Paul Davies, chemistry, and the origin of life

January 14, 2013 • 6:50 am

How the unique properties of life originated from inert matter is still one of the great unsolved problems of biology. Creationists, of course, claim that our failure to solve it means that God did it: as Ingersoll noted in yesterday’s quote, “Our ignorance is God; what we know is science.”

And perhaps we’ll never know precisely how life began, for it happened in the distant past and involved chemical reactions that could not fossilize.  But I have confidence in three things: life originated naturally and not through God’s fiat; that we will show that this was possible within 50 years or so by demonstrating the evolution of life-like systems in the laboratory under primitive Earth conditions; and that while life may have originated more than once, all living species descended from only a single proto-organism (lots of evidence for that one). If we can demonstrate the origin of life in the lab through realistic experiments, then—although we may not know how it really happened several billion years ago—we can say that it could have happened naturally, and therefore we need not invoke God.

Paul Davies is an astrophysicist who spends a lot of time promoting the comity between science and faith, and who won the Templeton Prize in 1995. As Wikipedia notes:

An opinion piece published in the New York Times,[2] generated controversy over its exploration of the role of faith in scientific inquiry. Davies argued that the faith scientists have in the immutability of physical laws has origins in Christian theology, and that the claim that science is “free of faith” is “manifestly bogus.”[2] The Edge Foundation presented a criticism of Davies’ article written by Jerry Coyne, Nathan Myhrvold, Lawrence Krauss, Scott Atran, Sean Carroll, Jeremy Bernstein, PZ Myers, Lee Smolin, John Horgan, Alan Sokal and a response by Davies beginning I was dismayed at how many of my detractors completely misunderstood what I had written. Indeed, their responses bore the hallmarks of a superficial knee-jerk reaction to the sight of the words “science” and “faith” juxtaposed.[3] While atheists Richard Dawkins[4] and Victor J. Stenger[5] have criticised Davies’ public stance on science and religion, others including the John Templeton Foundation, have praised his work.

I’ve written about Davies’ accommodationism several times on this site; one example, in which he finds God in quantum mechanics, is here (see also his 2007 New York Times article asserting that science rests on faith.) But, to give him credit, Davies is also something of a polymath, and has an interest not just in physics, but in the early Ediacaran fauna and especially in the origin of life (“abiogenesis”). He’s apparently done technical work on the last issue, which I must confess I haven’t read (the literature on abiogenesis is huge). But in yesterday’s Guardian, Davies summarizes his work in a piece called “The secret of life won’t be cooked up in a chemistry lab.

His basic claim is that science hasn’t solved the origin-of-life problem because it’s been left largely in the hands of chemists, and from-the-ground-up reductionist chemistry is not the way to solve it:

But a more fundamental obstacle stands in the way of attempts to cook up life in the chemistry lab. The language of chemistry simply does not mesh with that of biology. Chemistry is about substances and how they react, whereas biology appeals to concepts such as information and organisation. Informational narratives permeate biology. DNA is described as a genetic “database”, containing “instructions” on how to build an organism. The genetic “code” has to be “transcribed” and “translated” before it can act. And so on. If we cast the problem of life’s origin in computer jargon, attempts at chemical synthesis focus exclusively on the hardware – the chemical substrate of life – but ignore the software – the informational aspect. To explain how life began we need to understand how its unique management of information came about.

Well, he doesn’t mention that many biologists and biochemists have also worked on the problem, including Stanley Miller, Gerald Joyce, and Nobel Laureate Jack Szostak. But never mind. Davies goes on to say that the fundamental problem is not in the purview of chemistry, but in how biological information gets organized hierarchically:

Information theory has been extensively applied to biological systems at many levels from genomes to ecosystems, but rarely to the problem of how life actually began. Doing so opens up an entirely new perspective on the problem. Rather than the answer being buried in some baffling chemical transformation, the key to life’s origin lies instead with a transformation in the organisation of information flow.

Sara Walker, a Nasa astrobiologist working at Arizona State University, and I have proposed that the significant property of biological information is not its complexity, great though that may be, but the way it is organised hierarchically. In all physical systems there is a flow of information from the bottom upwards, in the sense that the components of a system serve to determine how the system as a whole behaves. Thus if a meteorologist wants to predict the weather, he may start with local information, such as temperature and air pressure, taken at various locations, and calculate how the weather system as a whole will move and change. In living organisms, this pattern of bottom-up information flow mingles with the inverse – top-down information flow – so that what happens at the local level can depend on the global environment, as well as vice versa.

. . . Walker and I propose that the key transition on the road to life occurred when top-down information flow first predominated. Based on simple mathematical models, we think it may have happened suddenly, analogously to a heated gas abruptly bursting into flame.

. . . The way life manages information involves a logical structure that differs fundamentally from mere complex chemistry. Therefore chemistry alone will not explain life’s origin, any more than a study of silicon, copper and plastic will explain how a computer can execute a program. Our work suggests that the answer will come from taking information seriously as a physical agency, with its own dynamics and causal relationships existing alongside those of the matter that embodies it – and that life’s origin can ultimately be explained by importing the language and concepts of biology into physics and chemistry, rather than the other way round.

Now I have to say that this is either way above my pay grade, is too abstruse for the general public, or is simply wrong. And I’ll take a chance here and say that the problem is not going to be solved using information theory or “top down” considerations. Yes, of course the environment played a role in the origin of life, but that doesn’t mean that we can’t take a reductionist approach. (As an example of “top down” control, Davies uses the fact that genes can be turned on by environmental factors, but of course that area of research progressed purely through reductionism, not through incorporating “environmental information theory.”)  To me, the problem of abiogenesis involves taking a set of original molecules that didn’t contain any information for a replicator, and then figure out what kind of processes enabled them to gain the ability to replicate. Along with that is the formidable problem of how the replicator became connected with a phenotype, in other words, how that replicator evolved to produce proteins and cells—and that is information.  All this can, I think, be attacked without any importation of “information theory” from the beginning. Information is a consequence of, rather than an a priori requirement for, understanding abiogenesis.

In fact I don’t know what Davies is suggesting here, and maybe his technical work has insights that are fruitful.  The Guardian article, though, is simply opaque, and there are hints from other work that Davies’ approach may not be fruitful.

One of these is a new book by chemist Addy Pross, What is Life? How Chemistry Becomes Biology. (The title, of course, is borrowed from Erwin Schrödinger’s famous book that had such a huge influence on the pioneers of molecular genetics.)

I have just received this book and have only glanced through it, but it’s a short volume describing the latest thought on the origin of life, and Pross’s own speculations on how it started. And it’s pure chemistry: no information theory needed. It does, in fact, what Davies says hasn’t been done: meshing the language of chemistry with that of biology.

I’m told by those who have read it that Pross has an intriguing and possibly correct idea about how a population of molecules can, through a process of biochemical natural selection, become replicators that get hitched to metabolism. The “information” of life then is a result of the process, not something that, as Davies asserts, needs be modeled in from the beginning. As Pross notes on p. 152:

A moment’s thought suggests therefore that the term “information” in its biological context is just “specific catalysis” when considered in a chemical context.

I’ll be reading this book in the next month or so, and will report back, but I wanted to say four things now:

1. If you’re interested in the origin of life, read Pross’s book. It’s written for the curious layperson, and is therefore perfectly suitable for the non-scientist readers of my site.
2. I suspect that Davis is wrong and that the reductionist approach, based on chemistry, is the way to solve the problem of life’s origin.
3. But I may be wrong, too, and somehow, in ways I don’t understand “top down information flow”—whatever that means—may be the key to the problem. But if the history of science is any guide, the problem will yield to chemists starting with a population of molecules existing under realistic early-Earth conditions.
4. The Guardian article is completely opaque to the nonscientific reader (indeed, to me as well), and seems more like p.r. for Davies’ work than a genuine attempt to explain to the layperson how life might have originated.

Picture 1

h/t: Michael

127 thoughts on “Paul Davies, chemistry, and the origin of life

  1. It seems to me that natural selection fulfills the role of top-down information flow. Replicators are “informed” by the environment about what works, and the population of replicators (the “gene pool”) is modified as a result.

  2. the answer will come from taking information seriously as a physical agency

    What the heck does it mean for an abstract property to have “physical agency”? Does “justice” have physical agency? Does “primeness” have physical agency? Information only has agency when it is instantiated in a physical system, and the kind of agency it has will be solely dependent on the nature of that physical system. For example, an abstract sequence involving four item can be physically represented as magnetized domains on a hard drive or as a series of nucleotides in a strand of DNA, but those two options have very different physical implications.

    1. In physics, information is sometimes treated like a property. “Information entropy” is involved in solutions to Maxwell’s demon, and a big question in physics is if information is conserved. (Caveat: Not at all a physicist, so I may be quite wrong.)

      1. Yes, but Tulse is correct. Information is always relative a system, but people like to reify it.

        Or in the case of creationists, deify it.

        1. I’m going to be perverse and defend Davies, only based on his words that Jerry posted. My defense rests on the idea that his ideas are obvious platitudes.

          He’s arguing, as far as I can tell, that information exists in a hierarchy of more-or-less abstract levels (genes to ecosystems, for example), that there are interactions between these levels, and that the interactions are not strictly bottom-up. Well, duh! Does anyone seriously argue with that? It’s not threat to reductionism, much less an appeal to religion.

          1. There’s agreement that gene-ecosystem interactions aren’t strictly bottom-up, but Davies is suggesting abiogenesis might be explained as top-down (chemical-informational) interactions.

            Coyne maintains that Informtn Theory is a non-starter for explaining abiogenesis, particularly whilst Pross et al are continuing to unpack basic chemical-biological interactions.

      2. Information is not a conserved quantity, but it is a real quantity.
        “…entropy is defined as the amount of additional information needed to specify the exact physical state of a system, given its thermodynamic specification. The second law is then a consequence of this definition and the fundamental postulate of statistical mechanics.” (Wikipedia)
        The Second Law deals in realities – information about the microstates of a system.

        1. Hmm… but that’s just our model; we chose to define entropy in that way, we choose to specify the exact physical state. (And the Second Law has been stated in different ways over time.)

          Reality just is.

          This is (partly) Atkins point, above.


          1. I can’t agree with that. Remember Eddington’s famous pronouncement:
            “If someone points out to you that your pet theory of the universe is in disagreement with Maxwell’s equations—then so much the worse for Maxwell’s equations. If it is found to be contradicted by observation—well these experimentalists do bungle things sometimes. But if your theory is found to be against the second law of thermodynamics [the law of entropy] I can give you no hope; there is nothing for it but to collapse in deepest humiliation.”
            The second law was stated differently once upon a time, before statistical mechanics. I don’t think the 19th century mechanics of heat engines is very relevant in the age of the quantum.

              1. Not at all. You suggest that it is arbitrary (“…we chose to define entropy in that way, we choose to specify the exact physical state.”) But (Wikipedia again) “In statistical mechanics, the Second Law is not a postulate, rather it is a consequence of the fundamental postulate…” which is not an arbitrary model. Check “fundamental postulate” in the same source to see the quantitatively exact consequences and equivalences in informational terms.

              2. Well, I’m not suggesting the second law is arbitrary either, just the way we describe/define it. Perhaps “notional”, rather than “arbitrary”.

                It’s still not evident to me that the second law depends on or validates the notion that information is a fundamental physical thing.


              3. The universe today is in a very different state from what it was last Friday. The conservation laws cannot account for this. Obviously. The Second Law or Law of Entropy does. Entropy is defined as the product of Boltzmann’s Constant and the natural logarithm of Omega. [quoting Wikipedia from here onwards]:
                “Omega is the number of microstates consistent with the given macrostate.
                The statistical entropy reduces to Boltzmann’s entropy when all the accessible microstates of the system are equally likely. It is also the configuration corresponding to the maximum of a system’s entropy for a given set of accessible microstates, in other words the macroscopic configuration in which the lack of information is maximal.”
                Hard stuff. See also:

                Maybe it would be clearer if I ask you to a)define and give an example of “a physical thing”
                b)state the modern Second Law in terms which do not reduce to information (i.e. which only refer to your “physical things”
                c)make it clear whether entropy is fundamental.
                (That’s not meant in a bossy or manipulative way. I just mean that thinking about those points will convince you better than having pages of mathematical physics waved under your nose. I don’t like browbeating people with technical stuff – especially when I’m no expert myself.)

              4. I’ve forgotten too much physics to argue at that level anyway… 

                By physical “things”, something like mass, energy, charge, spin, colour (in QCD), parity, &c.

                It just strikes me that taking information to have the same objective reality is confusing the map for the territory.

                I’m not trying to argue to prove you wrong, but I still don’t see that information has that objective reality.

                Maybe you can provide something insightful and persuasive, but for the moment I don’t see this going anywhere fruitful, so don’t feel obliged to reply at length.


              5. Hi Ant.

                I’ve tried to put in in a few hundred words. I hope it does it for you:

                Consider n molecules of a gas in an enclosure which can be divided into halves by an airtight shutter. Initially the gas is confined in the left half. Then the shutter is opened. After a short interval, the probability that any particular molecule is to the left has fallen from 1 to 1/2. Somewhat later, if you slam the shutter closed, what is the probability that all the molecules are again in the left half? If n=0, p=1. If n=1, p=1/2. If n=2, p=1/4. In general, p=1/(2^n). If we know which half a particular molecule is in, that is one “bit” of information more than if we don’t. The amount of information in bits equals minus one multiplied by the logarithm (to base 2) of that probability. (The “minus one” is just a sign convention – otherwise information would always be negative, and “more information” would be mahematically “less information”. It’s like avoiding saying that 10 electrons have “less” charge than 1 electron, because -10 < -1…) Opening the shutter had physical effects (halving the pressure, reducing the temperature…) which are irreversible unless energy is added from outside the system. (Only "statistically" irreversible. With n=10, if we closed and opened the shutter every second, then a few times every hour we'd find all the molecules back in the left half. For a thousand molecules we'd have to wait a century, and for a million the time since the Big Bang would be insufficient – the probability falls exponentially. A real litre of gas might contain a million million million million molecules…)

                When all the molecules are confined in the left half, the system can change, can do work. When they fill the whole enclosure, nothing much can happen. But in both cases there is the same amount of energy present and the same amount of mass – conservation laws don't make things happen. That requires information or negative entropy. Information IS negative entropy (that's a definition, not a rhetorical trick). When the shutter first opened, information (i.e. knowledge about the positions of molecules) was lost, and that drove the changes (Universal Gas Law). The Second Law of Thermodynamics states that during any change entropy increases, i.e. information decreases. We are so used to thinking of energy that we think it drives change. Not so. Change or work requires irregularities in the distribution of energy, and the measure is information.

          1. I haven’t seen it. I had heard that there was a version out recently, which got reasonably respectful comments. But for some reason (abroad, offshore, wife wanted to see something else?) I didn’t see it, so “no comment”.
            Stallone’s version was … getting to the level of “so bad, it’s good”. I liked the comics and it hurt.

            1. Information moves from place to place and requires energy to do so. That energy is accountable for, and is measurable. If the nebulous concept of “energy” thinks that information is real enough to interact with, then that’s good enough for me.

              1. Good question.
                But information can be here, and then be moved to somewhere else, and then be deleted from it’s original location. So something is moving. What that something is, is a different question. And I’ll have to cite the “Clinton Solipsism” : “It depends on what the meaning of the word `is` is.” Which is philosophy – not something I really worry about.

    2. Seems to me he’s making the very common mistrake of thinking of information as some sort of platonic ideal.

      We’ve known for quite some time that, though it’s often very useful to think of it as such, information just simply doesn’t exist.

      Instead, all we have is communication and computation.

      There is no information stored on your hard disk, just a particular pattern of polarized magnets. And, as far as what’s actually happening goes, the drive is just a form of delayed communication. You still have to use energy to record the signal onto the platter, and you still have to use energy to read the signal from the platters — all with the exact same ultimate limits as Claude Shannon demonstrated.

      Once you’ve communicated that time-delayed signal to the computer’s CPU, it can then perform computations on it, and those computations will have the same limits that Alan Turing figured out. And each step in computation will involve more communication, and that communication will still be subject to Shannon’s limits. At no point is there any “real” information, just real physical signals getting propagated and transformed.

      When all that gets translated into pictures projected by your computer monitor, a similar but much more complex and convoluted and messier process takes place in your head when your eyes communicate the signal transmitted by light from the display to your brain, which then does its own computation.

      When you look at information processing from this perspective, the “information processing” of the most primitive forms of life is so laughably simple that it’s not at all hard to imagine natural crystallization types of processes in the very messy natural environment stumbling on something “good enough” to bootstrap Evolution. Remember, it had the whole planet to work, it had hundreds of millions of years to work, and it only had to happen once.

      Indeed, the exact same processes are certainly still potentially happening all around us…it’s just that anything so simplistic would simply get eaten as abiotic nutrients long before it could establish itself tot he point that it could survive the onslaught of a billion years of very hungry evolutionary refinement.



      1. There is no information stored on your hard disk, just a particular pattern of polarized magnets.

        There is information in the sense that one chooses to interpret those magnetic patterns, but there is nothing inherent or objective about those patterns — one could just as well interpret the patterns of temperature or vibration.

        1. It’s even worse than that, actually.

          Every interpretation is equally valid.

          That’s the lesson from cryptography about one-time pads. You can pick any encoding you want, and it’s no more or less valid than any other encoding; it just works out to a different message.

          The sender and receiver have to agree upon the encoding, of course. But there’s nothing that says that two people can’t use two different encodings that result in the same message getting encoded entirely differently — or the reverse: two different messages getting encoded exactly the same.



      2. Information exist as a property though, a measure. The problem for platonists is that there is no single measure and the same measure means different things on different systems. For example, Shannon information on a radio channel means coded strings, while Shannon information that the environment channels into the genome means selection on variation.

        Then you get to codings and meaning (messages), but that is entirely independent of information theory. The maximum informative genome is the random according to the Kolmogorov complexity measure of information. In this case Shannon information goes to decrease Kolmogorov complexity information, by extinguishing alleles. But the meaning of the alleles are decided by fitness relative the environment, not SI or KC.

    3. What the heck does it mean for an abstract property to have “physical agency”?

      I guess that “physical” in this context means “agency in the physical world”, not that the agency itself is physical. In general what does it mean for anything to have agency? To the extend that agency presupposes an agent, all uses of that term in a reductionist, “materialist” framework are essentially metaphorical. No?

  3. History shows that the surest way for a great physicist to lose credibility is by embarking into the question of evolution and the origins of life.

  4. Chemical reactions are not at all random: they are directed by the laws of chemistry and, when optimized, are extremely specific and selective. That’s the source of this so-called “information”. Life is a lot of very well-optimized chemical reactions that got that way though chemical laws: they didn’t need any directions from above.

    If creationists (or crypto-creationists) would bother to examine any other chemical system, like the atmosphere or geosphere, they’d see the sort of specificity and selectivity that they attribute to “information” occuring there, too. Not as complex, but still there. There’s nothing special about life that requires information, all chemical systems, to some degree, self-organize.

  5. I also hesitate to say that he is wrong, as Paul is such a great intellect, yet like you I will risk that and say, he is wrong. There is nothing but nature to account for nature. Where many of us in physical sciences (I am an applied physicist), often fall short is something that you mention briefly here which is the use of information theory. Modern Computational complexity models, whether free-agent based, or by the use of finite, or non finite element analysis, has been limited due to some technical issues with storage, and processing power. This is rapidly changing, and the handful of cross disciplinary people working in this area, may take the work of biologists and physicists, and combine it with the powers of an AI solution to answer even these questions that Davies feels are out of our reach. The chess master Gary Kasparov made a very interesting point that a computer alone can play great chess, but a computer being aided by a human (even a mediocre player) is unbeatable. In science we will increasingly have to take this same approach in answering large questions, which have normally been reserved for philosophy. All of this is to say that as usual I agree with you Jerry.

  6. Like Bayesian Statistics, Information Theory is frequently over applied. Both approaches are definitely very useful in certain circumstances.

    Abiogenesis is not one of them.


    1. @ NoAstromoner – there are limits to Bayesian Statistics and Information Theory, but from my experience we have nowhere near come to the point where they are over applied. Most of the population doesn’t even know what these are, and if you ask even the best of undergrads at places like Columbia University (where I teach) they are not applied at all senior research projects, or in general problem solving, as students haven’t been taught these fundamental tools. If you learn these rather non intuitive concepts too late in your education, they are no longer feel natural, and we tend to seek other routes to interpretation. There are other ways, and better ways, but these are good, and I think under applied, rather than over applied.

      1. Are you two talking at cross purposes? mcputman that these techniques could usefully be more widely used; NoAstronomer that they’re sometimes used where they are not in fact appropriate?


  7. I will add this to my Amazon “Save for later” list!

    Contra Davies, here’s Peter Atkins (a professor of chemistry) from On Being (p. 29 of the OUP pb edition):

    … we have to be cautious in interpreting a change in DNA as the emergence of information. At a molecular level, everything is junk and all changes are random. … the structure of DNA itself has not been constructed with a message in mind. … Natural selection takes over once random modifications to molecular junk have delivered a viable message, and if the modified organism is successful in its reproductive capacity it is welded into the biosphere and we think of the DNA as having developed information.


  8. As for the concept of biological or genetic information:

    “In the precise sense in which one may speak of semantic information, genetic information can hardly count as an instance of it. It simply lacks all its typical features, including meaningfulness, intentionality, aboutness, and veridicality. DNA contains the genetic code, precisely in the sense that it physically contains the genes which code for the development of the phenotypes. So DNA does contain genetic information, like a CD may contain some software. But the genetic code or, better, the genes, are the information itself. Genes do not send information, in the sense in which a radio sends a signal. They work more or less successfully and, like a recipe for a cake, may only partly guarantee the end result, since the environment plays a crucial role. Genes do not contain information, like envelopes or emails do, nor do they describe it, like a blueprint; they are more like performatives: ‘I promise to come at 8 pm’ does not describe or contain a promise, it does something, namely it effects the promise itself through the uttered words. Genes do not carry information, as a pigeon may carry a message, no more than a key carries the information to open a door. They do not encode instructions, as a string of lines and dots may encode a message in Morse alphabet. True, genes are often said to be the bearers of information, or to carry instructions for the development and functioning organisms, and so forth, but this way of speaking says more about us than about genetics. We regularly talk about our current computers as if they were intelligent—when we know they are not—and we tend to attribute semantic features to genetic structures and processes, which of course are biochemical and not intentional at all. The ‘code’ vocabulary should not be taken too literally, as if genes were information in a semantic-descriptive sense, lest we run the risk of obfuscating our understanding of genetics. Rather, genes are instructions, and instructions are a type of predicative and effective/procedural information, like recipes, algorithms, and commands. So genes are dynamic procedural structures that, together with other indispensable environmental factors, contribute to control and guide the development of organisms. This is a perfectly respectable sense in which biological information is indeed a kind of information. Dynamic procedural structures are a special type of informational entities, those that are in themselves instructions, programs, or imperatives.”

    (Floridi, Luciano. Information: A Very Short Introduction. Oxford: Oxford University Press, 2010. pp. 79-80)

    1. That seems to me to be what Ben Goren said with greater precision, but Ben’s post was a lot easier to understand.

  9. “The secret of life won’t be cooked up in a chemistry lab.“

    It’s only when you’ve thrown up your hands that you can clasp them in prayer.

  10. It seems to me that “top down information flow” refers to orders from either the Almighty or the Designer.

  11. When I read this–

    Therefore chemistry alone will not explain life’s origin, any more than a study of silicon, copper and plastic will explain how a computer can execute a program.

    –I hear echoes of the watch and the watchmaker. Perhaps he’s laying the groundwork for arguing that the information had to be organized by someone?

  12. See the article ‘Chance or Necessity’ at

    Quote from the abstract:” The emergence of life is probable on any wet, rocky planet. Serpentinization gives rise to alkaline hydrothermal vents that form: (i) simple organics; (ii) catalysts that direct primordial metabolism (iii) micropores with cell-like properties; and (iv) proton gradients equivalent to the proton-motive force.”

    A somewhat different take to Paul Davies.

    1. The emergence of life is probable on any wet, rocky planet. haha..well until their sample size is greater than 1 (earth) or it happens in the lab that statement is a big faceplant.

    1. Whatever the origins of the inelegant designer, it’s still a skyhook as Richard Dawkins has so eloquently put it. You’ve explained nothing and merely added a level of obfuscation.


        1. Hmmm…even if Dennett is the original source for the skyhook idea — and it seems quite reasonable for that to be the case — it’s Dawkins whom I’ve always heard use it. See here, for example.

          …and, sure enough, Dawkins gives credit to Dennett right at the beginning.


          1. I’m not sure when Dennett appropriated the term, but skyhooks have been in use in the mountaineering world since the 1960s, and the idea is probably older.
            See, for example,
            (It should be pointed out that Petzl started as a caving equipment manufacturer, and are now a major industrial access equipment supplier. This is as un-metaphorical as you can get without using a hammer.)

            1. I think skyhook was grim RFC humour (predecessor to the RAF in WWI). If some young pilot asked what was the best strategy if you got your tail shot off, etc… Of course, they may have borrowed it from aircraft designers, say.

              1. Oh yeah, I’d forgotten those “structures”.
                I think that I’d rather sit on top of several tonnes of explosives and light the blue touch paper. I’m sure they’d work, but … you go first, OK? (I’ll follow with the cave-diving equipment.)

              2. Another Matt:

                “Sounds like a Petard” sounds like the English translation of a punchline for a French fart joke — though “Smells like a Petard” sounds like an even-funnier punchline for that French fart joke.

                Either one would have been a pretty cool name for a grunge-rock band.

  13. Templeton may like Davies because he is closer to deism than many scientists, but his track record is hardly accomodationist. See for example

    which expeesses similar ideas as the new Guardian piece, but makes it quite clear that God is not compatible with his hypothesis, nor is Intelligent Design.

    His view seems to be that there’s something implicitly life-friendly in the initial conditions of the universe, but which is not described in the known laws of physics. That is a perfectly reasonable view.

    The late Victorian attitude was expressed by Lord Kelvin: “There is nothing new to be discovered in physics now, All that remains is more and more precise measurement.”

    Then came radioactivity, quantum theory and relativity…

    There is a similar over-confidence now.

    1. No, it is not reasonable. Physical laws describes all of physics.

      Davies is also subscribing to the idea of finetuning, despite that we know of the weakless universe and other obvious cases that rejects it.

      1. Physical laws describe the whole of physics… That’s a tautology. The question is whether we KNOW all the laws. In particular, Davies suggests that some law applying to the initial conditions of the universe (governing evolution to a life-friendly state) remains unknown. Certainly there are laws relating to the initial state which are unknown. Whether they act as Davies suggests is yet to be seen.

        The weakless universe is purely hypothetical. We don’t live in it. We don’t know if it would work. It would have no neutrinos, no stars even, unless the correct proportion of deuterium was arbitrarily fine-tuned in at the big bang.

    2. The late Victorian attitude was expressed by Lord Kelvin: “There is nothing new to be discovered in physics now, All that remains is more and more precise measurement.”
      Then came radioactivity, quantum theory and relativity…
      There is a similar over-confidence now.

      I don’t know which physicists you’ve been talking to, but every one that I’ve talked to for a couple of decades has known that there are gaping holes in our understanding of physical reality, and that our best theories do not form a logically self-consistent whole. From the well-understood end, we know that quantum mechanics gives us a good understanding of how matter behaves on the small scale, and we know that general relativity gives us a good understanding of how matter behaves on a large scale ; these understandings are accurate to around 10 significant figures of measurement, if not better. But these two theories are mutually incompatible, and no-one has worked out a way to reconcile them (not for want of trying – there’s a Nobel gong there!). Meanwhile the astronomers point out that the matter which we know of represents something like 4% of the material of the universe.
      (Sorry the wife interrupted my chain of thought with that silly cross-at-work case on the news. There was something else … but that’s enough.)
      Physicists know that they’ve got two major, mutually incompatible theories that describe about one twentieth of the universe. That’s not “over-confidence”.
      What we do know is that whatever theory replaces QM and GR, will need to agree with them to around 11 significant figures in the experiments that we can carry out.
      Get to it – there are Nobel prizes waiting!

      1. Oh yeah – my train of thought was restored by de-icing the car :
        From the poorly-understood end, there’s the mystery of the “dark mass” whose gravitational consequences we can see – about 20% of the mass-energy in the universe – and the wholly obscure repulsive energy field called “dark energy” whose acceleration of the expansion of the universe was observed in the mid-1990s and which constitutes around 76% of the mass-energy of the universe. Leaving the 4% remainder which we have two incompatible theories describing.

        1. I completely agree with you, insofar as most major physicists are concerned. I was referring to the over-confidence here (in this and similar threads) and similarly in discussions at RDF. It would be crudely simplistic to say that some evolutionary biologists (or their followers) are full of gung-ho and naively materialist optimism about ultimate explanations – to the extent that theoretical physicists may be written off as “deists” or woo-mongers. That’s especially the case when, like Davies, they suggest that mathematical entities have some sort of reality. Try this:

          “The picture of quantum tunneling from nothing raises another intriguing question… [It] is governed by the same fundamental laws that describe the subsequent evolution of the universe. It follows that the laws should be ‘there’ even prior to the universe itself… The laws are expressed in the form of mathematical equations. If the medium of mathematics is the mind, does that mean that mind should predate the universe?” Alex Vilenkin.

  14. All the responses to Davies’s article (along with the original article) including Jerry Coyne’s response are here:

    Actually I think Davies brings up an interesting point in his reply to the criticism i.e. Do we have any good reason for thinking that the laws which govern the universe should of necessity be simple? I found Lee Smolin and Sean Carrol’s opposite perspectives on this issue, particularly interesting.

    1. Do we have any good reason for thinking that the laws which govern the universe should of necessity be simple?

      Yes. Very, very good reasons.

      First is that, empirically, all laws to date that have been discerned have been simple — for, of course, certain definitions of the word.

      Second is that, if the premise is that more complex sources are required to explain simpler consequences, we’re left with an infinite regression of ever-increasing complexity. As Richard Dawkins puts it, that’s a skyhook with nothing supporting it from above.

      Last, everywhere we look, we see complexity arising from simplicity. We know how complexity arises from simplicity; it’s the natural order of things, at least insofar as we observe it to be. We’ve yet to encounter even a single instance of something complex that didn’t have simplicity at its core, or of something simple that required complexity as its origins.



      1. I think most scientists, including Davies, would agree with what you say, as far as it goes, but that doesn’t answer Davies’s question. It’s very easy to miss the point here, as some of the responses appear to do IMOP.

        When talking about the origins of the laws of physics it’s pretty meaningless to talk about “the natural order of things”, which begs the question. If you read the responses would you agree more closely with Sean Caroll’s (instrumentalist) or Smolin’s (more realist) point of view?

  15. Davies is reintroducing vitalism, as identified by astrobiologist Caleb Scharf: “that life is somehow truly distinct from other phenomena, not so much part of a continuum of processes and complexity.” [ ]

    The vitalism is not surprising, Davies is a deist and the Beyond Center is financed by at least two religious organisations, the Center for Theology and the Natural Sciences and the Templeton Foundation (through its front the Foundational Questions in Science Institute). [ ]

    We already know it is unnecessary to inject a qualitative difference in the transition process between chemical and biological evolution. According to recent thermodynamic models a population of RNA replicators can crystallize out during chemical evolution. This will happen with forcing, so no stalling akin to how other crystals form out of a solution. [“Thermodynamic Basis for the Emergence of Genomes”, Woo et al, PLOS Comp Biol 2012.]

    Such a population can form out of triphosphate activated nucleotides in alkaline hydrothermal vent environments. If it happens inside liposomes spontaneously formed from vent lipids, biological evolution takes over in Shoztak’s protocell fashion. Faster replicators promote liposome growth by putting tension on the liposome membrane which will enable their eating smaller protocells, et cetera. Growing liposomes will spontaneously divide with a pool of replicators inside to continue evolution.

    The problem facing astrobiology today is not suggesting more pathways taking systems from chemical to biological evolution, as noted by some astrobiologists. It is instead how to test and reject nonviable pathways.

    Reading Wlaker & Davies paper, I can’t see that they suggest such tests. They identify “a transition from bottom-up to top-down causation and information flow” to test the model internally. But they don’t identify how to test the application of the model (exemplified by a toy model of “a lattice of coupled logistic maps”) to astrobiology. [ ]

      1. Ha! I have a file somewhere of papers with great author names, one of which is Woo & Woo, on something (that I’ve forgotten) biochemical. That one was long before woo had the current connotation.

        But, the all-time best au I ever encountered, while flipping thru bound hardcopies as a grad student, in search of something else, (the way we used to do things back then) is this one – NB second au! I’ve always wondered if this was a real person, or something some wag at Cambridge slipped in to see if reviewers would comment. Searching the name only brings up the one citation in the link.

        1. I particularly liked a researcher I encountered while following up on the “nylon eating bacterium” – Dr Ohno.
          I tried following up on the second author you cite to find out what the “F.” stood for – and he doesn’t seem to be at Cambridge any more. However Dr Fayle is still bearing up under his burden.

            1. “Satsuma” according to my memory, but I’m sure you’re right. One needs mental hooks to hang things on (including skyhooks).

    1. I don’t think that Davies is a deist in any meaningful sense of the term, he just thinks there may be some reason why the universe is governed by one set of rules rather than another. It’s an interesting question that others such as Einstein have also asked. As Lawrence Krauss puts it:

      “Einstein once said that what most interested him about the Universe was whether God had any choice in its creation. He was, of course, not referring to a deity here, but rather asking the very important question: Can there only be one set of physical laws that allow for a consistent physical universe, or are there many possibilities?”

      Interests me too…

    2. Thanks for the link to Davies’s paper. The paper seems to make more sense than the Guardian article. I find the language of Davies’ discussion of information to be unhelpful. Like others here, I object to his reification of information. I also object to his talk about “top-down causation” and “information flow” between levels of organization. However, it seems to me that much of this language can be deleted or charitably translated into other language, leaving many of the substantive points of the paper more-or-less intact. But that does little to save the Guardian article, which lacks these substantive points and contains little but the misguided information-speak.

      Imagine a football team (higher level) and its players (lower level). We could say that appointing a new coach to the team caused the team to perform better, or that it caused the players to perform better. It makes little sense to see causation (or information) going from one level to the other (as Davies does). We are just modelling the same events at two different levels of abstraction, and we can see causation at both levels.

      It’s often the case that one level of model is more useful (for our purposes) than another. If we want to explain why Napoleon lost the battle of Waterloo, we probably won’t want to look at the molecular level. Similarly, some aspects of life and evolution might be usefully modelled in terms of information, a higher-level more abstract concept. The usefulness of such higher-level explanations might tempt us to see the causation at this level of explanation as being somehow more important than causation at lower-levels, and so tempt us to see this as “top-down” causation. But that would be a misleading way of seeing things. It becomes even more misleading if we start speaking of the information flowing from one level to another, as Davies does.

      I’m happy for Davies to point out that organisms can be seen as information-processing systems, and for him to discuss their evolution in such terms. I have no problem with that. But some of the language he uses tends to obscure the real issues.

  16. ‘Now I have to say that this is either way above my pay grade, is too abstruse for the general public, or is simply wrong’
    Once you wrote that, you got me. Let the very science oriented minds deal with prof Davies, I will read Pross instead.

  17. Paul Davies has a history of making vaguely theistic suggestions, and then protesting that he is being misunderstood when confronted by other scientists. Nonetheless, I find him worth reading. He wrote a readable book on biogenesis in 2000 called “The Fifth Miracle”. Despite the titular allusion to the miracles of Genesis, it is quite readable and enjoyable for a non-specialist (like me). I just dismiss his accommodationist proclivities.

  18. Jerry, I recently watched the videos of the “Moving Naturalism Forward” conference you participated in. I wondered if you had anything to say about the approach Terry Deacon espouses in those sessions (and I assume elsewhere — eventually I’ll get to more of his work).

  19. By coincidence, I received an email today notifying of an American Chemical Society journal issue focused on the origins of chemical evolution that may be of interest. URL:

  20. It seems to me that Paul Davies is criticizing the chemistry of abiogenesis without looking into the questions it is trying to answer. He says that “chemistry alone will not explain life’s origin, any more than a study of silicon, copper and plastic will explain how a computer can execute a program”. But we don’t know how the biological equivalents of silicon, copper and plastic emerged without life. Sure, understanding how information arose and was transmitted is an important problem for abiogenesis, but Davies makes it sound like the only problem.

    How did metabolism or homochirality arise? Before the building blocks of life can store and process information they have to be present. How were sugars, purines and pyrimidines, amino acids and fatty acids produced abiotically? These are all questions that chemists in abiogenesis ask regularly.

    Who knows, maybe Davies is right when he says “the key transition on the road to life occurred when top-down information flow first predominated”. But in order to get to the point where this transition is possible would require a heck of a lot of chemistry.

    1. Yes, but what does he mean by “the key transition on the road to life occurred when top-down information flow first predominated”? It’s hand wavy don’t you think?

  21. Pross’s book looks like a gold mine for quote miners.

    I’m trying the Kindle app for the first time. Kindle has a cool feature that aggregates everybody’s highlighting, and marks the most popular passages. I’m not very far along, but quite a few people have been highlighting the sentences that appear anti-evolutionary when taken out of context.

    Oh, to be so flexible with the false-witness commandment.

  22. In re.:

    Rather than the answer being buried in some baffling chemical transformation, the key to life’s origin lies instead with a transformation in the organisation of information flow.

    Seems to me that there are these two things: a cart and a horse, and Davies has managed to arrange them (insisted on arranging them) ass-backwards.

  23. Paul Davies’ most often mined quote is probably from The Cosmic Blueprint: New Discoveries in Nature’s Creative Ability To Order the Universe [1988] :-

    “[There] is for me powerful evidence that there is something going on behind it all… It seems as though somebody has fine-tuned nature’s numbers to make the Universe… The impression of design is overwhelming.”

    All shades of creationist love that one as shown in this Google results page

    It’s statements like that which elevated him to the high table at the Templeton trough

    1. I’m a little dismayed by the vitriol directed at Paul Davies in these comments. I don’t know the man, but he apparently has opinions and views and speculations that seem not agree with yours. Get over it.

      1. Sorry, Mr. Barnard, but there’s no vitriol in this thread, which seem to me to comprise a rational discussion of the issue without any ad hominem arguments or scorn for Dr. Davies. There is some discussion of his penchant for the numinous, but it’s hardly scornful.

        Your comment is gratuitous, and if you think that this kind of discussion is “vitriolic,” I suggest that you don’t frequent this site. This is precisely the place to discuss opinions and views that differ from Davies’s.

    2. The thing is though that Einstein and Hawking among many others have made similar statements and they were (of course) metaphorical. I don’t think that Davies’s references to god are any different. No doubt Spinoza would get a Templeton prize, were he alive, too.

      Maybe, the question to ask about Davies is whether he intentionally used/uses misleading theistic language in order to sell his books and or get Templeton money. I’m not sure what the answer to that is, but it would be uncharitable to just assume that was the case.

      1. No! Maybe the question is ~ should Davies have refused the award?

        I assume you objected to me writing this:- “It’s statements like that which elevated him to the high table at the Templeton trough”. Your last paragraph goes beyond what I wrote. I am claiming only that his woo got him noticed by the Templeton selection committee ~ I can’t read his mind & YOU can’t read mine.

        The Templeton Trough stands at around $70,000,000 per year** dispensed to people & organisations who will toe the Templeton line. This is distorting research ~ especially as some of the awards are going to people whose methodology is highly suspect. Recently the awards have included monies to scientists doing hard science ~ look at the list of past award winners [including Davies] who are “Honory Advisors” on the Templeton NEW FRONTIERS IN ASTRONOMY scheme. The Templeton crew realise that this era of shrinking research budgets is an unmissable opportunity to get their bullshit into every large university in the world.

        New York Times March 9th 1995
        Quoted in full with bold emphasis by me:- Scientist Wins Religion Prize Of $1 Million

        Paul Davies, a mathematical physicist who has written and lectured on connections between science and theology, has won a $1 million prize for his contributions to religious thought and inquiry.

        Dr. Davies, a professor of natural philosophy at the University of Adelaide in Australia, was named yesterday as winner of the 1995 Templeton Prize for Progress in Religion. He is the author of more than 20 books, among them “The Mind of God” (Simon & Schuster, 1992), which discusses ideas about the origin of the universe, order in nature and the nature of human consciousness.

        “Most people think that as science advances, religion retreats,” he said in a telephone interview yesterday. “But the more we discover about the world, the more we find there’s a purpose or design behind it all.”

        The Templeton prize was created in 1973 by the investor Sir John Templeton, an American-born British subject who is founder of the Templeton Growth Fund and the Templeton World Fund.

        1. I’m not objecting to what you said (and sorry to give that impression); I only meant to make the point that Davies is almost certainly an atheist despite his metaphorical language. But, I can’t escape the impression that the language he uses is pitched to attract those who are drawn to mystical interpretations of science.

          I too feel somewhat queasy about the whole Templeton thing, there’s something a bit distasteful about it, similar to Murdoch’s control of British newspapers and political lobbying by rich self interested people & organisations. I suppose, though, that if we were to condemn people for greed at the expense of scruples, then we would have to condemn the entire capitalist system, where that is mostly considered to be a virtue.

          1. This is where I think Davies is coming from [or going to]…

            JOHN BARROW who won the Templeton prize in 2006, proposed the Groucho Marx Effect paradox:- “A universe simple enough to be understood is too simple to produce a mind capable of understanding it”

            FRANK TIPLER came up with Omega Point Cosmology which puts sentient life at the centre of creation ~ proposing that at the end of the universe all matter will be one sentient mind ~ god…

            JOHN WHEELER coined the term “Participatory Anthropic Principle” (PAP), a version of a Strong Anthropic Principle. From a transcript of a radio interview on “The anthropic universe”:-

            We are participators in bringing into being not only the near and here but the far away and long ago. We are in this sense, participators in bringing about something of the universe in the distant past and if we have one explanation for what’s happening in the distant past why should we need more?

            In summary, I’m sure that Davies is not a theist & I’m broadly in sympathy with what you’ve written, but I can see no metaphor in this:- “Most people think that as science advances religion retreats, but the more we discover about the world, the more we find there’s a purpose or design behind it all.”

            His statement suggests to me that he believes there is a guiding intelligence off stage somewhere & somewhen ~ perhaps along the lines of what his fellows above have suggested. To repeat ~ Davies is NOT being metaphorical he is being purposefully vague . For various reasons he seems to need there to be a purpose [an arc through time] to the universe, but he’s very careful to be sufficiently vague about it all to avoid being shot to pieces by more grounded thinkers.

            Deepak Chopra must be laughing his socks off.

            1. If he’s not a theist, he’s very likely a deist (rather than a pan[en]theist). There is certainly some kind of supernaturalism (Sastra’s mind-first reality) in his claim.


            2. I can see where you are coming from. But Davies also (in his response to the criticism mentioned by JC in the OP) says stuff such as this:

              “I set out to see how far we can go in describing the deepest properties of the physical universe without appealing *to anything outside it* [my emphasis] — such as an unexplained transcendent god, an unexplained set of magically-imposed Platonic mathematical laws, or infinitely many unseen alternative universes.”

              That would appear to be a fully atheist agenda to me; it’s also an interesting question. But, Davies doesn’t appear to have very much to say in addressing it (I haven’t read all his books though), whereas some other physicists (Krauss, Smolin, Deutsch for instance) have ideas relating to this and don’t see the necessity of framing them using teleological language.

              1. Hmm… it doesn’t sound like a scientific agenda to me. Science shouldn’t even have an agenda, except to better understand reality. Science presumes none of those things, but would take them if it found them – and, indeed, the last falls out as a prediction of “standard” physics.

                Davies may claim to have “set out” in that direction, but his other remarks suggest that he finds that you can’t go far enough without such an appeal, and do seem to favour some kind of “Cosmic Mind”.


              2. PS. Feynman(0:32-1:04):

                People say to me, “Are you looking for the ultimate laws of physics?” No, I’m not; I’m just looking to find out more about the world, and if it turns out that there is a simple, ultimate law that explains everything, so be it, that would be very nice to discover. … When we go to investigate it, we shouldn’t pre-decide what it is we’re trying to do, except to find out more about it.

              3. @Ant. That’s all very well, but scientific theories often start out as guesses (educated guesses of course!) – and all the cutting edge physics/cosmology ideas (string theory, eternal inflation etc.) are such guesses as things currently stand. So I don’t think you can castigate Davies (and others) for making this particular guess, which may lead to some productive ideas.

                And there are some good reasons for guessing this way since, if the universe is just a brute fact, that would mean that the principle of sufficient reason must be false. That may be the case, but it’s hard to know whether the idea that there can be *in principle* no explanation is meaningful (or true if you want to be more pragmatic). It would mean, for instance, that any set of fundamental rules of physics is just as likely as any other!? (See my earlier Krauss quote, on Einstein, in this thread). I find that hard to stomach, since all the great theories of physics have been unifications and if the universe is just a brute fact it is totally unclear even in principle, why that should be the case.

                Feynman is just saying things are as they are. And that’s a good point, but it shouldn’t be used as a reason not to think about things such as quantum reality or the origins of the rules of physics.

              4. Well, sure they do, and Feynman famously said so. But some guesses are “too big” to be well formed hypotheses. And while you might say that you might make certain guesses about the origin of the universe, you certainly shouldn’t, as Davies does, rule out any (naturalistic) explanation a priori.

                If he’d said, “I don’t find [say] the multiverse to be a compelling idea, can I develop an alternative hypothesis that is simpler, better fits the data, makes more-specific testable predictions, and has more explanatory power?” that’s fine. But he was very clearly putting the cart before the horse.

                The cosmos may indeed be a brute fact, but it seems that, within the cosmos, everything that can happen does happen — including different “laws” of physics.


  24. “Now I have to say that this is either way above my pay grade, is too abstruse for the general public, or is simply wrong.”

    As a chemist, I vote for the simply wrong. Of course it’s chemistry! One big challenge is to create small self-organizing molecules which may be a precursor to larger self-organizing molecules. Once someone does that, everyone will be making their own variants of the self-assemblers. How the simple molecules make functional mistakes in replication which may lead to ever more complex molecules is another matter – perhaps there will be many possibilities. The great thing about the laboratory is that once people are on the right track, this sort of thing can be done millions of times faster in the lab than in nature.

    That crap about “information” is meaningless drivel; I doubt Davies has a decent definition of information.

  25. I wonder why we should regard information separately from the physical. Information seems inseparable from the physical!

  26. I just bought this book as a complement to coursera’s Astrobiology course that starts next week.
    I am just two chapters into it but it is an easy read and it looks really promising.
    It also helped me understand a few concepts that were not explained very clearly in the recommended text for the course: Astrobiology: A brief introduction

  27. logicophilosophicus wrote:
    Information is not a conserved quantity, but it is a real quantity.

    That information is conserved is fundamental to quantum mechanics. The wave function evolves according to a unitary operator. Therefore, information is conserved. This is what makes the black hole information paradox so interesting. It’s a fundamental inconsistency between quantum mechanics and general relativity.

    1. As has been pointed out many times, we cannot measure/detect the hypothetical wave function – we only detect the actual universe, whose decoherent states indicate specifically a loss of information. What you say “is” true of the Many Worlds multiverse, if it exists – since the sum of all these worlds includes all possible states. But that is exactly Davies point. It seems that one of two main possibilities is true: either our world is just one of an infinite number of universes where every possibility is realised, or there is only one universe after all, whose properties are so outrageously finetuned that some explanation is required. Richard Dawkins (in the biological context) noted that what looks designed may actually be “designoid” – a number of world-class theoretical physicists, including Davies, have sought “designoid” explanations for finetuning, and found it an intractable problem.

  28. logicophilosophicus
    Logicophilosphicus described entropy and information in a thermodynamic context. Another point of view is Shannon’s information theory. They’re formally equivalent, but are interpreted and applied in very different ways. That in itself is interesting.

    Shannon’s information measure is not the same as the commonplace, colloquial meaning of information. To show this, I’ll suggest a thought experiment.

    Suppose you’re presented with three hour-long videos from Netflix. One is white noise (random static), one is an episode of Downton Abbey, and one is a blank screen. The white-noise video has the most information according to Shannon’s theory. It cannot be compressed effectively. The blank-screen video has the least information (nearly zero), and can be compressed very effectively. The Downton Abbey episode is somewhere in the middle, toward the low-information, low-entropy part of the scale, but it clearly has more of what we normally think of as “information.” It can be compressed by removing redundancies and transmitted fairly efficiently, which makes video streaming work. We find meaningful those images that have low but still substantial entropy and that have an internal structure that is consistent with preexisting mental concepts, schema and expectations, whether learned or innate.

    One way to think of perception is as lossy compression.

Leave a Reply