milk as well?—­Let us then maintain, the asatkaryavadin rejoins, that there is indeed an equal non-existence of any effect in any cause, but that at the same time each causal substance has a certain capacity reaching beyond itself (ati/s/aya) for some particular effect only and not for other effects; that, for instance, milk only, and not clay, has a certain capacity for curds; and clay only, and not milk, an analogous capacity for jars.—­What, we ask in return, do you understand by that ‘ati/s/aya?’ If you understand by it the antecedent condition of the effect (before its actual origination), you abandon your doctrine that the effect does not exist in the cause, and prove our doctrine according to which it does so exist.  If, on the other hand, you understand by the ati/s/aya a certain power of the cause assumed to the end of accounting for the fact that only one determined effect springs from the cause, you must admit that the power can determine the particular effect only if it neither is other (than cause and effect) nor non-existent; for if it were either, it would not be different from anything else which is either non-existent or other than cause and effect, (and how then should it alone be able to produce the particular effect?) Hence it follows that that power is identical with the Self of the cause, and that the effect is identical with the Self of that power.—­Moreover, as the ideas of cause and effect on the one hand and of substance and qualities on the other hand are not separate ones, as, for instance, the ideas of a horse and a buffalo, it follows that the identity of the cause and the effect as well as of the substance and its qualities has to be admitted.  Let it then be assumed, the opponent rejoins, that the cause and the effect, although really different, are not apprehended as such, because they are connected by the so-called samavaya connexion[293].—­If, we reply, you assume the samavaya connexion between cause and effect, you have either to admit that the samavaya itself is joined by a certain connexion to the two terms which are connected by samavaya, and then that connexion will again require a new connexion (joining it to the two terms which it binds together), and you will thus be compelled to postulate an infinite series of connexions; or else you will have to maintain that the samavaya is not joined by any connexion to the terms which it binds together, and from that will result the dissolution of the bond which connects the two terms of the samavaya relation[294].—­Well then, the opponent rejoins, let us assume that the samavaya connexion as itself being a connexion may be connected with the terms which it joins without the help of any further connexion.—­Then, we reply, conjunction (sa/m/yoga) also must be connected with the two terms which it joins without the help of the samavaya connexion; for conjunction also is a kind of connexion[295].—­Moreover, as substances, qualities, and so on are apprehended as standing in the relation of identity, the assumption of the samavaya relation has really no purport.