Nor can you quote in favour of your view--of the real being known through the unreal--the instance of the stroke and the letter. The letter being apprehended through the stroke (i.e. the written character) does not furnish a case of the real being apprehended through the unreal; for the stroke itself is real.--But the stroke causes the idea of the letter only in so far as it is apprehended as being a letter, and this 'being a letter' is untrue!--Not so, we rejoin. If this 'being a letter' were unreal it could not be a means of the apprehension of the letter; for we neither observe nor can prove that what is non-existent and indefinable constitutes a means.--Let then the idea of the letter constitute the means!--In that case, we rejoin, the apprehension of the real does not spring from the unreal; and besides, it would follow therefrom that the means and what is to be effected thereby would be one, i.e. both would be, without any distinction, the idea of the letter only. Moreover, if the means were constituted by the stroke in so far as it is _not_ the letter, the apprehension of all letters would result from the sight of one stroke; for one stroke may easily be conceived as _not_ being _any_ letter.--But, in the same way as the word 'Devadatta' conventionally denotes some particular man, so some particular stroke apprehended by the eye may conventionally symbolise some particular letter to be apprehended by the ear, and thus a particular stroke may be the cause of the idea of a particular letter!--Quite so, we reply, but on this explanation the real is known through the real; for both stroke and conventional power of symbolisation are real. The case is analogous to that of the idea of a buffalo being caused by the picture of a buffalo; that idea rests on the similarity of picture and thing depicted, and that similarity is something real. Nor can it be said (with a view to proving the pûrvapaksha by another analogous instance) that we meet with a cognition of the real by means of the unreal in the case of sound (sabda) which is essentially uniform, but causes the apprehension of different things by means of difference of tone (nâda). For sound is the cause of the apprehension of different things in so far only as we apprehend the connexion of sound manifesting itself in various tones, with the different things indicated by those various tones [FOOTNOTE 77:1]. And, moreover, it is not correct to argue on the ground of the uniformity of sound; for only particular significant sounds such as 'ga,' which can be apprehended by the ear, are really 'sound.'--All this proves that it is difficult indeed to show that the knowledge of a true thing, viz. Brahman, can be derived from Scripture, if Scripture--as based on Nescience--is itself untrue.