Lectures on the Science of Language

by F. Max Müller

We know certain forms of language which correspond to certain forms of thought. We know that the subject must assume the form of the nominative, the object that of the accusative. We know that the more remote object may be put in the dative, and that the predicate, in its most general form, may be rendered by the genitive. We are taught that whereas in English the genitive is marked by a final s, or by the preposition of, it is in Greek expressed by a final ος, in Latin by is. But what this ος and is represent, why they should have the power of changing a nominative into a genitive, a subject into a predicate, remains a riddle. It is self-evident that each language, in order to be a language, must be able to distinguish the subject from the object, the nominative from the accusative. But how a mere change of termination should suffice to convey so material a distinction would seem almost incomprehensible. If we look for a moment beyond Greek and Latin, we see that there are in reality but few languages which have distinct forms for these two categories of thought. Even in Greek and Latin there is no outward distinction between the nominative and accusative of neuters. The Chinese language, it is commonly said, has no grammar at all, that is to say, it has no inflections, no declension and conjugation, in our sense of these words; it makes no formal distinction of the various parts of speech, noun, [pg 118] verb, adjective, adverb, &c. Yet there is no shade of thought that cannot be rendered in Chinese. The Chinese have no more difficulty in distinguishing between “James beats John,” and “John beats James,” than the Greeks and Romans or we ourselves. They have no termination for the accusative, but they attain the same by always placing the subject before, and the object after the verb, or by employing words, before or after the noun, which clearly indicate that it is to be taken as the object of the verb.[101] There are other languages [pg 119] which have more terminations even than Greek and Latin. In Finnish there are fifteen cases, expressive of every possible relation between the subject and the object; but there is no accusative, no purely objective case. In English and French the distinctive terminations of the nominative and accusative have been worn off by phonetic corruption, and these languages are obliged, like Chinese, to mark the subject and object by the collocation of words. What we learn therefore at school in being taught that rex in the nominative becomes regem in the accusative, is simply a practical rule. We know when to say rex, and when to say regem. But why the king as a subject should be called rex, and as an object regem, remains entirely [pg 120] unexplained. In the same manner we learn that amo means I love, amavi I loved; but why that tragical change from love to no love should be represented by the simple change of o to avi, or, in English, by the addition of a mere d, is neither asked nor answered.

Now if there is a science of language, these are the questions which it will have to answer. If they cannot be answered, if we must be content with paradigms and rules, if the terminations of nouns and verbs must be looked upon either as conventional contrivances or as mysterious excrescences, there is no such thing as a science of language, and we must be satisfied with what has been called the art (τέχνη) of language, or grammar.

Before we either accept or decline the solution of any problem, it is right to determine what means there are for solving it. Beginning with English we should ask, what means have we for finding out why I love should mean I am actually loving, whereas I loved indicates that that feeling is past and gone? Or, if we look to languages richer in inflections than English, by what process can we discover under what circumstances amo, I love, was changed, through the mere addition of an r, into amor, expressing no longer I love, but I am loved? Did declensions and conjugations bud forth like the blossoms of a tree? Were they imparted to man ready made by some mysterious power? Or did some wise people invent them, assigning certain letters to certain phases of thought, as mathematicians express unknown quantities by freely chosen algebraic exponents? We are here brought at once face to face with the highest and most difficult problem of our science, the origin of language. But it will be well [pg 121] for the present to turn our eyes away from theories, and fix our attention at first entirely on facts.

Let us keep to the English perfect, I loved, as compared with the present, I love. We cannot embrace at once the whole English grammar, but if we can track one form to its true lair, we shall probably have no difficulty in digging out the rest of the brood. Now, if we ask how the addition of a final d could express the momentous transition from being in love to being indifferent, the first thing we have to do, before attempting any explanation, would be to establish the earliest and most original form of I loved. This is a rule which even Plato recognized in his philosophy of language, though, we must confess, he seldom obeyed it. We know what havoc phonetic corruption may make both in the dictionary and the grammar of a language, and it would be a pity to waste our conjectures on formations which a mere reference to the history of language would suffice to explain. Now a very slight acquaintance with the history of the English language teaches us that the grammar of modern English is not the same as the grammar of Wycliffe. Wycliffe's English again may be traced back to what, with Sir Frederick Madden, we may call Middle English, from 1500 to 1330; Middle English to Early English, from 1330 to 1230; Early English to Semi-Saxon from 1230 to 1100; and Semi-Saxon to Anglo-Saxon.[102] It is evident that if we are to discover the original intention of the syllable which changes I love into I loved, we must consult the original form of that syllable wherever we can find it. We should never [pg 122] have known that priest meant originally an elder, unless we had traced it back to its original form presbyter, in which a Greek scholar at once recognizes the comparative of presbys, old. If left to modern English alone, we might attempt to connect priest with praying or preaching, but we should not thus arrive at its true derivation. The modern word Gospel conveys no meaning at all. As soon as we trace it back to the original Goddspell, we see that it is a literal translation of Evangelium, or good news, good tidings.[103] Lord would be nothing but an empty title in English, unless we could discover its original form and meaning in the Anglo-Saxon hlafford, meaning a giver of bread, from hlaf, a loaf, and ford, to give.

But even after this is done, after we have traced a modern English word back to Anglo-Saxon, it follows by no means that we should there find it in its original form, or that we should succeed in forcing it to disclose its original intention. Anglo-Saxon is not an original or aboriginal language. It points by its very name to the Saxons and Angles of the continent. We have, therefore, to follow our word from Anglo-Saxon through the various Saxon and Low-German dialects, till we arrive at last at the earliest stage of German which is within our reach, the Gothic of the fourth century after Christ. Even here we cannot rest. For, although we cannot trace Gothic back to any earlier Teutonic language, we see at once that Gothic, too, is a modern language, and that it must have passed [pg 123] through numerous phases of growth before it became what it is in the mouth of Bishop Ulfilas.

What then are we to do?—We must try to do what is done when we have to deal with the modern Romance languages. If we could not trace a French word back to Latin, we should look for its corresponding form in Italian, and endeavor to trace the Italian to its Latin source. If, for instance, we were doubtful about the origin of the French word for fire, feu, we have but to look to the Italian fuoco, in order to see at once that both fuoco and feu are derived from the Latin focus. We can do this, because we know that French and Italian are cognate dialects, and because we have ascertained beforehand the exact degree of relationship in which they stand to each other. Had we, instead of looking to Italian, looked to German for an explanation of the French feu, we should have missed the right track; for the German feuer, though more like feu than the Italian fuoco, could never have assumed in French the form feu.

Again, in the case of the preposition hors, which in French means without, we can more easily determine its origin after we have found that hors corresponds with the Italian fuora, the Spanish fuera. The French fromage, cheese, derives no light from Latin. But as soon as we compare the Italian formaggio,[104] we see that formaggio and fromage are derived from forma; cheese being made in Italy by keeping the milk in small baskets or forms. Feeble, the French faible, is clearly derived from Latin; but it is not till we see the Italian fievole that we are reminded of the Latin flebilis, tearful. We should never have found the etymology, [pg 124] that is to say the origin, of the French payer, the English to pay, if we did not consult the dictionary of the cognate dialects, such as Italian and Spanish. Here we find that to pay is expressed in Italian by pagare, in Spanish by pagar, whereas in Provençal we actually find the two forms pagar and payar. Now pagar clearly points back to Latin pacare, which means to pacify, to appease. To appease a creditor meant to pay him; in the same manner as une quittance, a quittance or receipt, was originally quietantia, a quieting, from quietus, quiet.

If, therefore, we wish to follow up our researches,—if, not satisfied with having traced an English word back to Gothic, we want to know what it was at a still earlier period of its growth,—we must determine whether there are any languages that stand to Gothic in the same relation in which Italian and Spanish stand to French;—we must restore, as far as possible, the genealogical tree of the various families of human speech. In doing this we enter on the second or classificatory stage of our science; for genealogy, where it is applicable, is the most perfect form of classification.

Before we proceed to examine the results which have been obtained by the recent labors of Schlegel, Humboldt, Bopp, Burnouf, Pott, Benfey, Prichard, Grimm, Kuhn, Curtius, and others in this branch of the science of language, it will be well to glance at what had been achieved before their time in the classification of the numberless dialects of mankind.

The Greeks never thought of applying the principle of classification to the varieties of human speech. They only distinguished between Greek on one side, [pg 125] and all other languages on the other, comprehended under the convenient name of “Barbarous.” They succeeded, indeed, in classifying four of their own dialects with tolerable correctness,[105] but they applied the term “barbarous” so promiscuously to the other more distant relatives of Greek, (the dialects of the Pelasgians, Carians, Macedonians, Thracians, and Illyrians,) that, for the purposes of scientific classification, it is almost impossible to make any use of the statements of ancient writers about these so-called barbarous idioms.[106]