Let us begin with what is nearest to us, Philology. Its whole character has been changed as if by magic. The two classical languages, Greek and Latin, which looked as if they had fallen from the sky or been found behind the hedge, have now recovered their title-deeds, and have taken their legitimate place in that old and noble family which we call the Indo-European, the Indo-Germanic, or by a shorter, if not a better name, the Aryan. In this way not only have their antecedents been cleared up, but their mutual relationship, too, has for the first time been placed in its proper light. The idea that Latin was derived from Greek, an idea excusable in scholars of the Scipionic period, or that Latin was a language made up of Italic, Greek, and Pelasgic elements, a view that had maintained itself to the time of Niebuhr, all this has now been shown to be a physical impossibility. Greek and Latin stand together on terms of perfect equality; they are sisters, like French and Italian:—
“Facies non omnibus una,
Nec diversa tamen qualem decet esse sororum.”
If it could be a scientific question which of the two is the elder sister, Greek or Latin, Latin, I believe, could produce better claims of seniority than Greek. Now, as in the modern history of language we are able to explain many things that are obscure in French and Italian by calling in the Provençal, the Spanish, the Portuguese, nay, even the Wallachian and the Churwälsch, we can do the same in the ancient history of language, and get light for many things which are difficult and unintelligible in Greek and Latin, by consulting Sanskrit, Zend, Gothic, Irish, and even Old Bulgarian. We can hardly form an idea of the surprise which was occasioned among the scholars of Europe by the discovery of the Aryan family of languages, reaching with its branches from the Himalayan mountains to the Pyrenees. Not that scholars of any eminence believed at the end of the last century that Greek and Latin were derived from Hebrew: that prejudice had been disposed of once for all, in Germany at least, by Leibniz. But after that theory had been given up, no new truly scientific theory had taken its place. The languages of the world, with the exception of the Semitic, the family type of which was not to be mistaken, lay scattered about as disjecta membra poëtæ, and no one thought of uniting them again into one organic whole. It was the discovery of Sanskrit which led to the reunion of the Aryan languages, and if Sanskrit had taught us nothing else, this alone would establish its claim to a place among the academic sciences of our century.
When Greek and Latin had once been restored to their true place in the natural system of the Aryan languages, their special treatment, too, became necessarily a different one. In grammar, for instance, scholars were no longer satisfied to give forms and rules, and to place what was irregular by the side of what was regular. They wished to know the reasons of the rules as well as of the exceptions; they asked why the forms were such as they were, and not otherwise; they required not only a logical, but also an historical foundation of grammar. People asked themselves for the first time, why so small a change as mensa and mensæ could express the difference between one and many tables; why a single letter, like r, could possess the charm of changing I love, amo, into I am loved, amor. Instead of indulging in general speculations on the logic of grammar, the riddles of grammar received their solution from a study of the historical development of language. For every language there was to be a historical grammar, and in this way a revolution was produced in philological studies to be compared only to the revolution produced in chemistry by the discoveries of Lavoisier, or in geology by the theories of Lyell. For instance, instead of attempting an explanation why the genitive singular and the ablative plural of the first and second declensions could express rest in a place—Romæ, at Rome; Tarenti, at Tarentum; Athenis, at Athens; Gabiis, at Gabii—one glance at the past history of these languages showed that these so-called genitives were not and never had been genitives, but corresponded to the old locatives in i and su in Sanskrit. No doubt, a pupil can be made to learn anything that stands in a grammar; but I do not believe that it can conduce to a sound development of his intellectual powers if he first learns at school the real meaning of the genitive and ablative, and then has to accept on trust that, somehow or other, the same cases may express rest in a place. A well-known English divine, opposed to reform in spelling, as in everything else, once declared that the fearful orthography of English formed the best psychological foundation of English orthodoxy, because a child that had once been brought to believe that t-h-r-o-u-g-h sounded like “through,” t-h-o-u-g-h like “though,” r-o-u-g-h like “rough,” would afterwards believe anything. Be that as it may, I do not consider that grammatical rules like those just quoted on the genitive and ablative, assuming the power of the locative, are likely to strengthen the reasoning powers of any schoolboy.
Even more pernicious to the growth of sound ideas was the study of etymology, as formerly carried on in schools and universities. Everything here was left to chance or to authority, and it was not unusual that two or three etymologies of the same word had to be learnt, as if the same word might have had more than one parent. Yet it is many years since Otfried Müller told classical scholars that they must either surrender the whole subject of the historical growth of language, etymology, and grammatical morphology, or trust in these matters entirely to the guidance of Comparative Philology. As a student at Leipzig, I lived to see old Gottfried Hermann quoting the paradigms of Sanskrit grammar in one of his last Programs; and Boeckh declared in 1850, at the eleventh meeting of German philologists, that, in the present state of the science of language, the grammar of the classical languages cannot dispense with the coöperation of comparative grammar. And yet there are scholars even now who would exclude the Science of Language from schools and universities. What gigantic steps truly scientific etymology has made in Greek and Latin, every scholar may see in the excellent works of Curtius and Corssen. The essential difference between the old and the new systems consists here, too, in this, that while formerly people were satisfied if they knew, or imagined they knew, from what source a certain word was derived, little value is now attached to the mere etymology of a word, unless at the same time it is possible to account, according to fixed phonetic laws, for all the changes which a word has undergone in its passage through Latin, Greek, and Sanskrit. How far this conscientiousness may be carried is shown by the fact that the best comparative philologists decline to admit, on phonetic grounds, the identity of such words as the Latin Deus, and the Greek Θεός, although the strongest internal arguments may be urged in favor of the identity of these words.1,A
Let us go on to Mythology. If mythology is an old dialect, outliving itself, and, on the strength of its sacred character, carried on to a new period of language, it is easy to perceive that the historical method of the Science of Language would naturally lead here to most important results. Take only the one fact, which no one at present would dare to question, that the name of the highest deity among the Greeks and Romans, Ζεύς, and Jupiter, is the same as the Vedic Dyaus, the sky, and the old German Zio, Old Norse Tyr, whose name survives in the modern names of Dienstag or Tuesday. Does not this one word prove the union of those ancient races? Does it not show us, at the earliest dawn of history, the fathers of the Aryan race, the fathers of our own race, gathered together in the great temple of nature, like brothers of the same house, and looking up in adoration to the sky as the emblem of what they yearned for, a father and a God. Nay, can we not hear in that old name of Jupiter, i.e., Heaven-Father, the true key-note which still sounds on in our own prayer, “Our Father which art in heaven,” and which imparts to these words their deepest tone, and their fullest import? By an accurate study of these words we are able to draw the bonds of language and belief even more closely together. You know that the nom. sing. of Ζεύς has the acute, and so has the nom. sing. of Dyaus; but the vocative of Ζεύς has the circumflex, and so has likewise the vocative of Dyaus in the Veda.2,B Formerly the accent might have been considered as something late, artificial, and purely grammatical: the Science of Language has shown that it is as old as language itself, and it has rightly called it the very soul of words. Thus even in these faint pulsations of language, in the changes of accent in Greek and Sanskrit, may we feel the common blood that runs in the veins of the old Aryan dialects.
History, too, particularly the most ancient history, has received new light and life from a comparative study of languages. Nations and languages were in ancient times almost synonymous, and what constitutes the ideal unity of a nation lies far more in the intellectual factors, in religion and language, than in common descent and common blood. But for that very reason we must here be most cautious. It is but too easily forgotten that if we speak of Aryan and Semitic families, the ground of classification is language, and language only. There are Aryan and Semitic languages, but it is against all rules of logic to speak, without an expressed or implied qualification, of an Aryan race, of Aryan blood, of Aryan skulls, and to attempt ethnological classification on purely linguistic grounds. These two sciences, the Science of Language and the Science of Man, cannot, at least for the present, be kept too much asunder; and many misunderstandings, many controversies, would have been avoided, if scholars had not attempted to draw conclusions from language to blood, or from blood to language. When each of these sciences shall have carried out independently its own classification of men and of languages, then, and then only, will it be time to compare their results; but even then, I must repeat, what I have said many times before, it would be as wrong to speak of Aryan blood as of dolichocephalic grammar.3
We have all accustomed ourselves to look for the cradle of the Aryan languages in Asia, and to imagine these dialects flowing like streams from the centre of Asia to the South, the West, and the North. I must confess that Professor Benfey’s protest against this theory seems to me very opportune, and his arguments in favor of a more northern, if not European, origin of the whole Aryan family of speech, deserve, at all events, far more attention than they have hitherto received.
For the same reasons it seems to me at least a premature undertaking to use the greater or smaller number of coincidences between two or more of the Aryan languages as arguments in support of an earlier or later separation of the people who spoke them. First of all, there are few points on which the opinions of competent judges differ more decidedly than when the exact degrees of relationship between the single Aryan languages have to be settled. There is agreement on one point only, viz., that Sanskrit and Zend are more closely united than any other languages. But though on this point there can hardly be any doubt, no satisfactory explanation of this extraordinary agreement has as yet been given. In fact, it has been doubted whether what I called the “Southern Division” of the Aryan family could properly be called a division at all, as it consisted only of varieties of one and the same type of Aryan speech. As soon as we go beyond Sanskrit and Zend, the best authorities are found to be in open conflict. Bopp maintained that the Slavonic languages were most closely allied to Sanskrit, an opinion shared by Pott. Grimm, on the contrary, maintained a closer relationship between Slavonic and German. In this view he was supported by Lottner, Schleicher, and others, while Bopp to the last opposed it. After this, Schleicher (as, before him, Newman in England) endeavored to prove a closer contact between Celtic and Latin, and, accepting Greek as most closely united with Latin, he proceeded to establish a Southwestern European division, consisting of Celtic, Latin, and Greek, and running parallel with the Northwestern division, consisting of Teutonic and Slavonic; or, according to Ebel, of Celtic, Teutonic, and Slavonic.
But while these scholars classed Greek with Latin, others, such as Grassmann and Sonne, pointed out striking peculiarities which Greek shares with Sanskrit, and with Sanskrit only, as, for instance, the augment, the voiceless aspirates, the alpha privativum (a, not an), the mâ and μή prohibitivum, the tara and τερο as the suffix of the comparative, and some others. A most decided divergence of opinion manifested itself as touching the real relation of Greek and Latin. While some regarded these languages not only as sisters, but as twins, others were not inclined to concede to them any closer relationship than that which unites all the members of the Aryan family. While this conflict of opinions lasts (and they are not mere assertions, but opinions supported by arguments), it is clear that it would be premature to establish any historical conclusions, such, for instance, as that the Slaves remained longer united with the Indians and Persians than the Greeks, Romans, Germans, and Celts; or, if we follow Professor Sonne, that the Greeks remained longer united with the Indians than the other Aryan nations. I must confess that I doubt whether the whole problem admits of a scientific solution. If in a large family of languages we discover closer coincidences between some languages than between others, this is no more than we should expect, according to the working of what I call the Dialectic Process. All these languages sprang up and grew and diverged, before they were finally separated; some retained one form, others another, so that even the apparently most distant members of the same family might, on certain points, preserve relics in common which were lost in all the other dialects, and vice versâ. No two languages, not even Lithuanian and Old Slavonic, are so closely united as Sanskrit and Zend, which share together even technical terms, connected with a complicated sacrificial ceremonial. Yet there are words occurring in Zend, and absent in Sanskrit, which crop up again sometimes in Greek, sometimes in Latin, sometimes in German.4,C As soon as we attempt to draw from such coincidences and divergences historical conclusions as to the earlier or later separation of the nations who developed these languages, we fall into contradictions like those which I pointed out just now between Bopp, Grimm, Schleicher, Ebel, Grassmann, Sonne, and others. Much depends, in all scientific researches, on seeing that the question is properly put. To me the question, whether the closer relations between certain independent dialects furnish evidence as to the successive times of their separation, seems, by its very nature, fruitless. Nor have the answers been at all satisfactory. After a number of coincidences between the various members of the Aryan family have been carefully collected, we know no more in the end than what we knew at first, viz., that all the Aryan dialects are closely connected with each other. We know—
1. That Slavonic is most closely united with German (Grimm, Schleicher);
2. That German is most closely united with Celtic (Ebel, Lottner);
3. That Celtic is most closely united with Latin (Newman, Schleicher);
4. That Latin is most closely united with Greek (Mommsen, Curtius);
5. That Greek is most closely united with Sanskrit (Grassmann, Sonne, Kern);
6. That Sanskrit is most closely united with Zend (Burnouf).
Let a mathematician draw out the result, and it will be seen that we know in the end no more than we knew at the beginning. Far be it for me to use a mere trick in arguing, and to say that none of these conclusions can be right, because each is contradicted by others. Quite the contrary. I admit that there is some truth in every one of these conclusions, and I maintain, for that very reason, that the only way to reconcile them all is to admit that the single dialects of the Aryan family did not break off in regular succession, but that, after a long-continued community, they separated slowly, and, in some cases, contemporaneously, from their family-circle, till they established at last, under varying circumstances, their complete national independence. This seems to me all that at present one may say with a good conscience, and what is in keeping with the law of development in all dialects.
If now we turn away from the purely philological results of the Science of Language, in order to glance at the advantages which other sciences have derived from it, we shall find that they consist mostly in the light that has been shed on obscure words and old customs. This advantage is greater than, at first sight, it might seem to be. Every word has its history, and the beginning of this history, which is brought to light by etymology, leads us back far beyond its first historical appearance. Every word, as we know, had originally a predicative meaning, and that predicative meaning differs often very considerably from the later traditional or technical meaning. This predicative meaning, however, being the most original meaning of the word, allows us an insight into the most primitive ideas of a nation.
Let us take an instance from jurisprudence. Pœna, in classical Latin, means simply punishment, particularly what is either paid or suffered in order to atone for an injury. (Si injuriam faxit alteri, viginti quinque æris pœnœ sunto, fragm. xii. tab.) The word agrees so remarkably, both in form and meaning, with the Greek ποινή, that Mommsen assigned to it a place in what he calls Græco-Italic ideas.5 We might suppose, therefore, that the ancient Italians took pœna originally in the sense of ransom, simply as a civil act, by which he who had inflicted injury on another was, as far as he and the injured person were concerned, restored in integrum. The etymology of the word, however, leads us back into a far more distant past, and shows us that when the word pœna was first framed, punishment was conceived from a higher moral and religious point of view, as a purification from sin; for pœna, as first shown by Professor Pott (and what has he not been the first to show?) is closely connected with the root pu, to purify. Thus we read in the “Atharva-veda,” xix. 33, 3:—
“Tvám bhû́mim átyeshi ójasâ
Tvám védyâm sîdasi cấrur adhvaré
Tvấm pavítram ṛshayo bhárantas
Tvám puníhi duritấni asmát.”
“Thou, O God of Fire, goest mightily across the earth; thou sittest brilliantly on the altar at the sacrifice. The prophets carry Thee as the Purifier; purify us from all misdeeds.”
From this root pu we have, in Latin, pūrus, and pŭtus, as in argentum purum putum, fine silver, or in purus putus est ipse, Plaut. Ps. 4, 2, 31. From it we also have the verb purgare, for purigare, to purge, used particularly with reference to purification from crime by means of religious observances. If this transition from the idea of purging to that of punishing should seem strange, we have only to think of castigare, meaning originally to purify, but afterwards in such expressions as verbis et verberibus castigare, to chide and to chasten.
I cannot convince myself that the Latin crimen has anything in common with κρίνειν. The Greek κρίνειν is no doubt connected with Latin cer-no, from which cribrum, sieve. It means to separate, to sift, so that κρῖμα may well signify a judgment, but not a crime or misdeed. Crīmen, as every scholar knows or ought to know, meant originally an accusation, not a crime, and, in spite of all appearances to the contrary, has nothing whatever in common with discrīmen, which means what separates two things, a difference, a critical point. In crimen venire means to get into bad repute, to be calumniated; in discrimine esse means to be in a critical and dangerous position.
It is one of the fundamental laws of etymology that in tracing words back to their roots, we have to show that their primary, not their secondary meanings agree with the meaning of the root. Therefore, even if crīmen had assumed in later times the meaning of judgment, yet its derivation from the Greek κρίνειν would have to be rejected, because it would explain the secondary only, but not the primary meaning of crīmen. Nothing is clearer than the historical development of the meanings of crīmen, beginning with accusation, and ending with guilt.
I believe I have proved that crīmen is really and truly the same word as the German Verleumdung, calumny.6 Verleumdung comes from Leumund, the Old High-German hliumunt, and this hliumunt is the exact representative of the Vedic śromata, derived from the root śru, to hear, cluere, and signifying good report, glory, the Greek κλέος, the Old High-German hruom. The German word Leumund can be used in a good and a bad sense, as good or evil report, while the Latin crī-men, for croe-men (like liber for loeber), is used only in malam partem. It meant originally what is heard, report, on dit, gossip, accusation; lastly, the object of an accusation, a crime, but never judgment, in the technical sense of the word.
The only important objection that could be raised against tracing crīmen back to the root śru, is that this root has in the Northwestern branch of the Aryan family assumed the form clu, instead of cru, as in κλέος, cliens, gloria, O.Sl. slovo, A.S. hlûd, loud, inclutus. I myself hesitated for a long time on account of this phonetic difficulty, nor do I think it is quite removed by the fact that Bopp (“Comp. Gr.” § 20) identified the German scrir-u-mês, we cry (instead of scriw-u-mês), with Sk. śrâv-ayâ-mas, we make hear; nor by the r in in-cre-p-are, in κράζω, as compared with κλάζω, nor even by the r in ἀ-κρο-ά-ομαι, which Curtius seems inclined to derive from śru. The question is whether this phonetic difficulty is such as to force us to surrender the common origin of śromata, hliumunt, and crīmen; but even if this should be the case, the derivation of crīmen from cerno or κρίνειν would remain as impossible as ever.
This will give you an idea in what manner the Science of Language can open before our eyes a period in the history of law, customs, and manners, which hitherto was either entirely closed, or reached only by devious paths. Formerly, for instance, it was supposed that the Latin word lex, law, was connected with the Greek λόγος. This is wrong, for λόγος never means law in the sense in which lex does. λόγος, from λέγειν, to collect, to gather, signifies, like κατάλογος, a gathering, a collection, an ordering, be it of words or thoughts. The idea that there is a λόγος, an order or law, for instance, in nature, is not classical, but purely modern. It is not improbable that lex is connected with the English word law, only not by way of the Norman loi. English law is A.S. lagu (as saw corresponds both to the German Sage and Säge), and it meant originally what was laid down or settled, with exactly the same conception as the German Gesetz. It has been attempted to derive the Latin lex, too, from the same root, though there is this difficulty, that the root of liegen and legen does not elsewhere occur in Latin. The mere disappearance of the aspiration would be no serious obstacle. If, however, the Latin lex cannot be derived from that root, we must, with Corssen, refer it to the same cluster of words to which ligare, to bind, obligatio, binding, and the Oscan ablative lig-ud belong, and assign to it the original meaning of bond. On no account can it be derived from legere, to read, as if it meant a bill first read before the people, and afterwards receiving legal sanction by their approval.
From these considerations we gain at least this negative result, that, before their separation, the Aryan languages had no settled word for law; and even such negative results have their importance. The Sanskrit word for law is dharma, derived from dhar, to hold fast. The Greek word is νόμος, derived from νέμειν, to dispense, from which Nemesis, the dispensing deity, and perhaps even Numa, the name of the fabulous king and lawgiver of Rome.
Other words might easily be added which, by the disclosure of their original meaning, give us interesting hints as to the development of legal conceptions and customs, such as marriage, inheritance, ordeals, and the like. But it is time to cast a glance at theology, which, more even than jurisprudence, has experienced the influence of the Science of Language. What was said with regard to mythology, applies with equal force to theology. Here, too, words harden, and remain unchanged longer even than in other spheres of intellectual life; nay, their influence often becomes greater the more they harden, and the more their original meaning is forgotten. Here it is most important that an intelligent theologian should be able to follow up the historical development of the termini technici and sacrosancti of his science. Not only words like priest, bishop, sacrament, or testament, have to be correctly apprehended in that meaning which they had in the first century, but expressions like λόγος, πνεῦμα ἅγιον, δικαιοσύνη have to be traced historically to the beginnings of Christianity, and beyond, if we wish to gain a conception of their full purport.
In addition to this, the Philosophy of Religion, which must always form the true foundation of theological science, owes it to the Science of Language that the deepest germs of the consciousness of God among the different nations of the world have for the first time been laid open. We know now with perfect certainty that the names, that is, the most original conceptions, of the Deity among the Aryan nations, are as widely removed from coarse fetichism as from abstract idealism. The Aryans, as far as the annals of their language allow us to see, recognized the presence of the Divine in the bright and sunny aspects of nature, and they, therefore, called the blue sky, the fertile earth, the genial fire, the bright day, the golden dawn their Devas, that is, their bright ones. The same word, Deva in Sanskrit, Deus in Latin, remained unchanged in all their prayers, their rites, their superstitions, their philosophies, and even to-day it rises up to heaven from thousands of churches and cathedrals,—a word which, before there were Brahmans or Germans, had been framed in the dark workshop of the Aryan mind.
That the natural sciences, too, should have felt the electric shock of our new science is not surprising, considering that man is the crown of nature, the apex to which all other forces of nature point and tend. But that which makes man man, is language. Homo animal rationale, quia orationale, as Hobbes said. Buffon called the plant a sleeping animal; living philosophers speak of the animal as a dumb man. Both, however, forget that the plant would cease to be a plant if it awoke, and that the brute would cease to be a brute the moment it began to speak. There is, no doubt, in language a transition from the material to the spiritual: the raw material of language belongs to nature, but the form of language, that which really makes language, belongs to the spirit. Were it possible to trace human language directly back to natural sounds, to interjections or imitations, the question whether the Science of Language belongs to the sphere of the natural or the historical sciences would at once be solved. But I doubt whether this crude view of the origin of language counts one single supporter in Germany. With one foot language stands, no doubt, in the realm of nature, but with the other in the realm of the spirit. Some years ago, when I thought it necessary to bring out as clearly as possible the much neglected natural element in language, I tried to explain in what sense the Science of Language had a right to be called the last and the highest of the natural sciences. But I need hardly say that I did not lose sight, therefore, of the intellectual and historical character of language; and I may here express my conviction that the Science of Language will yet enable us to withstand the extreme theories of the evolutionists, and to draw a hard and fast line between spirit and matter, between man and brute.
This short survey must suffice to show you how omnipresent the Science of Language has become in all spheres of human knowledge, and how far its limits have been extended, so that it often seems impossible for one man to embrace the whole of its vast domain. From this I wish, in conclusion, to draw some necessary advice.
Whoever devotes himself to the study of so comprehensive a science must try never to lose sight of two virtues: conscientiousness and modesty. The older we grow, the more we feel the limits of human knowledge. “Good care is taken,” as Goethe said, “that trees should not grow into the sky.” Every one of us can make himself real master of a small field of knowledge only, and what we gain in extent, we inevitably lose in depth. It was impossible that Bopp should know Sanskrit like Colebrooke, Zend like Burnouf, Greek like Hermann, Latin like Lachmann, German like Grimm, Slavonic like Miklosich, Celtic like Zeuss. That drawback lies in the nature of all comparative studies. But it follows by no means that, as the French proverb says, qui trop embrasse, mal étreint. Bopp’s “Comparative Grammar” will always mark an epoch in linguistic studies, and no one has accused the old master of superficiality. There are, in fact, two kinds of knowledge; the one which we take in as real nourishment, which we convert in succum et sanguinem, which is always present, which we can never lose; the other which, if I may say so, we put into our pockets, in order to find it there whenever it is wanted. For comparative studies the second kind of knowledge is as important as the first, but in order to use it properly, the greatest conscientiousness is required. Not only ought we, whenever we have to use it, to go back to the original sources, to accept nothing on trust, to quote nothing at second-hand, and to verify every single point before we rely on it for comparative purposes, but, even after we have done everything to guard against error, we ought to proceed with the greatest caution and modesty. I consider, for instance, that an accurate knowledge of Sanskrit is a conditio sine quâ non in the study of Comparative Philology. According to my conviction, though I know it is not shared by others, Sanskrit must forever remain the central point of our studies. But it is clearly impossible for us, while engaged in a scholarlike study of Sanskrit, to follow at the same time the gigantic strides of Latin, Greek, German, Slavonic, and Celtic philology. Here we must learn to be satisfied with what is possible, and apply for advice whenever we want it, to those who are masters in these different departments of philology. Much has of late been said of the antagonism between comparative and classical philology. To me it seems that these two depend so much on each other for help and advice that their representatives ought to be united by the closest ties of fellowship. We must work on side by side, and accept counsel as readily as we give it. Without the help of Comparative Philology, for instance, Greek scholars would never have arrived at a correct understanding of the Digamma—nay, a freer intercourse with his colleague, Bopp, would have preserved Bekker from several mistakes in his restoration of the Digamma in Homer. Latin scholars would have felt far more hesitation in introducing the old d of the ablative in Plautus, if the analogy of Sanskrit had not so clearly proved its legitimacy.
On the other hand, we, comparative philologists, should readily ask and gladly accept the advice and help of our classical colleagues. Without their guidance, we can never advance securely; their warnings are to us of the greatest advantage, their approval our best reward. We are often too bold, we do not see all the difficulties that stand in the way of our speculations, we are too apt to forget that, in addition to its general Aryan character, every language has its peculiar genius. Let us all be on our guard against omniscience and infallibility. Only through a frank, honest, and truly brotherly coöperation can we hope for a true advancement of knowledge. We all want the same thing; we all are etymologists—that is, lovers of truth. For this, before all things, the spirit of truth, which is the living spirit of all science, must dwell within us. Whoever cannot yield to the voice of truth, whoever cannot say, “I was wrong,” knows little as yet of the true spirit of science.
Allow me, in conclusion, to recall to your remembrance another passage from Niebuhr. He belongs to the good old race of German scholars. “Above all things,” he writes, “we must in all scientific pursuits preserve our truthfulness so pure that we thoroughly eschew every false appearance; that we represent not even the smallest thing as certain of which we are not completely convinced; that if we have to propose a conjecture, we spare no effort in representing the exact degree of its probability. If we do not ourselves, when it is possible, indicate our errors, even such as no one else is likely to discover; if, in laying down our pen, we cannot say in the sight of God, ‘Upon strict examination, I have knowingly written nothing that is not true;’ and if, without deceiving either ourselves or others, we have not presented even our most odious opponents in such a light only that we could justify it upon our death-beds—if we cannot do this, study and literature serve only to make us unrighteous and sinful.”
Few, I fear, could add, with Niebuhr: “In this I am convinced that I do not require from others anything of which a higher spirit, if He could read my soul, could convict me of having done the contrary.” But all of us, young as well as old, should keep these words before our eyes and in our hearts. Thus, and thus only, will our studies not miss their highest goal: thus, and thus only, may we hope to become true etymologists—i.e., true lovers, seekers, and, I trust, finders of truth.
NOTES.
NOTE A.
Θεός and Deus.text
That Greek θ does not legitimately represent a Sanskrit, Latin, Slavonic, and Celtic d is a fact that ought never to have been overlooked by comparative philologists, and nothing could be more useful than the strong protest entered by Windischmann, Schleicher, Curtius, and others, against the favorite identification of Sk. deva, deus, and θεός. Considering it as one of the first duties, in all etymological researches, that we should pay implicit obedience to phonetic laws, I have never, so far as I remember, quoted θεός as identical with deus, together with the other derivatives of the root div, such as Dyaus, Ζεύς, Jupiter, deva, Lith. deva-s, Irish día.
But with all due respect for phonetic laws, I have never in my own heart doubted that θεός belonged to the same cluster of words which the early Aryans employed to express the brightness of the sky and of the day, and which helped them to utter their first conception of a god of the bright sky (Dyaus), of bright beings in heaven, as opposed to the powers of night and darkness and winter (deva), and, lastly, of deity in the abstract.7 I have never become an atheist; and though I did not undervalue the powerful arguments advanced against the identity of deus and θεός, I thought that other arguments also possessed their value, and could not be ignored with impunity. If, with our eyes shut, we submit to the dictates of phonetic laws, we are forced to believe that while the Greeks shared with the Hindus, the Italians, and Germans the name for the bright god of the sky Zeus, Dyaus, Jovis, Zio, and while they again shared with them such derivatives as δῖος, heavenly, Sk. divyas, they threw away the intermediate old Aryan word for god, deva, deus, and formed a new one from a different root, but agreeing with the word which they had rejected in all letters but one. I suppose that even the strongest supporters of the atheistic theory would have accepted δεός, if it existed in Greek, as a correlative of deva and deus; and I ask, would it not be an almost incredible coincidence, if the Greeks, after giving up the common Aryan word, which would have been δοιϝός or δειϝός or δεϝός, had coined a new word for god from a different root, yet coming so near to δεϝός as θεϝός? These internal difficulties seem to me nearly as great as the external: at all events it would not be right to attempt to extenuate either.
Now I think that, though much has been said against θεός for δεϝός, something may also be said in support of δεϝός assuming the form of θεός. Curtius is quite right in repelling all arguments derived from Sk. duhitar = θυγάτηρ, or Sk. dvâr = θύρ-α; but I think he does not do full justice to the argument derived from φιάλη and φιαρός. The Greek φιάλη has been explained as originally πιϝάλη, the lost digamma causing the aspiration of the initial π. Curtius says: “This etymology of φιάλη is wrecked on the fact that in Homer the word does not mean a vessel for drinking, but a kind of kettle.” That is true, but the fact remains that in later Greek φιάλη means a drinking cup. Thus Pindar (“Isthm.,” v. 58) says:—
Ἄνδωκε δ’ αὐτῷ φέρτατος
οἰνοδόκον φιάλαν χρυσῷ πεφρικυῖαν Τελαμών,
which refers clearly to a golden goblet, and not a kettle. Besides, we have an exactly analogous case in the Sk. pâtram. This, too, is clearly derived from pâ, to drink, but it is used far more frequently in the sense of vessel in general, and its etymological meaning vanishes altogether when it comes to mean a vessel for something, a fit person. I see no etymology for φιάλη, except πιϝάλη, a drinking vessel.
Secondly, as to φιαρός, which is supposed to be the same as πιαρός, and to represent the Sanskrit pîvaras, fat, Curtius says that it occurs in Alexandrian poets only, that it there means bright, resplendent, and is used as an adjective of the dawn, while πιαρός means fat, and fat only. Against this I venture to remark, first, that there are passages where φιαρός means sleek, as in Theocr. ii. 21, φιαρωτέρα ὄμφακος ὠμᾶς, said of a young plump girl, who in Sanskrit would be called pîvarî; secondly, that while πῖαρ is used for cream, φιαρός is used as an adjective of cream; and, thirdly, that the application of φιαρός to the dawn is hardly surprising, if we remember the change of meaning in λιπαρός in Greek, and the application in the Veda of such words as ghṛta pratîka, to the dawn. Lastly, as in φιάλη, I see no etymology for φιαρός, except πιϝαρός.
I think it is but fair therefore to admit that θεός for δεϝός would find some support by the analogy of φιάλη for πιϝάλη, and of φιαρός for πιϝαρός. There still remain difficulties enough to make us cautious in asserting the identity of θεός and deus; but in forming our own opinion these difficulties should be weighed impartially against the internal difficulties involved in placing θεός as a totally independent word, by the side of deva and deus. And, as in φιάλη and φιαρός, may we not say of θεός also that there is no etymology for it, if we separate it from Ζεύς and δῖος, from Dyaus and divyas? Curtius himself rejects Plato’s and Schleicher’s derivation of θεός from θέω, to run: likewise C. Hoffmann’s from dhava, man; likewise Bühler’s from a root dhi, to think or to shine; likewise that of Herodotus and A. Göbel from θες, a secondary form of θε, to settle. Ascoli’s analysis is highly sagacious, but it is too artificial. Ascoli8 identifies θεός, not with deva, but with divyá-s. Divyás becoming διϝεός (like satya, ἐτεός), the accent on the last syllable would produce the change to δϝεό-ς, ϝ would cause aspiration in the preceding consonant and then disappear, leaving θεός = divyás. All these changes are just possible phonetically, but, as Curtius observes, the point for which the theists contend is not gained, for we should still have to admit that the Greeks lost the common word for god, deva and deus, and that they alone replaced it by a derivative divya, meaning heavenly, not bright.
Curtius himself seems in favor of deriving θεός from θες, to implore, which we have in θεσ-σάμενοι, θέσσαντο, πολύθεστος, etc. Θεός, taken as a passive derivative, might, he thinks, have the meaning of ἀρητός in πολυάρητος, and mean the implored being. I cannot think that this is a satisfactory derivation. It might be defended phonetically and etymologically, though I cannot think of any analogous passive derivatives of a root ending in s. Where it fails to carry conviction is in leaving unexplained the loss of the common Aryan word for deity, and in putting in its place a name that savors of very modern thought.
I think the strongest argument against the supposed aspirating power of medial v, and its subsequent disappearance, lies in the fact that there are so many words having medial v, which show no trace of this phonetic process (Curtius, p. 507). On the other hand, it should be borne in mind, that the Greeks might have felt a natural objection to the forms which would have rendered deva with real exactness, I mean δοιός or δέος, the former conveying the meaning of double, the latter of fear. A mere wish to keep the name for god distinct from these words might have produced the phonetic anomaly of which we complain; and, after all, though I do not like to use that excuse, there are exceptions to phonetic laws. No one can explain how ὄγδοος was derived from ὀκτώ or ἕβδομος from ἑπτά, yet the internal evidence is too strong to be shaken by phonetic objections. In the case of θεός and deus the internal evidence seems to me nearly as strong as in ὄγδοος and ἕβδομος, and though unwilling to give a final verdict, I think the question of the loss in Greek of the Aryan word for god and its replacement by another word nearly identical in form, but totally distinct in origin, should be left for the present an open question in Comparative Philology.