Clearly though, at some point, this bucket passing function cannot be
reduced further and becomes irreducible.  There is in fact an
irreducible minimum part requirement for all functional systems. I'm
not saying that there is no flexibility to various systems of
function.  What I am saying is that all functional systems require a
minimum number of parts and a particular orientation of these parts in
order for that particular function to be realized.  In this sense
then, all functions are irreducibly complex at some point or another.
Beyond this point, the particular function in question ceases
instantly - and this is true for all functions.

What happens now is that at increasing levels of functional complexity
(requiring longer and longer DNA codes), the density of beneficial
functions as compared to the total number of potential sequences in
sequence space, becomes less and less in an exponential manner.  It is
this exponential decline in the density of beneficial sequences that
quickly isolates beneficial functions in a vast ocean of nonfunctional
sequences that is simply too great for the mindless nondirected random
walk of random mutations (without the help of natural selection -
since nature cannot select between equally nonfunctional functions) to
wander through in any sort of reasonable time frame (i.e., trillions
upon trillions of years). In fact, there are no examples of evolution
producing a function at the multi-protein level of complexity where
all the proteins work together at the same time in a specific
orientation with each other.  It seems that at such a level of
complexity, the beneficial density of functions is so miniscule and
the isolation of multi-protein functions so great that the mindless
natural processes of evolution are simply helpless to evolve anything
at such levels of complexity.

Oh really?  Take the lactase function, for example.  It seems like the
minimum part requirement for the lactase function in living things is
around 480 amino acids in a flexible, but fairly specified
arrangement.  Some scientists, like Ian Musgrave, suggest that the
total number of lactase enzymes in sequence space may be as high as
10e90.  Considering that there are only somewhere around 10e80 atoms
in the visible universe, 10e90 seems like an extraordinarily huge
number.  That is, until you compare this number with 10e600 - the
total number of sequences in sequence space at this level of
complexity.  What does a ratio like 1 in 10e500 mean?  It means that
on average, a given amino acid sequence with the lactase function is
extremely isolated from another 400aa sequence chosen at random.  Such
isolation means that most life forms and their respective population
gene pool will not have anything close enough to even one of these
10e90 potential lactase sequences to evolve the lactase function.
Experiments and observations have been done that verify this
hypothesis.  Many species of bacteria have been observed for hundreds
of thousands if not millions of generations and have never evolved
this relatively simple single protein lactase function during that
entire time - even though it would be of significant benefit to them
if they ever did evolve this function.  Experiments like those done by
Barry Hall are extremely interesting in this regard since even
bacteria that historically hydrolyze lactose quite well are incapable
of evolving this function back again if their lacZ genes and their one
"spare tire" ebg gene are deleted - even over the course of tens of
thousands of generations in a strongly selective environment.  Hall
himself said that these bacteria had, "limited evolutionary
potential."

Now, what is it, I ask you, that "limits" the evolutionary potential
of a life form if not for the neutral gaps that are created by a
relatively low density of beneficial sequences in sequence space?  You
say that an "'irreducibly complex system' can be arrived at by
reduction of a 'reducibly complex system'".  This is certainly true,
but how did you get the reducibly complex system without first having
the irreducibly complex system with that particular function to begin
with?  At the lowest levels of functional complexity, such as
antibiotic resistance and certain single protein based functions, this
is not much of a problem.  However, when you start talking about more
complex functions requiring longer amino acid sequence of large single
proteins and especially multi-protein functions, the problems for
evolution become extremely difficult to overcome, even theoretically,
much less experimentally.

Hmmmm - this isn't even a good cop-out argument.  Even evolutionary
scientists refer to many DNA sequences as non-functional and mutations
that occur in these sequences as "neutral".  In fact, most mutations
that affect the DNA of "higher" organisms, such as humans, are
referred to as "neutral" mutations - in that natural selection cannot
recognize these mutations in a selectable way.  For example, out of
the 100 to 300 mutations that change human DNA in each generation,
only 3 or 4 of them are felt to be functionally selectable.  All the
rest are thought to be completely neutral with regards to natural
selection.  Of the 3 or 4 that are selectable, there is thought to be
a ratio of around 1000 to 1 in favor of a negative selection due to a
"detrimental" mutation.

Obviously then, neutral sequences are not only a reality, they are
obviously extremely common in the sequence space at just about all
levels of functional complexity.  The question is not if they exist or
if they are common, but if they really do expand in an exponential
fashion with increasing levels of functional complexity - as I have
proposed.

The fact is that the weight of evidence, as we currently understand
it, is strongly in favor of the existence of rather large stretches of
DNA in humans where mutations are selectably neutral.  You can
hypothesize otherwise as much as you want, but what evidence do you
have to support your position?

LOL - you are too funny.  First you argue against the idea of neutral
mutations, and then you try to come across as actually believing in
and supporting Kimura's work?  Please! In any case, Behe's ideas are
very much in line with Kimura's Neutral Theory of Evolution.  My
ideas, though a bit different from Behe's, are actually dependent upon
the existence of non-functional or "neutral" sequences/mutations.

My hypotheses are also testable in a falsifiable way and have a high
predictive power. What genetic test/prediction can you make that would
falsify evolution if you were wrong? Really, if this or that function
or level of complexity fails to evolve, you will not give up on the
idea of evolution since you can always say, "Well, perhaps the
organism wasn't right, or maybe the environment wasn't ideal."  Really
then, evolution becomes a non-testable and therefore a non-falsifiable
position.  It is actually the Intelligent Design Theory that is the
most scientifically supported position since it can actually be tested
in a falsifiable manner.

You snipped the language analogy so there is nothing more to say since
the information coded in DNA and even protein systems of functions are
not just language analogies, they are actually real
language/information systems.  If you don't understand that language
analogy, you cannot understand the problems for genetic evolution of
new functions - which are separated from each other like tiny islands
on vast oceans of non-functional sequence arrangements.

Oh really - then perhaps you can explain to me how a multi-protein
function can evolve one mutation at a time?

Did you actually read these references and abstracts yourself?  If you
had you would quickly recognize that all of these examples of
evolution in humans are based on one or two point mutations to single
protein functions. This is not the issue.  Such examples are very
common.  They are actually less difficult to evolve than to evolve a
single protein function "de novo" - as occurred with the evolution of
the nylonase function in bacteria.  The success of such examples of
evolution is based on the fact, that starting with a particular
function, variations in the *level* of that function are easy to
achieve with one or two point mutations - like your line of firemen
passing buckets of water.  A mutation may remove a few firemen from
the line and make the water delivery slower than before.  This change
in the level of function may actually be beneficial in certain
circumstances.  However, the statistical random walk required to
achieve this change in level of function is extremely short since a
very high ratio of mutations could achieve this change in the level of
a pre-established function.  However, problems arise when you want to
move from one type of function to another type of function.  How are
you going to evolve a brand new function "de novo"?  That is the
question.  For functions at the lowest levels of complexity, this is
not too much of a problem due to the relatively high ratio of
beneficial vs. non-beneficial sequences.  However, for functions at
higher and higher levels of functional complexity, this beneficial
density drops off in an exponential manner until evolution quickly
becomes hopelessly lost on an almost endless sea of meaningless
sequences.

Again, no examples of multi-protein functions where all the proteins
are working together at the same time in a specific orientation with
each other, have ever been observed and documented or even theorized
on paper in a reasonable manner.  If you know of such an example,
please do share it with me - but do your reading first.  Don't just
send me a bunch of references that do not show what you think their
titles say.

Sean