[ Pobierz całość w formacie PDF ]

that once we understand the mechanisms which ensure this, the consequence is
unworrying. Consider an analogy. Given the way that  I and  am here work, there is no
room for someone to be in error when they judge that  I am here . It falls out of the
semantic constitution of these indexical terms that any judgements of this form must be
true. Yet we do not on this account think that there is anything amiss with the semantics
of  I or  am here . I take the same point to hold for phenomenal judgements like  this is
an instance of an experience of red , and for not dissimilar reasons. Once we understand
the semantic workings of phenomenal concepts, we can see why this kind of judgement
cannot possibly go astray.
Now consider a slightly different case, where we use a state of perceptual re-creation,
rather than of perceptual classification, both to name some past experience, and to form a
type concept which we use to characterize that same experience. I visually imagine
seeing an elephant, and then use this act of imagination (i) to refer to some particular past
experience, and (ii) to characterize that past experience with a type-phenomenal concept
formed from the same act of perceptual re-creation.  That particular past experience was
an instance of seeing something as an elephant.
Perhaps in these cases of perceptual re-creation there is room for one kind of error that
doesn't arise in the perceptual classification case. I might fail to name any particular past
experience when I form the relevant subject term from my act of perceptual re-creation
( that particular past seeing of an elephant ) maybe because I have seen more than one
elephant and can't distinguish the occasions, or for some such reason. But if we put such
cases of reference failure to one side, then there is no remaining room for error, for the
same reasons as in the case involving perceptual classification. Your subject term can't
help but name an instance of the relevant type, if it names anything at all. For this type is
picked out as consisting of instances which appropriately resemble your act of perceptual
re-creation, while the particular experience which features as the subject of the judgement
is picked out as a specific instance of just the same kind of resemblance. Once more, the
semantic workings of phenomenal concepts remove any possibility of error.
4.12.2 Phenomenal Judgements Which Use a State of Perceptual
Classification to Identify an Experience, and a Different State of
Perceptual Re-Creation to Classify It, or Vice Versa
Now consider a rather different kind of case. I use a current state of perceptual
classification to form the subject term, and a state of perceptual re-creation to form the
characterizing type concept.  My current experience [identified via a state of perceptual
classification] is like that [and here an imaginative phenomenal concept is exercised].
These phenomenal judgements do not enjoy the same guarantee as those considered in
subsection 4.12.1. The subject term is here formed from a current state of perceptual
classification, but is then
end p.136
characterized by a type concept formed using a quite different perceptual state, a state of
perceptual re-creation. Since different states are used to form the two terms, there is no
inbuilt connection. A judgement of this form could take a current visual state of seeing
something as red, and characterize it falsely as like an imaginative re-creation of seeing
something as green.
Still, perhaps another kind of epistemological guarantee is possible here. Consider the
hypothesis that perceptual re-creation and perceptual classification may both be
underpinned by the same mechanism.15 Some kind of stored neural  templates may both
(i) be reactivated in perceptual re-creation and (ii) used to establish matches with
currently incoming stimuli in perceptual classification. If this is right, then it may mean
that judgements which classify current perceptual classifications by imaginative
phenomenal concepts are immune to error after all.
Look at it like this. Could you be mistaken in judging that a current state of perceptual
classification constituted by incoming stimuli resonating with some stored template
A is of a certain type the type (faintly) exemplified by activations of template B?
Well, maybe your making such a phenomenal judgement simply consists in A and B
being one template rather than two. On this suggestion, phenomenally judging that your
current perceptual state is of some imaginative phenomenal type would simply be a
matter of the same stored pattern of activation being used both in your identification of
your current perceptual state and in your typing it by your imaginative phenomenal
concept. Conversely, to judge that your current perceptual state is not of some
imaginative phenomenal type would simply be for two different stored templates to be
involved here.
If this suggestion is right, then it rules out any possibility of error when you judge  my
current experience [identified as a state of perceptual classification] is like that [and here
an imaginative phenomenal concept is exercised] . The template identity which
constitutes the judgement will simultaneously ensure that the judgement is true: since the
same pattern of activation is involved
end p.137
twice, the current experience referred to will inevitably be an instance of the type picked [ Pobierz całość w formacie PDF ]