[ExI] Increasing coherence over increasing context? Or Truth?

[Jef Allbright]

On Mon, May 18, 2009 at 10:23 PM, Lee Corbin <lcorbin at rawbw.com> wrote:

> Jef writes
>>>> Lee, my purpose in raising this issue will be served when it's
>>>> recognized that you simply can't "model" an "unattainable final goal."
>>>> What is the behavior of this element?  Precisely how does it
>>>> constrain the behavior of the model as a whole?

<snip>

> (Eternal Truth #2: every statement must be
> further modified.
> Hence "around the sun" must be modified to
> be "about the common center of gravity",
> "planets" must be modified to include a
> great many other bodies in the vicinity,
> "around" must be understood to not be in
> defiance of general motion due to galactic
> rotation and speed of the galaxy itself,
> and so on, literally forever.)

This "framing problem", similar to the "grounding problem", should be
commonplace to any eight year old philosopher.  Pragmatically, it's
not an issue, just as nature doesn't have to compute the infinite
sequences of digits in pi each and every time a soap bubble is formed
[cf. Bucky Fuller.]


> But still, it is absolutely important to
> be able to embrace the realization "planets
> circle the sun"---and to embrace it as
> correct, without any beating around the
> bush.

My point is that it's /not/ "absolutely" important (from who's
perspective, able to judge relative importance as if outside the
system?), but like nature building soap bubbles, it happens however it
does, and we as embedded observers construct theories of "reality"
tending toward increasing coherence, adapting to increasing context of
observations.

Some people reading this will think "okay, but it's splitting hairs,
debating angels on the head of a pin, of no practical consequence", so
this seems a good place for an assurance that the increased
sophistication I'm aiming for is of extremely practical importance and
in any case we'll get there (or not) with a great deal of unnecessary
suffering (or not) assuming that our evolutionary branch continues.


>> Or mightn't you agree with me that
>> his theory was more *coherent* within
>> the broader context of observations
>> of his time?
>
> No offense, but to me that's an unnecessarily
> obscure way of describing him and his theories.
> He was way off track about how the solar system
> basically works, even though his approximations
> for the sake of predictions were praiseworthy.
> So, to answer your question: no.

So you say "his approximations for the sake of prediction were
praiseworthy" but to you that does't mean his new theory, compared to
the popular prior, wasn't more coherent within the broader context?
Can you explain to me what "coherent" means to you?


>> In the same light, you might recognize that evolutionary "progress",
>> of which scientific progress is an instance, does not move
>> teleologically toward any particular goal, but rather, proceeds by way
>> of exploring its adjacent possible, responding to (local) regularities
>> (not Truths) with persistent structures reflecting the increasingly
>> probable.
>
> Actually, I agreed more with what you wrote to
> Max.
>
> I believe that (society parameters allowing)
> our scientific progress does move closer to
> pinning down reality (though I hasten to add
> as always that this is a never ending process).

The point is not whether scientific progress does or doesn't tend to
move toward an improving model of reality.  I think we can both agree
that it makes sense to assume it does.

The point is that from any particular point of view in this evolving
system, i.e., at any *particular* location we may presently inhabit on
the path of evolutionary contingincies, there is no way to know
whether we are moving closer or farther from Truth (what actually
works in the bigger picture, were we to know it), nor how long we may
have been moving so.  To do so would require a context greater than
that available to us.

>From ***Ptolomey's point of view***, was he moving closer to absolute
Truth?  With sufficient sophistication, he'd have known that he had no
way to tell, nor did it matter,  but he could indeed demonstrate that
his new improved model of celestial mechanics was more coherent within
the greater context of observations available to him at that time,
with the pragmatic benefits that entailed. That same principle applies
here and now.


> To me, a term such as "the increasingly probable"
> hides more than it makes clear. Something is
> probable?  Probably what?  Probably so?  True?  Oh.

Probable in the sense of more likely to be observed, to exert an
influence, to be detected, to make a difference in the structure of
its surroundings.  Evidence of a likelihood function, not a statement
of Truth.

<snip>

> As I say, I find many of your descriptions and the
> concepts you use to be useful, but they come rather
> late in one's epistemology, resting upon more
> basic things. Your kids and employees, of course,
> often really are looking for "yes" and "no" (I'm
> hardly telling you anything, of course), black or
> white, or particular shade of purple-gray, (as a
> mauve) when you try to convey a more nuanced
> outlook. Well, most of us in daily life use the
> concepts of basic realism.
>
> And so do you, of course. If the cop pulls you
> over, you'll either argue that you were *not*
> speeding or admit that you were, or some mixture
> of the two. You won't dare employ any talk of
> his increasingly adaptive response to evolving
> traffic flows and his own perceptions on the
> one hand vis a vis yours, with nothing like the
> actual "real" (sic) velocity being admissible . :)

Yes, Lee, what works is entirely dependent on context.  But if I
wanted to get people thinking about systems of social decision-making
seen as increasingly ethical, then yes, I would be using phrasing and
concepts similar to those I share with my friends online.

And to the extent we are dealing with the regularities of daily life,
then we do well to exploit the fast and frugal heuristics of our
evolutionary heritage. But the local effectiveness of heuristics comes
at the expense of reduced context.  Thus we have the effectiveness of
belief in authority, the cohesiveness of the in-group and its codes
(religion, cults)... and belief in an absolute Truth.

[Was the Truth describing the nature and existence of Rush Limbaugh
present in the early phases of our Universe?  Is that conceivable from
an information-theoretic point of view?  I know, it's hard to conceive
even now...]

But to the extent that we are trying to make effective predictions
about an increasingly uncertain future, then there is a moral
imperative to strive for increasing coherence, within increasing
context--applicable to knowledge of our present but evolving values,
and to our present but evolving instrumental (scientific) methods for
their promotion.

It's the best we can do.  That's why it matters.  That's why I keep
pressing the point.