On 10/23/2020 3:48 AM, Bruno Marchal wrote:
On 20 Oct 2020, at 20:26, 'Brent Meeker' via Everything List
<[email protected]
<mailto:[email protected]>> wrote:
On 10/20/2020 5:44 AM, Bruno Marchal wrote:
On 15 Oct 2020, at 22:53, 'Brent Meeker' via Everything List
<[email protected]
<mailto:[email protected]>> wrote:
On 10/15/2020 12:46 PM, Jason Resch wrote:
On Thu, Oct 15, 2020 at 1:56 PM 'Brent Meeker' via Everything List
<[email protected]
<mailto:[email protected]>> wrote:
You should have read Vic Stenger's "The Fallacy of Fine
Tuning". Vic points out how many examples of fine tuning are
mis-conceived...including Hoyle's prediction of an excited
state of carbon. Vic also points out the fallacy of just
considering one parameter when the parameter space is high
dimensional.
Hi Brent,
Thanks for the suggestions. I did read Barnes's critique of
TFOFT ( https://arxiv.org/abs/1112.4647 ) and I just now read
Stenger's reply: https://arxiv.org/pdf/1202.4359.pdf
I think they both make some valid points. It may be that many
parameters we believe are fine tuned will turn out to have other
explanations. But I also think in domains where we do have
understandings, such as in computational models (such as
algorithmic information thery: what is the shortest program that
produces X), or in the set of all possible cellular automata that
only consider the states of adjacent cells, the number that are
interesting (neither too simple nor too chaotic) is a small
fraction of the total. So there is probably fine tuning, but it
is, as you mention, extremely hard to quantify.
But my general criticism of fine-tuning is two-fold. First,
the concept is not well defined. There is no apriori
probability distribution over possible values. If the
possible values are infinite, then any realized value is
improbable. Fine tuning is all in the intuition. Charts are
drawn showing little "we are here" zones to prove the fine
tuning. But the scales are sometimes linear, sometimes
logarithmic. And why those parameters and not the
square?...or the square root? Bayesian inference is not
invariant under change of parameters.
At least for the cosmological constant, there seems to be some
understanding of its probability distribution, and it is
relatively independent of the other parameters in that it is
unrelated to nucleosynthesis, chemistry, etc. Therefore it is our
best candidate to consider in isolation from the other parameters
in the high-dimensional space.
Second, calling it "fine-tuning" implies some kind of process
of "tuning" or "selection". But that's gratuitous. Absent
supernatural miracles, we must find ourselves in a universe in
which we are nomologically possible. And that is true whether
there is one universe or infinitely many. So it cannot be
evidence one way or the other for the number of universes.
Let's say we did have an understanding of the distribution of
possible universes and the fraction of which supported conscious
life. If we discover the fraction to be 1 in 1,000,000 would this
not motivate a belief in there being more than one universe?
No, because it is equally evidence that one universe (this one) was
realized out of the ensemble. You are relying on an intuition that
it is easier to explain why all 1,000,000 exist than to explain why
this one exists. But that's an intuition about explaining things,
not about any objective probability. Every day things happen that
are more improbable than a million-to-one.
You need to take all the histories, which we know exists in arithmetic,
I don't know what "exists in arithmetic" has to do with existence.
We cannot prove the existence of any universal machinery without
assuming at least one of them.
Which is the same as saying the "proof" means nothing since it is the
same assumption.
Any of them will do, so I can assume at the start the one most people
are already familiar with: very elementary arithmetic (Robinson
Arithmetic). So I assume 0, s(0), s(s(0)), etc.
By exist (fundamentally, or ontologically, or “really”) I mean the
existence of 0, 1, 2, 3, … Only that exists ontologically.
By an observer I mean a number n such that phi_n is a Turing universal
and Löbian function (or n is Turing Löbian machine, and I note “[]”
its provability predicate. Incompleteness imposes to that
machine/number to distinguish the 8 modes (that I have described many
times, OK?).
For each mode [i] we have a corresponding notion of phenomenological
existence. With [0] = Gödel’s beweisbar ([]), and [i] = one of the
seven remaining modes, for example [1]p = [0]p & p, [2]p = [0]p & <0>t
& p, etc. (p always sigma_1).
Psychological existence can then be defined by [1](Ex [1] P(x),
physical existence is something like [2]<2>(Ex [2]<2> P(x), etc.
There is only one reality (the “standard model of arithmetic”), but
with many internal modes which are the modes corresponding to the
modal variant of “[]”, which are imposed to incompleteness. Both
psychology and physics are given by different modes of view, on the
same (arithmetical) reality.
Again, I could use any universal machinery, they all gives the same 8
modes.
It's all very well to hypothesize a theory. But then you must show that
it agrees with experience and tells us something we didn't know.
Brent
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/everything-list/5e231fc5-1dda-6f18-8e50-a491791c4fa8%40verizon.net.
