SETI's Seth Shostak on Intelligent Design

December 2, 2005 7:41 AM

Why SETI's searchfor intelligent extraterrestrial life is different from the work of proponents of Intelligent Design. An interesting bit of argumentation regarding the distinction between the simple signals searched for by SETI and the complex signals used in arguments for ID.
posted by voltairemodern (55 comments total)

Some of his thoughts on the nature of biological, natural systems seem a bit off... I wouldn't compare the "DNA Junk" in a sea lion (or any other eukaryote) to random electromagnetic noise in a star.

For one thing, that "DNA Junk" seems to be there for some reason, because it's not there in bacterial DNA. Bacterial DNA is stripped down and has little redundancy. Even Yeast, which is a eukaryote, has a very simple genetic structure, probably because it lives like bacteria.

No body knows exactly why all that "junk" is there, but it might help moderate genetic expression.

I think part of the problem with debating "intelligent design" is that people take the tack that there is no intelligence involved in evolution. But can't you say that the evolutionary process itself is intelligent? It can solve problems, can't it? Isn't that how we define intelligence?

I mean, an individual neuron isn't intelligent, yet, when you put billions of them together, you get a human mind (which obviously can be intelligent).

If you ask me, the first thing the ID folks need to prove, if their goal is to cast doubt on Evolution, is that Evolution itself is not intelligent.
posted by delmoi at 7:56 AM on December 2, 2005

I was confused about why ID proponents would think that SETI had any relevance to them at all. Thanks to the article, I now understand that they're either utterly ignorant of what SETI actually does, or willfully misinterpret their mission and methods. I'm inclined to believe that ID relies on people who do each.
posted by lodurr at 7:57 AM on December 2, 2005

... is that Evolution itself is not intelligent.

You s.o.b., I was hoping to get some work done today, and you had to go and start something interesting....
posted by lodurr at 7:59 AM on December 2, 2005

In other words, it seems like the ID folks are saying "Look at the Eiffel Tower, how can you say that was designed by unintelligent Neurons?"
posted by delmoi at 7:59 AM on December 2, 2005

but seriously, folks....

Delmoi, I'm inclinced at first blush to say that you're making an anthropomorphic error -- you're attributing agency to a thing which isn't actually even a thing. But though in my gut I think that's true, I don't have any argument that doesn't boil down to emergence: Agency and "intelligence" are emergent phenomena.

I'm not going to go digging for it (it's a print thing, and I have no idea even what magazine I read it in), but I remember reading years ago a quick gloss on the idea that it was an error to generalize from the concept of entropy to a generalized trend toward chaos -- that in fact the observable trends are always toward organization, not disorganization, and that it was possible to produce an explanation that incorporated entropy. I don't remember how the argument went, exactly, though.
posted by lodurr at 8:03 AM on December 2, 2005

(I get the analogy, delmoi, it just...bothers me. but I can probably be persuaded to get past that.)
posted by lodurr at 8:05 AM on December 2, 2005

design IS intelligence!
posted by blue_beetle at 8:14 AM on December 2, 2005

The main point of attack against this article will be that in the context of an Earth-like planet, would one reasonably expect the complexity observed on Earth. And while the people doing the attacking will be mongoloids of the highest and most cynical order, rhetorically it's a great starting point.
posted by Captaintripps at 8:18 AM on December 2, 2005

And, as an side, why the hell did they have to switch to that half-intelligently designed BOINC thing from the "Classic" SETI@home program? BOINC sucks.
posted by Captaintripps at 8:19 AM on December 2, 2005

Delmoi, I'm inclinced at first blush to say that you're making an anthropomorphic error -- you're attributing agency to a thing which isn't actually even a thing.

Yes, but:

I don't have any argument that doesn't boil down to emergence: Agency and "intelligence" are emergent phenomena.

You haven't left much for me to add here. "Agency", as you say, is an emergent property which comes about from the physical reality of our brains. I'm not quite sure what you mean by "agency" but if I make you feel better, I wouldn't call evolution an "agent" in that it does not have any particular goals or desires.

(Of course you have a question of perspective. From the perspective of a particular germ-line, evolution does appear to have a 'goal', the continuation of that germ-line. But when you get to the nitty-gritty that analogy doesn't hold up, there's so much going on with DNA that looking at individual genes with desires to propagate themselves gets really confusing. For example, in some environments e.coli will evolve to have much less error correction in its DNA splicing, so that it can mutate into different stains more easily. It gets really confusing when you think about "who" benefits from this change, rather then simply a predictable change that happens. So this agency metaphor is useful for understanding evolution in a broad sense, it breaks down in a lot of places).

But I was under the impression that you could have intelligence, and certainly 'creativity' without any goals or desires. A genetic algorithm does not have a goal, only a fitness function, yet it can come up with novel ideas, and solve some problems much better then a human. But the genetic algorithm emerges from a random number generator and a fitness function which are not intelligent.
posted by delmoi at 8:29 AM on December 2, 2005

I haven't read the article, but I know enough about creationist thought to say that SETI and it's prospect of finding extraterrestrial life presents a conundrum for the ID crowd.

At it's essence, ID and all the various flavors of creationism are still using an earth-centric model. Obviously Olde Time fundamentalists with their six-day creation myth and geocentric Bible believe that earth and life are at the center of God's creation (although few would argue that the earth is the center of our solar system.) With "contemporary creationism" (ID) this earth-centric view is slightly obscured, but still present in that their theory does not allow for the possibility of "other earths" and still elevates our planet and life to a special place above the rest of inanimate matter.

This is all moot though - SETI is a pipe dream doomed for failure (given the rarity of complex AND intelligent life, we are no doubt effectively cut off from other civilizations by time and space) and it maybe hundreds of years (if ever) before humans are able to find and image (via a massive space telescope) an obvious life-supporting (blue-green) planet in our galaxy. Ugh... that last sentence was messy.
posted by wfrgms at 8:37 AM on December 2, 2005

I hadn't realised that the SETI program was restricted to searching for such simple signals, so it's interesting to have this pointed out in this context.
It seems different people have different ideas about the meaning of the word 'complex'. I guess that to some, a random pattern would be the most complex, as it can't be reproduced exactly from a simpler algorithm. In this case, it's pretty clear that that to say complexity=intelligence becomes a nonsense.
I'm trying to decide how to apply the argument in the linked article to DNA sequence. What would be the equivalent of the 'persistent, narrow-band whistle' that SETI are looking for? Not the sequences that code for proteins, or the kind of regulatory elements that have been characterized. The simplest 'signals' in our DNA must be the low-complexity 'satellite' DNA, part of the junk DNA that Delmoi describes. We don't know what, if anything, that does so it doesn't help us.
Another, similar, simple repeating signal can be found in the trinucleotide repeats that cause so many diseases. These are stretches of DNA of extremely low complexity. Is this proof that an Intelligence causes these diseases? Can Spaghetti Monsterists expect a wrathful Designer to strike them down with Huntington's disease?
posted by nowonmai at 8:37 AM on December 2, 2005

"It can solve problems, can't it? Isn't that how we define intelligence?"

A rolling rock solves the problem of getting down a slope. But it isn't intelligent.

Seriously though.. evolution is just trial and error in action. It's just the fact that we only see the stuff that works that leads some people to believe that it's by design.
posted by Harry at 8:38 AM on December 2, 2005

But can't you say that the evolutionary process itself is intelligent?

That's a moot point, because the "intelligent" in "intelligent design" is code for "supernatural."
posted by rxrfrx at 8:43 AM on December 2, 2005

I am with Delmoi here and looking around at the human world I see so many things supposedly designed that have been created and modified down the years by trial and error. The stuff that works gets the green light for production the rest is largely forgotten.

I will even limb it here and say that in a way the design process itself has been created through trial and error.
posted by Samuel Farrow at 8:52 AM on December 2, 2005

The simplest 'signals' in our DNA must be the low-complexity 'satellite' DNA, part of the junk DNA that Delmoi describes. We don't know what, if anything, that does so it doesn't help us.

Er, not exactly. I'm under the impression that the 'junk' DNA is much less ordered then the coding portions.

A rolling rock solves the problem of getting down a slope. But it isn't intelligent.

You're missing the point.

That's a moot point, because the "intelligent" in "intelligent design" is code for "supernatural."

It is to the ID crowd, but this guy seems to be saying that DNA is "natural" because it's full of junk, even though the DNA of bacteria isn't full of junk. People who run around saying intelligent design isn't possible by pointing something they think is dumb are just as bad as people who try to prove it by pointing out something they think is smart.
posted by delmoi at 8:58 AM on December 2, 2005

nowonmai: it's pretty clear that that to say complexity=intelligence becomes a nonsense.

Yes... It starts to look like a classic duality.

In data communication you have a tension between organization and randomness. If you don't have enough repetitiveness you can't figure out how to frame the data you are receiving, but any repetitiveness is wasted throughput. I used to joke around in my digital communication class that the cosmic background radiation was a sign of intelligent life - they are just using frequency hoping spread spectrum and we can't figure out how to decode it.
posted by Chuckles at 9:04 AM on December 2, 2005

I think there are two problems with interpreting intelligence as part of Evolution.

First, what do you mean by intelligence anyway?

Second, I would argue that intelligence implies something other than trial by error. Crystals form through an analogous process to evolution. Molecules with the right alignment become fixed into the crystal, molecules with wrong alignments remain in solution. And yet we don't say that crystallization is "intelligent."

I would argue that part of any meaningful definition of intelligence involves some kind of a goal-directed limiting of solutions attempted. Furthermore, intelligence also includes avoidance of wrong solutions in future situations. We don't have either with evolution. Genetic variance and the rate of mutations is pretty close to random, and some of that variance and mutation will result in individuals that are poorly adapted to their environment.

In addition, the notion that "evolution solves problems" is pseudo-science, and probably one of the big issues with trying to turn evolution into a narrative. It is not as if some heroic ancestor to E. coli considered the inability to metabolize lactose a "problem." After all, many of its competitors that do just fine in animal guts do not.

Delmoi: A genetic algorithm does not have a goal, only a fitness function, yet it can come up with novel ideas, and solve some problems much better then a human. But the genetic algorithm emerges from a random number generator and a fitness function which are not intelligent.

And I would argue that while genetic algorithms are certainly a great way to computationally solve some problems, they don't strike me as being especially intelligent. (And where does the fitness function for these algorithms come from?)

Samuel Farrow: I am with Delmoi here and looking around at the human world I see so many things supposedly designed that have been created and modified down the years by trial and error. The stuff that works gets the green light for production the rest is largely forgotten.

Really, I can't think of a single example. In just about every design process I know about, the designer is applying some learned heuristics. We don't see for example automobile designs that assume a 20-foot tall human being, or iterative electrical jobs in which the wires do a random walk until they just happen to find a power source. And coffee pots with an unstoppered opening at the base.

Furthermore, you can do a cladistic analysis of organisms, and a cladistic analysis of human artifacts, and match them with predicted results of "trial and error" and design processes where designers build on prior heuristics. The results of this kind of analysis is pretty conclusive that the processes of human design and evolution are strongly different.
posted by KirkJobSluder at 9:06 AM on December 2, 2005

I also think that what is meant by "intelligent" in ID is "goal directed." To use the classic false example of trial and error, Thomas Edison didn't throw a random collection of materials across an electrical current and picking the best results. He knew he wanted to build a commercially viable electric light, he knew that other people had created light using carbon as a resistor. (That's right, he didn't invent it.) So he used this heuristic to limit his number of trials. (I suspect that he also learned a bit from each trial.)

IDers don't argue with the claims that some pretty inventive designs came out of evolution. They argue that those designs are too complex to have developed without an a priori goal driving the process.
posted by KirkJobSluder at 9:17 AM on December 2, 2005

They argue that those designs are too complex to have developed without an a priori goal driving the process.

There is an a priori goal. The goal is to survive and replicate, or you won't exist anymore. I don't quite understand how this isn't obvious to everyone.
posted by rxrfrx at 9:27 AM on December 2, 2005

I still think the processes are analogous KirkJobSluder, and I would be very interested to see primary research on a cladistic approach to assessing the evolution of human artifacts.

And you are missing my point on trial and error design - and without resorting to arguing semantics this early in the morning - design 'intelligence' builds incrementally on what has gone before. I think our cognition is a product of an evolutionary process - therefor all products of that must be evolutionary as well.
posted by Samuel Farrow at 9:30 AM on December 2, 2005

you all are missing the point, being dyslexic this stand out ....
"Agency and "intelligence" are emergent phenomena."

Intelligence -Agency= SIGNIT! connect the dots!

It not out there its all down here!
posted by OU812 at 9:35 AM on December 2, 2005

I would argue that part of any meaningful definition of intelligence involves some kind of a goal-directed limiting of solutions attempted. Furthermore, intelligence also includes avoidance of wrong solutions in future situations.

You state that intelligence has two properties:
1) "goal-directed limiting of solutions attempted."
2) "intelligence also includes avoidance of wrong solutions in future situations."

First of all, I can't see how those two things are different. Second of all It seems clear to me that evolution does have that property. Evolution is not simply trying every single possible solution randomly, forever. Rather it (is one of a class of algorithms that) makes small changes to things that work, and keeps the changes if they are better. So a solution that is very far away from something that already works is very, very unlikely to be tried. In "trial and error" a solution that is very far away from something that is already known to work is equally likely to be tried as one that is similar.

Furthermore, you can do a cladistic analysis of organisms, and a cladistic analysis of human artifacts, and match them with predicted results of "trial and error" and design processes where designers build on prior heuristics. The results of this kind of analysis is pretty conclusive that the processes of human design and evolution are strongly different.

What? Do you have any evidence of this? (Like a published paper) It seems like one of the most ridiculous statements I've ever read.
posted by delmoi at 9:39 AM on December 2, 2005

There is an a priori goal. The goal is to survive and replicate, or you won't exist anymore. I don't quite understand how this isn't obvious to everyone.

But whose goal is that? Not the cell, which might kill itself so that other nearby cells can live, not the gene, which can change, not the individual, which dies anyway. It's a metaphor that can break down in some very strange ways.

I think it's almost impossible to come up with a hard, solid definition of a physical thing which actually has this goal.
posted by delmoi at 9:44 AM on December 2, 2005

I think it's almost impossible to come up with a hard, solid definition of a physical thing which actually has this goal.

It all depends on the rules of the game, no? If we look at a flock of geese and say "things that look like geese are here now, and therefore they've won the evolution game" then the geese as a species had the goal, and therefore won the game. If it's a self-replicating, mobile DNA element that we're looking for, then it's that element that has won the game.
posted by rxrfrx at 10:10 AM on December 2, 2005

There is an a priori goal. The goal is to survive and replicate, or you won't exist anymore. I don't quite understand how this isn't obvious to everyone.
posted by rxrfrx at 9:27 AM PST on December 2 [!]

The rhetorical problem is that evolution as a process has no directed goal to survive and replicate — evolution is just a process — although products of evolution may claim to have that goal, or exhibit said goal, by virtue of their innate behavior or reflective consciousness.
posted by Rothko at 10:13 AM on December 2, 2005

I'm under the impression that the 'junk' DNA is much less ordered then the coding portions.
Junk DNA is largely composed of repeated elements. Depends to an extent on what you mean by 'ordered', but I think tandem repeats are highly ordered (in the respect used in the SETI article) although they carry little information.
posted by nowonmai at 10:18 AM on December 2, 2005

The rhetorical problem is that evolution as a process has no directed goal to survive and replicate

Yeah. I'll admit that it was disingenous of me to say "I don't understand how people find this non-obvious" because it's pretty clear that a lot of people are uncomfortable with amoral processes "deciding" our existence, and that's why the idea of evolution is so unpopular.
posted by rxrfrx at 10:19 AM on December 2, 2005

Furthermore, you can do a cladistic analysis of organisms, and a cladistic analysis of human artifacts, and match them with predicted results of "trial and error" and design processes where designers build on prior heuristics. The results of this kind of analysis is pretty conclusive that the processes of human design and evolution are strongly different.
posted by KirkJobSluder at 9:06 AM PST on December 2 [!]

Evolution nee selection is not "trial and error" in a purely random sense — environment (food, weather, disease, etc.) shapes or directs the viability (reproducability) of offspring, in the same way that (in your example) average human height guides a range of viable/sellable automobile designs. There's not all that much difference, procedurally.
posted by Rothko at 10:19 AM on December 2, 2005

An attempt to support ID with SETI must have Carl Sagan rolling in his grave.
posted by apis mellifera at 10:28 AM on December 2, 2005

Of course, the main difference between SETI and ID is that no one is trying to sue school districts and force them to teach Alienism.
posted by darukaru at 10:30 AM on December 2, 2005

What would be the equivalent of the 'persistent, narrow-band whistle' that SETI are looking for?

Generally, the approach that the SETI author takes is interesting--I also hadn't known about the small bandwidth of signals they looked for--but I think you're right to say it's misleading because it can't be generalized. Some "natural" things do look designed, according to the author's criteria, and some don't.

It seemed more effective as a defense of SETI than as an attack on Intelligent Design. Of course, prior to reading the article I would not have considered comparing the two, so I'm not sure why SETI needed defending in this way.
posted by voltairemodern at 10:36 AM on December 2, 2005

Of course, the main difference between SETI and ID is that no one is trying to sue school districts and force them to teach Alienism.

I'd sign up for Alienism as long as we don't have to drink any kool-aid...
posted by voltairemodern at 10:38 AM on December 2, 2005

rxrfrx: There is an a priori goal. The goal is to survive and replicate, or you won't exist anymore. I don't quite understand how this isn't obvious to everyone.

Oh, yeah. An a priori goal beyond just survival. The ID folks would argue that birds developed feathers rather than hair because of the a priori goal of winged flight. Or that bacterium developed specialized proteins over generations due to the a priori goal of rotary flagella.

Basically, ID redevelops the idea of morphic fields.

Sammuel Farrow: And you are missing my point on trial and error design - and without resorting to arguing semantics this early in the morning - design 'intelligence' builds incrementally on what has gone before. I think our cognition is a product of an evolutionary process - therefor all products of that must be evolutionary as well.

All of our cognition is a product of a quantum process, therefore, all products of that must be quantum as well! It's not that I don't get your point, it's that your point does not match observed reality.

Delmoi: First of all, I can't see how those two things are different. Second of all It seems clear to me that evolution does have that property. Evolution is not simply trying every single possible solution randomly, forever. Rather it (is one of a class of algorithms that) makes small changes to things that work, and keeps the changes if they are better. So a solution that is very far away from something that already works is very, very unlikely to be tried. In "trial and error" a solution that is very far away from something that is already known to work is equally likely to be tried as one that is similar.

Well, I think the difference is obvious. Goal-direction means that a population has a specific idea of how to solve the problem before natural selection takes place. (Dinosaurs knew that the solution to their long-term survival was developing wings for flight.) Learning means that evolution does not attempt failed solutions again. Evolution shows neither of these properties.

The fact of the matter is that mutation and natural selection will keep moving towards solutions that have been previously tried and failed. The ability to metabolize lactose is a very successful strategy for E. coli but every population of E. coli will have members that are not able to metabolize lactose.

What? Do you have any evidence of this? (Like a published paper) It seems like one of the most ridiculous statements I've ever read.

I'll have to hit Lexus Nexus when I have time. I seem to remember New Scientist, about two years ago. The study was done by an evolutionary biologist who also studies 19th and early 20th century horn instruments.

But the differences should be obvious. For example, the convergence of GUI designs between Apple, Microsoft, Gnome and KDE as they borrow from concepts proven by other designers.
posted by KirkJobSluder at 10:51 AM on December 2, 2005

"There is an a priori goal. The goal is to survive and replicate, or you won't exist anymore. I don't quite understand how this isn't obvious to everyone."
Because it's a teleologic assumption that's ungrounded.

Mutations are random. Some of them aid in propogation (broadly), and most of them don't.
(And I think that there is a pretty substantial difference between human design and natural selection/evolution— the evidence of intent and artificiality. But what do I know, I'm a guy with an appendix.)
posted by klangklangston at 11:05 AM on December 2, 2005

Well, I think the difference is obvious. Goal-direction means that a population has a specific idea of how to solve the problem before natural selection takes place. (Dinosaurs knew that the solution to their long-term survival was developing wings for flight.) Learning means that evolution does not attempt failed solutions again. Evolution shows neither of these properties.

Well, I already said I didn't think evolution had a specific goal in mind. If you're criteria for 'intelligence' is the same as the stuff discussed above as "agency", then I don't think evolution would be intelligence by your standard.

But if we were talking to an IDer, clearly they define intelligence as some force which can create things that have some measurable property of efficiency or complexity or something, something solves a specific problem (like winged flight) regardless of whether or not some agent had the desire for that solution.

The fact of the matter is that mutation and natural selection will keep moving towards solutions that have been previously tried and failed. The ability to metabolize lactose is a very successful strategy for E. coli but every population of E. coli will have members that are not able to metabolize lactose.

Well, I just don't think that is relevant at all. Some people do make the same mistake twice.
posted by delmoi at 11:28 AM on December 2, 2005

Because it's a teleologic assumption that's ungrounded.

It's not ungrounded; by making the statement "the things we see before us right now are the things that have survived," we are dictating the "rules of the game."
posted by rxrfrx at 11:37 AM on December 2, 2005

It's not ungrounded; by making the statement "the things we see before us right now are the things that have survived," we are dictating the "rules of the game."

But there is no reason to think there is any natural force which seeks to win the game we are describing.

"The game" is just a metaphor that encapsulates a lot of evolutionary biology, but (as I said) breaks down in some cases. All that that does is make easier for people to think about.
posted by delmoi at 11:41 AM on December 2, 2005

It's like thermodynamics. Entropy doesn't "want" to increase, it simply does.
posted by delmoi at 11:41 AM on December 2, 2005

KirkJobSluder I am on principle against ad hominem but All of our cognition is a product of a quantum process, therefore, all products of that must be quantum as well! It's not that I don't get your point, it's that your point does not match observed reality.

Tests the strength of my principles.

Spelling it out in short words for you friend - my point is that the development of the design process that you have inherited from your ancestors is itself a product of an evolutionary process, be it one whose information is coded in a media other than molecules.

The pace by which this information replicates and the substrate that this information replicates in differs from the genes that build our bodies - but the same top level rules apply - the products of these processes are analogous.
posted by Samuel Farrow at 11:44 AM on December 2, 2005

klangklangston: For three different definitions of intelligence that are being discussed, I have two different answers as to whether evolution is intelligent.

1) Intelligence is some force which produces artifacts which we can distinguish from artifacts from non-intelligent forces.

2) Intelligence is something that can solve problems

3) Intelligence is something that can perceive problems, desire to solve them, and then solve them, sating their desire.

I think evolution would be intelligent by definition one or two, but not three. Three, however, would be better encapsulated by the term "Agency", I think.

The first definition is the only one that can really be measured, though.
posted by delmoi at 11:47 AM on December 2, 2005

Samuel Farrow: I don't think those words were all that short. :P
posted by delmoi at 11:48 AM on December 2, 2005

delmoi - it is a pun - I was short (angry) when I wrote it :)
posted by Samuel Farrow at 11:58 AM on December 2, 2005

But there is no reason to think there is any natural force which seeks to win the game we are describing.

I really can't tell if you think I don't understand this, or are just trying to be extra-clear.
posted by rxrfrx at 12:18 PM on December 2, 2005

delmoi: Well, I just don't think that is relevant at all. Some people do make the same mistake twice.

Certainly, however most mistakes involve the misapplication of a principle that would work well in other circumstances. Most people may repeat a mistake a few times, but will reduce the frequency of that mistake over time.

Because mutations and re-assortment of genes is pretty much random, you are assured that an evolutionary "mistake" will be repeated millions, perhaps billions of times over geologic periods of time. Which is a good thing because over geologic periods of time, the "mistake" might turn out to be good within a newly created niche.

When we see people who make the same mistakes over and over again in the hope of finding a situation where the results might turn out beneficial, we tend to call them "foolish" or "stupid." By that standard, evolution is extravagantly "foolish" and "stupid."

But if we were talking to an IDer, clearly they define intelligence as some force which can create things that have some measurable property of efficiency or complexity or something, something solves a specific problem (like winged flight) regardless of whether or not some agent had the desire for that solution.

Well, I think this is granting a bit too much to both ID and evolution. There is nothing in evolution that favors winged flight, intelligence, or even rotating flagella as ways for a population to maximize reproductive success.

Sammuel Farrow: Spelling it out in short words for you friend - my point is that the development of the design process that you have inherited from your ancestors is itself a product of an evolutionary process, be it one whose information is coded in a media other than molecules.

he pace by which this information replicates and the substrate that this information replicates in differs from the genes that build our bodies - but the same top level rules apply - the products of these processes are analogous.

Well, I'll apologize. I've spent much of the past month arguing against a person who insists that evolution provides the only explanation of moral beliefs for secular humanists. After having spent more time than is prudent trying to patiently explain that evolution fundamentally is ONLY about changes in gene frequencies over time, I'm getting a bit frustrated with this.

I would argue that evolution is a strong theory because it works from some basic facts regarding how genes work.
1) genes primarily transfer from parent to offspring, and not between siblings, cousins, or other species (*).
2) genes are relatively stable over time.
3) it is possible to link genes to phenotypes with minimal dependence on context.

However, none of these assumptions apply to ideas:
1) ideas transfer to anyone who finds them useful and compatible with previously held ideas. (Rogers)
2) ideas change every time they are remembered (Loftus)
3) the "meaning" of ideas is strongly dependent on context with other ideas (Piaget, Eco).

Ideas and genes are both information. For that matter, quantum mechanics can be viewed as an information theory. But that does not mean that you can use quantum mechanics to explain shifts in gene frequencies over time, nor can you use evolution to explain the diffusion of ideas within cultures, nor can you use cultural psychology to explain the motion of electrons.

The processes of evolution, and the processes of cultural diffusion are only distantly analogous to each other. Once you get past the superficial details, evolution can't explain adoption of an innovation by multiple software vendors creating unrelated products, and cultural diffusion can't explain the structural difference in wings of unrelated bats, birds and insects.

Confusing the issue by trying to apply the theory of evolution outside of its scope, changes in gene frequencies in populations over time, only hurts evolutionary biology.

(*) While plasmids and viruses do complicate this a bit, even in bacteria we are talking about the transfer of a few genes out of thousands. The assumption of parent->offspring transfer is strong enough to describe most patterns in evolutionary biology.
posted by KirkJobSluder at 1:30 PM on December 2, 2005

1) genes primarily transfer from parent to offspring, and not between siblings, cousins, or other species (*).

Only in sexual reproduction, which is a relatively new development in evolutionary time. Given most life on earth is not multicellular, sex is not a given assumption for how genes propagate.

2) genes are relatively stable over time.

What does this mean?

3) it is possible to link genes to phenotypes with minimal dependence on context.

Huh? Phenotypes are measurements of subjective values; we research and catagorize phenotypes into measureable categories of subjective importance to us, e.g. weight, height, etc. Reductionism of this sort is entirely context-based.
posted by Rothko at 1:43 PM on December 2, 2005

Certainly, however most mistakes involve the misapplication of a principle that would work well in other circumstances. Most people may repeat a mistake a few times, but will reduce the frequency of that mistake over time.

Well, that's not what you said initially. Initially you stated that a necessary condition of Intelegence that an earlier mistake would not be repeated, now you're saying it should be repeated less frequently. Evolution is capable of doing this, actually, by duplicating important sets of genes so that a mutation doesn't knock them out.

But besides, it can be proven that this cannot be a property of any natural Intelegence.

1) Any Intelegence made from a finite amount of matter (A brain, all the DNA in the world, etc)
2) For an Intelegence to avoid making a mistake again in the future, it must record it somewhere
3) A record of a mistake takes up some physical matter as well.
4) Eventually the matter that makes up the Intelegence would fill up with mistake records, and no more could be made without erasing old ones.

Obviously if the Intelegence dies before it 'fills up' with memory this isn't an issue. But for an 'immortal' Intelegence, such as evolution, humanity in general this property is impossible to have.

When we see people who make the same mistakes over and over again in the hope of finding a situation where the results might turn out beneficial, we tend to call them "foolish" or "stupid." By that standard, evolution is extravagantly "foolish" and "stupid."

Well, this is just a matter of scale. No one would call those people non-intelligent the same way they would describe a crystal or a flame non-intelligent. Evolution may in fact be a rather stupid Intelegence by some measures. Certainly humanity is much better at solving problems on 'geologic' time frames.

Well, I think this is granting a bit too much to both ID and evolution. There is nothing in evolution that favors winged flight, intelligence, or even rotating flagella as ways for a population to maximize reproductive success.

!!!!!!!!!

I think it's pretty clear that all of those things increase fitness, which is half of the fundamental basis of evolution (mutation and natural selection)

Well, I'll apologize. I've spent much of the past month arguing against a person who insists that evolution provides the only explanation of moral beliefs for secular humanists. After having spent more time than is prudent trying to patiently explain that evolution fundamentally is ONLY about changes in gene frequencies over time, I'm getting a bit frustrated with this.

Yeah, that person is stupid. It's easy to get pissed off at ideas if they're misused all the time. But I'm not talking about ethics here.

However, none of these assumptions apply to ideas:
1) ideas transfer to anyone who finds them useful and compatible with previously held ideas. (Rogers)
2) ideas change every time they are remembered (Loftus)
3) the "meaning" of ideas is strongly dependent on context with other ideas (Piaget, Eco).

This really has nothing to do with our discussion.

Ideas and genes are both information. For that matter, quantum mechanics can be viewed as an information theory. But that does not mean that you can use quantum mechanics to explain shifts in gene frequencies over time, nor can you use evolution to explain the diffusion of ideas within cultures, nor can you use cultural psychology to explain the motion of electrons.

Well, DNA is made of molecules, which are made of atoms, which are governed by quantum mechanics. For that matter, people and brains societies are all made of molecules too, so they are all "based" on quantum mechanics in that sense. That doesn't mean you can take poorly understood concept from quantum mechanics and apply it to anything else broadly, which what you're probably complaining about.
posted by delmoi at 2:42 PM on December 2, 2005

2) genes are relatively stable over time.

What does this mean?

I think he means simply that they are stable bearers of information, as opposed to (for example) memories.


3) it is possible to link genes to phenotypes with minimal dependence on context.

Huh? Phenotypes are measurements of subjective values; we research and catagorize phenotypes into measureable categories of subjective importance to us, e.g. weight, height, etc. Reductionism of this sort is entirely context-based.

Agreed.
posted by voltairemodern at 2:49 PM on December 2, 2005

By the way, this post was totaly single-link-Op-Ed.
posted by delmoi at 2:54 PM on December 2, 2005

I think it's pretty clear that all of those things increase fitness, which is half of the fundamental basis of evolution (mutation and natural selection)

My reading of his point may be off, but I understood KirkjobSluder not as asserting that those things don't increase fitness, but that any particular increase in fitness is due to the interaction of traits with the environment. It's pretty easy to imagine environments in which each of the traits you mentioned would be detrimental to reproduction, and "the fittest" only refers to organisms that make it that far. So evolution, in the grand scheme of things, doesn't absolutely favor any of those traits; it favors whatever traits allow the organisms carrying them to survive long enough to reproduce.
posted by voltairemodern at 2:56 PM on December 2, 2005

By the way, this post was totaly single-link-Op-Ed.

Oh, here we were, almost 50 comments down on a non-newsfilter post, and no one had called me on it yet. And then you had to go and say it. Well, you'll get no apology from me...
posted by voltairemodern at 3:01 PM on December 2, 2005

Thanks KirkJobSluder - and I guess we disagree on how important the ideas about "genetic evolution" can be applied to "cultural evolution" to gain insights on our cultural development.

I am in no way advocating social darwinism.

I think the answer lies somewhere in the relationship between the unit of the gene as it applies to the organism and the unit of cultural information as it applies to, originally the hunter gatherer family group in our distant past but nowadays society, civilization and global (and maybe even intergalactic) communication.
posted by Samuel Farrow at 3:32 PM on December 2, 2005

Seth Shostak, at the SETI Institute, has an interesting online article criticizing ID here. His main charge is that ID has no business looking to SETI for support. Why? Because whereas ID looks for complexity in biology to detect design, SETI in fact looks for very simple signals to detect design, namely, radio signals with narrow bandwidth transmissions (not long, complex sequences of prime numbers as in Carl Sagan’s novel Contact).

But in fact, my criterion for design detection applies to the very signals that Shostak’s SETI Institute is looking for. Yes, as narrow bandwidth transmissions, the signals are simple to describe. But they are difficult for purely material processes to reproduce by chance. So we have simplicity of description combined with complexity in the sense of improbability of the outcome. That’s specified complexity and that’s my criterion for detecting design. It’s the same reason we detect design in the 1×4×9 monolith in 2001, A Space Odyssey. The structure is easily described; yet it is hard for natural processes to produce such rectangular solids by purely undirected material forces. ~ William Dembski
posted by bevets at 10:12 PM on December 2, 2005

rothko: Only in sexual reproduction, which is a relatively new development in evolutionary time. Given most life on earth is not multicellular, sex is not a given assumption for how genes propagate.

Parent-offspring relationships apply just as well in asexual reproduction, even more so because you don't have the complicating factor of reassortment.

2) genes are relatively stable over time.

What does this mean?

The mutation rate of genes is low enough that genetic traits can be passed on from generation to generation relatively unchanged. This means that we can make a probable match about relationships by comparing both phenotypes and genotypes.

Huh? Phenotypes are measurements of subjective values; we research and catagorize phenotypes into measureable categories of subjective importance to us, e.g. weight, height, etc. Reductionism of this sort is entirely context-based.

Well, I would argue that this is entirely subjective. For every phenotype, the degree of genetic influences on the variance of that phenotype can be quantified, and those influences can be identified as specific genes. So it is possible to say, just as well as we can say there is a moon around pluto, that gene X accounts for 40% of the variance in phenotype Y.

But there is a basic assumption in evolutionary biology that a gene for lactase doesn't change function when moved next to a gene for luciferase. With hypothetical "memes" we have to address the problem that ideas in close proximity to each other change in function and meaning.

Delmoi: Evolution is capable of doing this, actually, by duplicating important sets of genes so that a mutation doesn't knock them out.

Certainly evolution can come up with such a "solution" (although duplication of genes causes its own problems). However, you are still going to have the "mistake" of single copies repeated over and over again over generations.

Obviously if the Intelegence dies before it 'fills up' with memory this isn't an issue. But for an 'immortal' Intelegence, such as evolution, humanity in general this property is impossible to have.

Well, I think one of the problems here is that you also are applying the theory of evolution beyond the scope and scale at which it applies. Which is again, I think a big problem discussed earlier with proposing evolution as a narrative rather than a mechanism.

Fundamentally, you have a range of phenotypes in a population. The reproductive success of phenotypes (or individuals possessing phenotypes) is going to be different across the range. This will result in a statistical shift of that range over time.

In the scale of geologic time, this results in wings, intelligence and flagella. However, natural selection does not work on that scale. Evolution works on the scale of generations. And what it is selecting for probably isn't as profound as wings, intelligence and flagella, but more likely the length of individual bones, organ size, and tweaks in protein function.

From my point of view, you are opening up a huge can of worms by saying that "evolution solves problems." "Problems" are a human interpretation the phenomena, and you can just as easily say that crystal formation and flames are "intelligent solutions" to problems of thermodynamics. You are anthropomorphizing evolution.

This really has nothing to do with our discussion.

It's relevant in that Samuel Farrow seems to be making claims that human design processes are analogous to evolutionary processes.

voltairemodern: So evolution, in the grand scheme of things, doesn't absolutely favor any of those traits; it favors whatever traits allow the organisms carrying them to survive long enough to reproduce.

Bingo. However in most cases, the "trait" that is being selected for is not a "wing" but something like, "feathers on the back side of a forelimb with a length 0.2 standard deviations above the mean." Really I think that a big killer in talking about evolution is the X-men effect in which we talk about birds evolving a new feature like wings, rather than quantitative genetics which models shifts in statistical ranges over time.

Sammuel Farrow: I think the answer lies somewhere in the relationship between the unit of the gene as it applies to the organism and the unit of cultural information as it applies to, originally the hunter gatherer family group in our distant past but nowadays society, civilization and global (and maybe even intergalactic) communication.

Well, I guess I don't see any reason why we should assume such a relationship. Biological systems included dozens, if not thousands of different types of "units of information." For example, while evolution describes changes in gene frequencies over time, it does not describe the process by which genes are turned on or off, the process by which DNA is translated into mRNA, modifications performed on mRNA, and the resulting protein complexes. Certainly the process by which protein is created is a product of evolution, but it works by fundamentally different rules.

The same can be said for the immune system. In fact, the immune system must handle information using fundamentally different rules from evolution because pathogens evolve faster than vertebrates.

Likewise, I don't see a reason to assume that even if human cognition is a product of evolution, that the units of information in human cognition (and cultural diffusion) are analogous to genes. The key advantage for learned animal behavior adaptation to rapidly changing and dynamic environments within a generation.

Also, I didn't accuse you of being a "social darwinist" which means something radically different from what you are proposing.
posted by KirkJobSluder at 6:21 AM on December 3, 2005

Crap, tags fixed.

rothko: Only in sexual reproduction, which is a relatively new development in evolutionary time. Given most life on earth is not multicellular, sex is not a given assumption for how genes propagate.

Parent-offspring relationships apply just as well in asexual reproduction, even more so because you don't have the complicating factor of reassortment.

2) genes are relatively stable over time.

What does this mean?

The mutation rate of genes is low enough that genetic traits can be passed on from generation to generation relatively unchanged. This means that we can make a probable match about relationships by comparing both phenotypes and genotypes.

Huh? Phenotypes are measurements of subjective values; we research and catagorize phenotypes into measureable categories of subjective importance to us, e.g. weight, height, etc. Reductionism of this sort is entirely context-based.

Well, I would argue that this is NOT entirely subjective. For every phenotype, the degree of genetic influences on the variance of that phenotype can be quantified, and those influences can be identified as specific genes. So it is possible to say, just as well as we can say there is a moon around pluto, that gene X accounts for 40% of the variance in phenotype Y.

But there is a basic assumption in evolutionary biology that a gene for lactase doesn't change function when moved next to a gene for luciferase. With hypothetical "memes" we have to address the problem that ideas in close proximity to each other change in function and meaning.

Delmoi: Evolution is capable of doing this, actually, by duplicating important sets of genes so that a mutation doesn't knock them out.

Certainly evolution can come up with such a "solution" (although duplication of genes causes its own problems). However, you are still going to have the "mistake" of single copies repeated over and over again over generations.

Obviously if the Intelegence dies before it 'fills up' with memory this isn't an issue. But for an 'immortal' Intelegence, such as evolution, humanity in general this property is impossible to have.

Well, I think one of the problems here is that you also are applying the theory of evolution beyond the scope and scale at which it applies. Which is again, I think a big problem discussed earlier with proposing evolution as a narrative rather than a mechanism.

Fundamentally, you have a range of phenotypes in a population. The reproductive success of phenotypes (or individuals possessing phenotypes) is going to be different across the range. This will result in a statistical shift of that range over time.

In the scale of geologic time, this results in wings, intelligence and flagella. However, natural selection does not work on that scale. Evolution works on the scale of generations. And what it is selecting for probably isn't as profound as wings, intelligence and flagella, but more likely the length of individual bones, organ size, and tweaks in protein function.

From my point of view, you are opening up a huge can of worms by saying that "evolution solves problems." "Problems" are a human interpretation the phenomena, and you can just as easily say that crystal formation and flames are "intelligent solutions" to problems of thermodynamics. You are anthropomorphizing evolution.

This really has nothing to do with our discussion.

It's relevant in that Samuel Farrow seems to be making claims that human design processes are analogous to evolutionary processes.

voltairemodern: So evolution, in the grand scheme of things, doesn't absolutely favor any of those traits; it favors whatever traits allow the organisms carrying them to survive long enough to reproduce.

Bingo. However in most cases, the "trait" that is being selected for is not a "wing" but something like, "feathers on the back side of a forelimb with a length 0.2 standard deviations above the mean." Really I think that a big killer in talking about evolution is the X-men effect in which we talk about birds evolving a new feature like wings, rather than quantitative genetics which models shifts in statistical ranges over time.

Sammuel Farrow: I think the answer lies somewhere in the relationship between the unit of the gene as it applies to the organism and the unit of cultural information as it applies to, originally the hunter gatherer family group in our distant past but nowadays society, civilization and global (and maybe even intergalactic) communication.

Well, I guess I don't see any reason why we should assume such a relationship. Biological systems include dozens, if not thousands of different types of "units of information." For example, while evolution describes changes in gene frequencies over time, it does not describe the process by which genes are turned on or off, the process by which DNA is translated into mRNA, modifications performed on mRNA, and the resulting protein complexes. Certainly the process by which protein is created is a product of evolution, but it works by fundamentally different rules.

The structure of genes in eukaryotes is another issue altogether and largely independent of evolution. As I noted before, evolution only assumes the inheritance of phenotypes. The actual structure of the genes involved can be considered a "black box" as far as evolutionary biology is concerned.

The same can be said for the immune system. In fact, the immune system must handle information using fundamentally different rules from evolution because pathogens evolve faster than vertebrates.

Likewise, I don't see a reason to assume that even if human cognition is a product of evolution, that the units of information in human cognition (and cultural diffusion) are analogous to genes. The key advantage for learned animal behavior is adaptation to rapidly changing and dynamic environments within a generation.

Also, I didn't accuse you of being a "social darwinist" which means something radically different from what you are proposing.
posted by KirkJobSluder at 6:27 AM on December 3, 2005