Mark: Are Mutations Random?

Dr. Swamidass,

Could you at least agree that SOME mutations aren’t random? If you can’t at least concede this, then it seems to me you have an enormously high burden of proof to bear. Would you really insist that the bacterial flagellum that regenerated a good deal of its parts in a matter of days constitutes a random process that is somehow ordered like white/brown noise? Whatever one thinks of Behe’s work, it seems to me that no one disagrees with him that the flagellum is an incredibly complex biological “machine.” If you blew it up a billion x its size, I’m sure if people today were to repair a replica of its parts, no one would describe this as a randomly ordered process on par with white/brown/pink noise. That seems absolutely absurd to me. I acknowledge that the people that saw this happening attribute it to the power of “natural selection,” and that you’re still in the majority on this. But it seems like any thinking person not inundated with the “random” paradigm in biology would say that the order involved in such a procedure was on a level much much higher than the order in white noise.

And would you say that the three layers of error correction in DNA are to correct errors that don’t conform to a random pattern? That also seems really far fetched. Why not just admit that some mutations aren’t random?

Also, what degree of order does something need to achieve NOT to be considered random anymore? Perhaps if this could be specified, tests (certainly not in this forum!) could be done to see if certain processes constitute random or directed processes.

If a way to decide whether or not something is random is spelled out, perhaps we could move forward on this issue.

Have you been following along?

Essentially all mutations (and just about all processes) are random and ordered at the same time. It is merely a matter of degree. The one exception I can imagine is a boundary case, a Dirac function, which is deterministic with no randomness.

Shapiro is entirely incorrect when he says that mutations are “not random” and then justifies this with evidence order. There is always order to random ones, so this an absurd argument against randomness, because both are usually true at the same time.

There was never a point to concede because I’ve been saying it all along. This is nothing novel, but visible in introductory texts in statistics and probability, and also in the literature too. Remember that @Perry_Marshall just learned that there is order in randomness, including in his field of electrical engineering. He claimed initially that random meant there was no order. I am glad he is catching up.

@Perry_Marshall and Shapiro’s argument appears to amount to arguing that a half full glass actually contains no water, because half the glass is empty. At this point they have a very large bar to clear. It might be better to just walk away.

To be clear, I’m fairly certain the flagellum is not self assembling, so your though experiment would not yield a functional flagellum. It is nearly trivial, however, to demonstrate:

  1. The flagellum is assembled in a process that includes randomness and order.

As to it’s function itself, you may be surprised to learn:

  1. The most detailed and effective model of protein dynamics is a Markov state model, a random model. What ever the superficial similarity to a turbine, that is not how a flagellum works.

As for Behe’s IC argument: Which Irreducible Complexity?

I never attributed this to the power of natural selection. Think again. Nor did I say it was due to white noise.

Did you know we use randomness to model free choice of intelligence too? Randomness could be precisely where God directions appear in our equations. It is merely the part of the distribution we cannot predict.

For example, I can model my wife’s preferences for dinner tonight as a random variable. She is pregnant, so her choices are far from deterministic, not even normally she is a grand unknown. There is perhaps 50% chance we will stay in. Perhaps 50% we will eat out. So my wife is a coin flip. Not that we have adequately modeled my wife with a random variable, it seems we have demonstrates she has no mind and is purposeless.

Or so does the logic of your objection goes…

As I said deterministic processes onstensibly have no randomness. In practice there is always a gap of randomness between deterministic models and reality. This means reality is full of randomness.

@mark, are you falling prey to the idol of the marketplace (see Bacon)?

We do have several classes of mutations and are discovering new ones all the time. Your objection is ungrounded.



First of all, I’m not arguing for irreducible complexity. As with most things, I find Marshall’s explanation of IC in his book to be the most plausible explanation that I’ve seen.

Secondly, I was referring to an actual experiment where the genes of a flagellum were deleted. See the exchange between Noble and Shuker on this here-

Thirdly, maybe you’re right. Noble STILL uses the term stochastic to describe the process, but he says “harnessed stochasticity” to imply almost a directed randomness, which does seem to be close to your definition.

Nevertheless, Noble denies randomness when it comes to LOCATION of mutations: “We develop our case in stages: first, to show how multicellular organisms use targeted evolution of their cells to respond to environmental challenge; second, to show how populations of microorganisms achieve similar targeted responses; third, to show how epigenetic inheritance occurs in multicellular organisms with separate germ-lines; fourth; to show how the evolution of behavior can use similar processes that have developed agency in their evolution. In all these cases, variation is not random with respect to genome location and/or organism functionality.”

1 Like

It’s pretty much basic chemical thermodynamics that ensures DNA replication and repair are prone to error. No site is really immune.

Edit: I hope the description of Noble’s view is a bit overstated.


Would you agree with Noble that mutations are not random with respect to location and functionality?

I take this to mean something like slot machine 5 realizes that it will be demolished if doesn’t get 3 things in a row in the next hr, so it starts to play its own slots constantly for the next hr till it gets what it needed.

If you are referencing a specific study, it is important to put the actual study. A few imprecise sentences are never enough to be able make sense of anything in biology. Here is the study:

The ability to adapt to changes in the function of gene regulators, as opposed to structural genes, is a crucial aspect of evolutionary change. Taylor et al. mutated a central regulator for the formation of flagella in the bacterium Pseudomonas fluorescens . They then put the mutated flagella-free bacteria under strong selection pressure to regain mobility. The mutated bacteria regained the lost flagella, and motility, within 4 days. Two stereotypical mutations diverted an evolutionarily related regulator that normally controls nitrogen uptake to control flagella biosynthesis. The mutations increased the levels of the co-opted regulator, then altered its specificity for the flagella pathway.

To be clear, the genes for flagella were not delete. A key regulator (a protein that controls production of the flagellum RNA from DNA) was deleted, and this protein is not part of the flagellum itself. It also was very easy to recover this. Given that most human-chimp differences are changes in regulation, this is an example of the sort of evidence that makes the molecular basis of human evolution seem very easy.

It is also an example of something called “exaptation” where parts of a cell are easily repurposed for a new use. This means that evolution often does not have to start from scratch, but can just compose things together. A very interesting study.

It has nothing to do with whether or not mutations are random or not.

Yes, it is true that he is playing word games. Conceptually he is talking about the same thing, but he does so with an abuse of terminology. Stochastic = random. Directed randomness just says there is an order to the randomness that is beneficial, which is obviously true.

This is where he becomes incoherent and is abusing terminology. This is demonstrably false. His imprecision in language is creating immense confusion. As I said several times, there is both order and randomness in mutations. It is true that some of that order places mutations more frequently in some locations, and can be biased towards benefit. This order, however, does not meant they are not random.

Both can be true at the same time, so it does not follow that “variation is not random with respect to location or function.” Such a statement is false.

Both can be true. Mutations are both ordered with respect to location and function, and also random with respect to them. We can even quantify how much. That is exactly what we do. Both are true at the same time. In fact, we can even study how the order of mutations evolves to be more beneficial over time (the evolution of evolvability). The fact there is order does not mean there is no randomness. There are patterns in the randomness that correlate with function and location, but there still remains a great deal of randomness.

Moreover, this is not my idiosyncratic definitions. This is, actually, part of information theory. We can actually quantify how much randomness (entropy!) there is in a distribution. In no distribution to which Nobel has referred that there is no randomness. Period. Therefore he is false if he “denies randomness with respect to location.”

That being said, you are only reporting his views. It is also possible you are misquoting him. Though, given what I’ve seen from Shapiro, @Perry_Marshall, a EES, it is more likely he is abusing the terminology for rhetorical effect, to the point of miscommunication science and creating an avoidable conflict.

I disagree, and have explained why. However…

I have no problem with this. There is still randomness, but the shape of the distribution is such that it improves likelihood of meeting a functional goal. There is still randomness, but there is a pattern in the randomness. If Nobel says there is no randomness, Nobel is unequivocally false.

1 Like

@Mark, I hope this has clarified your questions here. Thanks for engaging.

Before you go, can you tell us about yourself? You are drawn to EES and @Perry_Marshall’s work in Evolution 2.0. What more can you tell us?

1 Like

In the accompanied text, an author characterized this as follows: “Two stereotypical mutations diverted an evolutionarily related regulator that normally controls nitrogen uptake to control flagella biosynthesis.” In other words, a sort of case observed and well understood. I used to work with bacterial two-component regulatory systems as a post-doc, in particular, one related to the regulation of nitrogen uptake. Interesting to see NtrC being exapted in this manner. We know that many transcriptional factors will bind with differing affinities to sites outside their ‘normal’ scheme (promiscuous activation).

From the abstract:

Step 1 mutations increase intracellular levels of phosphorylated NtrC, a distant homolog of FleQ, which begins to commandeer control of the fleQ regulon at the cost of disrupting nitrogen uptake and assimilation. Step 2 is a switch-of-function mutation that redirects NtrC away from nitrogen uptake and toward its novel function as a flagellar regulator.

There are numerous ways to increase expression of NrtC – e.g. gene duplication, modified promoter, modified transcriptional mechanism – and those sorts of changes happen in bacteria at a rate of about 1E-6 to perhaps 1E-7 per cell per generation. Point mutations arise at about 1E-10 per nucleotide per generation. For a few milliliters of bacteria culture at medium density, it shouldn’t take to long to hit on that set of changes to restore some degree of fleQ transcription. One needn’t propose targeted mutations or some sort of cell intelligence to account for the observations.


I suspect that part of the miscommunication here, is that some people thing of “random” as being opposed to “intelligent”. But they need not be. The owners of Las Vegas casinos are using randomness intelligently, and earning a pretty good income from that. You mention Noble’s term “harnessed stochasticity”. And that would be good expression for describing how casinos use randomness. And yes, biology can reasonably be said to also use “harnessed stochasticity”.

1 Like


I was noticing how much you delve in the issues of randomness and order.

I find that i am not quite sure what your point is on randomness.

Is it possible that my personal mantra- - “God-Guided-Evolution” - - is a stronger statement than even you are willing to assert?

Only a few months ago… you were willing to agree with my mantra as a theological statement (even though the science would never be able to tell us this).

Are you still of that mindset ?

1 Like

I’ve always resisted the term “guided”, preferring “governed”. Though I think “guided” is perfectly valid as a theological position.

Also, remember…

1 Like

Shucks, @swamidass, this is the first i’ve heard that!

But i like it! Consider my mantra changed!: “God-Governed-Evolution”… and it is still E.G.G. !!!

1 Like

I’d say rather Providentially Governed Evolution. That is the stronger language, in my opinion. We are affirming the doctrine of providence, not guidance or governance, after all.


Yikes! Nope… you lost me on that last swerve! I find the term “providential” as way too confusing and vague.

I’ll stick with God-Governed!

1 Like

About myself? Well, I just finished my ThM and a year before that, completed my MDiv at St Vladimir’s Orthodox Theological Seminary. I wrote my ThM thesis on Maximus the Confessor’s theology of the fall (which needs to be considerably revised now that Fr Maximos Constas has translated loads of new Maximus material) in conversation with modern science.

I defended a non-historical Adam and used Venema to make my points about population genetics. Only after finishing my thesis did I have a chance to really look at what has happened to the pop. gen. debate since “Adam and the Genome.” Although I don’t plan on changing my position regarding Adam since I hold the position more for historical-critical reasons than scientific ones, the “genealogical Adam” approach is something that might help Orthodox Christians not ready to swallow the initially troubling things a non-historical Adam would mean for the inspiration of scripture, etc. So I value the work done here.

While writing my thesis, I stumbled onto Perry Marshall’s book and the EES, and things started to make sense for me. Reading through the debates between Venema and Behe/other IDers, I always felt like there was something missing that neither side was talking about, and I feel like the Third Way/EES is that “missing” thing. I would also include Michael Denton in the EES even though he considers himself an IDer and really appreciate everything that I’ve read from him (same with Richard Sternberg, poor guy).

I am currently writing an article I hope to get published in the St Vladimir’s Theological Quarterly about how Maximus the Confessor’s concept of the Logoi and essence and activities of God help to make sense of the EES, and the growing number of scientists who endorse some sort of Platonism (Andreas Wagner and Simon Conway Morris for instance).

I am really trying to understand the science before I send anything to the press so being on this forum really helps me out. If any PhD would be willing to take a look at my article to check the science out before I send it out, that would be great appreciated.

That’s my story I guess!

Thanks for asking.


When you are ready, I’m sure someone will be up for helping.

There is a lot of good in @Perry_Marshall’s work. The problem with EES is the pseudohistory of science, and that it treats old concepts as if they are new. If you can sort through that, ignore their polemics, a lot of it is grounded in good science.

It attracts a range of people. Wagner and Jablonka seem generally reasonable. And they tend to avoid polemics and hyperbole. Not a big fan of Shapiro’s arguments. To me it’s like he’s taking a bunch of different edge cases in biology, stringing them together and projecting that this formulation is actually the ‘norm’ for biological systems. It helps to note that some people use polemics and buff their facts as a means of promoting their own twist, seeking to create an (artificial) contrast with other work. I think what we’ll find is that an ‘extended evolutionary synthesis’ will be defined some years or decades after the fact and after the dust settles.

1 Like

@Argon, @swamidass,

Would you describe the following process as random? I acknowledge we are NOT talking about mutations. If not random, then what word would you use?

“As previously exemplified by the fidelity of DNA replication, cognitive checkpoint control rather than mechanical precision ensures the reliability of eukaryotic cell division. Consideration of the spindle checkpoint in the M phase of the cell cycle illustrates this argument [149–151]. It is sometimes claimed that chromosome distribution at mitosis is random because there is no way to predict which of the two chromosome copies will end up in a particular daughter cell. But the fact that one and only one chromosome copy goes to each daughter actually makes the process highly nonrandom. If chromosome distribution were truly random, only 50% of cell divisions would produce progeny with a single copy of each duplicated chromosome in each daughter cell. For an organism such as S. cerevisiae, with 16 chromosomes, random chromosome distribution to daughter cells would result in only 1/216 < 1/32,000 divisions producing two progeny cells with equal complete genomes.
The job of the spindle apparatus is to make sure that each daughter cell receives a single copy of each duplicated chromosome. This guarantees that each daughter cell has a complete genome. We now have a fairly good idea of how this checkpoint operates. The mitotic chromosome[…]”

Excerpt From: James A. Shapiro. “Evolution.” iBooks.
This material may be protected by copyright.

Check out this book on the iBooks Store:

Shapiro goes on to say that this takes place with 99 percent accuracy.

Would you say he’s just using the wrong words to describe the process because he’s defining random incorrectly, or…?


How accurate does a process need to be in order to be considered non-random anymore? I know you have been saying it’s a matter of degree, but would it not sound a little silly to say that if something is 75 percent accurate, it’s 25 percent random? Maybe I’m thinking about this in a completely wrong manner.

I’m assuming @Perry_Marshall will respond to you on his blog in a similar but more informed manner.