ciple of transliteration must be to represent letter by letter, not sound by sound.
Which letter in our physiological alphabet should be fixed upon as the fittest representative of another letter in Arabic or Sanskrit, in Hindustani or Canarese, depends on a special agreement. If we found that ह in Sanskrit had in most respects the nature of the guttural semi-vowel, we should have to write it h, even though in some words it may represent the guttural flatus. If ע in Hebrew can be proved to have been originally the same guttural semi-vowel, it will have to be written h, even where we know that it was pronounced as guttural media aspirata (ghain), or not pronounced at all. Likewise, if English were to be transliterated with our alphabet, we should not adopt any of the principles of the "Fonetic Nus;" but here also, if the letter h had been fixed upon as on the whole the fittest representative of the English letter h, we should have to write it even where it was not pronounced at all, as in honest.
It will be the duty of Academies and scientific societies to settle, for the principal languages, which letter of their alphabet could best be expressed by a certain letter in the Missionary alphabet.
The first question to ask would be, taking a type, for instance, of the Sanskrit alphabet, "What is its most usual and most original value?" If that be fixed, the next question would be, "Is there another type which has a better claim to this value?" If so, their claims must be weighed and adjusted; and after it is settled, and after the physiological category is found under which this Sanskrit type would have its proper place, we should then have to look for the exponent of this physiological category in the Missionary alphabet, and henceforth always to transliterate the one by the other.
The following lists will show how Hebrew, Arabic, and Sanskrit might be transliterated. I am aware that objections will be raised on several points, because the original character of several Hebrew and Sanskrit letters has been frequently controverted. If the disputed value of these letters can be clearly settled by argument, be it so; and as soon as it is, there will never be a difficulty to find the exponent of that physiological category to which it has been adjudged. But if there is no chance of settling it by argument, it must be settled by authority or agreement; for, of two views which are
