From: Robert Bradbury (bradbury@genebee.msu.su)
Date: Sat Jan 22 2000 - 05:23:05 MST
On Fri, 21 Jan 2000, Dana Hedberg wrote:
> I wrote:
>
> [Stuff about the structure and "value" of information that I will
> comment on in another note. snipped...]
>
> > If 50% or more of
> > natural conceptions end in abortion (according to one issue of
> > Science), then there would appear to be lots of faulty working sets.
>
> I'm a little confused on terminology here, Robert. The above paragraph
> suggests that you are indicating that natural abortions arise due to the
> incompatibility of the "working set" of genes that comprise the fetus. I
> might actually place more emphasis on the mother's fault tolerance as a
> "gestation machine" in these cases.
Presumably the mother's error checking code is much less "accurate"
than that of the growing organism itself. I think the fraction of
the genome implicated in development is at least in the 20-30% range
(i.e. if your mutation rate is high enough to result in a single
"significant" mutation in any one of 20% of the genes the development
process fails). More genes are probably involved in development
but the mutations don't stop development because there are redundant
backup systems that can function in spite of the mutations. This is
why many transgenic gene knockout animals are surprisingly (and annoyingly
to the scientists) "fine" from a health perspective.
While there is clearly some "fault tolerance" mechanism that ages
and leads to things like increases Down's Syndrome, I would doubt
very much it is the primary mechanism. It also may be acting
*before* actual conception. Presumably there is some failure
rate in the production of sperm and eggs and a "robust" system
would have some process for removing as many failures as possible.
But biology as we understand it now doesn't seem to have very
good mechanisms for allowing a potental parent to "inspect" the
sperm or eggs for problems.
Whether spontaneous abortions are due more to (a) gene sets that
"lethal" to themselves (due to double mutants in essential
genes or gene incompatibility) or (b) essential developmental
failures (due to stochastic errors) may be difficult to say at
this time. It would certainly require an extensive exploration
of the literature to get a better understanding of the processes
involved and their frequencies.
>
> I commented on old adults having roughly the information base
> of young adults resulting in equivalent information valuations...
>
> I would disagree with your assessment that killing an old adult is about
> the same as killing a young one, in terms of information loss. I agree
> that memories are probably gathered linearly, but the older adult has so
> much more base information with which to form new thoughts. A silly
> example:
> Young adult has items of information: A and B | Total info: A, B, and
> AB. Where AB is the set of relations between the two initial pieces.
> Old adult has items of information: A, B, and C | Total info: A, B, C,
> AB, AC, BC, and ABC.
>
> Now when each of these learns something new, D, then the older adult
> would seem to "get" much more out of that new piece of information, no?
This would presumably be true. My comments were primarily intended to
contrast the relative difference in rates of information accumulation
and relationship establishing in children vs. adults. Children may
be learning things like new words at the rate of 10/day, while adults
may be learning them a a rate of 10/year (as a gross example). Unless
the elderly adults are in highly stimulatory environments, I expect
the information accumulation to occur at a declining rate (because
they know more of what is required to survive and/or have constrained
their environment to match their survival knowledge base).
I suspect given the architecture of the brain, that it may be
difficult to establish the complete set of relationships that you
discuss. There is probably some limit on the number of links
a piece of information might have. Just as in computers where
files may be limited to 2^8 or 2^16 links, so to the brain has
geometry constraints. If you look at Calvin's hexagonal architecture,
the limit might be 6 neighbors. If you look at neuronal interconnects
it might be a few thousand neighbors. Since the "strength" of
neural connections seems to be correlated with their survival
significance and since in most "elderly" survival "crisises" are
rare (relative to children or adolescents), I suspect their information
relationship interconnect rates and strengths are lower.
Looking at the absolute information loss (due to death), I'd put the values
at something like: Embryo: 1, Child: 100, Young Adult: 1000, Old Adult: 1100.
Looking at the "potential" for information "gain" lost, its probably
more like: Embryo: 10, Child: 1000, Young Adult: 200, Old Adult: 100.
(The low value for the embryo being due to the high probability that
it would not have made it to the "child" stage anyway.)
You can quibble with the numbers but the point is to consider the
total information content or rate of information accumulation of
an individual. Our current legal/moral systems seem to only have
two or three slipper sizes when we probably need dozens.
Robert
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:26:22 MST