[Note: Readers may wonder about the long delay between this post and the previous posts in the series. Before completing it, I encountered an article by Jacques Monod, mentioned below. Monod makes many of the same points that I intended when I began. Having been thus scooped, I felt it necessary to redo the post accordingly. It also grew sufficiently long that I decided to separate out this part to post first.]
In the previous posts in this series, I
argued that the popular image of scientific prediction was too narrow,
discussed predictions that were actually “retrodictions”, and
identified a category of scientific predictions called “cross your fingers”. These were actually particular empirical issues raised by the theory, often critical to its validation or refutation, but remaining to be resolved. Thus, for example, Copernicus agreed with Aristotle that heliocentric motion would give rise to an apparent annual parallax – a shift in the apparent position of the stars as seen from opposite sides of the Earth’s orbit. The failure to see this parallax was a serious problem for Copernicus, but he explained it away by asserting that the stars must be much farther away than previously thought. This explanation was a problem for him, because many people were not comfortable with the celestial sphere (and, presumably, God) being so distant. But it was also a “cross your fingers” prediction. Eventually, someone should be able to make a sufficiently precise measurement to see parallax, which would confirm both heliocentric motion and the great distance of the stars.
With the increasing precision of measurements, failure to detect annual parallax became more frustrating. But as long as it was consistent with a great distance to the stars, it was still not evidence against heliocentrism. (Of course, had there been independent evidence that the stars were nearer, then it would have been a refutation.) As it happens, before parallaxes were detected, the observational effects of the Earth’s heliocentric motion were detected by James Bradley in the form of the aberration of starlight, in 1725. Bradley’s discovery was a prediction of the form “I should have thought of that” – an unexpected phenomenon for which the existing theory provides a satisfactory explanation. It wasn’t until 1838 that Friedrich Wilhelm Bessel detected the first annual parallax, in the apparent motion of the star 61 Cygni.
Broadly speaking, Darwin’s theory of common descent with modification by natural selection has the following basic elements: variation (of progeny), selection (by ecological fitness, sexual appeal, etc.), inheritance of varied characters (in the progeny of the selected progeny), and lots of time (over which selection pushes small variations to become large variations). Of these, I think the one that troubled Darwin most at the time, and yielded his most striking prediction, was inheritance.
As did most scientists of his time, Darwin accepted the blending theory of inheritance, which held that progeny merged the hereditary factors of their parents to form some intermediate characteristics. Thus, the child of a tall mother and short father would be intermediate in height, and would pass that intermediate height characteristic on to its children. This made Darwin’s model of descent with modification more difficult, as it implied that significant variations would be diluted away after a few generations. Indeed, this was the nub of a famous
criticism of
Origin of Species by Fleeming Jenkin. Jenkin was an engineer (famous for work on laying oceanic telegraph cables, and for inventing a variation on the cable car in which the cable carried the power for the motors). He was a business colleague and friend of Sir William Thompson, later Lord Kelvin, of whom more later.
A point that has subsequently been made by historians of science is that this criticism was particularly effective because people assumed that most characteristics varied continuously. Indeed, one reason why Mendel’s work on genetics did not get more notice at the time it was published was that it was felt that the discrete characteristics he studied (peas wrinkly or smooth, green or yellow) were not relevant to the continuous variations that were thought to be the most important. Ernst Mayr complains that even Jenkin should have appreciated discontinuous inheritance, because the example of significant variation that he used in his critique was children being born with six fingers. Surely, says Mayr in
The Growth of Biological Thought (1982, Harvard University Press),
Darwin could have easily refuted Jenkin by pointing out that six-fingered individuals do not have children with five-and-a-half fingers and grandchildren with five-and-a-quarter fingers…
However, the strength of the presumption of continuous variation was such that it was only at the turn of the century, as part of the New Synthesis, that Mendel’s empirical work was used to establish the general fact of “particulate” inheritance, rather than blending.
Darwin himself was fully aware of the problem, and even talked about it in a letter to T. H. Huxley in 1857, cited by R. A. Fisher in his book
The Genetical Theory of Natural Selection (1930, Oxford University Press; reprinted in M. Ridley (ed.),
Evolution - An Oxford Reader):
Approaching the subject from the side which attracts me most, viz., inheritance, I have lately been inclined to speculate, very crudely and indistinctly, that propagation by true fertilization will turn out to be a sort of mixture, and not true fusion, of two distinct individuals, or rather of innumerable individuals, as each parent has its parents and ancestors. I can understand on no other view the way in which crossed forms go back to so large an extent to ancestral forms.
Darwin later attempted a particulate theory, called pangenesis, in his 1868 book
Variation of Animals and Plants Under Domestication. What counts for the discussion here, though, is not the theory itself, but the fact that he clearly saw the need for it. That nature worked by particulate inheritance, rather than by blending inheritance, was a “cross your fingers” prediction that was not validated until the New Synthesis. It should be considered every bit as bold as Copernicus’ prediction of the great distance of the stars.
Jacques Monod makes this point R. Harre (ed.),
Problems of Scientific Revolution (Oxford University Press); also reprinted in M. Ridley (ed.),
Evolution - An Oxford Reader):
What Jenkin’s remarks called for was a theory of heredity by which inheritance would be essentially discrete, discontinuous, and ensured by units that could be transmitted from generation to generation without losing their somatogenic qualities. Such is the gene… [T]he selective theory of evolution as Darwin himself had stated it, required the discovery of Mendelian genetics, which of course, was made.
and
If it fits, a good theory or a good idea will always be much wider and much richer than even the inventor of the idea may know at his time. The theory may be judged precisely on this type of development, when more and more falls into its lap, even though it was not predictable that so much would come of it.
I will talk about other Darwinian predictions in the “cross your fingers” category in the next post.
