What Good Is Half A Machine?

December 9, 2008 5:24 AM

One of the classic arguments against evolution by natural selection is "what good is half an X?" where X is an eye, a wing or some other complex body part or system. Directly responding to the implicit challenge some researchers have been not just figuring out how X could have evolved, but actually evolving new complex machines (previously). The basic ideas are so simple that web versions (explanation and discussion) have been popping up.
posted by DU (67 comments total)

That Mona Lisa thing is amazing.
posted by seanyboy at 5:39 AM on December 9, 2008

Also interesting is the BreveCreatures screensaver which simulates evolution (previously)
posted by chillmost at 5:52 AM on December 9, 2008 [1 favorite]

I've like watching my little cart go go go.
posted by These Premises Are Alarmed at 6:07 AM on December 9, 2008

The car seems to eventually level off, I guess because there's a maximum speed set for the wheels, which can only take you so far in the allotted time.

The discussion link had some interesting improvement ideas. In particular, I liked the idea of an evolving terrain to try to get a Red Queen Effect going that makes both sides better and better.
posted by DU at 6:26 AM on December 9, 2008

I think they need to dumb these videos and stories down for the people who believe in intelligent design magic.
posted by chunking express at 6:26 AM on December 9, 2008

Evangelicals should ditch creationism and accept that God is watching us evolve in His flash app, debugged by His divine hand.
posted by East Manitoba Regional Junior Kabaddi Champion '94 at 6:27 AM on December 9, 2008 [2 favorites]

I just re-watched that clock evolution thing. I remember it being a LOT more gripping the first time around. I'd like to see this program actually running, not a text (+ annoying music) video of the result.
posted by DU at 6:37 AM on December 9, 2008

I'd always thought that the must-be-useful-at-every-stage was a limitation that evolution had to deal with but that, thankfully, humans dont have to deal with. We can design and conceive things that only have a purpose when they are complete. That may be in fact what distingushes human artifacts from natural artifacts.

More imaginatively, I've always wondered this: Can there exist some huge complex object that does something wonderful but that cant be dissected into sub-components? Something that somehow exists "all at once"? Taken out of the anti-evolution context are there really interesting things which are irreducibly complex which we would build, I suppose, not by designing, but by haphazardly stumbling upon them?

Or, perhaps they can arise out of pure chance combined with the deepness of time. Perhaps our own Universe is a Mymosh. A collusion of "junk" that just happened to work not something that in any way evolved from anything else.
posted by vacapinta at 7:06 AM on December 9, 2008

are there really interesting things which are irreducibly complex

It is certainly true that there are methods of construction that are irreducibly complex. For instance, computer programs. If you are evolving a program from the primitive level of single letters and numbers and you get the line "fot (x=1; x<10; x++) {...}" you don't get something that is "close to" a for loop. You get an error.

But that brittleness is an artifact of computer history, not a feature of the natural world. Still, it might be a feature of some aspect of the natural world. But we're still talking about methods of construction. I don't think it's fair to the question to say you have to build a car and it has to be built in such-and-such a way.

Another unfair constraint is a binary fitness function (any stepwise fitness function with the steps too far apart is a problem, really). For instance, try to evolve an equilateral triangle where the fitness function just returns True or False depending on if the sides are identical. You can't compare two attempts and ask which is more fit because they are both just False. With no continuous "slope" to climb you are reduced to hoping you just stumble upon it.

But again, that's a toy problem specifically constructed to answer your challenge. You asked for something interesting. My guess is that for sufficiently loose problem statements and if you get points for being half-right, there are no such problems.
posted by DU at 7:29 AM on December 9, 2008

How small are your sub-components? Do atoms count? Molecules?

At first I thought that a star might be a good example..but that doesn't work.

http://en.wikipedia.org/wiki/Stellar_evolution
posted by device55 at 7:38 AM on December 9, 2008

The question isn't really what good is half a machine, but what harm is half a machine.
posted by b1tr0t at 7:45 AM on December 9, 2008 [6 favorites]

It's also an absurd question to think about "half an eye" - as though there couldn't possibly be something eyelike that's simpler, and that might have developed into full-fledged 'modern' eyes. Total failure of imagination.
posted by Tomorrowful at 7:52 AM on December 9, 2008 [6 favorites]

It's also an absurd question to think about "half an eye" - as though there couldn't possibly be something eyelike that's simpler, and that might have developed into full-fledged 'modern' eyes. Total failure of imagination.

Well, sure, but there are an awful lot of people who haven't taken first year bio, or didn't pay attention in first year bio, or are, for socio-political reasons. cynically misrepresenting the basics of anatomy to those too poorly educated or lazy to understand it.
posted by aught at 8:00 AM on December 9, 2008

a wizard did it.
posted by blue_beetle at 8:02 AM on December 9, 2008 [3 favorites]

Actually, I recall reading somewhere (probably mefi-linked) that a few researchers did a simulation of the development of single to multi-celled organisms and found that the development of light-detecting organs of increasing sensitivity was not just likely, but necessary, and happened with rather clear-cut evolutionary steps from photo-sensitive cells that allow plant-esque things to grow towards the most energy, to more sensitive and specific cell clusters that allow new organisms to find and eat all those previous sun-lovers, and then more sensitive and such for the next generation to find and eat the previous herbivorous generation, and so forth... it was practically a guarantee that eye-type organs would crop up each time they ran the simulation.
posted by FatherDagon at 8:09 AM on December 9, 2008 [4 favorites]

It's also an absurd question to think about "half an eye" - as though there couldn't possibly be something eyelike that's simpler, and that might have developed into full-fledged 'modern' eyes.

Is that really what that argument boils down to? I guess it is. I always figured I just didn't understand the point, because that seemed to simply dismissed.

Go little cart! I have a one instance with fat wheels close together, and another with smaller wheels very spread. I think the two fat approach is working better.
posted by These Premises Are Alarmed at 8:23 AM on December 9, 2008

The religious arguments against human evolution through natural selection are not based on logic, so attempting to logically refute them is fairly silly.

My problem with evolution is that is often used simply as a replacement for god. How did X become Y? "God did it." "No, evolution did it." Neither is refutable.

---------------

Let me give you an example:

The Galapagos Islands were useful for Darwin because they are so biologically isolated. There are relatively few species, so you can actually make some sense of the interconnections between them. Overall they are a pretty desolate place. Cacti are abundant -- a very particular type of cacti. The lower parts are protected by what is very clearly tree bark.

So how did this happen? How did we come to have a clear hybrid of tree and cactus in the middle of the equatorial pacific?

The standard evolutionary explanation of this is that both trees and cacti share a common ancestor. This ancestor evolved bark at some point and then split into cacti and trees, but both evolutionary lines *retained the genetic information for growing bark*.

Then, many many years later a strain of the cactus line ended up in the Galapagos. Natural selection took place, and the cactus mutants that tended to activate their "bark genes" became dominant.

Tada! Tree and cactus hybrid out there in the middle of nowhere.

------------------

Fair enough, but here's the problem: A similar argument can be used to justify any mutation at all. Mammal with a duck bill? Common ancestors. Poinsettia that smells like Eucalyptus? Common ancestor. Pig with wings? Common ancestor.

And particularly on that pig: Pigs don't have to re-evolve wings. They already have the genetic code, so the next generation of pigs can sprout fully formed stork wings if chance allows.

So where does that leave us? After these millions of generations with genetic offshoots all over the place, *absolutely anything, no matter how unlikely, can be explained by evolution*.

The corollary to this is that there is no way to falsify the theory of evolution. That is the mark of philosophy (and theology) not science. If you're going to run around saying "I have a hypothesis that explains everything but can't be disproven", does it really matter if you call it god or evolution?
posted by tkolar at 8:55 AM on December 9, 2008 [3 favorites]

One of the favorite ID examples of "irreducible complexity" used to be the mousetrap. I present to you:
A reducibly complex mousetrap.
Mousetrap evolution through natural selection.
posted by parudox at 8:58 AM on December 9, 2008 [2 favorites]

tkolar- the nice thing is that now, we can actually sequence and manipulate the genes in question to check. We can selectively introduce or knock out genes in the winged pig and find out whether the responsible genes are also found in birds.

Evolution is testable.
posted by Jpfed at 9:01 AM on December 9, 2008 [4 favorites]

Biologist Adolf Portmann argued that the external appearance of organisms is in large measure derived from their own inner logic, what he calls self-representation. He discusses in his book, Animal Forms and Patterns: A Study of the Appearance of Animals, the duck speculum, writing:

If ... we look at the speculum on a duck's wing, we might imagine that an artist had drawn his brush across some ten blank feathers, which overlap sideways - making white, bluey-green, and black lines - so that the stroke of the brush touched only the exposed part of each feather. The pattern is a single whole, superimposed on the individual feathers, so that the design on each, seen by itself, no longer appears symmetrical. We realize the astonishing nature of such a combined pattern only when we consider that it develops inside several or many feather sheaths completely separated from one another; and that in each individual feather it appears at an early stage while it is still tightly rolled up, the join pattern not being produced until these feathers are unfolded. What sort of unknown forces direct the construction work in the 'painting' of these feather germs?—quotation from here

posted by No Robots at 9:01 AM on December 9, 2008 [1 favorite]

tkolar -- I'm afraid your ignorance is showing.

Every example of convergent evolution doesn't mean that the common ancestor is the source of the genetic material for the convergence. Basic morphology and cytology will indicate whether the tough bark-like substance on the cacti is actual "bark" resurrected from the genetic ashes of a distant ancestor or specially adapted, tougher cells that in a deep morphological sense share little in common with tree bark and look an awful lot like specialized tough cactus cells. Biologists can tell the difference, here.

Just because something has an "eye" doesn't mean that the "eye" is some pre-written genetic code in all animals' DNA. It just means that having an organ that can sense light, and perhaps even focus it, is an advantage. IIRC, the "eye" has evolved over 15 different ways from basic light cells to the big goopy eyeball of mammals to the freaky compound eye of the bee to the basic light-sensitive spots on the sides of some fish.
posted by chimaera at 9:11 AM on December 9, 2008

Also, "common ancestor" has nothing to do with natural selection. What you are talking about is genetics, which is a technology that natural selection happens to employ.
posted by DU at 9:20 AM on December 9, 2008

I would love to see a program like this that figures out how to build the tallest tower of goo.
posted by buriednexttoyou at 9:37 AM on December 9, 2008 [1 favorite]

I would love to see a program like this that figures out how to...

Coupling this program with Incredibots/Fantastic Contraption would be super sweet. Basically, you the player would have a bunch of physical measurements of the environment. You are required to create a fitness function that forces the evolution of a particular type of machine. That might be a little too open-ended and vague, though.
posted by DU at 9:43 AM on December 9, 2008

Jpfed wrote...
We can selectively introduce or knock out genes in the winged pig and find out whether the responsible genes are also found in birds.

Evolution is testable.

Hmm, to me that sounds like testing genetics, not evolution.
posted by tkolar at 9:58 AM on December 9, 2008

chimaera wrote...
Every example of convergent evolution doesn't mean that the common ancestor is the source of the genetic material for the convergence.

Sorry if I appeared to be arguing that was the case. That wasn't my intent.

Basic morphology and cytology will indicate whether the tough bark-like substance on the cacti is actual "bark" resurrected from the genetic ashes of a distant ancestor or specially adapted, tougher cells that in a deep morphological sense share little in common with tree bark and look an awful lot like specialized tough cactus cells. Biologists can tell the difference, here.

Yes, and to be specific that is what has been decided by biologists about the bark on the cacti in the Galapagos.

Convergent evolution doesn't provide any problems for me. It follows pretty directly out of natural selection.
posted by tkolar at 10:07 AM on December 9, 2008

Oh, and just to be pedantic: By definition, no examples of convergent evolution imply a common ancestor is the source for the genetic material. If that's the case then it wasn't convergent evolution.
posted by tkolar at 10:10 AM on December 9, 2008

Evolution is stupidly clever. It throws out -- births -- millions of copies, to find one that's just a tiny bit better. Profligate and cruel.

Watching this is so sad. All the configurations that are obviously ill-suited, watching until these "hopeful monsters" stumble and die. And even sadder, the ones with a somewhat better configuration that get halfway up the hill, they get stuck and dies a lonely death straining to gt out of a hole they can never ecape.

An octopus will brood hundreds of eggs, scores will survive to become juveniles, and only handful to adulthood.

Fish will brood thousands, most of which will die violently.

Even among mammals, most will end up eaten or if they're very lucky, elderly and infirm and weak, starving to death.

All struggling to breed as many offspring as they can, to continue teh cruel experiemnt.

It makes me wonder how many of us are Nature's mistakes, combinations of genes doomed to failure and frustration, so that somewhere someone else's combination of genes can show the way "forward".

So much death and suffering and pain and failure to progress by such small increments.
posted by orthogonality at 10:12 AM on December 9, 2008 [2 favorites]

So much death and suffering and pain and failure to progress by such small increments.

Doesn't it make more sense to see this organically as the gradual unfolding of life in all its forms, just as each individual organism unfolds its own form? I mean, we don't pine for our lost skin cells, do we?
posted by No Robots at 10:20 AM on December 9, 2008 [3 favorites]

Thanks for the clarification, tkolar.
posted by chimaera at 10:21 AM on December 9, 2008

Damn, it took me 10 minutes to figure out what's meant by the expression "what good is half an eye." The first google result eventually made the position clear but didn't account for the expression itself.

It sounds like an incredibly stupid thing to say. I can see some logic in wondering how complex structures evolved incrementally. There are some cases where it's intuitive to think everything would have had to pop into being all at once, together, and that a process of slow modifications would have found no path to the end result that was viable along the way.

But "half an eye?" Is this expression built on the assumption that structures evolve piece by piece? First one half of the eye, then the other half? What bullshit is that? The first "eye" was nothing more than a single-celled organism with some photoreceptive ganglia to help it dart out of the way if a shadow fell on it.
posted by scarabic at 10:32 AM on December 9, 2008

It is certainly true that there are methods of construction that are irreducibly complex. For instance, computer programs.

I think there may be a difference between fragile and irreducibly complex. It's true that getting one thing wrong in a computer program yields an error and not a result almost as good as desired. But you can break a computer program down into component parts. I think computers are actually a great example of incrementalism. Someone makes hardware, along comes an OS, then a programming lanugage, then libraries, then programs, then applications, then mashups, then sites, then... you get the idea. It's all layers built on top of layers. Right now, somewhere in the belly of MetaFilter I'm briefly invoking some UNIX code from 35 years ago, you know?
posted by scarabic at 10:38 AM on December 9, 2008 [2 favorites]

DU wrote...
Also, "common ancestor" has nothing to do with natural selection. What you are talking about is genetics, which is a technology that natural selection happens to employ.

I'm not following you here. If the argument is that current conditions result in a long dormant genetic mutation re-emerging, then common ancestors matter a lot. Regardless of how the traits are passed (we happen to know it's genetic) the point is that -- at this late stage -- absolutely any trait can appear fully formed in any species.
posted by tkolar at 10:39 AM on December 9, 2008

There is a hill that my cart just can't get over. I'm just going to let it keep going and see if iit figures out how to do it.
posted by Astro Zombie at 10:43 AM on December 9, 2008 [1 favorite]

We can selectively introduce or knock out genes in the winged pig and find out whether the responsible genes are also found in birds.

Evolution is testable.Hmm, to me that sounds like testing genetics, not evolution.

Although let me add that it would be a very interesting negative result if a pig suddenly ended up with fully formed wings that were not genetically related to any other winged animals. It wouldn't disprove evolution, but it would be a poke in the eye.
posted by tkolar at 10:48 AM on December 9, 2008

I'm not following you here.

The car evolution app exhibits natural selection but there is no "long dormant genetic" *anything*. All the "genes" for the car are in use, there's nothing on the back burner.

Genetics is a mechanism through which natural selection operates. Your critique of evolution based on genetics is like a critique of physics based on a stone cathedral. "They could just put another column in here anywhere--physics is unfalsifiable!" Wha?

...the point is that -- at this late stage -- absolutely any trait can appear fully formed in any species.

Not true, even in your limited examples. You can't just "turn on" bark genes in a cactus, even if they are there, and have them work right straight from the box. And if you do use them, you can't use an altered version for something else (i.e. it's one-use only) unless you make a copy, tune it correctly, etc and so forth. That is, evolve.
posted by DU at 10:52 AM on December 9, 2008

The question isn't really what good is half a machine, but what harm is half a machine.

True, but if a particular phenotype requires 1000 separate genetic mutations to arise, then for half of those mutations to happen purely by chance with nothing selecting for them is still very unlikely, even if nothing is selecting against them. And it seems like there can't be many mutations that are 100% neutral to reproductive fitness.

On the other hand, the Earth has been around for a very long time, so it's hard to have an intuition for what is and isn't likely to happen randomly during that time. You'd have to actually do some math.
posted by straight at 10:59 AM on December 9, 2008

the line "fot (x=1; x<1>

Then your langauge is too high level. Having studied the workings of biology in a class or two, DNA works closer to assembler. Translating, you have something akin to

mov 1, ax
loop:
inc ax
cmp 0
bnz loop

[etc]


If you add instructions inside the loop, you get a different loop; if you fiddle with the loop entirely, you may get a different loop or no loop at all. If you handle mutation at the binary level, you could get a jump to a random location in memory. There's still plenty of chance for infinite loop, if you change bnz to jmp or break the comparison. But the (albeit very large) chance of infinite loops doesn't mean a binary fitness function: if 10 gets changed to 9 you have a function that's close but not perfect. And since computer scientists care about the speed of algorithms, make the fitness function include runtime at some level.

I imagine people have thought about programming primitives more suitable for GA. For example, changing branching primitives to relative addressing rather than absolute. It's important to recognize that today's organisms have a lot of mechanisms for capitalizing on mutation and dealing with failure on a large scale. Sexual reproduction, for example, needs every gamete to be a functioning cell before offspring is created. So millions of mutations are fatal, but they get discarded out of hand before getting applied to a long "fitness function test", i.e. life.
posted by pwnguin at 11:26 AM on December 9, 2008

Then your langauge is too high level.

Right. Tierra and other artificial life "soup" programs take this approach. Define a bunch of primitive operations and use those as your DNA/program.

My point is that by inappropriate choice of methodology, you can easily manufacture an "irreducibly complex" problem.
posted by DU at 11:34 AM on December 9, 2008

You know, that worked in Live Preview...
posted by pwnguin at 11:39 AM on December 9, 2008

It's also an absurd question to think about "half an eye" - as though there couldn't possibly be something eyelike that's simpler, and that might have developed into full-fledged 'modern' eyes.

My favorite answer to the "half-an-eye" question is Nietzsche's:

...from time immemorial it had been believed that in understanding the ascertainable aims and use of a thing, a form, an institution, one also understood why it had come into existence — thus the eye was understood as made for seeing, the hand as made for grasping... But all aims, all uses are merely signs... the entire history of a "thing", an organ, a custom may take the form of an extended chain of signs, of ever-new interpretations and manipulations...

Good code gets reused, and watching one end-user deploy a bit of good code generally won't tell us much about its origin.
posted by dickymilk at 11:39 AM on December 9, 2008 [2 favorites]

DU: It is certainly true that there are methods of construction that are irreducibly complex. For instance, computer programs. If you are evolving a program from the primitive level of single letters and numbers and you get the line "fot (x=1; x

The obvious objection here is that the primitives of computer programming are not ascii characters in a text file. No more than the word "cytosine" is a primitive for building an enzyme.

tkolar: Fair enough, but here's the problem: A similar argument can be used to justify any mutation at all. Mammal with a duck bill? Common ancestors. Poinsettia that smells like Eucalyptus? Common ancestor. Pig with wings? Common ancestor.

Well, gee, there are so many problems and misconceptions in this short passage, I really don't know where to begin. To start with, "mammal with a duck bill" isn't a mutation, as it's a pretty complex feature. The results of a mutation would be much more modest, such as having a pair fewer teeth, or a somewhat wider jaw.

In fact, the appearance of a true duck bill or gull's wing on a mammal would be an obvious falsification of descent with modification via natural selection as a theory to explain such features. The most recent common ancestor for the duck and the platypus had neither a bill or wings, the two branches diverged long before there was any species that can be recignized a modern mammal or a modern bird.

Although superficially similar to a duck bill, it has obvious differences, not the least of which is that platypus snout is studded around the edges with electroperception organs for locating prey, while the duck bill is a filter for separating food from water. The evolutionary explanation for convergent features in unrelated lineages of organisms isn't "common ancestor." It's that natural selection converges on some common solutions for similar problems. Hammerhead sharks also have a broad head to improve electroperception. Winged flight requires light bodies with a large surface area, something that has evolved in birds, reptiles (twice), mamals (at least once), and insects (multiple times).

egardless of how the traits are passed (we happen to know it's genetic) the point is that -- at this late stage -- absolutely any trait can appear fully formed in any species.

Nonsense. Cats can't spontaneously develop the same reumen found in a sheep. The two lineages diverged long before modern sheep evolved a reumen, and modern cats evolved into obligate carnivores. Pigs can't develop insect wings because vetebrates and arthropods split in the precambrian long before either arthropods or vertebrates colonized dry land.

Although let me add that it would be a very interesting negative result if a pig suddenly ended up with fully formed wings that were not genetically related to any other winged animals. It wouldn't disprove evolution, but it would be a poke in the eye.

I certainly would think that the sudden de novo appearance of such a complex feature would require the investigation of an alternate explanation of observed biodiversity.
posted by KirkJobSluder at 12:29 PM on December 9, 2008 [3 favorites]

tkolar: Hmm, to me that sounds like testing genetics, not evolution.

Of course, genetic inheritance was one of the first predictions made by Darwin to be experimentally confirmed.
posted by KirkJobSluder at 12:47 PM on December 9, 2008

This SIGGRAPH video from 1994 is still one of the most impressive GA demonstrations I've seen.

The green box game reaches not un-WALL-E-esque levels of cute around 2:45.
posted by dickymilk at 12:54 PM on December 9, 2008 [1 favorite]

While I'm ranting, there is another issue why pigs can't spontaneously develop wings as a sudden expression of latent genes. It's not as if there is a "wing" gene that can be switched on or off. The pig is already using its scapula, radius, ulna, carpals and metacarpals for another purpose. These are the exact same bones used by Aves and Chioptera. The developmental genes that determine the length, width, and orientation of these bones have been repurposed.
posted by KirkJobSluder at 1:01 PM on December 9, 2008

KirkJobSluder writes "While I'm ranting, there is another issue why pigs can't spontaneously develop wings as a sudden expression of latent genes. "

Except that many examples of development involve a gene (or several) being duplicated and a re-purposed: color-sensing cones in the eye, entire wings in insects (where one set of wings maintain their ancestral form, and another air are modified), etc.
posted by orthogonality at 1:07 PM on December 9, 2008

orthogonality: Both of those appear at first blush to me at least to involve redundant structures. Color sensing cones in the eye exist alongside color-insensitive rods. Insect wings involve multiple body segments. At least with pigs, it's hard to see how you can get wings without changing the entire pectoral girdle. Certainly it may be possible that pigs have an entire spare set of alternative genes for pectoral girdle development tucked away somewhere. But I doubt it.
posted by KirkJobSluder at 1:18 PM on December 9, 2008

The standard evolutionary explanation of this is [that the organism] *retained the genetic information for growing bark*

You kill me tkolar. At least you give me acid reflux. Are you being serious when you say that pigs already have the genetic code for wings? That a mammal has teh genetic code for a duck bill? Really?

This is an Intelligent Design argument! It is the people in the I.D. camp that make these kind of claims!

If the galapagos cactus has what is "clearly tree bark" and is a "cactus tree hybrid" like you claim, then this cactus clearly has hair, and is a hybrid between a saguaro and your grandmother.

I will continue to believe that you are just trolling or trying to be funny. Because you know what, if you really want to falsify the theory of evolution, you could find a number of species that shared characteristics of different nested groupings, like pigs with bird wings.

Pigs and birds are both Amniotes, the branch for pigs is Amniota-Synapsida, for birds it is Amniota-Reptilia-Romeriida-Diapsida. The last common ancestor of birds and pigs most likely looked like a small lizard, some 130 million years ago. A long time before bird wings.
posted by dirty lies at 1:24 PM on December 9, 2008 [1 favorite]

And of course, one of the advantages to the combination of evolutionary biology and molecular biology is that we can make a testable hypothesis in regards to the origin of those vestigial wings.
posted by KirkJobSluder at 1:29 PM on December 9, 2008

so attempting to logically refute them is fairly silly.

Well, if we're talking science, logical refutation is not the metric. Science builds models that predict future behavior. To do so requires simplification and parameterization. All these models are wrong and don't need to be refuted, but some of these models are useful because they do predict future behavior. "A wizard did it" has zero predictive value and so is useless as a scientific model. Emotional valence I leave for others to discern.
posted by Mental Wimp at 1:44 PM on December 9, 2008

What good is answering a question if some incurious bugger is just going to move on to the next half-assed talking point?
posted by Artw at 1:52 PM on December 9, 2008

The religious arguments against human evolution through natural selection are not based on logic, so attempting to logically refute them is fairly silly.

I've heard this statement often enough to figure it's worth refuting in its own right.

You should be able to use logic to demonstrate that they are not based on logic. That's something. The reason non-logic-based arguments persist is that a lot of people can't tell good reasoning from bad. If you think they'll just go away because they're not based on logic, you're wrong.

Poor logic or non-logic passes for reason often enough, unfortunately, that it's far from needless or superfluous to go through the exercise of disproof for the record. It is also wrong to suggest that bullshit statements are logic-proof and cannot be undermined by reasoned discussion. If that were the case we wouldn't have logic, science, discourse, or a theory of evolution at all. We'd still be bludgeoning each other with stones.
posted by scarabic at 2:13 PM on December 9, 2008 [1 favorite]

And here I thought it was eclipses that triggered the evolution of flying in politicians.
posted by cytherea at 2:13 PM on December 9, 2008

I don't know what you guys are talking about, so here's a fruit fly with legs growing out of its head.
posted by parudox at 2:25 PM on December 9, 2008

I've let my little car ecosystem evolve for a few hours, and it has converged to about the exact car I would design using engineering judgement. What was really neat to see was how the cars evolved different solutions for different parts of the track:

• At the beginning, where no individuals made it past the first hill, the ones that just stood upright were fittest.
• Later, when the individuals were only climbing the first few small uphills, they tended to put their mass on the front of the car
• When the cars reached their first major downhill, they had to spend a great deal of time evolving that weight back to the center of the wheelbase (and low).

It's too bad the simulation fixes the speed and limits the lifetime of each individual. I'd like to see the little cars change to travel even further (or have a line-rider style environment editor).
posted by Popular Ethics at 2:36 PM on December 9, 2008

Of course, genetic inheritance was one of the first predictions made by Darwin to be experimentally confirmed

Wait a second. Darwin knew about genes?
posted by AdamCSnider at 3:44 PM on December 9, 2008

This post fucking rules.
posted by greenie2600 at 5:37 PM on December 9, 2008

Wait a second. Darwin knew about genes?

As 'genes'? Nope (Bateson, in the early 1900's was the first to use the term, IIRC). But he surmised that there were granular units of heritability* that accounted for the transmission of traits between generations, that these worked independently or together to influence traits in the resulting generation, and that they could be modified by outside influences.

He got the mechanics of these units, and the methods of influence on them, totally wrong - he was rather invested in the ideas of Lamark, and was also a strong proponent of social equity (even if it was only amongst and between the "better races" ;-), which lead to his idea that inherited traits were the result of a sort of "one cell, one vote" democratic choice (gemmules, proposed in Variation of Plants and Animals Under Domestication). But the basic idea of "heritable units being passed to later generations" was sound, and was essentially a prediction later demonstrated empirically (& dodgily? ;-) by Mendel, who (deservedly) gets most of the credit.

Interestingly enough, because Darwin was initially and foremost interested in geology, his lines of thought seem to owe a lot to the granular nature of rocks and the changes in form they undergo due to geological processes.

(* I forget the exact terminology he uses in The Origin of Species ..., though he's definitely thinking of it in there. It's in my pile to be re-read in the next few weeks.)
posted by Pinback at 7:54 PM on December 9, 2008

KirkJobSluder wrote...
Nonsense. Cats can't spontaneously develop the same reumen found in a sheep. The two lineages diverged long before modern sheep evolved a reumen, and modern cats evolved into obligate carnivores.

You could definitely boil down the problem to my very strong suspicion that, were a cat to turn up with the same reumen found in a sheep, the response would be "Hey look, it turns out that the shared ancestor of cats and sheep did have a reumen after all!".
posted by tkolar at 9:53 PM on December 9, 2008 [1 favorite]

dirty lies wrote...
I will continue to believe that you are just trolling...

You caught me. I actually did this bit live in the Harvard Library once.

My girlfriend was sitting across the table from me and I said sotto voce "I think the general presentation of the theory of evolution is too broad to be considered scientific in itself. I believe it's more of a philosophical statement of how the biological sciences should be approached."

Naturally you can imagine the total chaos that ensued. I'm not even allowed back in the state of Massachusetts until 2012.
posted by tkolar at 10:00 PM on December 9, 2008 [2 favorites]

But the basic idea of "heritable units being passed to later generations" was sound,

I strongly disagree that this idea could be sourced with Darwin or any of his contemporaries. Animal husbandry had been around many thousands of years before they came along.

And as you mentioned, he got the mechanism *way* wrong.
posted by tkolar at 10:04 PM on December 9, 2008

scarabic wrote...
You should be able to use logic to demonstrate that [nonsense arguments] are not based on logic. That's something.

Is it, though?

The reason non-logic-based arguments persist is that a lot of people can't tell good reasoning from bad.

I disagree. The reason non-logic-based arguments persist is that a lot of people don't *care* about logic. They know what they feel is true, and they base their life and decisions on that.

Logicians make lousy politicians.
posted by tkolar at 10:24 PM on December 9, 2008

I strongly disagree that this idea could be sourced with Darwin or any of his contemporaries. Animal husbandry had been around many thousands of years before they came along.

Heritability has historically had at least two explanations. One is digital (genes) and one is analog ("mixing"). I do not know that Darwin or a contemporary first came up with the digital idea (likely not), but the fact that people bred animals before that proves nothing whatsoever.

It may behoove you to read a little about evolution and genetics before spouting off in threads on those topics.
posted by DU at 4:33 AM on December 10, 2008

Thanks for the clarification, Pinback.
posted by AdamCSnider at 6:29 AM on December 10, 2008

Evolve your own image
posted by DU at 11:59 AM on December 10, 2008

It may behoove you to read a little about evolution and genetics before spouting off in threads on those topics.

Really? That's the direction you want to take?

Sigh.
posted by tkolar at 12:32 PM on December 10, 2008

Let me add that it would be a very interesting negative result if a pig suddenly ended up floating in midair with no visible means of support. It wouldn't disprove physics, but it would be a poke in the eye.
posted by Michael Roberts at 8:24 PM on December 10, 2008