More Articles on Evolution
The Limits of Darwinism
The parts of an irreducibly complex system--a padlock, say--must work together
if the system is to work at all. Such systems, Mike Behe has argued, are inaccessible
to a Darwinian mechanism.
H. Allen Orr confidently demurs, the shade of H. J. Muller having persuaded
him that irreducible complexity is itself an evolutionary property,
something that simply pops up in the evolutionary record, like the vermiform
appendix. Herewith, then, the Darwinian padlock. A key-like structure
(a slug, perhaps) makes an adventitious appearance, followed by the equally
adventitious appearance of a lock. The key then obligingly adapts itself to
the lock, creating the familiar irreducibly complex padlock.
It is a defect of this scheme that the requisite steps involve two useless
artifacts and a miracle.
I have, of course, rigged the example for rhetorical effect, but Orr's argument
is in any case an exercise is misapplied force, like a screw hammered into
wood. Whatever the definition of irreducible complexity, some systems must
remain inaccessible to a Darwinian mechanism if the underlying theory is to
have empirical content. Witness the lock and key. A universal evolutionary
mechanism is no more plausible than a universal differential equation. The
only question of interest is whether some biological structures are included
among the systems inaccessible to every Darwinian mechanism. The conceptual
question aside, the incontrovertible fact remains that in biochemistry the
requisite Darwinian intermediates are missing from the literature and nowhere
to be seen in life. This is the gravamen of Behe's argument. And it is to
Behe's credit that he has made it.
Having for many years declined to discuss the obvious, evolutionary biologists
now argue that the facts to which Behe is drawing attention represent merely
an irregularity in the otherwise smooth manifold of Darwinian successes.
Their response would, of course, be more persuasive if those successes were
more conspicuous. The achievements of contemporary biology are very considerable.
In this, Orr is entirely correct. But it has been the molecular biologists
and biochemists who have discovered an entirely new, entirely limpid, entirely
lovely and astonishing new world, this even as evolutionary biologists have
continued to trot restlessly across the same dark plain, baying at shadows
in the endless night.
I very much enjoyed the appetizers of Berwick's discussion, but found myself
disappointed that the main course appears to be missing.
I shall supply it myself. On n'est jamais aussi bien servi que par soi
mÍme. Richard Dawkins has set himself the formidable task of providing
an explanation of biological complexity. Yet Climbing Mount Improbable
contains neither a serious study of any organ nor a sober analytical discussion
of any theory. The argument in favor of random variation and natural selection
is sustained chiefly by the conviction that it has been defended elsewhere,
a point of view evident also in Dennett's Darwin's Dangerous Idea.
The problems confronting evolutionary theory are by no means trivial. We
are largely unable to characterize what evolution has accomplished
and so logically in no position to determine whether evolution has accomplished
it. No very precise definition of an organism is at hand: indeed, no definition
at all. Many properties of living creatures remain resistant to specification,
let alone explanation, a fact that becomes embarrassingly evident whenever
attempts are made to simulate aspects of animal life. Computer models exist
that mimic bird flocking, but no one has the faintest idea how to represent
algorithmically the fact that the behavior of higher organisms is in some
measure autonomous. Our understanding of various biological systems is profoundly
incomplete. Vision is a case in point. Inasmuch as the visual system issues
in a visual experience, a satisfactory theory of vision must resolve
the problem of consciousness. If we knew the laws governing evolution, we
should be able mathematically to describe a Darwinian mechanism and by specifying
its initial state and boundary conditions see certain portions of the evolutionary
record repeat themselves. This would offer biologists the opportunity to do
what is commonly done in astrophysics, and that is to compare such simulations
with the real world. We can do no such thing. The discipline hilariously described
as artificial life, a worthy successor, no doubt, to artificial
intelligence, has succeeded only in showing that where simulations proceed
according to Darwinian algorithms, nothing life-like can be expected; and
where something life-like can be expected, simulations have not proceeded
according to Darwinian algorithms.
From a mathematical point of view, Darwinian theories appear far too weak
to have brought about the remarkable structures evident in living creatures.
Finite state automata and finite state Markov processes are systems of generous
and demonstrable inadequacy in psychology, a circumstance that evolutionary
biologists tend to ignore, perhaps because it embodies facts with which they
are not acquainted.
Curiously enough, biologists are sometimes prepared to endorse these harsh
conclusions, as long as the gross deficiencies in Darwinian theory may respectfully
be represented as work in progress. In a recent issue of Science,1
a quartet of theoretical biologists survey what is currently absent from evolutionary
theory. Interesting mathematical models. An explanation for the evolutionary
increase in information and complexity. Any plausible account of biological
organization. An explanation of the major macro-evolutionary changes: the
emergence of life, the appearance of eukaryotic cells, the human capacity
for language. Theories of variation and theories of development.
In the largest sense, we do not know whether life is determined by the laws
of physics and chemistry, or whether it is consistent with but independent
of those laws, or whether it is the expression of a profound but inscrutable
design. Dogmatic assertions in the face of ignorance express nothing more
than a lack of interest in rational inquiry. It is for this reason that otherwise
sympathetic observers have come to the conclusion that Richard Dawkins is
motivated as much by a theological agenda as scientific curiosity.
Now that is the main course I would have wished to see.
1 Torbjorn Faserstrom et al., "Biologists Put on Mathematical
Glasses," in Science 274 (December 20, 1996): 2039-40.