The Vedanta Sutras of Badarayana, Commentary by Sankara (SBE38), tr. by George Thibaut  at sacred-texts.com
7. The absence of that (i.e. of dreams, i.e. dreamless sleep) takes place in the nâdîs and in the Self; according to scriptural statement.
The state of dream has been discussed; we are now going to enquire into the state of deep sleep. A number of scriptural passages refer to that state. In one place we read, 'When a man is asleep, reposing and at perfect rest so that he sees no dream, then he has entered into those nâdîs' (Kh. Up. VIII, 6, 3). In another place it is said with reference to the nâdîs, 'Through them he moves forth and rests in the surrounding body' (Bri. Up. II, 1, 19). So also in another place, 'In these the person is when sleeping he sees no dream. Then he becomes one with the prâna alone' (Kau. Up. IV, 20). Again in another place, 'That ether which is within the heart in that he reposes' (Bri. Up. IV, 4, 22). Again, 'Then he becomes united with that which is; he is gone to his Self (Kh. Up. VI, 8, i). And, 'Embraced by the highest Self (prâgña) he knows nothing that is without, nothing that is within' (Bri. Up. IV, 3, 21). Here the doubt arises whether the nâdîs, &c., mentioned in the above passages are independent from each other and constitute various places for the soul in the state of deep sleep, or if they stand in mutual relation so as to constitute one such place only. The pûrvapakshin takes the former view on account of the various places mentioned serving one and the same purpose. Things serving the same purpose, as, e.g. rice and barley 1, are never seen to be dependent
on each other. That the nâdîs, &c., actually serve the same purpose appears from the circumstance of their being all of them exhibited equally in the locative case, 'he has entered into the nâdîs,' 'he rests in the pericardium,' &c. 1--But in some of the passages quoted the locative case is not employed, so, e.g. in 'He becomes united with that which is' (satâ, instrumental case)!--That makes no difference, we reply, because there also the locative case is meant. For in the complementary passage the text states that the soul desirous of rest enters into the Self, 'Finding no rest elsewhere it settles down on breath' (Kh. Up. VI, 8, 2); a passage in which the word 'breath' refers to that which is (the sat). A place of rest of course implies the idea of the locative case. The latter case is, moreover, actually exhibited in a further complementary passage, 'When they have become merged in that which is (sati), they know not that they are merged in it.'--In all these passages one and the same state is referred to, viz. the state of deep sleep which is characterised by the suspension of all special cognition. Hence we conclude that in the state of deep sleep the soul optionally goes to any one of those places, either the nâdîs, or that which is, &c.
To this we make the following reply--'The absence of that,' i.e. the absence of dreams--which absence constitutes the essence of deep sleep-takes place 'in the nâdîs and in the Self;' i.e. in deep sleep the soul goes into both together, not optionally into either.--How is this known?--'From scripture.'--Scripture says of all those things, the nâdîs, &c., that they are the place of deep sleep; and those statements we must combine into one, as the hypothesis of option would involve partial refutation 2. The assertion
made above that we are compelled to allow option because the nâdîs, &c., serve one and the same purpose, is without foundation; for from the mere fact of two things being exhibited in the same case it does not follow by any means that they serve the same purpose, and that for that reason we have to choose between them. We on the contrary see that one and the same case is employed even where things serve different purposes and have to be combined; we say, e.g. 'he sleeps in the palace, he sleeps on the couch 1.' So in the present case also the different statements can be combined into one, 'He sleeps in the nâdîs, in the surrounding body, in Brahman.' Moreover, the scriptural passage, 'In these the person is when sleeping he sees no dream; then he becomes one with the prâna alone,' declares, by mentioning them together in one sentence, that the nâdîs and the prâna are to be combined in the state of deep sleep. That by prâna Brahman is meant we have already shown (I, 1, 28). Although in another text the nâdîs are spoken of as an independent place of deep sleep as it were ('then he has entered into those nâdîs'), yet, in order not to contradict other passages in which Brahman is spoken of as the place of deep sleep, we must explain that text to mean that the soul abides in Brahman through the nâdîs. Nor is this interpretation opposed to the employment of the locative case ('into--or in--those nâdîs'); for if the soul enters into Brahman by means of the nâdîs it is at the same time in the nâdîs; just as a man who descends to the sea by means of the river Gaṅgâ is at the same time on the Gaṅgâ.--Moreover that passage about the nâdîs, because its purpose is to describe the road, consisting of the rays and nâdîs, to the Brahma world, mentions the entering of the soul into the nâdîs in order to glorify the latter (not in order to describe the state of deep sleep); for the clause following upon the one which refers to the entering
praises the nâdîs, 'There no evil touches him.' The text, moreover, adds a reason for the absence of all evil, in the words, 'For then he has become united with the light.' That means that on account of the light contained in the nâdîs (which is called bile) having overpowered the organs the person no longer sees the sense-objects. Or else Brahman may be meant by the 'light;' which term is applied to Brahman in another passage also, 'It is Brahman only, light only' (Bri. Up. IV, 4, 7). The passage would then mean that the soul becomes, by means of the nâdîs, united with Brahman, and that hence no evil touches it. That the union with Brahman is the reason for the absence of all contact with evil, is known from other scriptural passages, such as, 'All evils turn back from it; for the world of Brahman is free from all evil' (Kh. Up. VIII, 4, 1). On that account we have to combine the nâdîs with Brahman, which from other passages is known to be the place of deep sleep.--Analogously we conclude that the pericardium also, because it is mentioned in a passage treating of Brahman, is a place of deep sleep only in subordination to Brahman. For the ether within the heart is at first spoken of as the place of sleep ('He lies in the ether which is in the heart,' Bri. Up. II, 1, 17), and with reference thereto it is said later on, 'He rests in the pericardium' (II, 1, 19). Pericardium (purîtat) is a name of that which envelops the heart; hence that which rests within the ether of the heart--which is contained in the pericardium--can itself be said to rest within the pericardium; just as a man living in a town surrounded by walls is said to live within the walls. That the ether within the heart is Brahman has already been shown (I, 3, l4).--That again the nâdîs and the pericardium have to be combined as places of deep sleep appears from their being mentioned together in one sentence ('Through them he moves forth and rests in the purîtat). That that which is (sat) and the intelligent Self (prâgña) are only names of Brahman is well known; hence scripture mentions only three places of deep sleep, viz. the nâdîs, the pericardium, and Brahman. Among these three again Brahman alone is the lasting place of deep sleep; the
nâdîs and the pericardium are mere roads leading to it. Moreover (to explain further the difference of the manner in which the soul, in deep sleep, enters into the nâdîs, the pericardium and Brahman respectively), the nâdîs and the pericardium are (in deep sleep) merely the abode of the limiting adjuncts of the soul; in them the soul's organs abide 1. For apart from its connexion with the limiting adjuncts it is impossible for the soul in itself to abide anywhere, because being non-different from Brahman it rests in its own glory. And if we say that, in deep sleep, it abides in Brahman we do not mean thereby that there is a difference between the abode and that which abides, but that there is absolute identity of the two. For the text says, 'With that which is he becomes united, he is gone to his Self;' which means that the sleeping person has entered into his true nature.--It cannot, moreover, be said that the soul is at any time not united with Brahman--for its true nature can never pass away--; but considering that in the state of waking and that of dreaming it passes, owing to the contact with its limiting adjuncts, into something else, as it were, it may be said that when those adjuncts cease in deep sleep it passes back into its true nature. Hence it would be entirely wrong to assume that, in deep sleep, it sometimes becomes united with Brahman and sometimes not 2. Moreover, even if we admit that there are different places for the soul in deep sleep, still there does not result, from that difference of place, any difference in the quality of deep sleep which is in all cases characterised by the cessation of special cognition; it is, therefore, more appropriate to say that the soul does (in deep sleep) not cognize on account of its oneness, having become united with Brahman; according to the Sruti, 'How should he know another?' (Bri. Up. IV, 5, l5).--If, further, the sleeping soul did rest in the nâdîs and the purîtat, it would be impossible
to assign any reason for its not cognizing, because in that case it would continue to have diversity for its object; according to the Sruti, 'When there is, as it were, duality, then one sees the other,' &c.--But in the case of him also who has diversity for his object, great distance and the like may be reasons for absence of cognition!--What you say might indeed apply to our case if the soul were acknowledged to be limited in itself; then its case would be analogous to that of Vishnumitra, who, when staying in a foreign land, cannot see his home. But, apart from its adjuncts, the soul knows no limitation.--Well, then, great distance, &c., residing in the adjuncts may be the reason of non-cognition!--Yes, but that leads us to the conclusion already arrived at, viz. that the soul does not cognize when, the limiting adjuncts having ceased, it has become one with Brahman.
Nor do we finally maintain that the nâdîs, the pericardium, and Brahman are to be added to each other as being equally places of deep sleep. For by the knowledge that the nâdîs and the pericardium are places of sleep, nothing is gained, as scripture teaches neither that some special fruit is connected with that knowledge nor that it is the subordinate member of some work, &c., connected with certain results. We, on the other hand, do want to prove that that Brahman is the lasting abode of the soul in the state of deep sleep; that is a knowledge which has its own uses, viz. the ascertainment of Brahman being the Self of the soul, and the ascertainment of the soul being essentially non-connected with the worlds that appear in the waking and in the dreaming state. Hence the Self alone is the place of deep sleep.
141:1 Either of which may be employed for making the sacrificial cake.
142:1 The argument of the pûrvapakshin is that the different places in which the soul is said to abide in the state of deep sleep are all exhibited by the text in the same case and are on that account co-ordinate. Mutual relation implying subordination would require them to be exhibited in different cases enabling us to infer the exact manner and degree of relation.
142:2 By allowing option between two Vedic statements we lessen the p. 143 authority of the Veda; for the adoption of either alternative sublates, for the time, the other alternative.
143:1 Where the two locatives are to be combined into one statement, 'he sleeps on the couch in the palace.'
145:1 Ân. Gi. explains karanâni by karmâni: nâdîshu purîtati ka gîvasyopâdhyantarbhûtani karanâni karmâni tishthantîty upâdhyâ-dhâratvam, gîvasya tv âdhâro brahmaiva.
145:2 But with the nâdîs or the pericardium only.