Looking for sources on the information argument

As I said above, the problem is probably not that blind evolution cannot favor a certain body shape but that it cannot achieve optimal physics when many other suboptimal design are possible. Here is a relevant passage in Helder piece:
To find out why the sponge skeleton has such interesting properties, the Fernandes team set up four different grid designs. Option D was a 2-dimensional square divided into quarters. Option C added cross bracing to each quarter so that each square was divided into four equal triangles. Option B added the cross bracing to only two squares (numbers one and four) of the design, while the other two squares were open. Option A, the most elaborate design, mimics the Venus Flower Basket design. It had double strands of cross bracing in squares one and four, while the others were open. When these designs were tested for load-bearing capacity, design A (like the Venus’ Flower Basket) displayed the best resistance to buckling stress. Having demonstrated the superior resilience of design A, the scientists wondered if they could improve on this glass sponge design. They conducted seven separate optimization experiments. Their studies demonstrated that “the sponge-inspired design provides a superior mechanism for withstanding loads.” 13

As a result of their studies the scientific team concluded that

> The results presented here therefore demonstrate that, by intelligently allocating material within a square lattice, it is possible to produce structure with optimal buckling resistance.

Here is a paper that describes biological functional information

in an academic way. Szostak won a Nobel prize in chemistry.

Optimization is done with respect to a function, and the value of a function depends on the environment–or to use the language of biologists, the value of a function depends on the selection pressures on a population.

A sloth moves slowly but a jaguar is extremely fast. You could say that the sloth design is suboptimal with respect to the function of locomotion. However, the sloth has a variety of optimizations for the niche it occupies as a climbing mammal. The jaguar simply doesn’t need all of the sloth’s optimizations, and the sloth doesn’t need all of the jaguar’s optimizations.

Humans are in-between with respect to locomotion speed. Running on 2 legs makes us slower than a jaguar in running short distances, but in compensation we humans get the advantage of better downrange vision from a standing position. We also have better endurance, which is an optimization for certain kinds of hunting.

In short, optimization is a very nuanced issue. Unfortunately, your generalization does not address the topic in a way I find helpful. But I appreciate that you tried. :slight_smile:



The Fernandes paper cited by Helder and quoted indirectly by you, Gilbert, did not touch upon the subject of evolutionary mechanisms. The reference to “intelligently allocating material” clearly refers to how humans can learn from designs present in nature; the paper does not pretend to deal with the question of the origins of the sponge’s design.

The question of whether the design can be reached through natural processes hinges on whether there is a functional gradient on the path to the optimal structure. A question for you, Gilbert: Did Helder even attempt to show that no such gradient exists? Or did she merely assume that it does not? As far as I can tell, she just assumed its non-existence.

So her logic is this, as far as I can tell:

Fact: The structural design of Euplectella aspergillum is highly optimized to deal with buckling strain.
Assumption: There is no functional gradient for evolutionary processes to produce such a design.
Conclusion: Therefore, evolution has no explanation for the structural design of Euplectella aspergillum

I find Helder’s assumption unconvincing for many reasons, among them that we see functional gradients everywhere we look in biology.

I also note this statement Helder makes:

their [i.e., sponges’] fossil location at the bottom of Cambrian rocks does not record early evolution but rather that they were among the first creatures engulfed by sediments from the global flood.

This is patently absurd, in my opinion. There is tremendous biological diversity at the bottom of oceans, so a massive sediment dump on ocean floors would yield a wide variety of fossilized fauna. Also, there is a tremendous diversity in geological structures that would be inundated at the start of a global, sediment-carrying flood (riverbeds, lake beds, etc.) and their associated fauna.

That Helder makes such an absurd statement greatly diminishes her scientific credibility in my eyes. But do you see it differently, Gilbert?



It is important to make distinction here. Imagine that many overall solutions/designs exist for a given biological problem. Let’s call them solutions A, B, C, …Z. Among them, only one is optimal (let’s say it is C), the other being suboptimal but still compatible with life. If they are many solutions, let’s pretend that natural selection will have the capacity to find one. Let’s imagine that it’s H. In that case, it is true that NS probably has the power to optimise solution H and turn it into solution H’. But in this scenario, solution H’ would still be suboptimal by comparison with solution C. The lesson here is that it is indeed really surprising to observe biological structures exhibiting optimal physics in nature.

Not as evident as you imagine, as shown by the chloroquine example in malaria.

There is no reason to think that there are functional gradients connecting different overall designs/solutions (in my illustration above, there is no reason to think there is a functional gradient connecting H or H’ with C).

Why is this relevant? Credential mongering much?

Multi-drug resistant Mycoplasma tuberculosis: hahahahahahaha


On the one hand we have here a statement that (a) “Chance processes like natural selection” (b) “show zero potential” for (c) “supplying the information needed” for (d) “such a profound connection between structure and function”. Point (a) shows the author to be under the delusion that natural selection is a “chance” process. Everyone knows that chance processes, ones that show no tendency to preferentially go in the direction of increased adaptation, could only accidentally result in any adaptation. As is indicated by (b) as well. These processes would not supply any information (c). Those three parts of the statement are perfectly in line with the usual creationist argument that natural selection cannot increase adaptive information. Giltil, in his usual way, points to (d) “such a profound connection” and claims that the statement is only about the actual achievement of optimality.

So let’s presume that Giltil would agree with the statement if it were modified to say that “Directional processes such as natural selection show considerable potential for supplying the information needed for substantial connection between structure and function.” Right? I somehow suspect that he wouldn’t, and if we were betting I would place a side bet against him agreeing with that.

At least he is using a definition of information (the Orgel-Szostak-Haszen concept of “functional information”) which is one of the sensible ways of defining information that is semantically meaningful in this context. So the disagreement is not based on using different definitions of information.


Why can’t it do that? Again you’re leaving out the reason why. You just say that it can’t get to optimal if sub-optimal designs are also possible. But if sub-optimal designs are sub-optimal, then any improvement on sub-optimal designs would have a survival and reproductive advantage?


Evolution did produce chloroquine resistance, so if you consider this an optimal solution to a problem it’s difficult to see how this is supposed to make us think evolution couldn’t produce optimal solutions to problems faced by an organism.

There are multiple mutational pathways to effective chloroquine resistance even within the transporter protein in which chloroquine resistance is known to have evolved(and, again, chloroquine resistance did evolve), so how is that supposed to constitute evidence against there being evolvable pathways to high levels of adaptation, much less evidence against there pathways to an adaptively “optimal” solution to some given physical challenge?

In what way is chloroquine resistance “optimal” by the way, aren’t you supposed to give us reason to think natural evolution can’t produce sufficient information for optimal solutions?

It seems to me there is, and I think the examples of things like hydrodynamics/drag coefficients, and load-bearing capacities are good examples where there must be virtually endless ways to gradually transform any given shape into the globally optimal solution, with substantial gains for selection all the way to the top.

In fact exactly for these kinds of relatively simple physical adaptations like shape, size, density, friction, strength and so on it seems totally obvious to me that these can be manipulated gradually and incrementally with significant pay-offs to selection all the way from the bottom to the top. If there’s an optimal length to some structure, it seems to me one can gradually increase or decrease length in small increments. If there’s an optimal density, you can gradually increase or decrease it. If there’s an optimal curvature, and one degree of curvature is better than another, you can gradually bend or unbend it and thus selection can get you from worst to best. Und so weiter.


I should add that I actually do agree that we don’t have reason to think ALL different adaptive solutions to any given challenge faced by an organism can be reached directly by natural selection, nor that any and all adaptations can be change gradually and adaptively into each other. And we even know of some such cases where distinct, unrelated, and non-optimal solutions to the exact same problem faced by different organisms have evolved.

I recall the example where bacteria and plants appear to have independently evolved entirely different enzymes that catalyze the exact same chemical reaction, and unsurprisingly the bacterial enzyme is catalytically superior to the plant enzyme(which makes sense given the vastly superior population sizes reachable by bacteria, and that bacteria have existed on earth for billions of years longer than plants):


NS and variation, Gil. IDcreationists go to amazing lengths to deny the obvious. It would seem that you completely disagree with the author of the paper you cited that NS is a “chance process,” correct? If so, why would you cite a paper with such a blatant falsehood?

It’s evident to those of us who read the primary literature in the field and don’t believe Behe’s representations. I eagerly await your detailed review of these data; this paper (and many more) demonstrates that Behe is not working in reality. Have you read it?

The graph in Figure 1C, shown right there in the preview above, tells you that there is a functional gradient for chloroquine resistance. It’s a great example, just not what you thought it was from taking Behe’s misrepresentations as reality.


Which is a model consistent with sometimes finding optimal solutions, and sometimes finding locally optimized, globally suboptimal solutions. Which is what we find. We should even expect some globally optimal solutions in systems with multiple local maxima, both by chance from starting on the best peak in the first place and by valley crossing from suboptimal peaks.

To say there is a problem you would need an assessment of, across all species, the probability of traits having multiple distinct local maxima, the proportion of those traits that are at the global maxima, the disparity of those maxima, and the probability of valley crossing (which can be highly dependent on a variety of factors, including allometric scaling from other traits under selection).

So far what you’ve presented is one organism that may be at a global peak, that may have discretely separated peaks. Meaning even taking you at your word for both, you’d still have said nearly nothing. And, well, I don’t think we’re apt to take you at your word on this.

I struggle to imagine how an example of natural selection doing a thing could be evidence against the power of natural selection doing a thing. Especially when, as @Mercer helpfully noted, it is an example of it doing precisely the thing you suggest it shouldn’t.

I said ‘optimization’, which may or may not result in the global optima. You haven’t presented evidence that there are sufficient examples of globally optimized, multsolution traits in the first place, as I described in the first part of this response. Until you are able to do so, you don’t have an argument.


Your error I think is to imagine that the structures of the glass sponges we are talking about are simple physical adaptations like shape, size, density, friction, strength, etc… when in fact they are built according to complex and optimal architectural principles. Here is a quote from the paper: Our computational results reveal a rich multifaceted role of the skeletal motifs for E. Aspergillus on the flow physics within and beyond its body cavity.
Margaret Helder, the author of the piece discussing the paper then goes on to say:
All the skillful (I would add optimal) design choices displayed by this glass sponge, therefore, have important things to teach our best designers of skyscrapers, bridges, ships, planes and anything that must respond safely to forces imposed by the flow of air or water.

@stlyankeefan There is another important aspect of information entering this discussion.The measures of information I described can also measure the difference between two sources of information. The concept here is Relative Information representing a change rather than addition or subtraction.

Darwin’s finches are a good example. All the various types of finches should contain roughly the same amount of information in their genomes. There are relative changes that make up the differences. The argument that new information cannot be added is like saying that information cannot be changed.

A very simple example of an information distance measure is the Hamming Distance, which is a count of the bit-wise differences in two binary sequences.

@Giltil might want to read this one too, as the first step towards optimal function is changing information.


Thank you. :grinning:
Do you agree the Hazen’s FI can change by RV+NS?
(see my previous comment about relative information)

With regard to Glass Sponges, I think there is a category mistake between sponges having an optimal physical form, and this representing Optimal Functional Information. This certainly has not been established in terms of Hazen’s FI, and I think that might be difficult.

Which brings me to my next question: How are you defining Optimal Functional Information?

My intuition here is the sponge will not maximize functional information. It has multiple objectives to achieve, strength, feeding, growth, reproduction, energy efficiency, etc., and maximizing one is likely to be at the cost of the others. It is likely the sponge is at a point of balance for all these factors, rather than an optimum for any one of them.


Thanks so much! This has all been very helpful. It’s probably way more than I need for my class, but I have a much better understanding of some of the many problems with the ID arguments about information.

I teach Gen Ed Biology to undergrads, mostly liberal arts majors. When they say things like “information requires a designer,” I suspect that haven’t really thought about what information is or how it’s measured. I do have a couple of computer geeks in my class who will appreciate some of the more technical aspects. But for most students, we can start with what is information, how do you measure it, and can evolution/natural selection create it.

Thanks to everyone who responded. :heart:


Yes, you should be OK. I thought it was important to add Relative Information to the discussion. Evolution is change, and Information can change in many ways (form and function). Adding and subtracting from a total is the wrong way to think about it.

1 Like

I think that’s an important point. I’ve already had one conversation about nylonase enzymes with a student who argued that those enzymes are not “new” information, just changes to existing information.

1 Like

Now I’m getting out of my expertise. To my understanding, changes to shape and form, like Darwin’s finches or the optimal shape for a skeleton of a sponge, require mostly regulatory changes. More of this, less of that, but mostly reshaping the existing parts. There is good website demonstrating this, but is uses Flash and may no longer function.


It occurs to me the functional changes, like new proteins, is also a matter of reshaping a protein, evolution should be good at that too, but I don’t know it the analogy holds.

Here is a Java version of Boxcar2D. This one doesn’t have as many features as the original, but it runs much faster - fast enough you could leave it running in the background during a lecture.