Reactionary fringe meets mutation-biased adaptation. 5.3. How history is distorted.

According to Bateson (1894). natural selection is “obviously” a “true cause.” Punnett (1905) explains that mutations are heritable while environmental fluctuations are not, concluding that “Evolution takes place through the action of selection on these mutations” (p. 53). De Vries begins his major 1905 English treatise by writing that

Darwin discovered the great principle which rules the evolution of organisms. It is the principle of natural selection. It is the sifting out of all organisms of minor worth through the struggle for life. It is only a sieve, and not a force of nature (p. 6)

Morgan (1916), in his closing summary, writes:

Evolution has taken place by the incorporation into the race of those mutations that are beneficial to the life and reproduction of the organism (p. 194)

The views of these early geneticists were published, read, and discussed. All 4 were awarded the Royal Society’s Darwin prize in the period from 1900 to 1930.

So, why does the evolutionary literature continue to tell a story in which the early geneticists (“mutationists”) reject selection and believe in evolution by mutation alone (e.g., Dawkins 1987, p. 305 of The Blind Watchmaker; Cronin 1991, p. 47 of The Ant and the Peacock; Ayala and Fitch 1997; Eldredge 2001, p. 67 of The Triumph of Evolution; Segerstråle 2002, Oxford Encyclopedia of Evolution 2, pp. 807 to 810; Charlesworth and Charlesworth 2009)?

How does history get distorted? Why do the distortions persist? Is there a process for correcting them?

Synthesis Historiography

According to another story, opponents of Darwinism suffered from a condition called “typological thinking” or “essentialism,” preventing them from using “population thinking” to understand systematics and evolution. Historian Mary Winsor (2006) begins her piece (below) by explaining that the Essentialism Story was “fabricated” by Ernst Mayr.

Mayr’s self-serving accounts of history are so notorious that historians have a name for them– “Synthesis Historiography” or SH (Amundson, 2005), i.e., telling history in ways that turn out right for the Modern Synthesis.

The Essentialism Story, the Mutationism Myth, and the Eclipse of Darwinism are all parts of SH. In SH, all truths must come from Darwin and his followers, and this makes the early geneticists problematic, because they were crucial in the development of genetical thinking about evolution. Between 1900 and 1920, they conceptualized the multiple-factor theory, the biological species concept, the Hardy-Weinberg equilibrium, the allelic selection model, and the probability of fixation (Stoltzfus and Cable, 2014; Provine, 1971).

So, SH turns history on its head: the early geneticists are said to have rejected selection; the period of dramatic discovery and innovation from 1900 to 1920 is depicted as a period of darkness and confusion, the Eclipse of Darwinism, when the world was deprived of His light; and the credit for introducing 20th-century views of mutation and heredity to evolutionary biology is awarded to 19th-century physiologist August Weismann (though genuine historians object that Mayr’s story does not reflect Weismann’s views: see Winther, 2001, or Griesemer, 1989).

The above timeline of “notable people who have contributed to evolutionary thought” from a popular education resource illustrates SH. I have added a pink rectangle enclosing the birth year of anyone 25 to 60 years old– the prime of a scientist’s life–  when genetics was discovered, thus implicating de Vries (1848), Johannsen (1857), Bateson (1861), Cuénot (1866), Morgan (1866), and Punnett (1875). No timelines begin in the box.

Normality drift and back-projection

Mayr’s historical fictions are brazen, but most distortions emerge in a subtle and passive way, when ordinary scientists assume, incorrectly, that the old language was intended to cover current results, that the old theories were broadly conceived to cover all reasonable possibilities, and that dead authorities looked at evolution the way we do. I have done it myself.

For instance, contemporary scientists may use “Darwinian evolution” or “Darwinian adaptation” (e.g., here or here) for lucky mutant models of de Vriesian aptation, not realizing that Darwin and his Synthesis-era followers deprecated this mode of change, previously called “pre-adaptation” or “random pre-adaptation” (see Stoltzfus, 2017). When Darwin was criticized for providing “not a theory of the Origin of Species at all, but only a theory on the causes which lead to the relative success and failure of such new forms as may be born into the world,” his response was “That may be a very good theory, but it is not mine” (for a scholarly analysis of Darwin’s position on the creativity of selection, see Beatty, 2010, 2016).

Another example is provided by the continued use of the phrase “Fisher’s geometrical model” (e.g., here, here, or here) for a model that, in its current form, rejects Fisher’s assumptions and subverts his conclusion (see Stoltzfus, 2017 supplement). In the original model, the size of an allelic effect determines the chance of being beneficial, which is negatively correlated with effect-size– the smallest effects having the highest chance. The model was used to argue that the smallest changes are the most likely, in support of natura non facit saltum.

Fifty years later, Kimura re-cast Fisher’s argument in a stochastic origin-fixation framework, which creates a second effect, positively correlated with effect size, due to the higher chance of fixation of beneficial mutations with larger effect-sizes. The combination of divergent effects results in an intermediate optimum, undermining Fisher’s original conclusion, and undermining the original gradualist intent of Fisher’s argument.

This distortion is mainly a matter of normality drift, but the literature also reveals signs of back-projection, i.e., the anachronistic belief that Fisher shared Kimura’s mutationist view (e.g., p. 121 of Orr, 2005).

Back-projection is both pernicious and difficult to avoid. I have done it myself, in regard to the following passages

“Probably the most effective aid in establishing new genes lies in their repeated production by independent mutations.  A gene produced twice by mutation has twice as good a chance to survive as if produced only once.” (Shull, 1936, Evolution, p. 140)

“There is another result, clearly established by the genetic work on Drosophila, that is favorable to the final establishment of a new type of character if it is beneficial. Most, perhaps all, of the mutations appear more than once. This improves their chances of becoming incorporated in the species, and if the mutation produces a character that favors survival the chance of its becoming established is still further increased.” (Morgan, 1925, Evolution and Genetics, p. 140 to 141).

Previously, I cited the “twice as good a chance” passage from Shull to suggest that the early geneticists intuited the proportional effect of biases in the introduction process. Indeed, the passages above indicate that Morgan and Shull thought that the chances of introduction are consequential for evolution. Other writings show that they were interested in the potential for internal biases in evolution, e.g., Shull (1935) wrote “What the world most needs, then, is not a good five-cent cigar, but a workable– and correct– theory of orthogenesis.” 

So, it would seem obvious that they would put the pieces together and propose a theory of variation-biased evolution, implicating biases in the introduction process.

Yet, they did not. In the passages above, Shull and Morgan are simply not working out a theory for variation-biased evolution: they are making a plausibility argument in which evolution by mutation and selection is rendered more plausible by the fact that mutant alleles may be introduced multiple times, rather than having only a single chance to succeed or fail. In retrospect, this is embarrassingly obvious.

Remember the parable of the blind men and the elephant, or the example of chaotic dynamics hidden in familiar equation systems: until we know a theory, we just don’t know it, even if we can see most of the parts.

The cure for back-projection

Finally, let us return to TREE’s hatchet piece, whose authors reject the idea that the theory of Yampolsky and Stoltzfus is new:

Haldane (1927, 1933) fully presents the implications of biases in the introduction process, 70 years before Yampolsky and Stoltzfus, and this theory was then invoked by Mayr and Simpson to explain how variation imposes biases on adaptation. Clearly, this is already part of the Modern Synthesis.

I’m just kidding! They do not say that, because






Instead, they say something indirect and misleading:

“It is simply not true that the architects of the modern synthesis were unaware of the potential role of novel mutations in the evolutionary process because mutational bias was already discussed in Haldane’s and Dobzhansky’s early papers [19,74,75]. For instance, Dobzhansky discussed the possible role of mutation bias (expressed as ‘similarity in germ plasms’) as a potential explanation for parallel geographic variation in various species of lady-bird beetles”

The first mistake in this framing is to focus on a part or a premise of the theory, while neglecting the inner logic that leads from premises to conclusions. They are implying that, if someone notices that not all mutation rates are the same (mutation bias), they automagically grasp the implications of biases in the introduction process. This is a fallacy. Before we know a theory, we are like the blind men in the parable who identify the parts of the elephant but not the whole creature.

The second false step involves the conflation of people and theories that is central to Synthesis apologetics. Theories are not made of people. Even if some historic person knows the implications of biases in the introduction process, and likes the Modern Synthesis, this does not mean that the former theory is magically incorporated into the latter theory.

Nevertheless, let us consider what the cited sources– Haldane (1927, 1933) and Dobzhansky (1933)– have to say about the possible role of biases in the appearance of novel mutations in determining the course of evolution.

Haldane (1927) addresses the probability of fixation for a new mutation, and a mutation-selection balance model. He invokes opposing-pressures thinking, not biases in the introduction process, concluding that:

“Mutation therefore determines the course of evolution as regards factors of negligible advantage or disadvantage to the species. It can only lead to results of importance when its frequency becomes large.”

That is, Haldane derives the classic view that mutational effects require neutral evolution or high mutation rates. Strike one.

The second piece, Haldane (1933), begins by listing its 3 topics:

  • “(1) Evolution due to a mutation rate so large as to cause the spread of a disadvantageous character . . .
  • (2) Primary effects of the spread of genes nearly neutral from the point of view of selection . . .
  • (3) Secondary effects of frequent disadvantageous mutations.”

None is the right topic. Strike two.

What about “Geographical Variation in Lady-beetles” by Dobzhansky (1933)? Though Dobzhansky was publishing work in experimental genetics at this time, here he is reporting a quantitative survey of variation in the patterns of spots on the elytra (wing-covers) of different species of lady-bird beetles. He finds similar patterns of variation in different species, and suggests an explanation:

“As stated above, there is observed a parallelism in the variability of the related species and genera. Homologous varieties of different species may be more similar to each other in appearance than the different varieties of the same species. This parallelism is, probably, due to the essential similarity of the germ-plasms of the related species. “

This is more promising. Dobzhansky is referring somewhat obliquely to the effect of parallel tendencies of variation in generating parallel patterns of intra-specific variation. The context makes clear that Dobzhansky (1933) likes the law of homologous series in variation proposed by Nikolai Vavilov (1922 [PDF]), cited as a “rule” on page 1:

“Large groups of related species and genera exhibit parallel series of patterns, upholding the rule of homologous series in variation, formulated by Vavilov (1922)”

What a “homologous” or “parallel” series means is that one species has varieties or races A, B, and C, and a second closely related species has corresponding varieties A’, B’, and C’. Vavilov’s (1922 [PDF]) theory held that

“Variation does not take place in all directions, by chance and without order, but in distinct systems and classes analogous to those of crystallography and chemistry.  The same great divisions into orders and classes manifest regularities and repetitions of systems” (p. 84)

That is, Vavilov imagined something like the periodic table of the elements, but for systematics. He proposed a research program of identifying the pattern of variant forms characteristic of a taxon, and then looking for the forms that complete the pattern.

In other words, Dobzhansky is not advocating the Modern Synthesis, but a pre-Synthesis mutationist theory. This is an excellent illustration of why it is a mistake to use people as proxies for theories.

Furthermore, Vavilov’s theory is not explicit about population genetics, so one does not know, for instance, whether the proposed effects require high mutation rates or neutral evolution.

Thus (returning to the argument in TREE’s hatchet piece), for Dobzhansky or Haldane to know and even to like some of the thinking of the mutationists does not magically induce knowledge of the theory of Yampolsky and Stoltzfus (2001), much less make it part of the Modern Synthesis, which clearly was a repudiation of mutationism. Strike three.

In conclusion, when the authors of TREE’s hatchet piece write

“It is simply not true that the architects of the modern synthesis were unaware of the potential role of novel mutations in the evolutionary process”

they have deceived the reader.

What is happening here? Why would the authors of a refereed paper go out of their way to make a historical argument that is complete bullshit?

The answer, of course, is that the authors have thoroughly deceived themselves. To understand how, imagine that we have a strong prior belief that everything valuable comes from tradition and traditional authorities, e.g., we might believe that

“. . . the major problems of the field have been solved. We understand both the nature of the mutational processes that generate novel genetic variants and the populational processes which cause them to change in frequency over time — most importantly, natural selection and random genetic drift, respectively . . . no serious evolutionist will now defend once prevalent views such as orthogenesis (predetermined evolution) or the inheritance of acquired characteristics . . . we will never again come up with concepts as fundamental as those formulated by the ‘founding fathers’ of population genetics (Fisher, Wright and Haldane), or do experiments as path-breaking as Dobzhansky’s demonstration of natural selection acting on polymorphic chromosome inversions.”

Charlesworth B. 1996. The good fairy godmother of evolutionary genetics. Curr Biol 6:220.

When our faith in the fullness and authority of tradition reaches this upper bound, we will respond to the claim of a new principle with unyielding skepticism, as if it were a claim of finding a unicorn or being abducted by aliens. When faced with evidence that Haldane found a unicorn (i.e., argued against the efficacy of tendencies of variation) or that Fisher was abducted by aliens (i.e., ignored the potential importance of biases in the introduction process), we will be confident that the evidence must have some other explanation, even if we can not prove it.

Thus, the deceptive attributions in TREE’s hatchet piece arise because the authors are operating near the Charlesworth limit. They have an unwavering confidence that contemporary scientists cannot possibly have formulated a new principle, and that traditional authorities must have known this earlier. Any ancient text that mentions mutation bias or new mutations simply strengthens this prior belief.


When we confront a scientific issue that is more than a generation old, we necessarily become historians. In such a context, getting the science right depends on getting the history right, and this requires an awareness of two factors. The first is that the literature features a multitude of individually minor instances of normality drift and back-projection, whose cumulative effects may be profoundly distorting.

The second is that the Synthesis literature is rife with outrageously self-serving accounts of history. When such stories are repeated countless times, people come to believe they are well established, and that they can only be overturned by extraordinary evidence. However, Synthesis Historiography is propaganda, not scholarship. The term to describe well meaning people who repeat propaganda is “useful idiot.” We have a scholarly obligation not to repeat dubious historical claims, until we see evidence.

What does evidence look like? In recent posts, we have seen multiple examples of statements from historic texts that clearly convey the inner logic of theories, and which reveal the know-like-love relationships of persons to theories.

Most scientists do not have time to pore over the historical literature searching for this kind of evidence. However, being a rigorous consumer of evidence is much easier than being a producer: one simply demands that textual evidence demonstrates exactly what is being claimed, i.e., one treats meta-scientific claims as carefully as one would treat a scientific claim.

Reactionary fringe meets mutation-biased adaptation. 5.1. Thinking about theories

A wikipedia page disambiguating “Modern Synthesis” defines neo-Darwinism as

“the state-of-the-art in evolutionary biology, as seen at any chosen time in history from the 1890s to the present day.”

Because “neo-Darwinism” and the “Synthesis” are conflated with whatever is widely accepted, they are now attacked on grounds that are completely unrelated to genuine neo-Darwinism or the original Modern Synthesis, e.g., as when a network of life (rather than a tree) is invoked as a contradiction of Darwinism. The attack by Noble (2015) on the

“conceptual framework of neo-Darwinism, including the concepts of ‘gene’, ‘selfish’, ‘code’, ‘program’, ‘blueprint’, ‘book of life’, ‘replicator’ and ‘vehicle’”

is entirely a critique of late-20th-century reductionism à la Dawkins, and addresses neither neo-Darwinism (selection and variation as the potter and the clay), nor the Modern Synthesis that emerged in the 1950s. The original Modern Synthesis is simply not a reductionist theory, but positively invokes emergent phenomena (population-level forces, the gene pool as dynamic buffer) in the service of selection as a high-level governing principle.

How did this confusion happen? Mayr, who died in 2005, never accepted neutral evolution. Yet, decades earlier, scientists intent on preserving a “Synthesis” began to claim that it somehow changes to accommodate new findings, making it a “moving target” (Smocovitis, 1996). This innovation defeats the purpose of falsifiable theories. What if the mutationists adopted this self-serving doctrine, claiming that mutationism is continually updated to cover new results?

Strangely, the latest defenders of orthodoxy no longer invoke “neo-Darwinism” or “Modern Synthesis,” but refer to the “Synthesis” or “Standard Evolutionary Theory” (e.g., Nature debate of 2014). What does this mean? Did the Modern Synthesis fail to update itself quickly enough? If neo-Darwinism was refuted, how precisely did that happen? What does the word “theory” even mean in these debates? What changes in evolutionary thinking have happened in the past 50 years?

To discuss such questions, we must begin with a basic understanding of what scientific theories are, and what they are not– issues established by generations of scientific usage.

TheoryA and TheoryC

Kimura’s Neutral Theory is the conjecture that the majority of changes at the molecular level result from the random fixation of selectively neutral alleles. The endosymbiotic theory holds that mitochondria and chloroplasts arose from bacterial endosymbionts. For hundreds of years, scientists have used “theory” to denote a major hypothesis or systematic conjecture. This is the primary meaning of “theory” in science.

Yet, this meaning is typically not what we have in mind when we use the word “theoretical” or “theoretician.” Population genetics theory is not the conjecture that populations have genetics, nor is music theory the conjecture that music exists.  Instead, these are bodies of abstract principles.

That is, scientists use “theory” for both (1) theoryC (concrete, conjectural), a grand conjecture or major hypothesis to account for some observed phenomena, and (2) theoryA (abstract, analytical), a body of principles relevant to some discipline, methodology or problem area (addressed at length here).

For instance, The Role of Theory in Advancing 21st Century Biology emphasizes the development of formalisms, and says that “a useful way to define theory in biology is as a collection of models,” clearly a reference to theoryA. The title and the quotation illustrate how English speakers often denote theoryA using the abstract noun, i.e., “theory” not preceded by “the” or “a.” Fisher wrote that

“No practical biologist interested in sexual reproduction would be led to work out the detailed consequences experienced by organisms having three or more sexes; yet what else should he do if he wishes to understand why the sexes are, in fact, always two?”

Familiar methods of science require theoryA statements covering imaginary, non-actual worlds, in order to support modus tollens reasoning, in which X is rejected based on the proposition if X then Y, and the observation that Y is absent. For instance, in the case of neutral models, X = neutrality, and Y is some neutral expectation about rates or patterns.  In order to have confidence in rejecting neutrality, we must be confident that, in a hypothetical world of neutrality, Y would occur.

Thus, we evaluate theoryA and theoryC differently. The standard of validity for a theoryC is verisimilitude: how well does it match the real world? By contrast, a statement in theoryA is valid if it is correctly derived from its assumptions, even if it invokes imaginary things, e.g., infinite populations.  Once a piece of theoryA is valid (correctly derived), it is valid forever. By contrast, a theoryC takes risks, and can be refuted by contrary facts.

Opponents of the Neutral TheoryC deny its verisimilitude, yet are quite happy to make use of its theoreticalA infrastructure in efforts to define and then reject neutral models, as in the review by Kreitman (1996). The paradox in Kreitman’s title “The neutral theory is dead. Long live the neutral theory” is resolved by the fact that it refers first to theoryC, then to theoryA.

Persons and theories

What is the possible relationship of a person P to a theoryC T?  In What makes it new?, we saw examples of the 3 basic types– knows, likes, loves (commits to accepting)–, each of which has positive and negative flavors:

1.  P knows (does not know) T.  That is, P expounds or applies T, revealing its inner logic. A special case is when P originates T, i.e., P expounds T for the first time. Merely citing a source or naming a theory does not show knowledge. For instance, Maynard Smith, et al. (1985) know the opposing pressures argument: they can recite it and follow the implications.

2.  P likes (dislikes) TP advocates for (against) T, arguing for (against) its likelihood on theoretical or empirical grounds. The authors of TREE’s hatchet piece dislike the theory of Yampolsky and Stoltzfus (2001), arguing against its relevance, whereas they like the correlated-selection-shapes-M theory of Box 3: they advocate this theory as being a more likely explanation for cases of an alignment between mutational and evolutionary tendencies.

3.  P commits to accepting (rejecting) T, possibly with some limits, as if compelled inescapably by logic or evidence. That is, P reasons about the world as if T (or not-T) is not merely conjectural, but a solid foundation for further reasoning. For instance, in the following passage from the 1959 Darwin centennial (when the victory of the Modern Synthesis was declared), Stebbins makes a falsifiable prediction based on his commitment to a theoryC:

“One very important point here is this: if we say that genetic recombination is necessary to generate new adaptive systems and then say that such highly adaptive and complex systems as the cell of an amoeba, or a euglena with its nucleus, chloroplasts, eyespots, flagella, etc., evolved without the aid of genetic recombination, we are contradicting ourselves. Even though we don’t know that genetic recombination exists in these one-celled organisms, we must postulate its existence at the time they evolved”

Stebbins, p. 115 to 116 of Vol 2, Tax and Calendar, Evolution After Darwin: Issues in Evolution

That is, if the ability of selection to create complex traits depends on recombination among alleles maintained in the gene pool– a defunct principle of the Modern Synthesis–, then single-celled organisms with complex adaptive structures must have evolved with recombination.

These are relations of persons to a theoryC. For theoryA, things are simpler. If we know a theoryA, we commit to its in-principle validity, unless we suspect a problem with its derivation. In fact, theoreticiansA occasionally have lengthy disputes about the precise assumptions underlying theoryA statements, e.g., Hamilton’s rule or Fisher’s fundamental theorem. 

Views and schools of thought

In the Origin of Species, Darwin invokes 3 major means of evolutionary modification, which we can denote with A (natural selection), B (use and disuse) and C (direct effects of environment), in a ratio of roughly 25:5:1. On this basis, we could define Darwin’s view as D ::= A > B > C

In this way, we could define a view for every scientist. 

We could cluster these views into schools of thought or research programs based on similarity, or based on a social network. Each cluster would have a spectrum of know-like-love relationships to various ideas.

However, these person-based concepts usually do not correspond to cohesive theories, e.g., most single-author books about evolution do not present grand unified theories. The typical scientist is an explorer and an opportunistic problem-solver. A scientist may value conflicting theories. Why not? Few scientists are like Pearson or Fisher, dedicated to ideological purity.  Before Kimura turned his vaulting ambition to neutral evolution, it was directed at the antithetical goal of unifying theoretical population genetics with a deterministic maximization principle (see Grodwohl, 2017).

Likewise, schools or traditions can be inconsistent or arbitrary.  DeLage and Goldsmith (1913) write that ”The truth is that [neo-]Lamarckism was never a real system” (p. 244) because it evolved as a counter-reaction, defined in contrast to the complete rejection of external heredity-modifying factors by the dominant neo-Darwinian school. 

Discerning theories

A theory exists to the extent that it causes different people to carry out the same kind of domain-specific reasoning. The way to understand a theory is to look for the reasoning it supports, going from inputs to outputs based on some internal logic.

For instance, the opposing-pressures argument was used to reject the general idea of variation-biased evolution, and to support the doctrine that selection is the sole source of direction in evolution. This conclusion emerges by comparing the magnitudes of two inputs, mutation rates and selection coefficients. The internal logic depends on the “forces” theory, which says that evolutionary causation can be understood in terms of mass-action pressures on allelic frequencies. Under this theory, the direction of mutation, when not aligned with the direction of selection, must be opposed to selection, and the mutation-selection balance equation tells us what happens in this case.

Fisher, Haldane, and others expressed the opposing-pressures theory in similar ways. By contrast, expressions of neo-Darwinism or the Modern Synthesis are quite diverse. We can think of them as the diverse phenotypic expressions of an underlying genotype. If the expressions are too diverse, it may mean that there is no common theory, or that the common theory is only a weak determining factor due to the influence of other theories.

Finally, because theories are often embedded in works of persuasive rhetoric, we must be careful to separate the product from the sales pitch, following the English aphorism to watch what they do, not what they say. We find the surest evidence for the content of a theory, not when scientists present an explicit description, but when they invoke the terms of the theory to conduct reasoning from premises to conclusions, exposing the inner logic.

In particular, scientists with the “love” relationship to a theory may exaggerate strengths, ignore weaknesses, and set up false comparisons to flatter the object of their affection.  Fisher’s Ch. 1 argues that, once the discrete basis of heredity is recognized, then neo-Darwinism follows and all other views must be cast aside. Yet, we do not define neo-Darwinism as the view that follows from genetics, because Fisher’s assertion is clearly false, e.g., neutrality, mutation-biased evolution, and saltations are all compatible with Mendelian population genetics.

Likewise, consider the following from Maynard Smith (1969)

”If one invents counter-examples, they seem absurd.  Thus if someone discovers a deep-sea fish with varying numbers of luminous dots on its tail, the number at any one time having the property of being always a prime number, I should regard this as rather strong evidence against neo-Darwinism.  And if the dots took up in turn the exact configuration of the various heavenly constellations, I should regard it as an adequate disproof.  The apparent absurdity of these examples only shows that what we know about existing organisms is consistent with neo-Darwinism.”

(p. 86 of “The Status of Neo-Darwinism” in  Waddington CH, editor. Towards a Theoretical Biology 2. Sketches. Edinburgh: Edinburgh Universeity Press)

This passage tells us nothing important about neo-Darwinism, because the counter-examples are absurd.  Actual historic counter-examples to neo-Darwinism include (1) one-step speciation by de Vriesian macromutations, (2) parallelism or trends due to biases in variation per Vavilov (1922) or Eimer (1898), (3) evolutionary change dependent on the timing or character of individual mutations, (4) rampant neutral evolution, and (5) saltations, i.e., changes that are not composed of infinitesimals.

Reactionary fringe meets mutation-biased adaptation. 5.2. The Modern Synthesis of 1959

As we learned in What makes it new?, the discovery that biases in the introduction process may impose biases on evolution is new, because classical thinking says that evolution is screened off from the dynamics of introduction, so that it can be understood as a process of shifting the frequencies of existing alleles. How did this position emerge? Was it a technical, mathematical issue?

The revolt of the clay

The crucial issue, following Ch. 4 of Provine’s (1971) seminal The Origins of Theoretical Population Genetics, was not a matter of mathematics at all, but whether selection is the kind of creative governing force required for a neo-Darwinian view.

Early in the 20th century, Johannsen experimentally refuted Darwin’s non-Mendelian theory of environmental fluctuations and blending inheritance (see Roll-Hansen 1989; Gayon, 1998). Bateson and the early geneticists called for a new understanding of evolution based on genetics (see Stoltzfus and Cable, 2014). They refashioned the concept of selection as a sieve, or as a force that shifts the frequencies of true-breeding types, as in the allelic selection model of 1915, which Provine (1971) calls “the perfect complement to Morgan’s theory of evolution by single gene replacement” (p. 141).

In a world of discrete genetics, they argued, selection is not creative: “mutation proposes, selection disposes,” referring to the popular aphorism “Man proposes, God disposes,” which has versions in English, French and German, and traces back 600 years to Thomas à Kempis’s Latin aphorism using proponit (proposes) and disponit (administers, manages). That is, the mutationist aphorism puts selection in the role of God, with the connotations of Edwin Landseer’s 1864 painting “Man proposes, God disposes” (above), a dire reflection on the thwarted ambitions of Franklin’s lost expedition.

The Mendelian argument against the creativity of selection brought belief in neo-Darwinism to a low point. Poulton (1908) describes this era as the “revolt of the clay against the power of the potter,” referring to the biblical admonition against challenging the creator:

“What sorrow awaits those who argue with their Creator. Does a clay pot argue with its maker? Does the clay dispute with the one who shapes it, saying, ‘Stop, you’re doing it wrong!’ Does the pot exclaim, ‘How clumsy can you be?’” (Isaiah 45:9, New Living translation)

The restoration of Darwinism

In Provine’s (1971) narrative, the neo-Darwinian revival, restoring selection as the potter and variation as the clay, is sparked by a famous series of selection experiments. Beginning with mottled black-and-white rats, Castle and his coworkers selected lines that were nearly all white, and nearly all black: “wholly new grades,” according to Castle. The scale of the experiment was only 20 generations, too short for new mutations to play an important role. Instead, color changed incrementally as new, more extreme genetic combinations of alleles at many loci came together by recombination and were selected.

As Provine concludes, “Castle had been able to produce new types by selection,” without mutation, providing an experimental basis to rebut the mutationists and argue for selection as the creative force in evolution.

A general theory was constructed from this paradigm of selection, recombination, and shifting gene frequencies, representing a formally complete Mendelian justification for neo-Darwinism. By the time of the 1959 Darwin centennial, this theory was invoked by Mayr, Simpson, Dobzhansky, Huxley, Stebbins and others. Viewed from the outside, in terms of visible features, evolution is a smooth shift in phenotypes driven by selection (figure, left). At the genetic level (center), frequencies of small-effect alleles at many loci (A1 vs. A2, B1 vs. B2, etc) are shifting simultaneously from the previous optimum to a new optimum. All of evolution follows from shifting gene frequencies. 

In mathematical models, the population moves in the topological interior of an allele-frequency space (right), without an introduction process.

Of course, every allele had to arise by mutation at some time in the past, but the theory holds that the dynamics of this process are not consequential, because selection does not wait for new mutations, but leverages variation available in an abundant gene pool.

Thus, the theory deliberately excludes the mutationist view that the course of evolution depends importantly on the timing and character of individual events of mutation. Mutations play no direct role, but simply replenish the gene pool, as Stebbins (1959) explains:

“mutation neither directs evolution, as the early mutationists believed, nor even serves as the immediate source of variability on which selection may act.  It is, rather, a reserve or potential source of variability which serves to replenish the gene pool as it becomes depleted through the action of selection” (p. 305) 

Likewise, Mayr (1960) says

“The real function of mutation is to replenish the gene pool and to provide material for recombination as a source of individual variability in populations.”  (p. 355 of Mayr E. 1960. The Emergence of Evolutionary Novelties. In: Tax S, Callender C, editors. Evolution After Darwin: The University of Chicago Centennial. Chicago: University of Chicago Press). 


“Evolution is not primarily a genetic event. Mutation merely supplies the gene pool with genetic variation; it is selection that induces evolutionary change.” (p. 613 of Animal Species and Evolution, 1963).

Mayr did not discover or study this process himself, but is merely invoking a way of thinking learned from others. This is how we know that a theory exists: it causes different people to carry out similar forms of reasoning. For additional statements of this theory, contrasted with contemporary thinking, see The shift to mutationism is documented in our language.

And its demise

Note that the quotations above are examples of good historical evidence. They show the operation of a theory at work, revealing how scientists invoke the logic of a theory to reach high-level conclusions. Previously (part 4), we saw that Huxley, Gould, Maynard Smith, et al. and Reeve and Sherman invoke the opposing-pressures argument. Above, Mayr and Stebbins invoke the gene-pool logic of the Modern Synthesis of 1959. In the case of Huxley, Gould, Mayr and Stebbins, these high-level conclusions are offered as powerful truths, without qualification, i.e., they exemplify the relationship we defined previously [LINK_PART_5.1] as “love” or “commitment.”

That is, although the Modern Synthesis of 1959 was a speculative theory, proponents believed that it was well supported, e.g., Stebbins (1959) follows his description of principles (cited above) by declaring that

“The factual evidence in support of these postulates, drawn from a wide variety of animals and plants, is now so extensive and firmly based upon observation and experiment that we who are familiar with it cannot imagine the appearance of new facts which will either overthrow any of them or seriously limit their validity.”

Stebbins’s confidence was misplaced: the theory was promptly contradicted in the 1960s by molecular sequence comparisons suggesting evolution as a Markov chain of mutation-fixation events. At the time, this contradiction was resolved by arguing that the Modern Synthesis correctly depicts the evolution of the kinds of visible features discussed by evolutionists for generations, and is not compromised by the discovery of unimportant changes at an invisible molecular level. The “paradox” of two distinct modes of evolution was a theme of researchers for many years (see Dietrich, 1998).

“Molecular” evolution was left to “molecular” evolutionists, who held that “we need new rules in order to understand the pattern and dynamics of molecular evolution” (King and Jukes, 1969). 

Among the new rules was a direct link between the rate of evolution and the rate of mutation that classical theory had not supplied (McCandlish and Stoltzfus, 2014).  Yet, mutationist models remained in the domain of molecular evolution, being used mainly by neutralists (McCandlish and Stoltzfus, 2014). The introduction process was not the focus of contending theories, which addressed other issues (see Crow, 2008).  The dominant mainstream conflict was neutralism vs. selectionism, not neo-Darwinism vs. mutationism (though Jack King largely grasped the issue, in an ahistorical way).

The role of the introduction process in evolution is due for re-examination, based on the evidence that the course of adaptation reflects biases in the introduction of variation.

The theory behind the Synthesis

In the Synthesis story, a new and powerful understanding of evolution emerges in the mid-20th century, (1) combining genetics with neo-Darwinism, (2) sweeping away all rivals, and (3) providing a unified basis for biologists to apply evolutionary principles to their work.

What theory satisfies this description? What theory plays the starring role in the Synthesis story?

The Modern Synthesis of 1959, as described above, (1) offers a Mendelian justification for neo-Darwinism, (2) excludes mutationism (via the gene-pool theory) and internally biased evolution (via the opposing-pressures argument), and (3) supports high-level principles that can be applied across disciplines, e.g., they can be applied to interpreting the fossil record without knowing details of genetics and development, because the theory says that these details are not determinative.

By contrast, the latest lists of alleged Synthesis principles from apologists for orthodoxy (e.g., Futuyma, 2017) fail to justify neo-Darwinism, fail to exclude mutationism, and fail to support the high-level principles that (for instance) allowed Simpson to infer that trends in the fossil record must be due to selection because there is no alternative. Such lists of principles are not theories designed to take bold risks or to inspire the imagination, but are merely the output of an algorithm: describe evolutionary thinking so as to maximize the apparent similarity with mainstream views of 1959.

As mainstream thinking diverges from 1959, the output of this algorithm has dwindled. Today it says merely that evolution is a process that takes place in populations through the action of mutation, selection, drift and recombination, with selection playing a very important role. Relative to the Modern Synthesis of 1959, this “theory” is weak, but in Synthesis apologetics, it is claimed to be powerful on the grounds of lasting so long. Of course, the description generated by running the Futuyma algorithm after X years always has the property that it has lasted X years. This algorithm is not informative about the theories that shape evolutionary discourse, because it necessarily leaves out discarded ideas such as the creativity of selection and the role of recombination in the gene pool.

To the extent that science proceeds in Popperian fashion, by rejection, these discarded ideas serve as the mile-posts of progress. For instance, when Provine reflected back on the Modern Synthesis in the 2001 re-issue of his 1971 classic, he said that it “came unraveled for me during the period since 1980.”  He rejects the notion that macroevolution is a simple extension of microevolution, rejects the idea that evolution depends on recombination rather than mutation as the proximate source of variation, and refers to the gene pool as “one of the most artificial concepts of population genetics.”

Nevertheless, the Modern Synthesis of 1959 remains valuable, as does neo-Darwinism, when we apply them as models– as conceptual tools with specific functions, in the manner that neutral models are applied by contemporary scientists. Such theories provide ongoing guides to thought and to hypothesis-generation. Their successes and failures may be used to demarcate our substantive knowledge of how evolution actually occurs.