testimony an act of telling, including all assertions apparently intended to impart information, regardless of social setting. In an extended sense personal letters and messages, books, and other published material purporting to contain factual information also constitute testimony. Testimony may be sincere or insincere, and may express knowledge or baseless prejudice. When it expresses knowledge, and it is rightly believed, this knowledge is disseminated to its recipients, near or remote. Secondhand knowledge can be passed on further, producing long chains of testimony; but these chains always begin with the report of an eyewitness or expert.
In any social group with a common language there is potential for the sharing, through testimony, of the fruits of individuals’ idiosyncratic acquisition of knowledge through perception and inference. In advanced societies specialization in the gathering and production of knowledge and its wider dissemination through spoken and written testimony is a fundamental socioepistemic fact, and a very large part of each person’s body of knowledge and belief stems from testimony.
Thus the question when a person may properly believe what another tells her, and what grounds her epistemic entitlement to do so, is a crucial one in epistemology. Reductionists about testimony insist that this entitlement must derive from our entitlement to believe what we perceive to be so, and to draw inferences from this according to familiar general principles. (See e.g., Hume’s classic discussion, in his Enquiry into Human Understanding, section X.) On this view, I can perceive that someone has told me that p, but can thereby come to know that p only by means of an inference – one that goes via additional, empirically grounded knowledge of the trustworthiness of that person. Anti-reductionists insist, by contrast, that there is a general entitlement to believe what one is told just as such – defeated by knowledge of one’s informant’s lack of trustworthiness (her mendacity or incompetence), but not needing to be bolstered positively by empirically based knowledge of her trustworthiness. Anti-reductionists thus see testimony as an autonomous source of knowledge on a par with perception, inference, and memory. One argument adduced for anti-reductionism is transcendental: We have many beliefs acquired from testimony, and these beliefs are knowledge; their status as knowledge cannot be accounted for in the way required by the reductionist – that is, the reliability of testimony cannot be independently confirmed; therefore the reductionist’s insistence on this is mistaken. However, while it is perhaps true that the reliability of all the beliefs one has that depend on past testimony cannot be simultaneously confirmed, one can certainly sometimes ascertain, without circularity, that a specific assertion by a particular person is likely to be correct – if, e.g.,one’s own experience has established that that person has a good track record of reliability about that kind of thing. See also EPISTEMOLOGY, HUME, INFEREN — TIAL KNOWLEDG. E.F.