I tried to send a reply with different content before, but apparently
it has not yet gone through or something. If it doesn't do so soon,
I'll have to investigate.
Here is what I posted to GMBM just recently. It's a better formulation
of the problem than anything I've published before, but not the formal
solution. This is not entirely the work of me, but I apparently have
permission to publish.
Mark,
okay. Why did you moderate away my report of the mistake in the CTMU?
I'll iterate it in better terms.
Langan believes the CTMU to satisfy the Law Without Law condition. But
he also believes it to have the M=R and MAP properties. Here he
implicitly assumes that concepts cannot be used ambiguously.
Ambiguous use of a concept means the concept is used in such a way
that essential information about the theory, in which the concept
belongs, is omitted. To be sure, according to the CTMU every concept
belongs to the CTMU, but this is not enough.
Even within the CTMU, we could, for example, use the concept "A number
whose successor is 0" ambiguously in such a way, that we would not
know whether the 0 is placed in the theory of natural numbers or in
the theory of integers. In this case, we could argue the concept to
have an empty extension, because the theory of natural numbers would
not include --1. But we could also argue the extension of the concept
to be --1, because 0 is the successor of that number in the theory of
integers. The reason why this is possible is that our use of the
concept does not include information that is necessary for defining
the concept unambiguously.
We could deem such use of a concept impossible. But that would be an
arbitrary declaration contrary to Law Without Law. There is no
inherent reason why it would be impossible to communicate in such a
way that information that eliminates ambiguity is omitted.
MAP is an attempt to deem ambiguous use of a concept as impossible.
The CTMU cannot have Law Without Law and MAP.
If we cannot prove that it is impossible to use a concept ambiguously,
we cannot state M=R except as an arbitrary axiomatic declaration. But
if we did that, CTMU again would not have Law Without Law.
-Tuukka