which need go unexplained. For even this freedom to conceptualise can
itselfbe explained on such an account, as we shall see.
So, we agree that absent or inverted phenomenal feelings are concep-
tually possible. Thus we can conceive of the possibility of undetectable
zombies. These would be people who arefunctionallyindistinguishable
from ourselves, who act and behave and speak just as we do; but who are
entirely lacking in any inner phenomenology. Equally, we can conceive of
the possibility of inverted phenomenologies (seeWgure 9.1 below). We can
conceive that other people, when they look at something red, have the kind
of subjective experience which I should describe as an experience of green;
and that when they look at something green, they have the sort of ex-
perience whichIget when I look at a ripe tomato. But becausethey
describe as ‘an experience of green’ whatIdescribe as ‘an experience of red’
(and vice versa), the diVerence never emerges in our behaviour. We both
say that grass is green, and causes experiences of green, and that tomatoes
are red, and cause experiences of red. Do these facts aboutconceivability
show that the subjective aspects of our experiences themselves must be
both non-representational and not functionally deWnable?
No. The best response to these arguments is to allow that absent and
inverted feelings areconceptuallypossible, but to point out that it does not
follow from this that they are logically (metaphysically) possible – even less
that they arenaturallypossible; and to claim that only the latter would
establish the actual existence of qualia. In fact the argument falls prey to
essentially the same weakness as the ‘what Mary didn’t know’ argument.
We can allow that there are recognitional concepts of experience and of the
way subjective states distinctively feel, and we can allow that thosecon-
ceptsare not relationally or causally deWned, while insisting that the
properties which those concepts pick outarerelational ones. It is because
the concepts are recognitional that absent and inverted feelings are concep-
tually possible. But it is because the properties which those concepts pick
out are actually relational ones, that absent and inverted feelings are,
arguably, neither naturally nor metaphysically possible.
To elaborate this thought is, in fact, to develop a higher-order thought
(HOT) account of phenomenal consciousness, to be discussed in section 3.
The idea is, that it is by virtue of having HOTs about our perceptual states
- and in particular, by deploying recognitional concepts of experience –
that those states come to possess their phenomenal properties. To aWrst
approximation: any creature which can perceive red and which can make
all the visual discriminations which I can, and which can recognise its own
perceptual representations of red as and when they occur, willipso facto
be a subject of just the same phenomenal feelings as me, on such an
account.
242 Consciousness: theWnal frontier?