itself was the product of a history, probably involving earlier migrations, certainly involving changes in society and economy. In all natural languages, history is reflected in inconsistencies, words whose meanings are undergoing change, words that some speakers misunderstand.
Still, linguists and archaeologists do try to pin down, from the vocabulary of proto-Indo-European, where its speakers lived, what kind of society they lived in, how they farmed, how they lived and even (Emile Benveniste, Indo-European language and society, 1973) how they thought. They try to identify a 'homeland' or (to use the German term) Heimat for the Indo-Europeans.
Heimat: the 'homeland'
Very different results have been reached. In the early 19th century some thought that Sanskrit itself, first spoken in north-west India, was the parent tongue. Sir Monier Monier-Williams, author of the great Sanskrit dictionary, saw the vast ruins of Balkh in Afghanistan as the city from which all Indo-Europeans traced their origin. Others looked to the north German plain.
In the 20th century none of these views has found favour. For several decades now the majority of scholars has placed the 'homeland' on the south Russian steppes. The theory is particularly identified with the work of Marija Gimbutas, who presses the identification of the proto-Indo-Europeans with the builders of kurgans - burial mounds - in the eastern Ukraine.
A persistent minority looks south of the Black Sea. But these do not agree among themselves. The archaeologist Colin Renfrew (Archaeology and language: the puzzle of Indo-European origins, 1987) identifies the proto-Indo-Europeans with the earliest Neolithic farmers of central Asia Minor, between 7000 and 6000 B.C. Tomas Gamkrelidze and Vyacheslav Ivanov argue for an origin in what is now Armenia, perhaps equally early, followed by migration to the Ukraine, which then became a secondary 'homeland' and point of dispersal for the speakers of early Indo-European dialects. Robert Drews (The coming of the Greeks, 1988) also looks to Armenia - but as late as 1700 BC, when, he suggests, a chariot-riding elite spread their Indo-European speech across southern Europe and southern Asia.
The dialect of proto-Indo-European
Meanwhile, there has been work on the lines of descent from proto-Indo-European: how early, unrecorded dialects gradually differentiated into the widely different languages later known, and how these dialects were interlinked. They did not suddenly split. Just as with modem languages, there must for a long time have been a dialect continuum within Indo-European, marked by a succession of changes that affected different groups of dialects. On a modem dialect map the divisions are called isoglosses.
One, immediately noticeable, is called in linguists' shorthand the centum/satem split (see below). The innovation of the satem dialects -a very common one in language history - was the regular change from velar stops to dental fricatives when followed by a front vowel: thus, from proto-Indo-European *kmtom, Avestan satem contrasts with Latin centum (pronounced kentum).
Germanic, Baltic and Slavonic share some interesting features, such as the way in which the numerals '20' to '90' are built up. In the other Indo-European languages old compound words exist, such as Latin triginta, Greek triakonta for '30'. The corresponding terms in these north European languages are newly coined, and they have the obvious meaning 'three tens': Gothic threis tigjus, Lithuanian trys desimtys, Old Slavonic triye desete, tri deseti.
What was proto-Indo-European like?
Like Sanskrit, Old Slavonic, Latin and Greek, proto-Indo-European had numerous forms for nouns (incorporating the notions of number and 'case') and for verbs (marking 'person', number, time and 'mood'). At some time in distant prehistory, an agglutinating language -in which these notions were separately identifiable as affixes - had become a synthetic language, in which they were fused with one another and with the noun or verb root. What is more, the fusion took different forms depending on the shape of the root. Students of Latin and Greek are all too familiar with the resulting noun declensions and verb conjugations whose forms cannot be wholly predicted and must be separately learnt. Proto-Indo-European, it seems, was already like this.
The later history of all the languages of the family has been of a gradual reduction in this complexity, and thus of a long term change of character. While the early Indo-European languages were largely synthetic, English and Hindi (to take two examples) are largely isolating languages.
The first ten numerals in proto-Indo-European may be reconstructed as #oinas, #dwo, #treyes, #qwetwor, #penkwe, #sweks, #septm, #okto, #enewen, #dekm.
How did Indo-European languages spread so widely?
Colin Renfrew has argued that the spread of Neolithic farming from central Anatolia to northern Greece and the Balkans is the archaeologically visible sign of the earliest Indo-European expansion. Most Indo-European specialists have not yet accepted Renfrew's argument, but it cannot be said to have been disproved. Renfrew's more tentative hypothesis that Indo-European languages spread also to Iran and northern India at this early period, along with Neolithic farming, goes against the grain of the evidence.
The theory accepted by many linguists, though not by all, is that around 4000 BC an early group of Indo-European dialects was spoken across a wide swathe of central and eastern Europe, perhaps extending into southern Siberia and central Asia. The theory is often linked to the domestication of the horse - and certainly horsemanship has helped the spread of languages and empires, in this same steppe region, ever since.
Not long after 2000 BC the first surviving written records of Indo-European languages document Hittite - the language of a powerful kingdom in Asia Minor - and the influence of an early form of Indo-Aryan on the Hurrian-speaking Mitanni in the Middle East. It is likely, meanwhile, that other Indo-European dialects were gradually spreading further into western Europe. Outlying and less accessible regions - Spain, Italy, southern Greece, northern Scandinavia - retained their non-loco-European speech for rather longer. The speakers of early Indo-Iranian dialects must by now have been approaching Iran and north-western India from the north. The hypothesis of their long-distance migration seems necessary to explain the links between Indo-Iranian languages on the one hand and Slavonic and URALIC LANGUAGES on the other.
At the time when the Romans were establishing their empire, early texts from Europe and Asia show with fair certainty the location of most of the then surviving branches of Indo-European: Celtic languages, Germanic languages, Baltic languages, Slavonic languages, Latin and the other Italic languages, Illyrian (possibly the ancestor of Albanian), Thracian, Greek, Armenian, the Iranian languages, the Indo-Aryan languages and - far away to the northeast - the Tocharian languages. The latter are not especially similar to Iranian and Indo-Aryan: their geographical location must result from a quite different migration.
Latin, spread widely by trade and empire, gradually evolved into the Romance languages. More recently English, French, Spanish and Russian have spread across the world: will they, in turn, split into daughter languages?