260 Rev. Dr. Watt on the different Kinds of Cuneiform Writing 
of holding private intercourse with each other, a sacred language, which conti- 
nued long unknown to all but the individuals of their own caste, and of which, as 
it would now appear, they took the principal framework, and part of the materials, 
from the above dialect? Still further, the close degree of congruity, in gram- 
matical structure, which has been observed between the Sanscrit of the Vedas and 
the language of the cuneiform legends in question, forbids our admitting so long 
an interval between those dialects, in the states of them exhibited in the compared 
records, as would suffice to account for the great difference in their ingredients 
by any process of alteration naturally effected in the course of time. It is true 
that the materials of a language are more liable to change than its forms, and 
that roots are altered more rapidly than inflexions; but yet a proportion between 
the natural rates of variation of the two classes of elements of words holds within 
certain bounds, which are here quite overpassed. The apparent inconsistency, 
however, of this case is at once removed, if the mode I have, in the second part 
of my Work, assigned for the formation of the Sanscrit, be conceded, namely, that 
it was made to consist of a selection of words from different foreign languages; 
since, according to this view of the matter, for every non-Persian term introduced, 
the corresponding Persian one was dropped and abandoned. In this way, and, I 
will venture to add, in this way alone, can be explained the extraordinary circum- 
stance of so many words of the legends in the first kind of cuneiform writing being 
lost, notwithstanding that their inflexions and grammatical structure are found 
still preserved very nearly in the dialect of the Vedas. 
4. In dissenting from the extreme antiquity which is very generally attri- 
buted to the Sanscrit and Zend, I am happy to find myself supported by Major 
Rawlinson, with regard to the latter dialect, in the following extract from his 
essay : * — although I conjecture the Zend to be a later language than that of the 
inscriptions, upon the debris of which, indeed, it was probably refined and syste- 
matized, yet I believe it to approach nearer to the Persian of the ante- Alexandrian 
ages than any other dialect of the family, except the Vedic Sanscrit, that is avail- 
able to modern research.”* But this deduction from the internal evidence of 
the case, with respect to one of the dialects in question, holds a@ fortiort with 
regard to the other, which is allowed to have a yet closer affinity to the language 
* Journal of the Royal Asiatic Society, vol. x. part i. pp. 8, 9. 
