"An inventor can advance step by step in the construction of an airplane even if he is only experimenting with sticks and scraps of metal in his own backyard. But he cannot watch the Missing Link evolving in his own backyard. If he has made a mistake in his calculations, the airplane will correct it by crashing to the ground. But if he has made a mistake about the arboreal habitat of his ancestor, he cannot see his arboreal ancestor falling off the tree. "
See the full link here.
I don't think this argument is as cogent as it once was, because with the advent of molecular biology and genetics we are in principle able to avail ourselves of "universals" (laws) governing the behaviour of some aspects of that which we are attempting to model.
However, the warning is true insofar as evolution is something taking place in vivo rather than in vitro. So if one oversimplifies the models, does not correctly model the environment, what have you, the specific predictions may well go awry, and one is left with something which isn't even semi-empirical, but is more ex post facto hand waving...
And it is there that Chesterton's warning still stands.
Speaking of which, Chesterton inluded in one of his short stories The Strange Crime of John Boulnois) the concept of catastrophism in which he posited (as a minor part of the plot) that evolution occurred more or less in fits and starts, rather than gradually and at a more or less constant rate...the book was written in 1914.
Cheers!
Thanks so much, grey_whiskers, for the link to Chesterton's Everlasting Man. I haven't read Chesterton in years; probably it's time to reread him, he is so penetrating and wise. I especially appreciate his remark that modern man cannot stand agnosticism; and that this word came into currency about the time of Darwin's theory. You've heard the old saw: Who ceases to believe in God does not then believe in nothing; He'll believe in anything.
BTW, I think Chesterton's argument is just as cogent as ever, notwithstanding the advent of molecular biology and genetics, which "in principle" give us access to universals that may apply to govern "some aspects" of behavior we are attempting to model. It seems the point is we can't model any system completely because we don't and can't know everything about it as it evolves in space and time. All we can do is model some aspects of its behavior occurring at the space/time point of observation. There is no way we can say we know the whole system, only this or that aspect of it.
The other thing I wonder about is experimental tests require us to isolate the object of study out of its context. So we lose all sense that the object is a contingent object -- i.e., it does not stand alone in nature, it is not self-given or independent, but part of a greater dynamic whole. Though we seek "certainty," it is always completely beyond our grasp. So agnosticism creeps in -- which humans apparently cannot abide. Sometimes they'll concoct "just-so stories" to relieve themselves of it. Chesterton traces the route from agnosticism to dogma in Everlasting Man....
Thank you ever so much for writing, grey-whiskers! And again, for the great link!