Re: Perpetually Perplexed
From: Jim McGinn (jimmcginn_at_yahoo.com)
Date: 01/20/05
- Next message: Robert Karl Stonjek: "Paper: Ancient DNA"
- Previous message: Jim McGinn: "Joe's Unconditional Frequency Obfuscation"
- In reply to: Jim McGinn: "Perpetually Perplexed"
- Next in thread: Perplexed in Peoria: "Re: Perpetually Perplexed"
- Reply: Perplexed in Peoria: "Re: Perpetually Perplexed"
- Messages sorted by: [ date ] [ thread ]
Date: Thu, 20 Jan 2005 13:01:56 -0500 (EST)
Guy Hoelzer wrote:
> in article cs95qs$1qp6$1@darwin.ediacara.org, Jim McGinn at
> jimmcginn@yahoo.com wrote on 1/14/05 11:16 AM:
>
> > Guy Hoelzer wrote:
> >> in article cs3lf4$17b$1@darwin.ediacara.org, Jim McGinn at
> >> jimmcginn@yahoo.com wrote on 1/12/05 9:06 AM:
> >>
> >>> 2) Explain the logic/rationality of the neoDarwinistic
> >>> assumption that R is determined by genes IBD.
> >>> That's it! It's that simple. Don't give up now.
> >>
> >> I'll take a shot at this one.
> >>
> >> I will start with the existence of an individual who tends to
behave
> >> altruistically as a result of having a particular allele.
> >
> > And, of course, this behavior would not be possible
> > without the existence of many other alleles.
> > Correct?
>
> Correct. The effect of any mutation depends upon the genomic
environment in
> which this new mutation finds itself.
So you admit that the first sentence of your premise
introduced a concept that was not very accurate? Right?
> >> The probability that this allele is present in a particular other
individual
> >> is correlated with (not entirely determined by) the degree of
relatedness
> >> with this other individual.
> >
> > We're supposed to just take your word for this?
>
> I didn't think that this statement would require an explanation.
I would think it was obvious that since this supposition
is central to your argument that you'd be bending over
backwards demonstrate it's accuracy. Why aren't you?
(Because you know that any discussion of the actual
evidence obviates the inanity of the supposition
that such a coorelation exists. So what will Guy
do. Guy will continue talking in vague generalities
and avoid any kind of DEFINITIVE discussion at all
of the causative aspects of the phenomena under
discussion.)
> The
> constraint of vertical transmission of genetic alleles, as opposed to
> infective horizontal spread that can be independent of relatedness,
> generates this correlation.
I can't make any sense of this sentence. (And
anybody that claims they can is lying.)
> I was careful to use the word "correlation",
> because there are certainly instances where an individual shares a
> particular allele with a distant relative while not sharing it with a
closer
> relative.
Certain instances? I think your sensibilities are
lost in some kind of neoDarwinistic haze. You know
for a fact that in the vast majority of instances
there is little coorelation is any at all.
> Nevertheless, my argument only relied on a frequentist
> correlation. Can you think of any circumstance in which such a
correlation
> might not exist?
It's plainly obvious to anybody that cares to look
into it that there are, in fact, relatively few
instances in which there is any coorelation at all.
And even then the coorelation is relatively low.
> If so, my argument would fail under those conditions.
But, of course, you will never actually come to
grips with such because you've managed to keep
your thinking so vague that in your mind the
truth of this supposition is only of secondary
consideration.
> >> The strength of this correlation depends on many factors,
including the
> >> frequency of the allele in the population, but it is reasonable to
expect a
> >> positive correlation to exist as long as this locus is polymorphic
in the
> >> population.
> >
> > Pretty much goes without saying.
>
> That is pretty much what I thought about my assertion of a
correlation
> above, but you seemed to reject that one. It is hard to know what
will be
> rejected by somebody on sbe, so I tried to build a simple argument
with
> non-controversial bits.
Sounds like little more than an excuse for not
directly confronting the abundant evidence that
shows no correlation (or very little).
>
> >> The logic of equating "r" with the
> >> probability that this allele will be shared IBD is based on this
> >> correlation. It is a standard sort of simplifying assumption
> >
> > Simplifying assumption? Are you suggesting that
> > since it's a simplifying assumption, to use your
> > words, that therefore we should ignore it's
> > inaccuracy?
>
> No. I don't think that it is inaccurate. My claim was for a
correlation,
> which allows for controverting instances.
It "allows for controverting instances?" I suppose
you're going to introduce this as a principle that
provides you the rationale to ignore the fact that
in the vast majority of instances there is little
or no correlation at all. Right? Isn't this what
you are leading up to?
> It is "simplifying" because it
> summarizes a diverse set of inferences as a mean-field correlation,
Mean field correlation? You're really getting
creative here, Guy. You're just digging yourself
deeper.
> ignoring
> the details of particular instances in the process.
Ignoring details? What details? What are you
talking about?
> >> to ignore other possible correlations (potential sources of
information about
> >> the possible presence of the allele in another individual),
> >
> > You're saying it's standard to ignore "other"
> > correlations? What are these other correlations
> > and why is it, supposedly, standard to ignore
> > them.
>
> My comment was a general one regarding the process of scientific
modeling.
> I am allowing for the existence of other unknown correlations. Other
> correlations are often "ignored" due to ignorance, and sometimes
purely to
> simplify the model for the sake of clarity.
Ignorance = Clarity
Isn't this, really, what you are saying?
> I agree that ignoring known
> correlations can lead to inaccuracy, but there is sometimes value in
trading
> off accuracy for clarity. In any case, my post was a response to
your plea
> to explain some of the logic of Hamilton's model, and that is what I
tried
> to do. You may not like the logic of heuristic modeling, but this is
the
> kind of logic used by Hamilton.
As you've described it here, it would seem that
"heuristic modeling," involves eluding to
nonexistent coorelations and employing ignorance
to bring about clarity.
>
> >> and allow the maximum possible predictive value of IBD
probabilities, to ask
> >> whether information about kinship alone could suffice as a
selective
> >> mechanism promoting an increase in frequency of this allele.
> >
> > Kinship a selective mechanism? I can't make
> > any sense of this.
>
> The vertical transmission of genetic information through the
reproduction of
> active agents constrains the possible distributions of genetic
variation in
> a biological population (see above). Constraints are universally
necessary
> to the manifestation of mechanism; otherwise cascades of cause and
effect
> would be free to follow any path. Mechanisms exist due to sets of
> constraints that channel cause-effect cascades along particular
paths, which
> is what makes outcomes predictable and mechanisms useful. The
constraint of
> vertical transmission generates kinship structure in a population and
> potentiates the process of kin selection.
You're making the broad observation that the
phenomena of kin selection exists. Nobody
disputes this. But you also seem to be insinuating
that this observation, somehow, serves as the
rationale for the supposition that R is determined
by genes IBD. You seem to not be getting it. If you
were to demonstrate such you'd have to start dealing
more directly with the cause and effects aspects of
fitness (fitness accounting). I don't see you making
any efforts in this direction.
> >> Hamilton showed that it
> >> can suffice,
> >
> > Isn't this, basically, to restate the whole
> > premise. I think you're supposed to be
> > demonstrating how Hamilton, supposedly,
> > showed such.
>
> I am trying to illuminate a piece of his logic, not reiterating how
Hamilton
> showed anything.
No. Read what you wrote. You stated that he, "showed
that information about kinship alone could suffice as
a selective mechanism promoting an increase in
frequency of an allele." Is it not true, in reality,
that Hamilton never did actually show any such thing?
Is it not true, in actuality, that all Hamilton did
was present a theory to that effect. It's you
neoDarwinists that *believe* he showed such. It's
you neoDarwinists that *believe* you can see the
king's new clothes.
> You can read what Hamilton "showed" for yourself, which I
> hope you have already done.
Like I said, I'm unaware that he showed anything at all.
The onus is on you if you want to assert such.
> You obviously were not persuaded by Hamilton's
> presentation, so I thought that explaining his logic from a different
angle
> might help.
It would help if you would simply answer the question
or being unable to answer the question (which is
really the case) be honest enough to admit that you
can't answer the question and avoid the pretense.
(Surely, even if you are unable to notice the pretense
in your own thinking [and I suspect you honestly
don't] I don't see how you could not recognize it in
what Joe's been submitting lately.)
> >> and he went on to derive his famous (infamous?) Rule indicating
> >> the precise conditions
> >
> > Precise? Up to this point in your post you've
> > been making one vague excuse after another for
> > the lack of precision in Hamilton's rule, now
> > you are declaring it precise?
>
> Precision is different from accuracy. Yes, I am declaring that
Hamilton's
> Rule indicates a PRECISE point below which his model predicts
selective
> frequency decrease and above which it predicts selective frequency
increase.
Introducing the word precision doesn't change anything.
Supporting one skyhook with a new skyhook does not
solve the problem.
> >> under which selection tends to increase the allele's
> >> frequency in his simplified and optimistic (regarding the extent
of
> >> phenotypic correlation) model.
> >
> > What precise conditions?
>
> b/c > r
>
> I am not claiming that it is clear how we could measure the values of
these
> parameters. I am merely stating that this is a precise prediction
stemming
> from Hamilton's model.
Let me get this straight. You admit we cannot measure
these values but your sure it's precise. (I guess
we're supposed to take your word for it.) And you're
saying that because it's precise that therefore . . .
you lost me.
> Ambiguity in how we should measure these parameters
> decreases the effective precision of the Rule in empirical
application
> (e.g., model testing).
>
> >> The explanatory power of the model depends on the extent of
influence over
> >> phenotypic altruism by other factors (e.g., environmental effects,
other gene
> >> effects, [gene X environment] interactions, etc.), and the actual
strength of
> >> the correlation between IBD values (based specifically on the
temporal depth
> >> of such calculations used) and the probability that the allele is
present in
> >> the other individual. IMHO, it is very surprising how often the
kin selection
> >> model seems to be consistent with observations given all of this
dependence
> >> on unconsidered factors.
> >
> > It doesn't seem consistent to me.
>
> I was not referring to "statistical consistency." I was merely
indicating
> that a large body of literature reveals that positive social
interaction is
> more common among closer relatives, especially immediate family
members,
> than among more distant relatives.
Yes. Nobody is disputing these observations.
Why do you keep referring to this?
> As I have posted many times on sbe, I
> think that there are other very plausible explanations for this
observation
> (e.g., reciprocation, or payoffs of social reputation), but it is
> nevertheless consistent with Hamilton's model.
Idiotic. You can't make the slightest bit of sense
of Hamilton's model and you're declaring the evidence
consistent with it. Hamilton's rule is nonsense.
The fact that there's evidence consistent with
it does not change the fact that it is nonsense.
The Ptolemaic notion of celestial motion is nonsense.
The fact that it can be employed to predict some
celestial events (and therefore can be said to have
heuristic value) doesn't change the fact that it is
nonsense. The same is true for Hamilton's rule. The
fact that it seems to be consistent with some
observations does not give us an excuse to ignore the
nonsensicality of its underlying assumptions.
>
> >> I should comment on one other important aspect of the logic,
> >
> > What logic?
> >
> >> which strikes me as tenuous. That is the leap from considering
the
> >> relationship between a particular pair of individuals and the
economics of an
> >> interaction between them to the net effect of the altruism allele
on total
> >> organism fitness, which is a integrated result of a lifetime's
activity.
> >> Hamilton's Rule relies on the assumption that the extent of
altruism (and
> >> spite) is perfectly graded in every interaction with every social
partner
> >> throughout the organism's lifetime. To the degree that altruism
is allotted
> >> imperfectly with regard to IBD probabilities (specifically
measured at the
> >> temporal depth used by the investigator), it would become
increasingly less
> >> likely that selection would favor the allele under Hamilton's
model.
> >
> > Not sure what you're getting at here.
>
> See my response to Perplexed in this thread.
>
> >> I hope this helps illuminate the "logic/rationality of the
neoDarwinistic
> >> assumption that R is determined by genes IBD."
Jim
- Next message: Robert Karl Stonjek: "Paper: Ancient DNA"
- Previous message: Jim McGinn: "Joe's Unconditional Frequency Obfuscation"
- In reply to: Jim McGinn: "Perpetually Perplexed"
- Next in thread: Perplexed in Peoria: "Re: Perpetually Perplexed"
- Reply: Perplexed in Peoria: "Re: Perpetually Perplexed"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
