symptoms enquired about. In Gujarati, for example, there is no specific word for ‘depressed mood’ and interpreters, faced with this difficulty, often attempt to get around it by asking about other related but distinct symptoms or states such as a lack of happiness but rarely tell the psychiatrist that they cannot translate the symptoms as asked.
Given these problems, two strategies present themselves depending on one’s underlying approach to epistemology: internalist of externalist. According to one, interpreters should be medically trained. The thinking behind this is as follows. If a clinician is to gain knowledge from a patient via an intermediary, then the intermediary had better have knowledge to give. But if that is the case and if knowledge has to be analysed in internalist terms then the intermediary has to be able to justify his or her own claims to know. Thus it would seem that he or she needs to have a full medically informed understanding of the significance of all terms employed by the patient. Only so will the interpreter be able fully to justify their account of the patient.
The alternative view is influenced by the externalist description of testimony set out in this brief paper and is suggested as a response to the experiences described above. On this view, gaining knowledge by testimony depends on a transparent transmission. Providing nothing impedes it, it is possible simply to inherit another’s knowledge, including a patient’s knowledge of her symptoms and their duration. On this second view, testimony is most reliable when it is least transformed in transmission.
This is not to say that a clinician can simply ignore factors that might impede such transmission. To be in a position to have knowledge, to have a standing in the space of reasons as McDowell puts it, they must be rationally responsive to such factors where they might practically exist. But they do not need to be able to provide an argument to the truth of their beliefs from risk-free descriptions of appearances in order to gain knowledge by testimony. That they should be able to is merely an unrealistic philosophical prejudice.
On this second view, interpreters are thus best who most transparently pass on the meaning of the words of patients. This suggests that their knowledge of medical theory should be as close as possible to, but no more than, that of the patient. The risk, otherwise, is that the theoretical apparatus of the interpreter can intervene between patient and clinician. It becomes an opaque screen, rather than a transparent medium, for the transmission of knowledge. Of course, if interpreters are fully medically trained, they can simply make diagnoses themselves. That is possible in the case of bilingual doctors. But on the assumption that an interpreter is not so fully qualified, it is better, according to this model, for them not to exceed the theoretical knowledge of the patient.
This is, of course, a claim with empirical consequences. It would be interesting to test the validity of diagnosis based on interpreters with these two kinds of skill. But the discussion has at least suggested that a prior unreflective commitment to the internalist view of knowledge should not blind one to this possibility and purely conceptual arguments would favour the second approach.
Internalist accounts of knowledge are reflected in lay thinking and historically by Descartes’ influential defence of knowledge against scepticism. Internalism, however, cannot account for the transmission of knowledge by testimony, which plays a key role in psychiatric diagnosis. Thus one important result of recent reflection on the status testimony is that it, and thus diagnosis as a whole, cannot be fitted within the influential internalist framework.
A second feature flows from the inapplicability of internalism to testimony. The model of knowledge by testimony implies that knowledge can be passed via an interpreter even though the interpreter himself or herself is not medically trained. Indeed, there is a reason to believe that medical training might introduce a barrier for the transparent communication of symptoms.