# Re: tautology

From: Russell Standish <R.Standish.domain.name.hidden>
Date: Fri, 5 Nov 1999 11:12:42 +1100 (EST)

>
> On Thu, 4 Nov 1999, Russell Standish wrote:
> > > On Tue, 26 Oct 1999, Russell Standish wrote:
> > > [JM wrote] [&BTW I am getting tired of RS omitting the attribution]
> >
> > ^^^ Blame my email software. I almost always leave the .signatures in
> > to make it obvious who I'm responding to.
>
>
> > > It is obvious that p(Y1&X) = p(Y1&Z), because in all instances in
> >
> > It is not obvious, for the same reason that p(Y1&X) = p(Y2&X) is not obvious.
> > If QTI is true, then it is clearly not true. Don't assume what you're
> > trying to prove.
>
> Perhaps I should have been a little more clear. I am discussing
> the ASSA, not trying to prove it but to show that it is self consistent.
> You are right in the sense that I left something out. I am
> assuming a reasonable measure distribution based on the physical
> situation. For example, the measure could be proprtional to the number of
> implementations of a computation, as I like to assume.
> It is also possible to assume an unreasonable measure
> distribution, like the RSSA. This of course would require new, strange
> and complicated laws of psycho-physics.
> So what I am really doing is showing that (ASSA + reasonable
> measure (RM)) is self consistent. However, the way we have been using the
> term ASSA, RM has almost always been assumed.
> In any case it is always true that some way of calculating the
> measure distribution is needed. Your claim was that the RSSA is needed.
> My example shows that RM does the job.
>

My understanding is that ASSA cannot assign a probability to p(Y1|X)
or p(Y2|Z). Your reasonable measure presumably gives values for p(Y1|X'),
p(Y2|Z'), p(X) and p(Z). Now p(Y1)=p(Y1|X)p(X)+p(Y1|X')(1-p(X)) - it
seems to me likely that p(Y1|X)p(X) is negligible (although clearly
there are circumstances where it is not (eg when there is only an
"Adam" and an "Eve")) compared with the other term, so that p(Y1)
\approx p(Y1|X').

The real problem, and I have long pointed this out, is that absolute
measure is completely irrelevant to what one observes about
oneself. QTI is the assumption that p(Y1|X)=p(Y2|Z)=1, under
appropriate definitions of what X and Z mean.

I don't think your measure argument is wrong, or that ASSA is wrong,
its just that it doesn't disprove QTI. I don't adhere to QTI as an
article of faith, however, it seems more likely to be the truth than
not. If someone can come up with a good counter-argument to QTI, then
of course I'll modify my beliefs. I have tried to falsify QTI, but not
succeeded so far.

> - - - - - - -
> Jacques Mallah (jqm1584.domain.name.hidden)
> "I know what no one else knows" - 'Runaway Train', Soul Asylum
> My URL: http://pages.nyu.edu/~jqm1584/
>
>

----------------------------------------------------------------------------
Dr. Russell Standish Director
High Performance Computing Support Unit,
University of NSW Phone 9385 6967
Sydney 2052 Fax 9385 6965
Australia R.Standish.domain.name.hidden
Room 2075, Red Centre http://parallel.hpc.unsw.edu.au/rks
----------------------------------------------------------------------------
Received on Thu Nov 04 1999 - 16:12:53 PST

This archive was generated by hypermail 2.3.0 : Fri Feb 16 2018 - 13:20:06 PST