Monday, July 28, 2014

On the mythology of natural selection. Part X: Finally: Traveling evolution's geodesics

Science recognizes Charles Darwin's contributions to knowledge because he was a deep thinker who largely transformed a whole area of human knowledge in ways that seem likely to be permanent.   One often refers to such inspirational figures by coining terms to acknowledge their views.  'Darwinism' is one.  To name a concept after him is entirely deserved (even his co-recognizer of the salient facts of life as an evolving history, Alfred Wallace, in 1889 referred to the idea with the very title of a book summarizing the field: Darwinism: An Exposition of the Theory of Natural Selection, with Some of Its Applications).

By greatly honoring his name, science and history recognize his contributions.  But his work is not to be read exegetically as a source of The Word.  He's respected, but not sanctified.  No scientist we know of  checks what Darwin said as a means of testing whether something is true or not. This is science; exegesis is for historians.

Charles Darwin would not want a halo!
Darwin would want it that way ("And get that halo off me, please!").  He was a brilliant, doggedly thorough and persistent gentleman scholar.  His works are a wonder to read, because he was a keen and incredibly patient observer and a master synthesizer of reams of data from various fields.  But he was human and lived in the times in which he lived and with the knowledge and technologies then available.  Even he, the developer of the selectionist 'law' of nature, mainly offered hand-waving assertion and circumstantial evidence, but definitely did not in the Origin of Species by Means of Natural Selection show that the origin of species was by means of natural selection.  If you doubt this, read the Origin yourself, especially Chapter IV.  But that's a topic for another day.

Nonetheless, we think that Darwin's view of natural selection as a universal determinative law to account for the evolution of the traits of organisms has become widely and without deep thought adopted as a given rather than a matter of science to be examined in the real world.  This is in many ways taken almost directly from Darwin's personal view, though most who use the term have done little if any actual reading of Darwin's voluminous works.  Such a worldview becomes dogma, or the assumption of a belief system, an ideology, rather than science, as we have tried to show in this series. That is the sense in which we discuss its mythology--the tales told and assumptions made about it that don't reflect what we actually know, and don't know.

The public and professional literature are awash in 'gene for' or 'evolved for' assertions that go largely unquestioned (thoughtful, pointed reviews of this season's recent pulp eugenics book excepted).  Such stories about genomic causation or reconstructions of evolutionary history are often expressed as if they explain traits in question, but by equating plausibility with truth in fact verge on tautology, and serve preconceptions rather than testing how the living world actually works.  The pervasive explanatory 'ether' that is invoked, is natural selection.  Most biologists probably don't quite realize how close they come to explanations that would be the same if God's will were used instead.  That universal acid or Darwin's 'dangerous idea', as one philosopher referred to it, is indeed dangerous both to human society and to science.

In this series, we have discussed various ways in which purely historical processes of proliferation and divergence from common ancestry can potentially account for aspects of the origin of complex traits within organisms, and by extension among organisms, that are not due, or not just due, to natural selection.

Our objective has not been to denigrate the idea of natural selection, but to show that there are other entirely natural processes that can lead to the evolution of complex traits.  Evolution and genomics are not just one-trick ponies.  As we have noted, how and when (or even whether) these processes are at work is a matter that can be studied by clever designs or choice of material--they are testable ideas.  In fact one of them suggests that natural evolutionary historical processes could generate complex traits without any involvement of force-like natural selection at all.  However, if one clings to a selectionist ideology, selection as an axiom rather than a scientific assertion, then of course one will never even ask the questions as to whether such ideas might be correct, much less test them.

We can, however, ask a far more fundamental question, which is how selection works when it does occur.  For this, and as elsewhere in this series, we suspend skepticism for the moment and take universal, Newtonian force-like natural selection as a given.  To try to make our point, an analogy to astrophysics might be instructive.

The 'spacetime geodesics' of evolution by natural selection
In cosmology--the real thing--travel in the universe is governed by relativistic spacetime.  An object at any time moves in a way that is determined by its motion at the time and the gravitation field--one might say the gravitational 'ecology' due to the physical objects in the area.  The bigger or closer another object, the more spacetime is curved.  But the net curvature is the total of these effects in this location in spacetime.  We use the term because on the cosmic scale, space and time are inseparably part of the nature of existence.  We can say that the past has existed and, from our vantage point, we can get a glimpse of it from the distribution of matter and energy at the present time.  In real cosmology, we can also 'see' the future to the extent that we can know and (in principle, at least) predict the state and location of the objects towards which our object is moving.  Any object in motion essentially integrates all of the forces acting upon it, and moves through spacetime accordingly.  Unless driven by some separate impulsive force, the path it follows is determined by these local spacetime properties, and is known as a geodesic, which one can think of as the shortest path between points spacetime.

SpaceTime trajectory real and imaginary (modified from GoogleImages)

If the object of our interest is affected by several small local objects in the area, it meanders geodesically past them.  If or when there is a large nearby object, it follows a more greatly affected curvature around it (blue inset).  At the end of a period of interest (which we call now), and if we knew the location of the point at our chosen starting time (then), and we're oblivious to what happened between then and now, we can always assert that what happened was that the object of our desire was shot by a specific force directly along that path (shown as dotted arrow).  

But that assertion is a fiction, and science is supposed to be about fact.  That fact is that our object got to now through a path that essentially had nothing to do with now--it was not aiming toward now, its movement at any time was locally determined, sometimes by major (one might say clear-cut) factors, often by a host of minor even seemingly trivially small ones, collectively setting up the position at the next major influence.  Note that in our cosmic analogy all of our object's meanderings are considered to be totally determined by the shape of gravitational spacetime--we assumed that no probabilities were involved.

This should be enough about spacetime geodesics in astrophysics.  The point of the analogy should hopefully be clear.   Even if the implicit complete determinism of Darwinian assumptions were true, the complex dynamic nature of earthly ecologies means that an evolutionary geodesic need not follow a retrospectively reconstructable path from then to now.  A species or trait need not have evolved 'for' its current use, not even in stages aimed in a particular direction, not with its various components evolving synchronously or even sympatrically.  Indeed, if and where ecologies are complex and dynamic, the meanderings of our object--a trait or species--may be essentially indistinguishable from random movement relative to any long-term 'purpose'.  

An organism is a collection of traits, often correlated by shared genomic mechanisms, that are continually being pulled every which-way by selection of various intensities for various reasons (the nearby gravitational objects).  These pulls are based on the local context, and that context (the 'stars' etc. in the area) change over time and space--over evolutionary 'spacetime'.  The position in the future may not be predictable from its local conditions at any given time: evolution has no 'momentum' by which the past actively propels species through the present--or if it does, we need a much-revised theory of its causal dynamics.

The shape of the head (something I've worked on with collaborators for years) is measurable in its many dimensions, and we can look at fossils and comparative species, some of whose evolutionary past history we shared through common ancestry.  But the head is not a unitary trait, perhaps not a very meaningful trait to ask about its evolutionary trajectory in selectionist terms. The length of the face, say, need not have evolved 'for' language, or reduced smell, or balance in upright posture, or change in diet, or forward-looking stereo vision, or.....  

These factors and their importance for 'fitness' need only be synchronized if they share genomic pathways, and otherwise need not be synchronous in time or space, and the 'same' trait can serve multiple or even changing functions.

As one zooms one's lens closer and closer, the picture dissolves into every more graininess. No two individuals have the identical trait or genotype. We don't see 'spandrels' or other states that set the stage for the trait as we see it today. We don't see a ladder of discrete improvements towards today. No punctuated equilibrium, saltations, and the like except as we choose to impose them on our observations. To the extent the picture given here is apt, each local point in evolutionary spacetime is so local, so dependent on its context at that instant in time, that it need not suggest the kind of longterm arrow from then to now that is so commonly explicit or explicit in our texts, papers, and so much of the rhetoric even in professional biology.

Realizing in this way how things are--or at least might be--under deterministic assumptions, it is now time to re-introduce the other complications discussed in this series, including chance in its many manifestations, at the scales of change that are relevant.  The 'geodesic' path of evolution meanders in no one direction, by no one cause.  Even if it is 'driven' by selection at every instant, it is a path whose determinants are problematic to identify in practice. Whatever is the truth, selection is not obviously simple in a force-like way. 

And here's something to think about:   Scientists often say that the simplest--the most 'parsimonious' explanation is the preferred one most likely to be true.   In the case of our cosmological analogy, I think it's right to say that the straight-arrow path from then to now is not the simplest explanation of the change!   In fact, it may be among the least simple explanations.  That's because the geodesic path is actually the simplest: it only involves the explanation that the path follows spacetime curvature (e.g., gravity). In that sense, that is the 'straight' path.  Given the complexity of space, you'd need a host of ad hoc reasons to try to account for the usual notion of a 'straight' path (like our dashed arrow).  

The same can be said of evolution.  The straightest path is that which follows the dynamic local 'pulls' of selection; to avoid following that path, as if somehow 'insisting' on a direct rather than geodesic path from then to now, it would have to resist the myriad local conditions to keep on that path--and that then becomes teleological!  Is that a suitable 'simpler' explanation?

As we navigate our search for truth, if we are not careful in our thinking we, like Odysseus, face the twin threats of Scylla and Charybdis:  the mesmerizing sirens of examples of strong recent or artificial selection that can lead us onto rocks of extrapolation to events on a slow, long time-scale, and there is a danger of being sucked into a whirlpool of complacency by taking criteria like DNA sequence conservation too casually.

Between Scylla and Charybdis; Wikimedia

Darwin's magnificent legacy
Darwin's realization that life today is the legacy of processes of descent with modification and differential proliferation, was profound and correct as far as anything we know.  (Somewhat separate from, but related to the effectiveness of selection, is that branching from common ancestry into distinct entities such as species that are isolated from each other was also part of the theory in Darwin's view, as it is of ours today.)  But none of this implies there is only one manner of differential proliferation nor that we cannot find and flesh out the basic concepts--even if there are no fundamentally different phenomena yet to be discovered.  For example, with organismal selection, genetic variation and its effects can be sorted out by the organisms in their various local areas, with no competitive differential reproduction required, and in that sense no nasty competition to the death, necessary.

There has been nothing spooky in our use of the analogy of an imaginary odyssey through evolutionary 'spacetime'.  We are not invoking heresies, that Darwin himself would not recognize. Indeed, what we have said may be imperfectly stated but is perfectly consistent with the long-standing insistence that evolution is locally contingent and not teleological.   But that in fact is inconsistent with the ubiquitous simplistic 'evolved for' and 'gene for' tales in the public as well as professional literature.

Just to be clear, with all of its nuances, including chance effects, evolution can be accounted for by historical, material processes; there is nothing that we know that requires any external force (as in religious arguments) to guide evolution. 

Just because a story is plausible does not make it either probable nor true.  Think of it this way:  The dashed arrow in our geodesic figure is the direct then-to-now path.  We can note that by comparing a fossil to its descendants we can always construct the usual simple, direct here's-how adaptation story. But not only is plausibility not the same as truth, but considering what we've tried to say in this series, perhaps such stories are least likely to be the truth!  Perhaps it's very naive to think, much less to assert, that such simplicity is what we should expect in our evolutionary tales.  Indeed, Darwin himself tacitly was led by temptation to make some similar assumptions, basically if implicitly and unintentionally teleological ones, in his books describing barnacle evolution.

What is important to us is the difference between a nuanced view, and the way people act and speak in practice (despite perhaps their being clever enough to include briefly-stated caveats).  One can always defend business-as-usual, the direct then-to-now Just-So story, by asserting that, "Yes, yes, all the details you mention are true, but they don't really matter in my example." But one should ask whether that's sincere rather than a dismissive acknowledgement-in-passing, a lazy excuse for clinging to a comfortable ideology--especially, but not only, when it comes to explaining human variation.  If so, the invocation of 'natural selection' is an exercise in mythology, in ideology.

Ideology assumes, science asks.


Holly Dunsworth said...

If I may... BRAVO!

Ken Weiss said...

Thanks, Holly! I just hope it makes sense, and the analogy is taken seriously....but not pushed too far.

Anonymous said...

Today's evolution lesson -

"The new model also predicts that information travels faster if the flock is well aligned—something else the team observed, Cavagna says. Other models don’t predict or explain that relationship. "This could be the evolutionary drive to have an ordered flock," he says, because the birds would be able to maneuver more rapidly and elude potential predators, among other things."


Ken Weiss said...

To Manoj
I assume that your point is that one can say that the quoted evolutionary explanation is just the sort of essentially empty Just-So story we wrote about. It assumes an evolutionary 'drive', among other things. Of course, it could be true, but what would the evidence be other than, if even that could be collected, that a bird not following the asserted rules didn't reproduce well.

Anonymous said...

Thanks. Yes, that was my point. Most of these researchers assume that 'evolution' is a well-proven and precise scientific theory for life and all one has to do is to think of (speculate on) the some kind of evolutionary scenario to explain their observation, molecular data notwithstanding.

And then the authors have the audacity to claim in the last sentence - "It's kind of reassuring we don't need to think about the telepathic explanation", when all he did was presented some kind of 'religious' explanation in lieu of telepathic explanation.


Ken Weiss said...

To Manoj,
Kevin Stacey pointed out a paper by Conradt et al. in the 2009 Am. Naturalist, that points out that sometimes an individual, for some reason, bolts from the regular flock behavior and becomes a 'leader' in changing the flock's direction.

It's a very standard kind of Darwinian selectionist argument about optimizing and the like, but the relevant point here, to me at least, is that every individual in the group follows the rule....except when they don't. That seems to border on tautological or untestable, even if it could be plausible.

In any case, it does all show how pliable adaptation arguments can be. Even such arguments were correctly intuited by the investigators, we need better criteria for showing not just that the explanation works today, but that it is also the true explanation for why it's that way, I think.

Sto-ology said...

Hey thought you might like this.

Ken Weiss said...

Reply to Sto-ology
Thanks for pointing this out. I have Gilbert and Epel's book and it's thoughtful as all things by Gilbert are (I don't know Epel). There are still issues to work out, if it can be done peacefully, to reconcile the kind of determinism that strong selectionists insist on with the more nuanced factors that affect genetic proliferation.