Consciousness and ConceptualismOne complication owes to the different re-identification constraints available (Chuard
2006a). It’s one thing to identify distinct objects—e.g., x at t and y at t+n—as falling under
the same concept: here, re-identification is nothing but repeated identification of instances of
a concept. It’s another to identify y as falling under concept c while remembering x in such a
way as to identify both x and y as being relevantly the same in that both fall under c. The latter
kind of re-identification is more demanding: it involves explicit memory of past encounters
with other instances, together with comparative judgments based on such memories. It’s not
clear we always meet such a requirement, even when it comes to non-demonstrative concepts: if
my memory is rather poor in recalling past encounters with a specific type (e.g., hexagon), this
needn’t undermine my ability to identify instances of that type. The less demanding notion of
“re-identification” (as mere repeated identification) is unproblematic in this respect. Yet Kelly’s
argument presupposes that the subject doesn’t just identify a determinate shade, but that they
also recognize the color thus identified as the same as one previously identified: it appears to rest
on the more demanding constraint.
Nor is it even clear, in fact, whether the less demanding re-identification constraint applies
to demonstrative concepts (see Chuard 2006a; Coliva 2003). As Brewer suggests in the passage cited
earlier, demonstrative concepts are meant to be context-dependent tools for categorizing objects
and their properties. And part of the context involves how one perceptually attends to the
relevant samples. In this sense, possession of such concepts may be quite fickle: demonstrative
concepts can be thought of as disposable classificatory devices, to be used in a given context, but not
beyond. Accordingly, if there’s any change in the perceptual scene, or if the perceiver shifts the
focus of their attention, etc., such contextual changes may suffice to ground distinct demonstra-
tive concepts (for the same sample, even). Hence, a subject may not be able to identify a color
as falling under the same demonstrative concept twice, simply because the first demonstrative
concept deployed isn’t available the second time around. This needn’t imply that demonstrative
concepts aren’t concepts, or that they aren’t concepts of kinds. Just like other concepts, they can
serve in conceptual discriminations or inferences (within a context), etc. But there’s no obvi-
ous reason to assume that re-identification ought to have a special status when it comes to the
possession-conditions of different types of concepts.^25
Finally, a worry about the extension of demonstrative concepts (Dokic and Pacherie 2001;
Kelly 2001a). Samples of highly similar but distinct colors can be arranged so as to be percep-
tually indiscriminable (despite their fine-grained differences): patch a’s color may be visually
indiscriminable from patch b’s, which is indiscriminable from c’s, even though a and c are visu-
ally discriminable, owing to their greater chromatic difference—perceptual indiscriminability is
non-transitive, that is.
Under the conceptualists’ proposal, b’s color falls under the demonstrative concept thisb
formed when attending to b’s specific shade. And since b’s color is perceptually indiscriminable
from a’s, it seems a should also fall under the demonstrative concept thisb—being indiscriminable,
they might be conceptually identified in the same way, one might think. But if a falls under thisb,
c should too, since c’s color is also indiscriminable from b’s. Hence, a, b, and c, fall in the extension
of thisb. However, the color of a and c are discriminable, which means that thisb applies, not just
to distinct colors, but to visually discriminable ones, even though thisb was supposed to be a fine-
grained concept of a determinate shade. Demonstrative concepts aren’t fine-grained enough,
after all (Dokic and Pacherie 2001: 195; see also Peacocke 1992: 83; Martin 1992: 757).
Conversely, since b is chromatically indiscriminable from both a and c, it seems it should also
fall under the respective demonstrative concepts formed when attending to a, thisa, and when
attending to c, thisc. Which means that b’s determinate color falls under at least three distinct
demonstrative color concepts: distinct, since thisa and thisc pick out different, and discriminable,