Gettier problem

3. Causal analyses and proper functioning

The preceding Gettier-type cases involve no causal relationship between one’s believing that p and the state of affairs that one believes to obtain. So various analyses requiring such specified causal connections have been proposed (see Knowledge, causal theory of). Alternatively, sometimes conditionals have been included in a fourth condition of knowing, relating believing that p to what one believes to obtain. For instance, Robert Nozick (1981) initiated a new line of research by requiring that if it were the case that p then one would believe that p. As it stands, this suggestion faces a difficulty that also affects simple versions of causal theories: these views fail to rule out knowledge in a refinement of another variant of the Nogot case that was devised by Keith Lehrer (1979). Here he imagines that Nogot is manifesting a compulsion to trick people into justifiedly believing truths by concocting Gettierized evidence of type e for those truths. Nozick’s conditional is satisfied if we add the details that: (1) Mr Nogot’s neurosis is highly specific to the type of information contained in the proposition that p; (2) his compulsion is not easily changed; (3) while in the office he has no other easy trick of the relevant type to play on one; and (4) one arrives at one’s belief that p not by reasoning through false beliefs but by basing the belief that p on a true existential generalization of one’s evidence (Shope 1992).

Further variations on the case of tricky Mr Nogot challenge Alvin Plantinga’s (1993b) proposal that one’s knowing that p requires not only that one have a true, justified, sufficiently strong belief that p, but that one also meets the following requirements: (1) the cognitive faculties involved in the production of one’s belief are functioning properly in an environment sufficiently similar to the one for which they were designed; (2) the portion of one’s design plan covering formation of beliefs when in the latter circumstances specifies that such formation directly serves the function of forming true beliefs; (3) if those circumstances include additional beliefs or testimony, then the latter are or express beliefs also satisfying (2) – and so on, backwards through any chain of input beliefs or testimony from one person to another; (4) there is a high statistical or objective probability that a belief produced in accordance with that portion of one’s design plan in one’s type of circumstances is true.

Plantinga notes that the original Nogot case involves credulity, that is believing for the most part what one’s fellows tell one. When Nogot says that he owns a Ford, credulity helps to produce one’s false belief that f, which serves as further input to one’s belief-forming faculties so as to generate the true belief that p. Plantinga suggests that credulity was designed to function in circumstances where testimony from our fellows does not deceive us and is intended to produce true beliefs in us. Both of the latter circumstances are absent when Nogot brags about his Ford and one comes to believe that f.

But in the case of tricky Mr Nogot the informant does intend to produce a true belief in one. Of course, the case does not satisfy (4), since it is likely that one will not only believe that p, but also believe that f. So consider instead a variant of the case where one is in addition a clever and cautious reasoner of the type described earlier, highly likely to avoid believing that f, and suppose that Nogot realizes this. He then takes no significant risk of one’s forming a false belief.

The example may violate (3) if the purpose specified in the design plan for the use of testimony is to induce or sustain belief on the part of another in what the testifier says. But Plantinga admits that Nogot provides a wide variety of evidence that supports both the proposition that f and the proposition that p. So the Nogot case can be altered by imagining that he provides a large quantity of evidence involving no false testimony. Then it is obscure how Plantinga can show that tricky Mr Nogot is providing evidence for a non-designed purpose, since evidence never supports only one proposition, and we rarely, if ever, come to believe everything for which we have evidence.

Suppose that Plantinga is defended by the suggestion that the circumstances in which one’s relevant cognitive faculties were designed to operate exclude the type found when tricky Mr Nogot meets the clever reasoner. This response exposes a ‘generality problem’ in Plantinga’s account: what level of generality is relevant to specifying the circumstances pertinent to conditions (2), (3) and (4)? Plantinga rejects the possibility of a naturalistic account of proper function, and focuses attention on what a theistic God would have in mind when designing us in his image, part of which would involve our powers as knowers being analogous to God’s in so far as he has beliefs that are true and he, although not designed, fulfils a specification that is an ideal for cognitive design plans, namely, to believe that p if and only if it is true that p. So what does it matter that a neurosis is what motivates the way in which evidence is provided by tricky Mr Nogot, since true beliefs are ensured?

In any event, Plantinga’s requirements are too weak to deal with the following variant of the original Nogot case: Nogot is entirely sincere but his Ford has, unknown to everyone in the office, been repossessed or destroyed by a meteorite since he parked it. Requirement (3) implies that when one’s belief depends upon a chain of testimony, then at each stage the input propositions must have the right kind of credentials, namely, being a belief that has warrant for the testifier. But the sincere Nogot would have warrant for believing the truth that f in a variant of the above case where nothing untoward happens to his Ford. So it is difficult to see how the repossession or meteorite strike removes such warrant in the former case. (Moreover, the case of unlucky, sincere Mr Nogot could also be varied so as not to include any false testimonial evidence.)