This book, as its title indicates, is the outcome of research which had as its object a considerable portion of the Indo-European vocabulary. But the nature of the expressions studied, the method applied, and the analysis carried out, need elucidation.

Among the languages of the world those belonging to the Indo-European family best lend themselves to wide-ranging study both in space and time. Such studies can also be pursued in the greatest variety and depth, since these languages extend from Central Asia to the Atlantic and are attested for a period of almost four millennia. Further, they are bound up with very ancient civilizations of varying degrees of development, some of them ranking among the richest which have ever existed. Finally, certain of these languages have produced an abundant literature of a very high order, and for this reason were for a long period the exclusive object of linguistic analysis.

Indo-European is defined as a family of languages, issuing from a common language, which have become differentiated by gradual separation. This constitutes a global event, immense in scope, which we are able to grasp in its entirety because in the course of time it broke up into a series of separate histories, each of them that of a particular language.

It is a very remarkable fact indeed that we are able to single out the peoples which partook in the original community and to designate them with certainty as Indo-Europeans to the exclusion of all others, because the stages of their migrations and their settlements remain unknown. The reason for this is language and language alone. The notion ‘Indo-European’ is primarily a linguistic one, and if we are in a position to extend it so as to include other aspects of their civilization, this again is due solely to language. The concept of genetic relationship has in no other linguistic domain so precise a sense and such clear justification. We find in Indo-European a model of the correspondence relationships which delimit a family of languages and which allow us to reconstruct their earlier stages back to the initial one.

For the last hundred years the comparative study of the Indo-European languages has been pursued in two opposed but complementary directions. On the one hand, reconstructions are made from simple or complex elements, be they phonemes, whole words, or inflections, which are susceptible to comparison in different languages and so can make their contribution to a reconstruction of the common prototype. In this manner models are devised which in their turn form the basis of new reconstructions. On the other hand, we may proceed in the opposite direction: we start with a well-established Indo-European form and trace the forms which are descended from it. This method traces the paths of dialectal differentiation which resulted in new unities. At this stage the elements inherited from a common language are found incorporated in independent structures which constitute individual languages. As such they are transformed and assume new values within the oppositions by which they are created and which they determine. Thus we must not merely study the possibilities of reconstruction which summarize vast series of correspondences and reveal the structure of common elements; we must also examine the development of individual languages, because it is here that we have the productive medium, the source of the innovations which transform the ancient system. The comparative linguist thus moves between these two poles and his efforts are precisely directed towards distinguishing between conservation and innovation; he must account both for identities and differences.

To these general considerations, which the principle of linguistic comparison imposes, must be added the specific traits within the lexical domain, which are those concerned in the present study.

Very early on it occurred to specialists in the Indo-European field that correspondences between the vocabularies of ancient languages illustrate the principal aspects of a common culture, particularly of material culture. Thus instances of the lexical inheritance were collected from expressions for family relationships, numbers, names of animals, metals, agricultural implements, etc. A series of authors, ranging from the nineteenth century until recent times, devoted themselves to the compilation of such lists of common expressions, which are of an evident utility.

Our enterprise is of a wholly different nature. No attempt has been made to compile yet one more inventory of Indo-European ‘facts’ in so far as they are defined by lexical correspondences. On the contrary, most of the material we are concerned with does not belong to the common vocabulary. The forms involved are specifically expressions relating to institutions, but in particular languages; and what we propose to analyse is their genesis and their Indo-European connexions. In other words, we propose to study the formation and organization of the vocabulary of institutions.

The expression ‘institution’ is here understood in a wider sense: it includes not only the institutions proper, such as justice, government, religion, but also less obvious ones which are found in various techniques, ways of life, social relationships and the processes of speech and thought. The subject is truly boundless, the aim of our study being precisely to throw light on the genesis of the vocabulary which relates to it. Our starting point is usually one or the other of the Indo-European languages and the examples chosen come from the terms of pregnant value. Round the chosen datum, by an examination of its peculiarities of form and sense, its connexions and oppositions and, following this, by comparison with related forms, we reconstruct the context in which it became specialized, often at the cost of profound transformations. In this way we endeavour to restore a unity dissolved by processes of evolution, bringing buried structures to light and harmonizing the divergencies of technical usages. In so doing we shall also demonstrate how languages reorganize their systems of distinctions and renew their semantic apparatus.

I leave to others the historical and sociological aspects of these processes. If we deal with the Greek verb hēgéomai and its derivative hēgemṓn, this is in order to see how the notion of ‘hegemony’ was established, but without regard to the fact that Greek hēgemonía came to mean successively the supremacy of an individual, or a nation, or the equivalent of the Roman imperium, etc. What concerns us is the connexion, difficult to account for, between an expression of authority such as hēgemṓn and the verb hēgéomai which means ‘to think, to judge’. In so doing we explain the signification, leaving to others the problem of designation. When we discuss the Germanic word feudum in connexion with the terms for animal husbandry, feudalism itself is mentioned merely by preterition. This approach will make it easier for historians and sociologists to see what use they can make of analyses presented here, precisely because no extra-linguistic presuppositions have intruded.

The task of the linguist is delimited in the following way. He takes his material from the vast store of established correspondences which are transferred without much change from one etymological dictionary to another. This material is of its very nature far from homogeneous. Each separate fact comes from a different language and constitutes part of a distinct system which develops along unpredictable lines. The primary task is to demonstrate that these forms correspond to one another and that they are all direct continuations of some original form. This established, we have next to account for the differences, sometimes considerable, which they may present with regard to their phonetic appearance, their morphology, or their meaning. Thus we may equate the Armenian word k^cun ‘sleep’ with Latin somnus ‘sleep’, because we know the rules of correspondences which allow us to reconstruct a common form *swopno-. It is possible to connect the Latin verb carpo ‘to gather’ with the German noun Herbst ‘autumn’, because Herbst in Old High German is herbist, and herbist may be traced back to a pre-Germanic form *karpisto- which means ‘(time) most appropriate for harvesting’ (cf. Engl. harvest); and this is further confirmed by a third datum, the Greek noun karpós ‘fruits of the earth, harvest produce’. But a simple comparison which seems acceptable at first sight, like the root teks- in Latin (in the verb texo) and the root takṣ- in Sanskrit, two forms which correspond exactly, runs into grave difficulties: Latin texo means ‘weave’, whereas Sanskrit takṣ means ‘cut with an axe’; and one cannot see how one meaning can be derived from the other, nor from which original meaning either could have evolved, since ‘weaving’ and ‘carpentry’ seem irreducible to a common technique.

Even within the corpus of a single language, forms of the same word can be divided into distinct groups which seem hardly reconcilable. Thus from the root *bher-, represented by fero in Latin, three separate groups of derivatives have evolved which form as many lexical families: (1) fero ‘to carry’ in the sense of gestation, from which forda ‘pregnant female’ is derived, linking up with gesto; (2) fero ‘carry’ in the sense of ‘bring about, involve, entail’ is used with reference to manifestations of chance, hence fors, fortuna and their numerous derivatives, which also include the notion of ‘fortune, riches’; (3) fero ‘carry’ in the sense of ‘carry off’ forms a group with ago and can be defined as referring to seizure and booty. If we compare with this the forms derived from bhar- in Sanskrit the picture becomes still more varied. To the senses just listed we must add those of ‘to carry’ in the sense of ‘support, take care of’, hence the derivative bhartr̥- ‘husband’; from ‘carry’ in connexion with horse riding comes ‘ride’, etc. Thus one only has to study in detail one of these groups to see that in every case they constitute a coherent lexical unit hinging on a central notion, readily supplying institutional expressions.

An attempt has been made to show how words which at first exhibited little differentiation progressively acquired specialized applications and evolved semantic subfamilies that reflect a profound evolution of institutions, as well as the emergence of new activities or ideas. Such developments within a particular language may also come to influence other languages through cultural contact. For instance, lexical relationships established by processes peculiar to Greek served as models through translation or simple borrowing for similar relationships in Latin.

We have tried to bring out the dual character peculiar to the phenomena here described. On the one hand we are faced with the tangled web of developments which may take centuries or even millennia and which the linguist must trace back to their primary causation; and on the other hand the investigator must try to bring out certain universal tendencies which govern these individual developments. We can understand them, apprehend their structure, and arrange them in a rational schema (1) if we are able to study them directly and avoid the pitfalls of simplified translations; (2) if we are able to establish certain essential distinctions, notably one on which we have repeatedly insisted, namely that between designation and signification. For without this distinction so many discussions of ‘meaning’ end in confusion. The task is, by comparison and diachronic analysis, to elicit a ‘signification’, whereas our starting point will merely be a ‘designation’. By such a method the chronological approach is tantamount to an explanatory one.

The nature of this research determines the manner of exposition. No discussions of detail or bibliographical references will be found. The material used in the analyses is to be found in all the etymological dictionaries, but we are not aware of much previous work with which we could have compared our arguments. Everything here said stems from first-hand study of the facts used. We have made every effort to be intelligible to the non-specialist reader with strict regard for the exigencies of demonstration. But it must be conceded that the ramifications and complex connexions which came to light in the course of this exploration make coherent exposition difficult. It is not easy to make neat distinctions between the various subjects under discussion. Inevitably there will be some overlapping between the various parts of this work because this is inherent in the facts of the vocabulary. All the same, we hope that those who are willing to follow this exposition of our researches to the end will find in it matter for more general considerations, especially on the possibility of applying certain of the proposed models to the study of languages or civilizations in which, because of the lack of written documentation, there is also a lack of historical perspective.

The present work is based on several series of lectures given at the Collége de France which M. Lucien Gerschel has been kind enough to collect. They have been thoroughly revised and recast, and the first draft has often been entirely rewritten and recent results have been added. Some parts had previously been the subject of detailed treatment in published articles, and references to these are given. In order to clarify the exposition we have followed the suggestion by M. Pierre Bourdieu, who read the whole manuscript and made some useful comments: each chapter is preceded by a brief resumé, which is the work of M. Jean Lallot. This scholar kindly prepared the manuscript for the press and also drew up the tables of languages and made the index. I should like to thank him here for his help and the meticulous execution of his task.