The Vedanta Sutras, commentary by Sankaracharya (SBE34), tr. by George Thibaut  at sacred-texts.com
3. (The omniscience of Brahman follows) from its being the source of Scripture.
Brahman is the source, i.e. the cause of the great body of Scripture, consisting of the Rig-veda and other branches, which is supported by various disciplines (such as grammar, nyâya, purâna, &c.); which lamp-like illuminates all things; which is itself all-knowing as it were. For the origin of a body of Scripture possessing the quality of omniscience cannot be sought elsewhere but in omniscience itself. It is generally understood that the man from whom some special body of doctrine referring to one province of knowledge only originates, as, for instance, grammar from Pânini possesses a more extensive knowledge than his work, comprehensive though it be; what idea, then, shall we have to form of the supreme omniscience and omnipotence of that great Being, which in sport as it were, easily as a man sends forth his breath, has produced the vast mass of holy texts known as the Rig-veda, &c., the mine of all knowledge, consisting of manifold branches, the cause of the distinction of all the different classes and conditions of gods, animals, and men! See what Scripture says about him, 'The Rig-veda, &c., have been breathed forth from that great Being' (Bri. Up. II, 4, 10).
Or else we may interpret the Sûtra to mean that Scripture consisting of the Rig-veda, &c., as described above, is the source or cause, i.e. the means of right knowledge through which we understand the nature of Brahman. So that the sense would be: through Scripture only as a means of knowledge Brahman is known to be the cause of the origin, &c., of the world. The special scriptural passage meant has been quoted under the preceding Sûtra 'from which these beings are born,' &c.--But as the preceding Sûtra already has pointed out a text showing that Scripture is the source of Brahman, of what use then is the present Sûtra?--The words of the preceding Sûtra, we reply, did not clearly indicate the scriptural passage, and room was thus left for the suspicion that the origin, &c., of the world were adduced merely as determining an inference (independent of Scripture). To obviate this suspicion the Sûtra under discussion has been propounded.
But, again, how can it be said that Scripture is the means of knowing Brahman? Since it has been declared that Scripture aims at action (according to the Pûrva Mîmâmsâ
Sûtra I, 2, 1, 'As the purport of Scripture is action, those scriptural passages whose purport is not action are purportless'), the Vedânta-passages whose purport is not action are purportless. Or else if they are to have some sense, they must either, by manifesting the agent, the divinity or the fruit of the action, form supplements to the passages enjoining actions, or serve the purpose of themselves enjoining a new class of actions, such as devout meditation and the like. For the Veda cannot possibly aim at conveying information regarding the nature of accomplished substances, since the latter are the objects of perception and the other means of proof (which give sufficient information about them; while it is the recognised object of the Veda to give information about what is not known from other sources). And if it did give such information, it would not be connected with things to be desired or shunned, and thus be of no use to man. For this very reason Vedic passages, such as 'he howled, &c.,' which at first sight appear purposeless, are shown to have a purpose in so far as they glorify certain actions (cp. Pû. Mî. Sû. I, 2, 7, 'Because they stand in syntactical connection with the injunctions, therefore their purport is to glorify the injunctions'). In the same way mantras are shown to stand in a certain relation to actions, in so far as they notify the actions themselves and the means by which they are accomplished. So, for instance, the mantra, 'For strength thee (I cut;' which accompanies the cutting of a branch employed in the darsapûrnamâsa-sacrifice). In short, no Vedic passage is seen or can be proved to have a meaning but in so far as it is related to an action. And injunctions which are defined as having actions for their objects cannot refer to accomplished existent things. Hence we maintain that the Vedânta-texts are mere supplements to those passages which enjoin actions; notifying the agents, divinities, and results connected with those actions. Or else, if this be not admitted, on the ground of its involving the introduction of a subject-matter foreign to the Vedânta-texts (viz. the subject-matter of the Karmakânda of the Veda), we must admit (the second of the two alternatives proposed above viz.) that the
[paragraph continues] Vedânta-texts refer to devout meditation (upâsanâ) and similar actions which are mentioned in those very (Vedânta) texts. The result of all of which is that Scripture is not the source of Brahman.
To this argumentation the Sûtrakâra replies as follows: