Open main menu

The Nature and Origin of the Noun Genders of the Indo-European Languages










Professor of Indogermanic Philology in the University of Leipsic

Translated by


Princeton University


Copyright, 1897,

By Charles Scribner's Sons.

University Press:

John Wilson and Son, Cambridge, U. S. A.


AMONG the many valuable contributions of William Dwight Whitney to linguistic science is one especially important and fundamental principle. It may be stated in these words. In explaining the prehistoric phenomena of language we must assume no other factors than those which we are able to observe and estimate in the historical period of language development. The factors that produced changes in human speech five thousand or ten thousand years ago cannot have been essentially different from those which are now operating to transform living languages. On the basis of this principle we look to-day at a much-discussed problem of Indo-European philology with views very different from the views held by the founders of Comparative Philology and their immediate successors. I refer to the problem, how the Indo-European people came to assign gender to nouns, to distinguish between masculine, feminine, and neuter. This question is of interest to others besides philologists. What man of culture who has learned languages such as the Greek, Latin, or French has not at times wondered that objects which have no possible connection with the natural gender of animals appear constantly in the language as male or female? In German, for example, it is der fuss, but die hand; der geist, but die seele; in Latin, hīc hortus, hīc animus, hlc amor, but haec planta, haec anima, haec felicitas; in Greek, πλοῦτος, ὁ οἶκος, but ἡ πενία, ἡ οἰκία.

This gender distinction pervades all the older Indo-European languages, and must therefore be regarded as having its origin in the time of the pro-ethnic Indo-European community. Not only is the subject itself full of interest, but also the treatment it has received from the philological research of our century. The various efforts made to solve the problem may very aptly illustrate an essential difference which exists between the theories of language development held in the beginning and middle of this century and those which prevail to-day,—a difference of method existing not in comparative linguistics alone, but also in other fields of philological and historical research that border on it.

Permit me, then, gentlemen, in this lecture, first to set before you the views of earlier investigators on this subject, and then the position taken by scholars of more recent times.

Let me neglect, for the moment, the so-called neuter gender, and consider only the distinction made in nouns between masculine and feminine. First of all, we must notice that there is a certain difference in the mode of expressing this gender distinction in the Indo-European languages, depending upon whether it is a real physical sex that is marked, or what is usually called "formal" or "grammatical" gender, which has to do with concepts possessing no natural, animal gender. In the case of natural sex there is to be noticed in all Indo-European languages a two-fold method of giving it expression. In a number of the words that denote living beings the name for the male and the name for the female are formed from different roots, and the mode of inflection may be the same for both roots. It is so in the case of Latin pater and māter, Greek πατήρ and μήτηρ. Here the root of the word distinguishes between the male and the female. Take, on the other hand, pairs such as Latin deus and dea, gallus and gallīna, Greek θεός, 'god', and θεά, 'goddess', λύκος, 'wolf', and λύκαινα, 'she-wolf', English god and goddess: here the word for the male and that for the female have the same root material and a common stem meaning; the inflectional ending only is different. The grammatical term for this in German is "motion". We say the word is "moviert" in order to mark the feminine sex. In cases of grammatical gender, on the contrary, there is but one way of making a distinction, viz.,—by inflection. The gender is made evident only by the inflectional endings, as in Latin animus, anima, Greek οἶκος, οἰκία. This fact shows us that the question as to how "formal" gender is related to natural gender, and how the history of both is to be investigated, depends entirely and exclusively on the terminations used to express gender, on the inflectional suffixes which mark sex.

In only two or three places in the whole circle of human languages has anything been found comparable with the formal gender of the Indo-European languages. In the Semitic-Hamitic group, especially, the whole language is pervaded with the idea of gender, but in a manner that is entirely different from the Indo-European, externally and internally. There are scholars who believe in a relationship between the Semitic-Hamitic family and the Indo-European, but up to the present it has not been proved; and the so-called gender of nouns is of all things least adapted to furnish an argument for a close genealogical connection. Everything goes to prove that in the matter of gender there was no common development, but that the genders had a separate history. It is accordingly correct method if we first investigate the history of noun genders in each family by itself.

"We have noticed that very few families of languages mark gender distinctions in their substantives. Even within the Indo-European, not all of the languages have preserved this peculiarity. The English, for example, has but a few remains. These languages without grammatical gender are just as well off. The category is entirely superfluous as regards the main purpose of language, which is to express thought in the clearest possible manner. Not only superfluous is it, but often even contradictory and foolish. Sophists like Protagoras held this opinion in antiquity. They ridiculed the gender distinctions of the Greek, and it is easy to see why. What real connection with animal gender have all those concepts which our primitive ancestors characterized as masculine or feminine? This peculiarity of our language does not usally cause any practical difficulty to us Indo-Europeans. We learn it, as we learn all other peculiarities of the language structure, in early childhood. It enters in sucum et sanguinem with the rest. People who are not ludo-Europeans, whose mother tongue has no formal gender, have a very different experience, when in more mature years they learn a language like the Greek or Latin. A new world opens itself to them as they find, for example, Latin animus called a masculine form, and anima a feminine form. They marvel at the imagination of the Indo-European people, who can look at everything, be it never so abstract and lifeless, as a concrete object, and as having a corporeal existence, and who, further, assign a sex in each case, masculine or feminine.

The grammarians of classical antiquity did little more with this problem than to become thoroughly perplexed over it. They contented themselves with the assertion that man uttering speech had the right to assign arbitrarily a sex to any object which had in this particular been neglected by nature. Not until the philosophic grammar of the eighteenth century took hold of the subject, was it treated in a scientific manner. Herder and Adelung were the first to attempt an explanation. They insisted that early man in his simplicity long considered everything he looked upon as animated, and treated it as a living being. Grammatical gender is, according to this, the result of the tendency of primitive man to individualize and personify. Adelung tried also to specify why in particular cases this or that gender was chosen. He says that everything which was characterized by activity, liveliness, strength, size, or had anything of the frightful or terrible in its nature was made masculine. Those objects, on the contrary, that were felt to be susceptible, fertile, delicate, passive, attractive, became feminine. Jacob Grimm followed in this track in the third volume of his German Grammar. He treats the question of the origin of noun genders in the Indo-European languages at great length—some two hundred and fifty pages—and with all his incomparable skill and grace in presentation. He believes with Adelung that grammatical gender had its origin in the creative imagination of the primitive folk. He thinks that in that remote pro-ethnic period, at a time when imagination, not reason, was the predominant faculty, man individualized and personified every possible lifeless object of the external world, and assigned to it masculine or feminine traits. Just as Adelung, Grimm also believes that whatever gave the impression of the larger, stronger, more rough, more active, was looked upon as masculine; on the contrary, whatever was felt to be smaller, finer, more gentle, soft, tender, or still, was made feminine. He tries to prove this by many special cases, and investigates with poetic spirit the characteristics of natural objects. One says die hand, haec manus, ἡ χείρ, but der fuss, hīc pēs, ὁ πούς, because the hand is thought of as the smaller, daintier, the foot as larger and stronger. All the philologists of that day accepted the Adelung-Grimm. hypothesis, and it remained unattacked until long past the middle of the present century. Wilhelm von Humboldt, Pott, Miklosich, Steinthal, Madvig, and Georg Curtius, for example, accepted it openly. It is unnecessary for me to describe at length how this theory stands in the closest relation to a belief still prevailing in the days of Humboldt and Grimm,—the belief in a golden age of mankind, where poetry beautified and simplified the whole life of primitive man. Nor need I dwell on its particularly close relation to the then current theory of the origin and nature of folk-poetry. The explanation of Adelung and Grimm has long out-lived those views and beliefs out of which it originated. Yet even Wilhelm Scherer called the chapter on gender the acme of Grimm's grammar. And only a few years ago, in 1890, this theory found a warm and eloquent defender in the person of Gustav Roethe, a young Germanic scholar of talent and repute. In the preface to the new edition of the third volume of Grimm's grammar, edited by him, this scholar declares Grimm's view of the origin of gender to be correct in all essential points. But opposition had arisen before Roethe's time. A calmer and more critical spirit began to pervade the science of language from about the year 1870. The more matter-of-fact learning of the newer linguistics, which supports itself on more solid foundations, was compelled to question seriously Grimm's hypothesis; and the decision had to be rendered that this theory, though idealistic and poetic, was not strictly scientific. Allow me to present to you in few words the reasons why I, as well as some other philologists, have come to the conclusion that it must be rejected.

Firstly. If we pursue a correct method and start from what we know empirically; if we confine ourselves to the facts that lie clearly before us and can be judged by the materials of our science,—facts that belong to the present or recent past of our Indo-European languages,—then we must assert that masculine and feminine as grammatical genders say and mean nothing for the speech of every-day life. And it is only the ordinary, every-day language that is of importance for this subject. By the grammatical gender, no idea of anything masculine or feminine, either in literal or figurative sense, is called up. The masculine and feminine suffixes differ entirely from other noun suffixes, to which grammatical terminology has assigned names on the basis of some definite signification. The Germans, for example, call -chen and -lein diminutive suffixes, and, in fact, every German understands by söhnchen and söhnlein a small son. So in English booklet is a small book, or lambkin is a little lamb. Nowhere, however, in the Indo-European languages can it be proved that, for example, the Indo-European "feminine suffix" , as it appears to-day in Lithuanian and Russian, e. g., Lithuanian rankà, Russian ruká, 'hand', and as the Romans had it in anima, casa, fuga, the Greeks in χώρᾱ, 'land', οἰκία, 'house', calls up, or has called up in any degree, the idea of female or of any especially feminine characteristic. And how can any one prove that it was different in the primitive community, when there must have been hundreds of substantives in -ā- which did not signify living beings? Among these, too, there must have been many that denoted concepts which were in no sense concrete, but purely abstract, as, for example, *qṷoinā, 'recompense ', from which comes Avestan kaēna, Greek ποινή, Old Church Slavonic cěna. That the formal gender in our Indo-European languages for thousands of years was not connected with the idea of the masculine or feminine, is shown by quite unmistakable evidence. I will call attention here to but one proof. Έπίκοινα (epicoena) is the term used by grammarians for those substantives which, although they denote animals, have for both physical genders only one language expression. The German says der hase, 'the hare', der adler, 'the eagle', and means by this both the male and female; again, die mans, 'the mouse ', die eule, 'the owl', for both the male and female. In like manner the Greeks said, for example, ὁ μῦς, 'the mouse' , but ἡ ἀλώπηξ, 'the fox'. If there had been any feeling that the real physical sex was expressed by the gender assigned to the word by the language, they would have had to understand by der hase always the male hare only, and by die maus always the female mouse. Further, when one said der weibliche hase or die männliche maus, he would feel that this manner of expression contained a downright contradiction. But this is nowhere the case.

A second point which speaks against Grimm's hypothesis is the following. The Indo-Europeans, from the very fact of being a primitive people without culture, are asserted to have had this remarkable impulse to personify and sexualize. This tendency accordingly made them take a fanciful view of the whole universe and think of the great majority of noun concepts as male or female. Now there are peoples to-day who still represent about the same degree of culture which we must suppose our ancestors possessed at the time when they began to differentiate nouns into masculine and feminine. Should we not find among these peoples some parallel to this mental attitude? Yet nothing has come to light at all comparable with the Indo-European sexualizing—in the sense in which it is presupposed by Grimm's hypothesis; this too, though some of these uncultured peoples look at everything in a very concrete fashion, and possess a very lively imagination, which displays itself in their language as well as in many other directions. It is not a valid objection to say here that the structure of the languages of these savage peoples is essentially different from that of the Indo-European languages, and also from that of the Semitic-Hamitic languages. In a psychological sense, the grammatical categories of our inflectional languages are to be found in every language of the earth; the mode of expression alone is different. Had the Indo-European gender suffixes originally meant male and female, or manlike and womanlike, the other languages would have been by no means without analogies in their manner and means of expressing gender distinctions.

In the third place, Grimm's theory is in itself psychologically improbable. It presupposes that noun concepts were always (1) individualized and thought of as a separate object, (2) conceived of as a living being, and (3) sexualized as male or female. Now, for primitive man the external world was mostly matter, material, just as it is for us to-day, but to him even more so perhaps than to us. Material and general concepts, such as gold, mud, water, fog, flesh, grain, were certainly not as a rule conceived of and named in the pro-ethnic period as individuals; yet they show in large part, since primitive times, either masculine or feminine gender. How is it that such substantive concepts came to be conceived of as male or female, if they were not even considered as an individual? Further than this, that which is individualized is not necessarily thought of as animated and personal. Even if we imagine to ourselves the fancy of the Indo-European as lively and active, creating for itself many mythical images, yet however active it may have been, it could have drawn only a small circle of objects into its scope. It is certainly true that our primitive ancestors thought more in the concrete than we moderns. But "concrete thinking" does not mean to consider as man or beast something which is not concrete, but which is in its very essence abstract. Animalization and personification, like poetry, have their origin in fantastic, exalted feeling, and there never has been a time when man stood continually on such a poetic height. Every-day life is hard and prosaic in modern times, and still more stern and prosaic was it in those primitive days to which Grimm's theory carries us back. Aside from this, the creative imagination of man produces not only anthropomorphic and theriomorphic beings, but also inanimate, material metaphors. The cloud floating across the heavens, for example, is looked upon in mythology as an animate being, as a giant, but is also considered as a garment of air, a cloak, or something similar. Why should one think that primitive man overloaded language with personal metaphors instead of impersonal? And, thirdly, each particular thing, even if it is animalized, is not necessarily at the same time sexualized. Very often our imagination discovers in a lifeless object attributes of a person, and for the moment, or for a longer time, personifies this object, and forms out of it a living being. But it is not necessary that all the characteristics of a living being be present in our consciousness, and that in each case we think of the objects accordingly as male or female. Language itself shows us, with its epicene nouns already mentioned, with its words like Greek λύκος, Latin lupus, German wolf, used alike for male or female, that often enough no notice is taken of distinction of sex.

One fact stands out clearly as the conclusion to all this: Grimm's theory ascribes to the Indo-Europeans a mental condition which we cannot harmonize with what we actually know of the mental life of man and of races. It may find a parallel, at best, in certain pathological states of the human intellect. But, you may ask, does not one thing argue very strongly in favor of Grimm's theory,—the fact, namely, that in the mythology and poetry of the Indo-European people, where lifeless concepts are personified, the sex of the mythological personage corresponds regularly to the grammatical gender of the words concerned? The Greeks thought of ὒπνος, 'Sleep', and θάνατος, 'Death', as male deities, not as female, and γαῖα, ' Earth ', and ἄτη, 'Folly ', as goddesses, not as gods. In the old Germanic mythology der Tag (New High German der tag) appears only as a god, die Nacht (New High German die nacht) only as a goddess. Among the Hindoos agni, the 'Fire', only as a god, usas, the 'Dawn', only as a goddess. And so in other cases. But this sort of personification in no way substantiates the theory that the origin of grammatical gender is to be sought in the particularly active imagination of the primitive Indo-European people. Grimm's theory deceived our mythologists and led them to a mistaken view,—a view that meets us in several places in the otherwise instructive book of H. Usener on the "Götternamen." In all the cases that come into consideration here the grammatical gender of the word, so far as we can judge, is the earlier. The imagination used this gender and allowed itself to be led by it. The procedure is the same whether it is uncultivated primitive man creating a myth unconsciously, or the poet doing it with conscious effort. When either personified a lifeless concept into a living being, it was the grammatical form of the noun that, through the psychological impulse of analogy, an impulse that was very strong, and was, indeed, almost compulsory, decided the definite direction of the gender,—whether it should be masculine or feminine. Our thoughts and conceptions cling close to the language form. We do not control and lead language, but language rules and directs us. The Greek ὕπνος, for example, was from Indo-European time a simple appellative, just as the corresponding words in the other Indo-European languages, Latin somnus, Sanskrit svápnas, Lithuanian sãpnas, etc. Because ὕπνος had the same inflectional form as the numerous masculine nouns in -ος, like ἀδελφός, 'brother', θεός, 'god', and was, in consequence, most closely associated in the consciousness with these, the Greeks made Sleep a god and not a goddess. So far as the concept itself was concerned, there was nothing to prevent Sleep from being personified as a female. Ὑγίεια, 'Health', became a female deity, because the word in its inflectional form corresponded to the numerous words denoting females,—such as γραῖα, 'old woman', πότνια, 'mistress'. There is nothing in the nature of health that compelled it to be of either gender as opposed to the other. For this reason the separate Indo-European peoples have often assigned different genders to the same deities; the cause being that the appellative itself was of different genders in the different languages. To the Greeks and Romans ἔρως and amor was a boy or youth; for both these words were masculine when used as appellatives. To the Germans, on the contrary, die Minne, die Liebe was a goddess, since the appellative was feminine. To the same cause are due the different conceptions of the sun and moon, as male or female, among the different peoples; it is always according to the nature of the appellative. In the pro-ethnic period the state of affairs must have been similar. Therefore, there is nothing to hinder the assumption that pro-ethnic *dyēus, the dyāuš of the ancient Hindoos, the Ζεύς of the Greeks, the Jūpiter (Juppiter) of the Romans, became a male deity, and not a female, because the name, originally meaning 'heaven' and 'bright day', was a masculine appellative. In every case where the mythological name is at the same time retained in its original appellative signification, we have a similar right to assume that only the grammatical gender was present at first, and that this decided the choice of sex for the personified conception. Even to-day, in the art of the Germans and other peoples who still have the grammatical gender, this gender receives recognition. I know of numerous pictures and statues which represent die elektricität as a person, and all personify the force as a female. Yet occasionally the grammatical gender is overlooked. At an art exhibition in one of the larger German cities about fifteen years ago, Hunger was to be seen represented in marble as a ragged, hollow-eyed old woman. At the base of the statue was the inscription Der Hunger. A critic claimed, in a newspaper article, that this was incorrect. Der Hunger should be made a man. The artist had perhaps followed a French prototype, whose creator had chosen a woman because of la faim. The sculptor got rid of the inaccuracy, but not by carving a new statue and making it a man; he simply wrote beneath his figure the words,—Die Hungersnot.

It holds good, then, for the historical period of the Indo-European languages, that in personifying lifeless things, the sex is usually determined by the grammatical gender; and no one can prove that in such cases the anthropomorphic conception is older than the word with which our ancestors named the thing. This fact, in my opinion, destroys the foundation of Grimm's hypothesis. Jacob Grimm, with poetic fancy, sought to recall to us a beautiful idyl of the past. Sentimentalists may lament the excessive sobriety and arid intellectuality of modern grammarians, which dares in its lack of appreciation to disturb this idyl. I, for my part, cannot but feel that, in declaring the beautiful idyl to be a mere poetic glorification of the imaginative faculties of our ancestors, and in explaining grammatical gender as a fiction, we are not robbing them of anything we could wish them to have possessed.

You will ask, What is the truth about grammatical gender? How came the Indo-Europeans to possess it, if Grimm is in the wrong? I must consider here for a moment the so-called neuter, which has thus far been left out of the discussion. You know that the neuters in all Indo-European languages, so far as their stem formation is concerned, belong with the masculines. Latin jugum, genitive jugī, etc., has the same o-suffix that appears in words like dolus, populus; mare, maris, etc., has the same i-suffix as avis, turris. The difference between the neuters and the other genders consists only in the different case-forms used for nominatives and accusatives. The masculine nominative singular has an -s, the neuter an -m as case-sign. This is connected with a fact that I cannot here dwell upon at greater length, viz., that while the forms with -s originally served as the subject of the sentence (which could also be represented by the -m form), these -s forms had, besides this, a more specific meaning. The termination -a in Latin plural juga was a mark of distinction for a collective signification of the noun stem. These are relationships of a very different character from the antithesis of masculine and feminine, which has up to this point received our attention. One cannot, to be sure, neglect entirely the neuter in studying the origin of the masculine and feminine, yet the development of the neuter is another question, and one subordinate to that which regards the origin of the other two genders. We can therefore omit it from this discussion.

The masculine and feminine gender is expressed by means of the so-called stem suffixes, as, for example, the contrast acc. sg. ani-mu-m and ani-ma-m shows. All suffixes which appear from primitive times in both masculines and feminines are here irrelevant to the question; such are -i-, -u-, -men-, -ter-, -es-. It is perfectly plain that these never had a specifically masculine or a specifically feminine significance, and had consequently nothing whatever to do with this distinction. Further it must be noticed, that the o-suffix in the so-called masculines, such as Latin animus, deus, Greek ἄνεμος, θεός, cannot originally have denoted a physical sex. This is shown by the universally recognized fact that those substantives in -o-s which denote men or beasts were used primitively as a general term for the animal without regard to sex distinction. Note, for example, the Indo-European word *eko-s = Latin equos, Greek ἵππος, Sanskrit aš̍vas. It signified originally horse in general, and did not have any special meaning like stallion. Not until there appeared by the side of such substantives in -o-s, forms with the suffix -a- or with the suffix -iē-, -ī- to denote the female, did the use of the o-stems suffer any limitation. It was then that the o-stem first came to be employed to signify specifically a male being. In this way Latin equos, by contrast with equa, 'mare', acquired the special meaning 'stallion'. So Sanskrit vŕ̥kas, in contrast with vr̥kī̍, 'she-wolf', was used to mean the male wolf. In a word, the whole problem that is at present claiming our attention, depends for its settlement on one question. What was the original function of the -ā- in words like Latin anima, equa, Greek χώρᾱ 'land', θεᾱ́, 'goddess', Sanskrit bhidā̍, 'split'? And what was the original function of the -iē-, -ī- in words like Latin acies, Greek γλῶσσα, 'tongue', πότνια, 'mistress', Sanskrit š̍ácī, 'strength', vṛkī̍, 'she-wolf'?

According to the older theory, as we have seen, these suffixes -ā-, -iē-(-ī-) originally carried the notion of female, or some special female characteristic. We are brought to a different decision. They did not, in all probability, have that original signification. It is a misuse of the grammatical terminology to call these two suffixes, in general, and in every case where they appear in the Indo-European languages, by the name "feminine suffixes." They are feminine only in some cases, and, indeed, in only a comparatively small proportion of the whole number, as in words like Greek θεά and πότνια, are they really what we ordinarily call them. Similar misuse and similar inaccuracy and inadequacy of the scientific terminology is to be found in many other cases. To mention but a single example: the suffix -to- in the Latin substantive formations such as datus, amātus, fīnītus, is called the suffix of the perfect "passive" participle. Yet the Latin has numerous -to- participles with signification that is not passive; for example, ratus, secūtus. These participles, as can readily be proved, never had, in Latin nor in the pro-ethnic Italic language nor in the Indo-European, passive meaning. The participial suffix -to- was originally used to form verbal adjectives which predicated an action as a distinguishing characteristic or peculiarity; for example, ὕδωρ ῥυτόν is water which has the characteristic that it flows, and in which the flowing is perceptible. This is a function that has nothing whatever to do with the distinction between active and passive. Recognizing this fact, we can understand how the suffix came to be used in all the older Indo-European languages in active as well as in passive forms. He who comprehends all Latin -to- participles under the name perfect "passive" participles makes a false use of this term. This misuse is wide-spread, and unfortunately so in its results, for it gave rise to the belief that the fundamental meaning of -to- was a passive one. In a very similar manner the circumstance that -ā- and -- denote the female animal in some of the substantives formed with them, has had the result that we speak of the "feminine" suffix in words like Latin anima, acies. In both cases there is an unjustifiable generalization of a term.

If one examines all the words of the Indo-European languages which are formed with the suffixes -ā-, -iē-(-ī-), he comes readily to the view that the original function of these suffixes was to form abstracts and collectives. This fundamental meaning would, in many cases, be preserved unchanged in all Indo-European languages. It remains in Latin fuga, 'flight', juventa, 'youth', aciēs, 'sharpness', materiēs, 'matter', 'building material'. The view can be well defended that our suffixes started with this original function, and acquired afterwards, though still of course in the period of the Indo-European community, the power to denote living beings of the female sex. Allow me to establish this position as briefly as possible. It is a peculiarity of the languages of the Indo-European family, modern as well as ancient, that names with abstract signification are often employed for concrete concepts. Terms expressing a quality come to be used to denote the individual person or thing which possesses that quality; further, terms of collective signification are employed to designate individuals. A good example is found in those words which mean youth and youthfulness. In several languages, words such as these have come to be used as implying a single youthful person. The English word youth is a case. Beauty in English, and the corresponding die schönheit in German, are both used to designate beautiful people, though chiefly, of course, those who belong κατ' εξοχήν, to the "fair sex." The German employs the abstracts bedienung, 'servic', aufwartung, 'attendance', for individuals who serve and attend. The German frauenzimmer in the older New High German meant those women who lived together in the part of the house reserved for females, viz., a number of women. Since the beginning of the eighteenth century, and up to the present time also, it means only a single woman. In this way, or in a similar way, certain abstracts and collectives in -ā- and -iē- may, in the Indo-European period, have become names for females. The common Indo-European word for woman, pro-ethnic *gṷenā (Greek γυνή, Gothic qinō, Old Church Slavonic žena), can originally have had the meaning 'bearing', 'parturition', and the transition to the signification 'the animal that bears' would have been the same as the transition of bedienung, 'service', to bedienendes wesen, bedienende person, 'one who serves', English colloquial 'help'. Pro-ethnic *ekṷā, Latin equa, can have meant originally 'a drove of horses', 'a stud '. The way it comes to mean 'mare' is shown by the German word huhn; this meant at first the cocks and the hens together, then the flock of female fowl, and finally the individual female fowl. If the suffixes -a- and -ie- implanted themselves in this manner in a number of words of feminine signification, the idea of feminine sex could attach itself to the suffixes, and they could acquire this additional shade of meaning. The final step was for the suffixes to become "productive" with this meaning inherent in them; and that, too, has sure parallels in both the newer and older Indo-European languages.

I quote two examples,—one from the German and one from the Greek. The suffix -iska-, used by Germans to form adjectives, equivalent to English -ish, New High German -isch, had originally a very general adjective signification. We have it in Gothic mannisks, 'manly', in English thievish, in New High German himmlisch, 'heavenly'. It appears with especial frequency in derivations from the names of persons and peoples, e. g., New High German kriegerisch, römisch, englisch. In a number of these adjectives it happens that the noun forming the base of the adjective is the name of a person whose rank or occupation is considered blameworthy or contemptible ; such as New High German diebisch, from dieb, 'thief', räuberisch from räuber, 'robber', närrisch from narr, 'fool'. In this way the element -isch has itself come to share in the idea of the contemptible, and particularly in this direction has become "productive" in New High German. New words have been coined with the suffix -isch to express the sense of contempt. Abgöttisch, 'idolatrous', teuflisch, 'devilish', ' selbstisch, 'selfish', linkisch, 'awkward', hämisch, 'knavish', are examples. The suffix has not, however, in all cases acquired the additional ethical significance. In himmlisch, 'heavenly', städtisch, 'urban', kriegerisch, 'warlike', and many other adjectives it has retained its ancient meaning, which implied no notion either of contempt or esteem. Just as these last-mentioned words have remained entirely uninfluenced by the idea of contempt, so many of the substantives formed with -a- and -ie- contain nothing of the idea of feminine sex: such are Latin fuga, anima, Greek φνγή, χώρα, etc. The parallelism goes still further. If the Romans, when they personified lūna or abundantia, thought of them as feminine, and made them female deities because they associated them with words for female beings like dea, fēmina, lupa, the process is analogous to the treatment of kindisch, 'childish', weibisch, 'womanish'. These last did not yet have in Luther's time any touch of disparagement in their meaning, but denoted what is to-day expressed by kindlich, 'child-like', and weiblich, 'womanly'. They received the secondary touch of disparagement in consequence of the influence of adjectives like diebisch, närrisch, teuflisch, and the like. The further example is the history of the primitive suffix -bho- in Greek. This suffix, whose original meaning is not clear, is found in Greek in names of animals, like ἔλαφος, 'stag', ἀσκάλαφος, 'owl', but also in words of entirely different signification, as κρόταφος, 'temple', κόλαφος, 'cuff on the ears', φλήναφος, 'gossip', 'chatter'. The Greek inherited two or three names for animals which had this termination from the time of the Indo-European community. They became models in the Greek language, and a large number of animals received names in -αφος, -αφη, formed on the analogy of these few. On the signification of the words like κρόταφος, which lay outside of this category, the spread of -αφος in the zoological terminology had no influence.

There are still, gentlemen, many other questions which I should answer, and these have no doubt occurred to you in the course of my discussion. Above all, there is the question as to how the Indo-European people came to express distinctions of gender in the forms of the adjectives, as in magnus,-a,-um. I cannot attempt to-day to enter upon these difficult and complex problems. The solution of the main problem does not depend upon them. It is sufficient for me to have shown that it is possible to take a historical view of the noun genders masculine and feminine without ascribing to our Indo-European ancestors a mental state that has no analogy in those periods that are familiar to us from historical tradition. The solution which I have presented to you can unfortunately never be absolutely proved; for we have to do with a period in the history of language in which we cannot go a step further than simple hypothesis. It can be said, however, of our explanation, and it is indeed its strongest claim over the theory of Adelung and Grimm, that it keeps within the limits of phenomena which are among the best substantiated in the history of the Indo-European language family, and which may be observed in the very latest phases of its development.

And so I return, in my conclusion, to that statement with which I began. I said that the different attempts which have been made to explain the problem of grammatical gender in the Indo-European speech illustrate well the difference between the methods of investigation employed by the older generation of linguistic students, and those of the generation at work in the present. In the time of Grimm and Bopp and their immediate successors, it was the custom to devote attention preferably to the prehistoric times, and to explain the peculiarities of the primitive language largely on subjective theories. Forces and tendencies were ascribed to the primitive tongue and to the prehistoric period which have no analogy in historical times, and pro-ethnic antiquity was thus surrounded with a fantastic and mystical glamour. The later investigation regards more the present. It considers it of the chief importance to understand that which is now before us, and that which belongs to the immediate past, investigating its growth and development. Its principle is this: to take as the starting-point what is known by experience, and to apply this to the unknown of the past, to the conditions of prehistoric times; to use it not without discretion, but yet as the main criterion for recognizing ancient conditions. In this way we may hope to throw some light even upon those most remote periods in the history of our language. Your own countryman Whitney was one of the first to insist on these principles of investigation. It is my hope that this spirit of genuine historical induction, which has prevailed but a single score of years in linguistic science, may never again be lost to Indo-European philology.

  1. This lecture is based chiefly on the following articles: Brugmann, Das Nominalgeschlecht in den indogermanischen Sprachen, Techmer's Internationale Zeitschrift für Sprachwissenschaft, IV. p. 100 sqq.; Zur Frage der Entstehung des grammatischen Geschlechtes, Paul und Braune's Beiträge zur Geschichte der deutschen Sprache und Literatur, XV. p. 523 sqq.; Michel, Zum Wechsel des Nominalgeschlechts im Deutschen, I. (Strassburg, 1889), p. 3 sqq.; Zur Beurtheilung von Jacob Grimm's Ansicht über das grammatische Geschlecht, Germania, XXXVI. p. 121 sqq. Other recent articles on the subject in hand are: Roethe, Vorrede zum Neudruck der Grimm'schen Grammatik, Band III. (1889), and Anzeiger für deutsches Altertum, XVII. p. 181 sqq.; Henning, Ueber die Entwicklung des grammatischen Geschlechts, Zeitschrift für vergleichende Sprachforschung, XXXIII. p. 402 sqq.