Use the following persistent identifier: http://nrs.harvard.edu/urn-3:hul.ebook:CHS_Benveniste.Indo-European_Language_and_Society.1973.
Chapter 4. The Indo-European Expression for “Marriage” 
“Marriage” has no Indo-European term. In speaking of the man it is simply said—and this in expressions which have often been remodeled in particular languages—that he “leads” (home) a woman whom another man has “given” him (Lat. uxorem ducere and nuptum dare; in speaking of the woman, that she enters into the “married state,” receiving a function rather than accomplishing an act (Lat. ire in matrimonium).
The Indo-European vocabulary of kinship, ever since it has been the object of study, has taught us that in conjugality the situation of the man and that of the woman have nothing in common, just as the terms designating their respective relationship were completely different.
That is why there is, properly speaking, no Indo-European term for “marriage.” As Aristotle observed for his own language, “the union of man and woman has no name,” ἀνώνυμος ἡ γυναικὸς καὶ ἀνδρὸς σύζευξις (Polit. I, 3, 2). In fact, the expressions encountered today are all secondary creations; this is true of Fr. mariage, German Ehe (literally “law”), Russian brak (derived from brat’sja ‘carry off’), etc. In the ancient languages the facts are more specific, and it will be of interest to consider them in all their diversity.
This diversity is not merely lexical, a testimony of independent designation in each separate language; it is also morphological, and this fact, which is less obvious, has not been noticed. We have to clarify this in order that the facts may be organized: the terms differ according to whether the man or the woman is concerned, but the important difference is that for the man the terms are verbal, and for the woman nominal.
In order to say that a man “takes a wife”, Indo-European employs forms of a verbal root *wedh– ‘lead’, especially “lead a woman to one’s home.” This particular sense emerges from close correspondence between the majority of languages: Celtic (Welsh) dy–weddio, Slavic vedǫ, Lithuanian vedù, Av. vādayeiti, with the Indo-Iranian derivatives vadhū– ‘newly married woman’, Greek héedna (ἕεδνα) ‘marriage gift’.
Such was the expression in the most ancient stage and when certain languages found new words to express the notion of “to lead”, the new verb also assumed the value of “marry (a woman).” This is what happened in Indo-Iranian.  The root *wedh– survived in a large part of Iranian in the form of the verb vad-. But Indic has not preserved it: it has only kept the derived noun vadhū– ‘newly married woman’. Instead of *vadh– which has disappeared, it employs nay– for “lead” and also for “marry.” The same substitution of nay– for vad– is manifested in certain dialects of Iranian from Old Persian on, so that nay-and vad– were for a long period in competition on Iranian territory. In Latin, too, we find a new verb for the sense of “lead.” This is ducere, which also takes on the sense of “marry” in uxorem ducere. Another verb is peculiar to Greek, gameîn (γαμεῖν), which has no certain correspondences.
Besides these verbs which denote the role of the husband we must place those which indicate the function of the father of the bride. The father, or in default of this his brother, has authority to “give” the young woman to her husband: πατρὸς δόντος ἒ ἀδελπιõ, as the Law of Gortyn, chapter viii, puts it. “Give” is the verb constantly used for this formal proceeding; it is found in various languages, generally with some variation in the preverb: Greek doûnai (δοῦναι), ekdoûnai (ἐκδοῦναι), Latin dare, Gothic fragiban, Slavic otŭdati, Lithuanian išduoti, Skt. pradā-. Avestan uses paradātā and aparadātā to distinguish between the girl who has been properly “given” by her father and one who has not been so given. This constancy of expression illustrates the persistence of usages inherited from a common past and of the same family structure, where the husband “led” the young woman, whom her father has “given” him, to his home.
If we now search for terms employed to designate the “marriage” from the woman’s point of view, we find that there exists no verb denoting in her case the fact of marrying which is the counterpart of the expressions mentioned. The only verb which can be cited is the Latin nubere. But apart from being confined to Latin, nubere properly applies only to the taking of the veil, a rite in the ceremony of marriage, not to the marriage itself, or only by implication. In fact the verb is never used outside certain special circumstances. It serves, for instance, to stress a difference in the social condition between man and woman, as in a passage of Plautus (Aul. 479f.), where a character proposes “that the rich marry the daughters of the poor citizens, who have no dowry,” opulentiores pauperiorum filias ut indotatas ducant uxores domum, but he anticipates the question: “whom will the rich and dowried daughters marry?” Quo illae nubent divites dotatae?; the opposition between uxorem ducere and nubere is intentional. Otherwise the verb is mainly poetical. Commonly used are only the participle nupta and the phrase nuptum dare ‘give (one’s daughter) in marriage’, that is to say, those verbal forms which make the woman the object and not the subject. Nor can we apply the Latin verb maritare to the function of the woman. Even at the late date at which it appears, maritare as active verb signifies “to match, to join”, and as an intransitive verb it is more often used of the man than of the woman.
This negative lexical situation, the absence of a special verb, indicates that the woman does not “marry”, she “is married.” She does not accomplish an act, she changes her condition. Now this is precisely what is shown, and this time in a positive way, by the terms which denote the change of status of the married woman. These are exclusively nominal forms which appear at two extreme points of the Indo-European territory, in Indo-Iranian and Latin.
These terms are used in a phrase which formally declares that the woman enters into the “stage of wifehood.” We have in Vedic two abstract nouns of very similar form, janitva– and janitvaná– ‘state of the married woman (jani-)’, both in a formulaic context: hastagrābhásya didhiṣós távedám pátyus janitvám abhi sám babhūtha ‘you have entered into this marriage (janitvám) with a husband who takes you by the hand and desires you’ (to the widow, Rig Veda X, 18, 8); janitvanā̂ya māmahe ‘he has offered (two young women) for marriage’ (VIII, 2, 42). We see in the first passage the connection between the set terms janitvam on the one hand and hastagrābhásya patyus on the other, the husband who, with a ritual gesture, takes his young wife by the hand; in the second, that janitvaná indicates the destination of the woman given to her husband in the forms required “to become a wife.” An equivalent to janitvá– is the symmetrical term patitvá-, patitvaná– ‘state of husband’ (X, 40, 9) when this designates the power to which the woman is submitted, thus patitvám … jagmúṣī ‘(the young woman) who has come under the power of the husband’ (I, 119, 5).
It is interesting to note a parallel fact in Old Iranian, where the same notion is expressed in an abstract derivative furnished with the same suffix, Avestan nāiriθwana-. The stem is here nāiri– = Vedic nāri– ‘woman, wife’, an Indo-Iranian feminine, which makes a pair with nar– in the traditional formulae: Ved. nŕ̥bhyo ná̄ribhyas (I, 43, 6 ; VIII, 77, 8) = Avestan nərəbyasča nāiribyasča (Y. 54, 1). In Avestan nāiriθwana formed, like Vedic janitvaná-, has exactly the same sense “state of wifehood”, and it also appears in a formulaic passage: xvaŋha va duγδa va … nərəbyō ašavabyō nāiriθwanāi upavādayaēta ‘a sister or woman might be led into marriage to pious men (Vd. XIV, 15); this attests a legal expression where nāiriθwanāi vādaya– ‘to lead (a young woman) into marriage’ appears with a verb vad(aya)-, the technical value of which was seen above.
To sum up, the term which we translate by “marriage,” Ved. janitvana, Avestan nāiriθwana– is only valid for the woman and signifies the accession of a young woman to the state of legal wifehood.
This justifies us in regarding it as a trait of great antiquity, bound up with the structure of the Indo-European “Grossfamilie,” because it recurs in Roman society. The Latin term matrimonium is of great significance in this respect. Taken literally, matrimonium signifies “legal status of the mater,” in conformity with the sense of derivatives in –monium, which are all legal terms (testimonium, vadimonium, mercimonium, and naturally, patrimonium). The reason for the creation of matrimonium is not the analogy of patrimonium, which conveys a quite different notion. It emerges from set expressions from which matrimonium gets its full sense, that is from the point of view of the father: dare filiam in matrimonium; from the husband’s point of view, alicuius filiam ducere in matrimonium; and finally from the point of view of the woman: ire in matrimonium. Thus matrimonium defines the condition to which the young woman accedes; that of mater(familias). This is what marriage means for her, not an act but a destination: she is “given and led,” with a view to matrimonium, in matrimonium, just as the similar terms of Indo-Iranian janitvaná-, nāiriθwana– figure in our formulae in the form of a dative of intention, designating the state to which the wife is intended. From this comes later matrimonia in the sense of “married women,” like servitia ‘slave women’.
The modern forms of matrimonium in the Romance languages, particularly in Italian matrimonio, have acquired the general sense of “marriage.” Better still, the derivative matrimonial functions today in French as a corresponding adjective to mariage, for instance in régime matrimonial, so that we might easily take matrimonial as the Latin derivative of mariage, like oculaire from oeil, or paternel from père. This, it must be stressed, would be pure illusion: mariage, a normal derivative from marier (Lat. maritare) has nothing in common with matrimonium. But the fact that the two are associated so closely as to seem related shows how far we have travelled from the ancient values.
We see here a type of Indo-European correspondence which is not once treated by traditional comparative grammar. The present analysis reveals the unity of terms which are etymologically diverse but are brought together by their content and constitute parallel series. The nominal forms which finally come to designate “marriage” all denoted at first the condition of the woman who became a wife. It was necessary for this specific sense to be blurred to enable the abstract concept of “marriage” to take shape, so that the end result was a designation for the legal union of man and woman.
[ back ] 1. This chapter has been already published in the Festschrift A Pedro Bosch-Gimpera, Mexico City, 1963, pp. 49ff.
[ back ] 2. These lexical developments have been analyzed in detail in our study Hittite et indo-europeén (Paris, 1962, pp. 33 ff).