individuals need a way of working out and keeping track of the costs and
beneWts involved in social exchange. And then that very same mechanism
will enable cheaters to be detected because it will be invoked both in
working out what is required by way of fulWlment of an obligation and in
discerning what is a case of defaulting. So perhaps it is something of a
misnomer to call this module the ‘cheater-detection module’ – ‘social-
contract module’ might be better. But we will retain the title since it seems
to have become accepted. Besides, it may well be that what provided the
initial impetus towards development of the module was selective pressure
towards cheater detection in cases of social exchange which werenot
innovative (possibly early forms of food-sharing).
It is important that the special reasoning mechanism postulated as a
cognitive adaptation for social exchange should not be (and is not taken by
Cosmides and Tooby to be) a processing system for reasoning about
prescriptive or deontic rules in general. For if all that we had in evidential
support for the hypothesis of a cheater-detection module were consider-
ations of evolutionary plausibility combined with data indicating a diver-
gence in performance on deontic and indicative versions of the selection
task, then the case for such a module would not be all that strong. This is
because there is a potential alternative explanation as to why people do
better on deontic tests. This is that deontic tests are easier because the
deontic selection task is adiVerenttask, and one which is easier to process.
The reason why this might be so is that in the indicative versions of the
task, subjects have toWnd the cards which need to be turned over in order
to establish whether the ‘rule’ or generalisationis true or not. That is why
the selection task is thought to show something about ability to test
hypotheses. But in deontic versions of the task there is no question as to
whether the rule is ‘true’ or not. In these cases it is a matter of stipulation
that a certain prescriptive rule applies. The subjects’ task is not toWnd out
whether a certain rule obtains. (That reallywouldbe diYcult. For while
philosophers of science have laid heavy emphasis on the underdetermina-
tion of theory by evidence, it is far more obvious that observed conduct
underdetermines prescriptive rules.) In the deontic cases subjects are asked
to detect violations of a rule, but there is no doubt as to what the rule is. In
the indicative cases, by contrast, subjects are presented with a rule or
generalisation, but the truth-value of that generalisation is uncertain. So in
eVect subjects need to engage in an extra level of processing in the indica-
tive selection tasks, because in order to solve them they have to ask
themselves: ‘Suppose this conditional applied. What would it rule out?’ In
the deontic tests that crucialWrst step is fed to the subjects, because they are
told that the rule applies and asked to spot what it forbids. So if what was
to be explained were just superior performance on deontic as compared
Psychological explanations of performance 121