This Week in Intelligent Design - 05/04/11

| 48 Comments

Intelligent design news from the 30th of March to the 5th of April, 2011.

Sometimes, keeping up with the intelligent design movement can feel like a full-time job. Other times… not so much. While some of the output by the Discovery Institute and its related organisations is somewhat novel, most of it is simply rehashed ideas from a limited pool. This week was a fairly good example of the latter. We had copy-and-paste arguments for the positive nature of the design argument, as well as an unsurprising plug for a blatantly religious debate, not to mention many, many posts on Uncommon Descent about- well, I don’t really know. The words just tend to blend together after a while, forming a soup of pseudo-philosophy and rhetoric.

The only really “novel” thing this week was something to do with April Fool’s Day… and it wasn’t novel in a good way.

48 Comments

So humans design stuff for various reasons or no reason at all, and so the intelligent designer(s) must have designed stuff for some reason or no reason at all (or some mysterious reason we can’t know about and shouldn’t even ask). I think I’ve got it!

If humans were designed, whoever did the design is not our friend.

2) Intelligent agents can rapidly infuse large amounts of information into systems.

True, but they can also take it away from systems. This happens far more frequently than you might at first think. Plus, the level of information change might not be large – human designers can work rather slowly at times, especially when they’re given access to the Internet.

I’ve developed the habit of asking that all statements about “information” or “complexity” be related to some accepted, objective defintion, such as Shannon information or Kolmogorov complexity, or, if necessary, a new one created by the claimant, but one that is coherent and can be used in a reproducible way.

For example, DNA sequences certainly are (not “contain”, but “are”) information, under Shannon information theory, if a human is trying to sequence DNA. However, that’s equally true of the iron content of a set of randomly gathered mud samples - if I human is trying to measure the iron contents of the mud samples, then in that context, the iron contents are (not “contain”, “are”) information.

Since information, under what is usually regarded as the most accepted, reproducible, and rigorous treatment, is created by the act of observation, the fact that a phenomenon is (transiently) information has nothing to do with whether or not it was “designed”.

I do think it is fair, when discussing with people of good faith, to use the terms in an informal way, as in “intestinal parasite worms seem to have evolved to be less complex than their free living ancestors”. This reminds us that evolution need not be “directional”. (Although more complex life can only come into being by evolving from less complex life, once a lineage is in existence, its descendants can be either more or less complex).

However, we need to remember that this is an intuitive, informal way of using the term “complex”. It is acceptable because the statements are not at odds with reality, and could be restated in more rigorous terms.

When someone comes along and starts playing bad faith word games, falsely claiming to overturn evidence and theory, then the first response should be to make sure that terms are very clearly defined.

In informal usage, “information” means simply “useful data”, whether it’s DNA sequence or chemical percentages.

he fact that a phenomenon is (transiently) information has nothing to do with whether or not it was “designed”.

Yep. The fact that somebody finds a use for something definitely does not mean that the something was designed deliberately engineered to be useful.

The one question ID fans never seem to answer is “what is the ‘irreducibly complex’ difference between a human and a chimpanzee?”. I mean, who really cares, theologically, whether God (or the Flying Spaghetti Monster) created the first DNA? If you can’t show that some added “information” was needed to make a human out of a common human-chimp ancestor, then you’re wasting your time as far as Christian fundamentalists are concerned.

In my field of endeavour, “information” is that which allows one system to gain useful knowledge of the state of another system.

“Useful” is, unsurprisingly, entirely context dependent.

And the “another system” may often be the supersystem of which the recipient system is a subsystem.

One could respond to Luskin’s hand-waving with a good many relevant facts and questions–like by asking why convergence, clearly something that only makes sense in evolutionary context, would be expected at all with “design,” which simply can and usually does actually copy designs without much regard to lines of descent.

But there’s a general question that they can’t answer, which is, “Why is life highly constrained by the past as entailed by evolution, and not relatively unconstrained by the past as expected from intelligence?”

His blithering is meant to obfuscate that question, rather than to deal with it.

Glen Davidson

But there’s a general question that they can’t answer, which is, “Why is life highly constrained by the past as entailed by evolution, and not relatively unconstrained by the past as expected from intelligence?”

I think the answer is that their intelligent designer just felt like doing it that way. Evidently he/she is a capricious goofball who likes to fool us.

Glen Davidson said: But there’s a general question that they can’t answer, which is, “Why is life highly constrained by the past as entailed by evolution, and not relatively unconstrained by the past as expected from intelligence?”

An engineering way of putting this is: “Why are there no ‘clean sheet of paper’ Designs in nature?”

Human designers often adapt what was available before in their designs, but sometimes they just come up with something entirely new.

Human designers often adapt what was available before in their designs, but sometimes they just come up with something entirely new.

Or a recombination of features from previously independent lines of development. (Which Glen already mentioned, so this post isn’t entirely new itself.)

Shebardigan said: In my field of endeavour, “information” is that which allows one system to gain useful knowledge of the state of another system.

If the output of a system is a string of random* numbers, then i have gained useful knowledge that the system is a random-number generator. Likewise, if a system produces a noisy signal, the noise itself tells me something about that system’s characteristics and performance. So design proponents would probably reject your definition as too inclusive of things they don’t consider real information. :)

*For the pedants; yeah we can’t ever tell for sure. Assume we’re talking a string that at least appears random to every test you give it.

eric said:

If the output of a system is a string of random* numbers, then i have gained useful knowledge that the system is a random-number generator.

Heh. In our cases, it would tell us that the communication channel was bad broke, unless we indeed knew ahead of time that that remote system was supposed to provide us with “random” numbers.

So design proponents would probably reject your definition as too inclusive of things they don’t consider real information. :)

Imagine my sorrow and/or discomfiture at this unwholesome turn of events.

mrg said:

An engineering way of putting this is: “Why are there no ‘clean sheet of paper’ Designs in nature?”

Human designers often adapt what was available before in their designs, but sometimes they just come up with something entirely new.

In this context, it is worth noting that no ancient hand-axe was actually “designed”. Various primates who were used to bashing things with rocks discovered that some shapes of rocks were better at separating stuff into bits than other rocks were, and they tended to keep those rocks around, and look for other rocks that had similar shape for use when the original “found” rocks wore out or broke.

In the event that no similar-shaped rocks were around, the primates who had played around with banging rocks together noted that, using the right kind of banging, you could make a rock with the same characteristics as the found-on-the-ground hand axe rocks. They were not designing, they were manufacturing copies.

Over time, certain inadvertent changes to the general axe-rock pattern turned out to be better at hacking things to bits than the original, so those tended to be more often duplicated by rock bangers.

Sophisticated hand axes were almost never designed, they were nearly always evolved.

(I had a lovely found-on-the-ground hand axe rock that I picked up on a hike when I was growing up in Western Colorado. Sharp, ready to hand, and obviously a product of natural events.)

So much for Intelligent Design.

Shebardigan said:

In this context, it is worth noting that no ancient hand-axe was actually “designed”. Various primates who were used to bashing things with rocks discovered that some shapes of rocks were better at separating stuff into bits than other rocks were, and they tended to keep those rocks around, and look for other rocks that had similar shape for use when the original “found” rocks wore out or broke.

In the event that no similar-shaped rocks were around, the primates who had played around with banging rocks together noted that, using the right kind of banging, you could make a rock with the same characteristics as the found-on-the-ground hand axe rocks. They were not designing, they were manufacturing copies.

Over time, certain inadvertent changes to the general axe-rock pattern turned out to be better at hacking things to bits than the original, so those tended to be more often duplicated by rock bangers.

Which makes me wonder when generalized, abstract pattern-recognition developed in our ancestors (hey! Sharp edge = better meat scraper!), or the discovery that we could actually make some rock/bone into some other shape that suited our needs.

Bio-Essays editor Andrew Moore has weighed in on whether we ought to acknowledge the reality of Design:

http://onlinelibrary.wiley.com/doi/[…]01190011/pdf

Not surprisingly, this resulted in the latest instance of breathtaking inanity from Dishonesty Institute mendacious intellectual pornographer Casey Luskin:

http://www.evolutionnews.org/2011/0[…]b045551.html

Am sure Luskin is probably aware that Ken Miller has defended the existence of Design in Nature, including evolutionary biology, by arguing that we need to acknowledge its existence and its emergence via natural processes like Natural Selection, so we can refute Intelligent Design creationist claims that the existence of Design “proves” the existence of an Intelligent Designer(s) (who could more likely be Klingons or the Time Lords, not Yahweh or Zeus IMHO).

Or perhaps an inter-planetary design coalition.

Moore may be slightly overstating a good case – I wince at some of the anthropomorphisms in documents by people who should know better, but I think terms like “evolutionary strategy” may be hard to displace.

However, Luskin is in unusually absurd form: “The nerve of those guys to suggest that they ought to stop explaining their ideas using clumsy and inaccurate terminology that we have found only too convenient to exploit.”

From the Bio-Essays article linked by John Kwok

In this issue Jacques Dubochet asserts that a major reason for the lack of public acceptance of evolution is a fundamental misunderstanding of its core concepts: many people think of evolution producing modifications with an aim, e.g. ‘eyes in order to see’, ‘legs for walking’, rather than in terms of directionless variation, and natural selection without motivation, design or strategy.

While I strongly agree with this, and with the overall suggestions in the article (in fact I have tried to use similar language to what they describe for years), this is the problem faced when people make good faith efforts to understand evolution.

The concept that “an unbelievably huge number of randomly* genetically variant offspring are being born per unit time, and a small fraction of the variation results in phenotypic changes which are either, in a probablistic way, selected for or against, or increase in frequency for purely random reasons, within a local environment” is hard to grasp.

(*Some types of mutations are more common than others, and for some, we can accurately forecast the long term frequency with which they will occur, but not exactly when they will occur. This is the characteristic of a “random variable” in mathematics, and individual mutations are random events relative to human observers. This does not imply that all mutations occur at equal frequency.)

The human brain sees frog ancestors as “knowing” that strong jumping legs and an insect-snapping tongue would be “desirable”, or even as being aware that they “needed” those features. The anthropomorphic language emerges over and over again.

Having said that, evolution denial is grounded in the social/political/religious desire to deny that all humans share common ancestry with other species and with each other. It’s hard to get polls on whether or not people deny evolution in plants and bacteria. That is because, for whatever reason, the polls are almost always constructed in a way the asks about human or animal evolution, AND sets up acceptance of evolution as a confrontational denial of mainstream religion. As I have noted, these biased polls seem to please both creationists and a subset of non-creationists who like to dwell on the presumed “stupidity” of a vast number of other people.

What these guys are talking about would improve understanding of evolution among the sincerely interested. However, most denialists don’t even care about correct understanding (and indeed actively seek to avoid it). So better language would help - but not with reducing ID/creationism.

harold said:

From the Bio-Essays article linked by John Kwok

In this issue Jacques Dubochet asserts that a major reason for the lack of public acceptance of evolution is a fundamental misunderstanding of its core concepts: many people think of evolution producing modifications with an aim, e.g. ‘eyes in order to see’, ‘legs for walking’, rather than in terms of directionless variation, and natural selection without motivation, design or strategy.

While I strongly agree with this, and with the overall suggestions in the article (in fact I have tried to use similar language to what they describe for years), this is the problem faced when people make good faith efforts to understand evolution.

The concept that “an unbelievably huge number of randomly* genetically variant offspring are being born per unit time, and a small fraction of the variation results in phenotypic changes which are either, in a probablistic way, selected for or against, or increase in frequency for purely random reasons, within a local environment” is hard to grasp.

(*Some types of mutations are more common than others, and for some, we can accurately forecast the long term frequency with which they will occur, but not exactly when they will occur. This is the characteristic of a “random variable” in mathematics, and individual mutations are random events relative to human observers. This does not imply that all mutations occur at equal frequency.)

There is widespread interst in tracing one’s geneaology. What should be done is to emphasize that these “random” mutations are constrained by the prior genealogical history - what we know as phylogenetic history - of the population that has such mutations and, as such, Natural Selection is using these mutations to promote sufficient shifts in the gene frequencies of that population that could arise in one or more speciation events within it. I would hope that if such connections might be made, then the public would have a better understanding of evolution, and perhaps, a more personal connection to it that would lead to their acceptance of it as valid mainstream science.

What should be done is to emphasize that these “random” mutations are constrained by the prior genealogical history

This is absolutely true, and important to remember.

In fact, random variables are sampled from distributions, which are usually constrained. The roll of a cubical die is a good example. The results can only be the integers one through six, but to a typical human observer, the results of future rolls still represent random variables.

In no way, shape or form does the word “random” imply “totally unconstrained by reality” or “only applicable to a distribution in which every possible result occurs with equal frequency”.

When a given strand of DNA is replicating, for example, only those mutations which are physically possible within the contest of that sequence of DNA can occur, and among those, some have much, much, much greater probability of happening than others.

The random part of mutation accumulation is somewhat like a Markov chain. http://en.wikipedia.org/wiki/Markov_chain. Of course, screening by selection is also important.

Note that you can correctly infer a great deal about the history of a genome, and its relationship to other genomes, from its sequence, and of course, in pure theory, you could also construct a set of possible offspring genomes. In fact, that set would be massively large, but much, much smaller than the 4^3,000,000,000 possible 3 billion base pair DNA sequences that could theoretically exist. However, what you can’t do, for the foreseeable future, is predict which of the possible mutations will occur at the next replication event.

And after that very next replication event, if there was mutation, then the “massively large” theoretical “possible offspring genomes” has changed for the next generation, and so on into the future…correct? And this “butterfly effect” of minor present changes constraining future possibilities would make long term evolutionary prediction impossible–except of course “moving in the direction of increased fitness for the environment at the time”.

harold said:

What should be done is to emphasize that these “random” mutations are constrained by the prior genealogical history

This is absolutely true, and important to remember.

In fact, random variables are sampled from distributions, which are usually constrained. The roll of a cubical die is a good example. The results can only be the integers one through six, but to a typical human observer, the results of future rolls still represent random variables.

In no way, shape or form does the word “random” imply “totally unconstrained by reality” or “only applicable to a distribution in which every possible result occurs with equal frequency”.

When a given strand of DNA is replicating, for example, only those mutations which are physically possible within the contest of that sequence of DNA can occur, and among those, some have much, much, much greater probability of happening than others.

The random part of mutation accumulation is somewhat like a Markov chain. http://en.wikipedia.org/wiki/Markov_chain. Of course, screening by selection is also important.

Note that you can correctly infer a great deal about the history of a genome, and its relationship to other genomes, from its sequence, and of course, in pure theory, you could also construct a set of possible offspring genomes. In fact, that set would be massively large, but much, much smaller than the 4^3,000,000,000 possible 3 billion base pair DNA sequences that could theoretically exist. However, what you can’t do, for the foreseeable future, is predict which of the possible mutations will occur at the next replication event.

I agree completely harold, and am commenting only to correct this as follows:

What should be done is to emphasize that these “random” mutations are constrained by the prior genealogical history - what we know as phylogenetic history - of the population that has such mutations and, as such, Natural Selection is using these mutations to promote sufficient shifts in the gene frequencies of that population which could lead to one or more branching events (speciation) of new populations diverging from the earlier ancestral one.

Just Bob and John Kwok -

I agree with everything both of you have said.

Although this is interesting -

would make long term evolutionary prediction impossible

It depends on what precision level you are trying to predict at.

Because of phylogenetic history constraints, we can make many general predictions. We can note that horses with four legs and two bird-like wings are very unlikely to arise, for example, because of deeply conserved body plan constraints in the lineage.

Less trivially, we can predict all sorts of things like insecticide resistance, antibiotic resistance, selection for relatively smaller large individuals in island habitats, and emergence of local ecological balance between herbivore and carnivores.

However, such predictions are inexact and imprecise. We certainly can’t predict the details perfectly.

harold said: Because of phylogenetic history constraints, we can make many general predictions.

See Dougal Dixon’s classic bestiary of the future, AFTER MAN – rat-derived carnivores up to analogues of polar bears, rabbit-derived large herbivores like llamas, and most intriguingly penguins evolved into analogues of baleen whales.

How about an analogy (FL, you won’t understand this, so just skip it).

The “randomness” that creationists think must go on in genetics (if it weren’t for “design” or other supernatural tinkering) is indeed like the rolling of that six-sided die: completely unpredictable, random, chaotic, with the previous roll(s) having no influence whatever on the next one. Thus predicting the exact sequence of, say, the next 100 numbers to come up becomes impossible, since any sequence is as likely as any other. (Yes, that means 100 6’s in a row are exactly as likely as a seemingly “random” sequence.)

But the true “randomness” in genetics is more like that in predicting future chess positions. A player’s next move is absolutely constrained to the legal moves possible, determined by the positions of other pieces on the board at the time. E.g., for the first move of the game there are only 20 possible actions; at some points there can be many more than that, and at some points only ONE legal move. But every possible “next position” in chess is completely constrained by the situation on the board at the moment.

Thus a chess computer could easily print out a list of every possible situation, say, five moves into the future. In midgame that might be a huge list, but it would be finite. We can be absolutely confident the board will be in one of those states. And for any one state, the computer could print out a list of the very few possible pathways that could have led to that situation. And knowing the skill level (and even the psychology) of an opponent, can narrow future positions down dramatically.

Rolling a die repeatedly is like the creationist misconception of the “randomness” in evolution: any damn thing could be the result. But RM+NS is much more like possible future chess positions: each next position is limited to what is legal after the last one, AND there’s “selection pressure” exerted by what the player thinks would be a GOOD move (and by what he perceives as likely from the opponent).

A chimpanzee isn’t going to produce a human baby, just like your queen isn’t going to cross the board on the first move. A horse descendant in the next million years developing wings is so unlikely as to be considered impossible with confidence, just as your winning the game with only a king, while your skilled opponent still has 16 pieces.

Hey, I like that analogy. Pieces that make good or indifferent moves stay on the board; those that make bad moves get taken off the board.

I like that analogy. However, there are absolutely no creationists I have ever encountered who even pretend to understand the concepts of historical contingency or cumulative selection. The words might as well be meaningless to them. That is the only thing that allows them to continue using meaningless crap such as tornado in a junkyard, so I don’t expect it to change anytime soon.

Any argument is good enough when your goal is simply to deny reality. It doesn’t matter if the argument makes any sense or not. It doesn’t matter if the argument is contrary to the facts or not. All that matters is that someone is dumb enough or ignorant enough to fall for it, at least for a little while. Somehow they never seem able to consider long term consequences of using such stupid arguments either. Coincidence? I don’t think so.

Just Bob said: Rolling a die repeatedly is like the creationist misconception of the “randomness” in evolution: any damn thing could be the result.

Actually, the result of a die roll is heavily constrained by its physical structure to be between 1 and 6. So they’re still apt. Creationist misconceptions (crocoducks and dogs with wings) seem to run along the lines of thinking a roll of 5 and a roll of 12,356,980 is equally likely.

AND there’s “selection pressure” exerted by what the player thinks would be a GOOD move (and by what he perceives as likely from the opponent).

Here, the chess analogy holds up better than the dice one. Dice have no memory, so ‘selecting’ some subset of dice based on their past roll will not alter the odds of getting some result on a future roll. However, selection does work on dice players. If you know the odds of various craps results, you can quickly confirm that the players who play the long odd bets tend to leave the table faster than the ones who play the pass line with odds.

Just Bob -

Yes, that’s pretty accurate. Maybe backgammon would be a better example. (In fact, computers learned backgammon via neural net methods, which are a variant of random choices acted on by selection.)

There are many constraints in the biological process.

The first constraint is the fact that you’re staring with what you start with. You have the DNA sequence that is the result of past evolution.

The second constraint is that mutations have to be chemically possible. Yes, I know, duh, but we are talking about chemical reactions not pure computer models. Important to remember that.

The constraints imposed by selection can kick in at any time. If a given offspring genome (of the DNA template genome) is not compatible with early development, for example, there you go right there.

As many of us have noted many times, the absence of croco-ducks and half-cat-half-dogs is evidence against creationism. Under creationism, omnipotent magic is a common cause of earthly events, and there is no reason why crcoducks, or rhinostriches, or anything else shouldn’t magically appear at any time. The theory of evolution does include the element of mutations that occur in a random (not perfectly predictable even with good information about frequency) way with respect to human observers. However, it also explains why there are constraints on what can happen, something than ID/creationism most certainly does not do.

harold said:

As many of us have noted many times, the absence of croco-ducks and half-cat-half-dogs is evidence against creationism.

So if creationists want to stick with “kinds,” then dragons - with their front and back legs along with their wings - had to have micro-evolved from insects.

eric said:

Just Bob said: Rolling a die repeatedly is like the creationist misconception of the “randomness” in evolution: any damn thing could be the result.

Actually, the result of a die roll is heavily constrained by its physical structure to be between 1 and 6. So they’re still apt. Creationist misconceptions (crocoducks and dogs with wings) seem to run along the lines of thinking a roll of 5 and a roll of 12,356,980 is equally likely.

Right, I should have stated that more clearly. I was thinking of the whole sequence, one roll after another, for 100 times, with 6 possibilities for each roll. A single die roll (6 possible outcomes) would be analogous to a single base pair mutation (3 possible outcomes that are actual CHANGES).

Actually (maybe I’m getting tangled up in the analogy), a single die roll is far, far more likely to produce a change from the previous roll (5 out of 6) than a mutation in a base pair: IF it changes it could be 1 of 3 outcomes BUT any change at all is VERY RARE. The chance of any particular base pair mutating in one copying event is very tiny, right? Thus a single base pair is very strongly (but not absolutely) constrained to look exactly like its immediate “ancestor.” And a whole genome is normally constrained to look very much (but not exactly) like its ancestor–just as a chessboard must look very much (but not exactly) like it did before the last move.

And hey, a single die will most likely look different after a roll. A single chess piece will most likely be in exactly the same position it was in last turn.

Dice have no memory. A chessboard and DNA molecules do.

Actually (maybe I’m getting tangled up in the analogy),

There’s a limit to how far analogy can take you, yes.

a single die roll is far, far more likely to produce a change from the previous roll (5 out of 6) than a mutation in a base pair: IF it changes it could be 1 of 3 outcomes BUT any change at all is VERY RARE.

Correct, but I will rephrase this and add a bit of information.

A point mutation, in which the identity of a given base pair at a given defined locus is changed, is usually a relatively rare event.

What is the frequency at which a given base pair will experience a point mutation? That depends on many factors, and can be measured either per unit time or per replication event (in humans, when talking about well-characterized mutations that have known medical impact, “new mutations per live births” is often used, which is implicitly similar to “per replication event”).

(Let me clarify why, in medicine, we may use the redundant term “new mutation” - because, and technically this is a poor usage but it is ingrained, we may refer to a medically significant allele as a “mutation, even if we know perfectly well that it was inherited from a parent and that the actual mutation event may have been generations ago. Thus, if it really wasn’t there in the parents, if it really is an example of mutation, we say “new mutation”.)

In normal cells with intact DNA repair mechanism, mutation frequency per base pair per time or replication is indeed a seemingly low number. This source gives an estimate of one mutation per 50 million bases added to the copy strand (which is the same thing as saying one per 50M base pairs in the template strand), but it’s talking about ALL mutations. This is obviously a rough but useful estimate. http://users.rcn.com/jkimball.ma.ul[…]tations.html

Let’s pretend we’re looking at one base pair locus. If there is no mutation at that locus, the die, so to speak, isn’t even rolled. If there is a point mutation that creates a change to another base, then there are three possible outcomes, although they don’t occur with quite equal frequency. One of them is actually much, much more likely than the other two. http://en.wikipedia.org/wiki/Point_mutation.

I guess you can think either of a die which comes up “no change” approximately 49,999,999/50,000,000 times, or a die that is only cast, on average, once every 50 million times a base pair is replicated.

Although these types of mutations are quite rare in terms of frequency per locus per unit time or replication event, every time a human cell divides, 6 billion base pairs have to be replicated (the genome is “3 billion base pairs”, but humans are diploid). So you can expect about “120 (new) mutations per human being”. If you could count the number of such mutations in every human and graph the frequency of various totals, you can take it to the bank that you would see a normal distribution, presumably centered around something like 120. (Because it’s the result of a lot of multinomial distributions, but ignore this sentence if it isn’t helpful.) I don’t know what the standard deviation would be, although below I will talk as if I have some idea, to clarify something.

The chance of any particular base pair mutating in one copying event is very tiny, right?

Yes, as discussed above.

Thus a single base pair is very strongly (but not absolutely) constrained to look exactly like its immediate “ancestor.”

No, by “constrained”, I mean things that more or less literally cannot happen at any reasonable frequency, in a single 13-14 billion year old universe with an estimated mere 2.5 x 10^89 elementary particles. Let’s go with the “something like 120 mutations per human zygote” estimate. There are 4^3,000,000,000 possible strings of nucleotides the length of a human genome, and we started with one, and, although in theory we could have any number of mutations, with a normal distribution centered on 120, with some kind of SD that is probably not many orders of magnitude different from the mean. So massively mutated genomes would be vanishingly rare, even with something like 130 million human infants born every year.

AND OF COURSE, you have to have a genome compatible with reasonably normal intrauterine development to become a human infant in the first place. So not only would massively mutated genomes be very, very rare to begin with, they would be vanishingly unlikely to result in viable births, or indeed, even in noticed pregnancies in most cases.

And of course, some ancient, highly conserved parts of the genome really can’t be fucked with all that much, because they code for genes that are essential for development. Are they immutable? Of course not, and it is changes in these regions that may drive the extreme morphologic diversity we see in life on earth (as well as cancer, but that’s another story). But they have to be the right kind of changes, in the right environment, and that happens very, very seldom. Furthermore, when this kind of stuff does happen, it still can seem very “gradual” by human standards. The apparent selection for larger brains over much of human history (*which is no probably longer going on, and I state that as a neutral educated opinion, not a commentary on Bristol Palin’s $262,000 income for running an “abstinence promoting ‘non-profit’”*) resulted in extremely rapid morphologic change in our lineage. But not from the day to day perspective of humans as it was going on. It’s unlikely that even the very elderly, if there were any, noticed the trend over their lifetimes.

And a whole genome is normally constrained to look very much (but not exactly) like its ancestor–just as a chessboard must look very much (but not exactly) like it did before the last move.

Yes, although the chess board is absolutely constrained by the fact that only a few moves are legal, whereas evolutionary constraints all, even the most extreme, have a stochastic flavor.

Thanks, as always, Harold.

It seems I had the gist, but you could put real numbers to it.

But I’m not sure I get this part (IANAS):

ME: Thus a single base pair is very strongly (but not absolutely) constrained to look exactly like its immediate “ancestor.”

HAROLD: No, by “constrained”, I mean things that more or less literally cannot happen at any reasonable frequency, in a single 13-14 billion year old universe with an estimated mere 2.5 x 10^89 elementary particles.

Why wouldn’t any particular base pair have the 1:50,000,000 chance of mutating in any particular replication?

eric said:

Actually, the result of a die roll is heavily constrained by its physical structure to be between 1 and 6. So they’re still apt. Creationist misconceptions (crocoducks and dogs with wings) seem to run along the lines of thinking a roll of 5 and a roll of 12,356,980 is equally likely.

Exactly the point I’ve often made: it’s a fallacy to assume that random actions always have random outcomes, because the outcomes are often “constrained” by physical properties. Another example is snowflakes - or more accurately, snow crystals. They form by a random accumulation of water molecules. But because of the physical properties of the water molecule, the crystal is “constrained” to form these six-sided shapes. A further example is sand on a beach. The ocean randomly pulverizes and dissolves rocks. But because the constituent minerals vary in abundance and hardness, the beach is often composed mostly of the very hard and abundant mineral, quartz. In fact, the latter is actually a good analogy for evolution: random destruction of rocks (mutation) followed by selective filtering by natural resistance to that destruction (natural selection).

Let’s not forget that mutations come in many flavors, not just point mutations. We have indels, tandem repeats, gene duplication, whole-genome duplication, horizontal gene transfer, positional effects, nondisjunction, changes in regulatory DNA, etc. About 40% of the human genome is made up of repetitive sequences - I’d love to know what “information” is contained therein.

Just Bob -

Here’s what I mean by constrained - a modern cat giving birth to a modern dog. Could you start with a cat genome on paper and identify massive numbers of mutations that would make it a dog genome? Of course, there would be many ways on paper, but as we noted, each one of the mutations needed has only a one in tens of millions chance of happening in one replication event. The chances of them all happening sufficiently perfectly at once (in both sperm and ovum contributed by each of two cats who happen to mate) is so massively small as to be incomprehensible. Even though there are vast numbers of DNA sequences that make up viable dog genomes, this is still true. And then the dog would have to be able to develop to birth in a cat uterine environment, which is probably impossible (although maybe not, and that’s more or less irrelevant).

Under the theory of evolution, this type of thing is impossible, because of the constraints imposed by former events. Sounds trivial? Remember, under ID/creationism, there are no such constraints. Fully formed modern organisms can magically appear from thin air.

John S. -

Exactly the point I’ve often made: it’s a fallacy to assume that random actions always have random outcomes, because the outcomes are often “constrained” by physical properties.

Semantic quibble here, but an important one.

“Constrained” and “random” are in no way whatsoever contradictory with one another, unless the constraint is purely deterministic from a human perspective.

The only people who have ever mis-defined “random” to mean “no constraints” are creationists, and when they do this, they are projecting the claims of creationism onto science.

If I tell a group of 5 guests to a dinner party “I’m going to assign your seats randomly”, NOBODY thinks it means “Some of you are going to turn into crocoducks”. EVERYBODY understands that there are a constrained number of orders in which 5 dinner guests can be arranged (5!, or 120), and that I am going to use some method that is non-deterministic, by human standards, such as drawing names from a hat, to generate the seating order.

The concept of a constrained set of outcomes, the probability* or frequency* of each being understandable, but prediction of exactly which will occur in each trial being impossible from a human perspective, has been understood for centuries.

http://en.wikipedia.org/wiki/Cardano

As far as I know, post-modern creationism, with its total reliance on verbal dishonesty and flawed logic, is the first school of thought to ever outright deny the intuitive and easily demonstrated concept of an outcome that is both constrained, yet also not perfectly predictable.

*There is a philosophical/mathematical debate about whether or not probability and expected frequency are always synonymous, which can be skipped here. In the context of things like mutations or dice rolls, they clearly are.

Oops, I don’t want to be accused of creating a straw man.

As far as I know, post-modern creationism, with its total reliance on verbal dishonesty and flawed logic,

So far, so good.

is the first school of thought to ever outright deny the intuitive and easily demonstrated concept of an outcome that is both constrained, yet also not perfectly predictable

This is defensible, but is not true of all creationists all the time, so I retract it.

What they definitively do is -

1) Misuse the term “random”.

2) Create straw man versions of the theory of evolution that imply, falsely, that the TOE is at odds with probability theory, when that is the opposite of the truth. And in doing so, they usually project their own belief in magic “poof” events onto the TOE.

Harold wrote:

“The only people who have ever mis-defined “random” to mean “no constraints” are creationists, and when they do this, they are projecting the claims of creationism onto science.”

What an great observation. This is why trying to argue with a creationist can be so frustrating. They are almost always arguing AGAINST their own position. They really don’t understand evolution at all. They generally have no clue how evolution is supposed to work. All they have is a distorted characterization of their own position. That’s all they can ever seem to understand. Then they argue vehemently that that can’t happen! You can’t agree with them because then they claim victory and you can’t disagree with them without reinforcing their misconceptions.

Of course they could easily do a little research, learn a little biology and study the actual theory of evolution. But of course when they do that they realize that they have no genuine arguments left and that any reasonable person can see that there is absolutely no reason to reject the actual theory of evolution.

harold said:

Just Bob -

Here’s what I mean by constrained - a modern cat giving birth to a modern dog.

Cool. You were talking about a particular–dare I say it, specified–massive set of mutations all at once, while I was thinking about a single point mutation. I know I’m stretching this way too far, but I was thinking of the likelihood of a first move of E2 to E3, while you were pointing out the impossibility of castling queenside on the first move.

The cat-to-dog in one generation scenario would be called a saltation, correct? Perhaps we need a more powerful and “impossible” term than ‘saltation’ for the kind of leap that creationists think evolution would make possible (but which would be ‘normal’ for magic creation).

To return to an earlier point about the random nature of mutations making long-term predictions impossible, I was thinking along these lines: If aliens investigated earthly life, say 350 million years ago, and found something that we might call the “first reptile,” given access to both the phenotype and genotype–the lizard and his DNA–they could not possibly have predicted that his descendants would include pteranodons, snapping turtles, albertosaurus, blindworms, triceratops, gavials, seismosaurus, macaws, plesiosaurus, horny toads, rattlesnakes, and hummingbirds. They certainly could confidently predict that any descendants might evolve to survive changing conditions and exploit available niches.

But could they predict with confidence that the ‘first reptile’ would leave any descendants at all? Given the history of major extinctions up to that time, wasn’t there a fair chance that the ‘first reptile’ would be the last?

DS -

Yes, this is exactly what you encounter when you deal with the semi-sincere education-deprived rank and file.

I say semi-sincere, since disdain for the intellectual achievements of others cannot really be justified at any education level. But in the case of the rank and file, the ignorance often has an involuntary character - for some combination of intense social pressure, economics, or academic ability, actually learning about the TOE may have been denied to them.

They’re told by their leaders “God did magic”.

Every other position is taken as “The same magic, but no God or a different God”.

They instinctively use the same defense mechanism against science that they would against the teachings of a different creationist sect or faith (the latter being what they perceive as a greater threat). They immediately construct an absurd straw man magical version of the other person’s claims (not noticing that in doing so they project their own beliefs), and then invoke the argument from incredulity against the straw man. “LOL, there are no crocoducks, 747’s don’t form in junkyards, LOL evilutionists are funny”.

This is nearly always the style of the creationist comments that come the closest to being honest.

They assume that you assume the magic, but are denying that it was done by their particular god.

The message “No magic necessary, gods have nothing to do with this particular issue” is heavily blocked by defenses and assumptions.

In order to prevent that message from getting through, the leaders rush in with their various more “sophisticated” semantic games, false accusations/demonization, science-y sounding word salad, etc.

Just Bob -

Great discussion.

Cool. You were talking about a particular–dare I say it, specified–massive set of mutations all at once, while I was thinking about a single point mutation. I know I’m stretching this way too far, but I was thinking of the likelihood of a first move of E2 to E3, while you were pointing out the impossibility of castling queenside on the first move.

Exactly.

The cat-to-dog in one generation scenario would be called a saltation, correct? Perhaps we need a more powerful and “impossible” term than ‘saltation’ for the kind of leap that creationists think evolution would make possible (but which would be ‘normal’ for magic creation).

Your first sentence here is incorrect, but your second sentence is correct.

“Saltatory” is a highly imprecise but useful word that people of good faith who make an effort to understand one another have used to describe the way in which human-recognized big morphologic variations or radiations appear in the fossil record, with respect to time.

The entire “gradual versus saltatory” dialog is highly subjective but of value. What “saltationists” like Steven J. Gould were pointing out was that there is no reason to assume that past morphologic evolution, as observed by humans via the fossil record, should have the subjective appearance of being smoothly distributed across time.

The word doesn’t necessarily refer to individual events, but if it did, the example of a “saltation” would be a possible and viable but unusual cat, who had high reproductive success, so that the novel allele or allele combinations that made that cat unusual quickly increased in frequency in the population.

For a cat giving birth to a dog, the word “impossible” is ideal.

To return to an earlier point about the random nature of mutations making long-term predictions impossible, I was thinking along these lines: If aliens investigated earthly life, say 350 million years ago, and found something that we might call the “first reptile,” given access to both the phenotype and genotype–the lizard and his DNA–they could not possibly have predicted that his descendants would include pteranodons, snapping turtles, albertosaurus, blindworms, triceratops, gavials, seismosaurus, macaws, plesiosaurus, horny toads, rattlesnakes, and hummingbirds.

Indeed not, although they might have made many less specific predictions correctly, if they had knowledge of terrestrial biological evolution (which would have been going on for a long, long time by then). I will also note that “the first reptile” would be highly similar to “the closest relatives of the first reptile”, to the point that distinguishing exactly who was and was not a “true reptile” might be arbitrary.

They certainly could confidently predict that any descendants might evolve to survive changing conditions and exploit available niches.

Yes.

But could they predict with confidence that the ‘first reptile’ would leave any descendants at all? Given the history of major extinctions up to that time, wasn’t there a fair chance that the ‘first reptile’ would be the last?

All lineages on earth could go extinct at any time given the right conditions, but the “first reptile” was probably a member of a “breeding population of individuals who were all quite similar to the first reptile”. Reptile traits still had to be selected for in some way, but the “first” of a lineage isn’t necessarily isolated. We get that impression by observing the reproductive isolation of many modern species, but take a look at this (someone posted a much nicer version a while ago but I couldn’t find it). http://www.origin-of-mitochondria.n[…]hment_id=153

Thanks again, Harold. And again I was being a little too colloquial rather than precise. By ‘first reptile’ I meant that species (assuming one could be identified as separate from other not-quite-reptiles), much as one might refer to ‘the gray wolf.’

Correct me if I’m wrong, but I believe saltations are *relatively* common (1 out of many thousands) among highly inbred domestic breeds, like purebred cats. And their survival and reproductive success is probably only because the new and maybe weird morphology (hairlessness, crinkly fur, extra toes, etc.) is an interesting novelty to us. So we protect them and breed for the new ‘cool’ trait. In the wild they would likely have frozen, starved, or become someone’s lunch before ever leaving progeny.

Correct me if I’m wrong, but I believe saltations are *relatively* common (1 out of many thousands) among highly inbred domestic breeds, like purebred cats. And their survival and reproductive success is probably only because the new and maybe weird morphology (hairlessness, crinkly fur, extra toes, etc.) is an interesting novelty to us. So we protect them and breed for the new ‘cool’ trait. In the wild they would likely have frozen, starved, or become someone’s lunch before ever leaving progeny

Well, with caution about the use of the word “saltation” this is true. A couple of points.

1) We often breed domestic animals to relatively close relatives. This process creates far more homozygotes at various loci, than self-directed breeding in a typical wild population. Some of the alleles in question may be recent, others may have been in the population for a long time; just not visible because homozygotes are more rare in wild populations (except if there are population bottlenecks).

2) Some domestic traits are clearly likely to be selected against in the wild, but many others are probably just neutral. And since selection is basically about reproductive rate, even some seemingly maladaptive traits can be highly persistent. White colored cats have some disadvantages, but white feral cats are not all that uncommon, nevertheless. However, what would usually happen in the wild is that even if some cool-looking phenotype (by human standards) occurred as a result of an unusual combination of alleles, feral cats wouldn’t care. Even if it weren’t selected against, the individual would probably just breed with ordinary looking mates anyway, and the trait would be diluted back out. What humans provide is both protection, but also, extremely strong selective pressure. We decide who breeds and how much they breed, and with whom they breed (in the case of purebred cats, that is).

I wonder if whiteness in feral cats would persist long in a truly “wild” environment, you know, with wolves, mountain lions, and bobcats around. Then they would also have to hunt and catch truly wild prey, rather than subsisting on human garbage. Presumably effective camouflage would be useful for both purposes.

harold said:

Just Bob -

Here’s what I mean by constrained - a modern cat giving birth to a modern dog. Could you start with a cat genome on paper and identify massive numbers of mutations that would make it a dog genome? Of course, there would be many ways on paper, but as we noted, each one of the mutations needed has only a one in tens of millions chance of happening in one replication event. The chances of them all happening sufficiently perfectly at once (in both sperm and ovum contributed by each of two cats who happen to mate) is so massively small as to be incomprehensible. Even though there are vast numbers of DNA sequences that make up viable dog genomes, this is still true. And then the dog would have to be able to develop to birth in a cat uterine environment, which is probably impossible (although maybe not, and that’s more or less irrelevant).

Under the theory of evolution, this type of thing is impossible, because of the constraints imposed by former events. Sounds trivial? Remember, under ID/creationism, there are no such constraints. Fully formed modern organisms can magically appear from thin air.

And of course harold, the biggest constraint is prior phylogenetic history. In the hypothetical example you cite, that is impossible simply because the feline (cat) and canid (dog) lineages diverged from each other tens of millions of years ago.

As an aside, it is absolutely ludicrous for Dishonesty Institute mendacious intellectual pornographer Stephen Meyer to assert - as he has done in his “Signature in the Cell” - that one could use the fossil record to test for “deviations” from some kind of “perfect” Design. Why? He has ignored the prior phylogenetic histories of these fossils.

Regards,

John

P. S. Am in complete agreement with your comments to John S.

harold said: The entire “gradual versus saltatory” dialog is highly subjective but of value. What “saltationists” like Steven J. Gould were pointing out was that there is no reason to assume that past morphologic evolution, as observed by humans via the fossil record, should have the subjective appearance of being smoothly distributed across time.

Just for the record, Gould was not a saltationist, which his critics - both scientists AND creationists - have accused him of. He knew clearly the different between true saltation - like the “hopeful monster” coined by geneticist Richard Goldschmidt - and what he and his long-time friend and colleague Niles Eldredge had proposed with respect to punctuated equilibrium, which, I might add, drew extensively on Ernst Mayr’s thinking with respect to allopatric speciation.

The fossil record is illusory in the sense that we often miss - for geological reasons - extremely fine microstratigraphic detail, that it is easy to assume that new species arose via some saltationist process. However, that is certainly not the case, as much as creationists might wish it was.

John Kwok -

Thank you for the clarification with regard to PE/saltation. As I’ve mentioned before, I’m a fan of Steven J. Gould’s writings. I don’t want to get side-tracked here, but my recollection is that there was a rather dogmatic school of “gradualists” circa the mid-twentieth century who were mildly hostile to Gould. Of course, all of this was a dialog within science, and no-one involved denied biological evolution or its basic mechanisms.

And of course harold, the biggest constraint is prior phylogenetic history. In the hypothetical example you cite, that is impossible simply because the feline (cat) and canid (dog) lineages diverged from each other tens of millions of years ago

There is certainly no disagreement between on this point; I phrased the explanation in terms of the molecular genetic differences, but those differences are there because of the divergence.

harold said:

John Kwok -

Thank you for the clarification with regard to PE/saltation. As I’ve mentioned before, I’m a fan of Steven J. Gould’s writings. I don’t want to get side-tracked here, but my recollection is that there was a rather dogmatic school of “gradualists” circa the mid-twentieth century who were mildly hostile to Gould. Of course, all of this was a dialog within science, and no-one involved denied biological evolution or its basic mechanisms.

And of course harold, the biggest constraint is prior phylogenetic history. In the hypothetical example you cite, that is impossible simply because the feline (cat) and canid (dog) lineages diverged from each other tens of millions of years ago

There is certainly no disagreement between on this point; I phrased the explanation in terms of the molecular genetic differences, but those differences are there because of the divergence.

I am in complete agreement, harold, except to note that it is Stephen Jay Gould, not what you have written. Please note that for future reference.

About this Entry

This page contains a single entry by Jack Scanlan published on April 5, 2011 8:23 AM.

Supreme Court upholds tax credits for religious schools was the previous entry in this blog.

Complete list of evidence against biological evolution is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.

Categories

Archives

Author Archives

Powered by Movable Type 4.38

Site Meter