If we go back to earlier days, we find that Sanskrit scholars who had discovered that one of the names of the god of love in Bengali was Dipuc, i.e. the inflamer, derived from it by inversion the name of the god of love in Latin, Cupid. Sir William Jones identified Janus with the Sanskrit Gaṇeśa, i.e., lord of hosts,[9] and even later scholars allowed themselves to be tempted to see the Indian prototype of Ganymedes in the Kaṇva-medhâtithi or Kaṇva-mesha of the Veda.[10]
After the phonetic laws of each language had been more carefully elaborated, it was but too frequently forgotten that words have a history as well as a growth, and that the history of a word must be explored first, before an attempt is made to unravel its growth. Thus it was extremely tempting to derive paradise from the Sanskrit paradeśa. The compound para-deśa was supposed to mean the highest or a distant country, and all the rest seemed so evident as to require no further elucidation. Paradeśa, however, does not mean the highest or a distant country in Sanskrit, but is always used in the sense of a foreign country, an enemy’s country. Further, as early as the Song of Solomon (iv. 13), the word occurs in Hebrew as pardés, and how it could have got there straight from Sanskrit requires, at all events, some historical explanation. In Hebrew the word might have been borrowed from Persian, but the Sanskrit word paradeśa, if it existed at all in Persian, would have been paradaesa, the s being a guttural, not a dental sibilant. Such a compound, however, does not exist in Persian, and therefore the Sanskrit word paradeśa could not have reached Hebrew viâ Persia.
It is true, nevertheless, that the ancient Hebrew word pardés is borrowed from Persian, viz.: from the Zend pairidaêza, which means circumvallatio, a piece of ground inclosed by high walls, afterwards a park, a garden.[11] The root in Sanskrit is DIH or DHIH (for Sanskrit h is Zend z), and means originally to knead, to squeeze together, to shape. From it we have the Sanskrit dehî, a wall, while in Greek the same root, according to the strictest phonetic rules, yielded τοῖχος, wall. In Latin our root is regularly changed into fig, and gives us figulus, a potter, figura, form or shape, and fingere. In Gothic it could only appear as deig-an, to knead, to form anything out of soft substances; hence daig-s, the English dough, German Deich.
But the Greek παράδεισος did not come from Hebrew, because here again there is no historical bridge between the two languages. In Greek we trace the word to Xenophon, who brought it back from his repeated journeys in Persia, and who uses it in the sense of pleasure-ground, or deer park.[12]
Lastly, we find the same word used in the LXX., as the name given to the garden of Eden, the word having been borrowed either a third time from Persia, or taken from the Greek, and indirectly from the works of Xenophon.
This is the real history of the word. It is an Aryan word, but it does not exist in Sanskrit. It was first formed in Zend, transferred from thence as a foreign word into Hebrew and again into Greek. Its modern Persian form is firdaus.
All this is matter of history rather than philology. Yet we read in one of the best classical dictionaries: “The root of παράδεισος appears to be Semitic, Arab. firdaus, Hebr. pardês: borrowed, also, in Sanskrit paradêśa.”[13] Nearly every word is wrong.
From the same root DIH springs the Sanskrit word deha, body; body, like figure, being conceived as that which is formed or shaped. Bopp identified this deha with Gothic leik, body, particularly dead body, the modern German Leiche and Leichnam, the English lich in lich-gate. In this case the master of Comparative Philology disregarded the phonetic laws which he had himself helped to establish. The transition of d into l is no doubt common enough as between Sanskrit, Latin, and Greek, but it has never been established as yet on good evidence as taking place between Sanskrit and Gothic. Besides, the Sanskrit h ought in Gothic to appear as g, as we have it in deig-s, dough, and not by a tenuis.
Another Sanskrit word for body is kalevara, and this proved again a stumbling-block to Bopp, who compares it with the Latin cadaver. Here one might plead that l and d are frequently interchanged in Sanskrit and Latin words, but, as far as our evidence goes at present, we have no doubt many cases where an original Sanskrit d is represented in Latin by l, but no really trustworthy instance in which an original Sanskrit l appears in Latin as d. Besides, the Sanskrit diphthong e cannot, as a rule, in Latin be represented by long â.
If such things could happen to Bopp, we must not be too severe on similar breaches of the peace committed by classical scholars. What classical scholars seem to find most difficult to learn is that there are various degrees of certainty in etymologies even in those proposed by our best comparative scholars, and that not everything that is mentioned by Bopp, or Pott, or Benfey as possible, as plausible, as probable, and even as more than probable, ought, therefore, to be set down, for instance, in a grammar or dictionary, as simply a matter of fact. With certain qualifications, an etymology may have a scientific value; without those qualifications, it may become not only unscientific but mischievous. Again, nothing seems a more difficult lesson for an etymologist to learn than to say, I do not know. Yet to my mind, nothing shows, for instance, the truly scholarlike mind of Professor Curtius better than the very fact for which he has been so often blamed, viz.: his passing over in silence the words about which he has nothing certain to say.