AI and cognitive psychology rant (getting more and more OT - tell me if I should shut up)

[Stephen Horne]

On 1 Nov 2003 22:19:11 -0800, mis6 at pitt.edu (Michele Simionato) wrote:

>> most current theory is so
>> abstract that the explanations should be taken as metaphors rather
>> than reality anyway.
>
>True.

Thank god - someone actually understood that bit!!!

Except you could have agreed with your own misunderstanding of what I
meant, I suppose - but lets agree to ignore that option ;-)

>According to the old school of Physics, there is a large distinction
>between fundamental (somewhat microscopic) Physics and
>non-fundamental (somewhat macroscopic) Physics. The idea
>is that once you know the fundamental Physics, you may in principle
>derive all the rest (not only Physics, but also Chemistry, Biology,
>Medicine, and every science in principle). This point of view, the
>reductionism, has never been popular between chemists of biologists, of 
>course, but it was quite popular between fundamental physicists with
>a large hubrys.

I think I understand what you mean.

I am aware of the idea, though I haven't really considered what it
implies. Certainly we haven't discovered any fundamental layer of
physics yet (AFAIK) and we may never do. And even if we do discover a
baseline level, it may be that we can never express the higher levels
deterministically in baseline level terms (as Goedel says, there are
relations that can never be proven or disproven - the incompleteness
theorem).

If forced to take a position, I would say that the key requirement is
that each model be consistent with all other models at all layers of
abstraction over the range where all are applicable. A higher level
layer may have features that cannot be derived from the lower level
layers, but they cannot contradict each other unless you go outside
the scope where one or more of the models is applicable.

>Now, things are changing. Nowadays most people agree with the effective 
>field theory point of view.  According to the effective field theory approach,
>the fundamental (microscopic) theory is not so important. Actually, for
>the description of most phenomena it is mostly irrelevant. The point is
>that macroscopic phenomena (here I have in mind (super)conductivity or
>superfluidity) are NOT simply microscopic effects en mass: and in
>certain circumstances they do NOT depend at all from the microscopic theory.

OK - but if you are describing superfluidity as a single macroscopic
effect then you must describe it within a macroscopic framework. At
which point it has nothing to do with quantum effects because it isn't
within a quantum framework - it is just that the macroscopic
phenomenon called electricity (distinct from electrons moving en
masse) is not subject to the macroscopic phenomenon called resistance
(distinct from energy loss through the electomagnetic interactions
between electrons and atoms en masse) when the macroscopic phenomenon
called temperature (distinct from the kinetic energy of atoms en
masse) is sufficiently low.

There is nothing wrong with this per se - it is the limit of most
peoples (mine included) understanding of superconductivity - but it
has nothing to do with the framework of quantum mechanics.

The quantum framework may give an explanation, of sorts, for why
superconductivity occurs (or perhaps Goedel has put his veto on this)
but I do understand why explaining something less abstract to our
perceptions in terms of something even more abstract might seem
counterproductive ;-)

>That's life, but it is more interesting this way ;)

Agreed ;-)


-- 
Steve Horne

steve at ninereeds dot fsnet dot co dot uk