Three Ways of Knowing

Nobody alive has checked most of what they know. The question is not whether to trust testimony. It is which testimony, and why.

The fastest objection to any argument from scripture is also the most reasonable-sounding one: that is not evidence, that is a book saying so. Science, the objection goes, runs on observation and reason. Scripture runs on authority.

The objection deserves a full answer, because everything on this site depends on it. Every ĀRK essay tags its claims with one of three labels: pratyakṣa, perception; anumāna, inference; and śabda, testimony. The first two are uncontroversial. This essay is about why the third belongs on the list, and where each of the three stops working.

The framework is old. The Nyāya school of Indian logic catalogued these instruments of knowledge, pramāṇas, many centuries before “unscientific” became an accusation, and Baladeva Vidyābhūṣaṇa’s Prameya-ratnāvalī lists the same three as the ninth of its established truths (see the Codex, Part II).

Perception Is Not Self-Certifying

Pratyakṣa, direct perception, is the court of first appeal for everything a laboratory does. It is also demonstrably unreliable on its own.

In 1995 the MIT vision scientist Edward Adelson published the checker-shadow illusion: two squares on a checkerboard, one in open light and one in a cylinder’s shadow, that look plainly different and are printed in exactly the same grey. Knowing this does not help. The eyes keep reporting the difference after the mind has been told it is false.

Instruments do not escape the problem; they move it. In 2015 the Open Science Collaboration, 270 researchers, reported in Science the results of repeating 100 published psychology experiments. Of the originals, 97 had reported statistically significant results. Of the repeats, 36 did.

The finding has a stated limit, and it is disputed. Daniel Gilbert and colleagues argued in the same journal in 2016 that some repeats departed from the original protocols. On either reading, the lesson stands: a measurement becomes knowledge only after other people check it, and most people who rely on it never do.

The tradition names the failure modes precisely. Jīva Gosvāmī’s Tattva-sandarbha, the opening volume of his sixteenth-century Ṣaṭ-sandarbha, lists four defects of every embodied knower: bhrama, the tendency to mistake; pramāda, inattention; vipralipsā, the tendency to deceive, oneself included; and karaṇāpāṭava, the limited power of the senses.

Inference Stands on Something

Anumāna, inference, extends perception. From smoke on a hill, fire: the Nyāya texts’ own stock example. It reaches what the eye cannot, which is why every science runs on it.

But inference stands on premises, and its premises come from perception or from testimony. When they are incomplete, the logic can be flawless and the conclusion wrong. For centuries European logicians used “all swans are white” as a model of a safe generalisation. In 1697 the Dutch navigator Willem de Vlamingh’s expedition saw black swans on the Swan River in Western Australia.

David Hume stated the general problem in 1748: no quantity of past observations logically guarantees the next one. Science lives with that by treating its conclusions as provisional. That is exactly the right policy for studying prakṛti, material nature. It is also an admission that inference on its own cannot close a question about what lies beyond every possible observation.

The Vedānta-sūtra makes the point in one compound: tarkāpratiṣṭhānāt (2.1.11), “because reasoning has no fixed ground.” An argument can always be met by a cleverer argument. Where perception gives no footing, reasoning alone never settles.

Everyone Runs on Testimony

Here is the part the objection leaves out. Almost everything any educated person knows arrives as śabda, testimony.

Very few readers of this sentence have measured the speed of light, sequenced a genome, dated a rock, or counted the population of a city. A working scientist reads a paper and builds on it without repeating it. Science is not the absence of testimony. It is an institution for making testimony more reliable, with peer review, replication and citation as its instruments.

Western philosophy took this seriously late. C. A. J. Coady’s Testimony: A Philosophical Study (Oxford, 1992) argued that testimony cannot be reduced to perception and inference, because the evidence needed to check it is itself mostly testimony. The Nyāya tradition had listed śabda as a separate pramāṇa from the start.

What Makes Testimony Good

Everyone already applies tests. A source is trusted more when it was in a position to know, when it has no motive to deceive, when it has been handed on without alteration, and when it agrees with itself across everything it says.

The Vedic claim for śabda is that the Veda passes those tests at their limit. It does not originate in an embodied knower’s perception, so the four defects do not enter at the source. It is carried by paramparā, a succession of teachers whose job is to hand it on unaltered. And it is held to a concordance rule, the Brahma Codex, that makes internal consistency a condition every reading must meet.

This is where Jīva Gosvāmī locates the root of knowledge: in śabda free of the four defects. The Vedānta-sūtra says the same of its own subject in its third sūtra, śāstra-yonitvāt, “because scripture is the source.” The absolute truth is the kind of thing that can be known only if it makes itself known.

Pramāṇam is the word: evidence. Kṛṣṇa does not tell Arjuna to set perception and inference aside. He tells him which instrument has jurisdiction over the question being asked, which is what to do and what not to do.

Jurisdiction, Not Rivalry

The three are not rivals. Each has a jurisdiction, and each is strongest inside it.

1. 01
   pratyakṣa — perception

   Governs what the senses and their instruments can reach. Reported with its sample, its method and its stated limit. Strong, correctable, and silent about anything no instrument can register.

2. 02
   anumāna — inference

   Extends perception by reasoning. Valid when its premises are stated and its weakest step can be named. Strong within its premises, and never stronger than them.

3. 03
   śabda — testimony

   Carries what neither of the others can reach. Trusted when its source is free of the four defects, received through an unbroken line, and consistent with itself. The only instrument that can speak about what lies beyond every possible observation.

This is why ĀRK never claims science has proved a scriptural proposition. Science investigates prakṛti; it can neither prove nor disprove Kṛṣṇa. Claiming a laboratory has confirmed a śabda claim is a surrender disguised as a victory: it concedes that only perception counts, then loses the argument the first time the data moves.

And it is why the tagging is not a hedge. Telling a reader which instrument a claim rests on is telling them exactly how firm the ground is under it.

The Teacher

Testimony has a delivery mechanism, and the Gītā names it.

Tattva-darśinaḥ: those who have seen the truth. The verse grounds the teacher’s authority in sight, not in office, and it builds the check into the method: paripraśna, questioning. Testimony in this line is not “believe because we said so.” It is a report from people who claim to have seen, handed on under conditions designed to keep it unaltered, open to questions, and checkable against the whole corpus it belongs to.

Where This Could Be Wrong

The strongest objection is circularity. The Veda is trusted because it is free of defects, and that it is free of defects is known from the Veda. Stated at full strength, this is a serious charge.

The answer is that every instrument meets the same wall. Perception can be checked only by more perception. Inference can be checked only by more inference, standing on premises it cannot itself prove. Every source of knowledge is finally trusted on its own authority, inside its own jurisdiction. What separates a good source is not that it escapes this, but that it survives testing wherever testing is possible: consistency, transmission, and whether acting on it produces what it says it will.

That answer will not satisfy everyone, and it does not need to. A reader who declines śabda can still use the other two labels, check every empirical claim on this site at full strictness, and decline only the conclusions that rest on testimony. The labels are there so the choice is made in the open.

The Horizon

A civilization that accepts only perception has to act as if the questions its instruments cannot reach do not exist. That is a choice, not a finding.

The Vedic corpus offers a third instrument and a set of rules for using it honestly. Where the three are used together, and in order, the Gītā states where they lead.

The last step is not an instrument at all. It is bhakti, devotion, which the verse makes the condition for knowing Kṛṣṇa as He is. Testimony tells a reader where to look. What is there is found by the one who looks.

FOUNDATIONS 001 — three ways of knowing.
— ĀRK. Vedic Futurist