Wednesday, July 16, 2014

On the mythology of natural selection. Part II: Classical Darwinism

Darwin didn't invent the idea of natural selection as a way of adaptive advance for traits in organisms that better suited their environmental conditions. But he basically coined the term and institutionalized the view that persists to this day, often invoked in a largely unchanged way despite 150 years of biological and evolutionary research. It was a strikingly perceptive idea, that others had had in previous decades (or, in some ways, even in the classical Islamic world; see our post on Ibn Khaldun), but only in rudimentary expression and not pursued, perhaps it seemed so obvious but also because there wasn't the kind of data, such as fossils and broad biogeographic knowledge, that led the insightful Darwin to generalize it into a basic worldview.

Darwin wanted to explain a purely material, historical process by which life could have evolved its remarkable complexity from rudimentary beginnings, no divine creationism involved (except possibly at the very beginning). The idea of a 'blind' screening force has penetrating appeal to anyone wanting a purely material understanding of our world. But to accept it you also had to accept the brutal heartlessness of that world. That of course went against religious promises of better things to come, of the truth of basic moral precepts, and so on.

To Darwin, and he was completely clear and explicit about this both in his writing (primarily, On the Origin of Species) as well as his private notes and letters, natural selection was a law of Nature, a universal force. That view is no surprise, because Darwin was a product of the post-Enlightenment world of science, ushered in by the likes of Galileo and Newton, in which science used examples and data to formulate precise laws that applied to everywhere beyond those examples. That also meant that in essential ways, it was a deterministic law of cause and effect. Caveats that might modify or soften that universal law were generally uttered in passing, but not really absorbed as an important part of the 'law'.



That dogmatic universalism still characterizes much that is in print in the public media but even in the professional journals, perhaps especially in the peripherally evolutionary fields like social science including our own Anthropology, but also in very technical fields like medicine, information sciences--and even a routinely invoked metaphor for almost anything.

The reason for this is that selection is a potentially all-powerful explanation that does not require foresight or immaterial factors and that to most people there is no alternative material way to explain the origin of complex organisms. But there are such ways, and there is much--very much I think--that we simply do not yet understand even about natural selection itself.

We'll discuss these alternatives, but first it's important to explain just what classical Darwinian selection is.

The conditions for natural selection
Natural selection is about cause and effect (that's what a law of Nature is). Here is a commonly accepted, widely used, textbook way to express that law:
Natural selection means the systematic differential reproductive success of competing organisms. The idea is simple: Since a species over-reproduces so that not all individuals in the next generation can go on to successfully reproduce, and since there is variation in form among that species, and since some forms of an organism do better in a particular environment than other forms, and since the reason for this is included in their heritable genome, and since the environment remains stable long enough over time for this form to be favored persistently, and since the favorable forms are also lucky enough to produce offspring who go on to reproduce, and since they produce more offspring than their competition, then those forms can become ever more common over time at the expense of their competition. Since all these contingencies do occur, indeed co-occur, then the more prolific life form will become more suited—better adapted—to the environment in question. Since the forms are sequestered from each other by some mating barrier, then they would diverge over time, and this was the explanation Darwin and Wallace proposed for the origin as well as specialization of species.
To many this may sound entirely correct and completely obvious. But in fact, surprisingly to many perhaps, it has zero scientific content. Since the assumption is that traits that are here are here for these reasons, whatever you see in the world must be fitted, and all you can do is try to find the details of any given case. It may or may not be a true description of how things are, but it is take-it or leave-it. 

Now compare that to this version:
Natural selection means the systematic differential reproductive success of competing organisms. The idea is simple: if a species over-reproduces so that not all individuals in the next generation can go on to successfully reproduce, and if there is variation in form among that species, and if some forms of an organism do better in a particular environment than other forms, and if the reason for this is included in their heritable genome, and if the environment remains stable long enough over time for this form to be favored persistently, and if the favorable forms are also lucky enough to produce offspring who go on to reproduce, and if they produce more offspring than their competition, then those forms can become ever more common over time at the expense of their competition. If all these contingencies do occur, indeed co-occur, then the more prolific life form will become more suited—better adapted—to the environment in question. If the forms are sequestered from each other by some mating barrier, then they would diverge over time, and this was the explanation Darwin and Wallace proposed for the origin as well as specialization of species.
This version asserts cause and effect in a scientific and testable way: If a cause is present, then the effect will follow. It is an assertion about the world that is so logical that if the conditions are correct, the conclusion must follow. In a sense it defines the conditions for adaptive change. But it doesn’t assert that those conditions actually exist: determining that is where the assertions become science.

Because of this, natural selection is always possible, and sometimes clearly occurring. But that doesn't mean it is occurring in any given case. But it is typically hard if not impossible to actually prove that the 'ifs' apply to a given situation. Indeed, weirdly and ironically, to say that selection is not occurring is almost a mystical and unprovable claim. How can you ever show that there is exactly zero difference in survival or reproductive output of one version of a trait or its underlying genotype compared to another? Since each individual and its circumstances (and its genotype) is different, what kind of evidence could you possibly collect?

In an experimental setting, you can approximate this with large samples, using clones (more or less genetically identical comparison strains) and so on. Even samples from the natural world can sometimes show sufficiently clear differences in fitness between traits or genotypes. Of course, even here the comparison is judged by some sort of inherently subjective statistical criterion, such as a 'significance' test--a topic we've written about before. But in some instances, nobody really argues, when it comes to clear-cut contemporary cases.

It's much more difficult when it comes to trying to give a classical Darwinian natural selection explanation for a trait that is here today, in terms of why it got here. Here we are often in the realm of almost pure story-telling (often, and often properly, derided as Just-So story-telling). We cannot directly observe the requisite ecological, population, competitive, and other aspects of the past, even at any one time, much less over the eons of time over which most complex adaptations seem to have taken place (the idea of Darwinian gradualism).

In fact, Darwin was trying to explain how organisms change to adapt to their changing environments, but by far most selection is about how organisms don’t change! Purifying selection, that is, selection against changes in the established traits (or their underlying genotypes) is far more pervasive and easier to detect (and more convincing) than ‘positive’ adaptive selection. That’s because we clearly find that known functional parts of the genome (that is, areas like regular gene protein codes) vary within and between species by far less than areas with less clear function. Finding DNA evidence for adaptive change is much more difficult—unless, of course, you view the genes being maintained by purifying selection as having got there by adaptive selection. But that view is often if not usually an assumption.

In fact, there are ways in which traits may sometimes, often, or even mainly be determined at any given time, and over evolutionary time, that would undermine even the invocation of conservation as one of, if not the criterion for biologically relevant function. We’ve written about that, but it is beyond our topic today.

When testable becomes ideological
Many biologists are enraptured by Darwinian selection, for various reasons. One is that it provides a satisfying dogma you can invoke without restraint and by assumption rather than any serious standard of proof. Making the assumption is often justified by stating some sort of physics-based general theory of the cosmos, such as that everything biological involves energy, and if something uses more energy than its competitor its fitness will necessarily be less--if you have to eat more than your competitor to maintain your body or genotype, your competitor will have an easier path to success. But an assumption cannot be offered as an explanation of any particular case, and cannot by its very nature be questioned (that's what an assumption is).  Even invoking the laws of thermodynamics or other physics theory were correct doesn't mean the consequence is detectable or meaningful in any given case--such as how much DNA a species' genome can tolerate that has no function, or trying to give an adaptive explanation for how many hairs there are on a rabbit's body.

Even the testable version of natural selection (the ‘if’ rather than ‘since’ version) is close to a tautology because when the ‘ifs’ are true then they are essentially the same as the ‘sinces’: it’s just another way of saying what is reproductively better reproduces better. This problem is widely known to philosophers of biology, and to some biologists, but the behavior and statements of many if not most biologists (including the human geneticists whose work is hyped daily by the media) suggests they’ve rather poorly understood this.

Of course, adaptive scenarios asserting competitive natural selection may be true, and indeed many are. And selection may be the explanation of a large fraction of traits that are here today. But that is not automatically so. First, we should be checking the strength of evidence for each of the ‘ifs’ that are at least implicitly invoked in any such story.

If the story is just offered invoking Darwin’s name or natural selection, without checking that evidence seriously, then what is being stated is an ideology, not science.

Even without being in any way incompatible with the general idea of adaptive evolution, there are a number of ways in which complex traits can arise and evolve in organisms that do not involve competitive Darwinian selection—nor creationism nor implausibly blind luck. We’ll turn to those next….

No comments: