1. The Difficulty of Reductionism

There are two possible positions one may take regarding testimony. (Here, testimony should be understood in its broad sense as encompassing the entire epistemic event of coming to know something through another person, rather than being restricted to courtroom or religious testimony.) The first position is: do not believe testimony unless one has evidence that it is certainly true. The second: believe testimony unless one has evidence that it is certainly false. The first view holds that testimony, being a derivative and secondary form of knowledge, must be reduced and inferred back into non-derivative, primary knowledge. Although this view appears rational at first glance, it imposes an excessively stringent criterion on genuine knowledge, given how extensively we rely on testimonial knowledge in everyday life. If testimony can be justified only through objectively reducible supporting evidence, then ultimately we cannot even believe the birthday our parents told us we have—because we cannot rule out the possibility that our parents lied, or that Martians manipulated their memories with a radio weapon.

2. The Advantages of Non-Reductionism

Reductionist views, which demand objectively reducible “evidence” that grounds testimony, hold that testimony is believable only insofar as there is a securely reliable evidential foundation. David Hume adopts such a view. By contrast, the non-reductionist position holds that testimony can yield knowledge in a non-inferential and non-reductive manner. According to this view, unless there is clear evidence that a given piece of testimony is false, one may—and sometimes ought to—believe it. Thomas Reid, the Scottish theologian, champions such a non-reductionist stance. One might suspect that this view rests on a religious motivation that one must accept certain “words” unconditionally. Yet if we grant that we can trust, and at times must trust, experiences that cannot be testified to with certainty—experiences we cannot clearly articulate, or cannot even determine whether they are our own—then it becomes true that epistemic events are possible only by provisionally treating something as true. Testimonial non-reductionism is therefore as practical as it is theological.

3. The Martian Community

If testimony can always be false unless supported by evidence that it is certainly true, then we cannot rule out the possibility of hearing nothing but false testimony generated by a completely unreliable community of reporters. This motivates the non-reductionist stance outlined above. Tony Coady argues that if we call such a hypothetical group of testifiers a Martian community—a community “whose speakers’ utterances bear no correlation to truth”—then such a community is impossible. The claim is that we are imagining something that cannot coherently be imagined. If a testimonial community does not use a vocabulary consistently, if the same word suddenly refers to many different things or different words unexpectedly refer to the same thing, then we cannot learn what items in our world their words correspond to, nor how to translate their language. This impossibility of translation results in the impossibility of the community itself.

4. Similarity to Belief in Fictional Propositions

Believing testimony sometimes requires imagining, believing, and accepting what seems unimaginable, unbelievable, or unacceptable. Some forms of testimony even transfigure propositions about facts and states of affairs—things we ordinarily take for granted—into counterfactuals. This is normally the function of fictional propositions. Thus, although we may not wish to admit it, believing testimony can resemble believing fiction. The ontological burden that arises when one takes fictional facts literally also arises when one believes testimonial facts.

5. Preservation of Life

Unlike belief in fiction, belief in testimony often provides concrete assistance in preserving our lives. We rely on innumerable pieces of knowledge whose origins we cannot recall or prove. Indeed, what ensures our survival at the most critical moments is not the kind of knowledge we can remember or demonstrate clearly, but the embodied knowledge we have internalized since childhood—knowledge held in muscle and viscera before its reasons are ever known. Children and animals are the most frequently cited examples in defense of non-reductionism about testimony. A child believes adult testimony about how the world is long before gaining the ability to verify it independently. If we were required to regard only demonstrably proven knowledge as true knowledge, a child would swallow harmful objects before having time to gather the necessary evidence or would run into traffic before learning the criteria for justification. If the facts that “some things must not be eaten” or “one must obey traffic signals” had to be known through introspective, direct experience alone, we would struggle to meet the conditions of being alive that are required for knowing anything at all.

6. Rational Distrust

Yet there is something unsettling about the non-reductionist position defended in the previous section. Questions of whether a piece of testimony is credible, consistent, or coherent with other non-testimonial facts (evidence) still remain unresolved. Our minds, unable to relinquish suspicion, continue to harbor distrust and various hypotheses concerning testimony. Sometimes the capacity to generate such hypotheses appears to express the active epistemic agency of someone genuinely seeking knowledge. When testimony seems “untrustworthy,” we often suspect that the testifier may be a deliberate liar or may have been in a compromised psychological or perceptual state at the time of the event being testified to. Such rational suspicion about causal origins and credibility keeps open the possibility that testimony might at any moment be revealed as false.

7. Belief–Event–Condition

We cannot inspect another person’s past as we would pull a book from a shelf. Thus, the past testified to by a speaker often remains opaque, despite the act of testimony. Even a clearly delineated past or an apparently self-evident future can become unknowable under certain conditions. This is precisely why—knowing it may be mistaken—we rely on testimony as we rely on prophecy. We choose to believe a speaker even when we do not truly know the speaker’s past. Where a coherent continuity is abandoned, belief enters. Faith, then, is not the expression of will but the relinquishing of will—the abandonment of the epistemic activity of rational doubt. Belief begins where the knower’s attempt to extract a universal principle from testimony is renounced. Testimonial listening made possible by the suspension of doubt takes the place of knowledge.

8. Agency

In cases of testimonial knowledge, where the ultimate bearer of knowledge is not the knower themselves, the knower’s exercise of epistemic effort, ability, virtue, or introspective judgment is not the ultimate requirement for the acquisition of knowledge. If epistemic agency is not necessary for attaining true knowledge, then the knower’s epistemic activity deserves no credit. If we link acquiring knowledge through testimony to “the ability to listen” or “the effort to listen,” then listening becomes an issue of agency. But we do not fail to hear because we exert insufficient effort or lack ability in some particular interval. To accuse a hidden agent that manipulates what becomes audible or inaudible, believable or unbelievable, would be to treat representation as always carrying human intention and purpose. Yet, as inattentional blindness shows, cognition itself functions like a sieve that renders even the plainly visible invisible. Thus, instead of asking why something is not seen or heard, it may be better to ask why something is seen or heard at all. / July–December 2021

References

Jennifer Lackey (2007), “Why We Don’t Deserve Credit for Everything We Know,” Synthese, 158, 345–361.
Dan O’Brien (2011), An Introduction to the Theory of Knowledge, trans. Han Sang-ki, Seogwangsa.
Han Sang-ki (2012), “Reid’s Non-Reductionism about Testimony,” Philosophical Inquiry, 31, 153–79.