The previous quotation continues:
Not that the subject does not [also] enjoy a lesser informational state. It cannot be true that he heard from so-and-so that things are thus and so unless it is true that he heard so-and-so say that things are that way - a truth that leaves it entirely open whether things are that way. [ibid: 436]
But McDowell suggests that the very idea of having heard that things are thus and so should not be analysed using this lesser state as a starting point. Similarly, seeing that things are thus and so implies that it seems or appears to the subject that things are thus and so, a truth that leaves it entirely open whether things are that way. And similarly that should not be the starting point of an analysis of ‘seeing that...’.
The net effect of this reorientation is not to offer a philosophical checklist of when testimony amounts to knowledge but rather to stop commitment a fallacious view of knowledge undermining the trivial claim that: ‘if a knowledgeable speaker gives intelligible expression to his knowledge, it may become available at second hand to those who understand what he says’ [ibid: 417].
Testimony and interpretation: some empirical issues
The picture presented so far is one according to which a key aspect of psychiatric diagnosis rests on matters that are not within the direct control of the clinician. This is not to say that he or she can make no further tests on the reliability of evidence presented to them second hand. But even the clinician’s best epistemic position is one in which factors lie outside his or her control. He or she may, for example, test the reliability of one witness against another but there is no prospect of justifying the use of testimony in independent terms which an individual can ensure non-question-beggingly.
The picture set out above also suggests that acquiring knowledge through testimony is much less of an intellectual matter than one might be tempted to suppose. One can acquire knowledge simply by taking someone else’s utterance at face value, providing that one understands it.
Given the attraction of the internalist approach to knowledge, partly due to Descartes’ influence, one response to the implausibility of an internalist account of testimony is to deny that testimony really can or does underpin knowledge. To take a non-medical example, one might claim that it is not possible, for example, to gain knowledge of the location of a station in a foreign city simply by asking someone. Although one may gain a true belief about where the station is, it cannot amount to knowledge, according to this view, because it depends on an element of luck in avoiding a practical joker who would have mislead one. Proper knowledge, according to this modification of the natural view, is limited to first-person perception, memory or reasoning.
This will not do, however, because, as mentioned above, even first person observations are framed using linguistic concepts that are indebted to testimony. To give up on testimony as a source of knowledge is, in fact, to give up on most claims to knowledge both second and first hand. If an internalist account of knowledge by testimony is impossible, so much the worse for internalism. Testimony as we all know - not least because we have been told this! - can provide us with knowledge.
On the non-intellectual description supported by discussion of McDowell, knowledge can rub off on other people, who have ears to hear. It can, however, be impeded. One blockage which is relevant to psychiatric diagnosis is if a speaker and hearer do not speak the same language. This can be the case in psychiatric diagnosis if, for example, a patient does not speak the same language as the clinician.
A successful diagnostic assessment is contingent upon good communication between the clinician, the patient and the carer [Bhalla and Blakemore 1981; Shah 1992; Jones and Gill 1998; Shah 1999]. Good communication between the clinician and the patient will depend