1. Startseite
  2. Briefe
  3. Nr. 1176

Wilhelm von Humboldt an Sir Alexander Johnston, 10.06.1828 ("An Essay on the best Means of ascertaining the Affinities of Oriental Languages")


I have the honour to return you Sir James Mackintosh’s interesting memoir. It possesses (like every thing which comes from the pen of that gifted and ingenious writer) the highest interest; and the ideas which are so luminously developed in it have the more merit, if we consider, that, at the period when this memoir was published, philosophical notions on the study and nature of languages were rarer and more novel than they are at present.

I would, in the first place, observe, that the Royal Asiatic Society could not direct its efforts to a point more important, and more intimately connected with the national glory, than that of endeavouring to throw further light on the relations which subsist among the different Indian dialects. Since we cannot doubt that this part of Asia was the cradle of the arts and sciences at an extremely remote period, it would be highly interesting to ascertain with greater certainty whether the Sanscrit be a primitive idiom belonging to those countries, or whether, on the contrary, as most of the learned are at present inclined to believe, it was introduced as a foreign language into India; and if so, the country whence it originated would naturally follow in the course of inquiry. It is equally curious to determine whether the primitive languages of India are to be traced over the Indian archipelago in dialects differing little from each other, and whether we are to assign their origin to these islands or to the continent. Mr. Ellis’s paper on the Malayálam language, with which you were so good as to furnish me, contains assertions on the affinity of the Tamul language to the idioms of Java, which it would be very important to verify.

It must be confessed that these problems are extremely difficult to solve; and it is probable that we shall never arrive at results which are quite certain: we should, however, carry these researches as far as possible, and the difficulty of the undertaking ought not to deter, but rather to induce us to select the most solid and certain means of insuring success. This is more particularly the point to which I wish to direct your attention, since you have been pleased to ask my opinion respecting the methods proposed by Sir James Mackintosh. It would assuredly have been very desirable to execute his plan, at the period when it was formed; we should then by this time have had more complete information regarding the languages of India; and should perhaps have been in the possession of dialects, of the existence of which we are now ignorant. There do exist, however, some works, such as Sir James calls for. Not to mention printed books, I have myself seen in the library of the East-India Company a MS. collection of Sanscrit words, compared in great numbers with those of the other languages of India, made under the direction of Mr. Colebrooke.[b] Some distinguished authors, as for instance Mr. Campbell, in his Telugu Dictionary, have been at pains to mark from what foreign idiom such words are derived, as are not proper to the language of which they form a part; and if these works do not embrace all the Indian idioms, they have, on the other hand, the advantage of comprehending entire languages, or at least of not being confined to a limited number of expressions. In the present state of our knowledge of the languages of India, which is very different from that of 1806, and possessing, as we now do, grammars and dictionaries of most of these idioms, I should not advise our confining ourselves to a plan which can only give a very imperfect idea of each of them. We can, and ought, to go farther at the present day. I confess that I am extremely averse to the system which proceeds on the supposition that we can judge of the affinity of languages merely by a certain number of ideas expressed in the different languages which we wish to compare. I beg you will not suppose, however, that I am insensible to the value and utility of these comparisons: on the contrary, when they are well executed, I appreciate all their importance; but I can never deem them sufficient to answer the end for which they have been undertaken; they certainly form a part of the data to be taken into account in deciding on the affinity of languages, but we should never be guided by them alone, if we wish to arrive at a solid, complete, and certain conclusion. If we would make ourselves acquainted with the relation which subsists between two languages, we ought to possess a thorough and profound knowledge of each of them. This is a principle dictated alike by common sense and by that precision acquired by the habit of scientific research.

I do not mean to say, that, if we are unable to attain a profound knowledge of each idiom, we should on this account entirely suspend our judgment: I only insist on it that we should not prescribe to ourselves arbitrary limits, and imagine that we are forming our judgment on a firm basis, while it is in reality insufficient.

The method of comparing a certain number of words of one existing language with those of several others, has always the two-fold inconvenience of neglecting entirely the grammatical relations, as if the grammar was not as essential a part of the language as the words; and of taking from the language which we wish to examine isolated words, selected, not according to their affinities and natural etymology, but according to the ideas which they express. Sir James Mackintosh very justly observes, that the affinity of two languages is much better proved when whole families of words resemble each other, than when this is the case with single words only. But how shall we recognize families of words in foreign languages, if we only select from them two or three hundred isolated terms? There undoubtedly subsists among words of the same language an analogy of meanings and forms of combination easy to be perceived. It is from this analogy, considered in its whole extent, and compared with the analogy of the words of another language, that we discover the affinity of two idioms, as far as it is recognizable in their vocabularies. It is in this manner alone, that we recognize the roots and the methods by which each language forms its derivatives. The comparison of two languages requires, that we should examine whether, and in what degree, the roots and derivative terms are common to both. It is not, then, by terms expressive of general ideas; such as sun, moon, man, woman, &c., that we must commence the comparison of two languages, but by their entire dictionary critically explained. The simple comparison of a certain number of words, by reducing the examination of languages too much to a mere mechanical labour, often leads us to omit examining sufficiently the words which form the subjects of our comparison: and to avoid this defect, we are forced to enter deeply into all the minutiæ of grammar, separating the words from their grammatical affixes, and comparing only what is really essential to the expression of the idea which they represent. The words, of which we seek a translation in different languages, often cannot be rendered except by a compound term. Thus the sun in some languages is called the father, the author, the star, &c. of day. It is evident, that, in these cases, we no longer compare the same words, but words altogether different. To conclude: it is impossible to form a correct judgment on the resemblance of sounds without having carefully studied the system of sounds of each of the languages which we would compare. There occur often between different languages, and still more frequently between different dialects, regular transformations of letters, by which we can discover the identity of words that at first view seem to have but a very slight resemblance in sound. On the other hand, a great resemblance of sound in two words will sometimes prove nothing, or leave the judgment in great uncertainty, if it be not supported by a train of analogies for the permutation of the same letters. What I have remarked proves, as I think, that even if we confine ourselves to the comparison of a certain number of words in different languages, it is still necessary to enter more deeply into their structure, and to apply ourselves to the study of their grammar. But further, I am quite convinced that it is only by an accurate examination of the grammar of languages that we can pronounce a decisive judgment on their true affinities.

Languages are the true images of the modes in which nations think and combine their ideas. The manner of this combination represented by the grammar, is altogether as essential and characteristic as are the sounds applied to objects, that is to say, the words. The form of language being quite inherent in the intellectual faculties of nations, it is very natural that one generation should transmit theirs to that which follows it; while words, being simple signs of ideas, may be adopted by races altogether distinct. If I attach great importance, however, under this view, to the grammar of a language, I do not refer to the system of grammar in general, but to grammatical forms, considered with respect to their system and their sounds taken conjointly.

If two languages, such for instance as the Sanscrit and the Greek, exhibit grammatical forms which are identical in arrangement, and have a close analogy in their sounds, we have an incontestable proof that these two languages belong to the same family.

If, on the contrary, two languages do contain a great number of words in common, but have no grammatical identity, their affinity becomes a matter of great doubt; and if their grammars have, like those of the Basque and the Latin, an essentially different character, these two languages certainly do not belong to the same family. The words of the one have been merely transplanted into the other, which has nevertheless retained its primitive forms.

If I assert that, in order to prove the affinity of languages, we should pay attention to the employment of grammatical forms and to their sounds taken together, it is because I would affirm that they must be considered not only in the abstract but in the concrete. Some examples will render this clearer.

Several American languages have two plural forms in the first person, an exclusive and an inclusive form, according as we would include or exclude the person addressed. It has been thought that this peculiarity belonged exclusively to the American languages; but it is also found in the Mantchu, the Tamul, and in all the dialects of the South Sea Islands. All these languages have indeed this grammatical form in common; but it is only in the abstract. Each of them expresses it by a different sound: the identity of this form, therefore, does not furnish any proof of the affinity of these languages.

On the other hand, the Sanscrit infinitive, or rather the affixes  Sankskrit {tuṃ} and  Sankskrit {tu}, as in  Sankskrit {jetukāma} [c] „desirous of vanquishing,“ correspond as grammatical forms with the Latin supines, and there is at the same time a perfect identity of sound in these forms in the two languages, as the Latin supines terminate invariably in tum and tu. The striking conformity of the Sanscrit auxiliary verb to that of the Greek and Lithuanian languages, has been ingeniously developed by Professor Bopp. The Sanscrit  Sankskrit {veda}, the Greek οἶδα, and the Gothic vait, are evidently of the same origin. In all these three words there is a conformity both of sound and signification: but further; all the three verbal forms have these two peculiarities in common, that though preterites, they are used in a present sense, and that in all three the short radical vowel, which is retained in the plural, is changed to a long vowel in the singular. The Lithuanian weizdmi, I know , and the Sanscrit  Sankskrit {veidma} [d], shew clearly at first view that this word is not only the same in the two languages (as bos and beef in Latin and English), but that the two languages have, in the termination mi, modelled these words on the same grammatical form; for they not only mark the persons of the verb by inflexions added to the end of the root, but the affix of the first person singular is in both cases the syllable mi.

There is then in the examples adduced a conformity in grammatical use, and at the same time in sound; and it is impossible to deny that the languages which possess these forms must be of the same family.

The difference between the real affinity of languages, which presumes a filiation as it were among the nations who speak them, and that degree of relation which is purely historical, and only indicates temporary and accidental connexions among nations, is, in my opinion, of the greatest importance. Now it appears to me impossible ever to ascertain that difference merely by the examination of words; especially, if we examine but a small number of them.

It is perhaps too much to assert, that words pass from age to age and from nation to nation; that they arise also from connexions (which though secret, are common to all men) between sounds and objects, and that they thus establish a certain identity between all languages; while the manner of casting and arranging these words, that is to say, the grammar, constitutes the particular differences of dialects. This assertion, I repeat, is perhaps too bold, when expressed in this general way; yet I am strongly inclined to consider it correct, provided the expression grammar be not taken vaguely, but with a due regard to the sounds of grammatical forms. But whatever opinion may be entertained with respect to this manner of considering the difference of languages, it appears to me at all events demonstrated:

First, that all research into the affinity of languages, which does not enter quite as much into the examination of the grammatical system as into that of words, is faulty and imperfect; and,

Secondly, that the proofs of the real affinity of languages, that is to say, the question whether two languages belong to the same family, ought to be principally deduced from the grammatical system, and can be deduced from that alone; since the identity of words only proves a resemblance such as may be purely historical and accidental.

Sir James Mackintosh rejects the examination of grammar, for this reason, that languages which are evidently of the same stock have very different grammars. But we must not be misled by this phenomenon, although it is in itself quite true. The grammatical form of languages depends, on the one hand, it is true, upon the nature of these languages; but it also depends, on the other hand, upon the changes which they experience in the course of ages, and in consequence of historical revolutions. Out of these changes it has arisen, that languages of the same family have a different grammatical system, and that languages really distinct resemble each other in some degree. But the slightest examination will suffice to shew the real relations which subsist between those languages, especially if by following the plan above laid down we proceed to the examination of forms which are alike identical in their uses and in their sounds. It is thus that we discover without difficulty that the English language is of Germanic origin, and that the Persian belongs to the Sanscrit family of languages, notwithstanding the very great difference which exists between the grammars of these idioms.

It is generally believed, that the affinity of two languages is undeniably proved, if words that are applied to objects which must have been known to the natives ever since their existence, exhibit a great degree of resemblance, and to a certain extent this is correct. But, notwithstanding this, such a method of judging of the affinity of languages seems to me by no means infallible. It often happens, that even the objects of our earliest perceptions, or of the first necessity, are represented by words taken from foreign languages, and which belong to a different class. If we only examine the list furnished by Sir James Mackintosh, we shall find there such words as people, countenance, touch, voice, labour, force, power, marriage, spirit, circle, tempest, autumn, time, mountain, valley, air, vapour, herb, verdure, and others of the same kind. Now all these words being evidently derived from the Latin, as it was transformed after the fall of the Roman empire, we ought, judging from these words, rather to assign to the English an origin similar to that of the Roman languages than to that of the German.

If what I have here advanced be well founded, it appears to me easy to point out the system which the Royal Asiatic Society would do well to pursue, in order to complete our knowledge of the Indian languages, and to resolve the grand problem which they present to the minds of philologists who endeavour to discover the origin and the filiation of languages.

It would be proper to commence by examining the country geographically, taking a review of every part of India, in order to know exactly in what parts we are still in want of sufficient materials to determine the nature of their idioms. Where deficiencies are discovered, efforts should be used for their supply, by encouraging those persons who are already employed on those languages, or may intend studying them, to form grammars and dictionaries, and to publish the principal works existing in these languages, for which every facility should be afforded them. If materials to a certain extent were thus collected, we should unquestionably not want men who would be able to deduce from them conclusions from which to prepare a critical view of the affinity of the Indian languages, and to determine, as far as the data which we might possess would admit, the manner in which the Sanscrit and other languages of India and its islands have reciprocally acted upon each other. I assume that the learned of the Continent would take their share in this work, M. E. Burnouf, of Paris, having already commenced a series of papers on the subject in the Nouveau Journal Asiatique.

There exists in England a vast quantity of manuscript materials relating to these languages. Dr. B. Babington, for instance, possesses alphabets altogether unknown in Europe up to the present time. In England, also, the great advantage is possessed of being able to direct works upon these languages to be undertaken in India itself, and to guide such labours by plans sent from this country. In India these are living languages, and literary men of the very nations in which they are spoken may be employed in the researches we wish to forward. No other nation possesses so valuable an advantage. It is important to profit by it. The deficiencies in our knowledge are numerous and evident. We possess scarcely any thing upon the Malayalim; and are in want of a printed dictionary of the Tamul. But while we keep this object strictly in view, and work upon a fixed plan, we shall insensibly fill up these vacancies. It is certainly difficult to find men who both can and will engage in a work like this, but they are undoubtedly to be found. Thus Dr. Babington has mentioned Mr. Whish[e] to me, as being profoundly acquainted with the Malayalim, and as being already employed in making it better known in Europe. Solid labours upon languages are, in their nature, slow. In an enterprize so vast as that of examining to the utmost possible extent each of the numerous languages of India, progress can only be made insensibly and step by step. But learned societies afford this advantage, that the same labour can be continued through a long series of years; and complete and perfect works upon two or three idioms are certainly preferable to notions, more or less superficial, upon all the dialects of India, hastily put forth for the purpose of coming at once to a general conclusion.

These, Sir, are my ideas upon the subject, upon which you wished to have my opinion. It is only in compliance with your request, that I have ventured to lay them before you; for I am well aware how much better able the distinguished members of the Royal Asiatic Society are to form a judgment of, and give an opinion upon, this matter than I am.

I request you, Sir, to accept the assurance of my highest respect.
(Signed) Humboldt.
London, June 10, 1828.


    1. a |Editor| Betitelt: "An Essay on the best Means of ascertaining the Affinities of Oriental Languages, by Baron William Humboldt, For. M.R.A.S. Contained in a Letter addressed to Sir Alexander Johnston, Knt., V.P.R.A.S. Read June 14, 1828." – Der Brief, ursprünglich auf Französisch verfasst, wurde von Benjamin Guy Babington ins Englische übersetzt, und erschien unter dem englischen Titel in den Transactions of the Royal Asiatic Society of Great Britain and Ireland des Jahres 1830. [FZ]
    2. b |Editor| Anmerkung von Colebrooke (S. 221): "The work to which allusion is made by Baron William de Humboldt, in the passage where I am named, was undertaken by me in furtherance of the views developed by Sir James Mackintosh. I thought that a more copious comparative vocabulary than he had proposed, would be practically useful; and would be instructive in more points of view than he had contemplated. Accordingly, at my instance, a Sanscrit vocabulary and a Persian one were printed with blank half pages, and distributed among gentlemen, whose situations were considered to afford the opportunity of having the blank column filled up, by competent persons, with a vocabulary of a provincial language. Vocabularies of the same vernacular tongue by a Pandit and a Munshi, would serve to correct mutually, and complete the information sought from them. Very few answers, however, were received: indeed scarcely any, except from Dr. Buchanan Hamilton. The compilation, to which Baron de Humboldt refers, comprises as many as I succeeded in collecting. H.T.C."
    3. c |Editor| Hier korrekt, im Sonderdruck von 1828 (und einem Teil der bereits gedruckten Auflage) als  Sankskrit . Zur Korrektur siehe auch den Brief von Friedrich August Rosen an Wilhelm von Humboldt vom 9. Januar 1829 und den Errata-Zusatz auf S. XII im 1830 vollständig gedruckt vorliegenden Band der Transactions. [FZ]
    4. d |Editor| Hier korrekt, im Sonderdruck von 1828 als  Sankskrit .
    5. e |Editor| Charles Matthew Whish (1794–1833), englischer ziviler Angestellter in der Madras-Niederlassung der East India Company. [FZ]

    Wilhelm von Humboldt an Sir Alexander Johnston, 10.06.1828 ("An Essay on the best Means of ascertaining the Affinities of Oriental Languages"). In: Wilhelm von Humboldt: Online-Edition der Sprachwissenschaftlichen Korrespondenz. Berlin. Version vom 15.03.2023. URL: https://wvh-briefe.bbaw.de/1176


    Dieses Dokument als TEI-XML herunterladen


    Frühere Version des Dokuments in der archivierten Webansicht ansehen