Computer simulations of population genetics --
Walter ReMine
I am seeking to survey the computer simulations of population genetics
that run forward in time, and that realistically model mutation rates
(with harmful and beneficial mutations), under selection and
reproduction. That is, the simulated offspring receive mutations,
followed by selection and sexual reproduction to form the next
generation. Any simulations that meet this general specification are
requested. Especially simulations from an academic, commericial, or
industrial setting, or simulations that have resulted in published
papers, or that are widely known.
Please direct me to those by providing the name of the simulation, a
person to contact, or a reference to a paper. There is quite a variety
of these simulations, and I don't want to miss any.
I am seeking to survey this field, to identify what has been done, and
not yet been done.
-- Walter ReMine
P.S.
Of extra interest are simulations that (in addition to the above) might
model any of the following types of phenomena:
sexual reproduction, with linkage and recombination,
particular selection models, or choices thereof,
variable population size,
large numbers of mutations circulating
concurrently within the population,
a wide range of selection coefficients,
tracking of each mutation until it reaches
elimination or fixation,
chromosomes,
diploidy,
recessive mutations.
inm...@susx.ac.uk
Walter ReMine wrote:
> Computer simulations of population genetics --
>
> I am seeking to survey the computer simulations of population genetics
> that run forward in time, and that realistically model mutation rates
> (with harmful and beneficial mutations), under selection and
> reproduction. That is, the simulated offspring receive mutations,
> followed by selection and sexual reproduction to form the next
> generation. Any simulations that meet this general specification are
> requested. Especially simulations from an academic, commericial, or
> industrial setting, or simulations that have resulted in published
> papers, or that are widely known.
>
> Please direct me to those by providing the name of the simulation, a
> person to contact, or a reference to a paper. There is quite a variety
> of these simulations, and I don't want to miss any.
>
Wowww. This gives the impression that you have no idea of the size of
this field (these fields) and no idea where to start. Are you familiar
with search engines such as Google and citeseer?
These will direct you to published works primarily on the genetic
algorithm side -- ie (theory of) technical applications for
engineering/design purposes, more so than scientific population
genetics studies -- but you seem to be asking for that.
At a guess, between 1,000 and 10,000 papers per year published in this
area, so when you say you don't want to miss any ....
If you Google "site:citeseer.ist.psu.edu genetic algorithm" you get
28,300 hits
If you Google "site:citeseer.ist.psu.edu evolutionary simulation" you
get 4,600 hits
<<<<<<<
> That is, the simulated offspring receive mutations,
> followed by selection and sexual reproduction to form the next
> generation
>>>>>>>
You will find most of these papers choose to account for the
generational cycle from a different starting point: (1) selection from
parental pool (2) recombination (unless asexual) (3) mutation (4)
resultant offspring into parental pool.
> I am seeking to survey this field, to identify what has been done, and
> not yet been done.
Because this is such an enormous field, any survey only makes sense if
done for some specific purpose, with an explicit agenda. Can you make
explicit your purpose?
Inman Harvey
Perplexed in Peoria
<inm...@susx.ac.uk> wrote in message news:elpudp$2cpv$1...@darwin.ediacara.org...
I tried the second suggested Google and was charmed to find one of the
top ten hits being a paper titled "Simulation models as opaque thought
experiments". That pretty well sums it up, IMHO.
I've seen a pretty good smattering of papers based on pop-gen simulations.
The simulation code is rarely published along with the opaque thoughts
that summarize the results. Instead, there is the standard 'contact the
lead author' citation.
By contrast, sequence based phylogeny papers almost always use named,
standard, public-domain software products with more useful citations in
the Materials and Methods section. There are manuals, names of contact
persons for support, and (most impressive to an ex-software producer)
release numbers and revision histories.
Why the difference? Is it fair to say that the phylogeny people are doing
real science, whereas the pop-gen people are only experimenting with their
own opaque thoughts?
John Edser
"Perplexed in Peoria" jimme...@sbcglobal.net wrote:-
> > If you Google "site:citeseer.ist.psu.edu genetic algorithm" you get
> > 28,300 hits
> > If you Google "site:citeseer.ist.psu.edu evolutionary simulation" you
> > get 4,600 hits
> I tried the second suggested Google and was charmed to find one of the
> top ten hits being a paper titled "Simulation models as opaque thought
> experiments". That pretty well sums it up, IMHO.
>
> I've seen a pretty good smattering of papers based on pop-gen simulations.
> The simulation code is rarely published along with the opaque thoughts
> that summarize the results. Instead, there is the standard 'contact the
> lead author' citation.
>
> By contrast, sequence based phylogeny papers almost always use named,
> standard, public-domain software products with more useful citations in
> the Materials and Methods section. There are manuals, names of contact
> persons for support, and (most impressive to an ex-software producer)
> release numbers and revision histories.
>
> Why the difference? Is it fair to say that the phylogeny people are doing
> real science, whereas the pop-gen people are only experimenting with their
> own opaque thoughts?
JE:-
Jim is correct: "the pop-gen people are only experimenting with their
own opaque thoughts" because they do not supply a frame of reference.
Indeed, they just flatly refuse to discuss:
1) What is a frame of reference?
2) Why is one required in the sciences but not in mathematics?
Over the last 50 years population geneticists have consistently misused
irrefutable simplified/oversimplified models of refutable Darwinian theory,
e.g. the recent debacle of Felsenstein's zero cost argument (posted here)
including his invalid critique of ReMine's paper which correctly criticized
Felsenstein's false zero cost reasoning but was banned by Felsenstein from
publication.
Regards,
John Edser
Independent Researcher
inm...@susx.ac.uk
> <inm...@susx.ac.uk> wrote in message news:elpudp$2cpv$1...@darwin.ediacara.org...
> > If you Google "site:citeseer.ist.psu.edu genetic algorithm" you get
> > 28,300 hits
> > If you Google "site:citeseer.ist.psu.edu evolutionary simulation" you
> > get 4,600 hits
>
> I tried the second suggested Google and was charmed to find one of the
> top ten hits being a paper titled "Simulation models as opaque thought
> experiments". That pretty well sums it up, IMHO.
>
That paper, written by a close colleague of mine, is to some extent
directed at the issue you raise below.
> I've seen a pretty good smattering of papers based on pop-gen simulations.
> The simulation code is rarely published along with the opaque thoughts
> that summarize the results. Instead, there is the standard 'contact the
> lead author' citation.
>
> By contrast, sequence based phylogeny papers almost always use named,
> standard, public-domain software products with more useful citations in
> the Materials and Methods section. There are manuals, names of contact
> persons for support, and (most impressive to an ex-software producer)
> release numbers and revision histories.
>
> Why the difference? Is it fair to say that the phylogeny people are doing
> real science, whereas the pop-gen people are only experimenting with their
> own opaque thoughts?
Yes and no. There is a common classic misunderstanding rife in these
areas:-
Writer A: "I am interested in question A, my paper is intended to
address issue A using the tools I have to hand."
Reviewer B: "I am interested in question B, as surely all right-minded
people are. This paper I am reviewing uses (perhaps misuses) language
that looks partly familiar to me, surely the author must be attempting
to address issue B yet failing completely to do so. Writer A is foolish
and missing the point -- because it goes without question that B must
be the point!"
So many, perhaps most, of the papers referred to above of course fail
to fit the criterion of "real science" because they were never intended
to be doing science at all! Many of them are what I would loosely call
(A1) engineering or (A2) applied mathematics/tool making.
(A1) Evolutionary simulations for engineering: If I want to design and
optimise the wing of plane, there are many different methods, genetic
algoritms being merely one amongst many.The yardstick for success here
is whether your design is efficient, beats the competition -- and if
e.g. genetic algorithms or e.g. divination by inspecting the entrails
of a sacrificed chicken produce efficient results, either would be
acceptable. The methods need to be reported so that someone practiced
in the arts of GAs, or of entrail-inspection, understands the
particular new twist -- but of course nobody expects the use of "named,
standard, public-domain software products" because that is irrelevant
to the issue at hand. And nobody expects the standards of "real
science" to apply because this is not the aim of the project -- it was
never meant to be science.
Let me amend the "nobody expects" to "nobody *should* expect" :^)
(A2) Evolutionary simulations as applied mathematics/tool making. To
take a grand example, Newton and Leibnitz invented the tools, the
calculus of differentiation and integration, with the aid of thought
experiments in physics. But the output of their thoughts was not
(directly) science -- it was those invaluable tools of calculus that
could be used for many purposes (... including scientific purposes). At
usually a rather less grand level, many
evolutionary simulations in eg Artificial Life are just such thought
experiments with the aim of producing conceptual tools; that, if useful
tools, may be put to later use in tackling engineering or scientific
issues. And nobody expects the standards of "real science" to apply
because this is not the aim of the project -- it was never meant to be
science.
Let me amend the "nobody expects" to "nobody *should* expect" :^)
Having offered that defence against the frequent claims of "fact-free
science", I will freely admit that most of the papers hit by that
google search are flawed by any standards, let alone by the standards
of "real science". As in any area of research.
Inman Harvey
--
Inman Harvey >> Evolutionary and Adaptive Systems Group
<<
>> COGS/Informatics, Univ. of Sussex, Brighton BN1 9QH, UK
<<
inm...@susx.ac.uk >> http://www.informatics.susx.ac.uk/users/inmanh/
<<
Walter ReMine
and got 230 hits, and most of those don't seem to be papers on
simulations of population genetics (forward in time). So far, the
field isn't looking big.
> Because this is such an enormous field, any survey only
> makes sense if done for some specific purpose, with an
> explicit agenda. Can you make explicit your purpose?
I have several purposes with the survey. One goal is to identify the
'best' pop-gen simulations suited to doing research -- identify the
state-of-the-art. If there exist some 'best', then are they generally
recognized as such? Or is it an un-advertised aspect of this field?
If the 'best' pop-gen simulations are pretty decent, then perhaps I and
others can hope to use them for research. On the other hand, if the
pop-gen simulations are not suited to doing serious research, then that
fact would also be worth noting.
Another goal would be to classify (or categorize) the various
simulations -- if they happen to fall into some natural groupings of
some kind. This would help lend structure to the field, and help to
understand how the various simulations relate, one-to-another. This
would also help identify which features would be new, and which
features would not. By identifying the state-of-the-art, one can
promote the state-of-the-art.
Another goal is historical -- to identify what was done, when -- to
form a time-line of progress. Are there any insights that arise from
comparing that time-line with what was happening in other related
fields of biology?
Those are the general purposes of the survey. But so far, there seems
rather little to survey. If you know of pop-gen simulations (forward in
time), then please let me know of them.
Walter ReMine
> I tried the second suggested Google and was charmed to
> find one of the top ten hits being a paper titled "Simulation
> models as opaque thought experiments". That pretty well
> sums it up, IMHO.
>
> pop-gen simulations. The simulation code is rarely
> published along with the opaque thoughts that summarize
> the results. Instead, there is the standard 'contact the
> lead author' citation.
>
> By contrast, sequence based phylogeny papers almost
> always use named, standard, public-domain software
> products with more useful citations in the Materials and
> Methods section. There are manuals, names of contact
> persons for support, and (most impressive to an ex-software
> producer) release numbers and revision histories.
>
> Why the difference? Is it fair to say that the phylogeny
> people are doing real science, whereas the pop-gen people
> are only experimenting with their own opaque thoughts?
He made an *excellent* observation, and I cannot currently improve on
it. His statement confirms my perception of this field.
That is, the population-genetics people do not seem proud of their
simulations -- they seem muted about their results -- "The simulation
code is rarely published along with the opaque thoughts that summarize
the results. Instead, there is the standard 'contact the lead author'
citation." There doesn't seem much effort into pressing the available
computing power to produce the best pop-gen simulations. And the most
well-known pop-gen simulations seem toy-like, compared to the
biological realities of, say, sexual reproduction, linkage,
recombination, recessive mutations, a wide-range of selection
coefficients, chromosomes, and so forth. There are many legitimate
population genetics questions that could be resolved, or moved forward,
through computer simulations. Why isn't that being done?
I contacted a well-known authority in Indiana, who was not able to
recommend any pop-gen simulations. He suggested writing my own.
By contrast, the phylogeny-reconstruction papers "almost always use
named, standard, public-domain software products with more useful
citations in the Materials and Methods section. There are manuals,
names of contact persons for support, and (most impressive to an
ex-software producer) release numbers and revision histories." These
people are proud of their results, and are eager to display their
results.
Why the difference? No, it's not because one computer project is "real
science" while the other is not -- both could lead to legitimate
scientific insight. Instead, I suspect it is because the pop-gen people
simply are not proud of their simulation results. The results are not
showing what they want, so instead they give "opaque thoughts that
summarize the results", and they give little in the manner of "named,
standard, public-domain software products" that would allow *other*
researchers to do serious pop-gen simulations.
Perhaps I am wrong about that -- indeed, I hope I am wrong about that.
But that is one of the questions my survey hopes to illuminate. Do
there exist serious pop-gen simulations that do honor to what can be
done? Or instead, are the current pop-gen simulations substantially
'toy-like' compared to what could be done? Is this field legitimate?
Or an embarrassment?
The first step is to survey this field. If you are aware of any
candidates as suitable pop-gen simulations (forward in time), then
please let me know of them.
-- Walter ReMine
ErikW
Perplexed in Peoria wrote:
> <inm...@susx.ac.uk> wrote in message news:elpudp$2cpv$1...@darwin.ediacara.org...
> >
> > > Computer simulations of population genetics --
> > >
> > > I am seeking to survey the computer simulations of population genetics
> > > that run forward in time, and that realistically model mutation rates
> > > (with harmful and beneficial mutations), under selection and
> > > reproduction. That is, the simulated offspring receive mutations,
> > > followed by selection and sexual reproduction to form the next
> > > generation. Any simulations that meet this general specification are
> > > requested. Especially simulations from an academic, commericial, or
> > > industrial setting, or simulations that have resulted in published
> > > papers, or that are widely known.
> > >
> > > Please direct me to those by providing the name of the simulation, a
> > > person to contact, or a reference to a paper. There is quite a variety
> > > of these simulations, and I don't want to miss any.
> > >
> >
> > this field (these fields) and no idea where to start. Are you familiar
> > with search engines such as Google and citeseer?
> >
> > These will direct you to published works primarily on the genetic
> > algorithm side -- ie (theory of) technical applications for
> > engineering/design purposes, more so than scientific population
> > genetics studies -- but you seem to be asking for that.
> >
> > At a guess, between 1,000 and 10,000 papers per year published in this
> >
> > If you Google "site:citeseer.ist.psu.edu genetic algorithm" you get
> > 28,300 hits
> > If you Google "site:citeseer.ist.psu.edu evolutionary simulation" you
> > get 4,600 hits
>
> top ten hits being a paper titled "Simulation models as opaque thought
> experiments". That pretty well sums it up, IMHO.
>
> I've seen a pretty good smattering of papers based on pop-gen simulations.
> The simulation code is rarely published along with the opaque thoughts
> that summarize the results. Instead, there is the standard 'contact the
> lead author' citation.
>
> By contrast, sequence based phylogeny papers almost always use named,
> standard, public-domain software products with more useful citations in
> the Materials and Methods section. There are manuals, names of contact
> persons for support, and (most impressive to an ex-software producer)
> release numbers and revision histories.
>
> Why the difference? Is it fair to say that the phylogeny people are doing
> real science, whereas the pop-gen people are only experimenting with their
> own opaque thoughts?
I have some disagreements with this I think, especially after reading
the full abstract of the opaque thought model paper. It refers mainly
to Artificial Life simulations, AFAICT, which, in my opinion doesn't
belong in traditional pop gene simulations, whereas you seem to refer
to the latter. Standard pop gen simulations are done with published and
usually fairly standard models like various forms of the coalescent and
are really quite easy to understand. The insights that they bring are
also fairly straight forward to interpret. Coalesent simulations are
also used to generate distributions for hypothesis testing and
parameter estimates of real data.
So I don't think it's fair to say that refering to populatoin genetics
whereas it might be regarding AL-simulations
On the other hand there are probably loads of pop-gen simulations where
I am left more in the dark.
Turning to the presumed (?) topic of this thread, here's a link to a
simulation paper concerning the cost of natural selection (Nunney
2003):
http://www.sekj.org/PDF/anz40-free/anz40-185.pdf
The model is rather clearly described and a quick look tells me that
two of the assumptions of Haldane's model that I consider interesting
(admittedly my understanding of this topic is limited) are not
explored, namely all loci are unlinked and independent. Instead he
focussed on other factors. But all in all real science :P
Btw, maybe someone knowledgable/interested person can comment on the
paper?
ErikW
Malcolm
One of my pet beefs is that people hold code in contempt. "There is only one
thing worse than pseudocode", one PhD examiner wrote, "and that is real
code". That is true if you are briefly scanning the PhD on the train up to
the viva, but if you are trying to extend the work, mathematical
descriptions of algorithms are a menace.
>
> By contrast, sequence based phylogeny papers almost always use named,
> standard, public-domain software products with more useful citations in
> the Materials and Methods section. There are manuals, names of contact
> persons for support, and (most impressive to an ex-software producer)
> release numbers and revision histories.
>
> Why the difference? Is it fair to say that the phylogeny people are doing
> real science, whereas the pop-gen people are only experimenting with their
> own opaque thoughts?
>
programmer and not a scientist, but then that was by an atheist medical
researcher who obviously wasn't familair with the term "bioinformatics" and
was in atheist mode. {Drifting OT here].
The sort of programs needed to analyse phylogenetic trees apply essentially
the same algorithm to different sets of data. Thus a good program can be
reused. Population genetics simulations apply different algorithms to
invented or model data, so the program needs to be rerun each time.
I'll try to get back to Walter with a list of good simulations. I can't
provide one from the top of my head.
--
www.personal.leeds.ac.uk/~bgy1mm
freeware games to download.
Walter ReMine
population-genetics simulations (that run forward in time) and simulate
known biological genetics.
(1) Not simulations of coalescent theory
(that run backward in time)
(2) Not "Genetic Algorithms"
-- that do not mimic known genetics
(3) Not "Artificial Life" simulations
-- that do not mimic known genetics
I am interested in identifying ANY simulations meeting these basic
requirements -- and especially the best ones, the ones you would
recommend for doing population-genetics research. Or any coming from an
academic, industrial, or commercial background, or which have resulted
in published papers, or are widely known and documented.
If you know of any pop-gen simulations meeting these requirements,
please let me know.
-- Walter ReMine
Joe Felsenstein
There are a great many papers presenting genetic simulations. Using my
Bibliography of Theoretical Population Genetics, for example, we find that
when the string "imulat" is used (to get papers with "Simulate", "simulate",
"Simulation" or "simulation" as well as "simulating" in the title) that up to
1981 when the Bibliography ended there had been about 200 papers. The URL for
such a search is:
http://www.wsu.edu/~mmorgan/interactivities/felsenbib.cgi
PubMed is another place to search with keywords:
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?CMD=search&DB=pubmed
There have of course been many more papers since. There are two books on
how to do genetic simulations, one by Fraser and Burnell, one by Crosby.
The difficulty from your standpoint is that most simulations are done to
test interactions of particular pairs of evolutionary forces, not to construct
a realistic model of a population. This similar to what many other sciences
do -- leaving out the full grit of reality to get a clearer view of what
particular forces are doing.
The difficulty of simulating "known biological genetics" is that we don't
"know" many of the things we'd like to -- how frequent advantageous mutations
are, what their distribution of selection coefficients is, etc. So it's
rare that anyone can claim to accurately model reality.
As for the posts that claimed that population geneticist are being strangely
reticent about publishing their programs, there is a grain of truth to that.
But it's not that they're embarassed and feel there is some problem that they
have to cover up. The phylogeny programs that are widely distributed (and I
know all about these, as I maintain the only web site summarizing them)
are analyzing real data, so there is pressure to get them available to users.
Theoretical population geneticists writing simulation programs are not
intending to use these to analyze data. (Also there is a tradition in
population genetics of valuing algebraic treatments more than numerical
calculation). But mostly the programs used for simulation are of use to their
writers but probably marginal to most other users, so the effort of polishing
them for wide distribution seems less important. Most authors will, if
asked, provide their programs in some source code form.
----
Joe Felsenstein j...@removethispart.gs.washington.edu
Department of Genome Sciences and Department of Biology,
University of Washington, Box 355065, Seattle, WA 98195-5065 USA
John Edser
"ErikW" bryo...@hotmail.com wrote:-
> Standard pop gen simulations are done with published and
> usually fairly standard models like various forms of the coalescent and
> are really quite easy to understand. The insights that they bring are
> also fairly straight forward to interpret. Coalesent simulations are
> also used to generate distributions for hypothesis testing and
> parameter estimates of real data.
JE:-
Sets of pop gen models have to be derived by the processes of
simplification/oversimplification from just the ONE parent theory.
As long as pop gen models are NOT:
1) Oversimplified i.e. delete or change an empirical constant assumed by
their parent theory (see below).
2) Become misused to compete against and/or replace their parent theory.
THEN these models can be used to provide _extremely useful but however, only
heuristic insights_ into the workings of just the one, same EMPIRICALLY
based theory from which they were all simplified/oversimplified.
Oversimplified models critically reverse cause and effect within the model
compared to the parent theory, inevitably providing forever ongoing mass
confusion, e.g. Hamilton's Rule.
Definitions:-
1) Simplification: a change in or the deletion of any empirically based
variable defined by the parent theory within any model of it.
2) Oversimplification: a change in or the deletion of any empirically based
constant defined by the parent theory within any model of it.
If nobody can be bothered to identify the parent theory anymore then the
inevitable misuse of models in place of it can only produce "a comedy of
errors". An immediate example that has become an important historical
document held within sbe archives was Felsenstein's false zero cost of
substitution argument. Felsenstein et al subsequent ban of Walter ReMine's
paper which correctly identified the inherent contradiction which
Felsenstein had employed to create a false zero cost has never been
addressed either rationally or ethically. Neither has the VERY basic
question which I asked: how can ANY rational proposition for a zero cost of
substitution allow TWO and not just the ONE point of zero reproductive
excess within the SAME proposition? I have asked (and will continue to ask
for as long as it takes to get an answer) how such a thing was ever
tolerated by today's evolutionary theory establishment. It should be noted
that the deafening silence which continues to surround this particular issue
speaks volumes about the importance of it and the integrity of all the
participants (ReMine included).
http://www.sekj.org/PDF/anz40-free/anz40-185.pdf
The first few sentences (this particular PDF document does not allow me to
past and copy them) verify that Haldane's never defined a valid frame of
reference: a population constant. His use of C as a "decreasing function of
the initial frequency of the beneficial alleles" attempts to employ just a
variable (variable allele relationships) as a frame of reference which is
simply not rational. Felsenstein's attempt to do this only produced
disastrous results.
John Edser
"Malcolm" regn...@btinternet.com wrote:
> One of my pet beefs is that people hold code in contempt. "There is only
> one
> thing worse than pseudocode", one PhD examiner wrote, "and that is real
> code". That is true if you are briefly scanning the PhD on the train up to
> the viva, but if you are trying to extend the work, mathematical
> descriptions of algorithms are a menace.
JE:-
This is because "mathematical descriptions of algorithms" mostly do not
provide any valid frame of reference.
Malcolm,
You previously mentioned that you would supply code to solve Felsenstein's
zero cost riddle. I mentioned that computer code could not be written to
solve it unless a population constant was firstly defined by Felsenstein.
Have you attempted to write this code?
John Edser
"Walter ReMine" sci...@minn.net wrote:-
> > Jim Menegay wrote:-
> > Why the difference? Is it fair to say that the phylogeny
> > people are doing real science, whereas the pop-gen people
> > are only experimenting with their own opaque thoughts?
> He made an *excellent* observation, and I cannot currently improve on
> it. His statement confirms my perception of this field.
JE:-
Like any paradox, all these negatives re: pop gen do not exist (all of them
are just the absence of positives). So what needs to be done is to identify
all the missing positives. I claim this can only be done by CORRECTLY
identifying just the ONE theory that all pop gen models were
simplified/oversimplified from. Once this critical parent theory has been
identified all that has to be done is list the simplifications and
oversimplifications assumed by pop gen models and appraise their
significance relative to the theory these models were derived from.
John Edser
inm...@susx.ac.uk wrote:-
> > Why the difference? Is it fair to say that the phylogeny people are
> doing
> > real science, whereas the pop-gen people are only experimenting with
> their
> > own opaque thoughts?
> Yes and no. There is a common classic misunderstanding rife in these
> areas:-
>
> Writer A: "I am interested in question A, my paper is intended to
> address issue A using the tools I have to hand."
>
> Reviewer B: "I am interested in question B, as surely all right-minded
> people are. This paper I am reviewing uses (perhaps misuses) language
> that looks partly familiar to me, surely the author must be attempting
> to address issue B yet failing completely to do so. Writer A is foolish
> and missing the point -- because it goes without question that B must
> be the point!"
>
> So many, perhaps most, of the papers referred to above of course fail
> to fit the criterion of "real science" because they were never intended
> to be doing science at all! Many of them are what I would loosely call
> (A1) engineering or (A2) applied mathematics/tool making.
JE:-
If any of these models were allowed to usurp the theory from which they were
simplified/oversimplified does this constitute a valid or an invalid use of
them?
John Edser
j...@removethispart.gs.washington.edu (Joe Felsenstein)wrote:-
> Subject: Re: Computer simulations of population genetics --
> The difficulty from your standpoint is that most simulations are done to
> test interactions of particular pairs of evolutionary forces, not to
> construct
> a realistic model of a population. This similar to what many other
> sciences
> do -- leaving out the full grit of reality to get a clearer view of what
> particular forces are doing.
JE:-
Deleting or changing an empirically based VARIABLE defined within a
refutable theory allowing just a hypothetical via a simplified version of a
parent theory is "similar to what many other sciences do". What they do NOT
do is proceed to oversimplify any parent theory by changing or deleting an
empirically based constant within that theory. This is because this changes
the theory out of all recognition. Cause and effect become critically
reversed within an oversimplification. If you maximally oversimplify (delete
all constant terms assumed within the parent theory) all you are left with
is just a tautological model, e.g. Hamilton's rule. Only pop gen (which
remains the basis of gene centric Neo Darwinism) oversimplifies on just a
routine basis. Gene centric Neo Darwinism does not even differentiate
between a simplification and an oversimplification allowing itself to act as
if no parent theory even existed. The net result: it provides for itself an
epistemological blank cheque, i.e. it can say and do just about anything it
wishes in the name of science. This was proven when the random process of
genetic drift was allowed to become redefined to provide "evolution" and not
just temporal variation. The absurdity of such a redefinition had previously
been demonstrated by the mutationists who replaced evolution by Darwinian
natural selection with "evolution" via random mutation.
The apparent paradox which YOU identified within YOUR reasoning of the cost
of substitution problem was provided by just an oversimplification of that
problem. Your subsequent "solution" was also, just another uncorrected
oversimplification. The paradoxical cost of substitution problem as you
argued it and the irrational solution which you provided for it contained no
frame of reference, i.e. both remained oversimplified and uncorrected.
Please explain to sbe readers your rational justification for allowing TWO
and not just the ONE point of zero reproductive excess within the SAME
problem within your uncorrected oversimplified "solution"?
>snip<