Back to Methuselah was published in 1921, a good 15 years before the modern synthesis got underway, so "Neo-Darwinism" as Shaw uses the term means something different: not Darwin-plus-Mendel, but Darwin-minus-Lamarck. Unlike Darwin himself, who was willing to grant that Lamarckian processes (the inheritance of acquired characteristics) might play some role, the Neo-Darwinians broke with Lamarck completely and insisted that evolution was driven almost exclusively by what Shaw -- not wishing to profane the name of Nature -- insists on calling Circumstantial Selection.
It's not that Shaw doesn't believe in natural selection -- he grants that it occurs and that it influences evolution -- but he considers it to be an incidental process. He thinks of natural selection the way a more orthodox evolutionist thinks of genetic drift: It undeniably happens, but it's not all that important and evolution could go on just fine without it. The real driving force behind evolution is voluntary change.
If you can turn a pedestrian into a cyclist, and a cyclist into a pianist or violinist, without the intervention of Circumstantial [that is, natural] Selection, you can turn an amoeba into a man, or a man into a superman, without it. All of which is rank heresy to the Neo-Darwinian, who imagines that if you stop Circumstantial Selection, you not only stop development but inaugurate a rapid and disastrous degeneration.The details of this process are admittedly a little sketchy, even if we take the heritability of acquired characteristics for granted. It's easy enough to imagine how a giraffe -- the canonical example, which Shaw dutifully trots out ("I do not remember how this animal imposed himself illustratively on the Evolution controversy; but there was no getting away from him then; and I am old-fashioned enough to be unable to get away from him now.") -- might want a longer neck, try to get one by stretching, and succeed in lengthening its neck a bit. But when one tries to picture a cartilaginous fish "trying" to have bones, or a monkey "trying" not to have a tail (not to mention a plant or an amoeba "wanting" or "trying" to do anything at all), the theory seems to break down.
Let us fix the Lamarckian evolutionary process well in our minds. You are alive; and you want to be more alive. You want an extension of consciousness and of power. You want, consequently, additional organs, or additional uses of your existing organs: that is, additional habits. You get them because you want them badly enough to keep trying for them until they come. Nobody knows how: nobody knows why: all we know is that the thing actually takes place.
For Shaw, the important thing about Lamarckism is not the inheritance of acquired characteristics, but the importance of will in the evolutionary process. In fact, Shaw seems to think that the former depends on the latter -- that acquirements are inherited if and only if they were acquired deliberately rather than by accident. (He expresses this in a rather confusing way, saying that only "habits" can be inherited, but makes it clear that he is using "habit" in a special sense which includes not only customary behavior patterns but anatomical features as well. A Shavian "habit" is any feature that is voluntarily acquired and thereafter becomes involuntary and automatic.)
Because of Shaw's focus on the inheritance of voluntarily acquired characteristics, he dismisses the experiments of August Weismann -- in which he cut off the tails of 20 successive generations of rats and observed that their offspring were nevertheless born with tails -- as missing the point. Shaw considers it self-evidently ridiculous to suppose "that injuries or accidents coming from external sources against the will of the victim could possibly establish a [heritable] habit: that, for instance, a family could acquire a habit of being killed in railway accidents." He proposes the following as an alternative experiment which, if it were practicable, would be more relevant to the Lamarckian hypothesis as he understands it.
The scientific form of his experiment would have been something like this. First, he should have procured a colony of mice highly susceptible to hypnotic suggestion. He should then have hypnotizes them into an urgent conviction that the fate of the musque [sic] world depended on the disappearance of its tail, just as some ancient and forgotten experimenter seems to have convinced the cats of the Isle of Man. Having thus made the mice desire to lose their tails with a life-or-death intensity, he would very soon have seen a few mice born with little or no tail. These would be recognized by the other mice as superior beings, and privileged in the division of food and in sexual selection. Ultimately the tailed mice would be put to death as monsters by their fellows, and the miracle of the tailless mouse completely achieved.The odd thing about this -- okay, there are a lot of odd things about it, but one of the odd things about it -- is how thoroughly Darwinian it is. A true Lamarckian would perhaps expect that, once the mice had been suitably hypnotized, they would somehow try very hard to reduce the length of their own tails and would succeed in doing so, if perhaps only to a very slight degree. (Exactly how this would be done is, as I have said, not so clear.) Their children would then be born with very slightly shorter tails, which they in their turn would shorten a bit by the same method, and after many repetitions of this process a generation of tailless mice would finally be produced.
The objection to this experiment is not that it seems too funny to be taken seriously, and is not cruel enough to overawe the mob, but simply that it is impossible because the human experimenter cannot get at the mouse's mind.
Shaw predicts something completely different. Instead of the mice changing their own bodies by willpower and then passing on those changes to their children, he imagines that the mice's desire for taillessness would somehow cause a few tailless mutants to appear a generation or two later, and that the tailless mutation would become the norm by means of a process which can only be described as eugenics -- that is, self-imposed artificial selection, which is nothing more than a special case of Darwinian natural selection. The bit about the mutation arising "very soon" as a result of the mice's desire (rather than arising eventually by chance) is the only hint of anything non-Darwinian in Shaw's story.
Unlike the mouse story in the preface, which is not Lamarckian at all, the Back to Methuselah plays themselves do feature the inheritance of acquired characteristics. In the story, a few people read a book arguing that the human lifespan must be extended to at least 300 years, and as a result they themselves -- not the next generation, as in the case of the mice -- go on to live for 300 years! Somehow their desire directly causes sweeping physiological changes, which are then inherited by their children. The implication is that, had they instead read a book arguing that humans all ought to be nine feet tall, they could simply have taken thought and added the requisite cubits to their stature. (The physiological changes implied in increased longevity are internal and invisible, which helps make the story seem a little less obviously ridiculous. That's probably why the mouse story, featuring a more obvious physical feature, used a different mechanism. It would be too clearly bogus if the mice's own tails had simply disappeared after the hypnosis.) After that, eugenics -- in the form of sexual selection and genocide -- once again takes over. The long-lived people seek each other out as mates "for the good of the race," and eventually they decide to kill off all the short-lived ones. No matter how hard he tries to be a good Lamarckian, Shaw's imagination keeps being drawn back to Darwinian mechanisms.