# [hist-analytic] Response to Steve's Latest

Bruce Aune aune at philos.umass.edu
Wed Jan 6 08:29:16 EST 2010

```Steve obviously spent a lot of time writing up his “Preliminary
Remarks on the ‘Two Color Problem’,” and because many of his
remarks are directed specifically to me, I have a strong inclination
to say something about it. But I have a contrary inclination as well.
Steve takes up many things in his document that I have already
addressed at length in previous contributions to our ongoing
exchanges, and since Steve appears simply to disregard much of what I
have said, I wonder if there is any point in saying more. Our
interchanges have been a lot like a conversation with a person who
listens politely to what you say but then goes on speaking as if you
haven’t said anything yourself. What you do say seems to go in one
ear of your interlocutor and out the other. Well, I have a free
afternoon. I’ll say a few things in the hope that Steve will actually
consider them, though he doesn’t have to respond. Our
“discussion” of the color-issue has probably gone on long enough.

1. My first remarks concern what Steve says at the beginning
of his paper about Putnam and meaning postulates. Intending to expose
an error he thinks Putnam makes in introducing meaning postulates for
a solution to the red/green color problem, Steve asks us to consider
“how a relation can
under one set of circumstances be transitive but under another
not.” I am not sure what error Steve thinks Putnam made (he doesn’t
clearly identify it), but Steve makes a gross error in setting forth
his example. If there is a single case in which, for a relation R,
there are individuals a, b, and c such that aRb, bRc, and not-aRc,
then the relation R is simply not transitive. To be transitive, R must
be such that for all x, y, and z, if xRy and yRz, then xRz. Steve also
errs in specifying the details of his example. If his relation is that
of having the same hammer, then if x does have the same hammer as y
and y does have the same hammer as z, then x and z would have to have
the same hammer. If the relation is otherwise—if, say, it is that of
being a co-owner of a (=some) hammer—it need not be transitive, but
Steve would have to say just what the relation is.

As I explain in my ch 3, meaning postulates (the label is Carnap’s)
are supposed to be specifications of meaning for descriptive
expressions, either partial or complete, in the context of a sentence.
(They are complete or partial contextual definitions.) For reasons
that I explain, Carnap regarded such specifications as essentially
stipulative rather than descriptive. If I stipulate that I am using
“R” to represent a reflexive relation, then if my stipulation is
properly introduced (as I put it in my appendix 4), no empirical
considerations could possibly falsify it. E.g. if I stipulate that I
am using “R*” to represent a symmetrical relation, then anyone
having evidence that not-(bR*a) (in the intended sense of “xR*y” )
can only conclude that “aR*b” is false). It is possible, of course,
that no pair of objects may stand in the relation R*, but this would
not invalidate my stipulation either. Quine, in his original “Two
Dogmas…,” made the unfortunate claim that the progress of quantum
mechanics might induce us to “revise” the law of excluded middle;
his claim was unfortunate because too many of his readers read him as
meaning that quantum-mechanical facts might falsify the law. This
would not be possible, because the law holds only for statements
satisfying bi-valence, and no such statement, no matter what its
subject matter or truth-value, could undermine the law. (Truth tables
show this.) Revising the law would amount to amending the meaning of
the truth-functional “or.” This of course could be done; there is
no a priori objection to that.

Putnam, in attempting to show by reference to certain meaning
postulates that certain vernacular statements are analytically true
(or true solely by virtue of their meaning) no doubt thought of the
ones he gave as descriptive for informed discourse about colors. Since
philosophers of most stripes agree that the same part of a thing’s
surface could not possibly possess two different color-shades at the
same time, it is not unreasonable to suppose that this incompatibility
may have a basis in the meaning they attach to color words. This
supposition could, of course, be false, but that is something Steve
would have to show.  If the postulates Putnam introduced do represent
meaning connections implicit in his or others’ talk, his claim that
the assertions of incompatibility he is concerned are analytically
true deserves to be accepted as holding true for those language-users.

As I explained in an earlier post, I don‘t think of analyticity as
Putnam did, or does, and I would not accept his version of the “two
color” problem. I do believe, as I said in my book, that conceptually
sophisticated speakers of English, the speakers Wittgenstein commonly
referred to as “we,” do generally conceive of specific visible
colors as being identical just when they are appropriately
indiscernible (not when they can’t be distinguished under special
conditions, e.g., the sort of conditions that Danny described in his
controversy with me). Because they generally conceive of colors this
way, they generally use their determinate color predicates to
represent properties having this identity condition. The fact that
they do use color words this way is not crucial to my position on
analyticity, however. People need not speak the way I do or in the way
I recommend. As I explain in my book (see pp. 101-105), the notion of
analytic truth should ideally be relativized to particular language-
users or thinkers in specified contexts of speech or thought.

2.  I have things to say about almost every paragraph of Steve’s
discussion, but I must restrict myself to the most significant ones.
In his section G he includes some quotations from Wittgenstein, which
None of W’s claims is problematic for me. Steve seems to have
forgotten my story about Tom, Mary, and Harry (see ch 2, pp. 64ff and
ch 3, pp. 102f). Here we have three people who classify an unusual
color in three different ways and whose concepts of green and yellow
partly diverge, though they overlap for most other cases. My claim,
remember, is that there is actually no metaphysical or conceptual
compatibility between red and green or any other generic color
qualities. Specific shades of color (determinate color qualities) are
conceptually incompatible, but shades disjoint on the color wheel
could conceivably be classifiable together under a common generic
color. And of course we can always speak of colored things with
varying degrees of strictness.

In his section G Steve also says, “Sometimes it seems to me that
Bruce thinks you can

solve certain problems, the ones under discussion having to do with
defining, etc. “determinate colors” by simply going to the paint
store and asking to view a color wheel.”  It should not seem this way
to him if he remembers why I spoke of going to a paint store.  I was
making an objection to a bad argument he introduced to show that there
are no determinate (or discrete, a word he also used) colors. As I
said in my response to his Nov. 10 post, “Steve, holding fast to the
continuity idea …, thinks that the existence of humanly imperceptible
color differences (which he accepts) shows that there are no
“discrete colors”.  As he puts it, “if it is always the case that
between any two colors there is a third, and if the discrimination of
individuals is limited, then there will be no discrete colors for
these individuals.”  But if discrete colors are recognizable colors,
this contention is absurd; it involves the logical howler I accused
him of [in an earlier post]. From the fact (if it is a fact) that
there are differences between shades of red that I cannot recognize,
it hardly follows that I cannot distinguish any shades of red at all--
green. The analogy I drew [in the earlier post] between real numbers
ordered by SMALLER THAN and shades of color ordered by the supposedly
dense relation Steve seems to have in mind is, in fact, sound. In both
cases we have a conditional assertion, “(x)(y)(xRy ® there is a z
such that xRz & zRy),” and to infer from this that some minimal term
z’ possesses a special minimal vale, we need a premise of the form
“aRb” that we can know to be true. For this premise, the value of
“a” and “b” need not [in fact, can’t] be minimal at all.”

In an earlier post concerned with the same argument I
mentioned going into a paint store.  Here is what I said: “I think
Steve has gone around the bend talking about color being a continuum.
Suppose I go to a paint store and buy a can of Forest Green paint. I
use it to plaint a lawn chair. (I have actually done this many times.)
Isn’t the chair I painted now Forest Green in color?  And isn’t
that a definite color? (It is in fact another example of what I call a
determinate color.)  Where is the color that is continuous with the
color of this chair? If you can find it for me, show me the color that
is between the two—and so on and so on and so on….  What reason is
there for believing that the colors we see belong to a color
continuum, a continuum of visible (seeable) colors?  I certainly
can’t make infinite discriminations.  My computer monitor is capable
of displaying “millions of colors,” but not infinitely many of
them.  Am I supposed to be capable of discriminating more colors than
my computer can display?”

3. Steve has some odd things to say about identity.  He says
he doesn’t believe that the identity relation applicable to objects
of visual awareness is transitive.  He supports his belief this way:
“Suppose two things are the same color (I am looking at them). But
there is a third thing that is identical in color to the second but
not the first.  Then it would appear that I have at least three
colors, although only two are distinguishable.” Why do we have three
colors? If the first two are identical in color, there is one color
that they both have. If the third thing is not identical in color to
the second thing, then we have two colors, not three. But how is this
supposed to be possible? Let the three things be A, B, and C, and let
“C(X)” mean “the color of X.” If C(A) = C(B), how can C(B)
possibly = C(D) but C(D) not = C(A)? If identity is not transitive,
what is? Steve can’t cast doubt on the transitivity of identity
merely by supposing that transitivity fails for the colors of three
particular objects.

4.  In spite of all I have said in response to Steve’s
it.  I have said (nearly a dozen times) that I use “determinate” as
a contrast to “generic” or (to use W.E. Johnson’s favored term)
understood what Steve’s problem is. Isn’t the distinction between
the generic and the specific (or determinate) elementary?  Could a
thing be a mammal without being a dog or a cat or a snake or some
other kind of mammal?  Obviously not. And could it be a mammal of a
certain kind without being a particular instance of that kind?
Obviously not. Well, could something be colored without being red or
blue or green or some other generic color?  Obviously not.  And could
it be red without being some shade of red—vermillion, scarlet, red
madder?  Of course not. If scarlet is not itself generic—if it is not
exemplified by things different shades of scarlet—it is determinate;
if it is generic, things possess it only by virtue of possessing
specific color-features appropriate to that genus.

5. Steve extracts from Putnam an interesting example that brings out
an important fact about indistinguishability as a criterion for
sameness of determinate colors, a fact that I have not yet commented
on: Groups of colored objects can be so similar to one another in
color that if they are compared in a certain order, their color-
differences will be pairwise indiscernible. As an example, consider
objects A, B, and D. If A is compared with B, even the keenest human
viewers will be unable to discern a color-difference between them. The
same is true of B and D. But if A is compared with D, a color
difference will be apparent. Does this show that the principle I
defended, “C(x)= C(y) iff IND[C(x), C(y)],” is false? I think not.
What it does show is that the discernibility of color differences may
be direct as well as indirect. I say this because in the case I
described, we discover the difference between C(A) and C(D) by their
distinguishability. This difference, together with the principle of
transitivity for color-identity, permits us to infer that either C(A)
does not = C(B )or C(B) does not = C(D).  So we know that one of the
pairs we cannot directly see to be different in color is different
nevertheless. We know this indirectly—by inference from a difference
we know directly. The criterion for color-sameness that I have
defended should therefore be understood to be indistinguishability
both direct and indirect. If a difference cannot be discerned either
directly or indirectly, the relevant colors are the same.

6. In his section “On The Significance of the ‘Two color
Problem’,” Steve considers what we should say if “we take” the
two-color proposition to be synthetic. What we should say should
depend, I think, on what we regard as true. The proposition contains a
modal auxiliary, “can,” together with “no,” which adds up to
the idea of impossibility. What kind of possibility is this supposed
to be? In my book I mentioned that, owing to the structure of our
eyes, it appears to be physically impossible for anything to be seen
as both red and green. So if the relevant “cannot” is understood as
connoting physical impossibility, the two-color proposition may well
be synthetic and true. But philosophers who think it is true do not
base their assessment on scientific evidence; they think of the
relevant “cannot” as representing a kind of possibility that can be
known a priori. Steve himself views it this way; he thinks of the
proposition so understood as representing an item of a priori
knowledge. Empiricists would not agree with this opinion; they do not
believe that this kind of knowledge is possible (to use Kant’s
language). Kant’s Critique of Pure Reason was explicitly concerned to
show how such knowledge is possible (in the three areas where Kant
thought we had such knowledge), but Kant’s efforts were pretty
clearly a failure. (I have taught the Critique at least twenty times,
chapter of my book to criticizing contemporary attempts to defend the
existence of such knowledge. So I am convinced that it is not
“possible.” But I did think I could show that the two-color
proposition, qualified to apply to specific rather than generic
colors, can reasonably be regarded as analytic. That, I believe, is an
important result; it solves a problem that has been around for a very
long time.

7.  A final point. In his section “EPISTEMOLOGY vs. ONTOLOGY” Steve
comments on the relative priority of epistemology and ontology. For me
the issue is clear: epistemology is the fundamental part of
philosophy. Why do I say this? Because I think philosophy (at least if
it is a worthwhile subject) is a critical subject: its tenets are not
matters of faith, like some religion, but conclusions resting on
evidence. Epistemology, as  see it, is primarily concerned with the
nature of knowledge and evidence. For that reason alone, it is the
fundamental part of philosophy. I might add that I have written a long
book on metaphysics, so my views on ontology are well considered
rather than impressionistic. I might also say, since Steve speaks of
certain crowds in philosophy, that I have never aspired to be a member
in good standing of some philosophical crowd. When one of my students
or another philosopher begins to speak about what “we” believe in
philosophy, I always get nervous. I don’t want to be a fellow
believer in some philosophical creed.

Bruce Aune

January 6, 2010
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://rbjones.com/pipermail/hist-analytic_rbjones.com/attachments/20100106/ec8aba0c/attachment-0002.html>
```