Wednesday, November 18, 2015

Intelligent Design's 2001 Space Odyssey Style Search for Intelligence of the Gaps


Bad Theology: ID's search for intelligence might have gone off into the wild black yonder; but perhaps it was right under their noses all along.

In a post on Panda’s Thumb Joe Felsenstein continues the same debate with IDists which I looked at in the following posts:

(Also relevant to the material I present below is this link:

The above series of posts are an analysis of Joe Felsenstein and Tom English’s reaction to the work of IDists William Dembski, Winston Ewert and Robert Marks (DEM for short).  Below I publish quotes from Felsenstein’s latest post and as usual interleave them with my own comments. At the start of his post Felsenstein makes it clear that…

FELSENSTEIN: The issue is not the correctness of their [DEM’s] theorems, but given that they are correct, what flows from them. Dembski, Ewert, and Marks (DEM) may object that they did not say anything about that in their paper….
We don’t think that it is a stretch to say that DEM want their audience to conclude that Design is needed.
Let’s look at what conclusions Dembski, Ewert, and Marks draw from their theorems. There is little or no discussion of this in their paper. Are they trying to persuade us that a Designer has “frontloaded” the Universe with instructions to make our present forms of life?

My Comment: I think I largely concur with that: As I constantly say on this blog, de facto ID is essentially God-of-the-Gaps (although they will deny it), or perhaps in this instance “God-of-the-frontload”. If as Felsenstein says DEM don’t discuss in their paper the origin of the information required to generate life that may be because DEM believe the major part of their epistemic task complete. The epistemic procedure of de facto ID’s “explanatory filter” prompts a default to “intelligent design” if no “natural causes” can be found. To this end DEM’s paper has the role of locating an explanatory gap which they know all too well will be filled in by their followers; as Felsenstein says: “Are they trying to persuade us that a Designer has “frontloaded” the Universe with instructions to make our present forms of life?” But the explanatory filter epistemic as formulated by Dembski and used by his ID community has its limitations, especially in theology.


FELSENSTEIN 1. Their space of “searches” includes all sorts of crazy searches that do not prefer to go to genotypes of higher fitness – most of them may prefer genotypes of lower fitness or just ignore fitness when searching. Once you require that there be genotypes that have different fitnesses, so that fitness affects survival and reproduction, you have narrowed down their “searches” to ones that have a much higher probability of finding genotypes that have higher fitness.

2. In addition, the laws of physics will mandate that small changes in genotype will usually not cause huge changes in fitness. This is true because the weakness of action at a distance means that many genes will not interact strongly with each other. So the fitness surface is smoother than a random assignment of fitnesses to genotypes. That makes it much more possible to find genotypes that have higher fitness.

In short, with their theorems, Design is not needed to explain why a reproducing organism whose genotypes have fitnesses might be able to improve its fitnesses substantially. Just having reproducing organisms, and having the laws of physics, gets an evolving system much farther than a random one of DEM’s “search

My Comment: Firstly, there is a good reason why DEM must consider the whole domain of possible searches. The point of their whole exercise is to show that whatever be the “search” (or better “process”) behind the generation of life in our cosmos, then within the full set of possible Dembskian searches it must be a very special case; that is, it is highly a-typical. The conclusion, then, is that given the principle of equal-a-priori-probabilities the cosmic search must be a highly improbable case and therefore of high information. So, I myself can see the point of this enumeration of the entire domain of possible searches. And yet as I have discussed in the previous parts of this series Felsenstein is also right; Viz: if one posits a) a differential in the fitness of possible configurations and b) our particular laws of physics which smooth out the fitness surface, then it follows that the cosmic physical regime goes a long way to providing the front loaded information needed for the generation of life. 

However, I must register here my dissatisfaction with the fitness surface model. In this post I gave reasons why this model makes implicit assumptions about the survivability and reproducibility of the organic structures that respond to the fitness surface. In consequence far more fundamental than the fitness surface is the mathematical object I call the “spongeam”. This is a conjectured fully connected but extremely tenuous sponge-like-set in configuration space. This conjectured abstract structure is defined by the requirement that it is composed entirely of (organic) forms which are stable and complex enough to survive and  replicate.  The point is that these forms need not be very fit, but nevertheless must be fit and complex enough and sufficiently connected to allow some kind of evolutionary diffusion (by replication) across the conjectured channels of the spongeam. In fact some regions of the spongeam may not even have any fitness slopes at all; the fitness could be unchanging across the spongeam in those regions. In these "flat" regions evolutionary diffusion will be unbiased although in other regions where fitness changes the diffusion will be biased. In the latter regions the idea of “sloping fitness surfaces” will apply. It follows then that “fitness” is not as fundamental as the spongeam; different levels of fitness may or may not be superimposed on the spongeam. If Felsenstein is right then it is the spongeam which is implicitly frontloaded into the cosmos via our physics. The final twist here, however, is that I don’t think the spongeam exists; therefore neither do fitness surfaces. (See the post I have already linked to)

Felsenstein quotes Dembski:

DEMBSKI: The term “evolutionary informatics” was chosen deliberately and was meant to signify that evolution, conceived as a search, requires information to be successful, in other words, to locate a target. This need for information can be demonstrated mathematically in the modeling of evolutionary processes. So, the question then becomes: Where does the information that enables evolutionary searches to be successful come from in the first place? We show that Darwinian processes at best shuffle around existing information, but can’t create it from scratch. [As it turns out this latter statement by Dembski doesn't do justice to the subject as I intend to show in due course - TVR]

I see this work as providing the theoretically most powerful ID challenge against Darwinian evolution to date. As for the attention this work has garnered, there has been some, but Darwinists are largely ignoring it. I’m justified in thinking this is because our methods leave them no loopholes. We’re not saying that evolution doesn’t happen. We’re saying that even if it happens, it requires an information source beyond the reach of conventional evolutionary mechanisms.

My Comment: The first paragraph here basically concurs with what I have just said: For evolution to successfully generate life there must be some kind of informational “frontloading” (Unless we are to accept interventional tinkering). Felsenstein is saying that this information is probably implicit in the laws of physics, laws which imply a fitness surface smooth enough for conventional evolution. Felsenstien might be right (although as I have said I personally have reservations about this conclusion). Interestingly, Dembski says We’re not saying that evolution doesn’t happen and is in effect admiting that evolution could conceivably be tapping into information from somewhere, perhaps the spongeam. So even if evolution does occur DEM’s conclusion that a practical and successful search requires a priori information still applies. And Felsenstein would agree!

You might think, then, that Dembski has got the “materialists” into a “heads I win tails you lose” impasse. But generally IDists are unwilling to exploit this advantage because lurking in the background is Western dualism, a dualism embodied in Dembski’s explanatory filter and which implicitly sets natural forces against divine intelligent design. It is therefore dangerous for IDists to even admit that evolution might be sufficiently provisioned with the requisite information (presumably via the spongeam, perhaps) to do the job; for if they do, then cranking the handle of the explanatory filter leads to an embarrassing answer. This has the effect of making “evilution” taboo in the ID community.  This is why in the second paragraph Dembski says:

I see this work as providing the theoretically most powerful ID challenge against Darwinian evolution to date.

“Yes” and “no” to that Dembski! “Yes” if you are going to depict “Darwinian evolution” as the straw man caricature of an unguided process, as do some interlocutors on both sides of the debate. And “No” if one understands, and certainly Felsenstein understands it, that the “fitness surfaces” which may be implicit in physics, provisions evolution with the requisite directional information.

Before I proceed with the next quotation I need to make the following disclaimer. I don’t accept the habitual assumption of the de facto ID community that natural processes “can’t create information” and that information only emerges from the mysterious black box of the so-called “intelligent agent”. This de-facto ID error is bound up with what is likely to be a misconception about the nature of probability. In fact my latest work (which I hope to post in due course) suggests that the creation of information is exactly the intended role of those so-called “natural processes”.....watch this space.  In an earlier post here I explore some of the complexities of the information concept which impact this matter.

Felsenstein also quotes Robert Marks

MARKS: By looking to information theory, a well-established branch of the engineering and mathematical sciences, evolutionary informatics shows that patterns we ordinarily ascribe to intelligence, when arising from an evolutionary process, must be referred to sources of information external to that process. Such sources of information may then themselves be the result of other, deeper evolutionary processes. But what enables these evolutionary processes in turn to produce such sources of information? Evolutionary informatics demonstrates a regress of information sources. At no place along the way need there be a violation of ordinary physical causality. And yet, the regress implies a fundamental incompleteness in physical causality’s ability to produce the required information. Evolutionary informatics, while falling squarely within the information sciences, thus points to the need for an ultimate information source qua intelligent designer.

My Comment: Firstly let me say that the average reasonably intelligent, yet non-technical Christian will be completely amazed and fazed by the likes of gurus like Dembski, Ewart and Marks and unable to ferret out the weaknesses in their position. It all looks oh-so-technically-expert and this in itself is heart-warming and reassuring to the average guru follower who can connect with the dualist idea that only black-box-intelligence creates information. And yet there is a deep issue with what Marks says above. Given Marks’ habituated mode of thought it doesn’t enter his head that in any practical sense of the word so called “natural processes” can create information. Instead he sees conversation of information working much like energy conservation.  From the perspective of Marks’ dualistic habits of mind it is taken for granted that physical causality is wholly different from the “intelligent designer”. To him and others in the de facto ID community the designer is the mysterious and analytically indivisible entity sourcing information at the end of his information regress. It never occurs to him to make the connection that perhaps physical causality may be that intelligence at work.

At one point Felsenstein quotes a question by ID supporter Casey Luskin.

LUSKIN: What is Active Information, and why does it point to the need for Intelligent Design to solve a problem, rather than an unguided evolutionary process? ……..Well, we appreciate the work that you [Marks] are doing and the papers that you’re publishing analyzing many of these evolutionary algorithms and asking whether they support a Darwinian view of life or an Intelligent Design view of life. (My emphasis)

My Comment: If the spongeam and the fitness surfaces which ride on its back exist, as Felsenstein thinks they do, then "Darwinism" is certainly not unguided!  DEM’s work in fact shows that conventional evolution cannot be unguided. It is ironic that the B-teams on both sides of the debate err on this notion of unguided evolution  - see here for example.

Felsenstein quotes Ewart:

EWART: While some processes are biased towards birds, many others are biased towards other configurations of matter. In fact, a configuration biased towards producing birds is at least as improbable as birds themselves, possibly more so.

Having postulated Darwinian evolution, the improbability of birds hasn’t gone away; we’ve merely switched focus to the improbability of the process that produced birds. Instead of having to explain the configuration of a bird, we have to explain the configuration of a bird-making process.


My Comment: This is certainly true and this is what DEM have successfully shown. And yet there is a deep implicit issue embedded in Ewart’s statements as to the significance of his claims. It is on that significance which the de facto ID movement is going astray.  The ulterior motive behind the above, a motive which is clear to Felsenstein and myself, is that Ewart thinks he is paving the way for the explanatory filter to default us to the “intelligent agent”, whatever he means by that. The big problem, as I will be proposing in my latest work, is that intelligence too classifies as a highly improbable configuration and this fact points to a major loophole in the work of the ID gurus.

How does Felsenstein react to Ewart’s statements?....

FELSENSTEIN: This example leaves it unclear what the “process” is. The reader may be tempted to conclude that it is the process that models an evolving population. And then the reader may think that if this evolutionary process succeeds in improving fitness, that some outside force is needed to set up the process so that it succeeds. But for their theorem to apply, the processes considered must include processes that make no sense as models of evolution. Processes that wander around among genotypes randomly, without being more likely to come up with higher fitnesses. Even processes that prefer to find genotypes with lower fitnesses. All of those are among the processes that must be eliminated before we get to processes in which genotypes have fitnesses, and those fitnesses affect the outcome of evolution.

My Comment: As I have already said DEM have rightly included all the possible searches in their enumeration and that includes all the highly disordered searches which practically speaking are fruitless. Disorder, by definition, has an overwhelming statistical weight and therefore a successful evolutionary search is a very rare case when set against the class of disordered searches.  Using the principle of equal a priori probabilities it follows, then, that a practical evolutionary process is a highly improbable search and this by definition implies a high information object. But then Felsenstein is also right; he presents a good prima facia case that physics implies different levels of fitness and a smooth fitness surface, which is where the information required by DEM lies according to Felsenstein. (Although I must once again register my reservations about the existence of the spongeam on which the existence of the fitness surface depends,)

Felsenstein further comments on Ewart:

FELSENSTEIN: In his reply, Ewert invokes the smoothness of the fitness landscape, and considers the smoothness to result from “laws or self-organization”

(EWART): *Quote* It is not sufficient to invoke the three-fold incantation of selection, replication, and mutation. You must also assume a suitable fitness landscape. You have to appeal to something beyond Darwinism, such as laws or self-organization, to account for a useful fitness landscape. *Unquote*

He does not seem to realize that those “laws” might simply be the laws of physics, and that the “self-organization” can simply be self-reproduction, something that all organisms do.

My Comment: Although DEM are right in asserting that any working conventional evolutionary process must have, a priori, a high information content, it is notable that Ewert doesn’t acknowledge that conceivably this information could, as Joe Felsenstein plausibly maintains, reside in the rarity of our familiar physical regime. One might think that by admitting this as a possibility at least, the IDists could have their cake and eat it; they could even claim that Felsenstein is admitting the existence of “active information”!  But no, the de facto ID movement has painted itself into a corner here: For if something along the lines Felsenstein is suggesting could be satisfactorily demonstrated then not only would that bugbear of ID, the explanatory filter, stab IDists in the back, but the whole thrust of IDism, which has been unequivocally against any hint of “Darwinism”, would make it look as though they have been defeated. The IDists have fostered the fearsome dualist spectre that if those loathed “natural forces” are doing the creation job all along then ID is worsted. Notice Ewert’s reference to so-called “self-organization”, a vague concept which has yet to produce any substantive input into the evolution debate.  And yet if Felsenstein is right, the solution could be staring the IDists in the face; namely, that if the requisite fitness surfaces are implicit in physics then in effect common-or-garden physics is doing the job of “self-organisation”. (Assuming the existence of the spongeam I must add). Even though this outcome to the debate would still be consistent with the work of DEM, such an outcome would cloud the tribal-clarity of the IDists shrill anti-Darwinist rallying call, a call which appeals to the dualist thinking of every Christian sect between here and the Watchtower’s Brooklyn HQ.


FELSENSTEIN: It is clear from these examples that Dembski and Ewert mean their theorems to be read as evidence for an Intelligent Designer either frontloading the evolutionary process, or for an Intelligent Designer intervening in it. But Tom English and I have shown that their Active Information can come about without that. It can come about simply by having a reproducing organism which has different genotypes, which have different phenotypes, and these have different fitnesses. And further Active Information can also come about by the predisposition of the laws of physics to bring about fitness surfaces smoother than “white noise” fitness surfaces.

Could that Active Information be enough to explain the evolution of, say, a bird? Do they have some argument that further “configuration of a bird-making process” is needed beyond that? There is actually nothing in their argument that requires that there be further Intelligent Design

My Comment: Yes, given the work of DEM it does follow that the generation of life demands frontloading; either that or the ad hoc fiat of tinkering and intervention. Felsenstein is plausibly maintaining that the mutually acceptable frontloading is down to physics.  Essentially then DEM and Felsenstein aren’t at odds; for they both see the need for some kind of frontloading (If not interventional tinkering). But they disagree over the significance and meaning of this fact.  What makes the situation more complex is that for reasons I’ve already outlined the IDists are unwilling to admit that this frontloading could be down to common-or-garden physics; they’d much prefer their opponents to talk of some exotic and speculative “self-organization”, a vague idea which currently has little real intellectual traction. But for polemical reasons the IDists are unwilling to entertain the prosaic physics solution; a choice imposed on them by the dualism implicit in the explanatory filter which excludes any middle ground in their intelligence versus natural forces dichotomy. The IDists have committed themselves to the idea that some special ingredient X is needed for life to exist. Felsentstein says that that ingredient could well be the physics we all know and love. The IDist is inclined to say “no!” to that because otherwise it would cut across his anti-darwinist raison d'etre he has fashioned for himself. For the IDist ingredient X is likely to be thought of as some inaccessible “black box” intelligence and not mere prosaic physics; to admit the latter would be a terrible anti-climax to the de facto IDists' 2001 space odyssey message; namely, that they have found an artifact not created by common-or-garden “natural forces”;  this is in spite of the fact that even if Felsenstein is right the IDist still have a case to argue!

The ID community's loathing of "Darwinism", even if it actually doesn't directly cut across DEM's ideas, nevertheless, goes deep enough to cause division within the ID community. See for example this post by Vincent Torley on Uncommon Descent where in the comments section Torley is accused of supporting “Darwinism”. See comment 86 where we read: It almost seems as if VJ Torley is turning Darwinist on us. Someone please correct me if I’m wrong. Regardless of whether it is consistent with conservation of information ideas or not as a rule the average right of centre ID follower hates "Darwinism" and can not abide by it.

The added irony is that Felsenstein himself takes for granted the same dichotomy of intelligence vs. natural forces. Given his outlook on life it is likely, of course, that he believes “natural forces”, whatever that means, have done the job of evolving life.  Since he has shown (plausibly) that physics could be the seat of so-called active information, then his conclusion, as per the explanatory filter, is that intelligent agency is not required as an explanation. He, like his IDist antagonists, sees it as a straight choice between natural forces and God. Felsenstein is a dualist in his conceptual categories when it come to thinking about God. As Felsenstein says above: "But Tom English and I have shown that their Active Information can come about without that", and by "that" Felsenstein means an "Intelligent agent". For him physics trumps intelligence.



Epilogue

Even if Felsenstien is winning the argument this still leaves us with the question of Why our particular physical regime with its miraculous fine tuning? For IDists, of course, this is the work of the God of the Gaps, but for Felsentstein it’s probably the work of Physics of the Gaps, perhaps some kind of multiverse. But whatever way we look at it, finite chains of human logic will always leave an inevitable grand logical hiatus unfilled. The irrational arbitrariness of an impenetrable wall of brute fact contingency faces us at the end of our quest for obliging reason; positing neither physics nor intelligence will rid us of this super gap (But see appendix).  Therefore I suggest we leave it and get back to the thing we do best and that is to describe the cosmos we have been provisioned with using the intellectual tools the good Lord has also provisioned us with. 

From where I’m standing the results of Dembski, Ewart and Marks are starting to look like a misinterpreted mathematical trivialism, I hope to expand on this topic in later posts. What ID is missing is that those much despised so-called “natural processes” are actually provisioned, in any practical sense of the word, to do exactly what IDists dread and fear in their darkest dreams; namely to create information. But then why should a Christian be surprised at that? God is immanent in his world.



Appendix (Added 21 Nov)
Is there any hope that the finite human mind could ever grasp the concept of Aseity? Two lines of inquiry respectively from the atheist and theist camps might be as follows:

Atheistic Aseity: This line or argumentation might be based on some kind of super-copernicanism; that is, the super-multiverse where all options are somehow realised, an idea having its strongest form in Max Tegmark’s mathematical universe. Because everything exists in the super-verse then it follows that everything has an existence probability of unity. The Shannon “suprisal value”, that is, the information value of the existence of any particular state of affairs then sinks to zero. Since the human intellectual demand for explanation comes in large part from our intuitive sense of surprise as to why particular things are as they are, then it may be argued that super-copernicanism goes someway to assuaging our sense of surprise at apparent contingencies;  for in the super-verse nothing conceivable is given preferential existential treatment; the only surprise left is why there is something rather than nothing. But it might be argued that if everything exists it is no more surprising than everything failing to exist at all!

Regarding the epistemic question as to why human beings can know anything at all in such an indifferent and dispassionate universe it might be argued that in a universe of indifference we aren’t going to be specially targeted for deception; hence errors average out and we can be reasonably sure that we can acquire knowledge about somethings if not everything. To claim that we could know nothing in an impersonal universe is tantamount to the inverted conceit of the conspiracy theorists who believe that they are being specially targeted for deception. One thing to be said for Copernicanism is that it seems to be an antidote to the narcissism of fundamentalist paranoia!

However there are problems with this view: Namely, the simulation argument and why we know as much as we do; we would expect the universe to be far more random and unknowable if some form of super-copernicanism held sway.

Theistic Aseity: This line of thought is potentially much more fruitful to my mind. Early on in my intellectual career I was attracted to positivism; the general idea that everything swings on observer experience to a high degree; in fact strong positivism suggests that all else besides experience is meaningless. Strong positivism is counter intuitive when it comes to in-practice and in-principle realities that cannot be experienced like the planets of distant galaxies or other minds. But nevertheless positivism has left me with the general feeling that without the presence of an experiencing sentience to apprehend it in some way “reality” is a meaningless and incoherent idea. This view is clearly related to Berkeley’s idealism. So, if reality is meaningless without a sentient apprehender then the organised high complexity of the cosmos immediately follows: The experiencing sentience has to be sufficiently complex in order to possess the coherence needed to cognitively apprehend the cosmos, But since coherent human observers are composed of the very stuff of the cosmos, then it follows that the cosmos must be sufficiently organised and complex to support the human sentience that apprehends it. When human's describe the cosmos they are in effect describing  themselves. I advance a related idea in the introduction of my book Gravity and Quantum Non-Linearity. Viz; that conscious sentience is described in its own terms, much like a computer language compiler is written in the language it compiles.

The foregoing line of thought is essentially the strong anthropic principle. It attempts to show that sentient observers are logically necessary because a cosmos without them is regarded as an unintelligible  notion. These prototype ideas on the aseity of sentience may throw light on the aseity of God.

  Atheistic visions of the cosmos which are founded on the elementary elemental such as bits and particles will always face a logical hiatus: Simplicity is simply too simple to self-explain. (I touch on this idea of elementary elementalism being unable to self-explain in the following posts:
http://quantumnonlinearity.blogspot.co.uk/2012/12/paul-nelson-computer-simulations-and.html)


Appendix II

Without the spongeam conventional evolution is a non-starter

9 comments:

Joe Felsenstein said...

I am glad to see that you concur with most of what Tom and I have written on the Dembski-Ewert-Marks argument. If it boils down to arguing about where the laws of physics come from, I will let others do that. For most scientists, it will be sufficient to establish that with our present laws of physics, here in this universe, evolution can work and natural selection can bring about remarkable adaptations.

I do not expect to be involved in discussions of spongeam or the aseity. Especially since I have no clue what those words mean.

Joe Felsenstein said...

OK, rereading, I see that "spongeam" is your name for sponge-like connected sets of highly-fit genotypes. One could also describe that as a set of moderately- to highly-fit genotypes on a fitness surface. (Yes, I agree with you that the fitness surface model has great limits, but at least it enables discussion).

Timothy V Reeves said...

Hi Joe,

Thanks for the comments and reading through that carefully.

The concept of the spongeam arises as follows.

We take any replicating organic configuration (for simplicity I assume the environment and physics are both fixed) and then imagine amending it in an arbitrary way in a series of increments. It is very likely that even given the constraints of physics by far and away the majority of these incremental paths through configuration space will lead to complete non-viability; that is, to configurations not fit enough to even replicate once. Perhaps, however, (and this is a very big “perhaps”) some of these paths will have viable organic configurations along the whole of their length and link very diverse organisms. Now, if I superimpose all these paths of viablity (assuming they exist) on configuration space then in my mind’s eye I see something not unlike the picture I’ve just published at the end of this blog post - a kind of spongey structure, interspaced with big black voids of non-viability. The reason why I don’t talk of fitness surfaces is because:

a) It conjures up a picture of a continuous surface when it fact it should be surface (or better “volume”) full of holes of non-viability.

b) The spongeam concept doesn’t commit to the idea of fitness slopes: It is conceivable that in some parts of the spongeam fitness doesn’t vary; hence my talk of unbiased as well as biased diffusion across the spongeam.

Without the spongeam (That is, without the class of viable organic configurations forming a fully connected set in configuration space) then conventional evolution, which depends on a kind of abstract diffusion motion, wouldn’t work. Therefore, a very pertinent question is Does the spongeam exist?. Frankly, I have my doubts, although proving it either way may be one of Wolfram’s computationally irreducible tasks: That is, to get an analytical handle on the question there is no choice but to interrogate a very literal biological “simulation”… in fact just as is being done in real biology when biologists and palaeontologists gather evidence for biological change.

Timothy V Reeves said...

But then think of this: Given the overwhelming statistical weight of disorder against order (even the complex order of biological configurations) then one would expect those black voids of the spongeam to be overwhelmingly large compared to volume of the thin network of spongeam fibrils. Asking real biology to form a fully connected set across configuration space has the touch and feel of being asked to network together all the stars in the visible universe given a few pounds of copper; since matter is discrete there will come a point when we will run out of atoms to complete the task. My guess is that this is what it is like trying to populate configuration space with thin strands of viability and survivability. Hence, I doubt the spongeam exists.


True, this is merely a gut feeling argument and not a proof but it sets the scene for my little thought experiment which starts by asking: What if the spongeam doesn’t exist? What then? How would one cope computationally with such a scenario and solve the problem of life? What would physics have to look like given that it must be capable of generating life without a spongeam? I’m not expecting anyone to follow me down this (garden?) path of speculation. I call this journey Melencolia I and The Thinknet Project and I am publishing it in stages on this blog as and when I make progress. Nice little hobby.

Joe Felsenstein said...

Certainly you are right about there being far more places in genotype space that are inviable than that are viable. (You seem to actually be describing phenotype space, but I am thinking of all possible genotypes).

The issue is how the viable points are arranged. If they are in big clumps, with large empty spaces between, we could evolve a lot within a clump and never cross the lacunae between. You instead think that instead of big connected sets we have a sponge-like bunch of fibrils with smaller holes between them. I not going to try to prove one or the other, just wanted to see whether I was understanding your argument.

Timothy V Reeves said...

If genotype implies phenotype then I suppose the configuration space could be discussed entirely in terms of a map to genotype.

Yes, conceivably there could be a big isolated clump of viability within which conventional evolution would be confined. Now, I’m not a biologist so I have to think rather abstractly and in terms that are general, perhaps a little vague and a bit broad brush. But here’s my take for what it’s worth, so bear with me.

If we take the extremes of the clump then let us assume that a minimum of m changes are needed to get from, say, a simple replicator to some multi-cellular organism. So roughly speaking m is a measure of the linear dimension of the clump. Let us now take some arbitrary organism within the clump and make in-the-order-of m arbitrary changes to it. At this point my intuitions about the relative statistical weights of order as against disorder suggest to me that the overwhelming number of ways m changes can be made are likely to lead into voids of non-viability. Therefore I expect the “clump” to have huge regions of non-viability; hence the spongeam. If this has been established we are then faced with the next question: Viz: Is the total number of viable structures enough to connect all the organisms in the conjected “clump”, which is in fact now to be thought of as a spongey clump! As I have said, I have my doubts whether this question is answered in the affirmative.

That’s probably all a bit hand-wavy but it’s currently the way the subject has assembled itself in my mind.

Jumping ahead a bit with some more hand waving: If there are completely disconnected regions of viability in the “clump” then it is here where I anticipate some variant of quantum mechanics doing its stuff of leaping the gaps. I don’t how it’s doing it at the moment, but something looking a lot like QM would be needed. In fact I’m reminded of a TV program by Jim Al-Khalili where he talked about the way excitons work in photosynthesis to transfer energy…apparently the wave function doesn’t collapse until destinations came into range making the whole “random walk” process much faster. That idea is currently waiting back stage with me at the moment; if it could be scaled up, conceived more precisely and made to work then our problems with evolution would be over. Basically I see evolution being able to look ahead. Ergo, it’s an expression of intelligence at work. Heretical? Too right! But then I’m an amateur, so I’ve got absolutely nothing to lose by being eccentrically amateurish! I’m completely free to be the fool that rushes in…. No strict terms of reference holding me back! Haha!

Joe Felsenstein said...

Suppose that the set of viable genotypes is not a "spongeam" but is a long noodle. Then there could be a path from one end of the noodle to the other, m steps long, that passed through viable genotypes. At the same time random walks with m steps would be overwhelmingly likely to wander off of the noodle into regions of genotype space that had no viability.

Based on examples like these, I don't think that the disastrous result of these random walks is itself sufficient evidence for a "spongeam".

Timothy V Reeves said...

I’m not sure I understand your point there Joe: What I’m actually saying is that the existence of the spongeam is unlikely! Conventional evolution requires that the set of viable replicators form a connected set in configuration space. Given that this set is likely to be such a tiny proportion of configuration space my intuitions tell me that populating this space with a density sufficient for the set to fully connect is problematical. That is, I’m saying the spongeam probably doesn’t exist!

Your single-track “noodle” is much less demanding on the relatively small number of viable replicators. But I take it you appreciate that the branching pattern which is the tree of life requires some kind of network of replicator connectivity in configuration space. So I assume you introduce the noodle for the sake of making a point.

But if for arguments sake we assume the existence of the noodle I have a question. Configuration space has n dimensions. If a replicator steps off the one-dimensional noodle its line becomes extinct. There are 2n ways of stepping off at every point on the noodle to 2 ways of keeping on it. So if we set a replicator going at one end of the noodle I wonder how many would arrive at the far end? Thinking about it, it might actually work, because replication is an exponentially increasing process and this may offset the exponential decay caused by replicators stepping off and dying out. Moreover, if the noodle branches just a few times it will give us a tree......might be an idea there! Question is, though, can we prove the existence of this branching pattern from first principles?

Timothy V Reeves said...

As I think I have already said the sheer complexity and variety of biological objects along with their environments probably means that apart from some toy-town-models we are unlikely to ever get a full blown first principles theoretical handle on the question of the pattern real replicators form in configuration space. So, taking a leaf out of your book and limiting my terms of reference I’m prepared to leave this matter to biologists and their necessarily empirical approach to the subject: If they feel there is enough observational evidence to back up the idea that there exists a class of replicators forming a connected network in configuration space big enough to support standard evolution then I can run with that.

But in the meantime suffer me my little speculative thought experiment which entails positing a disconnected set of replicators and then asking myself what would the necessary physics look like needed to support evolution? I don’t think anyone need lose any sleep over this harmless hobby!