James Shapiro’s Evolution: A View from the 21st Century Offers a Stunning Look at Biological Complexity and Non-Darwinian Evolution
University of Chicago molecular biologist James A. Shapiro is not a proponent of intelligent design (ID). And he is an evolutionist. But his new book Evolution: A View from the 21st Century is recommended reading for ID proponents who have an interest in biological complexity.
Shapiro is not a promoter of the classical modern neo-Darwinian model of evolution. Rather, he believes that in a sense, biological organisms are programmed to evolve. That programming, he believes, ultimately did arise through Darwinian processes — unguided mutations preserved by natural selection. But he believes that once that programming was in place, it fostered much of the rest of the evolution of both prokaryotes and eukaryotes
A Change Would Do You Good
Sheryl Crow may not be a science expert (remember the toilet paper episode?), but her song title “A Change Would Do You Good” would make an apt description of Shapiro’s evolutionary model. He opens his book by asking “How does novelty arise in evolution?,” observing that “Innovation, not selection, is the critical issue in evolutionary change. Without variation and novelty, selection has nothing to act upon.” (p. 1) But how does that change arise in the first place?
According to Shapiro, Darwinian evolution’s insistence upon “countless random small changes over long periods of time” that are “accidental and random with respect to biological function or need” arose as a reaction to “the perceived need to reject supernatural intervention.” (pp. 1-2) Shapiro believes this tradition extends into the present day: “the accidental, stochastic nature of mutations is still the prevailing and widely accepted wisdom on the subject.” (p. 1)
But Shapiro wants to resurrect forms of goal-directed evolution, where novelty and variation are not produced randomly with respect to biological needs, and in fact organisms activate mechanisms to induce genetic and epigenetic changes at times when ‘a change would do you good.’ He takes this view because (1) modern biology has uncovered various mechanisms by which organisms can “rewrite” their own genomes, especially in response to stress, and (2) many biological pathways and structures do not seem amenable to stepwise Darwinian evolution. Shapiro explains:
(1) “The perceived need to reject supernatural intervention unfortunately led the pioneers of evolutionary theory to erect an a priori philosophical distinction between the “blind” processes of hereditary variation and all other adaptive functions. But the capacity to change is itself adaptive. Over time, conditions inevitably change, and the organisms that can best acquire novel inherited functions have the greatest potential to survive. The capacity of living organisms to alter their own heredity is undeniable. Our current ideas about evolution have to incorporate this basic fact of life.” (p. 2)
(2) “Do the sequences of contemporary genomes fit the predictions of change by ‘numerous, successive, slight variations,’ as Darwin stated, or do they contain evidence of other, more abrupt processes, as numerous other thinkers have asserted? The data are overwhelmingly in favor of the saltationist school that postulated major genomic changes at key moments in evolution. … [L]ittle evidence fits unequivocally with the theory that evolution occurs through the gradual accumulation of “numerous, successive, slight modifications.” (pp. 89, 128)
Evolution: A View from the 21st Century includes lucid (though at times somewhat technical) explanations of the latest findings in cellular complexity — complexity which in the opinion of many pro-ID scientists defies a classical Darwinian explanation. Whether or not Shapiro intended it (and I’m very confident he didn’t intend it), his book contains stunning descriptions of biochemical complexity and complex cellular regulation pathways that provide compelling arguments for biological fine-tuning that indicates intelligent design. For example:
- “One of the great scientific ironies of the last century is the fact that molecular biology, which its pioneers expected to provide a firm chemical and physical basis for understanding life, instead uncovered powerful sensor and communication networks essential to all vital processes , such as metabolism, growth, the cell cycle, cellular differentiation, and multicellular morphogenesis. … [T]he life sciences have converged with other disciplines to focus on questions of acquiring, processing, and transmitting information to ensure the correct operation of complex vital systems.” (p. 4)
- “Genomes are sophisticated data storage organelles integrated into the cellular and multicellular life cycles of each distinct organism. Thinking about genomes from an informatics perspective, it appears that systems engineering is a better metaphor for the evolutionary process than the conventional view of evolution as a selection-biased random walk through the limitless space of possible DNA configurations.” (p. 6)
- “Because the interactions in any cell process invariably grow more complex and involve more molecules as we investigate them in greater detail, most biologists agree that we are now in the systems biology era of research. Although this term is subject to various interpretations, a widespread view is that systems biology implies understanding how groups of molecules work coordinately (as a system) to achieve some useful function dependent upon conditions. Gone is the atomistic view that molecules act independently and automatically.” (p. 8)
- “There are a number of attempts to describe cellular information processing from a semiotic or linguistic perspective.” (p. 10)
His discussions of the biochemical mechanisms that regulate accurate DNA replication show the complexity of basic biological systems that neo-Darwinians often take for granted.
We can think of this two-level proofreading process as equivalent to a quality-control system in human manufacturing. Like human quality-control systems, it is based on surveillance and correction (cognitive processes) rather than mechanical precision. The multistep nature of proofreading is typical of many control processes in cells, where final precision is achieved by a sequence of two or more interactions that are each themselves inherently less precise. In this regard, the most applicable cybernetic models are fuzzy logic control systems. In such systems, accurate regulation occurs by overlaying multiple imprecise (“fuzzy”) feedback controls arranged so that each successive event results in greater precision. (p. 14)
Natural Selection Replaced by “Cognitive Networks” and “Self-Modification”
Shapiro describes his model of evolution as follows: Instead of “gradual selection of localized random changes” he proposes “sudden genome restructuring by sensory network-influenced cell systems … It replaces the ‘invisible hands’ of geological time and natural selection with cognitive networks and cellular functions for self-modification.” (pp. 145-146)
Rather than assembling by “random” mutation, Shapiro believes that much biological complexity arises through directed mutagenesis of the genome, often in response to outside stresses. Shapiro gives many examples of how organisms use what he calls “cellular regulation of natural genetic engineering” (p. 69) to reorganize the genome in times of stress. For example:
In the starvation-induced rearrangements that I studied in the 1980s and 1990s, the increase in transposon-mediated events increased by at least five orders of magnitude (that is, by a factor of over 100,000). They went from undetectable in more than 1010 bacteria under normal growth conditions to more than once per 105 bacteria on starvation plates. (p. 74)
According to Shapiro, such processes show that “the cell rewrites its genome” (p. 5) when needed.
But how did these supposed “natural genetic engineering” abilities arise? There are three questions that Shapiro’s model must face:
- (1) Does Shapiro’s model really reject Darwinian evolution and random variation?
- (2) Can we explain the origin of these “self-modification” mechanisms?
- (3) Can these “self-modification” mechanisms explain observed biological complexity?
Let’s briefly look at these questions.
Question 1: Does Shapiro’s Model Completely Reject Darwinian Evolution and Random Variation?
According to Shapiro, since certain environmental stresses or triggers can induce mutations in organisms, mutations are not “random.” By ‘not random,’ he means that he disagrees with the classical Darwinian view that variation arises blindly, without regard to the needs of the organism. I see two problems:
First, in an important sense, even within Shapiro’s model some of these mechanisms of change still have a random element. Some of his examples show mutations being induced at non-random times. While these mutations are not completely random, in an important sense, their effects may still be “random” with respect to the needs of the organism. For example, in the transposon example above, bacteria increase the rate of transposon-mediated events in respones to stress. As Shapiro notes, some transposons have targets that are “virtually ubiquitous throughout the genome” (p. 48). In another case he notes what he calls “indiscriminate LINE-mediated integration” (p. 52). The hope is that some will do something helpful and stick. But in an important sense this is still very much like a trial and error process of Darwinian selection, where not all change is adaptive. This becomes very clear when he writes that under his model:
The role of selection is to eliminate evolutionary novelties that prove to be non-functional and interfere with adaptive needs. (p.144)
While he notes that many of the mechanisms he cites provide a higher “probability of success” (p. 130) than mere point mutations, we’re still dealing with a trial and error process where variation arises, in an important sense, randomly without regard to the needs of the organism. The hope is that some of those changes will stick.
Some of the mechanisms he cites are like having monkeys sitting next to typewriters — they aren’t always typing but they are on standby, ready to start typing at a moment’s notice. The typewriting monkeys might be nonrandomly “switched on” at the right moment when you might need a letter, but there’s no guarantee that their random banging on the keyboard will produce a useful letter.
Second, though Shapiro doesn’t directly address this specific question, his model would seem to still require conventional random mutation and natural selection to explain how the “capacity to change” itself arose in the first place.
Shapiro argues convincingly that “the capacity to change is itself adaptive,” and thus the capacity to change is a selectable trait. But his argument implies that at some level, Darwinian evolution and random change must be invoked to explain the origin of the capacity to change. After all, he cannot invoke mechanisms of directed change to explain how those same mechanisms of directed change arose; those mechanisms could not have produced themselves, and had to come from something else. Under Shapiro’s model, at various points the capacity to change must have arisen by chance, and was preserved by natural selection. In Shapiro’s undesigned world, there is no way to ultimately divorce the evolutionary process from randomness, because at some point, “the capacity to change” arose randomly and blindly, without regard to the needs of the organism.
Here, Shapiro would seem to be confronted with a difficulty: He doesn’t want to rely on blind and random mechanisms for evolution because biological systems simply appear too complex to have arisen in such a fashion. However, if he doesn’t rely on blind and random mechanisms for at some point along the process, then he’s forced into the realm of intelligent design — which is exactly where he doesn’t want to go. In a sense, Shapiro’s thesis is no different from neo-Darwinism because he just pushes the selected trait back to the “capacity to change” rather than the “change itself.”
Question 2: Can we explain the origin of these “self-modification” mechanisms?
If the final complex “goal-directed” (p. 138) systems require non-random “genetic engineering” processes to form, does it make sense to believe that the processes which created those systems have a completely random and undirected origin?
Shapiro is very cognizant of this question, for he writes:
[T]he phrase natural genetic engineering has proven troublesome to many scientists because they believe it supports the Intelligent Design argument. As one Nobel Laureate put it after a seminar, “If there is natural genetic engineering, that means there has to be an engineer.” This empirically derived concept seems to many scientists to violate the principles of naturalism that exclude any role for a guiding intelligence outside of nature. (p. 134)
Clearly the specter of intelligent design is haunting Shapiro as he makes his case to other scientists. But his explanation for why ID isn’t needed is to simply assert that natural genetic engineering principles would provide a “distinct evolutionary advantage.” (p. 135) Here, Shapiro ignores his own advice.
While the capacity to change might be advantageous and “selectable,” as Shapiro reminds us at the opening of his book, that’s not enough to explain how something arose. Thus, as we saw above, Shapiro emphasizes that selection can only preserve the novelty provided by mutations. But Shapiro does not even attempt to explain how complex capacities to change — which he claims amount to a system of “cognitive networks and cellular functions for self-modification” could be produced by random mutations for selection to preserve.
It’s a truism to observe that there were no “natural genetic engineering” mechanisms in existence before “natural genetic engineering” mechanisms evolved. Thus, presumably he must rely upon more primitive and less dramatic Darwinian evolutionary mechanisms to explain how these mechanisms gained the capability to do things like “insert cis-regulatory signals and swap exons … simultaneously on more than one genetic locus encoding functionally related proteins.” (p. 135) Sure, the ability to do that precisely might sometimes provide a selective advantage, but Shapiro hasn’t established that the ability to effect such complex natural genetic engineering could evolve through step-by-step blind and unguided evolution.
Shapiro’s model goes partway in recognizing that biological systems operate in a “goal-oriented” (p. 138)–even “teleological” (p. 137) manner. Ultimately, however, Shapiro cannot escape from one vexing problem: goal-oriented processes don’t arise from random processes; goal-directed biological systems would seem to require a goal directed cause from the very start.
Question 3: Can these “self-modification” mechanisms explain observed biological complexity?
Shapiro’s excellent summaries of the mechanisms of non-random genetic and epigenetic biological change are entirely consistent with an ID view. For example, consider this claim from Shapiro:
It is generally necessary to integrate the expression of different regions of the genome in a coordinated fashion to execute a particular phenotypic trait. This regulatory integration is often achieved by reusing the same binding sites at multiple locations. (p. 31)
But how these binding sites come to be reused in multiple locations?
In Shapiro’s materialist view, the very existence of these coordinated systems is taken as evidence that transcriptional regulatory circuitry “can be formed and taken apart easily.” But if multiple different regions of the genome must be coordinated to produce such new traits, doesn’t that imply it’s actually quite complicate to form such features? Are the “self-modification” mechanisms Shapiro cites always up to the tasks he assigns them, or is he taking their mere existence as evidence that they must be able to build observed features?
Let me put the question another way: Shapiro prefers to describe the evolutionary process as “engineering” rather than “tinkering,” but is that because he has observed and demonstrated material mechanisms that truly can “engineer” the genome, or because he has observed features in biology that require engineering to arise?
After all, many of the “natural genetic engineering rearrangements” he cites entail observations like “LINE-1 elements associated with deletions in human genome variation” or “Many inversions associated with L1 repeats.” (pp. 122-123) No doubt those are accurate observations, but is that evidence that some complex mutational mechanism can produce (and has produced) new traits by coordinating multiple regions of the genome?
Shapiro’s ideas are certainly interesting, but his claim that these complex mutational mechanisms like “horizontal transfers and the movement of transposable elements through chromosome rearrangements [and] whole genome duplications and cell fusions” (p. 128) can spontaneously produce radically complex traits is not conclusively established. Much here is speculative, and much is inadequate: “domain swapping” (p. 130) cannot explain the origin of domains in the first place.
Thinking Outside the Box
I’m sure Shapiro would disagree with much of what I’m writing. But in the end, he should be commended for his out-of-the-box thinking. He understands what it means to be making bold proposals that buck the consensus
The common impulse is to declare “impossible” what does not agree with the assumptions or prejudices of a particular school of thought. Because it is the business of science to turn what was once thought to be impossible into reality, noting the perils of excessive generalization reminds us to keep our ideas fresh, creative, and inclusive rather than rigid and exclusive. (p. 80)
Those are words that ID proponents can certainly agree with.