Skip to content

Yet another “post-Darwinism”

Last updated on 22 Jun 2018

Coverhighresfinal 1 Over the years there have been many books that purport to “radically revise” or “supplant” Darwinian evolutionary biology; they come with predictable regularity. Usually they are of three kinds: something is wrong with natural selection, something is wrong with inheritance, or something is wrong with phylogeny. This book, by geneticist James A. Shapiro, exemplifies all three.

It is, I must admit, chock full of information not usually found in books about evolution. Part I has way more than I can assess regarding cell signalling, cell death and cellular reproductive strategies. Part II discusses the genetic system of eukaryotes, mostly, and introduces the core notion of the book – “natural genetic engineering”. Part III considers the evolutionary import of molecular genetics and genome sequencing, including some of the usual “post-evolutionary” suspects such as endosymbiosis, lateral transfer, and structural genomics. Part IV is where the “new conceptual basis for evolution” is presented. It is deeply disappointing.

He introduces the notion of natural genetic engineering thus:

A major assertion of many traditional thinkers about evolution and mutation is that living cells cannot make specific, adaptive use of their natural genetic engineering capacities. They make this assertion to protect their view of evolution as the product of random, undirected genome changes. But their position is philosophical, not scientific, nor is it based on empirical observations. This section demonstrates that in a large number of well-documented cases, natural genetic engineering capabilities have been utilized as part of the normal organism life cycle. In many of these cases, utilization involves the integration of different natural genetic engineering processes into a highly targeted and well-regulated series of changes with a clear adaptive benefit. The operation of a tightly regulated sequence of natural genetic engineering events in the adaptive immune system is probably the most elaborate example we have of purposeful genome manipulations. [66]

Immediately the contrast is suspect: “traditional thinkers”. Moreover, the framing – “living cells cannot make specific, adaptive use of their natural genetic engineering capacities” – sets up a strawman. I do not know of anyone in the field who asserts that living cells cannot employ any mechanisms of which they are capable. More on this below.

I shall presume that the science is correct, and the choice of apparent counterexamples to the ruling paradigm (which seems to be far more fluid than many of these books expect. Lateral transfer, endosymbiosis and jumping genes are many “post-Darwinian” ideas that have been easily inserted into the consensus) is illuminating. What is the illumination thus gained?

It is basically this:

Although they may go through many trial-and-error steps, human engineers do not work blindly. They are trying to accomplish defined functional goals. Can such function-oriented capacities be attributed to cells? Is this not the kind of teleological thinking that scientists have been taught to avoid at all costs? The answer to both questions is yes. [123f]

Shapiro is returing to the older style of teleology that might have satisfied an Aristotelian (such as the wonderful D’Arcy Thompson or Agnes Arber in the 1920s). The worry is that he does it in a rather hamfisted and unnecessary manner. Where they engaged with the best philosophical arguments of the day (and let there be no mistake, this is a philosophical claim being made, not a scientific one), Shapiro seems to think he can get away with statements like the one just quoted without any engagement at all.

Engineers do not accomplish defined functional goals. Instead they employ the results of prior experience on the presumption that what worked in the past will work now (and that includes the choice of goals themselves). No engineer has some cognitive “noetic ray” that allows them to see into the future; they are using the tested results of the past. When they make assumptions that fail to hold in the present case, the result is the Tacoma Narrows Bridge. Here engineers used prior art, especially with respect to stress analyses, to build an fantastic suspension bridge. They simply did not do resonance analyses, and this particular spot had winds that set thebridge resonating exactly at the frequency that tore it apart, spectacularly. Now engineers do resonance analyses.

Shapiro’s book is not, despite its subtitle, evolution for the 21st century. It is instead evolution for the 16th century, before Francis Bacon had written:

For the inquisition of Final Causes is barren, and like a virgin consecrated to God produces nothing. [The Advancement of Learning. iii. 5]

Bacon’s dictum long precedes Darwin, but the attempt to find a natural process which delivers design is treated as some kind of novel dscovery, when in fact it is simply unsupportable in the domains Shapiro appeals to. I must tell a personal story: at a workshop on microbial biology and philosophy at Exeter a few years back, I gave a talk on microbial species. Shapiro gave a talk on how bacterial colonies and biofilms were like (in fact, were) intelligent problem solving systems. Now I can accept the metaphor, if properly hedged, but really they exhibit only the sort of post hoc intelligence that all populations do, of responding to their environment in ways that either succeed to maintain the population, or fail to (in which case they remove themselves from the subject of investigation).

But Shapiro wanted to call this just intelligence. No metaphor, no analogy, the real thing. It struck me as over-reaching and so I asked “Why add intelligence to this? What is lost if we simply describe the dynamics of populational and community adaptation?” I was abruptly shut down. This is the core of the problem. What does it add to our understanding to call this “engineering” or “intelligence”?

At best I can only think this is a case of a scientist who wants to extend what Dennett once called the “white picket fence” around human specialness to the natural world. If so, it is either a case of clubfooted Aristotelianism or a kind of panpsychism, or both. Either way, this is no more a matter of a new approach than a return to animism might be.

Shapiro has been employed by more than a few intelligent design (ID) theorists on the strength of a paper he published in 1997. It is unclear whether he supports ID, from this book, despite a somewhat disparaging reference at the start of Part IV [p115]. I think he should be seen as a traditional Aristotelian rather than a modern ID theorist (which would have made him one of the few actual ID theorists in existence!), but then he fails to deal with how functional or teleological explanations explain if they are not the result of a trial and error – that is to say, natural selection – process. Since Bacon we have known that causes should not come after their effects. He does not deal with this problem, a staple of teleological discussions since then. Simply asserting something like Kant’s “purposiveness” (Zweckmässigkeit] as an explanation of living systems is not enough.

I cannot recommend this book.

Shapiro, James A. Evolution: A View from the 21st Century. FT Press Science, ISBN: 0-13-278093-3, $34.99 Publisher’s site

Late note: I messed up and confused him with Robert Shapiro. Now corrected.

Later change: fixed link to Shapiro’s site and corrected his profession. Thanks to Maureen O’Malley for spotting this.



  1. David David

    There are actually 2 Shapiro’s often cited by creationists. The one who wrote this book is James, trainer of von Sternberg. Robert, the chemist, is the one critical of the RNA world and other OoL theories. Can’t say I’m fond of either, they both play the skeptic without really acknowledging that many of their criticisms aren’t original, and hyping their preferred theories without acknowledging their significant lack of any experimental support.

  2. bob koepp bob koepp

    Yeah, metaphors are great, as long as they are taken metaphorically. But in fairness to panpsychists, don’t assume they embrace a sort of pansapience. The evidence that there are non-intelligent psyches is overwhelming.

  3. But Shapiro wanted to call this just intelligence.

    The trouble is, we have no agreed meaning for “intelligence”. If AI people can say that mechanistic computation is intelligence, then it’s hard to criticize Shapiro for saying that adaptive behavior is intelligence.

  4. Geoff Weil Geoff Weil

    “Engineers do not accomplish defined functional goals.”
    Really? Tell my boss. I spent a lot of my career being handed a spec and told to design something to meet it. Unless you have a very different definition of “functional goal”.

    • That was a deliberately provocative claim. It can be unpacked like this: the field of engineering does not achieve functional goals that are defined prior to there being experience within the field of how to achieve them except through trial and error. Once the field has this experience, then it can solve defined functional problems, and only then. The rest of the paragraph makes this clear, I would have thought. If not, here is a defined functional goal: make a space elevator. We know the physics and functionality involved – go do it. Can’t? Why not? Surely as an engineer you can just solve this ahead of time?

      • Geoff Weil Geoff Weil

        Know many engineers? That all seems way off base to me. If engineers only could achieve what was already known through trial and error we would all still be in the banging rocks together stage. First off, any good engineering education contains a boatload of theory. Applying physical theory towards achieving defined goals is more like the usual engineering process. (At least if it’s any good, lots of dumb engineers out there) Blind tinkering is for basement inventors. “OK, Ugg, let’s try hitting the rock with your head, next. Hmmm, that didn’t work. Let’s try hitting the rock with your aunt Uggtilda next. Hmmm, didn’t work, either. Let’s hit the rock on that mastodon next.”

        People are working on space elevators. Materials science isn’t there yet, but carbon nanotubes show a lot of promise.

      • Leviathan Leviathan

        Am I missing something? Even if making a space elevator would proceed purely by trial and error, isn’t the point that there was a prior intention to make a space elevator, followed by the actual making of a space elevator, regardless of the process involved? Does that not amount to accomplishing a functional goal in a sense that, say, evolving a wing does not?

        • 13 years in engineering school, a doctorate, and some 20 yrs of academic research in engineering design – and I’m still not quite getting John’s original text.

          Leviathan is closest to my (current) thinking. While engineers do use trial and error, I also think we accomplish functional goals. Tacoma Narrows is a great example of what happens when engineering misses something important, but it hadn’t been an issue till then. That is, bounded rationality reared its head. Put another way, if we had had access to more accurate information from other bridges in the past, then we would have realized that a good wind can really ruin your day, and the TN bridge would have been designed better.

          Another point: if the trial & error thing is true of engineering, then it’s true of all of science too – no? (Which would then beg the question: why pick on us poor engineers? 🙂

        • Let us suppose there was an engineer in the Neolithic. Aksed to build a bridge, the engineer does not work out, ahead of time, all the stress analyses, resonance analyses, and so forth, because although these are defined functions, they are not yet known. The metaphor Shapiro is making is that somehow genetic engineering proceeds by some cognitive clairvoyance, in which the act of defining the functions to be filled gives us a way of planning and designing. In fact, all knowledge in engineering (yea, and all sciences) is post hoc, not pre hoc, but a good many engineers, scientists and people who intuit about design seem to think that the act of conceiving a goal gives the solutions. In most cases the very goals themselves are achieved through trial and error. If our Neolithic engineer had been asked the best way to get to the moon, he would have thought (as the writers of Genesis in fact did) that the best way was to make a tower.

          Knowledge – all knowledge – proceeds through iterative incremental testing of ideas, goals and methods. If bacteria do any genetic engineering, they do it through the operation of mechanisms achieved by natural selection, just as engineering (and science) itself does.

          The problem with Shapiro’s and the ID crowd’s views is they seem to think design is somehow different to natural selection, when it is itself just the same sort of thing. There is nothing magical about design that explains why modern designers can reach the moon – the techniques, knowledge and the very goal itself has evolved through ordinary processes of trial and error. As recently as 500 years ago a great many people thought the moon was actually unreachable on account of the crystal spheres. Now we know better and have worked out how to get there. We didn’t do this by sitting down in front of a piece of paper with a pen and ruler. That was involved, but only after the knowledge had been achieved by engineering (in this case rocketry) and testing.

        • Engineers do have a tiny advantage over bacteria: they can do some of their trial and error on paper and also avail themselves of generalizable knowledge such as the laws of physics. One proposes a bridge design and then does finite element analysis to see if an actual bridge built to those specs would fall over. Of course, if one’s assumptions are wrong, disasters occur that wouldn’t happen to bacteria unless they also turn out to possess something analogous to language and insight.

          I don’t think people are magic animals, but it does appear to be the case that the pace of cultural evolution increased enormously at some point in the last hundred thousand years as witness the much faster turnover of artistic styles. It seems to me that the arguments about intelligent design in non-human evolution would be greatly advanced if we had a better handle on intelligent design in human cultural evolution.

        • ckc (not kc) ckc (not kc)

          …trial and error – in some ways it’s too bad that, in both science and engineering, the errors (at least the small ones, and certainly the recent and less egregious ones) are wiped from the record. Seeing only the big errors gives us all, scientists and engineers included, a false sense of how we got to where we are.

        • Oh I don’t want to ignore the obvious differences human engineers have over simpler trial and error systems like bacterial colonies and populations. Of course we have symbolic storage and transmission, and an (evolved) ability to model internally. If that’s what you mean by “insight”, then sure.

          But as to the supposed increase in capacity and cultural knowledge, almost all of that can be put down to population density permitting network complexity and enough transmission to prevent information loss over time. And what can’t be is due to new technologies like writing. It’s not that cultural evolution has picked up – there’s just a bit more of it happening.

  5. sparc sparc


    The operation of a tightly regulated sequence of natural genetic engineering events in the adaptive immune system is probably the most elaborate example we have of purposeful genome manipulations.

    If this would really be similar to enginnering the industry would have to produce millions of screws find one that fits a given nut. In addition, it may happen that the only one fitting is not among those being generated. Indeed a working system in nature but surely not purposeful in the sense Shapiro says. In addition, as how intelligent would Shapiro judge the introduction of hypermutations in V-genes of rearranged immunoglobulin loci introduced by the action AID?

  6. ted lawry ted lawry

    From the review, Shapiro bases his whole theory on the use of a form of natural selection by the adaptive immune system, which has long been known. As sparc says, that is pretty mindless way to do engineering. Is the book really that bad? (I wouldn’t be surprised if it were, I am just asking for completeness.)

  7. Lino Lino

    As with G. Weil, having been an engineer, I don’t understand your categorical statement that engineers “do not accomplish defined functional goals.” This seems to defy any common sense understanding of what engineers do.

    The very example you provide seems to bear this all out.

    The engineers who had the task of constructing the Tacoma Narrows Bridge, did, in fact, based on know principles of physics, build a bridge. No one can deny that fact. And, of course, building a bridge was indeed their “defined functional goal.” Somebody—probably Washington State—asked the engineers to build a bridge over the Tacoma Narrows. And this is exactly what they did. Mission accomplished.

    But it didn’t work. Why? Because a physical force (wind) acting in a vibrational manner was able to vibrate the steel at such a frequency that the frequencies became additive, and hence increasingly more powerful, until finally the bridge collapsed. Now, the idea of “trial and error” enters. That is, their construction failed, and through this failure something new was learned about the construction of bridges.

    But is that what Shapiro was talking about? Was he talking about cells that learned new skills through “trial and error” so that, over time, their set of skills changed (evolutionarily); or was he, instead, talking about how the static, constant over time, built-in cellular mechanisms operated in a “trial and error” fashion to ‘fix’ certain problems adaptively well? Obviously the latter, and not the former.

    The engineering of NGE is more at the level of robotics, where design engineers attempt to develop software so that the robots have a set of tools (software routines) that can be employed as the robot encounters a changed environment. You see, that’s how ‘artificial life’ is engineered. And the parallel to nature should readily be apparent.

  8. Rolf Aalberg Rolf Aalberg

    In a comment at Pandas Thumb, I used the impressive food processing capability of the human intestine as an example of ‘intelligence in my book’, and was promptly adviced to ‘ read another book’.

    To make myself clear, what was simmering in my mind was the idea that ‘intelligence’ is a rather subtle concept, and I have never toyed with ideas about the existence of an intelligence in nature with predictive powers.

    Therefore to

    What does it add to our understanding to call this “engineering” or “intelligence”?

    I say ‘nothing’.
    The idea I want to express is that in a broader sense, intelligence may be said to exist everywhere in nature; is natural. In that sense, the effect or result of the natural process of random mutation and natural selection, devoid of prescience or intent, may be indistinguishable from “Intelligent Design”.

  9. Hi Dr Wilkins,

    What is your opinion on the idea that Darwin was a teleologist precisely because of his conception of natural selection?
    Some of it is discussed here:
    Natural Selection Teleologists

    I have a few questions regarding natural selection and it is primary derived from the manner in which the concept is defended (scroll down to the last paragraph for the question). Briefly, is natural selection a mechanism, cause, a process, an outcome, force? There seem to be no agreement on these basic issues. Well that is the way I perceive it anyway.


    • Darwin was not, contrary to Asa Gray’s comment, a teleologist, nor did he reintroduce teleology into biology just because someone drew a diagram that puts him there. He was very careful not to treat purposiveness in biology as the result of an intrinsically imposed purpose, or as the imposition of purpose by a purposive being. His “teleology” is a subset of mechanism, not a superset.

      As to the nature of natural selection I demur to the Walsh and Ariew approach: I have argued before (also here) that natural selection is a schematic form of post hoc explanation that needs to be fleshed out in mechanistic terms to be explanatory.

      • The argument is not that Darwin viewed NS as intrinsically imposed. Instead the argument is that it is an extrinsic teleological force or agent or cause that preserves elements of Aristotle’s functional teleology and formal causality but is based on a mechanistic view of matter. The teleology is not intrinsic in matter (as in Aristotle’s hylemorphic substances) but extrinsically imposed by NS.

        It is in this sense that the final cause of natural selection (extrinsically imposed onto matter if you want) is just that it tends to “maximize reproductive success in particular environmental niches” or “maintains” the prevalence of beneficial mutations, or “limits” or “favours” some variations over other variations, or “steers” biological change toward the local maxima in the “fitness landscape”.

        You can argue that the teleology is a “subset of mechanism, not a superset”, sure, but that does not take away from the observations that it preserves elements of Aristotle’s teleology.

        John S. Wilkins: “I have argued before (also here) that natural selection is a schematic form of post hoc explanation that needs to be fleshed out in mechanistic terms to be explanatory.”

        Fair enough. From what I understand, it looks like you support the view that natural selection is to be used as a descriptive term and not a prescriptive term. Not prescriptive in the sense that natural selection is an agent (albeit impersonal and blind, as in non-directional) that causally interacts with something else. For example, natural selection is an agent that causally interacts with evolution by “maintaining” or “favour” or “produce fitter” etc.

        Descriptive in the sense that natural selection is just a descriptive term to describe when you have individuals in a population that have some kind of variation and fitness differences and are able to pass on their traits. The descriptive view makes sense IMO.

  10. Jud Jud

    If I recall correctly, Shapiro many years ago published, and was associated with other researchers who published, papers on the capabilities of microorganisms in groups qualitatively exceeding their capabilities as individuals. (Compare with work on social insects, e.g., ants.) This work was initially derided, but at least some of it is now widely accepted.
    It seems to me his current position, which certainly appears to outstrip the supporting evidence, may be an attempt to reprise this history.

  11. Is Evolution Predictable

    Of course it is.
    Nearly. Approximation, proportional to extent of included factors.
    And AFTER comprehending what evolution is…

    Is Evolution Predictable

    From DH comment on

    Origin And Nature Of Natural Selection

    Life is another mass format, a self-replicating mass format.
    All mass formats are subject to natural selection.
    Natural selection is the delaying conversion of mass to the energy fueling cosmic expansion.
    Cosmic expansion is the reconversion of all the Big Bang singularity mass to energy.

    Natural Selection Updated 2010, Beyond Historical Concepts:

    Natural Selection applies to ALL mass formats. Life, a self-replicating format, is just one of them.
    Natural Selection Defined:

    Natural selection is E (energy) temporarily constrained in an m (mass) format. Period.

    Natural selection is a ubiquitous property of each and every and all cosmic mass, spin array, formats, from the biggest black hole to the smallest physical particle. Mass strives to increase its constrained energy content in attempt to postpone its reconversion to energy and to postpone addition of its constitutional energy to the totality of the cosmic energy that fuels the cosmic expansion going on since Big Bang.

    Dov Henis
    (comments from 22nd century)

Comments are closed.