Haploid vs. Diploid DNA
Michelle Welcks
Does anyone know for sure whether Norn DNA is haploid or diploid?
It seems ashame if it is really haploid, since the programers at
cyberlife have said again and again in their documentation that Norns
can be bread for recessive genes.
Thanks,
Michelle
(((da...@mindspring.com)))
slink
It is haploid, and according to Toby Simpson the statements in the US
version manual referring to recessive genes was put there by marketing
and he never saw it before publication. The online help more
accurately describes norn genetics. The documentation that comes with
the Preview Kit is quite specific about this issue.
Sandra
Toby Simpson
> It is haploid, and according to Toby Simpson the statements in the US
> version manual referring to recessive genes was put there by marketing
> and he never saw it before publication.
Erm... I'll just clarify this. The manual was written by people within the
UK, and is an extremely good document. It is just one of "those things",
and is not something that we at CyberLife have ever claimed directly--it
was just one of those times when the Creative people writing the text
misinterpreted the technology documents and probably were not qualified
genetic experts (I think that is a safe assumption ;-)). Either way, it is
certainly not the end of the world:
Haploid genetics doesn't cause the huge range of problems that everyone
seems to think! It really is quite common in Nature, and diploid genetics
are considerably more complex to do. Here is a little paragraph on the
subject in handy "cut-out-n-keep" format which may be helpful:
"Haploid genetics might not be as much "fun" as diploid, but it doesn't
imply any significant restrictions on evolution. Diploid genetics is rare
or non-existent in computer models (at least, in practical systems, as
opposed to theoretical models of diploidy!). This is because diploidy
implies a very complex system of suppression, enhancement and interaction
that really requires a protein-level model of physiology (complete with
morphogenesis). Such a system would be a nightmare to implement a
hand-coded genotype in!"
... and hand coded genotype is exactly what we have with Creatures.
Best Regards,
Toby
--
Toby Simpson. Executive Producer/Manager--Creatures Products
CyberLife Technology Ltd., Cambridge, England.
http://www.cyberlife.co.uk
Alexander Laemmle
Hi!
>Norns are definitely haploids.
Yes...
>So the only surprises you can get are
>caused by mutations. This is a shame since diploid genotype would have
>been very easy to program and the gene pool in a game would have been
>twice as big.
How many diploid genotype realtime AL systems did you program in your
sparetime? Must have been very much that you think it would be easy to
do...
No, I think it is impossible today to write a complex diploid genetic
system that works in realtime. Such systems are far too complex to run
on a PC.
MfG
Alex
Visit the home of the Creatures Object Editor :
http://members.aon.at/alexanderlaemmle/index.html
http://www.geocities.com/siliconvalley/vista/6474/creatures.html
If you are looking for Creatures(c) addons, COBs, Tips, ...!
Thomas Andre Hafsaas
Thomas Andre Hafsaas wrote:
>
> Nothings impossible. You just have to program it, and find a good enough
> computer to run it on. :) BTW, if someone wants to make a computer model
> AI that works as a human brain one have to know exactly what a human
> brain does and why. We (humans that is) know a lot about the human brain
Just think my brain took a vacation. because although the brain and the
genes are linked it wasn't what the post was really about. Though I do
mean that nothings impossible not even coding a diploid gene system. But
that may take some time....
--
Thomas Hafsaas. thaf...@online.no. http://www.sn.no/~thafsaas
http://www.sn.no/~thafsaas/creatures
"I swore, * "When in doubt;
and the dying ember of * jump out of the nearest window.
the night leaving me cold * At least it keeps the surprise
beneath the stars, * on their side!"
I swore, *
to go with you." * Also known as Zatar
Michelle Welcks
Thank you very much Toby for the reply. --It is a testament to the
quality of people at Cyberlife that you took the time, and a big help
too! :)
Thanks again,
>Toby Simpson" <to...@cyberlife.co.uk> wrote:
Michelle
(((da...@mindspring.com)))
slink
>Thomas Andre Hafsaas wrote:
>>
>> Nothings impossible. You just have to program it, and find a good enough
>> computer to run it on. :) BTW, if someone wants to make a computer model
>> AI that works as a human brain one have to know exactly what a human
>> brain does and why. We (humans that is) know a lot about the human brain
>
>Just think my brain took a vacation. because although the brain and the
>genes are linked it wasn't what the post was really about. Though I do
>mean that nothings impossible not even coding a diploid gene system. But
>that may take some time....
>
I think what Toby's meant (clobber me if I'm wrong, Toby) is that it
would be too complex to model on the desktop computers to which the
average home user has access. Unless you want to run it on your Cray
laptop. <grin>
On the Cyberlife side of the Cyberlife website is a paper by Grand,
Cliff and Malhotra. There are also two other easier-to-understand
papers at the same location, but the one I cited above is the
technical description of how norns work inside. I strongly recommend
reading it if you want to gene-tailor your norns, or even if you just
want an appreciation of what went into creating them and why certain
limitations have been built into them.
Sandra
Thomas Andre Hafsaas
Toby Simpson wrote:
> engine as far as we can, learn as much as we can, and then develop better
> brain models that are considerably more "intelligent" than the ones we have
> today. The ultimate goal must be true human level sentience.
Hmm, I could just imagine the church having a bad day about that one :)
(Please everybody don't flame over that comment from me..)
Though do you at Cyberlife think that it's possible to make true human
level sentience?
> As I've waffled on more than long enough, I'll stop here, but as a
> footnote: Bear in mind a snail has only 400 neurons, and it uses them
> considerably more efficiently than a Norn uses its neurons. Why? How?
Ohh, me me I know that one... it's because well, erh just glad I am not
on jeapardy :) Erh, maybe better coding. (on a side note I *GOT* a snail
in a jar on my desk, the jar also got a plant emerged in water which
looks really cool, and the snail just followed the plant..) Now, back to
seriousness....
A snail has had advantage of evolution, and they have been here a lot
longer than creatures have. Norns are controlled by digital code, the
snail is controlled by natures code. (Hope I don't step on any toes by
writing this, that is not my intend) Humans have a hard time competing
with nature. Or it could be that the code isn't streamlined enough yet.
(oh great going meself that would only get my mailbox full of mail from
angry programmers :) okey, they are far superior programmers than me.) I
try to write something more indepth tomorow as it's way to late now.
BTW, did anyone know a snail could walk under water, erh, I mean it
floats upside down just noticed, no it's not dead.
> ... and for your next trick, invent a time machine, anti-gravity device and
> a teleporter :-)
Hmm, didn't I read something about that in my "physics 101" book? though
already got the time machine though ;)
Does anyone here have read "The Hitchikers Guide To The Galaxy"
trilogy?? just curious about the subject line up there.
Although the answer to life, the universe and everything isn't quite
what everyone expected.
Keith F. Goodnight
In article <01bca0fb$9592e9c0$d738...@camlt015.cyberlife.co.uk>, "Toby
Simpson" <to...@cyberlife.co.uk> wrote:
> "Haploid genetics might not be as much "fun" as diploid, but it doesn't
> imply any significant restrictions on evolution. Diploid genetics is rare
> or non-existent in computer models (at least, in practical systems, as
> opposed to theoretical models of diploidy!). This is because diploidy
> implies a very complex system of suppression, enhancement and interaction
> that really requires a protein-level model of physiology (complete with
> morphogenesis). Such a system would be a nightmare to implement a
> hand-coded genotype in!"
This is a puzzling paragraph. My professional field is evolutionary
biology & my research centers around computer simulations of social
behavior. Just about every model I've ever used in my research (mine, or
someone else's) has been diploid-- it really introduces no modelling
problems at all, there are several ways to handle it:
1. Assume complete dominance-- of the two alleles an individual has at
each locus, only one can actually be active.
The offspring "chooses" the active allele at each locus at
fertilization, either randomly (so that no allele is *inherently* dominant
over any other), or you can model Mendelian-style dominance by giving
alleles an associated "dominance value" (they will be dominant over alleles
with lower values of this number).
2. If alleles represent numerical values, you produce the "phenotype" by
averaging the two alleles: (N1 + N2)/2 = P -- then use P in place of the
value you're currently taking from the haploid genome.
If alleles represent qualitatively different states, you can't do this,
but you can still revert to method 1.
Both of these options are perfectly equivalent to haploidy when it comes to
the complexity of the indvidual Norn simulation itself.
Complexities such as the quoted paragraph worries about only arise if the
modeller wants them to.
If you choose to do "a protein-level model of physiology", you can. But
there's no special feature of diploidy that requires it-- you could do it
as easily in a haploid system, and the daunting complexities that make it
impractical would arise just as severely in a haploid system as in a
diploid. And you could also do a diploid system without it just as easily
as you could a haploid system.
The "very complex system of suppression, enhancement and interaction"
(which I agree would be a hopeless nightmare to try to fully model) arises
from the choice to do a gene product (protein) physiology model rather than
having loci translate "directly" to a phenotype. It doesn't have anything
particular to do with haploidy vs. diploidy.
Having said all that, however, it's true that there's nothing particularly
wrong with haploidy. Haploid mating can produce true recombination just as
easily as diploid mating can: in fact, it's exactly the same as option (1)
above except that, instead of choosing an allele to make active, the
offspring at fertilization chooses and allele to *keep*-- and the other is
discarded instead of silently hanging around.
In nature, there are haploid sexual species with fully recombinant
mating; there are also diploid asexual species that never recombine their
genomes at all: they give rise to clones of themselves.
So the interesting feature of mating-- recombination-- doesn't require
diploidy.
Diploidy in nature is more likely an adaptation to "defend" against
recessive lethal genes (alleles that fail to produce any product at all,
when the product they should have made was vital to survival) by keeping a
backup copy around, or perhaps an adaptation to increase the speed with
with gene product can be synthesized, by having two copies of the "source
material" for transcription enzymes to work on. This can enhance growth
rate, but it hardly applies to an artificial simulation, in which explicit
transcription-rate constraints probably aren't present to begin with.
The issue of recessive genes is about the only way in which diploidy
could have an effect noticeable in Creatures. You could have silent
mutations hanging around until suddenly you see them when you get the right
cross-- but if mutation rates are fairly high then the user's not likely to
be able to tell the difference between that and a de novo mutation anyway.
I have the distinc impression that Norn mutation rates are high enough that
even with a diploid model, you'd see new traits more often by mutation than
by the exposure of rare recessives anyway.
It is recessive genes that are the reason my professional research uses
diploid models: in Game-Theory analyses of social behavior, where selection
on a trait or strategy is often frequency-dependent, the dominance of an
allele (and whether that dominance uses a penetrance model or not, i.e.
whether it uses option 1 or option 2 from above) can affect the
evolutionary fate of the system.
But this isn't likely to be of any concern to Norn breeders-- even if
someone wants to do a rigorous evolutionary study of Norn social trait
evolution, a haploid model is just as valid if you're studying haploid
organisms.
Whew! I do seem to have droned on a bit! (I guess you can all tell I'm in
Academia...)
--Keith F. Goodnight
kei...@lucybraun.rice.edu
Benjamin Hsu
Let me get this straight,
Toby Simpson <to...@cyberlife.co.uk> 次寫入到主題
<01bca0fb$9592e9c0$d738...@camlt015.cyberlife.co.uk>...
> "Haploid genetics might not be as much "fun" as diploid, but it doesn't
> imply any significant restrictions on evolution. Diploid genetics is rare
> or non-existent in computer models (at least, in practical systems, as
> opposed to theoretical models of diploidy!). This is because diploidy
> implies a very complex system of suppression, enhancement and interaction
> that really requires a protein-level model of physiology (complete with
> morphogenesis). Such a system would be a nightmare to implement a
> hand-coded genotype in!"
So what your are saying is, that because of the suppression/expression
system of diploid genetics is wwwwwwwwwwwaaaaaaaaaaaaaaaaaaayyyyyyyy to
complicated, you have to code Creatures in haploid genetics. But you did
mention that in theoretical model of diploidy, it is possible to create a
computer model that mimics the diploid genetics. Couldn't you have
programed Creatures using the theoretical model or would that create
problems in the "evolution" aspect of the game?
Regards,
Ben
Benjamin Hsu
(I am sorry for any misinterpretation in the study of molecular genetics,
since I still need to retake that class =P)
From Toby Simpson's comment, I think the programmers were looking at the
complex model of expression/suppression found in real diploid genetics
involving multiple loci. And programming that kinda of system is down
right impossible, unless some sort of simplified model can be used to
represent the protein interactions.
But this also raise the question is there any sex-linked gene?
Pierre Belanger <bela...@quebectel.co> 次寫入到主題
<5s67a9$qqk$1...@news.quebectel.com>...
> da...@mindspring.com (Michelle Welcks) wrote:
>
> >Does anyone know for sure whether Norn DNA is haploid or diploid?
>
> >It seems ashame if it is really haploid, since the programers at
> >cyberlife have said again and again in their documentation that Norns
> >can be bread for recessive genes.
>
> Norns are definitely haploids. So the only surprises you can get are
> caused by mutations. This is a shame since diploid genotype would have
> been very easy to program and the gene pool in a game would have been
> twice as big. Plus plenty of nice surprises at each birth. This makes
> the artificial intelligence part of the game much more interesting
> than the phenotype variations.
>
> Pierre
>
>
>
Thomas Andre Hafsaas
Alexander Laemmle wrote:
> No, I think it is impossible today to write a complex diploid genetic
> system that works in realtime. Such systems are far too complex to run
> on a PC.
Nothings impossible. You just have to program it, and find a good enough
computer to run it on. :) BTW, if someone wants to make a computer model
AI that works as a human brain one have to know exactly what a human
brain does and why. We (humans that is) know a lot about the human brain
but not enough by far. (Glad I am not a human <evil grin>) (nah, just
kidding. ;))
Pierre Belanger
Toby Simpson
Michelle Welcks <da...@mindspring.com> wrote in article
<33e66e48...@news.mindspring.com>...
>
> Thank you very much Toby for the reply. --It is a testament to the
> quality of people at Cyberlife that you took the time, and a big help
> too! :)
It is my pleasure. By the way -- there will be a Creatures book over the
coming few weeks which contains quite a large amount of information on this
subject, as well as a guide to using the Genetics Kit which some users may
find quite handy. It also has a gene reference, chemical reference and a
few other things too.
Toby
Toby Simpson
Keith F. Goodnight <kei...@lucybraun.rice.edu> wrote in article
<keithg-ya02408000...@news.rice.edu>...
> In article <01bca0fb$9592e9c0$d738...@camlt015.cyberlife.co.uk>, "Toby
> Simpson" <to...@cyberlife.co.uk> wrote:
> This is a puzzling paragraph.
Thanks, I tried :-)
> 1. Assume complete dominance
> 2. If alleles represent numerical values, you produce the "phenotype"
by
> averaging the two alleles
> Complexities such as the quoted paragraph worries about only arise if the
> modeller wants them to.
Obviously, as you know, this is a highly complex issue. We were trying to
achieve a number of things simulaneously, the largest of which was coming
up with a "complete genetically specificed organism", of which "complete"
was the operative word.
To be honest, the entire experience has been enlightening (for both us, and
our customers). When you build a complex life-form inside a computer for
the first time, where do you draw the line? At a cellular level? Protein
level? Amino acid level? Which bits do you lose, which bits do you keep?
Losing what affects what, and how? Emergence helps in that we no longer
have to understand the behaviour of the whole system in order to get it,
but compromising the component parts that lead to the emergent behaviour is
not easy. When you have the advantage of being able to hand-code individual
genes from generation to generation, observe and act on changes, etc.,
diploid genetics sort of vanishes from the top of the pile when the
complexities involved in doing it properly are considered.
I'll pass all of your comments across to our scientists (they know more
than I do :-)). I'm finding this conversation most interesting--it is
precisely this sort of interest and user knowledge that justifys the
release of the Genetics Kit and enhances the product for everyone.
Incidentally, we're considering starting a new area on our web site
dedicated to the goal of generating the best Norn that is feasibily
possible with the genetic structure we have. If anyone out there is
interested in participating in this process, I'd be delighted to hear from
you. I can't promise a reply to each e-mail, but I promise to log your
interest when deciding how to handle our "CyberLab" in the future (If
you're going to respond, please use the e-mail address at the base of this
message rather than hitting reply). I'll keep everyone posted in this
newsgroup.
As a footnote--is anyone interested in hearing an official CyberLife
monthly summary as to what is going on in the Creatures community
world-wide? This would not affect or detract from all the excellent 3rd
party stuff that is already out there, more of just a bullet point list of
what is going on--a page or so. Strictly speaking, this newsgroup is not
"ours", it was created from the interest of Creatures users in the
"comp.ai.alife" newsgroup, so we don't have any god-like power to influence
or dictate the discussions here--which is the way it should be. If any
Creatures users are interested in this, e-mail me at the address in my
.signature file and if we get enough responses, we'll start at the end of
this month (I can't promise a reply to each e-mail, but they'll all be
counted).
> Whew! I do seem to have droned on a bit! (I guess you can all tell I'm in
> Academia...)
Yup, I can tell, but I can drone at least as well as you :-)
> --Keith F. Goodnight
> kei...@lucybraun.rice.edu
Toby
--
All these opinions are my own, etc.
Toby Simpson - Executive Producer/Manager - Creatures Products
CyberLife Technology Ltd., Cambridge, England.
to...@cyberlife.co.uk http://www.cyberlife.co.uk
Lummox JR
slink wrote:
>
> On 5 Aug 1997 19:34:40 GMT, "Toby Simpson"
> <to...@lobster.compulink.co.uk> wrote:
>
> <snip more good stuff>
>
> >
> >Anyway, so, we're working on a Norn to submit as a candidate for the next
> >American presidential elections. We're not quite there yet (apparently :-))
> >but you never know: blink, and CyberLife could rule the world ;-)
> >
> >Toby
> >--
> >These opinions are my own, for what that is worth :-)
> >Toby Simpson - Executive Producer/Manager - Creatures Products
> >CyberLife Technology Ltd., Cambridge, England.
> >to...@cyberlife.co.uk http://www.cyberlife.co.uk
> >
>
> could already be one of a number of past American presidential
> candidates. <grin>
>
> Sandra
Well, I don't know about that. After all, I don't think it's possible to
make a Norn (or a grendel, for that matter!) devolved enough to become a
politician.
Lummox JR
-------------------------
Visit the Norn Underground at
http://www.dreamscape.com/lummoxjr/creatures
Sir Ghoul
On Tue, 05 Aug 1997 21:09:05 GMT, sl...@netins.net (slink) wrote:
>On 5 Aug 1997 19:34:40 GMT, "Toby Simpson"
><to...@lobster.compulink.co.uk> wrote:
>
><snip more good stuff>
>
>>
>>Anyway, so, we're working on a Norn to submit as a candidate for the next
>>American presidential elections. We're not quite there yet (apparently :-))
>>but you never know: blink, and CyberLife could rule the world ;-)
>>
>>Toby
>>--
>>These opinions are my own, for what that is worth :-)
>>Toby Simpson - Executive Producer/Manager - Creatures Products
>>CyberLife Technology Ltd., Cambridge, England.
>>to...@cyberlife.co.uk http://www.cyberlife.co.uk
>>
One problem: By National law, he'd have to be 'born' on an American computer. No
Non-native citizen can become President.
Benjamin Hsu
Personally, I think the Genetic Kit is just a way to cross-breed norns
without the hassle of raising them to adulthood. And breeding is the main
reason I got this game in the first place because I am sick of breeding
fruit flys.
But I have stopped playing Creatures for a while, mainly because it's too
time consuming (and the little incompatiblity with Chinese Win95, ignor me,
I am just bitching). Although Norns are cleaner than fruit flies, Norns
require 100 times more attention than fruit flies if I want to keep them
happy and healthy (you just need rotten banana and cool area for the
flies).
With the up-coming Genetic Kit, I can artificially produce eggs with the
.GEN files. Instead of spending 1 hour and 20 minute to raise each Norn.
I can just produce generations after generations of Norns. Then select the
few offsprings with the trait I am looking for and back-up the rest. To
some Norn breeders, I appeared to have miss the whole point of the game.
But if I want to waste my time taking care a bunch of electrons, I will
just go get a tamagotchi (which I think is Bandai's source as to Pippin is
its sink, hehe).
I have nothing against haploid (it's easier to trace genes of haploid
organisms), but I agree with Keith F. Goodnight. Diploid model is not
impossible to program, it's just depend on how detailed/accurate do you
want the model to be. Without saying, the more detailed it is, the more
time you have to spend to develope/program it. Haploid organism can still
be very interesting if you have phenotype that require multiple loci
expression as in the case of Creatures (I think).
BTW, I wonder how popular is Creature in Japan, home of the Tamagotchi =P
--
Benjamin Hsu
dgu...@aol.com
mac...@ucla.edu
Benjamin Hsu
Wooohhoooo! I think I smell smoke in the air, someone better call the fire
department. Tell them a flame war is in the works =P
Seriously, a simple diploid model, as Keith F. Goodnight said, is not more
difficult to program than a haploid. But if you want a realistic diploid
model that knows which gene to express when you have multiple loci
specifying a phenotype, (plus linked-gene cross-over, and co-dominace! with
mutation!!!) You will need a nice protein level model just to sort out
what eye color your norn will have. But if your model only uses complete
dominance in expression and random cross-over, diploid just make a
genetist's life more miserable (breeding the offspring with the right
resessive trait, <shiver>)
It's just too bad that Norn's don't usually have any serious/fatal genetic
defects that makes you go "darn, if he had diploidic genetics, he might
have lived to see what a knock-out mate I have chosen for him!"
--
Benjamin Hsu
dgu...@aol.com
mac...@ucla.edu
Pierre Belanger <belanger@_globetrotter.qc.ca> 次寫入到主題
<5s8cur$o8a$2...@news.quebectel.com>...
> alexande...@aon.at (Alexander Laemmle) wrote:
>
> >>This is a shame since diploid genotype would have
> >>been very easy to program and the gene pool in a game would have been
> >>twice as big.
>
> >No, I think it is impossible today to write a complex diploid genetic
> >system that works in realtime. Such systems are far too complex to run
> >on a PC.
>
> Read carefully before answering. Who's talking about a _complex_
> diploid genetic system?
>
> >How many diploid genotype realtime AL systems did you program in your
> >sparetime?
>
> Your point is?
>
> Pierre
>
>
>
Anthony Pounds-Cornish
On Wed, 06 Aug 1997 00:32:36 GMT, sl...@netins.net (slink) wrote:
>I think it's because snails use a lot more "instinct", whatever that
>is in a living system as opposed to a norn. Maybe they don't have to
>have so many neurons because they don't use up so many combinations of
>neurons and dendrites finding out which thing to do every time. Maybe
>they are born with the dendrites in the right place and fully
>reinforced.
I agree with you, millenniums of evolution have brought the
number of neurones and their use in a snail down to what essentially
allows it to survive to reproduce for the next generation. Which moves
on the question of Norn goals? Are they same as most species, that is
to survive and reproduce to continue existence into the next
generation.
>I don't know a lot about snail chemistry except that they use
>hemocyanin (copper-based blood) instead of hemoglobin (iron-based
>blood), which says to me that they've been evolving for a long, long
>time. I doubt if they have a lot of flexibility left in their code.
This is a point which will become important in Norn genetics.
The way Norns turn out as this experiment continues will reflect on
society as a whole (or at least the part of it that uses Creatures).
Which brings an interesting question: If people have doubts about
bringing children into this world, does anybody have moral pangs about
bringing Norn's into the world?
---
<*> Anthony
Keith F. Goodnight
In article <33e86b14...@news.demon.co.uk>,
ant...@poundsms.demon.co.uk (Anthony Pounds-Cornish) wrote:
> I agree with you, millenniums of evolution have brought the
> number of neurones and their use in a snail down to what essentially
> allows it to survive to reproduce for the next generation. Which moves
> on the question of Norn goals? Are they same as most species, that is
> to survive and reproduce to continue existence into the next
> generation.
One thing to remember is that (apart from wolfling experiments) Norns
are generally under artificial selection rather than natural selection. The
users of the program have considerable control over which Norns breed and
which don't (particularly in the larger arena of which are considered worth
exporting to others, and which ones others consider worth importing).
From a gene's point of view, artificial selection is really exactly the
same as natural selection-- a gene is favored if it produces traits that
enhance reproduction *in the current environment* (the asterisked phrase is
often forgotten in non-professional discussions of evolution, leading to
frequent confusion) and the gene, being just a length of a molecule, has
neither awareness nor concern about whether the environment's constraints
result from nature or from a breeder's active intervention.
But from a wider perspective, artificial and natural selection have
greatly different outcomes. Natural selection is relentlessly, ruthlessly
practical. Nothing uselss can be tolerated; an organism will devote no
resources (not even the resource of time) to something that doesn't produce
a benefit. And by "benefit" I mean the very restrictive kind that is all
natural selection can "see": increased transmission of genes to the next
generation.
Artificial selection, by contrast, though the genes are being as
single-"mindedly" practical as always, can produce traits that have no
function to please the intelligent agent which controls the breeding:
whether those improve the organisms health & survival, or just make it more
useful to the breeder, or just suit the breeder's aesthetic whims.
Norns are not just under artificial selection presently; they are also
the products of a completely (or mostly) non-evolutionary design, which has
equipped them with some traits that probably serve no function except to
make them more interesting for humans to play with: the motivation to
speak, for instance, or the whole matter of having toys in Albia that the
Norns want to play with.
A snail doesn't play because playing requires neurons, metabolic energy
and time to be devoted both to creating the motivational state that
triggers it and the "wiring" that performs it-- and a snail has nothing to
gain from devoting those resources. Or at least, nothing important enough
to repay the cost.
Organisms with more complex brains have found play to be profitable--
with a more complex brain to start with, the "marginal cost" of building a
"boredom-avoider" is less, and the very same selective forces that are
likely to favor a large brain also make the payoff for playing better: a
complex and variable environment, unpredictable in its specifics, so that
enhanced learning and memory are favored, and that also favors gaining
physical skills by practice-- play-- instead of hard-wiring them
beforehand. "Play" in natural organisms is not profitless recreation, it's
the serious business of gaining survival skills.
I don't know if Norns learn anything from their play that makes their
survival more likely. It could be they do. The simulation could certainly
be so designed as to force it to be so. For example, "boredom" could be
designed so that it actually damages a Norn's health, and the amount of
boredom could be kept outside of the Norn genotype's ability to modify.
Natural selection would then favor Norns who were good at avoiding boredom.
(On the other hand, if a Norn's tendency to be bored was under genetic
influence, and if playing with toys had no other benefit than to reduce
boredom, then selection would favor even more strongly Norns that just
didn't get bored in the first place, and so didn't need to waste time on
toys to maintain their health.)
I suspect, however, that Norns have boredom (and the toys that cure it)
just because this makes them more interesting to humans-- without it (and
other comparable motivations that they have) they'd just sit motionless
until they were hungry or felt the urge to mate. This is what most natural
organisms at a Norn's level of complexity do (the other common approach is
to cruise steadily, gathering up food as you go by). And artificial
selection-- because humans will like active Norns better than boring,
practical Norns-- will maintain it.
I'm pretty happy with that-- I like more active Norns too.
[Slight topic change]
The whole question of long-term Norn evolution is interesting to me. Norns
have a population structure unlike anything found in nature: extremely tiny
population sizes (on each player's computer), linked by occasional
migration (exporting/importing). I'm sure that Norn genetics in this
situation would be of great interest to Conservation Biology, where the
issue of the consequences of dividing up species into small, isolated
populations is an important subject of research.
There is considerable debate currently over whether it's better to pool
surviving animals into a small number of large populations, or divide them
into many small ones. Issues of inbreeding, disease transmission, and
genetic diversity all come into play. Small populations are subject to
genetic drift that is stronger than selection, and will fix different
genes-- so your global genetic diversity might be greater under the small,
isolated population approach-- but also these differences might cause
mating incompatibilities or failures of disease resistance, so perhaps it's
better to keep a single large population that will maintain genetic
diversity within itself.
This, by the way, would be another argument for using diploid genetic in
Norns-- even without any epistatic effects or other complexities in the
model, study of inbreeding levels and drift would be more applicable to
conservation questions if Norns were diploid. But I digress.
Is there a prospect for macro-evolutionary change in Norns? Not really, I
think. Even with a ten-hour lifespan (and a generation time of around 1.5
to 2 hours), it'll still take longer than our lifetimes for us to see major
changes. The large total population, thanks to the many users, creates
plenty of opportunity for us to see new & interesting mutations, and we'll
no doubt see-- indeed, players' comments show that we have seen--
micro-evolutionary change, that is "fine-tuning" improvements in features
like obedience, feeding, etc.
But real macro-evolution-- the *accumulation* of changes, one after
another, each building on the last, until something truly different is
produced-- that takes thousands or millions of generations and you can't
run those generations "in parallel". They have to build on each other,
parents to offspring.
Even then there's no guarantee: if the "basic" Norn is pretty much okay,
then you can run the simulation for a thousand years and you won't get
anything more than fine-tuning. Albia itself doesn't change, and drive
evolution, the way the real world does. One of the most important driving
forces in macro-evolution is the fact that prey, predators and competitors
are all evolving *with* the species you're looking at, forcing it to keep
changing to keep pace (as it forces the others to do the same) and that's a
force entirely absent from Albia (at least in its current form).
> Which brings an interesting question: If people have doubts about
> bringing children into this world, does anybody have moral pangs about
> bringing Norn's into the world?
Not severe ones, no, but I do feel I have a certain degree of
responsiblity to not just (for example) export and discard any Norn I don't
like. I feel about the same degree of responsiblity to them that I would to
any living organism under my care. (The cute graphics, of course, ignite
the human tendency to anthropomorphize and make it easier for me to feel
responsible for them. If the designers had made them look like flatworms--
or as fluctuating numerical data (which is what they really are)-- then
even if the actual features of the model were identical I'd probably feel
nothing at all beyond intellectual interest.)
--Keith F. Goodnight
kei...@lucybraun.rice.edu
(Who was supposed to be working this afternoon... sheesh...)
Anthony Pounds-Cornish
On 6 Aug 1997 11:25:29 GMT, "Benjamin Hsu" <dgu...@aol.com> wrote:
>Personally, I think the Genetic Kit is just a way to cross-breed norns
>without the hassle of raising them to adulthood. And breeding is the main
>reason I got this game in the first place because I am sick of breeding
>fruit flys.
Ahh, Drosophila - memories are made of the damn buggers
breeding everywhere :)
---
<*> Anthony