From: Julian Suggate <julian.domain.name.hidden>
Date: Mon, 03 Nov 2003 10:06:34 +1300

Brent Meeker wrote:

>>> Even the probability of observing a single large scale violation of the laws of probability is vanishingly small.
>>
>> According to *our* laws of probability, that is.
>>
>> But how can you make recourse to our laws of probability if there
>> are infinitely many universes which have different laws?
>
> What are the laws of probability that might differ? That less
> probable things happen more often than more probable one?

Heh. You are trying to define probability in terms of itself. I imagine
a corner of existence where (as a previous poster described) a die
returns 6 every time. My whole point is: our stubborn insistence that 1,
2, 3, 4, 5 *or* 6 are just as probable as each other would be laughed at
in such a corner of existence.

If we happened to live in this other corner of existence, our
statistical models would have developed along very different lines.

And isn't it a little naive to assume that us humans of Earth have the
only 'correct' statistical model, just because it happens to reconcile
with the results of our experiments?

I guess I am saying that math is not devoid of our physical context...
how can it be? That it has empirical elements to it.

Consider the case for us: For the sake of the argument, imagine that I
used math to produce some absurd result, then the mathematical reasoning
must be flawed. But presume I used only accepted rules of inference.
Then the rules of inference must be inconsistent. Reductio ad absurdum.

But to apply this argument, we must determine that the result is absurd.
There's the rub. Often an argument is absurd 'by investigation' or 'by
inspection'.

Sounds like empirical evidence to me!

So if I used our statistical laws to derive a 'proof' in this other
universe whereby it was just as likely to roll a 1 as it was to roll a
6, wouldn't the inhabitants of this universe use investigation and r.a.a
to disprove my inference laws and their conclusion?

And who would be 'correct' do you think?

Jules
Received on Sun Nov 02 2003 - 16:04:53 PST

This archive was generated by hypermail 2.3.0 : Fri Feb 16 2018 - 13:20:08 PST