11
the adult scientist. How much of the scientist’s data would have to be misleading for the scientist
to be unable to arrive at a satisfactory theory? “[S]uppose that a scientist were presented with
data, two per cent of which are wrong (he doesn’t know which two per cent). Then he faces
some serious difficulties, which would be incomparably more serious if the data were simply
uncontrolled experience, rather than the results of controlled experiment, devised for its relevance
to theoretical hypotheses” (Chomsky quoted in Lyons, 1991: 135-36).
In the case of all movement being structural, or sensitive to parts of speech, we have an
example of what Chomsky calls a linguistic universal, something common to all natural
languages. The child is able to compensate for incomplete or otherwise misleading input.
Chomsky’s conclusion is that certain principles for constructing sentences are innately known,
being forced by genetic constraints. The innate linguistic universals constitute what he calls
“universal grammar,” the system of rules and principles common to all natural languages. For
Chomsky, it is a linguistic universal that movement relies on relations between parts of speech in
a sentence and not on, for example, word order (1971) (3).
A poverty-of-the-stimulus argument can be applied to moral knowledge. In reference to
moral competence, “whenever we see a very rich, intricate system developing in a more or less
uniform way on the basis of rather restricted stimulus conditions, we have to assume that there is
a very powerful, very rich, highly structured innate component that is operating in such a way as
to create that highly specific system on the basis of the limited data available to it...” (Chomsky,
1978: 241; see also 1988a: 152-53; McGinn, 1997: 44-9; Dwyer 1999, 2004). This is just an
appeal to common sense, but there are also controlled data suggesting that infants have rich
knowledge which is at least proto-moral.
One important element of moral judgment is the ability to distinguish objects which have
goals from those which do not. Infants can recognize goal-directedness insofar as they can