From Bacteria to Bach and Back

The basic framework for understanding evolution by natural selection was provided by Darwin in Origin of Species in his summary to chapter 4. It is worth repeating.

If during the long course of ages and under varying conditions of life, organic beings vary at all in the several parts of their organisation, and I think this cannot be disputed; if there be, owing to the high geometrical powers of increase of each species, at some age, season, or year, a severe struggle for life, and this certainly cannot be disputed; then, considering the infinite complexity of the relations of all organic beings to each other and to their conditions of existence, causing an infinite diversity in structure, constitution, and habits, to be advantageous to them, I think it would be a most extraordinary fact if no variation ever had occurred useful to each being’s own welfare, in the same way as so many variations have occurred useful to man. But if variations useful to any organic being do occur, assuredly individuals thus characterised will have the best chance of being preserved in the struggle for life; and from the strong principle of inheritance they will tend to produce offspring similarly characterised. This principle of preservation, I have called, for the sake of brevity, Natural Selection.

Over the years, this great insight has been refined and generalized, and various compact formulations have been composed. The best, for simplicity, generality, and clarity is probably philosopher of biology Peter Godfrey-Smith’s (2007) trio:

Evolution by natural selection is change in a population due to

(i) variation in the characteristics of members of the population,

(ii) which causes different rates of reproduction, and

(iii) which is heritable.

Whenever all three factors are present, evolution by natural selection is the inevitable result, whether the population is organisms, viruses, computer programs, words, or some other variety of things that generate copies of themselves one way or another. We can put it anachronistically by saying that Darwin discovered the fundamental algorithm of evolution by natural selection, an abstract structure that can be implemented or “realized” in different materials or media.

We will be looking at the evolution by natural selection of culture—cultural items, such as words, methods, rituals, styles—and to aid our imaginations in thinking about this treacherous topic, I want to introduce a thinking tool that I find particularly valuable for orienting ourselves in this tricky terrain: the Darwinian Spaces invented by Godfrey-Smith in his book Darwinian Populations and Natural Selection (2009).

It is tempting to say that the three conditions above define the essence of Darwinian natural selection, and the classic examples of natural selection used to illustrate the theory in action are perfect fits. But one of Darwin’s most important contributions to thought was his denial of essentialism, the ancient philosophical doctrine that claimed that for each type of thing, each natural kind, there is an essence, a set of necessary and sufficient properties for being that kind of thing. Darwin showed that different species are historically connected by a chain of variations that differed so gradually that there was simply no principled way of drawing a line and saying (for instance) dinosaurs to the left, birds to the right. What then about Darwinism itself? Does it have an essence, or is it kin to variations that blend imperceptibly into non-Darwinian explanations? What about phenomena that are not quite paradigm cases of Darwinian evolution? There are many, and Godfrey-Smith shows how to organize them so that their differences and similarities stand out clearly and even point to their own explanations.

For instance, all evolving entities need to “inherit” something—a copy of something—from their parents, but some varieties of copying are low fidelity, with lots of distortion or loss, and others produce near-perfect replicas. We can imagine lining up the different cases in order, from low-fidelity copying to high-fidelity copying, on the x-axis of a simple graph:

Evolution depends on the existence of high-fidelity copying but not perfect copying, since mutations (copying errors) are the ultimate source of all novelty. Digital copying technology is perfect, for all practical purposes: if you copy a copy of a copy of a copy … of a Word file, it will be letter-for-letter identical to the original file. Don’t expect mutations to accumulate, for better or for worse. DNA copies itself almost perfectly, but without its very occasional errors (not one in a billion nucleotides), evolution would grind to a halt.

Here’s another example of variation: the differences in fitness between members of a population may depend on “luck” or “talent” or any combination in between. In a selective environment where most of those who don’t reproduce are struck by lightning before they get an opportunity to mate, evolution cannot act on whatever differences there are in prowess. We could line up varying mixtures of luck and talent on another axis and note that just as copying “noise” (low-fidelity replication) hinders evolutionary R&D, so does environmental “noise” (lightning strikes and other accidents that remove otherwise worthy candidates from the reproduction tournament). While we’re at it, we could also add a third dimension and draw a cube with x, y, and z axes, on which we can locate various phenomena. (Unfortunately, most of us can’t visualize four or more dimensions gracefully, so we have to stop with three, and look at only three kinds of variation at once, swapping them in and out of the diagram to suit our purposes.)

Using these three-dimensional arrays, we can display pure Darwinian phenomena, quasi-Darwinian phenomena, proto-Darwinian phenomena, and (beyond the fuzzy boundaries) phenomena that aren’t at all Darwinian. Being able to see at a glance how these phenomena resemble each other and differ from each other is a fine aid to thinking about evolution. Instead of trying to draw bright lines that separate mere pseudo-Darwinian phenomena from phenomena that exhibit all three of the “essential” features of natural selection, we contrive gradients on which we can place things that are sorta Darwinian in some regard or another. Then we can see whether Darwinian theory applies to the intermediate cases, looking for trade-offs and interactions that can explain how and why these phenomena occur as they do. Godfrey-Smith reminds us that “evolutionary processes are themselves evolutionary products” and as a result emerge gradually and transform gradually. In short, his Darwinian Spaces give us a valuable tool to help us maintain and elaborate our “Darwinism about Darwinism,” to use Glenn Adelson’s excellent phrase.

In all these diagrams, the dimensions are treated as taking values between 0 (not Darwinian at all) and 1 (maximally Darwinian), so the paradigmatic, “most Darwinian” phenomena appear in the upper right corner, maxing out the three dimensions (1,1,1), and all phenomena that utterly lack the features represented belong at (0,0,0). We can choose any three variable features we like and assign them to the three dimensions, x, y, and z. In figure 7.1, H, fidelity of heredity, is mapped from left to right on the x dimension, and if fidelity is too low to maintain signal in the presence of noise, evolution can’t happen at all; useful mutations may arise but disappear in the noise before selection can drive them into succeeding generations. Even if the other dimensions shown are at good Darwinian values (close to 1), we get the “error catastrophes” (see figure 7.1) that line up near the left wall. (And as we just noted, if fidelity is perfect, evolution stalls as well, for lack of new variation, so the paradigm cases should be very near, but not at, the right wall.) The vertical y dimension represents continuity or “smoothness of fitness landscape.” Natural selection is a gradual process; it depends on blindly taking “small steps,” and if the selective environment is organized so that small steps occur on a steady or smooth slope (so that a small step sideways in any direction is also a small step up or down or neutral in fitness), then a sequence of small steps can do “hill-climbing” and arrive, in spite of its myopia (the blind watchmaker), at a peak. It will arrive at the only peak, the global optimum, in a particularly smooth fitness landscape, and at a local optimum in a landscape with several peaks. When the landscape is “rugged,” evolution is next to impossible since small steps are uncorrelated with progress or even maintaining one’s current fitness.

FIGURE 7.1: Darwinian space. © Godfrey-Smith.

On the z dimension, Godfrey-Smith has placed S, standing (don’t ask why “S”) for dependence on “intrinsic properties,” and this feature captures the luck-versus-talent dimension. In a tennis tournament, the better athletes tend to advance, thanks to their competence and strength; in a coin-tossing tournament mere luck, not any intrinsic property of the contestants, determines the outcome, and the winner is no more likely to beat any of the losers in a rematch than any other contestant. For example, in genetic “drift” (especially in small populations where the “sampling error” will be relatively large) the “winning” feature in a population is no better than the alternatives; luck has just happened to accumulate in one direction, boosting one feature into fixation in that population. For instance, in a small population of rabbits a few of which are darker gray than the others, it just happens that some of the darker ones get wiped out as road kill, another drowns, and the last carrier of the gene for dark gray falls off a cliff on its way to mating, eliminating the gene from the local gene pool, but not for cause.

One of the most valuable features of these Darwinian Spaces is that they help us see the phenomenon of “de-Darwinizing,” in which a lineage that evolved for generations under paradigmatic Darwinian conditions moves into a new environment where its future comes to be determined by a less Darwinian process. Human cells, shown in the diagram, are a fine example. These cells are all direct but very distant descendants of single-celled eukaryotes (“microbes” or “protists”) that fended for themselves as independent individuals. These distant ancestors were paradigmatic Darwinian evolvers (up in the 1,1,1 corner) driven by all three features, full strength. The cells that currently make up your body are the direct descendants (the “daughter cells”) of cells that are themselves descendants of earlier cells, going back to the zygote that formed when ovum and sperm were united at your conception. During development, in your mother’s womb and in infancy, there is a proliferation of cells, many more than will be needed ultimately to compose your organs, and a selection process ruthlessly culls the extras, leaving in place the winners who get the various jobs, while the losers are recycled to make raw materials for the next “generation.” This is particularly clear in the brain, where many brand new neurons are given the opportunity to wire themselves up usefully (e.g., as part of a path from some point on one of your retinas to some corresponding point in your visual cortex). It’s rather like a race, with many neurons trying to grow from A to B, following molecular clues like a trail of breadcrumbs. Those that get there first win, and survive, while the rest die and become feedstock for the next wave.

How do the winners “know” how to grow from A to B? They don’t; it’s luck (like winning a coin toss). Many neurons attempt the growth journey and most fail; those that make the right connection are saved. The cells that survive may be “intrinsically” just like their competitors (no stronger, no faster); they just happen to be in the right place at the right time. So the developmental process that wires up your brain is a de-Darwinized version of the process that evolved the eukaryotes that joined forces a billion years ago to form multicellular creatures. Notice that genetic drift, in which a feature goes to fixation as a result of sheer luck (like the human cells), is also shown in the diagram as low on hill climbing; human cells are selected for location, location, location, as the real estate agents say, but the winners in genetic drift are just lucky. Genetic drift has been around as long as evolution has, so it is not a case of de-Darwinizing.

Depicted in figure 7.2 is another of Godfrey-Smith’s Darwinian Spaces. This time the x dimension is B for bottleneck: Does reproduction funnel through a narrowing of one kind or another—in the extreme case, a single cell? A single sperm cell and a single ovum, out of all the several trillion cells that composed your father and mother at the time of your conception, united to inaugurate your generation. In contrast, a buffalo herd can become two herds (which can later become four herds) by simply dividing in two, or a pair of buffalo may get isolated somehow and become the founders of a new herd, much less genetically diverse. Any part of a sponge can also become a descendant sponge, but a sponge is somewhat more integrated—more an independent thing—than a herd is. An oak tree is highly integrated and has a reproductive bottleneck—the acorn—but like plants in general can also be propagated from a “scion” twig planted in the ground. There is no way (yet) to make a descendant human being—a clone—from a sliver of ear or toe. We are high on all three dimensions shown. Aspen groves are interesting since they are all connected underground by root systems so that they are not genetically individuals but more like conjoined identical mega-twins, a single “plant” covering a whole hilltop.

FIGURE 7.2: Darwinian space with other dimensions. © Godfrey-Smith.

Figure 7.3 is an application of Darwinian Spaces to the origin of life, as discussed in the first three chapters. This must be a set of processes that takes us from pre-Darwinian to proto-Darwinian to Darwinian.

When we get to bacteria, in the upper right corner, we have full-fledged reproduction, energy capture, and lots of complexity. The unanswered question concerns what route might have been taken to get there. We could plot the trajectories of the components (membrane, metabolism, replicator mechanisms, …) and see which joined up first and which were late additions or refinements. Other dimensions could be plotted, of course; two important dimensions might turn out to be size and efficiency. As noted in chapter 2, perhaps the first things/collections that (sorta) reproduced were relatively huge Rube-Goldberg contraptions that were then stable enough for natural selection to prune or streamline into highly efficient, compact bacteria. Populating the space with examples would be premature at this point, and even my placement of bacteria in the upper right corner might turn out to be misleading if archaea (or something else) turn out to have been the first clearly Darwinian reproducers.

FIGURE 7.3: Darwinian space for origin of life.

Cultural evolution: inverting a Darwinian Space

Now let’s consider an application of Darwinian Spaces to the phenomena of cultural evolution, leaving the details of explanation and defense for later.

On the x axis in figure 7.4 I put growth versus reproduction. Aspen groves grow larger “bodies” instead of having “offspring” (usually). It’s sorta “just growth” and sorta reproduction. Mushrooms are another example, but I don’t know of any animals that can double themselves, Siamese twin style, instead of reproducing the usual way. On the z axis I put degree of internal complexity, and on the vertical y axis is cultural versus genetic evolution. (Are there phenomena midway between? Yes, as we shall see.) Slime molds, like aspen groves, mix differential reproduction with differential growth—getting larger and larger instead of breaking up into individuated descendants. In culture the Roman Catholic Church grows and grows (until recently) but seldom spawns descendants these days (though it had some spectacular offspring in the sixteenth century). In contrast, the Hutterites are designed to send off daughter communities whenever a community gets big enough (for the evolutionary details, see Wilson and Sober 1995). Religions (or religious communities) are large, complex social entities. Words are more like viruses, simpler, unliving, and dependent on their hosts for reproduction. (Don’t make the common mistake of thinking that all viruses are bad; for every toxic virus there are millions in us that are apparently harmless, and some may be helpful or even essential.)

FIGURE 7.4: Darwinian space with religions.

Trust is (mainly) a cultural phenomenon—it no doubt gets a boost from our genes—and, like the air we breathe, is not very thinglike until we run out of it. It is (largely) a product of cultural evolution, just the way the air we breathe is a product of genetic evolution over billions of years. When life began, it was anaerobic (it didn’t require oxygen), and the atmosphere was almost oxygen-free, but once photosynthesis evolved, living things began pumping oxygen (in the form of CO₂ and O₂) into the atmosphere. This took several billion years, and some of the O₂ in the upper atmosphere was turned into O₃, or ozone, and without it, deadly radiation would reach the Earth’s surface and make our kind of life impossible. The oxygen level 600 million years ago was only 10% of its current level, so although the change is imperceptibly slow, it is dramatic over time. We can consider the atmosphere as a fixed feature of the selective environment in which evolution takes place, but it is actually also evolving and a product of earlier evolution on the planet. These phenomena—we might call them the biological and cultural “atmosphere” in which evolution can occur—don’t themselves reproduce, but they wax and wane locally and evolve over time in a non-Darwinian way.

Figure 7.5 is another preview.

In this diagram, I have inverted the poles of the earlier Darwinian diagrams; the paradigmatic Darwinian phenomena are in the (0,0,0) corner, and the paradigmatically non-Darwinian phenomena, the instances of intelligent design (not Intelligent Design—this has nothing to do with religion) are in the upper right corner.

FIGURE 7.5: Inverted Darwinian space with Darwinian phenomena at (0,0,0) and intelligent design at (1,1,1).

The claim that I defend is that human culture started out profoundly Darwinian, with uncomprehending competences yielding various valuable structures in roughly the way termites build their castles, and then gradually de-Darwinized, becoming ever more comprehending, ever more capable of top-down organization, ever more efficient in its ways of searching Design Space. In short, as human culture evolved, it fed on the fruits of its own evolution, increasing its design powers by utilizing information in ever more powerful ways.

In the upper right corner, the intelligent design extreme, we have a well-known ideal that is in fact never reached: the Godlike genius, the Intelligent Designer who bestowed gifts on the rest of us mortals from the top down. All the real cultural phenomena occupy the middle ground, involving imperfect comprehension, imperfect search, and much middling collaboration.

I put Picasso at the pinnacle not because I think he was smarter than the other geniuses, but because he once said “Je ne cherche pas. Je trouve.” (“I don’t search. I find!”) This is a perfect boast of genius, compressing into a few syllables the following message: “I don’t have to grub around with trial and error! I don’t trudge my way up the slope to the global summit of designs. I just leap to the top of Mount Everest, every time! I comprehend all; I even comprehend my comprehension!” Baloney, of course, but inspired baloney. In the case of Picasso, a barefaced lie. He often made hundreds of sketches with variations on a theme, nibbling away in Design Space until he found something he thought was a good stopping point for his journey. (He was a very great artist but part of his genius lay in signing—and selling—so many of his trials instead of discarding them.)

There is much more to be explored in this evolution of cultural evolution and its role in creating our minds, but first we should look more closely at how it got started. Like the origin of life, this is an unsolved problem, and a very difficult one. Our minds are in some regards as different from other minds as living things are from nonliving things, and finding even one defensible path from our common ancestor with the chimpanzees is a challenging task. There is no shortage of competing hypotheses, and we will consider the best (by my lights). Which came first: language or cooperation or tool making or fire tending or stone throwing or confrontational scavenging or trade or …? We shouldn’t be surprised if there proves to be no single “magic bullet” answer, but rather a coevolutionary process with lots of contributing factors feeding on each other. The fact remains that we are the only species so far that has developed explosively cumulative culture—yet another replication bomb—and any speculative story that explains why (both how come and what for) we have culture must also explain why we alone have culture. Culture has obviously been a Good Trick for us, but eyesight and flight are also obvious Good Tricks, and they have each evolved several times in different species. What barriers have stood in the way of other evolutionary lineages developing the same Good Trick?