The Mīmāṃsā school takes the analysis of śabda in a dramatically different direction. For Mīmāṃsā, the primary concern is the authority of the Vedas, the ancient Sanskrit scriptures that prescribe rituals, moral duties, and the structure of the cosmos. But their philosophical strategy for defending that authority is not theological; it is epistemological, and it is remarkably sophisticated.
Mīmāṃsā begins with what it calls svataḥ prāmāṇya, the thesis of intrinsic validity: all cognition is, by default, valid in itself and requires no external verification to establish its validity. Invalidity, on this view, must be positively demonstrated; validity is the default presumption. This is close to what Western epistemologists call dogmatism or default justification, you are entitled to believe what cognition presents until you have positive reason to doubt it.
Applied to the Vedas, the implication is striking: the Vedas are not valid because a reliable God authored them. In fact, the Pūrva-Mīmāṃsā (the older school associated with Kumārila Bhaṭṭa and Prabhākara, c. 7th century CE) is famously atheistic, it holds that there is no God as author of the Vedas. Instead, the Vedas are apauruṣeya, "without a person": uncreated, beginningless, impersonal. Their authority does not rest on an author's reliability (which could always be questioned) but on the texts' own intrinsic validity, which is never successfully defeated.
Kumārila distinguishes between two kinds of testimony: laukika (ordinary human testimony) and vedic (scriptural testimony). Laukika testimony is reliable conditionally, it depends on the speaker's reliability and can be defeated by counterevidence. Vedic testimony is categorically different: it carries no author's personal interest or possible error because it has no author. Its truth is structural, not contingent.
Prabhākara pushes further: he restricts genuine śabda-pramāṇa to Vedic injunctions, the imperative mood commands of the Vedas ("Do the Agnihotra sacrifice"). Other statements, including ordinary testimony, are valid only insofar as they serve or support these injunctions. This reflects Mīmāṃsā's core concern: the Vedas are essentially prescriptive, not descriptive, their function is to direct action, not to describe the world.
The Advaita Vedānta school of Śaṅkara (c. 8th century CE) accepts śabda as a pramāṇa but extends its scope dramatically: Vedic testimony is the only valid means of knowing brahman, the ultimate reality, because brahman is beyond both perception and inference. You cannot see brahman; you cannot infer it from empirical premises. The Upaniṣads, which declare "tat tvam asi" ("That art thou", the ultimate reality and your deepest self are identical) are the sole epistemic access to this truth. For Vedānta, śabda is not merely one pramāṇa among four, it is the pramāṇa uniquely adequate to the highest knowledge.
The debate between Nyāya and Mīmāṃsā on śabda illuminates a deep tension in epistemology that appears in Western thought too: is testimony valid because of facts about specific reliable testifiers (a reliabilist answer), or does some testimony carry intrinsic, self-grounding authority that does not depend on the testifier's track record? The Mīmāṃsā position is philosophically equivalent to the anti-reductionist position in Western social epistemology, but grounded not in a general principle about testimony as a basic source, but in a highly specific theological-epistemological claim about scriptural authority.