The Vedanta Sutras, commentary by Sankaracharya (SBE34), tr. by George Thibaut  at sacred-texts.com
23. For the same reason breath (is Brahman).
Concerning the udgîtha it is said (Kh. Up. I, 10, 9), 'Prastotri, that deity which belongs to the prastâva, &c.,' and, further on (I,11,4; 5), 'Which then is that deity? He said: Breath. For all these beings merge into breath alone, and from breath they arise. This is the deity belonging to the prastâva.' With reference to this passage doubt and decision are to be considered as analogous to those stated under the preceding Sûtra. For while in some passages--as, for instance, 'For indeed, my son, mind is fastened to prâna,' Kh. Up. VI, 8, 2; and, 'the prâna of prâna,' Bri. Up. IV, 4, 18--the word 'breath' is seen to denote Brahman, its use
in the sense of a certain modification of air is better established in common as well as in Vedic language. Hence there arises a doubt whether in the passage under discussion the word prâna denotes Brahman or (ordinary) breath. In favour of which meaning have we then to decide?
Here the pûrvapakshin maintains that the word must be held to denote the fivefold vital breath, which is a peculiar modification of wind (or air); because, as has been remarked already, that sense of the word prâna is the better established one.--But no, an objector will say, just as in the case of the preceding Sûtra, so here also Brahman is meant, on account of characteristic marks being mentioned; for here also a complementary passage gives us to understand that all beings spring from and merge into prâna; a process which can take place in connexion with the highest Lord only.--This objection, the pûrvapakshin replies, is futile, since we see that the beings enter into and proceed from the principal vital air also. For Scripture makes the following statement (Sat. Br. X, 3, 3, 6), 'When man sleeps, then into breath indeed speech merges, into breath the eye, into breath the ear, into breath the mind; when he awakes then they spring again from breath alone.' What the Veda here states is, moreover, a matter of observation, for during sleep, while the process of breathing goes on uninterruptedly, the activity of the sense organs is interrupted and again becomes manifest at the time of awaking only. And as the sense organs are the essence of all material beings, the complementary passage which speaks of the merging and emerging of the beings can be reconciled with the principal vital air also. Moreover, subsequently to prâna being mentioned as the divinity of the prastâva the sun and food are designated as the divinities of the udgîtha and the pratihâra. Now as they are not Brahman, the prâna also, by parity of reasoning, cannot be Brahman.
To this argumentation the author of the Sûtras replies: For the same reason prâna--that means: on account of the presence of characteristic marks--which constituted the reason stated in the preceding Sûtra--the word prâna also
must be held to denote Brahman. For Scripture says of prâna also, that it is connected with marks characteristic of Brahman. The sentence, 'All these beings merge into breath alone, and from breath they arise,' which declares that the origination and retractation of all beings depend on prâna, clearly shows prâna to be Brahman. In reply to the assertion that the origination and retractation of all beings can be reconciled equally well with the assumption of prâna denoting the chief vital air, because origination and retractation take place in the state of waking and of sleep also, we remark that in those two states only the senses are merged into, and emerge from, the chief vital air, while, according to the scriptural passage, 'For all these beings, &c.,' all beings whatever into which a living Self has entered, together with their senses and bodies, merge and emerge by turns. And even if the word 'beings' were taken (not in the sense of animated beings, but) in the sense of material elements in general, there would be nothing in the way of interpreting the passage as referring to Brahman.--But. it may be said, that the senses together with their objects do, during sleep, enter into prâna, and again issue from it at the time of waking, we distinctly learn from another scriptural passage, viz. Kau. Up. III, 3, 'When a man being thus asleep sees no dream whatever, he becomes one with that prâna alone. Then speech goes to him with all names,' &c.--True, we reply, but there also the word prâna denotes (not the vital air) but Brahman, as we conclude from characteristic marks of Brahman being mentioned. The objection, again, that the word prâna cannot denote Brahman because it occurs in proximity to the words 'food' and 'sun' (which do not refer to Brahman), is altogether baseless; for proximity is of no avail against the force of the complementary passage which intimates that prâna is Brahman. That argument, finally, which rests on the fact that the word prâna commonly denotes the vital air with its five modifications, is to be refuted in the same way as the parallel argument which the pûrvapakshin brought forward with reference to the word 'ether.' From all this it follows that the prâna, which is the deity of the prastâva, is Brahman.
Some (commentators) 1 quote under the present Sûtra the following passages, 'the prâna of prâna' (Bri. Up. IV, 4, 18), and 'for to prâna mind is fastened' (Kh. Up. VI, 8, 2). But that is wrong since these two passages offer no opportunity for any discussion, the former on account of the separation of the words, the latter on account of the general topic. When we meet with a phrase such as 'the father of the father' we understand at once that the genitive denotes a father different from the father denoted by the nominative. Analogously we infer from the separation of words contained in the phrase, 'the breath of breath,' that the 'breath of breath' is different from the ordinary breath (denoted by the genitive 'of breath'). For one and the same thing cannot, by means of a genitive, be predicated of--and thus distinguished from--itself. Concerning the second passage we remark that, if the matter constituting the general topic of some chapter is referred to in that chapter under a different name, we yet conclude, from the general topic, that that special matter is meant. For instance, when we meet in the section which treats of the gyotishtoma sacrifice with the passage, 'in every spring he is to offer the gyotis sacrifice,' we at once understand that the word denotes the gyotishtoma. If we therefore meet with the clause 'to prâna mind is fastened' in a section of which the highest Brahman is the topic, we do not for a moment suppose that the word prâna should there denote the ordinary breath which is a mere modification of air. The two passages thus do not offer any matter for discussion, and hence do not furnish appropriate instances for the Sûtra. We have shown, on the other hand, that the passage about the prâna, which is the deity of the prastâva, allows room for doubt, pûrvapaksha and final decision.
87:1 The vrittikâra, the commentators say.