156 J.J.C. Smart
information about the world in general could have evolved by natural selec-
tion. Whether or not the ‘homunculus’ strategy in neuroscience is the right
one I cannot see any philosophical objection to Haldane’s quotation from
Kathy Wilkes. Detecting horizontals (see p. 94) falls far short of intentional
or goal-directed behaviour.
A simple case of goal-directed behaviour is that of a predictor-controlled
gun. However the target twists or turns the gun will orient itself so as to have
a high probability of hitting the target. If the target had twisted differently
the gun would have moved itself differently. Similarly, consider a robot which
moves around obstacles in a room, however they are situated, to get to a plug
to recharge its batteries. All this is quite mechanistic. Such robots have of
course been built. In the case of organisms the mechanisms will have arisen
by natural selection. Now the goal-seeking behaviour ‘seeking to orient itself
to hit the target’ or ‘moving to recharge batteries’ is on any particular occasion
a sequence of gun movements or robot movements. However, the interesting
thing is that for the behaviour to be intentional there must be something
about the gun plus predictor or the robot’s inner construction that ensures
that if the target had moved differently the gun would have moved differently
so as to maximize the chance of hitting the target, and if the furniture had
been arranged differently the robot’s route to the plug would similarly have
been different. In more sophisticated sorts of intentional behaviour, as in
the human case of writing an article, say, a lot of pen movements are involved
but many alternative sequences of such movements would count as writing
the article. I agree that writing an article cannot be defined in terms of a
sequence of hand movements and even a particular instance of such writing
would be inadequately described as a sequence of hand movements. Things go
on in the mind. A materialist would not deny this, though he or she would
contingently identify the mental occurrences and control mechanisms with
brain events and brain states.
There is indeed something philosophically puzzling about intentionality
(with a ‘t’). This is because it is related to intensionality (with an ‘s’).
Intensionality is a matter of context. Consider ‘Joe believes that the head of
the philosophy department is the dean of the faculty of arts’. Suppose that the
head of the philosophy department is in fact the dean of the faculty of arts.
We cannot substitute for identicals here and retain the same meaning. To say
that Joe believes that the head of the philosophy department is the head
of the philosophy department is to ascribe only a trivial belief to Joe. Or again
consider ‘Joe wants a lawn-mower’. It does not follow that there is a lawn-
mower that Joe wants. Any lawn-mower might do. And what about ‘Joe
wants a unicorn’? There is no unicorn that Joe wants because there are no
unicorns. We cannot say that ‘wants’ signifies a relation between Joe and a
unicorn: it is not like ‘Joe kicks a football’ where there must be both Joe and