The Economy in Indo-European Language and Society

Indo-European Migration / Wikimedia Commons

By Dr. Émile Benveniste (1902-1976)
Former Professor of Linguistics
Collège de France

Male and Sire


Contrary to traditional etymologies we have to distinguish between two ideas on the Indo-European level: (1) on the physical side that of the “male,” i.e. *ers-, and (2) on the functional side that of the “sire,” i.e. *wers-. A semantic rapprochement between these two roots is found only in Sanskrit and may be regarded as secondary.


Bantengs, a wild cattle, forage in the jungle in Mondulkiri province in December 2010 / Creative Commons

We shall first consider some typical expressions relating to stock breeding. The object of study will be the differentiations characteristic of special techniques: on the lexical level, as elsewhere in linguistics, the differences are instructive, whether they are immediately apparent or come to light only after the analysis of a unitary group. An obvious and necessary distinction in a society of stock breeders is that between males and females. This is expressed in the vocabulary by words which can be regarded as common, since they appear in several languages, though not always with the same applications.

For the first word which we are going to study we have a series of correspondences which are relatively stable, although they admit of variations. They concern the word for “male”:

r̥ṣabha arəšan
Skt. Av. Gr. ársēn, árrēn
vr̥ṣabha *varəšan

We postulate for Avestan a word which happens not to be attested but which is implied by its derivatives, i.e. Av. varəšna– ‘masculine’, varəšni– ‘male’, ‘ram’.

In Greek, again, we find slightly deviant forms in the group e(w)érsē ((w)έρση), hérsai (ἕρσαι) (cf. the form with ν in Indo-Iranian); the meaning is (1) “rain, dew” (in the singular), whereas (2) the plural is applied to animals. To this family belongs Lat. verrēs, the male of a particular species, with its corresponding forms in Baltic, Lit. ver̃šis, Lett. versis. All these derive from the verbal root *wers– exemplified in the Skt. varṣati, which means in the impersonal “it rains” (cf. eérsē); we may also adduce Irl. frass ‘rain’ < *wr̥stā.

There is a morphological difference between the last forms and the preceding nominal forms, but this has not prevented etymologists from grouping them together. But this should give us pause: we have on the one hand forms with and without an initial w in Indo-Iranian. Similarly in Greek, whereas árrēn (ἄρρην) never has a w, Homeric metre implies that eérsē = ewérsē, which develops to hérsai.

Comparatists have interpreted this disagreement as an alternation. But since there are no compelling reasons to follow them, we should practice the utmost economy in the use of hypothetical “alternations.”

In Indo-European morphology there is no principle which would permit us to associate forms without w– with those containing a w-. To postulate a unified group here is gratuitous; there is no other example of this alternation w-/zero. As for the meaning of the words thus associated, where an analysis is possible, it will be seen that there are difficulties in bringing the words together.

In Sanskrit, vr̥ṣabha– and r̥ṣabha– attest the same manner of formation and the same notion. This is that of the “mythological bull” and “the male in general,” the epithet of gods and heroes alike. In Avestan, on the other hand, the two words (with or without w) have divergent meanings, and this disaccord is instructive outside Indo-Iranian: in Iranian arəšan and *varəšan are absolutely separate words. Arəšan in the Avestan texts is always opposed to a word which designates the female, this being sometimes xšaθrī (a purely Iranian term), but usually daēnu. This latter expression, which is Indo-Iranian (cf. Skt. dhenu), belongs to the group of Greek thē̂lus (cf. the Sanskrit root dhay– ‘suckle, nourish’). Thus we have here a specific designation, a functional one, for the female animal.

The opposition of arəšan– : daēnu– is constant. In the lists of animals we find the two series of terms enumerated in the same order:

“horse” | aspaarəšan– | aspadaēnu

“camel” | uštraarəšan– | uštradaēnu

“bovine” | gauarəšan– | gaudaēnu

The Avestan arəšan never designates any particular species, as does the Sanskrit r̥ṣabha which, without being the exclusive word for bull, frequently has this meaning. This is quite different from arəšan; it simply denotes the male as opposed to the female.

This opposition male/female may appear in a slightly different lexical guise in Avestan. For human beings, nar/xšaθrī are used, where the latter term looks like the feminine form of the adjective meaning ‘royal’, that is, ‘queen’. This may appear somewhat strange, but it is not inconceivable if we think of the correspondence between Greek gunḗ ‘woman’ and English queen. There are some slight variants such as nar/strī, where the second term is the Indo-Iranian name for “woman,” cf. in the compounds strīnāman (cf. Lat. nōmen) ‘of female sex’, while xšaθrī is sometimes transferred to the animal world. All this is quite clear; the opposition is unambiguous. Outside Iranian, arəšan has an exact equivalent in the Greek ársēnárrēn with precisely the same sense as in Avestan: it denotes the male as opposed to the female, árrēn contrasts with thē̂lus. The etymological identity of the two terms argues an Indo-European origin.

Let us now consider the Avestan word *varəšan. It expresses a different notion, that of the sire. It is not the characteristic of a special class of beings, but an epithet of functional value. *Varəšan (the actual form is varəšni-) is used with the name for sheep to designate the “ram”: maēšavarəšni-. This combination leaves no doubt as to its meaning. Apart from this, there is also historical testimony: *varəšan, by regular sound development, yielded Persian gušan, and this signifies not the “male” (represented in Persian by a form derived from nar) but the “sire.”

Outside Iranian, Latin verrēs is the exact counterpart in form and meaning. It does not denote the “male,” the male pig being called sūs (a word to which we shall return later) but the “sire.” Verrēs, ‘boar’, is used in exactly the same way as the corresponding Avestan form *varəšan.

What conclusion can we draw from these observations? *Ers– and *wers-, which were regarded as identical, are two different forms, absolutely distinct both in meaning and morphology. Here we have two words which rhyme, which may be superimposed, but which in reality belong to two independent families. One designates the “male” as opposed to the “female”; the other denotes a function, that of the “sire” of a flock or herd and not a species, like the first. It is only in Sanskrit that there was a close rapprochement between r̥ṣabha– and vr̥ṣabha-. Because of a mythology in which the bull has a prominent place and in virtue of a style in which high-flown epithets abound, the two terms became so far assimilated that the first assumed a suffix which belongs properly only to the second.

Such is our first conclusion. It can be given further precision by recourse to a distinct lexical development. There is probably some connection between Greek eérsē and hérsai. How can this be defined? The singular eérsē denotes the light rain of the morning, dew. Apart from this we have the Homeric plural form hérsai, which is only attested once (Od. 9,222): in the cave of Polyphemus there is a sheep-fold in which the animals are arranged in age groups, from the adults to the very youngest—the hérsai. Now, hérsai is the plural of eérsē. To understand this peculiar association, we can adduce some parallels in Greek: drósos means “dew drop,” but in Aeschylus drósos in the plural denotes young animals. There is a third example of the same kind: psakás, which means “fine rain,” has a derivative psákalon, ‘the newly-born of an animal’. This lexical relationship may be explained as follows. The tiny newly-born animals are like dew, the fresh little drops which have just fallen. Such a development of meaning, peculiar to Greek, would probably not have taken place if *wers– had first been the name of an animal, considered as the “male.” It seems therefore now to be established that we must posit for Indo-European a distinction between the two different notions and two series of terms. It was only in Indic that a rapprochement was effected with the result that they became similar in form. Everywhere else we find two distinct lexical items: one, *ers-, designating the male, (e.g. Greek árrēn), and the other *wers– in which the original notion of rain as a fertilizing liquid was transformed into that of “sire.”

A Lexical Opposition in Need of Revision: sūs and porcus


It is usually held that: 1) IE *porko– (Latin porcus) denotes the domestic pig as opposed to the wild animal, *– (Lat. sūs); 2) The dialect distribution of *porko– leads to the conclusion that only the European tribes practiced pig-breeding.

However, a careful examination shows 1) that in all languages, and particularly in Latin, where the opposition *– : *porko– was maintained, both these terms applied to the domesticated species, *porko– designating the piglet as opposed to the adult *-; 2) that *porko– is in fact also attested in the oriental part of the Indo-European world. Consequently pig-breeding must be attributed to the Indo-Europeans, but it was eliminated at an early date in India and Iran.


Wild boar / Photo by Valentin Panzirsch, Wikimedia Commons

The Latin term verrēs forms part of a group of words which refer to a particular species of animals, the pig. An attempt will be made to define the relations between the terms in this series of animal words in Latin, i.e. verrēssūs, and porcus.

Sūs and porcus have equal claim to Indo-European status, since both have correspondents in the majority of Indo-European languages. What is the relation between their senses? The distinction is generally held to be between the wild and the domesticated animal: sūs meaning the pig-species in general in its wild form, the wild boar, while porcus denoted exclusively the domesticated animal.

This is supposed to be a very important distinction from the point of view of the material culture of the Indo-Europeans, because whereas sūs is common to all dialects from Indo-Iranian to Irish, porcus is restricted to the European area of Indo-European and does not occur in Indo-Iranian. This difference suggests that Indo-Europeans were not acquainted with the domestic pig and that domestication did not take place until after the unity of the Indo-European people had been disrupted and some tribes had established themselves in Europe.

Today we might wonder how it came about that this interpretation was regarded as self-evident so that scholars came to believe that the difference between sūs and porcus reflected a distinction between the wild and the domesticated pig. Let us scrutinize those Latin authors who wrote on agricultural themes—Cato, Varro, Columella—and who used the language of the countryside. For them, sūs denoted both the domestic and the wild animal. Sūs is certainly used with reference to the wild pig, but the same word in Varro is always applied to the domestic species: the minores pecudes, the small animals, comprise ovis ‘sheep’, capra‘goat’, sūs ‘pig’, and they are all domestic animals.

A further proof is found in the term suovetaurilia, which designated the great sacrifice of the triple lustration, in which three symbolic animals figure. This technical term combines two species (ovistaurus), which were certainly domesticated, with sūs, and this presumably indicates that this was likewise a domesticated animal. This conclusion is confirmed by the fact that in Rome no wild animals were ever sacrificed.

Similarly in Greek, there is an abundance of examples in which hū̂s (ὗς = Lat. sūs) applies to the domestic animal. Certainly a distinction was made between the wild and domestic species, but only by means of an added epithet. The wild pig is called hū̂s ágrios as contrasted with the domesticated animal. We must conclude that it was in prehistoric times before the emergence of Latin that Indo-European *– = Greek hū̂s became applied to the useful species, i.e. the domesticated one.

In the other Indo-European dialects, the word is used in a different way. In Indo-Iranian – denoted the wild pig. The historic forms Sanskrit sūkara, Avestan – are formed from an identical stem. According to Bloomfield, one must begin with sūka-, this being an ancient stem which received a suffix –ra, attached on the model of other animal names, such as vyaghra ‘tiger’. Sūka-ra was analyzed as +kara ‘the animal which makes ’ by a kind of popular etymology. Besides Av. , a form xūk is met in Iranian, and this presupposes *hūkka. Thus Indo-Iranian had a form with a suffix –k which, over the domain of Indic and Avestan, referred only to the wild species. The reason is that neither in India nor in Persia were pigs bred in ancient times. There is no mention of pig breeding in our texts. Yet against this, from the evidence of Latin, we have seen that in the European sector the domestication of the pig took place well before Latin was constituted, the generic name being already employed for the domesticated animal. It is this sense of “domesticated pig” which is almost exclusively used in Latin. Sūs refers to the wild boar only in those contexts where the generic term suffices.

In studying the meaning of words which are peculiar to Latin with reference to the pig, a problem emerges: a minor one at first sight, but with consequences which turn out to be of considerable importance. Since sūs designates the species in general and more especially the domestic species, the distinction usually drawn disappears. Since both words refer to the domestic pig, sūs and porcus become synonymous. This pleonasm is surprising and provokes closer examination of the testimony by which the meaning of porcus is established (and not the translations, which are unanimous on this point).

We may begin with one of the terms in which the name of the animal appears in a stock expression, suovetaurilia, an expression already quoted, which implies the sacral combination of three animals sacrificed on the occasion of a lustration ceremony. The expression suovetaurilia is said to be irregular in a number of ways. We have

  1. a compound containing a group of three terms; but similar compounds are attested in the Indo-European languages, cf. Gr. nukhth-ḗmeron, ‘a night and a day’. Thus the objection is invalid.
  2. a phonetic difficulty, because the form is ove instead of ovi. This can be resolved if we give an exact determination of its signification and site it in the conditions in which it was constituted. It is no ordinary compound word, but a juxtaposition comprising not nominal stems, but case forms. It is a series of three ablatives: *, the ancient ablative of sūs (cf. sūbus, the ancient plural form); ove, a regular ablative, and finally taurō. There are thus three ablatives in juxtaposition and the whole being treated as a single word with attachment of the adjectival suffix –ilis, –ilia added to the last word with elision of the case ending. Why this juxtaposition? Because it is taken from the ritual expression in which the name of the sacrificial animal is in the ablative:  facere ‘to sacrifice by means of an animal’ and not the animal itself. Facere + the ablative is certainly the ancient construction. Therefore it meant to perform the cult act by means of these three animals, an ancient sacrificial grouping of these three species, where sūs is the name for the porcine species. We must reread a chapter of the De Agricultura by Cato (141), the famous text which describes the way in which the lustration of the fields, a ceremony of a private nature, was carried out. In this text, which has often been read, quoted and used, we are expressly concerned with the suovetaurilia. In proceeding to the sacrifice, the owner of the field must pronounce these words: macte suovetaurilibus lactentibus esto. This is a prayer of Mars that he should accept these suovetaurilia lactentia, three “suckling” animals, that is, young ones. This prayer is repeated a second time in these terms: Mars pater, eiusdem rei ergomacte hisce suovetaurilibus lactentibus esto. Cato continues: “when you sacrifice the porcus, the agnus, the vitulus, you must…” (ubi porcum immolabisagnum vitulumqueoportet…) The sacrifices in fact comprise three animals which this time are called porcusagnusvitulus. Let us compare the terms of the nominal sacrifice sūsovistaurus with that of the actual offering, porcusagnusvitulus. These expressions follow each other in exactly the same order and they indicate the sacrificial animals. It follows that vitulus is the young of the taurusagnus the young of the ovis, and porcus the young of the sūs. This is deduced in quasi-mathematical manner by superimposing the ritual expressions on the actual species of the sacrifice. The conclusion is inescapable that porcus can only mean piglet. The difference between sūs and porcus is not between the wild animal and the domesticated one: it is a difference in age, sūs being the adult and porcus the young animal.

We have another text which makes this point. In the De re rustica of Varro (Book II, ch. 1) the author gives advice to breeders on the raising of animals. Some months must elapse before the young animals are weaned: the agni at four months, the haediat three months, and the porci at two months. Thus porcus is paralleled with agnus and haedus. There are so many examples of this kind that the greater part of the chapter could be quoted. Varro makes the point that one can tell sues of good stock a progeniesi multos porcos pariunt, ‘if they produce plenty of porci’. As to feeding, it is the custom to leave the porci two months cum matribus. A little further on we read: porci qui nati hieme fiunt exiles propter frigora, ‘the porci born in the winter…’ Here the association of porcus and mater speaks for itself.

Piglets / Creative Commons

In an archaic expression of the religious vocabulary, the porci which are ten days old habentur puri ‘are considered pure’, and for this reason they are called “sacres” (the ancient form, instead of sacri, from the adjective *sacris); sacres porci, a very old expression, “the pigs which are ten days old.” Similarly, lactens porcus appears frequently, but we never encounter *lactens sūs. A diminutive porculus or porcellus exists, just as one finds agnus/agnellusvitus/vitellus; but there is no word *sūculus, since the name for the adult animal does not admit a diminutive. Thus the meaning of porcus, which is found perhaps forty times in this text, is constant. The meaning does not vary in later usage. Cicero uses it in the same sense: with reference to a villa (‘estate’) he writes: “abundat porcohaedoagno,” an expression where porci figure along with the other young animals, haedi and agni, kids and lambs. We know two words for swine-herd: sūbulcus ‘he who occupies himself with sues’ (parallel with būbulcus) and porculator. What reason was there to coin two separate expressions if the two words sūs and porcus had the same meaning? In fact the porculator looks after the young pigs (piglets), which need special treatment, while the sūbulcus looks after the adult pigs. We have thus established that throughout ancient Latin down to the classical period porcus designated only the piglet. The difference is now clear. What is astonishing is that this fact was not seen earlier and that an erroneous translation of such a common term as porcus has endured for so long. The relation of sūs to porcus is exactly the same as that of Greek hû̄ssūs(ὗςσῦς) to khoîros (χοῖρος). This difference is of great importance. In public and private cult there were no animals more commonly offered than the porcus, the young pig.

The Romans already knew what we have just discovered. Varro gives us, with a fanciful etymology, precisely the equivalents in the two languages: R.R. II, 1: “porcus graecum est nomenquod nunc eum vocant khoîron .” He thus knew that porcus meant the same as khoîros. But porcus exists not only in Latin; it is also found in Italic. The contrast between si and purka is the same in Umbrian in a ritual text where both figure. We must see what this opposition signifies in Umbrian.

The translation of the Iguvine Tables is usually expressed in Latin so that it is not particularly lucid. But we must consider the adjectives which accompany si and porkoSi is found with kumia, translated as ‘gravida’, and also with filiu, translated as ‘lactens’; and on the other hand there is purka. Now the combination of lactens with sūs is impossible in Latin and the difference in Umbrian becomes incomprehensible. If the Umbrian word si can signify an animal which may be gravida ‘pregnant’ as well as lactens ‘suckling’, what can porko mean? If the same word applies both to the adult and to the newly born animal, the difference of designation is no longer justifiable, and the other word purka becomes redundant.

In a ritual text of such precision, why is there this difference, in one place si and in another purka? The crux of the problem lies in the meaning of filiu. There is another possibility than that of the traditional translation. Two interpretations of filiu are possible: one as lactens ‘suckling’, but it is also possible to think of lactans ‘in milk’ (“she who suckles”). In fact the Umbrian filiu is related to Greek thê̄lus and to fēmina, which in Latin means “she who suckles,” also the meaning of Greek thê̄lus. In Irish and Lithuanian a form with the suffix l made from the same root *dhē– is used with reference to the mother: Lith. pirm-delú ‘animal which suckles for the first time’. Thus we may take the Umbrian filiu not as ‘lactens’ but as ‘lactans’. The sow is sometimes spoken of as “gravida” and sometimes as “lactans,” according to whether the animal is still pregnant or has already farrowed. It follows that purka is the term for the young pig; it is the piglet, just like the Latin porcus, and the situation which at first was quite incomprehensible becomes intelligible. We may thus be assured that the difference illustrated by both Latin and Umbrian is an inherited lexical difference. It is in fact prior to Italic.

In Celtic, the corresponding word for porcus, phonetically Irl. orc, is always cited with the group of porcus and given the translation “pig.” But the precision which we expect is given by the detailed dictionary of the Irish Academy, which translates orcas “young pig.” Thus we see that both Italic and Celtic show solidarity in offering one and the same meaning.

In Germanic, the two corresponding words are represented by derivatives, on the one hand by swein (German Schwein) and on the other by farhfarhili (German Ferkel). Here the modern forms already indicate the distinction: Ferkel is the piglet, specifically a diminutive form, whereas swein ‘pig’, derived from -, does not show a diminutive formation. The Germanic word corresponding to porcus immediately attests the sense of “young pig” which it has preserved. Finally, in Slavic and Baltic, the Lithuanian paršas, Sl. praęs (from which comes the Russian porosënok, which is a diminutive) is opposed to svin. Now the Slavic and Baltic *parša– corresponding to porcus has the sense of “piglet.” We thus have the same contrast in Slavic as in Germanic. This demonstration could have been pursued from two different angles, but whether we start from Germanic or Slavic, the same conclusion is reached as emerged from an unprejudiced study of the Latin evidence. At all events the testimony is consistent and the lexical situation seems identical in all western dialects of Indo-European.

It is, however, on the Indo-European level that the contrast between the two terms poses a new problem. The distribution of the two forms is unequal. The form *– is common Indo-European. It is attested in Indo-Iranian as well as all the strictly “European” dialects, whereas *porko– does not appear in Indo-Iranian but only in the European languages.

From this dialect distribution and from the meaning attributed to *– and *porko– the conclusion has been drawn that the Indo-European community was not acquainted with the pig except as a wild species. The very meaning of porcus, so it was believed, implied that the domestication had only begun in Europe after the settlement of certain ethnic groups.

But the restored signification of these terms transforms the problem. It assumes a new aspect, since the opposition is adult/newly-born and not wild/domesticated. Why then is the name for the newly-born animal (*porko-) not co-extensive with that of the adult (*-)? But is there in fact this unequal distribution of sūs and porcus? The whole chain of reasoning rests on the allegation that no trace of porcus has been found in Indo-Iranian territory. In fact, the problem has been much advanced and today the traditional affirmation must be challenged.

The same word *porkos is attested in an area that is geographically adjoining but linguistically quite different, in Finno-Ugrian, by the Finnish word porsas, Mordvinian purts and Zyrenian porś. Scholars are agreed in regarding this as a common borrowing by the Finno-Ugrian languages from a form in –s at some stage of Indo-European, but at what date did this word penetrate into Finno-Ugrian?

We may begin by noting that the meaning is certain: “piglet, small pig” in Finnish; in the other languages, the lexica are less precise, but this meaning is probable. The connection with Indo-European forms has been noted and the possible chronology of the borrowing has been discussed. What seems certain is that porsas in Finnish presupposes a stem in –o; the final –as is a Finnish adaptation of a stem in –o, replaced by a, because, from ancestral Finno-Ugrian times, ο was not permitted in the second syllable: *porso becomes porsa. The root *porso exhibits a characteristic palatalization of the k into s. The original form borrowed into Finno-Ugrian was marked by this palatalization before the change of the root ο into a, which is characteristic of Indo-Iranian. The theoretical Indo-Iranian form is *parśa, and this would appear in Indic as *parśa and in Iranian as *parsa. The phonetic shape of the Finno-Ugrian borrowing takes us back to the stage prior to Indo-Iranian, but posterior to the common Indo-European, where the word possessed a-k-. It was therefore an ancient dialect form which would precede the separation of Indo-Iranian. This is the conclusion reached by Finno-Ugrian specialists. But one difficulty has given them pause: the pre-Indo-Iranian form implied by the borrowing is not attested in Indo-Iranian. They have therefore hesitated to draw any firm conclusion.

Khotanese verses / Wikimedia Commons

But now we have the form in the oriental region. A Middle Iranian dialect of Eastern Iranian, called Khotanese, the knowledge of which goes back only a few years, has yielded evidence for the existence and the meaning of a word pasa, gen. pasä, which designates the pig. The meaning is certain because the texts are translated from Sanskrit or Tibetan, in which there occur expressions for dates borrowed from the animal cycle: there is a year or month of “the pig.” Thus Khotanese has restored to us the expected Indo-Iranian form: parśa, and it furnishes the proof that *porko– was also known in Indo-Iranian territory.

The negative argument can thus no longer be sustained. True, there is no trace of *porko– in Indic, but a word of this kind is exposed to accidents. There are peoples who, for religious reasons, exclude this animal from sacrifice and consumption, whereas it was esteemed by the peoples of Europe. We now know that the word did exist in Iranian. There is now no difficulty in admitting in principle that the Indo-European stem *porko– was common to all dialects. We have established its presence in Eastern Iranian and confirmation has been given by the Finno-Ugrian borrowings.

True, we are not yet able to define the exact meaning of the term in Khotanese, a language not attested until the seventh or eighth century of our era. But since *– is common to Indo-Iranian and the European languages, if *porko– was also used in Iranian, it must have been distinct from the word *-. The features which are presumed or are established indirectly accord with those taken from textual usage.

All this, namely the existence of two words employed since the European period, and the difference of sense which we have underlined, allows us to state that the common Indo-European word *porko– meant “the young pig.” The negative conclusion of traditional doctrine can no longer be upheld: there was after all Indo-European domestication of the pig. This is what the vocabulary reveals by the distinction made between sūs and porcus, which runs parallel with that encountered in the names for the other domestic animals.

Another point may be made, namely that the lexical distinction made between sūs and porcus may later be expressed by different terms. The opposition sūs : porcus persists throughout the whole of Latinity until after the classical period; but later the proper sense of sūs was transferred to porcus, which took over the function of sūs. At that moment sūs disappeared.

In the Glosses of Reichenau, which give such precious testimony for the transition of Latin to French, the term sūs is glossed as ‘porcus salvaticus’ (= wild pig). Thus sūs had been limited to the meaning of “wild pig” and porcus had taken its place as the name for “pig.” But it was necessary to coin a term to replace porcus in its original sense: hence porcellus, French pourceau.

Later, under the influence of the language of the Gospels, where porcellus means “pig,” recourse was had to a technical term for the young animal, French goret. There is an innovation in the expression for the distinction, but the distinction is preserved, for it is important to maintain a distinction which is anchored in an extra-linguistic reality—animal husbandry.

Próbaton and the Homeric Economy


It has been maintained that the term próbaton, created by the Greeks, meant small animals, especially the “sheep,” since in a mixed flock the sheep tend to walk in front (pro-baínein).

It will be shown that this thesis is untenable; 1) próbaton, to begin with, designated the large as well as the small animals. 2) the Greeks had no mixed flocks. 3) probaínein does not mean “walk in front.”

In fact, próbaton, a singular of próbata, is to be connected with próbasis ‘(movable) wealth’. It was because the sheep constituted “movable wealth” par excellence as opposed to possession which were stored in chests (keimḗlia), that it was called “próbaton.”


Flock of sheep / Wikimedia Commons

We have just considered a problem which is raised by the coexistence of several terms which appear to have the same meaning within one and the same language or in a number of Indo-European languages.

An analogous situation is present in Greek, where we also find two terms for the name of another species, the sheep: ówis(ὄwις) and próbaton (πρόβατον). The two terms both designate the sheep from the time of the earliest texts.

The first is an ancient word of the common vocabulary, exactly preserved in Greek, Latin and Sanskrit, which is now attested in Luvian in the form of hawi-. The second is confined to Greek and the form itself gives grounds for believing that it is a relatively late creation.

In Homer, ówis and próbaton coexist, but subsequently ówis disappeared in favor of próbaton, which was the only one to survive until modern times. The problem which poses itself is why there should be two distinct terms. What was the meaning of the new term? As for the first, we can do no more than note that it was a common Indo-European word of the ancestral Indo-European vocabulary, and is not susceptible to further analysis.

As for the second word, próbaton, considered on its own without regard to its meaning, there is an evident connection with probaínō (προβαίνω) ‘to walk in advance’. But what exactly is this connection between “sheep” and “walking” and how can we interpret it? The explanation given by the comparative linguist Lommel [1] has won general acceptance: probaínō means “walk in front”; próbaton designates the small animals because they “walk in front”; but in front of what? In certain African countries, we are told, herds and flocks are formed by assembling animals of various species and it is the sheep which walk at the head. As a consequence of this próbaton would have designated the animals which walk at the head of a mixed herd of animals. This explanation, approved by Wackernagel, has achieved orthodoxy; for instance, it figures in Liddell and Scott’s lexicon.

It is the history of this term which we will now take up again to see whether, from a study of its usage, the development of its meaning in the course of an evolution which we can follow step by step confirms the proposed explanation.

It must be noted at the outset that the form próbaton is not the most common one. The first examples are in the plural, tà próbata, and the singular is unknown at an early date. Only the plural is used in Homer and Herodotus. Especially in Herodotus, thirty-one examples of the plural are found but not a single one of the singular. In the Homeric poems, if one animal is referred to, it is óïs which is used and never próbaton; in fact, the only Homeric form is próbata—and this is not merely a morphological detail. We should not speak of a plural but rather of a collective:  próbata. It follows that the form próbaton is what is called a singulative; we may compare the relationship tálanta to tálanton and dákrua to dákruon. The generic names for animals are more frequently collectives, e.g. tà zô̄a, which occurs earlier than tò zō̂on.

A new term of Greek coinage which has persisted down to modern times is tò álogon which, early in our era, occurs with the specialized meaning of “horse” in the papyri. We must regard tò álogon as the singulative of tà áloga ‘the beasts’, those “deprived of reason,” the term given to the most common or most useful of animals, that is, the horse. Similarly, in Latin, animalia is older than animal. This is a very common type of designation: a large proportion of animal names are collectives.

It remains to give precision to the morphological relationships between  próbata and probaínō. At first sight próbaton or próbata seems to be a compound form in –batos, this being a verbal adjective derived from baínō. But if this were so, it would not have its normal meaning: for instance, ábatosdúsbatosdiábatos all have a passive sense, that is to say, “that which is crossed,” with a restriction of sense indicated by the first member of the composite word, or rather “that which can be crossed.” The passive voice is also apparent in the simple adjective batós (βατός) ‘accessible’. A different meaning appears in the composites like hupsíbatos, where –batos has an active meaning (“one who has climbed high, has gone on high”).

But neither the active nor the passive sense fits the suggested interpretation of próbaton, in which the second element functions as a present participle “which walks.” The fact is that the ancient grammarians make a distinction between próbaton and the adjectives in -batos: according to them, the plural dative of próbaton is próbasi (πρόβασι). Here we have a consonantal stem: probat– (προ-βατ). This is the only form which explains the dative and it is this which must be postulated. Such a form can be justified from a morphological point of view because there are root forms suffixed by –t– (cf. Skr. –jit-, kṣit-) which Greek adapted to a suffixal type and to an inflectional category which was more familiar. Compared with the Sanskrit parikṣit-, we have Greek periktitai (Od. 11, 288); cf. Lat. sacerdōt-. Where the Greek had –thet– this became normalized as –thétēs, this being one of the processes for converting archaic and aberrant forms to a more normal type. An analogous phenomenon, though by a different process, is seen in the case of próbaton: here recourse was had to thematization (facilitated by probatá) to normalize the original form in –bat– which is implicit in the dative plural próbasi and also in the present participial function of the word.

Now that we have considered the morphology with greater precision, we may turn to the problem of meaning. As we have seen, according to Lommel, próbata designated small animals, the sheep, so named because “they walk at the head” of the herd. What is thus essential to Lommel’s thesis is that próbata designated the “small animals.” But is this really the use of the word? Far from it! We have at our disposal many examples in the literary texts and in ancient dialect epigraphy.

First in Homer, Il. 23, 550: “you have in your house much gold, bronze, próbata and servants.” What does próbata mean here? Evidently “animals” in general, since no species is mentioned. Herodotus writes τὰ λεπτὰ τῶν προβάτων to specify “the small animals,” which would be absurd if próbata already meant “small animals.” Consequently what is meant are animals as such without any specification as to kind of size. After scrutiny of all the examples in Herodotus we can affirm that it is applied to live-stock, large or small. In Hippocrates, who wrote in the ancient Ionian dialect and whose vocabulary is of great interest, we find a clear opposition between próbata and ánthrōpoi, live-stock and men.

Next comes a decisive fact from an Arcadian inscription relating to Athene Alea at Tegea, τὸ μὲν μέζον πρόβατον…τὸ δὲ μεῖον ‘the large and small próbaton’, and there is another similar example with μεῖος and μεῖζων. All this is a clear indication that the word designates live-stock in general without further specification. It is possible to fix the moment when the sense became restricted to mean “small animals,” and it was in Attic that this semantic restriction took place.

There is no need to labor this point further: if próbata originally and everywhere designated “live-stock” in general, it becomes impossible to base the prehistory of the term on the sense “small live-stock,” this being a comparatively late development. A second point may be made: what warrant have we that in ancient Greece large mixed herds existed, at the head of which the sheep walked? This custom can be observed, we are told, in Africa. But was it pastoral custom precisely in Greece to assemble large herds of different animals?

We have no testimony about the composition of flocks, and all we have to do is to recall some familiar facts of Greek vocabulary. There is no single noun or a single compound for an assembly of animals. Greek uses different specific terms according to the kind of animals, with specific words for the respective herdsmen:

  • pô̄ü is exclusively a flock of sheep (shepherd = oiopólos)
  • agélē … a herd of cattle (cowherd = boukólos)
  • subósion … a herd of pigs (swineherd = subṓtēs)
  • aipólion … a herd of goats (goatherd = aipólos)

It should be noted that the name of the shepherd is based on ówis, not próbaton.

This distinction exists in other languages: in Latin, pecudes refers to the sheep (cf. pô̄ü), whereas amenta are “the large animals.” The English flock and herd may also be compared; indeed, English has a whole series of words for assemblages varying according to the species of animal.

If we only encounter special names for particular assemblages this must mean that mixed herds did not exist. Each species had its own special herdsman and was pastured separately.

Relief of a Herdsman and Oxen Roman 1st century CE Thasian Marble / Walters Art Museum

This is therefore a decisive objection to Lommel’s explanation. The practice of stock breeding was so old in Greece that long before the time of Homer there was a division of labor among the various specialized herdsmen. We find even in Mycenaean Greek a suqota, corresponding to Homeric subṓtēs and a qoukoro, who corresponds to boukólos. The name of the goatherd is also known in Mycenaean: aikipata. Thus there is nothing either in the traditional practice or in the vocabulary which would allow us to posit the existence of mixed herds: the second argument of Lommel falls to the ground.

However, there is still the etymological relationship between próbata and probaínō, which would seem to impose on próbata the meaning “those who walk at the head of.” But even for a verb of so transparent a form as probaínō we must not neglect verification. Now if one re-reads the examples, it emerges that probaínō never means “walk at the head of” even though all the dictionaries affirm it. We must scrutinize the sort of example from which this sense is deduced. The most frequent sense is in fact “to advance, progress, move forwards.” This sense is beyond all argument, for the examples are immediately apparent. In Homer (Il. 13, 18) κραιπνὰ ποσὶ προβιβάς ‘advancing with rapid steps’; Lysias (169, 38) προβεβηκῶς τῇ ἡλικίᾳ ‘of advanced age’. The meaning is thus invariably “to advance.”

But a second sense is posited: “to walk in front of somebody”—which is quite a different thing. This meaning is based on three examples from Homer, all of the same type: ὅ τε κράτεϊ προβεβήκῃ (Il. 16, 54) ‘who surpasses in might the others, who surpasses the others in power’, which has to be understood as “superior in might”; cf. Il. 6, 125; 23, 890. But it is the perfect tense which occurs in all these passages, and much confusion has arisen between the sense of the perfect and the sense of the verb: probaínō ‘I advance, I proceed forward’; thus the perfect probébēka means “I find myself in an advanced position,” e.g. Il. 10, 252 ἄστρα δὲ δὴ προβέβηκε, meaning “the night is advanced.” So an expression like προβέβηκε ἁπάντων or κράτεϊ means “someone is in an advanced position with respect to all” or “in respect to might.” In fact in Homer we find (Il. 6, 125) πολὺ προβέβηκας ἁπάντων, which means literally “you are far in advance of all.” It is because probaínō does not mean “to walk at the head of” but “to advance” that lexicographers have had to rely on these examples in the perfect in order to extract the sense of “to be in front of.” That sense does no more than illustrate the normal value of the perfect; as for the notion of superiority this simply results from the genitive-ablative, which indicates the point of departure from which an advanced position has been reached. Thus there is no difference in the meaning of the verb in the phrase ἄστρα προβέβηκε and in the three examples cited. The sense is one and the same, so that there is no need to subdivide into categories to distinguish between univocal examples. There is, however, a difference in Latin between progredior, which is the exact equivalent of probaínō, and praegredior ‘I walk ahead of the others’. But probaínō corresponds only to progredior.

Accordingly próbata does not mean “those who walk at the head of the herd.” One by one all the reasons which have been advanced in support of this explanation have crumbled: (1) próbata does not designate the small animals; (2) the Greeks did not keep mixed herds; (3) the meaning of probaínō is not “walk at the head of” but “proceed.”

What remains? Simply, a relationship between próbata and probaínō. To understand this relationship our starting point must be the meaning “advance, proceed”: próbata are those which advance, or proceed. But what then? The designation appears most peculiar and not a little puzzling. Is this a special attribute of live-stock or do not all animals “proceed” normally?

The solution is given in an expression morphologically related to próbata which we have not yet considered. It is the Homeric word próbasis (πρόβασις), an abstract word derived by the suffix –ti– from the same verb probaínō, which occurs only once in Homer, but in conditions which are ideal for our purpose. Od. 2, 75: keimḗliá te próbasín te. The Homeric expression denotes wealth: próbasis is a word in –sis of the class of abstract nouns capable of expressing collective meaning. This tendency is illustrated by such words as árosis which means “plowing” but also “arable land,” “corn-land” (cf. the French expression labourin “marcher dans les labours”); ktê̄sis ‘possession’ and also ‘the totality of ktḗmata’, just as árosis is the totality of ploughed land.

Thus próbasis indicates the totality of próbata, and the opposition keimḗlia/próbasis refers to possessions of two different categories, a distinction which seems to be essential in the economy of the Homeric world: Immovable or “lying” (keimḗlia from κεῖμαι ‘lie’), i.e. immovable property, and movable property (hósa probaínei).

This way of regarding property in its two categories has a rough resemblance to the French distinction between meubles(mobilia) and immeubles (immobilia). But immeubles are buildings, whereas meubles are chattels. In Homeric Greece the division was different: all that “lies” (keîtai), keimḗlia, precious metals in ingots, gold, copper and iron, is opposed to  próbata, property on the hoof, consisting of the herds and live-stock in general. Such is the sense of próbata as we have established from the textual evidence.

This explanation puts the economy of the Homeric world in a new perspective. Lommel conjured up an extraordinarily primitive type of herd composed of large numbers of animals. In fact próbata, connected as it is with próbasis, implies a much more developed social organization. In Homeric society wealth was a composite thing with a broad distinction on two different levels, between keimḗlia and próbata.

The same distinction was preserved until a much later age in Germanic. In the Scandinavian world we find a term which reminds us of próbata. This is the Icelandic gangandi , German gehendes Vieh (‘walking animals’); but here  represents pecus in the Germanic sense, that is to say “wealth.” Got. faihu translates argúrion ‘money’. The literal meaning of the expression is “wealth which moves” and this refers to live-stock (see below, Book One, Chapter Four). A further possible parallel, which we do not press, is offered by the Hittite iyant– ‘sheep’, for the word can be analyzed as the participle of the verb ai– (cf. Gr. eîmi) ‘go, walk’. It is however not yet certain that this is the word for sheep in general and not that of a particular variety. If the sense were confirmed, the parallel would be striking.

These are the essential facts. As for the rest of the semantic development, there is little point in illustrating ramifications of meaning represented by many examples in all languages at all periods.

The meaning to which the generic terms becomes restricted is determined by the most important species. This fact is universal and well attested, thus:

Lat. bestia > Fr. biche ‘hind, doe’
> Engandine becha ‘sheep’
Lat. animal > North Ital. dialects: nimal ‘pig’
> other regions: nemal ‘ox’

It is always the animal par excellence, the best represented species, the most useful locally which takes the generic name: Ital. pecora, ‘sheep’.

We may thus cite próbata among the groups of words subject to constant innovation. The special sense of próbaton derives from the local conditions of animal husbandry. The primary meaning, connected as it is with probaínō, cannot be interpreted except within the framework of a definite economic structure. [2]

pecu and pecunia


For all comparative philologists, Indo-European *peku means “live-stock” or, in a narrow sense, “sheep.” The meaning of “wealth” (e.g. Lat. pecūnia) is consequently regarded as secondary and this is explained as the result of a semantic extension of the term which originally referred to the main type of wealth, i.e. live-stock.

A study of *peku and its derivatives in the three great dialects where it is represented—Indo-Iranian, Italic and Germanic—leads to a reversal of the traditional interpretation: *peku originally meant “personal chattels, movables” and it was only as a result of successive specifications that it came to mean in certain languages “live-stock,” “smaller live-stock” and “sheep.” The evolution runs parallel with that of próbata (Book One, Chapter Three).

Cradle of Civilization blog / Creative Commons


In the vocabulary of the Indo-European economy, which was of a pastoral character, there is one term of central importance, *peku, attested in three great dialect regions: Indo-Iranian, Italic, and Germanic (Lithuanian pekus is most probably a loanword from Germanic or some other occidental language).

All comparative linguists are agreed in regarding *peku as the Indo-European name for “live-stock” and in deriving it from a root *pek– ‘to shear’. Thus, on this view, the term was applied to the sheep as the bearer of the fleece, and it was only as the result of a secondary development that the term came to be used for “live-stock” in general. Such is the explanation put forward in the early stages of comparative grammar.

An attempt will now be made to show that this conception of *peku is untenable and that a renewed examination of the evidence is necessary. The investigation will be concerned successively with Indo-Iranian, Latin and Germanic and will lead to conclusions which go beyond the problem under consideration.


Wikimedia Commons

The forms to be studied are Vedic paśu and Avestan pasu. In Vedic, the meaning is by and large “live-stock,” and this is confirmed by the various circumstances of its employment, its connection with vraja ‘cow-pen, fold, stall’, with gopā ‘shepherd’, with yūtha ‘flock’, etc. It must, however, be observed that:

  1. paśu is a collective term which covers the types of domestic animal (horses, cattle) and only those: aśvavantam gomantampaśum (Rig Veda I, 83, 4), paśum aśvyam gavyam (V, 61, 5), etc.;
  2. paśu even includes man, who is regarded as a biped paśu, on a par with the quadruped paśudvipáde cátus padeca paśáve(III, 62, 14). It is not only from this passage that this can be inferred, it is the explicit teaching of the Satapatha-Brahmana (VI, 2, 1, 2) on the five paśupuruṣam aśvam gām avim ajam ‘man, horse, ox, sheep, goat’. Other texts transpose this definition into a theory of sacrifice.

The inclusion of man among the paśu is indicative of a pastoral society in which movable wealth was composed of both men and animals, and in which the term paśu, which at first denoted movables, could stand both for bipeds and quadrupeds.

Iranian confirms this view. The association of men and animals, implicit in the Vedic definition, is expressed by the Avestic formula pasu vīra “livestock-men,” the antiquity of which has long been recognized.

What is the real meaning of vīra ‘man’ in the Avestic formula pasu vīra, which is echoed at the other end of Indo-European by the uiro pequo of the Iguvine tables? For Sanskrit, Lüders has shown that vīra, in a context where it is connected with livestock, means “slave.” This meaning, whether taken in a strict sense or merely as “house personnel, domestics,” is also valid for the Avestic vīra in pasu vīra.

We may adduce further confirmation taken from a gāthā of Zarathustra. In a strophe of a pathetic tone (Y. 46.2) Zarathustra complains of his impotence in overcoming the hostility which surrounds him on all sides: “I know why I am without power, Ο Mazda; it is because I am kamnafšu (=I have few pasu) and because I am kamnanar– (=I have few men).” The two qualifications kamnafšu ‘who has few pasu’ and kamnanar– ‘who has few men’ evidently come from the formula pasu vīra, with a replacement of vīra by nar-, which is also known in the Avesta. It is the fact that he has few pasu and few nar– that makes Zarathustra “powerless.” These possessions, which constitute the two species of movable wealth, together confer power. We may now add the Gathic expression kamnafšukamnanar– to the Avestan repertory of compounds based on the expression pasu vīra with their characteristic pairing.

The diversity of the linguistic evidence reflects the importance of pasu for the pastoral society of the northeast of Iran, the ideology of which has inspired the most ancient parts of the Avesta.

We shall restrict ourselves to the most ancient phase without following the later development of pasu, which is in any case well known. The ancient term has become today the name for “sheep” in one part of the Iranian world. A further specialization has thus followed on a much earlier one which conferred on pasu the meaning “livestock.”

All the same, it is in the sense of movable wealth that the Avestan vīra in pasu vīra has to be understood. This turn of phrase designates the totality of private movable possessions, whether human or animal, the men being sometimes included in paśu(pasu) but sometimes mentioned separately.

The same interpretation might be extended to uiro in Umbrian, not only because the formula uiro pequo comes from a common Indo-European heritage, but because of a specific indication peculiar to the two Italic peoples, the Umbrians and the Latins. Not enough attention has been paid to the striking similarity between the Umbrian formula and a passage in an ancient prayer by Cato. In Umbrian a certain ritual expression appears eleven times: uiro pequosalua seritu ‘salva servato’. Compare this with Cato: pastores pecuaque salva servassis. It suffices to superimpose the two texts:

  1. Umbrian uiro pequosalua seritu
  2. Latin pastores pecuaque salva servassis

to bring out the close correspondence of the two formulae. All the successive terms are etymologically related, except the first, where the same meaning is expressed by separate terms: it is precisely the Umbrian term uiro for which the Latin equivalent is not viros but pastores. From this we may conclude that Umbrian uiro, linked with pequo, designated the men whose task it was to look after the livestock. Thus we have in Umbrian an exact parallel to the notion of vīra associated with pasu in Indo-Iranian.

That pasu in the first instance had an economic sense can be confirmed from the term kṣu, which, although related to paśu– as Av. fšu– is to pasu-, became detached early on and kept the original sense better. The adjective purukṣu means “abounding in riches,” “in possessions,” but not specifically in livestock. This is an epithet of the gods Agni, Indra, Soma, and is often found associated with words meaning “wealth.”

All the indications point to the fact that the sense of “livestock” is a restriction of the more ancient comprehensive term “movable wealth,” applied as it was to the principal form of property in a pastoral society.


Creative Commons

The formation of pecūnia is unique in Latin. This gives it its value, but also its difficulty. It must be stressed that up to now the problem of its morphology has not been considered. The formal relation of pecūnia to pecū is that of a secondary derivative, which resulted in a lengthening of the final vowel of the stem. The essential question is that of the suffix. A parallel to the formation of Lat. pecūnia has been pointed out by Meillet among others: it is the O.Sl. –ynji (< *-unia). The suffix –ynji in Old Slavic makes abstract nouns from adjectives, e.g. dobrynji ‘goodness’ : dobrŭ ‘good’; or female names derived from corresponding male ones: bogynji ‘goddess’ : bogŭ ‘god’. We may even adduce a Slavic derivative in –ynji from a stem in *-u-: this is lĭgynji ‘lightening’ : lĭgŭkŭ ‘light’ (cf. Skr. laghú-, raghú– ‘light’).

This connection may be accepted, but we must draw certain conclusions. Since Latin pecūnia is an abstract noun, we have to posit an adjective as its basic form, just as with the Slavic abstract nouns in –ynji. We should then have to regard *peku as the neuter of a very archaic adjective which has not been preserved in any language. If this conclusion, inescapable as it is, seems too bold and if we think that it postulates a form the existence of which cannot otherwise be demonstrated, there still remains the alternative of explaining pecūnia from the resources of Latin.

We can link pecūnia with feminine derivatives in –nus, –– which are formed from nouns in –u-: thus fortūna, which is derived from *fortu– (cf. fortu-itus), or Portūnusopportūnus from portu-. We should then have to admit (1) that the correspondence between Latin pecūnia and the Slavic form in –ynji is only apparent and is due to a secondary process, and (2) that pecūnia is an abstract in –ia formed in Latin itself from a derivative –nus/- analogous to portūnusfortūna (cf. portus and fortu-itus), or possibly from a feminine form in *--.

This is the dilemma which confronts us in the analysis of this abstract noun for which there exists no parallel in Latin. Either pecūnia is an example of the same type of formation as the Slavic words in *-ūnyā and it must be linked up with an ancient adjective and not with the historical neuter pecū; or pecūnia is derived directly from the neuter form pecū, but by a process of suffixation which is not immediately comparable to the Slavic abstract noun in –ynji.

The other substantive which is derived from pecū is pecūlium. Here again we have an isolated form without analogous formations among the neuter words in –ium. Nevertheless it is possible to unravel its formation. Between pecū and pecūlium we have to posit an intermediary *pecūlis, which stands to pecū as īdūlis stands to īdūs and tribūlis to tribus. For the relationship between *pecūlis and pecūlium we might compare edūlis and edūlia (whence edūlium). From pecūlium is formed the denominative verb peculo(r), from which comes the noun peculātus, –ūs. Thus the series pecūlium : peculo(r) : peculātusbecomes parallel with dominium : dominor : dominātus. The whole string of derivatives which are grouped around pecūlium are now rationally organized.

Now the essence of the problem is, however, the meaning of pecūnia, that of pecūlium, and their relation to pecū. In the eyes of comparatists, pecū means livestock, pecūnia ‘wealth in the form of livestock’ and pecūlium ‘the animals given to a slave’. This is the information found in all etymological dictionaries and in works on Latin morphology, all of which repeat the interpretation of the three terms pecūpecūniapecūlium, an interpretation which goes back for centuries and even millennia because it comes to us from the Roman etymologists. The formal relationship between the three words is assured. The problem is how to interpret it. To this end we have to begin by establishing the sense of pecūnia and pecūlium.


Creative Commons

It is not enough to have explained the formal link which exists between pecūnia and pecū. We must also elucidate the sense relationship which follows from the derivation. Yet we shall peruse in vain the works of early and classical Latinity; equally fruitless will it be to scrutinize the examples quoted in Latin dictionaries; nowhere shall we discover a link between the meaning of pecūnia and that of pecū ‘flock, live-stock’. In all the examples quoted pecūnia means exclusively “fortune, money” and it is defined as “copia nummorum.” We thus have no option but to proceed by methodical inference without regard for traditional views.

If from the outset pecūnia had the exclusive meaning of “money, fortune, χρήματα,” this is because pecū has exclusively an economic sense and means “movable possessions.” Only in this way can we account for the constant meaning of pecūniawhich as a collective abstract noun generalizes the specific sense of pecū.

It was only by a special development of a pragmatic and secondary kind that *peku, which meant “movable wealth,” became applied in particular to an item of the real world, “live-stock.” In this analysis we must distinguish two different theoretical planes: (a) that of “signification” and (b) that of “designation.” Consequently we must distinguish (a) the proper sense of *peku as revealed by its ancient derivations and (b) the historical use of the word to designate “live-stock.” Once the semantic link between the particular term *peku and the particular reality “live-stock” was effected, the designation became fixed for a time. But history does not stand still and new specifications can always come about. This is what happened with the differentiations which were effected in Latin between pecūpecus, –orispecus, –udis. They form part of Latin lexical history and do not affect the fundamental relationships which it is our task to bring to light.

It is precisely these relations which have been misunderstood. The result is that both pecū and pecūnia have been wrongly interpreted. And these inexact ideas inspired first Romans and then modern scholars to offer the naïve translation of pecūnia as “wealth of live-stock,” which goes against all the evidence. On the contrary, we may posit that the real nature of the prehistoric pecū is elucidated by the real meaning of the historic pecūnia.

The notion “movable possessions,” expressed by pecūnia, may embrace other types of property than live-stock. Some idea of its original extent can be gained from a notice of Festus which may refer to an archaic expression: pecunia sacrificium fieri dicebatur cum fruges fructusque offerebanturquia ex his rebus constant quam nunc pecuniam dicimus (‘a sacrifice was said to be made with pecūnia, when fruits and produce were offered, because what we now call pecūnia consists of such things’).

For this glossator, fruges fructusque constituted the pecūnia. We record this extended meaning of pecūnia without rejecting, but rather by reinterpreting, the definitions of Varro: pecuniosus a pecunia magnapecuniam a pecu : a pastoribus enim horum vocabulorum origo (‘“pecuniosus” from “great pecunia”; “pecunia” from “pecu”: for these words originally belonged to the vocabulary of herdsmen’).

We only need to read Varro (L.L.) to realize what was understood in his time by pecūnia : under pecūnia he includes words like dos ‘dowry’, arrabo ‘deposit’, merces ‘salary’, corollarium ‘tip’ (V, 175); then multa ‘fine’ (177); sacramentum ‘sacred deposit’ (180); tributum ‘tribute’ (181); sors ‘pecūnia in faenore’ (‘capital bearing interest’) (VI, 65); sponsio ‘a deposit guaranteeing a promise of marriage’ (VI, 70). In addition there existed pecūnia signata ‘minted money’ (V, 169), the nuncupatae pecuniae of legal texts (VI, 60). In short, pecūnia covers all possible uses of money as an economic value or as a monetary symbol; but, we again repeat, it never refers to possession of “live-stock.” This means that in Latin usage, pecū and pecūnia had become separate terms owing to the fact that when pecū became specialized as the designation for live-stock this did not affect pecūnia, which preserved its original sense of “movable possessions.”


Lullingstone Roman Villa / Creative Commons

What has been said of pecūnia is to a large extent also true of pecūlium. We have here a term which, we may say straight away, is still further removed from pecū than pecūnia was. It is known that pecūlium denotes possessions granted to those who had no legal right to possessions as such: personal savings granted by the master to his slave and by the father to his son. The notion of “personal possessions” is the key notion, and they always consisted of movable goods: money or sheep. It is no task of ours to enquire why pecūlium refers to the savings of the slave and pecūnia to the fortune of the master; this is a problem which concerns the history of institutions and not the linguistic form. The distribution stated, we shall recognize the meaning of pecūlium in the derivative pecūliāris ‘pertaining to pecūlium’ or ‘given as pecūlium’. In fact, pecūliāris is only an adjective of pecūlium, and any movable possession can become pecūlium. This is seen as early as Plautus: a young boy can be given as pecūlium to the son of the master and will be called his pecūliāris puer. This is one of the elements in the comedy of the Captivi(v. 20, 982, 988, 1013). In ordinary conditions of life the slave could hardly amass a pecūlium except with what was within his reach: a little money and a few sheep. But this limitation did not imply that pecūlium designated an item of live-stock or a coin.

We thus find in pecūlium a second proof that the basic notion, that of pecū, did not refer specifically to live-stock. In pecūlium, even more than in pecūnia, the connection with personal property is underlined, even if it was restricted to a certain social class. But the possessions concerned are invariably movable ones, whether pecūlium is taken in the strict sense or in the figurative sense. These two notions, personal possession and movable possessions, also apply to the derived verb peculo(r), which yielded pecūlātus ‘(fraudulent) appropriation of public money’. Between this legal term and the basic term pecū a functional continuity can be re-established, pari passu with the link of morphological derivation. We may here argue from analogy. In the same way as we work back from edūlium ‘a tasty dish’ to edūlis ‘edible’ and thence to *edu, roughly “edibles,” so from pecūlium ‘personal movable possessions’ we work back to *pecūlis *‘what may be possessed’ and from *pecūlis to pecū, which we must now define as “(movable) property.” Whatever route we choose, we always arrive at the same conclusion: pecūsignifies “movable property” (personal chattels).


Wikimedia Commons

The word *peku is attested in all the ancient Germanic languages, but the sense varies according to the dialect and it is precisely these variations which are illuminating for the real sense of the term. We must scrutinize it in the proper context of each of the ancient dialects. It so happens that within the Germanic group the Old High German form fihu (variants fehofehu) is the only one which denotes “live-stock.” In texts translated from Latin this word renders pecuspecudes, and more generally iumenta. We may deduce further fëhelîh ‘tierisch’ (animal-like), fihu-stërbo ‘animal-plague’, fîhu-wart ‘Viehhirt’ (herdsman), fihuwiari ‘Viehweiher’ (animal pond). But these are Latinisms and here, as in many other instances, the Latin models were the determining factor. In fact we shall see that Old High German fihu was very far separated from the meaning which the word had preserved in the rest of the Germanic group, and the innovation or specialization must be laid at the door of Old High German, contrary to what is generally believed. Otherwise it would be impossible to understand the situation of *peku in all the other dialects, still to be described. Nor could we understand the role which this Old High German term played in the genesis of mediaeval Latin feudum ‘fief’.

We must first examine the Gothic evidence. The Gothic neuter faihu means only “money,” “fortune” and never refers to the animal world. An example follows. Gahaihaitun imma faihu giban ‘they promised to give him money, epēggeílanto autō̂i argúrion doûnai, promiserunt ei pecuniam se daturos’ (Mark 14, 11).

This example should suffice to demonstrate that faihu, the term chosen to translate Gr. khrḗmataargúrion, Lat. pecūniapossessiones refers exclusively to money, to wealth. This is also shown by the Gothic compounds of faihu, such as faihufriks‘avaricious, pleonéktēsphilárguros’, faihufrikei ‘cupidity, pleonexía’, faihugairns ‘desirous of money, philárguros’, etc.

It is clear that faihu was completely foreign to the pastoral vocabulary which includes quite different expressions, such as hairda‘herd, poímnēagélē’; hairdeis ‘shepherd, poimḗn’; aþwei “flock, poímnē’; wripus ‘flock, agélē’, lamb ‘sheep, ewe, próbaton’. The semantic associates of faihu are the terms which designate money and wealth: gabei ‘wealth, ploûtos’, gabeigs (gabigs) ‘rich, ploúsios’, and the denominative verbs gabigjan ‘to enrich, ploutízein’ and gabignan ‘to enrich oneself, plouteîn’; further, silubr ‘money, argúrion’ (metal and money), skatts ‘the coins, dēnárionmnâ’, in the plural ‘pieces of silver, argúria’.

A further proof that Got. faihu had no connection with the sphere of animal husbandry is furnished by a lexical relationship which has escaped notice and which must be established.

There exists in Gothic a verb gafaihonbifaihon, which translates the Greek pleonekteîn; from this verb is derived the noun bifaih ‘pleonexía’. In the Second Epistle to the Corinthians, which contains all the examples, St. Paul uses “pleonekteîn” for “getting the better of somebody, to enrich oneself at his expense, to exploit him.” This is what the Gothic bifaihongafaihonrenders.

The explanation of faihon is to be found within Gothic itself; faihon is the denominative of faihu. Its morphology is that of verbs made from nouns in –u-, e.g. sidon : sidus, or luston : lustus. The semantic connection between faihon and faihu is seen from the use of compounds of faihu. Since faihu denotes “money, wealth” and since faihufriks translates pleonektēs, just as faihu-frikei and faihu-geiro translate pleonexía, a verb faihon (bi-, ga-) was created as the equivalent of pleonekteîn in the particular sense of “to enrich oneself (at someone else’s expense).”

We now examine the Nordic evidence. The usual translation for Old Norse  ‘Vieh, Besitz, Geld’ (in German—live-stock, possession, money) must be rectified: basic and primary is the notion of “wealth, movables.” This emerges from three circumstances:

  1. the expression gangandi fé for “live-stock” evidently implies that  alone did not signify “live-stock,” but “wealth, possessions”; gangandi fé was used with reference to “wealth on the hoof,” the “live-stock”; cf. Gr. próbasispróbaton.
  2. The compound félag ‘common possession’, from which comes félagi ‘comrade, companion’ (this passed into Old English as feolaga ‘fellow’) also required the sense of “fortune, goods,” for  and not that of “herd.”
  3. The denominative verb féna means “to enrich oneself,” hence “acquire a fortune ().” From this is derived fénadr ‘riches’, which eventually came to mean “live-stock” as the result of a new specialization.

For Old English, it is sufficient to consult the Concise Anglo-Saxon Dictionary by J. R. Clark Hall and Meritt to see that féoh in the sense of “cattle, herd,” traditionally put at the head of the article, is attested only in a few examples, which incidentally would now require careful reconsideration, while the great majority of the examples are found among the headings “movable goods, property” and especially “money, riches, treasure.” We may say, then, that in Old English féoh was applied first and foremost to riches in general or to movable goods and only in the second instance, and then very rarely, to that form of movable property which consists of live-stock. In Beowulf it means solely “riches” or “treasure” and in Aelfric the expression wi liegendum fēo ‘for ready money’ confirms the antiquity of the sense. Finally, there are only three compounds where féoh means “animals” as against about thirty where it means “money, riches.”

The same observation can be made for Middle English by studying the articles on fẹ in the Middle English Dictionary of Kurath-Kuhn (III, 430). There are very few examples meaning “live-stock” but many more of fe in the sense of “movable property; possessions in live-stock, goods or money, riches, treasure, wealth,” and of “money as a medium of exchange or used for taxes, tributes, ransom, bribes etc.”

It was necessary to examine afresh the examples and to classify the usages according to their exact contextual value, liberating ourselves from the traditional schema which imposed “live-stock” as the primary meaning at all costs. This revision would probably be of some consequence for the history of English fee and that of French fief, Old French feu. According to the traditional explanation the Frankish fehu ‘live-stock’ is derived from Latin feus, ‘movable wealth’. It would rather seem that fihu, like Gothic faihu, designated all forms of movable property and that it kept that sense when it passed into Latin. Here, too, a new examination would appear to be called for.


What has been outlined above shows that the traditional conceptions of *peku in Indo-European must be entirely rethought. Our first conclusion is that *peku signifies “movable personal possession.” That this possession may in fact take the form of live-stock is a separate datum which concerns social structure and the forms of production. It is only in virtue of this frequent association between the term *peku and the material reality of animal husbandry that, by a generalization which took place outside the class of producers, *peku came to mean “live-stock” (the first specialization), then specifically “small live-stock” (the second specialization), and finally “sheep” (the third and last specialization). But intrinsically *peku does not designate either the flock or any animal species.

We are now able to establish a correlation between the proper sense of *peku, thus restored, and its dialect distribution. It is interesting to note the fact that *peku is lacking in Greek. This is no accident. Such an important notion could not simply disappear. The Indo-European term was in fact replaced in Greek by a new designation, which had the same sense. This is the Homeric próbasis with its far more common equivalent, próbata. Our study of this term (see above, Book One, Chapter Three) has revealed the evolutionary model which we have posited for *peku: it was, to begin with, an expression which designated “movable possessions.” For extra-linguistic reasons this term was frequently applied to the possession of “live-stock.” It thus became the word for live-stock and subsequently for the predominant species, “the sheep.”

But as was shown above, this specialization, which took place at an early date in Indo-Iranian, did not take place everywhere. We have in Latin and in a large part of Germanic testimony of great antiquity which shows that the initial meaning was “movable possession” and this explains the derivatives. This evolution is not reversible. It is in the highest degree improbable that *peku, if it had really signified “live-stock,” could have come to designate “money” and “fortune” in general, which is the exclusive meaning of próbata. Similarly the specific English term cattle, Fr. cheptel, goes back to Latin capitale ‘principal property’; already in a text of 1114, captale means “chattel, cattle, movable goods.” [3] But in the Middle Ages it still has the meaning “fortune, goods, income,” and the Spanish caudal signifies “goods, riches.” The progress from “movable possession” > “live-stock” is characteristic. But once accomplished, it is irreversible. Thus “livestock” is very often designated by expressions which refer to possessions in general, that is, it is simply called “property”; but the reverse never happens.

Our interpretation of *peku and its evolution thus conforms to what might be called the norm with regard to the terms of possession: a general or generic term is used by a certain class of producers as the designation for the typical object or element. In this sense it spreads outside its original milieu and becomes the usual designation of the object or element in question. Such is the case here. We have been able by comparative study of the evidence presented in three dialect groups to follow the stages of the process in the case of *peku, and to verify to a certain extent this internal reconstruction.

A last conclusion concerns the etymology of *peku. If the present demonstration is considered acceptable, it destroys the traditional rapprochement with *pek(t)- ‘to shear’. It is evident that *peku, a term with an economic sense which does not denote any animal, can have nothing in common with terms derived from *pek(t)-, which are concerned with the technique of shearing and combing wool: Gr. pékō ‘comb, card’, pókos ‘fleece’, pektéō ‘shear’, pékos n. ‘fleece’, pokízō ‘shear the wool’, kteís ‘comb’; Lat. pecto ‘comb, card’, pecten ‘comb’, pexus ‘hairy, downy’, Arm. asr ‘wool’. Between these forms and *peku the resemblance amounts to no more than simply homophony. The connection must be abandoned, and *peku-, a vestige of the most ancient Indo-European vocabulary, seems irreducible to any known root. [4]


1. Zeitschrift für Vergleichende Sprachforschung, 1914, 46–54.
2. For the whole of chapters 1, 2, 3 reference may be made to our article “Noms d’animaux en indo-européen” in Bulletin de la Société de Linguistique de Paris, 1949, pp. 74–103.
3. Baxter-Johnson: Mediaeval Latin Word-List, 1934, p. 64.
4.. A much more detailed version of the present study has been published in the USA in the conference proceedings entitled Indo-European and Indo-Europeans, 1970, University of Pennsylvania Press, 307–320.

Originally published by The Center for Hellenic Studies under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported license. From “Indo-European Language and Society”, by Émile Benveniste (University of Miami Press, 1973).