ascription. It is the practices of belief-ascription which need to track the
actual content of beliefs. This they do fairly well, though far from infal-
libly, in ordinary folk psychology.
In the last chapter we touched upon an aspect of mind-reading which
may account for the erroneous idea that rationality of belief is somehow
guaranteed by being built into principles of interpretation. We propose an
error theorywhich promises to explain why such an assumption of
rationality might seem to be at work. The error arises from the need to use
simulation as a way of enriching the stock of inferred beliefs which we
attribute to others. As we saw in the last chapter, simulators will be relying
upon their own inferential transitions. In any normal case (excluding
short-term doubts about one’s sanity) it must seem to simulators that their
own inferential processes are rational, otherwise they would not be
drawing the inferences they are. So it is true that simulators will generally
(unless discounting for known foibles and idiosyncrasies) represent the
target of their mental simulations as reasoning in a rational wayby the
simulator’s own lights. The error which our error theory diagnoses is that of
identifying rationality with the simulator’s own inferential dispositions.
That is understandable, of course, if the business of belief-ascription is
viewed from the perspective of the interpreter. But all that is required in
order for the process of inferential simulation to work well enough for the
purposes of folk psychology is that target and simulator should reasonin
much the same way. Correspondence may be achieved by their reasoning
equally well, but for purposes of prediction and interpretation they can
reasonequally badlyjust so long as they make the same mistakes. So after
seeingBlackcome up several times in succession an onlooker may be right
in thinking that one of the gamblers will expectRedto predominate in the
next few spins of the wheel – because they are both wrong in expecting this!
So predictive and interpretative success may be achieved through common
deWciencies, rather than perfect rationality.
Another variant on the argument from interpretation attempts to estab-
lish that believers’ true beliefs must vastly outnumber their false beliefs
(Davidson, 1975). The idea is that we can only ascribe false beliefs to
thinkers in cases in which we can identify what it is that they have false
beliefsabout. So in order for someone to believe something falsely, such as
thatthe gun in Edgar’s hand is unloaded, they have to possess numerous
true beliefs, such as thatthere is a gun in Edgar’s hand,Edgar has a hand,
and so on. However, this argument would only show that thinkers must
have very many more true beliefs than false beliefs (true byourlights, that
is), if wehaveto be able to identify and report what it is that they falsely
believe in a referential vocabulary to which we, the interpreters, are com-
mitted. It does not eliminate the possibility of a thinker having indeWnitely
Philosophical arguments in defence of rationality 113