PREFACE.

It is much to be regretted that all writers who have entered into etymological discussions, or who have employed etymology as the medium of other researches, should have permitted their judgments to be guided and influenced by some favourite hypothesis. For, however anxious an author may be to discover truth, still, if his mind be occupied by preconceived opinions, it is impossible for him to avoid giving more attention and more force to such circumstances as support these opinions, than to such as oppose them. Too many writers, also, in conducting an argument respecting the origin and affinity of nations, or even respecting their idolatry, have indulged in such absurdity of etymologies, and such mis-selection and perversion of authorities, as must render their love of truth extremely questionable. The ridicule, therefore, that is thrown on etymology, and the distrust with which it is received as proof, are the natural consequences of its having been employed so improperly. But, as it is illogical to argue from the abuse to the use, no work ought to be condemned on mere inspection of its titlepage, because erroneous methods have been adopted in the previous discussion of the same subject.

The following RESEARCHES, also, whatever other defects may be attributable to them, are at least free from the spirit of hypothesis.

For, having occasion to compile a Maratha dictionary, I amused myself, while collecting materials for that work, in noting down the Sanscrit words which I recognised as belonging to any language with which I was acquainted; and it was not until I had collected five hundred such words, that I began to enquire into the causes which could have introduced them into five distinct languages. Until then I had acquiesced in the correctness of the usual opinions entertained respecting the origin and affinity of languages, although doubts of their justness had often occurred to me. But, on further examining the subject, I found that none of the systems which had been proposed could adequately explain the causes of that intimate connection which must have existed, at some remote period, between a people speaking Sanscrit and the ancestors of the Greeks, Romans, and Goths. It was, therefore, necessary to discover some more probable and satisfactory explanation of so remarkable a circumstance, and I accordingly stated the conclusions to which its investigation had led me, in a paper which I laid before the Literary Society of Bombay, in November, 1822. This paper, however, I afterwards withdrew, as it occurred to me that neither its limits allowed the subject to be fully discussed, nor had I myself obtained all the information respecting it which was requisite. For I conceive it incumbent on every writer to ascertain, as far as possible, what may have been previously published on the topic which he intends to discuss. But the want of books prevented me, for some time, from having it in my power to enlarge and improve the paper just mentioned in the manner that I wished. Having at length, however, made myself, I believe, sufficiently acquainted with the principal opinions which prevail respecting the origin and affinity of languages, I now venture to lay the following RESEARCHES before the public.

PREFACE.

The original object of this work was merely to exhibit the remarkable affinity which exists between the Greek, Latin, Persian, Gothic, and Sanscrit languages, and to explain the causes which had, in my opinion, produced it. But, on further consideration, it appeared to me that neither of these points could be satisfactorily demonstrated, until the prevailing hypothesis respecting the existence of a primitive tongue, and respecting the origin of the Greeks, Romans, and Goths, had been first examined and refuted. I have, in consequence, been obliged not only to enter into a review of these subjects on which so much has been already written, but, also, in considering them, to differ in opinion, less or more, from every author by whom they have been previously discussed. But no person has hitherto applied a competent knowledge of Arabic, Persian, and Sanscrit to etymological purposes, and from new data, therefore, it may be permitted to draw new conclusions.

One writer, indeed, Dr. A. Murray, in his History of European Languages, has pretended to an acquaintance with Sanscrit and Persian; but the very erroneous judgment of the origin and nature of these languages which he has expressed, evinces that his knowledge of them must have been extremely superficial. He has himself, at the same time, admitted that he had not the Sanscrit language completely before him*; nor was it possible that he could, as no Sanscrit dictionary was then published. But Persian was perfectly accessible in grammars, dictionaries, and editions of works containing together the original text and its translation; and the ignorance, therefore, of this language betrayed by Dr. Murray is altogether inexcusable. It is not, however, so much the errors contained in this work, as the dogmatic tone in which the opinions are expressed, that are Hist. of European Languages, vol. ii. p. 381.

particularly censurable. For nothing but the most indisputable proofs
could warrant such positive assertions as these: —“
"The Medes, Per-
sians, and Indians spoke the same language. They were allied to
one another in the degree of the Ionic and Doric Greeks. This
important fact is established, 1. by the close resemblance of the
ancient Median names to the Sanscrit in form and sense; 2. by the
perfect coincidence of the remains of the Zend with the Sanscrit ;
3. by the easy derivation of almost every modern Persian word (the
Arabic terms excepted) from the Sanscrit.*.... The modern Persic
is Sanscrit, humbled and corrupted in a high degree.
It is simple,
elegant, perspicuous; but, at the same time, not capable of greater
powers of expression, than those which genius may impart to any
dialect, however defective by nature.†.... Ocular inspection, assisted
by such knowledge as the comparison requires, demonstrates the
ancient identity of the Sanscrit and Chaldee letters." That is, an
alphabet composed of fifty-two letters was derived from one con-
sisting of twenty-two letters only! The reputation acquired by Dr.
Murray as a philologist has induced me to notice his work here, in
order to explain the reason why I have scarcely ever quoted it in
the following pages, either for the purpose of approbation or refutation.
But for the first of these purposes it is much too erroneous; and, with
regard to the latter, I perfectly agree in opinion with Pinkerton, that
to confute absolute nonsense is surely as ridiculous as to write it. That
the reader, however, may not consider these remarks as too harsh,
I will leave it to him to decide whether that philologist is entitled
to any attention who, in the very commencement of his work, makes
such an assertion as this: -" By a careful study of the Anglo-Saxon,

* Hist. of European Languages, vol. ii. p. 222.
+ Ibid. p. 391.

+ Ibid. p. 227.

PREFACE.

Visigothic, and the elder English writers, more knowledge may be obtained of the original structure of the Greek, Latin, Celtic, or Sanscrit, than the deepest erudition can possibly supply!!"*

With respect to the conclusions contained in this work, which are deduced from etymological premises, the principles on which they depend are sufficiently explained in the Second Part. I shall here, therefore, merely observe that, in comparing together the words of any two languages, I conceive that correspondence in signification and in sound, subject to such slight permutations in the letters and slight contractions of the syllables as are proved to be admissible on clear and fixed principles, are the only criteria by which the identity of the words compared can be determined. These RESEARCHES, therefore, differ materially from other etymological works: because they contain no wearisome discussions respecting the changes which words may have undergone in passing from one language into another; nor any tedious reasoning to prove that some particular word in one language, notwithstanding dissimilarity of sound and meaning, may still be identical with some other term of another language. For the Comparative Table inserted in Part II. is the pièce justificative of the whole work; and as all the words compared together in it correspond in meaning, except in a few instances which I have noted at the bottom of the page, the reader, if unacquainted with the languages compared, has merely to determine whether the agreement of the words in sound is sufficient to prove their identity. Should he, then, be convinced that 900 Sanscrit words have passed into five

Hist. of European Languages, vol. i. p. 17.

No words can better characterise Dr. Murray's work, than those which he has himselt applied to Mr. Bryant's Analysis of Ancient Mythology: -"A fanciful work, of which the etymological part is false, the historical dubious, and the theoretical imaginary."—Vol. ii. p. 223.