ABOUT THE AUTHOR
Daniel C. Dennett is Distinguished Professor of Arts and Sciences and
Director of the Center for Cognitive Studies at Tufts University,
Massachusetts. He is also the author of Content and Consciousness (1969);
Brainstorms (1978; Penguin, 1997); Elbow Room (1984); The Intentional
Stance (1987); Consciousness Explained (1992; Penguin, 1993); and Kinds
of Minds (1996).
DARWIN'S
DANGEROUS
IDEA
EVOLUTION AND THE
MEANINGS OF LIFE
Daniel C. Dennett

PENGUIN BOOKS
Published by the Penguin Group
Penguin Books Ltd, 27 Wrights Lane, London W8 5TZ, England
Penguin Books USA Inc., 375 Hudson Street, New York, New York 10014, USA
Penguin Books Australia Ltd, Ringwood, Victoria, Australia
Penguin Books Canada Ltd, 10 Alcorn Avenue, Toronto, Ontario, Canada M4V 3B2
Penguin Books (NZ) Ltd, 182-190 Wairau Road, Auckland 10, New Zealand
Penguin Books Ltd, Registered Offices: Harmondsworth, Middlesex, England
First published in the USA by Simon & Schuster 1995
First published in Great Britain by Allen Lane The Penguin Press 1995
Published in Penguin Books 1996
3579 10 864
Copyright © Daniel C. Dennett, 1995
All rights reserved
The acknowledgements on p. 587 constitute an extension of this copyright page
The moral right of the author has been asserted
Printed in England by Clays Ltd, St Ives pic
Except in the United States of America, this book is sold subject
to the condition that it shall not, by way of trade or otherwise, be lent,
re-sold, hired out, or otherwise circulated without the publisher's
prior consent in any form of binding or cover other than that in
which it is published and without a similar condition including this
condition being imposed on the subsequent purchaser
To VAN QUINE
teacher and friend

Contents
Preface
PART I: STARTING IN THE MIDDLE
CHAPTER ONE
Tell Me Why
1. Is Nothing Sacred? 17
2. What, Where, When, Why—and How? 23
3. Locke's "Proof" of the Primacy of Mind 26
4. Hume's Close Encounter 28
CHAPTER TWO
An Idea Is Born
1. What Is So Special About Species? 35
2. Natural Selection—an Awful Stretcher 39
3. Did Darwin Explain the Origin of Species? 42
4. Natural Selection as an Algorithmic Process 48
5. Processes as Algorithms 52
CHAPTER THREE
Universal Acid
1. Early Reactions 61
2. Darwin's Assault on the Cosmic Pyramid 64
3. The Principle of the Accumulation of Design 68
4. The Tools for R and D: Skyhooks or Cranes? 73
5. Who's Afraid of Reductionism? 80

8 CONTENTS Contents 9

CHAPTER FOUR
The Tree of Life 85
1. How Should We Visualize the Tree of Life? 85
2. Color-coding a Species on the Tree 91
3. Retrospective Coronations: Mitochondrial Eve and
Invisible Beginnings 96
4. Patterns, Oversimplification, and Explanation 100
CHAPTER FIVE
The Possible and the Actual 104
1. Grades of Possibility? 104
2. The Library of Mendel 107
3. The Complex Relation Between Genome and Organism 113
4. Possibility Naturalized 118
CHAPTER SIX
Threads of Actuality in Design Space 124
1. Drifting and Lifting Through Design Space 124
2. Forced Moves in the Game of Design 128
3. The Unity of Design Space 135
PART II: DARWINIAN THINKING IN BIOLOGY
CHAPTER SEVEN
Priming Darwin's Pump 149
1. Back Beyond Darwin's Frontier 149
2. Molecular Evolution 155
3. The Laws of the Game of Life 163
4. Eternal Recurrence—Life Without Foundations? 181
CHAPTER EIGHT
Biology Is Engineering 187
1. The Sciences of the Artificial 187
2. Darwin Is Dead—Long Live Darwin! 190
3. Function and Specification 195
4. Original Sin and the Birth of Meaning 200
5. The Computer That Learned to Play Checkers 207
6. Artifact Hermeneutics, or Reverse Engineering 212
7. Stuart Kauffman as Meta-Engineer 220
CHAPTER NINE
Searching for Quality
1. The Power of Adaptationist Thinking 229
2. The Leibnizian Paradigm 238
3. Playing with Constraints 251
CHAPTER TEN
Bully for Brontosaurus 262
1. The Boy Who Cried Wolf? 262
2. The Spandrel's Thumb 267
3. Punctuated Equilibrium: A Hopeful Monster 282
4. Tinker to Evers to Chance: The Burgess Shale
Double-Play Mystery 299
CHAPTER ELEVEN
Controversies Contained 313
1. A Clutch of Harmless Heresies 313
2. Three Losers: Teilhard, Lamarck, and Directed
Mutation 320
3. CuiBono? 324
PART III: MIND, MEANING, MATHEMATICS, AND MORALITY
CHAPTER TWELVE
The Cranes of Culture 335
1. The Monkey's Uncle Meets the Meme 335
2. Invasion of the Body-Snatchers 342
3. Could There Be a Science of Memetics? 352
4. The Philosophical Importance of Memes 361
CHAPTER THIRTEEN
Losing Our Minds to Darwin 370
1. The Role of Language in Intelligence 370
2. Chomsky Contra Darwin: Four Episodes 384 3.
Nice Tries 393
CHAPTER FOURTEEN
The Evolution of Meanings 401
1. The Quest for Real Meaning 401
2. Two Black Boxes 412

10 CONTENTS
3. Blocking the Exits 419
4. Safe Passage to the Future 422
CHAPTER FIFTEEN
The Emperor's New Mind, and Other Fables 428
1. The Sword in the Stone 428
2. The Library of Toshiba 437
3. The Phantom Quantum-Gravity Computer:
Lessons from Lapland 444
CHAPTER SIXTEEN
On the Origin of Morality 452
1. E Pluribus Unum? 453
2. Friedrich Nietzsche's Just So Stories 461
3. Some Varieties of Greedy Ethical Reductionism 467
4. Sociobiology: Good and Bad, Good and Evil 481
CHAPTER SEVENTEEN
Redesigning Morality 494
1. Can Ethics Be Naturalized? 494
2. Judging the Competition 501
3. The Moral First Aid Manual 505
CHAPTER EIGHTEEN
The Future of an Idea 511
1. In Praise of Biodiversity 511
2. Universal Acid: Handle with Care 521
Darwin's theory of evolution by natural selection has always fascinated me,
but over the years I have found a surprising variety of thinkers who cannot
conceal their discomfort with his great idea, ranging from nagging skepti-
cism to outright hostility. I have found not just lay people and religious
thinkers, but secular philosophers, psychologists, physicists, and even biol-
ogists who would prefer, it seems, that Darwin were wrong. This book is
about why Darwin's idea is so powerful, and why it promises—not threat-
ens—to put our most cherished visions of life on a new foundation.
A few words about method. This book is largely about science but is not
itself a work of science. Science is not done by quoting authorities, however
eloquent and eminent, and then evaluating their arguments. Scientists do,
however, quite properly persist in holding forth, in popular and not-so-
popular books and essays, putting forward their interpretations of the work
in the lab and the field, and trying to influence their fellow scientists. When
I quote them, rhetoric and all, I am doing what they are doing: engaging in
persuasion. There is no such thing as a sound Argument from Authority, but
authorities can be persuasive, sometimes rightly and sometimes wrongly. I
try to sort this all out, and I myself do not understand all the science that is
relevant to the theories I discuss, but, then, neither do the scientists (with
perhaps a few polymath exceptions). Interdisciplinary work has its risks. I
have gone into the details of the various scientific issues far enough, I hope,
to let the uninformed reader see just what the issues are, and why I put the
interpretation on them that I do, and I have provided plenty of references.
Names with dates refer to full references given in the bibliography at the
back of the book. Instead of providing a glossary of the technical terms used,
I define them briefly when I first use them, and then often clarify their
meaning in later discussion, so there is a very extensive index, which will let
you survey all occurrences of any term or idea in the book. Footnotes are
for digressions that some but not all readers will appreciate or require.
Preface

12 PREFACE
One thing I have tried to do in this book is to make it possible for you to
read the scientific literature I cite, by providing a unified vision of the field,
along with suggestions about the importance or non-importance of the
controversies that rage. Some of the disputes I boldly adjudicate, and others I
leave wide open but place in a framework so that you can see what the issues
are, and whether it matters—to you—how they come out. I hope you will
read this literature, for it is packed with wonderful ideas. Some of the books I
cite are among the most difficult books I have ever read. I think of the books
by Stuart Kauffman and Roger Penrose, for instance, but they are
pedagogical tours deforce of highly advanced materials, and they can and
should be read by anyone who wants to have an informed opinion about the
important issues they raise. Others are less demanding—clear, informative,
well worth some serious effort—and still others are not just easy to read but a
great delight—superb examples of Art in the service of Science. Since you
are reading this book, you have prqbably already read several of them, so my
grouping them together here will be recommendation enough: the books by
Graham Cairns-Smith, Bill Calvin, Richard Dawkins, Jared Diamond, Manfred
Eigen, Steve Gould, John Maynard Smith, Steve Pinker, Mark Ridley, and Matt
Ridley. No area of science has been better served by its writers than
evolutionary theory.
Highly technical philosophical arguments of the sort many philosophers
favor are absent here. That is because I have a prior problem to deal with. I
have learned that arguments, no matter how watertight, often fall on deaf
ears. I am myself the author of arguments that I consider rigorous and
unanswerable but that are often not so much rebutted or even dismissed as
simply ignored. I am not complaining about injustice—we all must ignore
arguments, and no doubt we all ignore arguments that history will tell us we
should have taken seriously. Rather, I want to play a more direct role in
changing what is ignorable by whom. I want to get thinkers in other disci-
plines to take evolutionary thinking seriously, to show them how they have
been underestimating it, and to show them why they have been listening to
the wrong sirens. For this, I have to use more artful methods. I have to tell a
story. You don't want to be swayed by a story? Well, I know you won't be
swayed by a formal argument; you won't even listen to a formal argument for
my conclusion, so I start where I have to start.
The story I tell is mostly new, but it also pulls together bits and pieces
from a wide assortment of analyses I've written over the last twenty-five
years, directed at various controversies and quandaries. Some of these pieces
are incorporated into the book almost whole, with improvements, and others
are only alluded to. What I have made visible here is enough of the tip of the
iceberg, I hope, to inform and even persuade the newcomer and at least
challenge my opponents fairly and crisply. I have tried to navigate between
the Scylla of glib dismissal and the Charybdis of grindingly detailed
Preface 13
infighting, and whenever I glide swiftly by a controversy, I warn that I am
doing so, and give the reader references to the opposition. The bibliography
could easily have been doubled, but I have chosen on the principle that any
serious reader needs only one or two entry points into the literature and can
find die rest from there.

In the front of his marvelous new book, Metaphysical Myths, Mathematical
Practices: The Ontology and Epistemology of the Exact Sciences (Cam-
bridge: Cambridge University Press, 1994), my colleague Jody Azzouni
thanks "the philosophy department at Tufts University for providing a near-
perfect environment in which to do philosophy." I want to second both the
thanks and the evaluation. At many universities, philosophy is studied but not
done—"philosophy appreciation," one might call it—and at many other
universities, philosophical research is an arcane activity conducted out of
sight of the undergraduates and all but the most advanced postgraduates. At
Tufts, we do philosophy, in the classroom and among our colleagues, and the
results, I think, show that Azzouni's assessment is correct. Tufts has provided
me with excellent students and colleagues, and an ideal setting in which to
work with them. In recent years I have taught an undergraduate seminar on
Darwin and philosophy, in which most of the ideas in this book were
hammered out. The penultimate draft was probed, criticized, and polished by
a particularly strong seminar of graduate and undergraduate students, for
whose help I am grateful: Karen Bailey, Pascal Buckley, John Cabral, Brian
Cavoto, Tim Chambers, Shiraz Cupala, Jennifer Fox, Angela Giles, Patrick
Hawley, Dien Ho, Matthew Kessler, Chris Lerner, Kristin McGuire, Michael
Ridge, John Roberts, Lee Rosenberg, Stacey Schmidt, Rhett Smith, Laura
Spiliatakou, and Scott Tanona. The seminar was also enriched by frequent
visitors: Marcel Kinsbourne, Bo Dahlbom, David Haig, Cynthia
Schossberger, Jeff McConnell, David Stipp. I also want to thank my
colleagues, especially Hugo Bedau, George Smith, and Stephen White, for a
variety of valuable suggestions. And I must especially thank Alicia Smith, the
secretary at the Center for Cognitive Studies, whose virtuoso performance as
a reference-finder, fact-checker, permission-seeker, draft-updater/printer/
mailer, and general coordinator of the whole project put wings on my heels.
I have also benefited from detailed comments from those who read most or
all the penultimate-draft chapters: Bo Dahlbom, Richard Dawkins, David
Haig, Doug Hofstadter, Nick Humphrey, Ray Jackendoff, Philip Kitcher, Jus-
tin Leiber, Ernst Mayr, Jeff McConnell, Steve Pinker, Sue Stafford, and Kim
Sterelny. As usual, they are not responsible for any errors they failed to
dissuade me from. (And if you can't write a good book about evolution witii
the help of this sterling group of editors, you should give up!)
Many others answered crucial questions, and clarified my thinking in

14 PREFACE
dozens of conversations: Ron Amundsen, Robert Axelrod, Jonathan Bennett,
Robert Brandon, Madeline Caviness, Tim Clutton-Brock, Leda Cosmides,
Helena Cronin, Arthur Danto, Mark De Voto, Marc Feldman, Murray Gell-
Mann, Peter Godfrey-Smith, Steve Gould, Danny Hillis, John Holland, Alas-
tair Houston, David Hoy, Bredo Johnsen, Stu Kauffman, Chris Langton, Dick
Lewontin, John Maynard Smith, Jim Moore, Roger Penrose, Joanne Phillips,
Robert Richards, Mark and Matt (the Ridley conspecifics), Dick Schacht, Jeff
Schank, Elliot Sober, John Tooby, Robert Trivers, Peter Van Inwagen, George
Williams, David Sloan Wilson, Edward O. Wilson, and BUI Wimsatt.
I want to thank my agent, John Brockman, for steering this big project past
many shoals, and helping me see ways of making it a better book. Thanks
also go to Terry Zaroff, whose expert copyediting caught many slips and
inconsistencies, and clarified and unified the expression of many points. And
Ilavenil Subbiah, who drew the figures, except for Figures 10.3 and 10.4,
which were created by Mark McConnell on a Hewlett-Packard Apollo
workstation, using I-dea.
Last and most important: thanks and love to my wife, Susan, for her
advice, love, and support.
DANIEL DENNETT
September 1994

PART 1
STARTING IN THE
MIDDLE
Neurath has likened science to a boat which, if we are to rebuild it, we
must rebuild plank by plank while staying afloat in it. The philosopher
and the scientist are in the same boat....
Analyze theory-building how we will, we all must start in die middle.
Our conceptual firsts are middle-sized, middle-distanced objects, and
our introduction to diem and to everything comes midway in the
cultural evolution of die race. In assimilating this cultural fare we are
litde more aware of a distinction between report and invention, sub-
stance and style, cues and conceptualization, than we are of a distinc-
tion between die proteins and the carbohydrates of our material intake.
Retrospectively we may distinguish the components of theory-building,
as we distinguish the proteins and carbohydrates while subsisting on
diem.
—WILURD VAN ORMAN QUINE I960, pp. 4-6

1. Is NOTHING SACRED?
We used to sing a lot when I was a child, around the campfire at summer
camp, at school and Sunday school, or gathered around the piano at home.
One of my favorite songs was "Tell Me Why." (For those whose personal
memories don't already embrace this little treasure, the music is provided in
the appendix. The simple melody and easy harmony line are surprisingly
beautiful.)
Tell me why the stars do shine,
Tell me why the ivy twines,
Tell me why die sky's so blue.
Then I will tell you just why I love you.
Because God made the stars to shine, Because
God made the ivy twine, Because God made
the sky so blue. Because God made you, that's
why I love you.
This straightforward, sentimental declaration still brings a lump to my
throat—so sweet, so innocent, so reassuring a vision of life!
And then along comes Darwin and spoils the picnic. Or does he? That is
the topic of this book. From the moment of the publication of Origin of
Species in 1859, Charles Darwin's fundamental idea has inspired intense
reactions ranging from ferocious condemnation to ecstatic allegiance, some-
times tantamount to religious zeal. Darwin's theory has been abused and
misrepresented by friend and foe alike. It has been misappropriated to lend
scientific respectability to appalling political and social doctrines. It has been
pilloried in caricature by opponents, some of whom would have it
CHAPTER ONE
Tell Me Why

20 TELL ME WHY Is Nothing Sacred? 21

tists and other thinkers of his day, there really were large gaps in his theory
that have only recently begun to be properly filled in. The biggest gap looks
almost comical in retrospect. In all his brilliant musings, Darwin never hit
upon the central concept, without which the theory of evolution is hopeless:
the concept of a gene. Darwin had no proper unit of heredity, and so his
account of the process of natural selection was plagued with entirely rea-
sonable doubts about whether it would work. Darwin supposed that offspring
would always exhibit a sort of blend or average of their parents' features.
Wouldn't such "blending inheritance" always simply average out all differ-
ences, turning everything into uniform gray? How could diversity survive
such relentless averaging? Darwin recognized the seriousness of this chal-
lenge, and neither he nor his many ardent supporters succeeded in responding
with a description of a convincing and well-documented mechanism of
heredity that could combine traits of parents while maintaining an underlying
and unchanged identity. The idea they needed was right at hand, uncovered
("formulated" would be too strong) by the monk Gregor Mendel and
published in a relatively obscure Austrian journal in 1865, but, in the best-
savored irony in the history of science, it lay there unnoticed until its im-
portance was appreciated (at first dimly) around 1900. Its triumphant
establishment at the heart of the "Modern Synthesis" (in effect, the synthesis
of Mendel and Darwin) was eventually made secure in the 1940s, thanks to
the work of Theodosius Dobzhansky, Julian Huxley, Ernst Mayr, and others.
It has taken another half-century to iron out most of the wrinkles of that new
fabric.
The fundamental core of contemporary Darwinism, the theory of DNA-
based reproduction and evolution, is now beyond dispute among scientists. It
demonstrates its power every day, contributing crucially to the explanation of
planet-sized facts of geology and meteorology, through middle-sized facts of
ecology and agronomy, down to the latest microscopic facts of genetic
engineering. It unifies all of biology and the history of our planet into a
single grand story. Like Gulliver tied down in Lilliput, it is unbudge-able, not
because of some one or two huge chains of argument that might— hope
against hope—have weak links in them, but because it is securely tied by
hundreds of thousands of threads of evidence anchoring it to virtually every
other area of human knowledge. New discoveries may conceivably lead to
dramatic, even "revolutionary" shifts in the Darwinian theory, but the hope
that it will be "refuted" by some shattering breakthrough is about as
reasonable as the hope that we will return to a geocentric vision and discard
Copernicus.
Still, the theory is embroiled in remarkably hot-tempered controversy, and
one of the reasons for this incandescence is that these debates about scientific
matters are usually distorted by fears that the "wrong" answer would have
intolerable moral implications. So great are these fears that they
are carefully left unarticulated, displaced from attention by several layers of
distracting rebuttal and counter-rebuttal. The disputants are forever changing
the subject slightly, conveniently keeping the bogeys in the shadows. It is
this misdirection that is mainly responsible for postponing the day when we
can all live as comfortably with our new biological perspective as we do with
the astronomical perspective Copernicus gave us.
Whenever Darwinism is the topic, the temperature rises, because more is at
stake than just the empirical facts about how life on Earth evolved, or the
correct logic of the theory that accounts for those facts. One of the precious
things that is at stake is a vision of what it means to ask, and answer, the
question "Why?" Darwin's new perspective turns several traditional assump-
tions upside down, undermining our standard ideas about what ought to count
as satisfying answers to this ancient and inescapable question. Here science
and philosophy get completely intertwined. Scientists sometimes deceive
themselves into thinking that philosophical ideas are only, at best,
decorations or parasitic commentaries on the hard, objective triumphs of
science, and that they themselves are immune to the confusions that phi-
losophers devote their lives to dissolving. But there is no such thing as
philosophy-free science; there is only science whose philosophical baggage
is taken on board without examination.
The Darwinian Revolution is both a scientific and a philosophical revo-
lution, and neither revolution could have occurred without the other. As we
shall see, it was the philosophical prejudices of the scientists, more than their
lack of scientific evidence, that prevented them from seeing how the theory
could actually work, but those philosophical prejudices that had to be
overthrown were too deeply entrenched to be dislodged by mere philo-
sophical brilliance. It took an irresistible parade of hard-won scientific facts
to force thinkers to take seriously the weird new outlook that Darwin
proposed. Those who are still ill-acquainted with that beautiful procession
can be forgiven their continued allegiance to the pre-Darwinian ideas. And
the battle is not yet over; even among the scientists, there are pockets of
resistance.
Let me lay my cards on the table. If I were to give an award for the single
best idea anyone has ever had, I'd give it to Darwin, ahead of Newton and
Einstein and everyone else. In a single stroke, the idea of evolution by
natural selection unifies the realm of life, meaning, and purpose with the
realm of space and time, cause and effect, mechanism and physical law. But
it is not just a wonderful scientific idea. It is a dangerous idea. My admiration
for Darwin's magnificent idea is unbounded, but I, too, cherish many of the
ideas and ideals that it seems to challenge, and want to protect them. For
instance, I want to protect the campfire song, and what is beautiful and true
in it, for my little grandson and his friends, and for their children when they
grow up. There are many more magnificent ideas that are also jeopardized,

30 TELL ME WHY
Hume's Close Encounter 31

ideas in a human mind, we see, by an unknown, inexplicable economy,
arrange themselves so as to form the plan of a watch or house. Experience,
therefore, proves, that there is an original principle of order in mind, not in
matter" (Pt. II).
Note that the Argument from Design depends on an inductive inference:
where there's smoke, there's fire; and where there's design, there's mind. But
this is a dubious inference, Philo observes: human intelligence is
no more than one of the springs and principles of the universe, as well
as heat or cold, attraction or repulsion, and a hundred others, which fall
under daily observation__ But can a conclusion, with any propriety, be
transferred from parts to the whole?... From observing the growth of a
hair, can we learn any thing concerning the generation of a man?...
What peculiar privilege has this little agitation of the brain which we
call thought, that we must thus make it the model of the whole
universe?... Admirable conclusion! Stone, wood, brick, iron, brass have
not, at this time, in this minute globe of earth, an order or arrangement
without human art and contrivance: Therefore the universe could not
originally attain its order and arrangement, without something similar to
human art. [Pt. II.]
Besides, Philo observes, if we put mind as the first cause, with its "unknown,
inexplicable economy," this only postpones the problem:
We are still obliged to mount higher, in order to find the cause of this
cause, which you had assigned as satisfactory and conclusive ___ How
therefore shall we satisfy ourselves concerning the cause of that Being,
whom you suppose the Author of nature, or, according to your system of
anthropomorphism, the ideal world, into which you trace the material?
Have we not the same reason to trace that ideal world into another ideal
world, or new intelligent principle? But if we stop, and go no farther;
why go so far? Why not stop at the material world? How can we satisfy
ourselves without going on in infinitum? And after all, what satisfaction
is there in that infinite progression? [Pt. IV.)
Cleanthes has no satisfactory responses to these rhetorical questions, and
there is worse to come. Cleanthes insists that God's mind is like the human—
and agrees when Philo adds "the liker the better." But, then, Philo presses on,
is God's mind perfect, "free from every error, mistake, or incoherence in his
undertakings" (Pt. V)? There is a rival hypothesis to rule out:
And what surprise must we entertain, when we find him a stupid
mechanic, who imitated others, and copied an art, which, through a long
succession of ages, after multiplied trials, mistakes, corrections,
deliberations, and controversies, had been gradually improving? Many
worlds might have
been botched and bungled, throughout an eternity, ere this system was
struck out: Much labour lost: Many fruitless trials made: And a slow,
but continued improvement carried on during infinite ages of world-
making. (Pt. V.]
When Philo presents this fanciful alternative, with its breathtaking anticipa-
tions of Darwin's insight, he doesn't take it seriously except as a debating foil
to Cleanthes' vision of an all-wise Artificer. Hume uses it only to make a
point about what he saw as the limitations on our knowledge: "In such
subjects, who can determine, where the truth; nay, who can conjecture where
the probability, lies; amidst a great number of hypotheses which may be
proposed, and a still greater number which may be imagined" (Pt. V).
Imagination runs riot, and, exploiting that fecundity, Philo ties Cleanthes up
in knots, devising weird and comical variations on Cleanthes' own hy-
potheses, defying Cleanthes to show why his own version should be pre-
ferred. "Why may not several Deities combine in contriving and framing a
world?... And why not become a perfect anthropomorphite? Why not assert
the Deity or Deities to be corporeal, and to have eyes, a nose, mouth, ears,
etc.?" (Pt. V). At one point, Philo anticipates the Gaia hypothesis: the
universe
bears a great resemblance to an animal or organized body, and seems
actuated with a like principle of life and motion. A continual circulation
of
matter in it produces no disorder ___The world, therefore, I infer, is an
animal, and the Deity is the SOUL of the world, actuating it and actuated
by it. [Pt. VI.]
Or perhaps isn't the world really more like a vegetable than an animal?
In like manner as a tree sheds its seed into the neighboring fields, and
produces other trees; so the great vegetable, the world, or this planetary
system, produces within itself certain seeds, which, being scattered into
the surrounding chaos, vegetate into new worlds. A comet, for instance,
is the seed of a world.... [Pt. VII.]
One more wild possibility for good measure:
The Brahmins assert, that the world arose from an infinite spider, who
spun this whole complicated mass from his bowels, and annihilates
afterwards the whole or any part of it, by absorbing it again, and
resolving it into his own essence. Here is a species of cosmogony,
which appears to us ridiculous; because a spider is a little contemptible
animal, whose operation we are never likely to take for a model of the
whole universe. But still here is

36 AN IDEA IS BORN What Is So Special About Species? 37

Species could also be subdivided, of course, into subspecies or varieties—
cocker spaniels and basset hounds are different varieties of a single species-,
dogs, or Canis familiaris.
How many different kinds of organisms were there? Since no two organ-
isms are exactly alike—not even identical twins—there were as many dif-
ferent kinds of organisms as there were organisms, but it seemed obvious that
the differences could be graded, sorted into minor and major, or accidental
and essential. Thus Aristotle had taught, and this was one bit of philosophy
that had permeated the thinking of just about everybody, from cardinals to
chemists to costermongers. All things—not just living things— had two kinds
of properties: essential properties, without which they wouldn't be the
particular kind of thing they were, and accidental properties, which were free
to vary within the kind. A lump of gold could change shape ad lib and still be
gold; what made it gold were its essential properties, not its accidents. With
each kind went an essence. Essences were definitive, and as such they were
timeless, unchanging, and all-or-nothing. A thing couldn't be rather silver or
quasi-gold or a semi'-mammal.
Aristotle had developed his theory of essences as an improvement on
Plato's theory of Ideas, according to which every earthly thing is a sort of
imperfect copy or reflection of an ideal exemplar or Form that existed
timelessly in the Platonic realm of Ideas, reigned over by God. This Platonic
heaven of abstractions was not visible, of course, but was accessible to Mind
through deductive thought. What geometers thought about, and proved
theorems about, for instance, were the Forms of the circle and the triangle.
Since there were also Forms for the eagle and the elephant, a deductive
science of nature was also worth a try. But just as no earthly circle, no matter
how carefully drawn with a compass, or thrown on a potter's wheel, could
actually be one of the perfect circles of Euclidean geometry, so no actual
eagle could perfectly manifest the essence of eaglehood, though every eagle
strove to do so. Everything that existed had a divine specification, which
captured its essence. The taxonomy of living things Darwin inherited was
thus itself a direct descendant, via Aristotle, of Plato's essen-tialism. In fact,
the word "species" was at one point a standard translation of Plato's Greek
word for Form or Idea, eidos.
We post-Darwinians are so used to thinking in historical terms about the
development of life forms that it takes a special effort to remind ourselves
that in Darwin's day species of organisms were deemed to be as timeless as
the perfect triangles and circles of Euclidean geometry. Their individual
members came and went, but the species itself remained unchanged and
unchangeable. This was part of a philosophical heritage, but it was not an idle
or ill-motivated dogma. The triumphs of modern science, from Copernicus
and Kepler, Descartes and Newton, had all involved the application of precise
mathematics to the material world, and this apparently requires
abstracting away from the grubby accidental properties of things to find their
secret mathematical essences. It makes no difference what color or shape a
thing is when it comes to the thing's obeying Newton's inverse-square law of
gravitational attraction. All that matters is its mass. Similarly, alchemy had
been succeeded by chemistry once chemists settled on their fundamental
creed: There were a finite number of basic, immutable elements, such as
carbon, oxygen, hydrogen, and iron. These might be mixed and united in
endless combinations over time, but the fundamental building blocks were
identifiable by their changeless essential properties.
The doctrine of essences looked like a powerful organizer of the world's
phenomena in many areas, but was it true of every classification scheme one
could devise? Were there essential differences between hills and mountains,
snow and sleet, mansions and palaces, violins and violas? John Locke and
others had developed elaborate doctrines distinguishing real essences from
merely nominal essences; the latter were simply parasitic on the names or
words we chose to use. You could set up any classification scheme you
wanted; for instance, a kennel club could vote on a defining list of necessary
conditions for a dog to be a genuine Ourkind Spaniel, but this would be a
mere nominal essence, not a real essence. Real essences were discoverable
by scientific investigation into the internal nature of things, where essence
and accident could be distinguished according to principles. It was hard to
say just what the principled principles were, but with chemistry and physics
so handsomely falling into line, it seemed to stand to reason that there had to
be denning marks of the real essences of living things as well.
From the perspective of this deliciously crisp and systematic vision of the
hierarchy of living things, there were a considerable number of awkward and
puzzling facts. These apparent exceptions were almost as troubling to
naturalists as the discovery of a triangle whose angles didn't quite add up to
180 degrees would have been to a geometer. Although many of the taxo-
nomic boundaries were sharp and apparently exceptionless, there were all
manner of hard-to-classify intermediate creatures, who seemed to have por-
tions of more than one essence. There were also the curious higher-order
patterns of shared and unshared features: why should it be backbones rather
than feathers that birds and fish shared, and why shouldn't creature with eyes
or carnivore be as important a classifier as warmblooded creature? Although
the broad outlines and most of die specific rulings of taxonomy were
undisputed (and remain so today, of course), there were heated controversies
about the problem cases. Were all these lizards members of die same species,
or of several different species? Which principle of classification should
"count"? In Plato's famous image, which system "carved nature at the
joints"?
Before Darwin, these controversies were fundamentally ill-formed, and
could not yield a stable, well-motivated answer because there was no back-

38 AN IDEA IS BORN Natural Selection—an Awful Stretcher 39

ground theory of why one classification scheme would count as getting the
joints right—the way things really were. Today bookstores face the same sort
of ill-formed problem: how should the following categories be cross-
organized: best-sellers, science fiction, horror, garden, biography, novels,
collections, sports, illustrated books? If horror is a genus of fiction, then true
tales of horror present a problem. Must all novels be fiction? Then the
bookseller cannot honor Truman Capote's own description of In Cold Blood
(1965) as a nonfiction novel, but the book doesn't sit comfortably amid either
the biographies or the history books. In what section of the bookstore should
the book you are reading be shelved? Obviously there is no one Right Way to
categorize books—nominal essences are all we will ever find in this domain.
But many naturalists were convinced on general principles that there were
real essences to be found among the categories of their Natural System of
living things. As Darwin put it, "They believe that it reveals the plan of the
Creator; but unless it be specified whether order in time or space, or what
else is meant by the plan of the Creator, it seems to me that nothing is thus
added to our knowledge" (Origin, p. 413).
Problems in science are sometimes made easier by adding complications.
The development of the science of geology and the discovery of fossils of
manifestly extinct species gave the taxonomists further curiosities to con-
found them, but these curiosities were also the very pieces of the puzzle that
enabled Darwin, working alongside hundreds of other scientists, to discover
the key to its solution: species were not eternal and immutable; they had
evolved over time. Unlike carbon atoms, which, for all one knew, had been
around forever in exactly the form they now exhibited, species had births in
time, could change over time, and could give birth to new species in turn.
This idea itself was not new; many versions of it had been seriously
discussed, going back to the ancient Greeks. But there was a powerful
Platonic bias against it: essences were unchanging, and a thing couldn't
change its essence, and new essences couldn't be born—except of course by
God's command in episodes of Special Creation. Reptiles could no more turn
into birds than copper could turn into gold.
It isn't easy today to sympathize with this conviction, but the effort can be
helped along by a fantasy: consider what your attitude would be towards a
theory that purported to show how the number 7 had once been an even
number, long, long ago, and had gradually acquired its oddness through an
arrangement whereby it exchanged some properties with the ancestors of the
number 10 (which had once been a prime number). Utter nonsense, of course.
Inconceivable. Darwin knew that a parallel attitude was deeply ingrained
among his contemporaries, and that he would have to labor mightily to
overcome it. Indeed, he more or less conceded that the elder authorities of his
day would tend to be as immutable as the species they believed
in, so in the conclusion of his book he went so far as to beseech the support
of his younger readers: "Whoever is led to believe that species are mutable
will do good service by conscientiously expressing his conviction; for only
thus can the load of prejudice by which this subject is overwhelmed be
removed" (Origin, p. 482).
Even today Darwin's overthrow of essentialism has not been completely
assimilated. For instance, there is much discussion in philosophy these days
about "natural kinds," an ancient term the philosopher W. V. O. Quine
(1969) quite cautiously resurrected for limited use in distinguishing good
scientific categories from bad ones. But in the writings of other philosophers,
"natural kind" is often sheep's clothing for the wolf of real essence. The
essentialist urge is still with us, and not always for bad reasons. Science does
aspire to carve nature at its joints, and it often seems that we need essences,
or something like essences, to do the job. On this one point, the two great
kingdoms of philosophical thought, the Platonic and the Aristotelian, agree.
But the Darwinian mutation, which at first seemed to be just a new way of
thinking about kinds in biology, can spread to other phenomena and other
disciplines, as we shall see. There are persistent problems both inside and
outside biology that readily dissolve once we adopt the Darwinian
perspective on what makes a thing the sort of thing it is, but the tradition-
bound resistance to this idea persists.
2. NATURAL SELECTION—AN AWFUL STRETCHER
It is an awful stretcher to believe that a peacock's tail was thus formed;
but, believing it, I believe in the same principle somewhat modified
applied to man.
—CHARLES DARWIN, letter quoted in Desmond and
Moore 1991, p. 553
Darwin's project in Origin can be divided in two: to prove that modern
species were revised descendants of earlier species—species had evolved—
and to show how this process of "descent with modification" had occurred. If
Darwin hadn't had a vision of a mechanism, natural selection, by which this
well-nigh-inconceivable historical transformation could have been ac-
complished, he would probably not have had the motivation to assemble all
the circumstantial evidence that it had actually occurred. Today we can
readily enough imagine proving Darwin's first case—the brute historic fact
of descent with modification—quite independently of any consideration of
Natural selection or indeed any other mechanism for bringing these brute
events about, but for Darwin the idea of the mechanism was both the

42 AN IDEA IS BORN
Did Darwin Explain the Origin of Species? 43

think this cannot be disputed; if there be, owing to the high geometric
powers of increase of each species, at some age, season, or year, a severe
struggle for life, and this certainly cannot be disputed; then, considering
the infinite complexity of the relations of all organic beings to each other
and to their conditions of existence, causing an infinite diversity in struc-
ture, constitution, and habits, to be advantageous to them, I think it would
be a most extraordinary fact if no variation ever had occurred useful to
each being's own welfare, in the same way as so many variations have
occurred useful to man. But if variations useful to any organic being do
occur, assuredly individuals thus characterized will have the best chance
of being preserved in the struggle for life; and from the strong principle of
inheritance they will tend to produce offspring similarly characterized.
This principle of preservation, I have called, for the sake of brevity, Natural
Selection. [Origin, p. 127.]
This was Darwin's great idea, not the idea of evolution, but the idea of
evolution by natural selection, an idea he himself could never formulate with
sufficient rigor and detail to prove, though he presented a brilliant case for it.
The next two sections will concentrate on curious and crucial features of this
summary statement of Darwin's.
3. DID DARWIN EXPLAIN THE ORIGIN OF SPECIES?
Darwin did wrestle brilliantly and triumphantly with the problem of
adaptation, but he had limited success with the issue of diversity— even
though he titled his book with reference to his relative failure: the origin
of species.
—STEPHEN JAY GOULD 1992a, p. 54
Thus die grand fact in natural history of the subordination of group
under group, which, from its familiarity, does not always sufficiently
strike us, is in my judgment fully explained.
—CHARLES DARWIN, Origin, p. 413
Notice that Darwin's summary does not mention speciation at all. It is en-
tirely about the adaptation of organisms, the excellence of their design, not
the diversity. Moreover, on the face of it, this summary takes the diversity of
species as an assumption: "the infinite [sic] complexity of the relations of all
organic beings to each other and to their conditions of existence." What
makes for this stupendous (if not actually infinite ) complexity is the presence
at one and the same time (and competing for the same living space) of so
many different life forms, with so many different needs and strategies. Darwin
doesn't even purport to offer an explanation of the origin of the first species,
or of life itself; he begins in the middle, supposing many different species with
many different talents already present, and claims that starting from such a
mid-stage point, the process he has described will inevitably hone and di-
versify the talents of the species already existing. And will that process create
still further species? The summary is silent on that score, but the book is not.
In fact, Darwin saw his idea explaining both great sources of wonder in a
single stroke. The generation of adaptations and the generation of diversity
were different aspects of a single complex phenomenon, and the unifying
insight, he claimed, was the principle of natural selection.
Natural selection would inevitably produce adaptation, as the summary
makes clear, and under the right circumstances, he argued, accumulated
adaptation would create speciation. Darwin knew full well that explaining
variation is not explaining speciation. The animal-breeders he pumped so
vigorously for their lore knew about how to breed variety within a single
species, but had apparently never created a new species, and scoffed at the
idea that their particular different breeds might have a common ancestor.
"Ask, as 1 have asked, a celebrated raiser of Hereford cattle, whether his
cattle might not have descended from longhorns, and he will laugh you to
scorn." Why? Because "though they well know that each race varies slightly,
for they win their prizes by selecting such slight differences, yet they ignore
all general arguments and refuse to sum up in their minds slight differences
accumulated during many successive generations" (Origin, p. 29).
The further diversification into species would occur, Darwin argued, be-
cause if there was a variety of heritable skills or equipment in a population
(of a single species), these different skills or equipment would tend to have
different payoffs for different subgroups of the population, and hence these
subpopulations would tend to diverge, each one pursuing its favored sort of
excellence, until eventually there would be a complete parting of the ways.
Why, Darwin asked himself, would this divergence lead to separation or
clumping of the variations instead of remaining a more or less continuous
fan-out of slight differences? Simple geographical isolation was part of his
answer; when a population got split by a major geological or climatic event,
or by haphazard emigration to an isolated range such as an island, this
discontinuity in the environment ought to become mirrored eventually in a
discontinuity in the useful variations observable in the two populations. And
once discontinuity got a foothold, it would be self-reinforcing, all the way to
separation into distinct species. Another, rather different, idea of his was that
in intraspecific infighting, a "winner take all" principle would tend to
operate:
For it should be remembered that the competition will generally be most
severe between those forms which are most nearly related to each other
inhabits, constitution and structure. Hence all the intermediate forms

44 AN IDEA IS BORN Did Darwin Explain the Origin of Species? 45

between the earlier and later states, that is between die less and more
improved state of a species, as well as the original parent-species itself, will
generally tend to become extinct. [Origin, p. 121.]
He formulated a variety of other ingenious and plausible speculations on
how and why the relentless culling of natural selection would actually create
species boundaries, but they remain speculations to this day. It has taken a
century of further work to replace Darwin's brilliant but inconclusive
musings on the mechanisms of speciation with accounts that are to some
degree demonstrable. Controversy about the mechanisms and principles of
speciation still persists, so in one sense neither Darwin nor any subsequent
Darwinian has explained the origin of species. As the geneticist Steve Jones
(1993) has remarked, had Darwin published his masterpiece under its
existing title today, "he would have been in trouble with the Trades
Description Act because if there is one thing which Origin of Species is not
about, it is the origin of species. Darwin knew nothing about genetics. Now
we know a great deal, and although the way in which species begin is still a
mystery, it is one with the details filled in."
But the fact of speciation itself is incontestable, as Darwin showed, build-
ing an irresistible case out of literally hundreds of carefully studied and
closely argued instances. That is how species originate: by "descent with
modification" from earlier species—not by Special Creation. So in another
sense Darwin undeniably did explain the origin of species. Whatever the
mechanisms are that operate, they manifestly begin with the emergence of
variety within a species, and end, after modifications have accumulated, with
the birth of a new, descendant species. What start as "well-marked varieties"
turn gradually into "the doubtful category of subspecies; but we have only to
suppose the steps in the process of modification to be more numerous or
greater in amount, to convert these... forms into well-defined species"
(Origin, p. 120).
Notice that Darwin is careful to describe the eventual outcome as the
creation of "well-defined" species. Eventually, he is saying, the divergence
becomes so great that there is just no reason to deny that what we have are
two different species, not merely two different varieties. But he declines to
play the traditional game of declaring what the "essential" difference is:
... it will be seen that I look at the term species, as one arbitrarily given for
the sake of convenience to a set of individuals closely resembling each
other, and that it does not essentially differ from the term variety, which is
given to less distinct and more fluctuating forms. [Origin, p. 52.]
One of the standard marks of species difference, as Darwin fully recog-
nized, is reproductive isolation—there is no interbreeding. It is interbreed-
ing that reunites the splitting groups, mixing their genes and "frustrating" the
process of speciation. It is not that anything wants speciation to happen, of
course (Dawkins 1986a, p. 237), but if the irreversible divorce that marks
speciation is to happen, it must be preceded by a sort of trial separation
period in which interbreeding ceases for one reason or another, so that the
parting groups can move further apart. The criterion of reproductive isolation
is vague at the edges. Do organisms belong to different species when they
can't interbreed, or when they just don't interbreed? Wolves and coyotes and
dogs are considered to be different species, and yet interbreeding does occur,
and—unlike mules, the offspring of horse and donkey—their offspring are not
in general sterile. Dachshunds and Irish wolfhounds are deemed to be of the
same species, but unless their owners provide some distinctly unnatural
arrangements, they are about as reproductively isolated as bats are from
dolphins. The white-tailed deer in Maine don't in fact interbreed with the
white-tailed deer in Massachusetts, since they don't travel that far, but they
surely could if transported, and naturally they count as of the same species.
And finally—a true-life example seemingly made to order for philoso-
phers—consider the herring gulls that live in the Northern Hemisphere, their
range forming a broad ring around the North Pole.
As we look at the herring gull, moving westwards from Great Britain to
North America, we see gulls that are recognizably herring gulls, although
they are a little different from the British form. We can follow them, as
their appearance gradually changes, as far as Siberia. At about this point in
the continuum, the gull looks more like the form that in Great Britain is
called the lesser black-backed gull. From Siberia, across Russia, to northern
Europe, the gull gradually changes to look more and more like the British
lesser black-backed gull. Finally, in Europe, the ring is complete; the two
geographically extreme forms meet, to form two perfectly good species:
die herring and lesser black-backed gull can be both distinguished by their
appearance and do not naturally interbreed. [Mark Ridley 1985, p. 5]
"Well-defined" species certainly do exist—it is the purpose of Darwin's
book to explain their origin—but he discourages us from trying to find a
"principled" definition of the concept of a species. Varieties, Darwin keeps
insisting, are just "incipient species," and what normally turns two varieties
into two species is not the presence of something (a new essence for each
group, for instance ) but the absence of something: the intermediate cases,
which used to be there—which were necessary stepping-stones, you might
say—but have eventually gone extinct, leaving two groups that are in fact
reproductively isolated as well as different in their characteristics.
Origin of Species presents an overwhelmingly persuasive case for Dar-
win's first thesis—the historical fact of evolution as the cause of the origin

48 AN IDEA IS BORN Natural Selection as an Algorithmic Process 49

4. NATURAL SELECTION AS AN ALGORITHMIC PROCESS
What limit can be put to this power, acting during long ages and rigidly
scrutinising the whole constitution, structure, and habits of each crea-
ture,—favouring the good and rejecting the bad? I can see no limit to
this power, in slowly and beautifully adapting each form to the most
complex relations of life.
—CHARLES DARWIN, Origin, p. 469
The second point to notice in Darwin's summary is that he presents his
principle as deducible by a formal argument—if the conditions are met, a
certain outcome is assured.6 Here is the summary again, with some key
terms in boldface.
If, during the long course of ages and under varying conditions of life,
organic beings vary at all in the several parts of their organization, and I
think this cannot be disputed; if there be, owing to the high geometric
powers of increase of each species, at some age, season, or year, a severe
struggle for life, and this certainly cannot be disputed; then, considering
the infinite complexity of the relations of all organic beings to each other
and to their conditions of existence, causing an infinite diversity in struc-
ture, constitution, and habits, to be advantageous to them, I think it
would be a most extraordinary fact if no variation ever had occurred
useful to each being's own welfare, in the same way as so many variations
have occurred useful to man. But if variations useful to any organic
being do occur, assuredly individuals thus characterized will have the
best chance of being preserved in the struggle for life; and from the strong
principle of inheritance they will tend to produce offspring similarly
characterized. This principle of preservation, I have called, for the sake of
brevity, Natural Selection. [Origin, p. 127 (facs. ed. of 1st ed.).]
The basic deductive argument is short and sweet, but Darwin himself
described Origin of Species as "one long argument." That is because it
6. The ideal of a deductive ( or "nomologico-deductive" ) science, modeled on Newtonian
or Galilean physics, was quite standard until fairly recently in the philosophy of science,
so it is not surprising that much effort has been devoted to devising and criticizing various
axiomatizations of Darwin's theory—since it was presumed that in such a formalization
lay scientific vindication. The idea, introduced in this section, that Darwin should be seen,
rather, as postulating that evolution is an algorithmic process, permits us to do justice to
the undeniable a priori flavor of Darwin's thinking without forcing it into the Procrustean
(and obsolete) bed of the nomologico-deductive model. See Sober 1984a and Kitcher
1985a.
consists of two sorts of demonstrations-, the logical demonstration that a
certain sort of process would necessarily have a certain sort of outcome, and
the empirical demonstration that the requisite conditions for that sort of
process had in fact been met in nature. He bolsters up his logical dem-
onstration with thought experiments—"imaginary instances" {Origin, p.
95)—that show how the meeting of these conditions might actually account
for the effects he claimed to be explaining, but his whole argument extends
to book length because he presents a wealth of hard-won empirical detail to
convince the reader that these conditions have been met over and over again.
Stephen Jay Gould (1985) gives us a fine glimpse of the importance of this
feature of Darwin's argument in an anecdote about Patrick Matthew, a
Scottish naturalist who as a matter of curious historical fact had scooped
Darwin's account of natural selection by many years—in an appendix to his
1831 book, Naval Timber and Arboriculture. In the wake of Darwin's ascent
to fame, Matthew published a letter (in Gardeners' Chronicle?) proclaiming
his priority, which Darwin graciously conceded, excusing his ignorance by
noting the obscurity of Matthew's choice of venue. Responding to Darwin's
published apology, Matthew wrote:
To me the conception of this law of Nature came intuitively as a self-
evident fact, almost without an effort of concentrated thought. Mr. Darwin
here seems to have more merit in the discovery than I have had—to me it
did not appear a discovery. He seems to have worked it out by inductive
reason, slowly and with due caution to have made his way synthetically
from fact to fact onwards; while with me it was by a general glance at the
scheme of Nature that I estimated this select production of species as an a
priori recognizable fact—an axiom, requiring only to be pointed out to be
admitted by unprejudiced minds of sufficient grasp. [Quoted in Gould
1985, pp. 345-46.]
Unprejudiced minds may well resist a new idea out of sound conservatism,
however. Deductive arguments are notoriously treacherous; what seems to
stand to reason" can be betrayed by an overlooked detail. Darwin appre-
ciated that only a relentlessly detailed survey of the evidence for the his-
torical processes he was postulating would—or should—persuade scientists
to abandon their traditional convictions and take on his revolutionary vision,
even if it was in fact "deducible from first principles."

Gardeners' Chronicle, April 7, I860. See Hardin 1964 for more details.

50 AN IDEA IS BORN Natural Selection as an Algorithmic Process 51

From the outset, there were those who viewed Darwin's novel mixture of
detailed naturalism and abstract reasoning about processes as a dubious and
inviable hybrid. It had a tremendous air of plausibility, but so do many get-
rich-quick schemes that turn out to be empty tricks. Compare it to the
following stock-market principle. Buy Low, Sell High. This is guaranteed to
make you wealthy. You cannot fail to get rich if you follow this advice. Why
doesn't it work? It does work—for everybody who is fortunate enough to act
according to it, but, alas, there is no way of determining that the conditions
are met until it is too late to act on them. Darwin was offering a skeptical
world what we might call a get-rich-slow scheme, a scheme for creating
Design out of Chaos without the aid of Mind.
The theoretical power of Darwin's abstract scheme was due to several
features that Darwin quite firmly identified, and appreciated better than many
of his supporters, but lacked the terminology to describe explicitly. Today we
could capture these features under a single term. Darwin had discovered the
power of an algorithm. An algorithm is a certain sort of formal process that
can be counted on—logically—to yield a certain sort of result whenever it is
"run" or instantiated. Algorithms are not new, and were not new in Darwin's
day. Many familiar arithmetic procedures, such as long division or balancing
your checkbook, are algorithms, and so are the decision procedures for
playing perfect tic-tac-toe, and for putting a list of words into alphabetical
order. What is relatively new—permitting us valuable hindsight on Darwin's
discovery—is the theoretical reflection by mathematicians and logicians on
the nature and power of algorithms in general, a twentieth-century
development which led to the birth of the computer, which has led in turn, of
course, to a much deeper and more lively understanding of the powers of
algorithms in general.
The term algorithm descends, via Latin (algorismus) to early English
(algorisme and, mistakenly therefrom, algorithm), from the name of a
Persian mathematician, Muusa al-Khowarizm, whose book on arithmetical
procedures, written about 835 A.D., was translated into Latin in the twelfth
century by Adelard of Bath or Robert of Chester. The idea that an algorithm
is a foolproof and somehow "mechanical" procedure has been present for
centuries, but it was the pioneering work of Alan Turing, Kurt Godel, and
Alonzo Church in the 1930s that more or less fixed our current understanding
of the term. Three key features of algorithms will be important to us, and
each is somewhat difficult to define. Each, moreover, has given rise to
confusions (and anxieties ) that continue to beset our thinking about Darwin's
revolutionary discovery, so we will have to revisit and reconsider these
introductory characterizations several times before we are through:
(1) substrate neutrality: The procedure for long division works equally
well with pencil or pen, paper or parchment, neon lights or skywrit-
ing, using any symbol system you like. The power of the procedure is
due to its logical structure, not the causal powers of the materials used
in the instantiation, just so long as those causal powers permit the
prescribed steps to be followed exactly.
(2) underlying mindlessness: Although the overall design of the proce-
dure may be brilliant, or yield brilliant results, each constituent step,
as well as the transition between steps, is utterly simple. How simple?
Simple enough for a dutiful idiot to perform—or for a straightforward
mechanical device to perform. The standard textbook analogy notes
that algorithms are recipes of sorts, designed to be followed by novice
cooks. A recipe book written for great chefs might include the phrase
"Poach the fish in a suitable wine until almost done," but an algorithm
for the same process might begin, "Choose a white wine that says 'dry'
on the label; take a corkscrew and open the bottle; pour an inch of
wine in the bottom of a pan; turn the burner under the pan on high; ...
"—a tedious breakdown of the process into dead-simple steps,
requiring no wise decisions or delicate judgments or intuitions on the
part of the recipe-reader.
(3) guaranteed results: Whatever it is that an algorithm does, it always
does it, if it is executed without misstep. An algorithm is a foolproof
recipe.
It is easy to see how these features made the computer possible. Every
computer program is an algorithm, ultimately composed of simple steps that
can be executed with stupendous reliability by one simple mechanism or
another. Electronic circuits are the usual choice, but the power of computers
owes nothing (save speed) to the causal peculiarities of electrons darting
about on silicon chips. The very same algorithms can be performed (even
faster) by devices shunting photons in glass fibers, or (much, much slower)
by teams of people using paper and pencil. And as we shall see, the capacity
of computers to run algorithms with tremendous speed and reliability is now
permitting theoreticians to explore Darwin's dangerous idea in ways
heretofore impossible, with fascinating results.
What Darwin discovered was not really one algorithm but, rather, a large
class of related algorithms that he had no clear way to distinguish. We can
now reformulate his fundamental idea as follows:
Life on Earth has been generated over billions of years in a single branching
tree—the Tree of Life—by o'ne algorithmic process or another.
What this claim means will become clear gradually, as we sort through he
various ways people have tried to express it. In some versions it is utterly
vacuous and uninformative; in others it is manifestly false. In be-

52 AN IDEA IS BORN Processes as Algorithms 53

tween lie the versions that really do explain the origin of species and promise
to explain much else besides. These versions are becoming clearer all the
time, thanks as much to the determined criticisms of those who frankly hate
the idea of evolution as an algorithm, as to the rebuttals of those who love it.
5. PROCESSES AS ALGORITHMS
When theorists think of algorithms, they often have in mind kinds of algo-
rithms with properties that are not shared by the algorithms that will concern
us. When mathematicians think about algorithms, for instance, they usually
have in mind algorithms that can be proven to compute particular
mathematical functions of interest to them. (Long division is a homely
example. A procedure for breaking down a huge number into its prime
factors is one that attracts attention in the exotic world of cryptography.) But
the algorithms that will concern us have nothing particular to do with the
number system or other mathematical objects; they are algorithms for
sorting, winnowing, and building things.8
Because most mathematical discussions of algorithms focus on their guar-
anteed or mathematically provable powers, people sometimes make the
elementary mistake of thinking that a process that makes use of chance or
randomness is not an algorithm. But even long division makes good use of
randomness!
7? 47)
326574
Does the divisor go into the dividend six or seven or eight times? Who
knows? Who cares? You don't have to know; you don't have to have any wit
or discernment to do long division. The algorithm directs you just to choose a
digit—at random, if you like—and check out the result. If die chosen number
turns out to be too small, increase it by one and start over; if too large,
decrease it. The good thing about long division is that it always works

8. Computer scientists sometimes restrict the term algorithm to programs that can be
proven to terminate—that have no infinite loops in them, for instance. But this special
sense, valuable as it is for some mathematical purposes, is not of much use to us. Indeed,
few of the computer programs in daily use around the world would qualify as algorithms
in this restricted sense; most are designed to cycle indefinitely, patiently waiting for
instructions (including the instruction to terminate, without which they keep on going).
Their subroutines, however, are algorithms in this strict sense—except where undetec-
ted "bugs" lurk that can cause the program to "hang."
eventually, even if you are maximally stupid in making your first choice, in
which case it just takes a little longer. Achieving success on hard tasks in
spite of utter stupidity is what makes computers seem magical—how could
something as mindless as a machine do something as smart as that? Not
surprisingly, then, the tactic of finessing ignorance by randomly generating a
candidate and then testing it out mechanically is a ubiquitous feature of
interesting algorithms. Not only does it not interfere with their provable
powers as algorithms; it is often the key to their power. (See Dennett 1984,
pp 149-52, on the particularly interesting powers of Michael Rabin's random
algorithms.)
We can begin zeroing in on the phylum of evolutionary algorithms by con-
sidering everyday algorithms that share important properties with them. Dar-
win draws our attention to repeated waves of competition and selection, so
consider the standard algorithm for organizing an elimination tournament,
such as a tennis tournament, which eventually culminates with quarter-finals,
semi-finals, and then a final, determining the solitary winner.

Notice that this procedure meets the three conditions. It is the same
procedure whether drawn in chalk on a blackboard, or updated in a computer
file, or—a weird possibility—not written down anywhere, but simply
enforced by building a huge fan of fenced-off tennis courts each with two
entrance gates and a single exit gate leading the winner to the court where
the next match is to be played. (The losers are shot and buried where they
fall) It doesn't take a genius to march the contestants through the drill, filling
in the blanks at the end of each match ( or identifying and shooting the
losers). And it always works.
But what, exactly, does this algorithm do? It takes as input a set of com-
petitors and guarantees to terminate by identifying a single winner. But what
is a winner? It all depends on the competition. Suppose the tournament in
question is not tennis but coin-tossing. One player tosses and the other calls;
the winner advances. The winner of this tournament will be that single player
who has won n consecutive coin-tosses without a loss, depending on how
many rounds it takes to complete the tournament.

54 AN IDEA IS BORN Processes as Algorithms 55

There is something strange and trivial about this tournament, but what is
it? The winner does have a rather remarkable property. How often have you
ever met anyone who just won, say, ten consecutive coin-tosses without a
loss? Probably never. The odds against there being such a person might seem
enormous, and in the normal course of events, they surely are. If some
gambler offered you ten-to-one odds that he could produce someone who
before your very eyes would proceed to win ten consecutive coin-tosses
using a fair coin, you might be inclined to think this a good bet. If so, you
had better hope the gambler doesn't have 1,024 accomplices (they don't have
to cheat—they play fair and square). For that is all it takes (210 competitors)
to form a ten-round tournament. The gambler wouldn't have a clue, as the
tournament started, which person would end up being the exhibit A that
would guarantee his winning the wager, but the tournament algorithm is sure
to produce such a person in short order—it is a sucker bet with a surefire win
for the gambler. (I am not responsible for any injuries you may sustain if you
attempt to get rich by putting this tidbit of practical philosophy into use.)
Any elimination tournament produces a winner, who "automatically" has
whatever property was required to advance through the rounds, but, as the
coin-tossing tournament demonstrates, the property in question may be
"merely historical"—a trivial fact about the competitor's past history that has
no bearing at all on his or her future prospects. Suppose, for instance, the
United Nations were to decide that all future international conflicts would be
settled by a coin-toss to which each nation sends a representative (if more
than one nation is involved, it will have to be some sort of tournament—it
might be a "round robin," which is a different algorithm ). Whom should we
designate as our national representative? The best coin-toss caller in the land,
obviously. Suppose we organized every man, woman, and child in the U.S.A.
into a giant elimination tournament. Somebody would have to win, and that
person would have just won twenty-eight consecutive coin-tosses without a
loss! This would be an irrefutable historical fact about that person, but since
calling a coin-toss is just a matter of luck, there is absolutely no reason to
believe that the winner of such a tournament would do any better in
international competition than somebody else who lost in an earlier round of
the tournament. Chance has no memory. A person who holds the winning
lottery ticket has certainly been lucky, and, thanks to the millions she has just
won, she may never need to be lucky again—which is just as well, since there
is no reason to think she is more likely than anyone else to win the lottery a
second time, or to win the next coin-toss she calls. ( Failing to appreciate the
fact that chance has no memory is known as the Gambler's Fallacy; it is
surprisingly popular—so popular that I should probably stress that it is a
fallacy, beyond any doubt or controversy.)
In contrast to tournaments of pure luck, like the coin-toss tournament,
there are tournaments of skill, like tennis tournaments. Here there is reason to
believe that the players in the later rounds would do better again if they
played the players who lost in the early rounds. There is reason to believe—
but no guarantee—that the winner of such a tournament is the best player of
them all, not just today but tomorrow. Yet, though any well-run tournament
is guaranteed to produce a winner, there is no guarantee that a tournament of
skill will identify the best player as the winner in any nontrivial sense. That's
why we sometimes say, in the opening ceremonies, "May the best man
win!"—because it is not guaranteed by the procedure. The best player—the
one who is best by "engineering" standards (has the most reliable backhand,
fastest serve, most stamina, etc.)—may have an off day, or sprain his ankle, or
get hit by lightning. Then, trivially, he may be bested in competition by a
player who is not really as good as he is. But nobody would bother
organizing or entering tournaments of skill if it weren't the case that in the
long run, tournaments of skill are won by the best players. That is guaranteed
by the very definition of a fair tournament of skill; if there were no probability
greater than half that the better players would win each round, it would be a
tournament of luck, not of skill.
Skill and luck intermingle naturally and inevitably in any real competition,
but their ratios may vary widely. A tennis tournament played on very bumpy
courts would raise the luck ratio, as would an innovation in which the players
were required to play Russian roulette with a loaded revolver before
continuing after the first set. But even in such a luck-ridden contest, more of
the better players would tend, statistically, to get to the late rounds. The
power of a tournament to "discriminate" skill differences in the long run may
be diminished by haphazard catastrophe, but it is not in general reduced to
zero. This fact, which is as true of evolutionary algorithms in nature as of
elimination tournaments in sports, is sometimes overlooked by
commentators on evolution.
Skill, in contrast to luck, is protectable; in the same or similar circum-
stances, it can be counted on to give repeat performances. This relativity to
circumstances shows us another way in which a tournament might be weird.
What if the conditions of competition kept changing (like the croquet game
in Alice in Wonderland)? If you play tennis the first round, chess in the
second round, golf in the third round, and billiards in the fourth round, there
is no reason to suppose the eventual winner will be particularly good,
compared with the whole field, in any of these endeavors—all the good
golfers may lose in the chess round and never get a chance to demonstrate
their prowess, and even if luck plays no role in the fourth-round billiards
final, the winner might turn out to be the second-worst billiards player in the
whole field. Thus there has to be some measure of uniformity of the
conditions of competition for there to be any interesting outcome to a
tournament.

76 UNIVERSAL ACID
The Tools for R and D: Skyhooks or Cranes? 77

skyhooks. But then along have come the cranes, discovered in many cases by
the very skeptics who were hoping to find a skyhook.
It is time for some more careful definitions. Let us understand that a
skyhook is a "mind-first" force or power or process, an exception to the
principle that all design, and apparent design, is ultimately the result of
mindless, motiveless mechanicity. A crane, in contrast, is a subprocess or
special feature of a design process that can be demonstrated to permit the
local speeding up of the basic, slow process of natural selection, and that can
be demonstrated to be itself the predictable (or retrospectively explicable )
product of the basic process. Some cranes are obvious and uncon-troversial;
others are still being argued about, very fruitfully. Just to give a general sense
of the breadth and application of the concept, let me point to three very
different examples.
It is now generally agreed among evolutionary theorists that sex is a crane.
That is, species that reproduce sexually can move through Design Space at a
much greater speed than that achieved by organisms that reproduce asexually.
Moreover, they can "discern" design improvements along the way that are all
but "invisible" to asexually reproducing organisms ( Holland 1975 ). This
cannot be the raison d'etre of sex, however. Evolution cannot see way down
the road, so anything it builds must have an immediate payoff to
counterbalance the cost. As recent theorists have insisted, the "choice" of
reproducing sexually carries a huge immediate cost: organisms send along
only 50 percent of their genes in any one transaction (to say nothing of the
effort and risk involved in securing a transaction in the first place). So the
long-term payoff of heightened efficiency, acuity, and speed of the redesign
process—the features that make sex a magnificent crane—is as nothing to the
myopic, local competitions that must determine which organisms get favored
in the very next generation. Some other, short-term, benefit must have
maintained the positive selection pressure required to make sexual
reproduction an offer few species could refuse. There are a variety of
compelling—and competing—hypotheses that might solve this puzzle, which
was first forcefully posed for biologists by John Maynard Smith ( 1978). For
a lucid introduction to the current state of play, see Matt Ridley 1993- (More
on this later.)
What we learn from the example of sex is that a crane of great power may
exist that was not created in order to exploit that power, but for other reasons,
although its power as a crane may help explain why it has been maintained
ever since. A crane that was obviously created to be a crane is genetic
engineering. Genetic engineers—human beings who engage in recombinant-
DNA tinkering—can now unquestionably take huge leaps through Design
Space, creating organisms that would never have evolved by "ordinary"
means. This is no miracle—provided that genetic engineers (and the artifacts
they use in their trade) are themselves wholly the products of
earlier, slower evolutionary processes. If the creationists were right that
mankind is a species unto itself, divine and inaccessible via brute Darwinian
paths, then genetic engineering would not be a crane after all, having been
created with the help of a major skyhook. I don't imagine that any genetic
engineers think of themselves this way, but it is a logically available perch,
however precarious. Less obviously silly is this idea: if the bodies of genetic
engineers are products of evolution, but their minds can do creative things
that are irreducibly nonalgorithmic or inaccessible by all algorithmic paths,
then the leaps of genetic engineering might involve a skyhook. Exploring
this prospect will be the central topic of chapter 15.
A crane with a particularly interesting history is theBaldwin-Effect, named
for one of its discoverers, James Mark Baldwin (1896), but more or less
simultaneously discovered by two other early Darwinians, Conwy Lloyd
Morgan (famed for Lloyd Morgan's Canon of Parsimony [for discussion, see
Dennett 1983]) and H. F. Osborn. Baldwin was an enthusiastic Darwinian,
but he was oppressed by the prospect that Darwin's theory would leave Mind
with an insufficiently important and originating role in the (redesign of
organisms. So he set out to demonstrate that animals, by dint of their own
clever activities in the world, might hasten or guide the further evolution of
their species. Here is what he asked himself: how could it be that individual
animals, by solving problems in their own lifetimes, could change the
conditions of competition for their own offspring, making those problems
easier to solve in the future? And he came to realize that this was in fact
possible, under certain conditions, which we can illustrate with a simple
example (drawn, with revisions, from Dennett 1991a).
Consider a population of a species in which there is considerable variation
at birth in the way their brains are wired up. Just one of the ways, we may
suppose, endows its possessor with a Good Trick—a behavioral talent that
protects it or enhances its chances dramatically. The standard way of
representing such differences in fitness between individual members of a
population is known as an "adaptive landscape" or a "fitness landscape" (S.
Wright 1931). The altitude in such a diagram stands for fitness (higher is
better), and the longitude and latitude stand for some factors of individual
design—in this case, features of brain-wiring. Each different way a brain
might be wired is represented by one of the rods that compose the land-
scape—each rod is a different genotype. The fact that just one of the com-
binations of features is any good—that is, any better than run-of-the-mill—is
illustrated by the way it stands out like a telephone pole in the desert.
As figure 3.1 makes clear, only one wiring is favored; the others, no matter
how "close" to being the good wiring, are about equal in fitness. So such an
isolated peak is indeed a needle in the haystack: it will be practically invis-
ible to natural selection. Those few individuals in the population that are
lucky enough to have the Good Trick genotype will typically have difficulty

80 UNIVERSAL ACID Who's Afraid of Reductionism? 81

Although it has been regularly described and acknowledged in biology
textbooks, it has typically been shunned by overcautious thinkers, because
they thought it smacked of the Lamarckian heresy (the presumed possibility
of inheritance of acquired characteristics—see chapter 11 for a detailed
discussion). This rejection is particularly ironic, since, as Richards notes, it
was intended by Baldwin to be—and truly is—an acceptable substitute for
Lamarckian mechanisms.
The principle certainly seemed to dispatch Lamarckism, while supplying
that positive factor in evolution for which even staunch Darwinists like
Lloyd Morgan longed. And to those of metaphysical appetite, it revealed
that under the clanking, mechanical vesture of Darwinian nature, mind
could be found. [R. J. Richards 1987, p. 487]
Well, not Mind—if by that we mean a full-fledged, intrinsic, original,
skyhook-type Mind—but only a nifty mechanistic, behavioristic, crane-style
mind. That is not nothing, however; Baldwin discovered an effect that gen-
uinely increases the power—locally—of the underlying process of natural
selection wherever it operates. It shows how the "blind" process of the basic
phenomenon of natural selection can be abetted by a limited amount of
"look-ahead" in the activities of individual organisms, which create fitness
differences that natural selection can then act upon. This is a welcome
complication, a wrinkle in evolutionary theory that removes one reasonable
and compelling source of doubt, and enhances our vision of the power of
Darwin's idea, especially when it is cascaded in multiple, nested applications.
And it is typical of the outcome of other searches and controversies we will
explore: the motivation, the passion that drove the research, was the hope of
finding a skyhook; the triumph was finding how the same work could be
done with a crane.
5. WHO'S AFRAID OF REDUCTIONISM?
Reductionism is a dirty word, and a kind of 'holistier than thou' self-
righteousness has become fashionable.
—RICHARD DAWKINS 1982, p. 113
The term that is most often bandied about in these conflicts, typically as a
term of abuse, is "reductionism." Those who yearn for skyhooks call those
who eagerly settle for cranes "reductionists," and they can often make
reductionism seem philistine and heartless, if not downright evil. But like
most terms of abuse, "reductionism" has no fixed meaning. The central
image is of somebody claiming that one science "reduces" to another: that
chemistry reduces to physics, that biology reduces to chemistry, that the
social sciences reduce to biology, for instance. The problem is that there are
both bland readings and preposterous readings of any such claim. According
to the bland readings, it is possible (and desirable ) to unify chemistry and
physics, biology and chemistry, and, yes, even the social sciences and biol-
ogy. After all, societies are composed of human beings, who, as mammals,
must fall under the principles of biology that cover all mammals. Mammals,
in turn, are composed of molecules, which must obey the laws of chemistry,
which in turn must answer to the regularities of the underlying physics. No
sane scientist disputes this bland reading; the assembled Justices of the
Supreme Court are as bound by the law of gravity as is any avalanche,
because they are, in the end, also a collection of physical objects. According
to the preposterous readings, reductionists want to abandon the principles,
theories, vocabulary, laws of the higher-level sciences, in favor of the lower-
level terms. A reductionist dream, on such a preposterous reading, might be
to write "A Comparison of Keats and Shelley from the Molecular Point of
View" or "The Role of Oxygen Atoms in Supply-Side Economics," or "Ex-
plaining the Decisions of the Rehnquist Court in Terms of Entropy Fluctu-
ations." Probably nobody is a reductionist in the preposterous sense, and
everybody should be a reductionist in the bland sense, so the "charge" of
reductionism is too vague to merit a response. If somebody says to you, "But
that's so reductionistic!" you would do well to respond, "That's such a quaint,
old-fashioned complaint! What on Earth did you have in mind?"
I am happy to say that in recent years, some of the thinkers I most admire
have come out in defense of one or another version of reductionism, care-
fully circumscribed. The cognitive scientist Douglas Hofstadter, in Godel
Escher Bach, composed a "Prelude ... Ant Fugue" (Hofstadter 1979, pp. 275-
336) that is an analytical hymn to the virtues of reductionism in its proper
place. George C. Williams, one of the pre-eminent evolutionists of the day,
published "A Defense of Reductionism in Evolutionary Biology" (1985).
The zoologist Richard Dawkins has distinguished what he calls hierarchical
or gradual reductionism from precipice reductionism; he rejects only the
precipice version (Dawkins 1986b, p. 74 ).8 More recently the physicist
Steven Weinberg, in Dreams of a Final Theory (1992), has written a chapter
entitled "Two Cheers for Reductionism," in which he distinguishes between
uncompromising reductionism (a bad thing) and compromising reductionism
(which he ringingly endorses). Here is my own version. We must distinguish
reductionism, which is in general a good

• See also his discussion of Lewontin, Rose, and Kamin's (1984 ) idiosyncratic version of
reductionism—Dawkins aptly calls it their "private bogey"—in the second edition of The
Se!ftsh Gene (I989z\ p. 331.

82 UNIVERSAL ACID Who's Afraid of Reductionism? 83

thing, from greedy reductionism, which is not. The difference, in the context
of Darwin's theory, is simple: greedy reductionists think that everything can
be explained without cranes; good reductionists think that everything can be
explained without skyhooks.
There is no reason to be compromising about what I call good reduc-
tionism. It is simply the commitment to non-question-begging science with-
out any cheating by embracing mysteries or miracles at the outset. (For
another perspective on this, see Dennett 1991a, pp. 33-39.) Three cheers for
that brand of reductionism—and I'm sure Weinberg would agree. But in their
eagerness for a bargain, in their zeal to explain too much too fast, scientists
and philosophers often underestimate the complexities, trying to skip whole
layers or levels of theory in their rush to fasten everything securely and
neatly to the foundation. That is the sin of greedy reductionism, but notice
that it is only when overzealousness leads to falsification of the phenomena
that we should condemn it. In itself, the desire to reduce, to unite, to explain
it all in one big overarching theory, is no more to be condemned as immoral
than the contrary urge that drove Baldwin to his discovery. It is not wrong to
yearn for simple theories, or to yearn for phenomena that no simple (or
complex!) theory could ever explain; what is wrong is zealous
misrepresentation, in either direction.
Darwin's dangerous idea is reductionism incarnate,9 promising to unite and
explain just about everything in one magnificent vision. Its being the idea of
an algorithmic process makes it all the more powerful, since the substrate
neutrality it thereby possesses permits us to consider its application to just
about anything. It is no respecter of material boundaries. It applies, as we
have already begun to see, even to itself. The most common fear about
Darwin's idea is that it will not just explain but explain away the Minds and
Purposes and Meanings that we all hold dear. People fear that once this
universal acid has passed through the monuments we cherish, they will cease
to exist, dissolved in an unrecognizable and unlovable puddle of scientistic
destruction. This cannot be a sound fear; a proper reductionists explanation
of these phenomena would leave them still standing but just demystified,
unified, placed on more secure foundations. We might learn some surprising
or even shocking things about these treasures, but unless our valuing these
things was based all along on confusion or mistaken identity, how could
increased understanding of them diminish their value in
-.10
our eyes?

9. Yes, incarnate. Think about it: would we want to say it was reductionism in spirit?
10. Everybody knows how to answer this rhetorical question with another: "Are you so
in love with Truth at all costs that you would want to know if your lover were unfaithful
to you?" We are back where we started. I for one answer that I love the world so much
that I am sure I want to know the truth about it.
A more reasonable and realistic fear is that the greedy abuse of Darwinian
reasoning might lead us to deny the existence of real levels, real complex-
ities, real phenomena. By our own misguided efforts, we might indeed come
to discard or destroy something valuable. We must work hard to keep these
two fears separate, and we can begin by acknowledging the pressures that
tend to distort the very description of the issues. For instance, there is a
strong tendency among many who are uncomfortable with evolutionary
theory to exaggerate the amount of disagreement among scientists ("It's just a
theory, and there are many reputable scientists who don't accept this"), and I
must try hard not to overstate the compensating case for what "science has
shown." Along the way, we will encounter plenty of examples of genuine
ongoing scientific disagreement, and unsettled questions of fact. There is no
reason for me to conceal or downplay these quandaries, for no matter how
they come out, a certain amount of corrosive work has already been done by
Darwin's dangerous idea, and can never be undone.
We should be able to agree about one result already. Even if Darwin's
relatively modest idea about the origin of species came to be rejected by
science—yes, utterly discredited and replaced by some vastly more powerful
(and currently unimaginable) vision—it would still have irremediably sapped
conviction in any reflective defender of the tradition expressed by Locke. It
has done this by opening up new possibilities of imagination, and thus utterly
destroying any illusions anyone might have had about the soundness of an
argument such as Locke's a priori proof of the inconceivability of Design
without Mind. Before Darwin, this was inconceivable in the pejorative sense
that no one knew how to take the hypothesis seriously. Proving it is another
matter, but the evidence does in fact mount, and we certainly can and must
take it seriously. So whatever else you may think of Locke's argument, it is
now as obsolete as the quill pen with which it was written, a fascinating
museum piece, a curiosity that can do no real work in the intellectual world
today.
CHAPTER 3: Darwin's dangerous idea is that Design can emerge from mere
Order via an algorithmic process that makes no use of pre-existing Mind.
Skeptics have hoped to show that at least somewhere in this process, a
helping hand (more accurately, a helping Mind) must have been provided—a
skyhook to do some of the lifting. In their attempts to prove a role for
skyhooks, they have often discovered cranes: products of earlier algorithmic
processes that can amplify the power of the basic Darwinian algorithm,
making the process locally swifter and more efficient in a nonmiraculous
way. Good reductionists suppose that all Design can be explained without
skyhooks; greedy reductionists suppose it can all be explained without
cranes.

88 THE TREE OF LIFE How Should We Visualize the Tree of Life? 89

gradually lose Design? Is there a possible world in which bacteria are the?
descendants of mammals and not vice versa? These questions about possi-
bility will be easier to answer if we first look a bit more closely at what has
actually happened on our planet. So let us be clear that for the time being, the
vertical dimension in the diagrams below stands for time, and time alone,
with early at the bottom and late at the top. Following standard practice, the
left-right dimension is taken as a sort of single-plane summary of diversity.
Each individual organism has to have its time line, distinct from all others,
so, even if two organisms are exact atom-for-atom duplicates of each other,
they will have to appear side by side at best. How we line them all up,
however, can be according to some measure or family of measures of
difference in individual body shape—morphology, to use the technical term.
So, to return to our question, what would the overall shape of the entire
Tree of Life look like, if we could take it all in at a glance? Wouldn't it look
rather like a palm tree, as in figure 4.1?
This is the first of many trees, or dendrograms, we will consider, and of
course the limited resolution of the ink on the page blurs quadrillions of
separate lines together. I have left the "root" of the tree deliberately fuzzy and
indistinct for the time being. We are still exploring the middle, saving the
ultimate beginnings for a later chapter. If we were to zoom in on the trunk of
this tree and look at any cross-section of it—an "instant" in
time—we would see billions upon billions of individual unicellular organ-
isms, a fraction of which would have trails leading to progeny slightly higher
up the trunk. (In those early days, reproduction was by budding or fission;
somewhat later, a kind of unicellular sex evolved, but pollen-wafting and
egg-laying and the other phenomena of our kind of sexual reproduction have
to wait for the multicellular revolution in the fronds.) There would be some
diversity, and some revision of design over time, so perhaps the whole trunk
should be shown leaning left or right, or spreading more than I have shown.
Is it just our ignorance that prevents us from differentiating this "trunk" of
unicellular varieties into salient streams? Perhaps it should be shown with
various dead-end branches large enough to be visible, as in figure 4.2,
marking various hundred-million-year experiments in alternative unicellular
design that eventually all ended in extinction.

EARTH FORMED --------------------------------------------------------------------
FIGURE 4.3
There must have been billions of failed design experiments, but perhaps
none ever became very distant departures from a single unicellular norm. In
any event, if we were to zoom way in on the trunk, we would see a luxuriant
growth of short-lived alternatives, as in figure 4.3, all but invisible against
the norm of conservative replication. How can we be sure of this? Because,
as we shall see, the odds are heavily against any mutation's being more
viable than the theme on which it is a variation.

92 THE TREE OF LIFE Color-coding a Species on the Tree 93

that reproduce asexually, groupings of one sort or another may interest us for
various good reasons—groupings of shared morphology or behavior or of
genetic similarity, for instance—and we might choose to call the resulting
group a species, but there may very well be no theoretically important sharp
edges that would delimit such a species. So let us concentrate on sexually
reproducing species, all of which are to be found up in the multicellular
fronds of the Tree. How might we go about coloring all the life-lines of a
single such species red? We could start by looking at individuals at random
until we found one with lots of descendants. Call her Lulu, and color her red.
(Red is represented by the thick lines in figure 4.5.) Now move stepwise up
the Tree, coloring all Lulu's descendants red; these will all be members of
one species unless we find our red ink spreading into two distinct higher
branches, none of whose members form junctions across the void. If that
happens, we know there has been speciation, and we will have to back up and
make several decisions. We must first choose whether to keep one of the
branches red (the "parent" species continues red and the other branch is
considered the new daughter species ) or to stop the red ink altogether as
soon as the branching happens (the "parent" species has gone extinct,
fissioning into two daughter species).
If the organisms in the branch on the left are all pretty much the same in
appearance, equipment, and habits as Lulu's contemporaries, while the or-
ganisms in the right branch almost all sport novel horns, or webbed feet, or
stripes, then it is pretty obvious that we should label the left branch as the
continuing, parent species, and the right branch the new offshoot. If both
branches soon show major changes, our color-coding decision is not so
obvious. There are no secret facts that could tell us which choice is right,
which choice carves nature at the joints, for we are looking right at the places
where the joints would have to be, and there aren't any. There is nothing
more to being a species than being one of these branches of interbreeding
organisms, and nothing more to being the conspecific of some other organism
(contemporary or not) than being part of the same branch. The choice we
make will then have to depend on pragmatic or aesthetic considerations: Is it
ungainly to keep the same label for this branch as for its parent trunk? Would
it be misleading for one reason or another to say the branch on the right
rather than the branch on the left was the new species?2

2. The cladists (whose views will be briefly discussed later) are a school of taxonomists
that reject, for various reasons, the concept of a "parent" species' persisting. Every
speciation event, in their terms, results in a pair of daughter species and the extinction
of their common parent, no matter how closely one surviving branch resembles the
parent, compared with the other branch.

FIGURE 4.5
The same sort of quandary faces us when we try to complete the task of
color-coding the whole species by carrying our red ink down the Tree to
include all Lulu's ancestors. We will encounter no gaps or joints on this
downward path, which will take us all the way to the prokaryotes at the base
of the Tree if we persist. But if we also color sideways as we go down, filling
in the cousins, aunts, and uncles of Lulu and her ancestors, and then color up
from these sideways spreaders, we will eventually fill in a whole branch on
which Lulu resides down to the point where coloring any lower ( earlier )
nodes (for instance, at A in figure 4.6) causes "leakage" of red into neigh-
boring branches that clearly belong to other species.
If we stop there, we can be sure that only members of Lulu's species have
been colored red. It will be arguable that we have left out some that deserve
to be colored, but only arguable, for there are, again, no hidden facts, no
essences that could settle the issue. As Darwin pointed out, if it weren't for
the separations that time and the extinction of the intermediate stepping-
stones has created, although we could put the life forms into a "natural

94 THE TREE OF LIFE Colorcoding a Species on the Tree 95


FIGURE 4.6
arrangement" (of descent), we could not put them into a "natural classifi-
cation"—we need the biggish gaps between extant forms to form the
"boundaries" of any such classes.
The theoretical concept of species that predates Darwin's theory had two
fundamental ideas: that species members have different essences, and that
"therefore" they don't/can't interbreed. What we have subsequently figured
out is that in principle there could be two subpopulations that were different
only in that their pairings were sterile due to a tiny genetic incompatibility.
Would these be different species? They could look alike, feed alike, live
together in the same niche, and be genetically very, very similar, yet
reproductively isolated. They would not be different enough to count as
salient varieties, but they would satisfy the primary condition for being two
different species. In fact, there are cases of "cryptic sibling species" that
approximate this extreme. As we already noted, at the other extreme we have
the dogs, readily distinguished into morphological types by the naked eye,
adapted to vastly different environments, but not reproductively iso-
lated. Where should we draw the line? Darwin shows that we don't need to
draw the line in an essentialist way in order to get on with our science. We
have the best of reasons to realize that these extremes are improbable: in
general, where there is genetic speciation there is marked morphological
difference, or marked difference in geographical distribution, or (most likely)
both. If this generalization weren't largely true, the concept of species would
not be important, but we need not ask exactly how much difference (in
addition to reproductive isolation) is essential for a case of real species-
difference.3
Darwin shows us that questions like "What is the difference between a
variety and a species?" are like the question "What is the difference between
a peninsula and an island?"4 Suppose you see an island half a mile offshore
at high tide. If you can walk to it at low tide without getting your feet wet, is
it still an island? If you build a bridge to it, does it cease to be an island?
What if you build a solid causeway? If you cut a canal across a peninsula
(like the Cape Cod Canal), do you turn it into an island? What if a hurricane
does the excavation work? This sort of inquiry is familiar to philosophers. It
is the Socratic activity of definition-mongering or essence-hunting: looking
for the "necessary and sufficient conditions" for being-an-X. Sometimes al-
most everyone can see the pointlessness of the quest—islands obviously
don't have real essences, but only nominal essences at best. But at other
times there can still seem to be a serious scientific question that needs
answering.
More than a century after Darwin, there are still serious debates among
biologists (and even more so among philosophers of biology ) about how to
define species. Shouldn't scientists define their terms? Yes, of course, but
only up to a point. It turns out that there are different species concepts with
different uses in biology—what works for paleontologists is not much use to
ecologists, for instance—and no clean way of uniting them or putting them
in an order of importance that would crown one of them (the most important
one) as the concept of species. So I am inclined to interpret the persisting
debates as more a matter of vestigial Aristotelian tidiness than a useful
disciplinary trait. (This is all controversial, but see Kitcher 1984 and G. C.
Williams 1992 for further support and concurring arguments, and the recent
anthology on the topic, Ereshefsky 1992, and Sterelny 1994, an insightful
review essay on that anthology.)

3. The issues are further complicated by the existence of hybridization—in which mem-
bers of two different species do have fertile offspring—a phenomenon that raises inter-
esting issues that are off the track we are exploring.
4. The evolutionary epistemologist and psychologist Donald Campbell has been the most
vigorous developer of the implications of this side of Darwin's legacy.

96 THE TREE OF LIFE Retrospective Coronations 97

3. RETROSPECTIVE CORONATIONS: MITOCHONDRIAL EVE AND
INVISIBLE BEGINNINGS
When we tried to see whether Lulu's descendants split into more than one
species, we had to look ahead to see if any large branches appeared, and then
back up if we deemed that somewhere along the line a speciation event must
have happened. We never addressed the presumably important question of
exactly when speciation should be said to occur. Speciation can now be seen
to be a phenomenon in nature that has a curious property: you can't tell that it
is occurring at the time it occurs! You can only tell much later that it has
occurred, retrospectively crowning an event when you discover that its
sequels have a certain property. This is not a point about our epistemic
limitations—as if we would be able to tell when speciation occurs if only we
had better microscopes, or even if we could get in a time machine and go
back in time to observe the appropriate moments. This is a point about the
objective property of being a speciation event. It is not a property that an
event has simply by virtue of its spatio-temporally local properties.
Other concepts exhibit similar curiosities. I once read about a comically
bad historical novel in which a French doctor came home to supper one
evening in 1802 and said to his wife-. "Guess what / did today! I assisted at
the birth of Victor Hugo!" What is wrong with that story? Or consider the
property of being a widow. A woman in New York City may suddenly
acquire that property by virtue of the effects that a bullet has just had on some
man's brain in Dodge City, over a thousand miles away. (In the days of the
Wild West, there was a revolver nicknamed the Widowmaker. Whether a
particular revolver lived up to its nickname on a particular occasion might be
a fact that could not be settled by any spatio-temporally local examination of
its effects.) This case gets its curious capacity to leap through space and time
from the conventional nature of the relation of marriage, in which a past
historical event, a wedding, is deemed to create a permanent relation—a
formal relation—of interest in spite of subsequent wanderings and concrete
misfortunes (the accidental loss of a ring, or the destruction of the marriage
certificate, for instance.)
The systematicity of genetic reproduction is not conventional but natural,
but that very systematicity permits us to think formally about causal chains
extending over millions of years, causal chains that would otherwise be
virtually impossible to designate or refer to or track. This permits us to
become interested in, and reason rigorously about, even more distant and
locally invisible relationships than the formal relationship of marriage. Spe-
ciation is, like marriage, a concept anchored within a tight, formally defin-
able system of thought, but, unlike marriage, it has no conventional
saliencies—weddings, rings, certificates—by which it can be observed. We
can see this feature of speciation in a better light by looking first at another
instance of retrospective crowning, the conferring of the title of Mitochon-
drial Eve.
Mitochondrial Eve is the woman who is the most recent direct ancestor, in
the female line, of every human being alive today. People have a hard time
thinking about this individual woman, so let's just review the reasoning.
Consider the set A, of all human beings alive today. Each was born of one
and only one mother, so consider next the set, B, of all the mothers of those
alive today. B is of necessity smaller than A, since no one has more than one
mother, and some mothers have more than one child. Continue with the set
C, of mothers of all those mothers in set B. It is smaller still. Continue on
with sets D and E and so forth. The sets must contract as we go back each
generation. Notice that as we move back through the years, we exclude many
women who were contemporaries of those in our set. Among these excluded
women are those who either lived and died childless or whose female
progeny did. Eventually, this set must funnel down to one— the woman who
is the closest direct female ancestor of everybody alive on earth today. She is
Mitochondrial Eve, so named (by Cann et al. 1987) because since the
mitochondria in our cells are passed through the maternal line alone, all the
mitochondria in all the cells in all the people alive today are direct
descendants of the mitochondria in her cells!
The same logical argument establishes that there is—must be—an Adam
as well: the closest direct male ancestor of everybody alive today. We could
call him F-Chromosome Adam, since all our F-chromosomes pass down
through the paternal line just the way our mitochondria pass through the
maternal line.5 Was F-Chromosome Adam the husband or lover of Mito-
chondrial Eve? Almost certainly not. There is only a tiny probability that
these two individuals were alive at the same time. (Paternity being a much
less time-and-energy-consuming business than maternity, what is logically
possible is that F-Chromosome Adam lived very recently, and was very, very
busy in the bedroom—leaving Errol Flynn in his, um, dust. He could, in
principle, be the great-grandfather of us all. This is about as unlikely as the
case in which F-Chromosome Adam and Mitochondrial Eve were a couple.)
Mitochondrial Eve has been in the news recently because the scientists
who christened her think they can analyze the patterns in the mitochondrial

5. Note one important difference between the legacies of Mitochondrial Eve and Y-
Chromosome Adam: we all, male and female, have mitochondria in our cells, but they
all come from our mothers; if you are male, you have a V-chromosome and got it from
your father, but most—virtually all, but not quite all—females have no Y-
chromosome at all.

100 THE TREE OF LIFE
which it followed that they were—as they later turn out to be—the founders
of a new species. We can imagine, if we want, an extreme (and improbable)
case in which a single mutation guarantees reproductive isolation in a single
generation, but, of course, whether or not the individual who has that
mutation counts as a species-founder or simply as a freak of nature depends
on nothing in its individual makeup or biography, but on what happens to
subsequent generations—if any—of its offspring.
Darwin was not able to present a single instance of speciation by natural
selection in Origin of Species. His strategy in that book was to develop in
detail the evidence that artificial selection by dog- and pigeon-breeders could
build up large differences by a series of gradual changes. He then pointed out
that deliberate choice by title animals' keepers was inessential; the runts of
the litter tended not to be valued, and hence tended not to reproduce as much
as their more valued siblings, so, without any conscious policy of breeding,
human animal-keepers presided unwittingly over a steady process of design
revision. He offered the nice example of the King Charles spaniel, "which
has been unconsciously modified to a large extent since the time of that
monarch" (Origin, p. 35)—as can be confirmed by a careful examination of
the dogs in various portraits of King Charles. He called such cases
"unconscious selection" by human domesticators, and he used it as a
persuasive bridge to get his readers to the hypothesis of even more
unconscious selection by the impersonal environment. But he had to admit,
when challenged, that he could provide no cases of animal-breeders'
producing a new species. Such breeding had definitely produced different
varieties, but not a single new species. Dachshund and St. Bernard were not
different species, however different in appearance. Darwin admitted as much,
but he might quite correctly have gone on to point out that it was simply too
early to tell whether he had given any examples of speciation accomplished
by artificial selection. Any lady's lapdog could at some future date be
discovered to have been the founding member of a species that split off from
Canis familiaris.
The same moral applies to the creation of new genera, families, and even
kingdoms, of course. The major branching that we would retrospectively
crown as the parting of the plants from the animals began as a segregation of
two gene pools every bit as inscrutable and unremarkable at the time as any
other temporary drifting apart of members of a single population.
4. PATTERNS, OVERSIMPLIFICATION, AND EXPLANATION
Much more interesting than the question of how to draw the species bound-
ary are all the questions about the shapes of the branches—and even more
interesting, the shapes of the empty spaces between the branches. What
Patterns, Oversimplification, and Explanation 101
trends, forces, principles—or historical events—have influenced these
shapes or made them possible? Eyes have evolved independently in dozens
of lineages, but feathers probably only once. As John Maynard Smith ob-
serves, mammals go in for horns but birds do not. "Why should the pattern
of variation be limited in this way? The short answer is that we do not know"
(Maynard Smith 1986, p. 41).
We can't rewind the tope of life and replay it to see what happens next
time, alas, so the only way to answer questions about such huge and ex-
perimentally inaccessible patterns is to leap boldly into the void with the
risky tactic of deliberate oversimplification. This tactic has a long and dis-
tinguished history in science, but it tends to provoke controversy, since
scientists have different thresholds at which they get nervous about playing
fast and loose with the recalcitrant details. Newtonian physics was over-
thrown by Einstein, but it is still a good approximation for almost all pur-
poses. No physicist objects when NASA uses Newtonian physics to calculate
the forces at liftoff and the orbital trajectory of the space shuttle, but, strictly
speaking, this is a deliberate use of a false theory in order to make calculation
feasible. In the same spirit, physiologists studying, say, mechanisms for
changing the rate of metabolism try in general to avoid the bizarre com-
plexities of subatomic quantum physics, hoping that any quantum effects will
cancel out or in other ways be beneath the threshold of their models. In
general, the tactic pays off handsomely, but one can never be sure when one
scientist's grubby complication will be elevated into another scientist's Key to
the Mystery. And it can just as well work the other way around: the Key is
often discovered by climbing out of the trenches and going for the panoramic
view.
I once got in a debate with Francis Crick about the virtues and vices of
Connectionism—the movement in cognitive science that models psycho-
logical phenomena by building up patterns in the connection-strengths
between the nodes in very unrealistic and oversimplified "neural nets" sim-
ulated on computers. "These people may be good engineers," Crick averred
(as best I recall), "but what they are doing is terrible science! These people
willfully turn their backs on what we already know about how neurons
interact, so their models are utterly useless as models of brain function." This
criticism somewhat surprised me, for Crick is famous for his own brilliant
opportunism in uncovering the structure of DNA; while others struggled up
the straight and narrow path of strict construction from the evidence, he and
Watson took a few daring and optimistic sidesteps, with gratifying results.
But in any case, I was curious to know how widely he would cast his
denunciation. Would he say the same thing about population geneticists? The
derogatory term for some of their models is "bean-bag genetics," for they
pretend that genes for this and that are like so many color-coded beads on a
string. What they call a gene (or an allele at a locus)

102 THE TREE OF LIFE
bears only a passing resemblance to the intricate machinery of the codon
sequences on DNA molecules. But thanks to these deliberate simplifications,
their models are computationally tractable, enabling them to discover and
confirm many large-scale patterns in gene flow that would otherwise be
utterly invisible. Adding complications would tend to bring their research to
a grinding halt. But is their research good science? Crick replied that he had
himself thought about the comparison, and had to say that population
genetics wasn't science either!
My tastes in science are more indulgent, as perhaps you would expect
from a philosopher, but I do have my reasons: I think the case is strong that
not only do "over"-simplified models often actually explain just what needs
explaining, but no more complicated model could do the job. When what
provokes our curiosity are the large patterns in phenomena, we need an
explanation at the right level. In many instances this is obvious. If you want
to know why traffic jams tend to happen at a certain hour every day, you will
still be baffled after you have painstakingly reconstructed the steering, brak-
ing, and accelerating processes of the thousands of drivers whose various
trajectories have summed to create those traffic jams.
Or imagine tracing all the electrons through a hand calculator as it mul-
tiplies two numbers together and gets the correct answer. You could be 100
percent sure you understood each of the millions of causal microsteps in the
process and yet still be utterly baffled about why or even how it always got
the right answer to the questions you posed it. If this is not obvious, imagine
that somebody made—as a sort of expensive prank—a hand calculator that
usually gave the wrong answers! It would obey exactly the same physical
laws as the good calculator, and would cycle through the same sorts of
microprocesses. You could have perfect explanations of how both calculators
worked at the electronic level, and still be utterly unable to explain the
intensely interesting fact that one of them got the answers right and the other
got them wrong. This is the sort of case that shows what would be silly about
the preposterous forms of reductionism; of course you can't explain all the
patterns that interest us at the level of physics (or chemistry, or any one low
level). This is undeniably true of such mundane and unperplexing
phenomena as traffic jams and pocket calculators; we should expect it to be
true of biological phenomena as well. (For more on this topic, see Dennett
1991b.)
Now consider a parallel question in biology, a textbook standard: why do
giraffes have long necks? There is one answer that could in principle be
"read off" the total Tree of Life, if we had it to look at: Each giraffe has a neck
of the length it has because its parents had necks of the lengths they had, and
so forth back through the generations. If you check them off one by one, you
will see that the long neck of each living giraffe has been traced back
through long-necked ancestors all the way back... to ancestors who didn't
Patterns, Oversimplification, and Explanation 103
even have necks. So that's how come giraffes have long necks. End of ex-
planation. (And if that doesn't satisfy you, note that you will be even less
satisfied if the answer throws in all the details about the individual devel-
opmental and nutritional history of each giraffe in the lineage.)
Any acceptable explanation of the patterns we observe in the Tree of Life
must be contrastive: why do we see this actual pattern rattier than that one—
or no pattern at all? What are the nonactualized alternatives that need to be
considered, and how are they organized? To answer such questions, we need
to be able to talk about what is possible in addition to what is actual.
CHAPTER 4: There are patterns in the unimaginably detailed Tree of Life,
highlighting crucial events that made the later flourishing of the Tree pos-
sible. The eukaryotic revolution and the multicellular revolution are the most
important, followed by the speciation events, invisible at the time, but later
seen to mark even such major divisions as those between plants and animals.
If science is to explain the patterns discernible in all this complexity, it must
rise above the microscopic view to other levels, taking on idealizations when
necessary so we can see the woods for die trees.
CHAPTER 5: The contrast between the actual and the possible is fundamental
to all explanation in biology. It seems we need to distinguish different grades
of possibility, and Darwin provides a framework for a unified treatment of
biological possibility in terms of accessibility in "the Library of Mendel," the
space of all genomes. In order to construct this useful idealization, we must
acknowledge and then set aside certain complications in the relations
between a genome and a viable organism.

Grades of Possibility? 105
CHAPTER FIVE
The Possible and the
Actual

1. GRADES OF POSSIBILITY?
However many ways there may be of being alive, it is certain that there
are vastly more ways of being dead, or rather not alive.
—RICHARD DAWKINS 1986A, P. 9
Any particular non-existent form of life may owe its absence to one of
two reasons. One is negative selection. The other is that the necessary
mutations have never appeared.
—MARK RIDLEY 1985, P. 56
Take, for instance, the possible fat man in that doorway; and, again, the
possible bald man in diat doorway. Are they the same possible man, or
two possible men? How do we decide? How many possible men are
there in mat doorway? Are there more possible thin ones than fat ones?
How many of them are alike? Or would their being alike make them
one? Are no two possible things alike? Is this the same as saying that it
is impossible for two things to be alike? Or, finally, is the concept of
identity simply inapplicable to unactualized possibles?
—WILLARD VAN ORMAN QLINE 1953, P. 4
There seem to be at least four different kinds or grades of possibility:
logical, physical, biological, and historical, nested in that order. The most
lenient is mere logical possibility, which according to philosophical tradition
is simply a matter of being describable without contradiction. Super-
man, who flies faster than the speed of light, is logically possible, but
Duperman, who flies faster than the speed of light without moving anywhere,
is not even logically possible. Superman, however, is not physically possible,
since a law of physics proclaims that nothing can move faster than the speed
of light. There is no dearth of difficulties with this superficially
straightforward distinction. How do we distinguish fundamental physical
laws from logical laws? Is it physically or logically impossible to travel
backwards in time, for instance? How could we tell for sure whether a
description that is apparently coherent—such as the story in the film Back to
the Future—is subtly self-contradictory or merely denies a very fundamental
(but not logically necessary ) assumption of physics? There is also no dearth
of philosophy dealing with these difficulties, so we will just acknowledge
them and pass on to the next grade.
Superman flies by simply leaping into the air and striking a gallant midair
pose, a talent which is certainly physically impossible. Is a flying horse
physically possible? The standard model from mythology would never get
off the ground—a fact from physics (aerodynamics), not biology—but a
horse with suitable wingspan could presumably stay aloft. It might have to be
a tiny horse, something aeronautical engineers might calculate from
considerations of weight-strength ratios, the density of air, and so forth. But
now we are descending into the third grade of possibility, biological pos-
sibility, for once we begin considering the strength of bones, and the pay-
load requirements for keeping the flapping machinery going, we concern
ourselves with development and growth, metabolism, and other clearly
biological phenomena. Still, the verdict may appear to be that of course
flying horses are biologically possible, since bats are actual. Maybe even
full-sized flying horses are possible, since there once were pteranodons and
other flying creatures approaching that size. There is nothing to beat actu-
ality, present or past, for clinching possibility. Whatever is or has been actual
is obviously possible. Or is it?
The lessons of actuality are hard to read. Could such flying horses really
be viable? Would they perhaps need to be carnivorous to store enough
energy and carry it aloft? Perhaps—in spite of fruit-eating bats—only a
carnivorous horse could get off the ground. Is a carnivorous horse possible?
Perhaps a carnivorous horse would be biologically possible if it could evolve,
but would such a diet shift be accessible from where horses would have to
start? And, short of radical constructive surgery, could a horse-descendant
have both front legs and wings? Bats, after all, make wings of their arms. Is
there any possible evolutionary history of skeletal revision that would yield a
six-limbed mammal?
This brings us to our fourth grade of possibility, historical possibility.
There might have been a time, in the very distant past, when the possibility
of six-limbed mammals on Earth had not yet been foreclosed, but it might

106 THE POSSIBLE AND THE ACTUAL The Library of Mendel 107

also be true that once our four-finned fishy ancestors got selected for moving
onto the land, the basic four-limbed architecture was so deeply anchored in
our developmental routines that alteration at this time is no longer possible.
But even that distinction may not be sharp-edged. Is such an alteration in
fundamental building-plan flat impossible, or just highly unlikely, so
resistant to change that only an astronomically improbable sequence of
selective blows could drive it into existence? It seems there might be two
kinds or grades of biological impossibility: violation of a biological law of
nature (if there are any), and "mere" biohistorical consignment to oblivion.
Historical impossibility is simply a matter of opportunities passed up.
There was a time when many of us worried about the possibility of President
Barry Goldwater, but it didn't happen, and after 1964, the odds against such a
thing's ever happening lengthened reassuringly. When lottery tickets are put
on sale, this creates an opportunity for you: you may choose to buy one,
provided you act by a certain date. If you buy one, this creates a further
opportunity for you—the opportunity to win—but soon it slides into the past,
and it is no longer possible for you to win those millions of dollars. Is this
everyday vision we have of opportunities—real opportunities—an illusion?
In what sense could you have won? Does it make a difference if the winning
lottery number is chosen after you buy your ticket, or do you still have an
opportunity to win, a real opportunity, if the winning number is sealed in a
vault before the tickets are put on sale (Dennett 1984)? Is there ever really
any opportunity at all? Could anything happen other than what actually
happens? This dread hypothesis, the idea that only the actual is possible, has
been called actualism (Ayers 1968). It is generally ignored, for good reasons,
but these reasons are seldom discussed. (Dennett 1984, and Lewis 1986, pp.
36-38, offer good reasons for dismissing actualism.)
These familiar and prima facie reliable ideas about possibility can be
summed up in a diagram, but every boundary in it is embattled. As Quine's
questions suggest, there is something fishy about casual catalogues of merely
possible objects, but since science cannot even express—let alone confirm—
the sorts of explanations we crave without drawing such a distinction, there is
little chance that we can simply renounce all such talk. When biologists
wonder whether a horned bird—or even a giraffe with stripes instead of
blotches—is possible, the questions they are addressing epitomize what we
want biology to discover for us. Alerted by Quine, we can be struck by the
dubious metaphysical implications of Richard Dawkins' vivid claim that
there are many more ways of being dead than of being alive, but manifestly
he is getting at something important. We should try to find a way of recasting
such claims in a metaphysically more modest and less contentious
framework—and Darwin's starting in the middle gives us just the foothold
we need. First we can deal with the relation between historical and
FIGURE 5.1
biological possibility, and then perhaps it will suggest some payoffs for how
to make sense of the grander varieties.1
2. THE LIBRARY OF MENDEL
The Argentine poet Jorge Luis Borges is not typically classified as a philos-
opher, but in his short stories he has given philosophy some of its most
valuable thought experiments, most of them gathered in the stunning col-
lection Labyrinths (1962). Among the best is the fantasy—actually, it is
more a philosophical reflection than a narrative—that describes the Library
of Babel. For us, the Library of Babel will be an anchoring vision for helping
to answer very difficult questions about the scope of biological possibility, so
we will pause to explore it at some length. Borges tells of the forlorn
explorations and speculations of some people who find themselves living in
1. Back in 1982, Francois Jacob, the Nobel laureate biologist, published a book entitled
The Possible and the Actual, and I rushed to read it, expecting it to be an eye-opening
essay on how biologists should think about some of these conundrums about possibility.
To my disappointment, the book had very little to say on this topic. It is a fine book, and
has a great title, but the two don't go together, in my humble opinion. The book I was
eager to read hasn't yet been written, apparently, so I'll have to try to write part of it
myself, in this chapter.

108 THE POSSIBLE AND THE ACTUAL The Library of Mendel 109

a vast storehouse of books, structured like a honeycomb, composed of
thousands (or millions or billions) of hexagonal air shafts surrounded by
balconies lined with shelves. Standing at a railing and looking up or down,
one sees no top or bottom to these shafts. Nobody has ever found a shaft that
isn't surrounded by six neighboring shafts. They wonder: is the warehouse
infinite? Eventually, they decide that it is not, but it might as well be, for it
seems that on its shelves—in no order, alas—lie all the possible books.
Suppose that each book is 500 pages long, and each page consists of 40
lines of 50 spaces, so there are two thousand character-spaces per page. Each
space either is blank, or has a character printed on it, chosen from a set of
100 (the upper- and lowercase letters of English and other European
languages, plus the blank and punctuation marks).2 Somewhere in the Li-
brary of Babel is a volume consisting entirely of blank pages, and another
volume is all question marks, but the vast majority consist of typographical
gibberish; no rules of spelling or grammar, to say nothing of sense, prohibit
the inclusion of a volume. Five hundred pages times 2,000 characters per
page gives 1,000,000 character-spaces per book, so there are 1001,000,000 books
in the Library of Babel. Since it is estimated3 that there are only 10040 (give or
take a few) particles (protons, neutrons, and electrons) in the region of the
universe we can observe, the Library of Babel is not remotely a physically
possible object, but, thanks to the strict rules with which Borges constructed
it in his imagination, we can think about it clearly.
Is this truly the set of all possible books? Obviously not—since they are
restricted to being printed from "only" 100 different characters, excluding,
we may suppose, the characters of Greek, Russian, Chinese, Japanese, and
Arabic, thereby overlooking many of the most important actual books. Of
course, the Library does contain superb translations of all these actual books
into English, French, German, Italian,..., as well as uncountable trillions of
shoddy translations of each book. Books of more than 500 pages are there,
beginning in one volume and continuing without a break in some other
volume or volumes.
It is amusing to think about some of the volumes that must be in the
Library of Babel somewhere. One of them is the best, most accurate 500-
page biography of you, from the moment of your birth until the moment of
your death. Locating it, however, would be all but impossible (that slippery
word), since the Library also contains kazillions of volumes that are mag-
nificently accurate biographies of you up till your tenth, twentieth, thirtieth,
fortieth ... birthday, and completely false about subsequent events of your
life—in a kazillion different and diverting ways. But even finding one read-
able volume in this huge storehouse is unlikely in the extreme.
We need some terms for the quantities involved. The Library of Babel is
not infinite, so the chance of finding anything interesting in it is not literally
infinitesimal.4 These words exaggerate in a familiar way—we caught Darwin
doing it in his summary, where he helped himself to an illicit "infinitely"—
but we should avoid them. Unfortunately, all the standard metaphors—
"astronomically large," "a needle in a haystack," "a drop in the ocean"—fall
comically short. No actual astronomical quantity (such as the number of
elementary particles in the universe, or the time since the Big Bang measured
in nanoseconds) is even visible against the backdrop of these huge but finite
numbers. If a readable volume in the Library were as easy to find as a
particular drop in the ocean, we'd be in business! If you were dropped at
random into the Library, your chance of ever encountering a volume with so
much as a grammatical sentence in it would be so vanishingly small that we
might do well to capitalize the term—"Vanishingly" small—and give it a
mate, "Vastly," short for "Very-much-more-than-astronomically."5
Moby Dick is in the Library of Babel, of course, but so are 100,000,000
mutant impostors that differ from the canonical Moby Dick by a single


2. Borges chose slightly different figures: books 410 pages long, with 40 lines of 80
characters each. The total number of characters per book is close enough to mine
(1,312,000 versus 1,000,000) to make no difference. 1 chose my rounder numbers for
ease of handling. Borges chose a character set with only 25 members, which is enough
for uppercase Spanish (with a blank, a comma, and a period as the only punctuation ), but
not for English. I chose the more commodious 100 to make room without any doubt for
the upper- and lowercase letters and punctuation of all the Roman-alphabet languages.
3. Stephen Hawking (1988, p. 129) insists on putting it this way: "There are something
like ten million million million million million million million million million million
million million million (1 with eighty zeroes after it) particles in the region of the
universe that we can observe." Denton (1985 ) provides the estimate of 1070 atoms in the
observable universe. Eigen (1992, p. 10) calculates the volume of the universe as 1084
cubic centimeters.
4. The Library of Babel is finite, but, curiously enough, it contains all the grammatical
sentences of English within its walls. But that's an infinite set, and the library is finite! Still,
any sentence of English, of whatever length, can be broken down into 500-page chunks,
each of which is somewhere in the library! How is this possible? Some books may get
used more than once. The most profligate case is the easiest to understand: since there
are volumes that each contain a single character and are otherwise blank, repeated use of
these 100 volumes will create any text of any length. As Quine points out in his infor-
mative and amusing essay "Universal Library" (in Quine 1987), if you avail yourself of
this strategy of re-using volumes, and translate everything into the ASCII code your word-
processor uses, you can store the whole Library of Babel in two extremely slender
volumes, in one of which is printed a 0 and in the other of which appears a 1! (Quine also
points out that the psychologist Theodor Fechner propounded the fantasy of the univer-
sal library long before Borges.)
5. Quine (1987) coins the term "hyperastronomic" for the same purpose.

116 THE POSSIBLE AND THE ACTUAL
the other organisms extant, viability is a constantly changing property, a
moving target, not a fixed condition. This problem is minimized if we join
Darwin in starting in the middle, with currently existing environments, and
extrapolate cautiously to earlier and later possibilities. We can leave till later
a consideration of the initial bootstrapping that may (or must) have happened
to set this coevolution of organisms and their environments in motion.
The third complexity concerns the relationship between the texts of the
genomes that do determine viable organisms, and the features those organ-
isms exhibit. As we have already noted several times in passing, there is no
simple mapping of nucleotide "words" onto Mendelian genes—putative
carriers of the "specs" (as an engineer would say) for one feature or another.
It is simply not the case that there is a sequence of nucleotides that spells
"blue eyes" or "webbed feet" or "homosexual" in any descriptive language.
And you can't spell "firm" or "flavorful" in the language of tomato DNA—
even though you can revise the nucleotide sequence in that language so that
the effect is firmer, more flavorful tomatoes.
When this complication is acknowledged, it is usually pointed out that
genomes are not like descriptions or blueprints of finished products, but
more like recipes for building them. This does not mean, as some critics have
contended, that it is always—or even ever—a mistake to speak of a gene for
this or that. The presence or absence of an instruction in a recipe can make a
typical and important difference, and whatever difference it makes may be
correctly described as what the instruction—the gene—is "for." This point
has been so frequently and influentially missed by the critics that it is worth
pausing to expose its error vividly. Richard Dawkins has come up with an
example that does this so well that it is worth quoting in full (it also
highlights the importance of the second of our complications, the relativity of
viability to environment):
Reading is a learned skill of prodigious complexity, but this provides no
reason in itself for scepticism about the possible existence of a gene for
reading. All we would need in order to establish the existence of a gene for
reading is to discover a gene for not reading, say a gene which induced a
brain lesion causing specific dyslexia. Such a dyslexic person might be
normal and intelligent in all respects except that he could not read. No
geneticist would be particularly surprised if this type of dyslexia turned
out to breed true in some Mendelian fashion. Obviously, in this event, the
gene would only exhibit its effect in an environment which included
normal education. In a prehistoric environment it might have had no
detectable effect, or it might have had some different effect and have been
known to cave-dwelling geneticists as, say, a gene for inability to read
animal footprints. In our educated environment it would properly be called
a gene 'for' dyslexia, since dyslexia would be its most salient consequence.
The Complex Relation Between Genome and Organism 117
Similarly, a gene which caused total blindness would also prevent reading,
but it would not usefully be regarded as a gene for not reading. This is
simply because preventing reading would not be its most obvious or de-
bilitating phenotypic effect. [Dawkins 1982, p. 23. See also Dawkins 1989a,
pp. 281-82, and Sterelny and Kitcher 1988.]
The indirect way in which groups of codons—triplets of DNA nucle-
otides—instruct the building process does not prohibit us, then, from speak-
ing of a gene for x or for y, using the familiar geneticists' shorthand, and
bearing in mind that that is what we are doing. But it does mean that there
may be fundamental differences between the space of genomes and the space
of "possible" organisms. The fact that we can consistently describe a finished
product—say, a giraffe with green stripes instead of brown blotches —does
not guarantee that there is a DNA recipe for making it. It may just be that,
because of the peculiar requirements of development, there simply is no
starting point in DNA that has such a giraffe as its destination.
This may seem very implausible. What could be impossible about a giraffe
with green stripes? Zebras have stripes, drakes have green feathers on their
heads—there is nothing biologically impossible about the properties in
isolation, and surely they can be put together in one giraffe! So you'd think.
But you must not count on it. You'd probably also think a striped animal with
a spotted tail was possible, but it may well not be. James Murray (1989) has
developed mathematical models that show how the developmental process of
distributing color on animals could readily make a spotted animal with a
striped tail, but not vice versa. This is suggestive, but not yet—as some have
rashly said—a strict proof of impossibility. Anyone who had learned how to
build a tiny ship in a bottle—a hard enough trick— might think it was flat
impossible to put a whole fresh pear in a narrow-necked bottle, but it isn't;
witness the bottles of Poire William liqueur. How is it done? Could the
molten glass somehow be blown around a pear without scorching it? No, the
bottles are hung on the trees in the spring so that the pears can grow inside
them. Proving that there is no straightforward way for biology to accomplish
some trick is never a proof of impossibility. Remember Orgel's Second Rule!
In his account of Biomorph Land, Dawkins stresses that a tiny—indeed
minimal—change in the genotype (the recipe) can produce a strikingly large
change in the phenotype (the resulting individual organism), but he tends to
slight one of the major implications of this: if a single step in the genotype
can produce a giant step in the phenotype, intermediate steps for the
phenotype may be simply unavailable, given the mapping rules. To take a
deliberately extreme and fanciful example, you might think that if a beast
could have twenty-centimeter tusks and forty-centimeter tusks, it would
stand to reason that it could also have thirty-centimeter tusks, but the rules

122 THE POSSIBLE AND THE ACTUAL
space of the Library of Mendel, but it does put a difficult burden of proof on
anyone who thinks that there are laws of biology over and above the laws of
mathematics and physics. Consider the fate of "Dollo's Law," for instance.
'Dollo's Law' states that evolution is irreversible....[But] There is no reason
why general trends in evolution shouldn't be reversed. If there is a trend
towards large antlers for a while in evolution, there can easily be a subse-
quent trend towards smaller antlers again. Dollo's Law is really just a state-
ment about the statistical improbability of following exactly the same
evolutionary trajectory twice (or indeed any particular trajectory), in
either direction. A single mutational step can easily be reversed. But for
larger numbers of mutational steps... the mathematical space of all possible
trajectories is so vast that the chance of two trajectories ever arriving at the
same point becomes vanishingly small __ There is nothing mysterious or
mystical about Dollo's Law, nor is it something that we go out and 'test' in
nature. It follows simply from the elementary laws of probability. [Dawkins
1986a, p. 94.]
There is no shortage of candidates for the role of "irreducible biological
law." For instance, many have argued that there are "developmental laws" or
"laws of form" that constrain the relation between genotype and pheno-type.
In due course we will consider their status, but already we can locate at least
some of the most salient constraints on biological possibility as not "laws of
biology" but just inescapable features of the geometry of design space, like
Dollo's Law (or the Hardy-Weinberg Law of gene frequency, which is
another application of probability theory, pure and simple).
Take the case of the horned birds. As Maynard Smith notes, there aren't
any, and we don't know why. Might it be because they are ruled out by a
biological law? Are horned birds flat impossible? Would they have to be
inviable in any and all possible environments, or is there simply no path at all
"from here to there" because of restrictions on the genome-reading process?
As we have already noted, we should be impressed by the severe restrictions
encountered by this process, but we should not be carried away. Those
restrictions may not be a universal feature, but a temporally and spatially
local feature, analogous to what Seymour Papert has dubbed the QWERTY
phenomenon in the culture of computers and keyboards.
The top row of alphabetic keys of the standard typewriter reads QWERTY.
For me this symbolizes the way in which technology can all too often serve
not as a force for progress but for keeping things stuck. The QWERTY
arrangement has no rational explanation, only a historical one. It was
introduced in response to a problem in the early days of the typewriter:
The keys used to jam. The idea was to minimize the collision problem by
separating those keys that followed one another frequently.... Once
Possibility Naturalized 123
adopted, it resulted in many millions of typewriters and ... the social cost
of change ... mounted with the vested interest created by the fact that so
many fingers now knew how to follow the QWERTY keyboard. QWERTY
has stayed on despite the existence of other, more "rational" systems.
[Papert 1980, p. 33.]12
The imperious restrictions we encounter inside the Library of Mendel may
look like universal laws of nature from our myopic perspective, but from a
different perspective they may appear to count as merely local conditions,
with historical explanations.13 If so, then a restricted concept of biological
possibility is the sort we want; the ideal of a universal concept of biological
possibility will be misguided. But as I have already allowed, this does not
rule out biological laws; it merely sets the burden of proof for those who want
to propose any. And in the meantime, it gives us a frame-work for describing
large and important classes of regularity we discover in the patterns in our
biosphere.
CHAPTER 5: Biological possibility is best seen in terms of accessibility (from
some stipulated location) in the Library of Mendel, the logical space of all
genomes. This concept of possibility treats the connectedness of the Tree of
Life as a fundamental feature of biology, while leaving it open that there may
also be biological laws that will also constrain accessibility.
CHAPTER 6: The R and D done by natural selection in the course of creating
actual trajectories in the Vast space of possibilities can be measured to some
extent. Among the important features of this search space are the solutions to
problems that are perennially attractive and hence predictable, like forced
moves in chess. This explains some of our intuitions about originality,
discovery, and invention, and also clarifies the logic of Darwinian inference
about die past. There is a single, unified Design Space in which the processes
of both biological and human creativity make their tracks, using similar
methods.

12. Others have exploited the QWERTY phenomenon to make similar points: David
1985, Gould 1991a.
13. George Williams (1985, p. 20) puts it this way: "1 once insisted that'... the laws of
Physical science plus natural selection can furnish a complete explanation for any bio-
logical phenomenon' [Williams 1966, pp. 6-7]. I wish now I had taken a less extreme
view and merely identified natural selection as the only theory that a biologist needs in
addition to those of the physical scientist. Both the biologist and the physical scientist
need to reckon with historical legacies to explain any real-world phenomenon."

Drifting and Lifting Through Design Space 125
CHAPTER SIX
Threads of Actuality in
Design Space

1. DRIFTING AND LIFTING THROUGH DESIGN SPACE
The actual animals that have ever lived on Earth are a tiny subset of the
theoretical animals that could exist. These real animals are the products
of a very small number of evolutionary trajectories dvough genetic
space. The vast majority of theoretical trajectories through animal space
give rise to impossible monsters. Real animals are dotted around here
and there among the hypothetical monsters, each perched in its own
unique place in genetic hyperspace. Each real animal is surrounded by
a little cluster of neighbours, most of whom have never existed, but a
few of whom are its ancestors, its descendants and its cousins.
—RICHARD DAWKINS 1986a, p. 73
The actual genomes that have ever existed are a Vanishingly small subset
of the combinatorially possible genomes, just as the actual books in the
world's libraries are a Vanishingly small subset of the books in the imaginary
Library of Babel. As we survey the Library of Babel, we may be struck by
how hard it is to specify a category of books that isn't Vast in membership,
however Vanishingly small it is in relation to the whole. The set of books
composed entirely of grammatical English sentences is a Vast but Vanishing
subset, and the set of readable, sense-making books is a Vast but Vanishing
subset of it. Vanishingly hidden in that subset is the Vast set of books about
people named Charles, and within that set (though Vanishingly hard to find)
is the Vast set of books purporting to tell the truth about Charles Darwin, and
a Vast but Vanishing subset of these consists of books composed en-
tirely in limericks. So it goes. The number of actual books about Charles
Darwin is a huge number, but not a Vast number, and we won't get down to
that set (that set as of today, or as of the year 3000 A.D. ) by just piling on the
restricting adjectives in this fashion. To get to the actual books, we have to
turn to the historical process that created them, in all its grubby particularity.
The same is true of the actual organisms, or their actual genomes.
We don't need laws of biology to "prevent" most of the physical possi-
bilities from becoming actualities; sheer absence of opportunity will account
for most of them. The only "reason" all your nonactual aunts and uncles
never came into existence is that your grandparents didn't have time or
energy (to say nothing of the inclination) to create more than a few of the
nearby genomes. Among the many nonactual possibles, some are—or were—
"more possible" than others: that is, their appearance was more probable than
the appearance of others, simply because they were neighbors of actual
genomes, only a few choices away in the random zipping-up process that
puts together the new DNA volume from the parent drafts, or only one or a
few random typos away in the great copying process. Why didn't the near-
misses happen? No reason; they just didn't happen to happen. And then, as
the actual genomes that did happen to happen began to move away from the
locations in Design Space of the near-misses, their probability of ever
happening grew smaller. They were so close to becoming actual, and then
their moment passed! Will they get another chance? It is possible, but Vastly
improbable, given the Vast size of the space in which they reside.
But what forces, if any, bend the paths of actuality farther and farther away
from their locations? The motion that occurs if there are no forces at all is
called random genetic drift. You might think that drift, being random, would
tend always to cancel itself out, bringing the path back to the same genomes
again and again in the absence of any selective forces, but the very fact that
there is only limited sampling in the huge space (which has a million
dimensions, remember!) leads inevitably to the accumulation of "distance"
between actual genomes (the upshot of "Dollo's Law").
Darwin's central claim is that when the force of natural selection is im-
posed on this random meandering, in addition to drifting there is lifting. Any
motion in Design Space can be measured, but the motion of random drift is,
intuitively, merely sideways; it doesn't get us anywhere important. Consid-
ered as R-and-D work, it is idle, leading to the accumulation of mere typo-
graphical change, but not to the accumulation of design. In fact, it is worse
than that, for most mutations—typos—will be neutral, and most of the typos
that aren't neutral will be deleterious. In the absence of natural selection, the
drift is inexorably downward in Design Space. The situation in the Library of
Mendel is thus precisely like the situation in the Library of Babel. Most
typographical changes to Moby Dick can be supposed to be practically

130 THREADS OF ACTUALITY IN DESIGN SPACE Forced Moves in the Game of Design 131

explored this curious question, and his ingeniously reasoned answer is No. In
"Why Intelligent Aliens Will Be Intelligible," he offers grounds for believing
in something he calls the
Sparseness Principle: Whenever two relatively simple processes have prod-
ucts which are similar, those products are likely to be completely identi-
cal! [Minsky 1985a, p. 119, exclamation point in the original.]
Consider the set of all possible processes, which Minsky interprets a la the
Library of Babel as all permutations of all possible computers. (Any com-
puter can be identified, abstractly, as one "Turing machine" or another, and
these can be given unique identifying numbers, and then put in numerical
order, just like the alphabetical order in the Library of Babel.) Except for a
Vanishing few, the Vast majority of these processes "do scarcely anything at
all." So if you find "two" that do something similar (and worth noticing), they
are almost bound to be one and the same process, at some level of analysis.
Minsky (p. 122) applies the principle to arithmetic:
From all this, I conclude that any entity who searches through the simplest
processes will soon find fragments which do not merely resemble arith-
metic but are arithmetic. It is not a matter of inventiveness or imagination,
only a fact about the geography of the universe of computation, a world far
more constrained than that of real things.
The point is clearly not restricted to arithmetic, but to all "necessary
truths"—what philosophers since Plato have called a priori knowledge. As
Minsky (p. 119 ) says, "We can expect certain 'apriori' structures to appear,
almost always, whenever a computation system evolves by selection from a
universe of possible processes." It has often been pointed out that Plato's
curious theory of reincarnation and reminiscence, which he offers as an
explanation of the source of our a priori knowledge, bears a striking re-
semblance to Darwin's theory, and this resemblance is particularly striking
from our current vantage point. Darwin himself famously noted the resem-
blance in a remark in one of his notebooks. Commenting on the claim that
Plato thought our "necessary ideas" arise from the pre-existence of the soul,
Darwin wrote: "read monkeys for preexistence" (Desmond and Moore 1991,
p. 263).
We would not be surprised, then, to find that extra-terrestrials had the
same unshakable grip on "2 + 2 = 4" and its kin that we do, but we would be
surprised, wouldn't we, if we found them using the decimal system for
expressing their truths of arithmetic. We are inclined to believe that our
fondness for it is something of a historical accident, derived from counting
on our two five-digit hands. But suppose they, too, have a pair of hands, each
with five subunits. The "solution" of using-whatever-you've-got to count on
is a fairly obvious one, if not quite in the forced-move category.3 It would not
be particularly surprising to find that our aliens had a pair of prehensile
appendages, considering the good reasons there are for bodily symmetry, and
the frequency of problems that require one thing to be manipulated relative to
another. But that there should be five subunits on each appendage looks like a
QWERTY phenomenon that has been deeply rooted for hundreds of millions
of years—a mere historical happenstance that has restricted our options, but
should not be expected to have restricted theirs. But perhaps we
underestimate the Tightness, the rationality, of having five subunits. For
reasons we have not yet fathomed, it may be a Good Idea in general, and not
merely something we are stuck with. Then it would not be amazing after all
to find that our interlocutors from outer space had converged on the same
Good Idea, and counted in tens, hundreds, and thousands.
We would be flabbergasted, however, to find them using the very symbols
we use, the so-called arabic numerals: "1," "2," "3" ... We know that right
here on Earth there are perfectly fine alternatives, such as the Arabic nu-
merals, " I," "v," " v," "i" ... as well as some not-so-viable alternatives, such
as roman numerals, "i," "ii," "iii," "iv" ... If we found the inhabitants of
another planet using our arabic numerals, we would be quite sure that it was
no coincidence—there had to be a historical connection. Why? Because the
space of possible numeral shapes in which there is no reason for choosing
one over the others is Vast; the likelihood of two independent "searches"
ending up in the same place is Vanishing.
Students often have a hard time keeping clear about the distinction be-
tween numbers and numerals. Numbers are the abstract, "Platonic" objects
that numerals are the names of. The arabic numeral "4" and the roman
numeral "IV" are simply different names for one and the same thing—the
number 4. (I can't talk about the number without naming it in one way or
another, any more than I can talk about Clinton without using some word

3. Seymour Papert (1993, p. 90) describes observing a "learning disabled" boy in a
classroom in which counting on your fingers was forbidden: "As he sat in the resource
room I could see him itching to do finger manipulations. But he knew better. Then I saw
him look around for something else to count with. Nothing was at hand. I could see his
frustration grow. What could I do?... Inspiration came! I walked casually up to the boy
and said out loud: 'Did you think about your teeth?' I knew instantly from his face that he
got the point, and from the aide's face that she didn't. 'Learning disability indeed!' I said
to myself. He did his sums with a half-concealed smile, obviously delighted with the
subversive idea." (When considering using-whatever-you've-got as a possible forced
move, it is worth recalling that not all peoples of our Earth have used the decimal system;
the Mayans, for instance, used a base-20 system.)

134 THREADS OF ACTUALITY IN DESIGN SPACE The Unity of Design Space 135

There is no single summit in Design Space, nor a single staircase or ladder
with calibrated steps, so we cannot expect to find a scale for comparing
amounts of design work across distant developing branches. Thanks to the
vagaries and digressions of different "methods adopted," something that is in
some sense just one problem can have both hard and easy solutions,
requiring more or less work. There is a famous story about the mathema-
tician and physicist (and coinventor of the computer) John von Neumann,
who was legendary for his lightning capacity to do prodigious calculations in
his head. (Like most famous stories, this one has many versions, of which I
choose the one that best makes the point I am pursuing.) One day a colleague
approached him with a puzzle that had two paths to solution, a laborious,
complicated calculation and an elegant, Aha!-type solution. This colleague
had a theory: in such a case, mathematicians work out the laborious solution
while the (lazier, but smarter) physicists pause and find the quick and easy
solution. Which solution would von Neumann find? You know the sort of
puzzle: Two trains, 100 miles apart, are approaching each other on the same
track, one going 30 miles per hour, the other going 20 miles per hour. A bird
flying 120 miles per hour starts at train A (when they are 100 miles apart),
flies to train B, turns around and flies back to the approaching train A, and so
forth, until the trains collide. How far has the bird flown when the collision
occurs? "Two hundred forty miles," Von Neumann answered almost
instantly. "Darn," replied his colleague, "I predicted you'd do it the hard
way." "Ay!" von Neumann cried in embarrassment, smiting his forehead.
"There's an easy way!" (Hint: how long till the trains collide?)
Eyes are the standard example of a problem that has been solved many
times, but eyes that may look just the same (and see just the same) may have
been achieved by R-and-D projects that involved different amounts of work,
thanks to the historical peculiarities of the difficulties encountered along the
way. And the creatures that don't have eyes at all are neither better nor worse
on any absolute scale of design; their lineage has just never been given this
problem to solve. It is this same variability in luck in the various lineages that
makes it impossible to define a single Archimedean point from which global
progress could be measured. Is it progress when you have to work an extra
job to pay for the high-priced mechanic you have to hire to fix your car when
it breaks because it is too complex for you to fix in the way you used to fix
your old clunker? Who is to say? Some lineages get trapped in (or are lucky
enough to wander into—take your pick) a path in Design Space in which
complexity begets complexity, in an arms race of competitive design. Others
are fortunate enough (or unfortunate enough-take your pick) to have hit upon
a relatively simple solution to life's problems at the outset and, having nailed
it a billion years ago, have had nothing much to do in the way of design work
ever since. We human beings,
complicated creatures that we are, tend to appreciate complexity, but that
may well be just an aesthetic preference that goes with our sort of lineage;
other lineages may be as happy as clams with their ration of simplicity.
3. THE UNITY OF DESIGN SPACE
The formation of different languages and of distinct species, and the
proofs that both have been developed through a gradual process, are
curiously the same.
—CHARLES DARWIN 1871, p. 59
It will not have gone unnoticed that my examples in this chapter have
wandered back and forth between the domain of organisms or biological
design, on the one hand, and the domain of human artifacts—books, prob-
lems solved, and engineering triumphs on the other. This was by design, not
accident, of course. It was to help set the stage for, and provide lots of
ammunition for, a Central Salvo: there is only one Design Space, and ev-
erything actual in it is united with everything else. And I hardly need add that
it was Darwin who taught us this, whether he quite realized it or not.
Now I want to go back over the ground we have covered, highlighting the
evidence for this claim, and drawing out a few more implications of it and
grounds for believing it. The similarities and continuities are of tremendous
importance, I think, but in later chapters I will also point to some important
dissimilarities between the human-made portions of the designed world and
the portions that were created without benefit of the sort of locally con-
centrated, foresighted intelligence we human artificers bring to a problem.
We noted at the outset that the Library of Mendel (in the form of printed
volumes of the letters A, C, G, T ) is contained within the Library of Babel, but
we should also note that at least a very large portion of the Library of Babel
(What portion? See chapter 15) is in turn "contained" in the Library of Men-
del, because we are in the Library of Mendel ( our genomes are, and so are the
genomes of all the life forms our lives depend on). The Library of Babel de-
scribes one aspect of our "extended phenotype" (Dawkins 1982 ). That is, in
the same way that spiders make webs and beavers make dams, we make
(among many other things) books. You can't assess the spider's genome for
viability without a consideration of the web that is part of the normal equip-
ment of the spider, and you can't assess the viability of our genomes (not any
longer, you can't) without recognizing that we are a species with culture, a
representative part of which is in the form of books. We are not just designed,
we are designers, and all our talents as designers, and our products, must
emerge non-miraculously from the blind, mechanical processes of Darwinian

142 THREADS OF ACTUALITY IN DESIGN SPACE The Unity of Design Space 143

puncture the egg of a fly of the Drosophila willistoni species, and in the
process picked up some of that species' characteristic DNA, which it then
inadvertently transmitted to the egg of a (wild) Drosophila melanogaster fly!
This could explain the sudden explosion in the wild of a particular DNA
element common in D. willistoni but previously unheard of in D. melano-
gaster populations. She might add: What else could explain it? It sure looks
like species plagiarism.
Other researchers are looking at other possible vehicles for speedy design
travel in the world of natural (as opposed to artificial) genetics. If they find
them, they will be fascinating—but no doubt rare—exceptions to the
orthodox pattern: genetic transmission of design by chains of direct descent
only.6 We are inclined, as just noted, to contrast this feature sharply with
what we find in the freewheeling world of cultural evolution, but even there
we can detect a powerful dependence on the combination of luck and
copying.
Consider all the wonderful books in the Library of Babel that will never be
written, even though the process that could create each of them involves no
violation or abridgment of the laws of nature. Consider some book in the
Library of Babel that you yourself might love to write—and that only you
could write—for instance, the poetically expressed autobiographical tale of
your childhood that would bring tears and laughter to all readers. We know
that there are Vast numbers of books with just these features in the Library of
Babel, and each is composable in only a million keystrokes. At the daw-dling
rate of five hundred strokes a day, the whole project shouldn't take you much
longer than six years, with generous vacations. Well, what's stopping you?
You have fingers that work, and all the keys on your word-processor can be
depressed independently.
Nothing is stopping you. That is, there needn't be any identifiable forces,
or laws of physics or biology or psychology, or salient disabilities brought on
by identifiable circumstances (such as an ax embedded in your brain, or a
gun pointed at you by a credible threatener). There are Vastly many books
that you are never going to write "for no reason at all." Thanks to the myriad
particular twists and turns of your life to date, you just don't happen to be
well disposed to compose those sequences of keystrokes.
If we want to get some perspective—limited, to be sure—on what patterns
go into creating your own authorial dispositions, we will have to consider the
transmission of Design to you from the books you have read. The books that
actually come to exist in the world's libraries are deeply
dependent not just on their authors' biological inheritance, but on the books
that have come before them. This dependence is conditioned by coincidences
or accidents at every turning. Just look at my bibliography to discover the
main lines of genealogy of this book. 1 have been reading and writing about
evolution since I was an undergraduate, but if I had not been encouraged by
Doug Hofstadter in 1980 to read Dawkins' The Selfish Gene, I probably
would not have begun coalescing some of the interests and reading habits
that have been major shapers of this book. And if Hofstadter had not been
asked by The New York Review of Books to review my book Brainstorms
(1978), he probably would never have hit upon the bright idea of proposing
that we collaborate on a book, The Mind's I (1981), and then we would not
have had the opportunity for mutual book-recommending that led me to
Dawkins, and so forth. Even if I had read the same books and articles by
following other paths, in a different order, I would not be conditioned in
exactly the same way by that reading, and hence would have been unlikely to
have composed (and edited, and re-edited) just the string of symbols you are
now reading.
Can we measure this transmission of Design in culture? Are there units of
cultural transmission analogous to the genes of biological evolution? Daw-
kins (1976 ) has proposed that there are, and has given them a name: memes.
Like genes, memes are supposed to be replicators, in a different medium, but
subject to much the same principles of evolution as genes. The idea that there
might be a scientific theory, memetics, strongly parallel to genetics, strikes
many thinkers as absurd, but at least a large part of their skepticism is based
on misunderstanding. This is a controversial idea, which will get careful
consideration in chapter 12, but in the meantime we can set aside the
controversies and just use the term as a handy word for a salient ( mem-
orable) cultural item, something with enough Design to be worth saving—or
stealing or replicating.
The Library of Mendel (or its twin, the Library of Babel—they
are contained in each other, after all) is as good an approximate model of
Universal Design Space as we could ever need to think about. For the last
four billion years or so, the Tree of Life has been zigzagging through this
Vast multidimensional space, branching and blooming with virtually un-
imaginable fecundity, but nevertheless managing to fill only a Vanishingly
small portion of that space of the Possible with Actual designs.7 According


6. The genetic elements transferred in Drosophila are "intragenomic parasites" and
probably have a negative effect on the adaptedness of their host organisms, so we
shouldn't get our hopes up unduly. See Engels 1992.
7. "I confess that I believe the emptiness of phenotypic space is filled with red her-
rings. ... Under the null hypothesis that no constraints at all exist, the branching path-
ways through space taken by this process constitute a random-branching walk in a

144 THREADS OF ACTUALITY IN DESIGN SPACE
to Darwin's dangerous idea, all possible explorations of Design Space are
connected. Not only all your children and your children's children, but all
your brainchildren and your brainchildren's brainchildren must grow from
the common stock of Design elements, genes and memes, that have so far
been accumulated and conserved by the inexorable lifting algorithms, the
ramps and cranes and cranes-atop-cranes of natural selection and its
products.
If this is right, then all the achievements of human culture—language, art,
religion, ethics, science itself—are themselves artifacts ( of artifacts of arti-
facts ...) of the same fundamental process that developed the bacteria, the
mammals, and Homo sapiens. There is no Special Creation of language, and
neither art nor religion has a literally divine inspiration. If there are no
skyhooks needed to make a skylark, there are also no skyhooks needed to
make an ode to a nightingale. No meme is an island.
Life and all its glories are thus united under a single perspective, but some
people find this vision hateful, barren, odious. They want to cry out against
it, and above all, they want to be magnificent exceptions to it. They, if not
the rest, are made in God's image by God, or, if they are not religious, they
want to be skyhooks themselves. They want somehow to be intrinsic sources
of Intelligence or Design, not "mere" artifacts of the same processes that
mindlessly produced the rest of the biosphere.
So a lot is at stake. Before we turn, in part HI, to examine in detail the
implications of the upward spread of universal acid through human culture,
we need to secure the base camp, by considering a variety of deep challenges
to Darwinian thinking within biology itself. In the process, our vision of the
intricacy and power of the underlying ideas will be enhanced.
CHAPTER 6: There is one Design Space, and in it the Tree of Life has grown a
branch that has recently begun casting its own exploratory threads into that
Space, in the form of human artifacts. Forced moves and other good ideas
are like beacons in Design Space, discovered again and again, by the
ultimately algorithmic search processes of natural selection and human
investigation. As Darwin appreciated, we can retrospectively detect the
historical fact of descent, anywhere in Design Space, when we find shared
design features that would be Vastly unlikely to coexist unless there was a
thread of descent between them. Historical reasoning about evolution dius
depends on accepting Paley's premise: the world is full of good Design,
which took work to create.
This completes die introduction to Darwin's dangerous idea. Now we
The Unity of Design Space 145
must secure its base camp in biology, in part 11, before looking at its power
to transform our understanding of the human world, in part HI.
CHAPTER 7: How did the Tree of Life get started? Skeptics have thought a
stroke of Special Creation—a skyhook—must be needed to get the evolu-
tionary process going. There is a Darwinian answer to this challenge, how-
ever, which exhibits the power of Darwin's universal acid to work its way
down through the lowest levels of the Cosmic Pyramid, showing how even
the laws of physics might emerge from chaos or nothingness without re-
course to a Special Creator, or even a Lawgiver. This dizzying prospect is
one of the most feared aspects of Darwin's dangerous idea, but the fear is
misguided.

high-dimensional space. The typical property of such a walk in a high-dimensional space
is that most of the space is empty" (Kauffman 1993, p. 19).

PART II
DARWINIAN THINKING
IN BIOLOGY
Evolution is a change from a no-howish untalkaboutable all-alikeness by
continuous sticktogetherations and somethingelsifications.
—WILUAM JAMES 1880
Nothing in biology makes sense except in the light of evolution.
—THEODOSIUS DOBZHANSKY 1973

CHAPTER SEVEN
Priming Darwin's Pump

1. BACK BEYOND DARWIN'S FRONTIER
And God said, Let the earth bring forth grass, the herb yielding seed,
and the fruit tree yielding fruit after his kind, whose seed is in itself,
upon the earth: and it was so.
And the earth brought forth grass, and herb yielding seed after his
kind, and the tree yielding fruit, whose seed was in itself, after his kind:
and God saw tint it was good.
—GENESIS 1:11—12
From what sort of seed could the Tree of Life get started? That all life on
Earth has been produced by such a branching process of generation is now
established beyond any reasonable doubt. It is as secure an example of a
scientific fact as the roundness of the Earth, thanks in large part to Darwin.
But how did the whole process get started in the first place? As we saw in
chapter 3, Darwin not only started in the middle; he cautiously refrained
from pushing his own published thinking back to the beginning—the ulti-
mate origin of life and its preconditions. When pressed by correspondents,
he had little more to say in private. In a famous letter, he surmised that it
was quite possible that life began in "a warm little pond," but he had no
details to offer about the likely recipe for this primeval preorganic soup. And
in response to Asa Gray, as we saw (see page 67), he left wide open the
possibility that the laws that would govern this Earth-shattering move were
themselves designed—presumably by God.
His reticence on this score was wise on several counts. First, no one knew
better than he the importance of anchoring a revolutionary theory in the
bedrock of empirical facts, and he knew that he could only speculate, with
scant hope in his own day of getting any substantive feedback. After all, as
we have already seen, he didn't even have the Mendelian concept of the

150 PRIMING DARWIN'S PUMP Back Beyond Darwin's Frontier 151

gene, let alone any of the molecular machinery underlying it. Darwin was an
intrepid deducer, but he also knew when he didn't have enough premises to
go on. Besides, there was his concern for his beloved wife, Emma, who
desperately wanted to cling to her religious beliefs, and who could already
see the threat looming in her husband's work. Yet his reluctance to push any
farther into this dangerous territory, at least in public, went beyond his
consideration for her feelings. There is a wider ethical consideration at stake,
which Darwin certainly appreciated.
Much has been written about the moral dilemmas that scientists face when
the discovery of a potentially dangerous fact puts their love of truth at odds
with their concern for the welfare of others. Under what conditions, if any,
would they be obliged to conceal the truth? These can be real dilemmas, with
powerful and hard-to-plumb considerations on both sides. But there is no
controversy at all about what a scientist's ( or philosopher's) moral
obligations should be regarding his or her speculations. Science doesn't often
advance by the methodical piling up of demonstrable facts; the "cutting edge"
is almost always composed of several rival edges, sharply competing and
boldly speculative. Many of these speculations soon prove to be misbegotten,
however compelling at the outset, and these necessary by-products of
scientific investigation should be considered to be as potentially hazardous as
any other laboratory wastes. One must consider their environmental impact.
If their misapprehension by the public would be apt to cause suffering—by
misleading people into dangerous courses of action, or by undercutting their
allegiance to some socially desirable principle or creed—scientists should be
particularly cautious about how they proceed, scrupulous about labeling
speculations as such, and keeping the rhetoric of persuasion confined to its
proper targets.
But ideas, unlike toxic fumes or chemical residues, are almost impossible
to quarantine, particularly when they concern themes of abiding human
curiosity, so, whereas there is no controversy at all about the principle of
responsibility here, there has been scant agreement, then or now, about how
to honor it. Darwin did the best he could: he kept his speculations pretty
much to himself.
We can do better. The physics and chemistry of life are now understood in
dazzling detail, so that much more can be deduced about the necessary and
(perhaps) sufficient conditions for life. The answers to the big questions must
still involve a large measure of speculation, but we can mark the speculations
as such, and note how they could be confirmed or discon-firmed. There
would be no point any more in trying to pursue Darwin's policy of reticence;
too many very interesting cats are already out of the bag. We may not yet
know exactly how to take all these ideas seriously, but thanks to Darwin's
secure beachhead in biology, we know that we can and must.
It is small wonder that Darwin didn't hit upon a suitable mechanism of
heredity. What do you suppose his attitude would have been to the spec-
ulation that within the nucleus of each of the cells in his body there was a
copy of a set of instructions, written on huge macromolecules, in the form of
double helixes tightly kinked into snarls to form a set of forty-six chro-
mosomes? The DNA in your body, unsnarled and linked, would stretch to the
sun and back several—ten or a hundred—times. Of course, Darwin is the
man who painstakingly uncovered a host of jaw-dropping complexities in the
lives and bodies of barnacles, orchids, and earthworms, and described them
with obvious relish. Had he had a prophetic dream back in 1859 about the
wonders of DNA, he would no doubt have reveled in it, but I wonder if he
could have recounted it with a straight face. Even to those of us accustomed
to the "engineering miracles" of the computer age, the facts are hard to
encompass. Not only molecule-sized copying machines, but proofreading
enzymes that correct mistakes, all at blinding speed, on a scale that super-
computers still cannot match. "Biological macromolecules have a storage
capacity that exceeds that of the best present-day information stores by
several orders of magnitude. For example, the information density' in the
genome of E coli, is about 1027 bits/m3" (Kiippers 1990, p. 180).
In chapter 5, we arrived at a Darwinian definition of biological possibility
in terms of accessibility within the Library of Mendel, but the precondition
for that Library, as we noted, was the existence of genetic mechanisms of
staggering complexity and efiiciency. William Paley would have been trans-
ported with admiration and wonder at the atomic-level intricacies that make
life possible at all. How could they themselves have evolved if they are the
precondition for Darwinian evolution?
Skeptics about evolution have argued that this is the fatal flaw in Darwin-
ism. As we have seen, the power of the Darwinian idea comes from the way
it distributes the huge task of Design through vast amounts of time and space,
preserving the partial products as it proceeds. In Evolution: A Theory in
Crisis, Michael Denton puts it this way: the Darwinian assumes "that islands
of function are common, easily found in the first place, and that it is easy to
go from island to island through functional intermediates" (Denton 1985, p.
317). This is almost right, but not quite. Indeed, the central claim of
Darwinism is that the Tree of Life spreads out its branches, connecting
"islands of function" with isthmuses of intermediate cases, but nobody said
the passage would be "easy" or the safe stopping places "common." There is
only one strained sense of "easy" in which Darwinism is committed to these
isthmus-crossings' being easy: since every living thing is a descendant of a
living thing, it has a tremendous leg up; all but the tiniest fraction of its
recipe is guaranteed to have time-tested viability. The lines of genealogy are
lifelines indeed; according to Darwinism, the only hope of entering this
cosmic maze of junk and staying alive is to stay on die isthmuses.

152 PRIMING DARWIN'S PUMP Back Beyond Darwin's Frontier 153


FIGURE 7.1
But how could this process get started? Denton (p. 323) goes to some
lengths to calculate the improbability of such a start-up, and arrives at a
suitably mind-numbing number.
To get a cell by chance would require at least one hundred functional
proteins to appear simultaneously in one place. That is one hundred si-
multaneous events each of an independent probability which could hardly
be more than 10-20 giving a maximum combined probability of 102000
This probability is Vanishing indeed—next to impossible. And it looks at
first as if the standard Darwinian response to such a challenge could not as a
matter of logic avail us, since the very preconditions for its success—a
system of replication with variation—are precisely what only its success
would permit us to explain. Evolutionary theory appears to have dug itself
into a deep pit, from which it cannot escape. Surely the only thing that could
save it would be a skyhook! This was Asa Gray's fond hope, and the more we
have learned about the intricacies of DNA replication, the more enticing this
idea has become to those who are searching for a place to bail out science
with some help from religion. One might say that it has appeared to many to
be a godsend. Forget it, says Richard Dawkins:
Maybe, it is argued, the Creator does not control the day-to-day succession
of evolutionary events, maybe he did not frame the tiger and the lamb,
maybe he did not make a tree, but he did set up the original machinery of
replication and replicator power, the original machinery of DNA and pro-
tein that made cumulative selection, and hence all of evolution, possible.
This is a transparently feeble argument, indeed it is obviously self-
defeating. Organized complexity is the thing we are having difficulty ex-
plaining. Once we are allowed simply to postulate organized complexity, if
only the organized complexity of the DNA/protein replicating engine, it is
relatively easy to invoke it as a generator of yet more organized complex-
ity.... But of course any God capable of intelligently designing something
as complex as the DNA/protein replicating machine must have been at least
as complex and organized as the machine itself. [Dawkins 1986a, p. 141.]
As Dawkins goes on to say (p. 316), "The one thing that makes evolution
such a neat theory is that it explains how organized complexity can arise out
of primeval simplicity." This is one of the key strengths of Darwin's idea, and
the key weakness of the alternatives. In fact, I once argued, it is unlikely that
any other theory could have this strength:
Darwin explains a world of final causes and teleological laws with a prin-
ciple that is, to be sure, mechanistic but—more fundamentally—utterly
independent of "meaning" or "purpose". It assumes a world that is absurd
in the existentialist's sense of the term: not ludicrous but pointless, and this
assumption is a necessary condition of any non-question-begging account
of purpose. Whether we can imagine a non-mechanistic but also non-
question-begging principle for explaining design in the biological world is
doubtful; it is tempting to see the commitment to non-question-begging
accounts here as tantamount to a commitment to mechanistic materialism,
but the priority of these commitments is clear ___ One argues: Darwin's
materialistic theory may not be the only non-question-begging theory of
these matters, but it is one such theory, and the only one we have found,
which is quite a good reason for espousing materialism. [Dennett 1975, pp.
171-72.]
Is that a fair or even an appropriate criticism of the religious alternatives?
One reader of an early draft of this chapter complained at this point, saying
that by treating the hypothesis of God as just one more scientific hypothesis,
to be evaluated by the standards of science in particular and rational thought
in general, Dawkins and I are ignoring the very widespread claim by be-
lievers in God that their faith is quite beyond reason, not a matter to which
such mundane methods of testing applies. It is not just unsympathetic, he
claimed, but strictly unwarranted for me simply to assume that the scientific
method continues to apply with full force in this domain of faith.

156 PRIMING DARWIN'S PUMP Molecular Evolution 157

plexity, that of the biological macromolecules, an almost unlimited variety
of structures is possible.
—BERND-OIAF KUPPERS 1990, p. 11
Our task is to find an algorithm, a natural law that leads to the origin of
information.
—MANFRED EIGEN 1992, p. 12
In describing the power of the central claim of Darwinism in the previous
section, I helped myself to a slight (!) exaggeration: I said that every living
thing is the descendant of a living thing. This cannot be true, for it implies an
infinity of living things, a set with no first member. Since we know that the
total number of living things (on Earth, up till now) is large but finite, we
seem to be obliged, logically, to identify a first member—Adam the
Protobacterium, if you like. But how could such a first member come to
exist? A whole bacterium is much, much too complicated just to happen into
existence by cosmic accident. The DNA of a bacterium such as E coli has
around four million nucleotides in it, almost all of them precisely in order. It
is quite clear, moreover, that a bacterium could not get by with much less. So
here is a quandary: since living tilings have existed for only a finite time,
there must have been a first one, but since all living things are complex, there
couldn't have been a first one!
There could only be one solution, and we know it well in outline: before
there were bacteria, with autonomous metabolisms, there were much simpler,
quasi-living things, like viruses, but unlike them in not (yet) having any
living things to live off parasitically. From the chemist's point of view,
viruses are "just" huge, complex crystals, but thanks to their complexity, they
don't just sit there; they "do things." In particular, they reproduce or self-
replicate, with variations. A virus travels light, packing no metabolic
machinery, so it either stumbles upon the energy and materials required for
self-replication or self-repair, or eventually it succumbs to the Second Law of
Thermodynamics and falls apart. Nowadays, living cells provide concen-
trated storehouses for viruses, and viruses have evolved to exploit them, but
in the early days, they had to scrounge for less efficient ways of making more
copies of themselves. Viruses today don't all use double-stranded DNA;
some use an ancestral language, composed of single-stranded RNA (which of
course still plays a role in our own reproductive system, as an intermediary
"messenger" system during "expression"). If we follow standard practice and
reserve the term virus for a parasitic macromolecule, we need a name for
these earliest ancestors. Computer programmers call a cobbled-together
fragment of coded instructions that performs a particular task a "macro," so I
propose to call these pioneers macros, to stress that although they are "just"
huge macromolecules, they are also bits of program or
algorithm, bare, minimal self-reproducing mechanisms—remarkably like the
computer viruses that have recently emerged to fascinate and plague us (Ray
1992, Dawkins 1993)1 Since these pioneer macros reproduced, they met the
necessary Darwinian conditions for evolution, and it is now clear that they
spent the better part of a billion years evolving on Earth before there were
any living things.
Even the simplest replicating macro is far from simple, however, a com-
position with thousands or millions of parts, depending on how we count the
raw materials that go to make it. The alphabet letters Adenine, Cytosine,
Guanine, Thymine, and Uracil are bases that are not too complex to arise in
the normal course of prebiotic affairs. (RNA, which came before DNA, has
Uracil, whereas DNA has Thymine.) Expert opinion differs, however, on
whether these blocks could synthesize themselves by a series of coincidences
into something as fancy as a self-replicator. The chemist Graham Cairns-
Smith (1982, 1985) presents an updated version of Paley's argument, aimed
at the molecular level: The process of synthesizing DNA fragments, even by
the advanced methods of modern organic chemists, is highly elaborate; this
shows that their chance creation is as improbable as Paley's watch in a
windstorm. "Nucleotides are too expensive" (Cairns-Smith 1985, pp. 45-49).
DNA exhibits too much design work to be a mere product of chance, Cairns-
Smith argues, but he then proceeds to deduce an ingenious—if speculative
and controversial—account of how that work might have been done. Whether
or not Cairns-Smith's theory is eventually confirmed, it is well worth sharing
simply because it so perfectly instantiates the fundamental Darwinian
strategy.2
A good Darwinian, faced yet again with the problem of finding a needle in
a haystack of Design Space, would cast about for a still simpler form of

1. Warning: biologists already use the term macroevolution, in contrast to microevolu-
tion, to refer to large-scale evolutionary phenomena—the patterns of speciation and
extinction, for instance, in contrast to the refinement of wings or changes in resistance to
toxins within a species. What I am calling the evolution of macros has nothing much to
do with macroevolution in that established sense. The term macro is so apt for my
purposes, however, that I have decided to stick with it, and try to offset its shortcomings
with this patch—a tactic Mother Nature also often uses.
2. For just this reason, Richard Dawkins also presents a discussion and elaboration of
Cairns-Smith's ideas in TheBlind Watchmaker (1986a, pp. 148-58). Since Cairns-Smith's
1985 account and Dawkins' elaboration are such good reading for nonexperts, I will refer
you to them for the delicious details, and provide just enough summary here to whet
your appetite, adding the warning that there are problems with Cairns-Smith's hypothe-
ses, and balancing the warning with the reassurance that even if his hypotheses are all
ultimately rejected—an open question—there are other, less readily understandable,
alternatives to take seriously next.

158 PRIMING DARWIN'S PUMP Molecular Evolution 159

replicator that could somehow serve as a temporary scaffolding to hold the
protein parts or nucleotide bases in place until the whole protein or macro
could get assembled. Wondrous to say, there is a candidate with just the right
properties, and more wondrous still, it is just what the Bible ordered: clay!
Cairns-Smith shows that in addition to the carbon-based self-replicating
crystals of DNA and RNA, there are also much simpler (he calls them "low-
tech") silicon-based self-replicating crystals, and these silicates, as they are
called, could themselves be the product of an evolutionary process. They
form the ultra-fine particles of clay, of the sort that builds up just outside the
strong currents and turbulent eddies in streams, and the individual crystals
differ subtly at the level of molecular structure in ways that they pass on
when they "seed" the processes of crystallization that achieve their self-
replication.
Cairns-Smith develops intricate arguments to show how fragments of
protein and RNA, which would be naturally attracted to the surfaces of these
crystals like so many fleas, could eventually come to be used by the silicate
crystals as "tools" in furthering their own replication processes. According to
this hypothesis (which, like all really fertile ideas, has many neighboring
variations, any one of which might prove to be the eventual winner), the
building blocks of life began their careers as quasi-parasites of sorts, clinging
to replicating clay particles and growing in complexity in the furtherance of
the "needs" of the clay particles until they reached a point where they could
fend for themselves. No skyhook—just a ladder that could be thrown away,
as Wittgenstein once said in another context, once it had been climbed.
But this cannot be close to the whole story, even if it is all true. Suppose
that short self-replicating strings of RNA got created by this low-tech pro-
cess. Cairns-Smith calls these entirely self-involved replicators "naked
genes," because they aren't for anything except their own replication, which
they do without outside help. We are still left with a major problem: How did
these naked genes ever come to be clothed? How did these solipsistic self-
reproducers ever come to specify particular proteins, the tiny enzyme-
machines that build the huge bodies that carry today's genes from generation
to generation? But the problem is worse than that, for these proteins don't just
build bodies; they are needed to assist in the very process of self-replication
once a string of RNA or DNA gets long. Although short strings of RNA can
replicate themselves without enzyme assistants, longer strings need a retinue
of helpers, and specifying them requires a very long sequence—longer than
could be replicated with high-enough fidelity until those very enzymes were
already present. We seem to face paradox once again, in a vicious circle
succinctly described by John Maynard Smith: "One cannot have accurate
replication without a length of RNA of, say, 2000 base pairs, and one cannot
have that much RNA without accurate replication" (Maynard Smith 1979, p.
445).
One of the leading researchers on this period of evolutionary history is
Manfred Eigen. In his elegant little book, Steps Towards Life (1992)—a good
place to continue your exploration of these ideas—he shows how the macros
gradually built up what he calls the "molecular tool-kit" that living cells use
to re-create themselves, while also building around themselves the sorts of
structures that became, in due course, the protective membranes of the first
prokaryotic cells. This long period of precellular evolution has left no fossil
traces, but it has left plenty of clues of its history in the "texts" that have been
transmitted to us through its descendants, including, of course, the viruses
that swarm around us today. By studying the actual surviving texts, the
specific sequences of A, C, G, and T in the DNA of higher organisms and the
A, C, G, and U of their RNA counterparts, researchers can deduce a great
deal about the actual identity of the earliest self-replicating texts, using
refined versions of the same techniques the philologists used to reconstruct
the words that Plato actually wrote. Some sequences in our own DNA are
truly ancient, even traceable (via translation back into the earlier RNA
language) to sequences that were composed in the earliest days of macro
evolution!
Let's go back to the time when the nucleotide bases (A, C, G, T, and U)
were occasionally present here and there in varying amounts, possibly con-
gregated around some of Cairns-Smith's clay crystals. The twenty different
amino acids, the building blocks for all proteins, also occur with some
frequency under a wide range of nonbiotic conditions, so we can help
ourselves to them as well. Moreover, it has been shown by Sidney Fox (Fox
and Dose 1972) that individual amino acids can condense into "protein-oids,"
protein-like substances that have a very modest catalytic ability ( Eigen 1992,
p. 32). This is a small but important step up, since catalytic ability— the
capacity to facilitate a chemical reaction—is the fundamental talent of any
protein.
Now suppose some of the bases come to pair up, C with G, and A with U,
and make smallish complementary sequences of RNA—less than a hundred
pairs long—that can replicate, crudely, without enzymatic helpers. In terms
of the Library of Babel, we would now have a printing press and a book-
bindery, but the books would be too short to be good for anything except
making more of themselves, with lots of misprints. And they would not be
about anything. We may seem to be right back where we started—or even
worse. When we bottom out at the level of molecular building blocks, we
face a design problem that is more like construction out of Tinker Toy than
gradual sculpting in modeling clay. Under the rigid rules of physics, either
the atoms jump together into stable patterns or they don't.
Fortunately for us—indeed, fortunately for all living things—scattered in
the Vast space of possible proteins there happen to be protein constructions
that—if found—permit life to go forward. How might they get found? Some-
how we have to get those proteins together with the protein-hunters, the

160 PRIMING DARWIN'S PUMP Molecular Evolution 161

fragments of self-replicating nucleotide strings that will eventually come to
"specify" them in the macros they compose. Eigen shows how the vicious
circle can turn friendly if it is expanded into a "hypercycle" with more than
two elements (Eigen and Schuster 1977). This is a difficult technical concept,
but the underlying idea is clear enough: imagine a circumstance in which
fragments of type A can enhance the prospects of hunks of B, which in turn
promote the well-being of bits of C, which, completing the loop, permit the
replication of more fragments of A, and so forth, in a mutually reinforcing
community of elements, until the point is reached where the whole process
can take off, creating environments that normally serve to replicate longer
and longer strings of genetic material. (Maynard Smith 1979 is a great help in
understanding the idea of a hypercycle; see also Eigen 1983.)
But even if this is possible in principle, how could it get started? If all
possible proteins and all possible nucleotide "texts" were truly equiproba-ble,
then it would be hard to see how the process could ever get going. Somehow,
the bland, mixed-up confetti of ingredients has to get some structure imposed
on it, concentrating a few "likely-to-succeed" candidates and thereby making
them still more likely to succeed. Remember the coin-tossing tournament in
chapter 2? Somebody has to win, but the winner wins in virtue of no virtue,
but simply in virtue of historical accident. The winner is not bigger or
stronger or better than the other contestants, but is still the winner. It now
seems that something similar happens in prebiotic molecular evolution, with
a Darwinian twist: winners get to make extra copies of themselves for the
next round, so that, without any selection "for cause" (as they say when
dismissing potential jurors), dynasties of sheer replicative prowess begin to
emerge. If we start with a purely random assortment of "contestants" drawn
from the pool of self-replicating fragments, even if they are not initially
distinguishable in terms of their replicative prowess, those that happen to win
in the early rounds will occupy more of the slots in the subsequent rounds,
flooding the space with trails of highly similar (short) texts, but still leaving
vast hypervolumes of the space utterly empty and inaccessible for good. The
initial threads of proto-life can emerge before there is any difference in skill,
becoming the actuality from which the Tree of Life can then grow, thanks to
tournaments of skill. As Eigen's colleague Bernd-Olaf Küppers (1990, p.
150) puts it, "The theory predicts that biological structures exist, but not what
biological structures exist."3 This is

3. Kiippers (1990, pp. 137-46) borrows an example from Eigen (1976) to illustrate the
underlying idea: a game of "non-Darwinian selection" you can play on a checkerboard
with differently colored marbles. Start by randomly placing the marbles on all the squares,
creating the initial confetti effect. Now throw two (eight-sided!) dice to determine a
square (column 5, row 7, for instance) on which to act. Remove the marble on that
enough to build plenty of bias into the probability space from the outset.
So some of the possible macros, inevitably, are more probable—more
likely to be stumbled upon in the Vast space of possibilities—than others.
Which ones? The "fitter" ones? Not in any nontrivial sense, but just in the
tautological sense of being identical to (or nearly identical to) previous
"winners," who in turn tended to be almost identical to still earlier "winners."
(In the million-dimension Library of Mendel, sequences that differ at a single
locus are shelved "next to" each other in some dimension; the distance of any
one volume from another is technically known as the Hamming distance.
This process spreads "winners" out gradually—taking leaps of small
Hamming distances—from any initial starting point in any and all directions
in the Library.) This is the most rudimentary possible case of "the rich get
richer," and since the success of the string has an explanation with no
reference beyond the string itself and its resemblance as a string to its parent
string, this is a purely syntactic definition of fitness, as opposed to a semantic
definition of fitness (Kiippers 1990, p. 141). That is, you don't have to
consider what the string means in order to determine its fitness. We saw in
chapter 6 that mere typographical change could never explain the Design that
needs explaining, any more than you could explain the difference in quality
between two books by comparing their relative frequencies of alphabetic
characters, but before we can have the meaningful self-replicating codes that
make this possible, we have to have self-replicating codes that don't mean a
thing; their only "function" is to replicate themselves. As Eigen (1992, p. 15)
puts it, "The structural stability of the molecule has no bearing upon the
semantic information which it carries, and which is not expressed until the
product of translation appears."
This is the birth of the ultimate QWERTY phenomenon, but, like the
cultural case that gives it its name, it was not entirely without point even
from the outset. Perfect equiprobability could have dissolved into a mo-
nopoly by a purely random process, as we have just seen, but perfect
equiprobability is hard to come by in nature at any point, and at the very
beginnings of this process of text generation, a bias was present. Of the four
bases—A, C, G, and T—G and C are the most structurally stable: "Calcula-
tion of the necessary binding energies, along with experiments on binding

square. Throw the dice again; go to the square they name and check the color of the
marble on this square and put a marble of that color on the just-vacated square ( "repro-
duction" of that marble). Repeat the process, over and over. Eventually, it has the effect
of unrandomizing the initial distribution of colors, so that one color ends up "winning"
but for no reason at all—just historical luck. He calls this "non-Darwinian selection"
because it is selection in the absence of a biasing cause; selection without adaptation
would be the more familiar term. It is non-Darwinian only in the sense that Darwin didn't
see the importance of allowing for it, not in the sense that Darwin ( or Darwinism ) cannot
accommodate it. Manifestly it can.

166 PRIMING DARWIN'S PUMP
The Laws of the Game of Life 167

quite strong that this is often the case, and not just in discussions of the
Anthropic Principle. Consider the related confusions that surround Darwinian
deduction in general. Darwin deduces that human beings must have evolved
from a common ancestor of the chimpanzee, or that all life must have arisen
from a single beginning, and some people, unaccountably, take these
deductions as claims that human beings are somehow a necessary product of
evolution, or that life is a necessary feature of our planet, but nothing of the
kind follows from Darwin's deductions properly construed. What must be the
case is not that we are here, but that since we are here, we evolved from
primates. Suppose John is a bachelor. Then he must be single, right? (That's a
truth of logic.) Poor John—he can never get married! The fallacy is obvious
in this example, and it is worth keeping it in the back of your mind as a
template to compare other arguments with.
Believers in any of the proposed strong versions of the Anthropic Prin-
ciple think they can deduce something wonderful and surprising from the
fact that we conscious observers are here—for instance, that in some sense
the universe exists for us, or perhaps that we exist so that the universe as a
whole can exist, or even that God created the universe the way He did so that
we would be possible. Construed in this way, these proposals are attempts to
restore Paley's Argument from Design, readdressing it to the Design of the
universe's most general laws of physics, not the particular constructions those
laws make possible. Here, once again, Darwinian coun-termoves are
available.
These are deep waters, and most of the discussions of the issues wallow in
technicalities, but the logical force of these Darwinian responses can be
brought out vividly by considering a much simpler case. First, I must in-
troduce you to the Game of Life, a nifty meme whose principal author is the
mathematician John Horton Conway. (I will be putting this valuable thinking
tool to several more uses, as we go along. This game does an excellent job of
taking in a complicated issue and reflecting back only the dead-simple
essence or skeleton of the issue, ready to be understood and appreciated.)
Life is played on a two-dimensional grid, such as a checkerboard, using
simple counters, such as pebbles or pennies—or one could go high-tech and
play it on a computer screen. It is not a game one plays to win; if it is a game
at all, it is solitaire.4 The grid divides space into square cells, and each cell

FIGURE 7.2
is either ON or OFF at each moment. (If it is ON, place a penny on the square;
if it is OFF, leave the square empty.) Notice in figure 7.2 that each cell has
eight neighbors: the four adjacent cells—north, south, east, and west—and
the four diagonals—northeast, southeast, southwest, and northwest.
Time in the Life world is discrete, not continuous; it advances in ticks, and
the state of the world changes between each two ticks according to the
following rule:
Life Physics: For each cell in the grid, count how many of its eight neigh-
bors are ON at the present instant. If the answer is exactly two, the cell stays
in its present state ( ON or OFF ) in the next instant. If the answer is exactly
three, the cell is ON in the next instant whatever its current state. Under all
other conditions, the cell is OFF.
That's it—that's the only rule of the game. You now know all there is to
know about how to play Life. The entire physics of the Life world is captured
in that single, unexceptioned law. Although this is the fundamental law of
the "physics" of the Life world, it helps at first to conceive this curious
physics in biological terms: think of cells going ON as births, cells going OFF
as deaths, and succeeding instants as generations. Either overcrowding (more
than three inhabited neighbors) or isolation (fewer than two inhabited
neighbors) leads to death. Consider a few simple cases.
In the configuration in figure 7.3, only cells d and / each have exactly
three neighbors ON, so they will be the only birth cells in the next generation.
Cells b and h each have only one neighbor ON, SO they die in the next
generation. Cell e has two neighbors ON, SO it stays on. Thus the next
"instant" will be the configuration shown in figure 7.4.

4. This description of Life is adapted from an eariier exposition of mine (1991b). Martin
Gardner introduced the Game of Life to a wide audience in two of his "Mathematical
Games" columns in Scientific American, in October 1970 and February 1971. Pound-
stone 1985 is an excellent exploration of the game and its philosophical implications.

176 PRIMING DARWIN'S PUMP The Laws of the Game of Life 177

ical howler if they argued that since they existed, die Life world, with its
particular physics, had to exist—for after all, Conway might have decided to
be a plumber or play bridge instead of hunting for this world. But what if
they deduced that their world was just too wonderful, with its elegant, Life-
sustaining physics, to have come into existence without an Intelligent
Creator? If they jumped to the conclusion that they owed their existence to
the activities of a wise Lawgiver, they'd be right! There is a God and his
name is Conway.
But they would be jumping to a conclusion. The existence of a universe
obeying a set of laws even as elegant as the Life law (or the laws of our own
physics) does not logically require an intelligent Lawgiver. Notice first how
the actual history of the Game of Life divided the intellectual labor in two: on
the one hand there was the initial exploratory work that led to the physical
law promulgated by the Lawgiver, and on the other hand there was the
engineering work of the law-exploiters, the Artificers. These might have
happened in that temporal order—first Conway, in a stroke of inspired
genius, promulgates the physics of the Life world, and then he and his
students design and build the wonderful denizens of that world according to
the law laid down. But in fact the two tasks were intermixed; many trial-and-
error attempts to make things that were interesting provided the guidance for
Conway's legislative search. Notice, second, that this postulated division of
labor illustrates a fundamental Darwinian theme from the previous chapter.
The task of the wise God required to put this world into motion is a task of
discovery, not creation, a job for a Newton, not a Shakespeare. What Newton
found—and what Conway found—are eternal Platonic fixed points that
anybody else in principle could have discovered, not idiosyncratic creations
that depend in any way on the particularities of the minds of their authors. If
Conway had never turned his hand to designing cellular-automata worlds—if
Conway had never even existed—some other mathematician might very well
have hit upon exactly the Life world that Conway gets the credit for. So, as
we follow the Darwinian down this path, God the Artificer turns first into
God the Lawgiver, who now can be seen to merge with God the Lawfinder.
God's hypothesized contribution is thereby becoming less personal—and
hence more readily performable by something dogged and mindless!
Hume has already shown us how the argument runs, and now, bolstered by
our experience with Darwinian thinking in more familiar terrain, we can

his friends and colleagues in this quest. I have always found the prospect of such a proof
mouth-watering, but the paths to it are totally beyond me. So far as I know, nothing
substantive has yet been published on this most interesting epistemological question, but
I want to encourage others to address it. The same thought experiment is posed, inde-
pendently, in Stewart and Golubitsky 1992, pp. 261-62.
extrapolate a positive Darwinian alternative to the hypothesis that our laws
are a gift from God. What would the Darwinian alternative have to be? That
there has been an evolution of worlds (in the sense of whole universes), and
the world we find ourselves in is simply one among countless others that
have existed through eternity. There are two quite different ways of thinking
about the evolution of laws, one of them stronger, more "Darwinian," than
the other in that it involves something like natural selection.
Might it be that there has been some sort of differential reproduction of
universes, with some varieties having more "offspring" than others? Hume's
Philo toyed with this idea, as we saw in chapter 1:
And what surprise must we entertain, when we find him a stupid mechanic,
who imitated others, and copied an art, which, through a long succession
of ages, after multiplied trials, mistakes, corrections, deliberations, and
controversies, had been gradually improving? Many worlds might have
been botched and bungled, throughout an eternity, ere this system was
struck out: Much labour lost: Many fruitless trials made: And a slow, but
continued improvement carried on during infinite ages of world-making.
[Pt. V.]
Hume imputes the "continued improvement" to the minimal selective bias
of a "stupid mechanic," but we can replace the stupid mechanic with
something even stupider without dissipating the lifting power: a purely
algorithmic Darwinian process of world-trying. Though Hume obviously
didn't think this was anything but an amusing philosophical fantasy, the idea
has recently been developed in some detail by the physicist Lee Smolin
(1992). The basic idea is that the singularities known as black holes are in
effect the birthplaces of offspring universes, in which the fundamental phys-
ical constants would differ slightly, in random ways, from the physical con-
stants in the parent universe. So, according to Smolin's hypothesis, we have
both differential reproduction and mutation, the two essential features of any
Darwinian selection algorithm. Those universes that just happened to have
physical constants that encouraged the development of black holes would
ipso facto have more offspring, which would have more offspring, and so
forth—that's the selection step. Note that there is no grim reaper of universes
in this scenario; they all live and "die" in due course, but some merely have
more offspring. According to this idea, then, it is no mere interesting
coincidence that we live in a universe in which there are black holes, nor is it
an absolute logical necessity. It is, rather, the sort of conditional near-
necessity you find in any evolutionary account. The link, Smolin claims, is
carbon, which plays a role both in the collapse of gaseous clouds (or in other
words, the birth of stars, a precursor to the birth of black holes) and, of
course, in our molecular engineering.

180 PRIMING DARWIN'S PUMP
consecutive coin-tosses without a loss!" You then arrange for your dupes to
meet, pairwise, until you have a final winner. (You never let the contestants
discuss your relation to them, and you kiss off the 1,02 3 losers along the way
with some sotto voce gibe to the effect that they were pretty gullible to
believe your claim about being Mephistopheles!) The winner—and there
must be one—will certainly have been given evidence of being a Chosen
One, but if he falls for it, this is simply an illusion of what we might call
retrospective myopia. The winner doesn't see that the situation was struc-
tured so that somebody simply had to be the lucky one—and he just hap-
pened to be it.
Now if the universe were structured in such a way that an infinity of
different "laws of physics" got tried out in the fullness of time, we would be
succumbing to the same temptation were we to draw any conclusions about
the laws of nature being prepared especially for us. This is not an argument
for the conclusion that the universe is, or must be, so structured, but just for
the more modest conclusion that no feature of the observable "laws of
nature" could be invulnerable to this alternative, deflationary interpretation.
Once these ever more speculative, ever more attenuated Darwinian hy-
potheses are formulated, they serve—in classic Darwinian fashion—to di-
minish by small steps the explanatory task facing us. All that is left over in
need of explanation at this point is a certain perceived elegance or won-
derfulness in the observed laws of physics. If you doubt that the hypothesis
of an infinity of variant universes could actually explain this elegance, you
should reflect that this has at least as much claim to being a non-question-
begging explanation as any traditional alternative; by the time God has been
depersonalized to the point of being some abstract and timeless principle of
beauty or goodness, it is hard to see how the existence of God could explain
anything. What would be asserted by the "explanation" that was not already
given in the description of the wonderful phenomenon to be explained?
Darwin began his attack on the Cosmic Pyramid in the middle: Give me
Order, and time, and I will explain Design. We have now seen how the
downward path of universal acid flows: if we give his successors Chaos (in
the old-fashioned sense of pure meaningless randomness), and eternity, they
will explain Order—the very Order needed to account for the Design. Does
utter Chaos in turn need an explanation? What is there left to explain? Some
people think there is still one leftover "why" question: Why is there
something rather than nothing? Opinions differ on whether the question
makes any intelligible demand at all.10 If it does, the answer "Because God
Eternal Recurrence—Life Without Foundations? 181
exists" is probably as good an answer as any, but look at its competition:
"Why not?"
4. ETERNAL RECURRENCE—LIFE WITHOUT FOUNDATIONS?
Science is wonderful at destroying metaphysical answers, but incapable
of providing substitute ones. Science takes away foundations without
providing a replacement. Whether we want to be there or not, science
has put us in a position of having to live without foundations. It was
shocking when Nietzsche said this, but today it is commonplace; our
historical position—and no end to it is in sight—is that of having to
philosophize without 'foundations'.
—HIIARY PUTNAM 1987, p. 29
The sense that the meaning of the universe had evaporated was what
seemed to escape those who welcomed Darwin as a benefactor of
mankind. Nietzsche considered that evolution presented a correct pic-
ture of the world, but that it was a disastrous picture. His philosophy
was an attempt to produce a new world-picture which took Darwinism
into account but was not nullified by it.
—R. J. HOLLINGDALE 1965, p. 90
In the wake of Darwin's publication of Origin of Species, Friedrich
Nietzsche rediscovered what Hume had already toyed with: the idea that an
eternal recurrence of blind, meaningless variation—chaotic, pointless shuf-
fling of matter and law—would inevitably spew up worlds whose evolution
through time would yield the apparently meaningful stories of our lives. This
idea of eternal recurrence became a cornerstone of his nihilism, and thus part
of the foundation of what became existentialism.
The idea that what is happening now has all happened before must be as
old as the dejd-vu phenomenon that so often inspires superstitious versions of
it. Cyclical cosmogonies are not uncommon in the catalogue of human
cultures. But when Nietzsche hit upon a version of Hume's—and John
Archibald Wheeler's—vision, he took it to be much more than an amusing
thought experiment or an elaboration of ancient superstitions. He thought—at
least for a while—he had stumbled upon a scientific proof of


10. For an engaging examination of the question, see ch. 2 of Robert Nozick's Philosoph-
ical Explanation. Nozick offers several different candidate answers, all of them admit-
tedly bizarre, but notes, disarmingly. "The question cuts so deep, however, that any
approach that stands a chance of yielding an answer will look extremely weird. Someone
who proposes a non-strange answer shows he didn't understand the question" (Nozick
1981, p. 116).

196 BIOLOGY IS ENGINEERING Function and Specification 197

acid sequences can be used as "philological" clues in re-creating the evo-
lutionary history of the production and use of lysozyme.
And here is a puzzle, first noted by Walter Elsasser (1958, 1966), but quite
conclusively solved by Jacques Monod (1971). Considered very abstractly,
the fact that a one-dimensional code can be "for" a three-dimensional
structure shows that information is added. Indeed, value is added. The
individual amino acids have value (by contributing to the functional prowess
of a protein) not just in virtue of their location in the one-dimensional
sequence that forms the string, but in virtue of their location in three-
dimensional space once the string is folded up.
Thus there is a seeming contradiction between the statement that the
genome 'entirely defines' the function of a protein and the fact that this
function is linked to a three-dimensional structure whose data content is
richer than the direct contribution made to the structure by the genome.
[Monod 1971, p. 94.]
As Küppers (1990, p. 120) points out, Monod's solution is straightforward:
"The seemingly irreducible, or excess, information is contained in the
specific conditions of the protein's environment, and only together with these
can the genetic information determine unambiguously the structure and thus
the function of the protein molecule." Monod (1971, p. 94) puts it this way:
... of all the structures possible only one is actually realized. Initial con-
ditions hence enter among the items of information finally enclosed within
the ... structure. Without specifying it, they contribute to the realization
of a unique shape by eliminating all alternative structures, in this way
proposing—or rather imposing—an unequivocal interpretation of a poten-
tially equivocal message.2
What does this mean? It means—not surprisingly—that the language of
DNA and the "readers" of that language have to evolve together; neither can
work on its own. When the deconstructionists say that the reader brings
something to the text, they are saying something that applies just as surely to
DNA as to poetry; the something that the reader brings can be charac-

2. Philosophers will recognize, I trust, that Monod thus both posed and solved
Putnam's (1975) problem of Twin Earth, at least in the context of the "toy problem" of
molecular evolution. Meaning "ain't in the head," as Putnam famously observed, and it
ain't (all) in the DNA either. Twin Earth, otherwise known as the problem of broad
versus narrow content, will get exhumed briefly in chapter 14, so I can give it its proper
Darwinian funeral.
terized most generally and abstracdy as information, and only the combi-
nation of information from the code and the code-reading environment
suffices to create an organism.3 As we noted in chapter 5, some critics have
fastened on this fact as if it were somehow the refutation of "gene centrism,"
the doctrine that the DNA is the sole information store for inheritance, but
that idea was always only a handy oversimplification. Though libraries are
commonly allowed to be storehouses of information, of course it is really
only libraies-plus readers that preserve and store the information. Since
libraries have not—up till now, at any rate—contained among their volumes
the information needed to create more readers, their capacity to store
information (effectively) has been dependent on there being another
information-storage system—the human genetic system, of which DNA is the
principle medium. When we apply the same reasoning to DNA itself, we see
that it, too, requires a continuing supply of "readers" that it does not itself
entirely specify. Where does the rest of the information come from to specify
these readers? The short answer is that it comes from the very continuities of
the environment—the persistence in the environment of the necessary raw
(and partially constructed) materials, and the conditions in which they can be
exploited. Every time you make sure that your dishrag gets properly dry in
between uses, you break the chain of environmental continuity (e.g., lots of
moisture) that is part of the informational background presupposed by the
DNA of the bacteria in the dishrag whose demise you seek.
We see here a special case of a very general principle: any functioning
structure carries implicit information about the environment in which its
function "works." The wings of a seagull magnificently embody principles of
aerodynamic design, and thereby also imply that the creature whose wings
these are is excellently adapted for flight in a medium having the specific
density and viscosity of the atmosphere within a thousand meters or so of the
surface of the Earth. Recall the example in chapter 5 of sending the score of
Beethoven's Fifth Symphony to "Martians." Suppose we carefully preserved
the body of a seagull and sent it off into space (without any accompanying
explanation), to be discovered by these Martians. If they

3. David Haig (personal communication) has drawn my attention to a fascinating new
wrinkle in this unfolding story about folding proteins: molecular chaperones. "Chaper-
ones are molecular cranes par excellence. They are proteins with which an amino acid
chain associates while it is folding that allows the chain to adopt a conformation that
would be unavailable in the absence of the chaperone. The chaperone is then discarded
by the folded protein. Chaperones are highly conserved.... Molecular chaperones were
named by analogy to the functions of chaperones at a debutante ball: their role was to
encourage some interactions and to discourage others." For recent details, see Martin et
al. 1993, Ellis and van der Vies 1991.

198 BIOLOGY IS ENGINEERING Function and Specification 199

made the fundamental assumption that the wings were functional, and that
their function was flight (which might not be as obvious to them as we, who
have seen them do it, think), they could use this assumption to "read off' the
implicit information about an environment for which these wings would be
well designed. Suppose they then asked themselves how all this aerodynamic
theory came to be implicit in the structure, or, in other words: How did all
this information get into these wings? The answer must be: By an interaction
between the environment and the seagull's ancestors. ( Dawkins 1983a
explores these issues in more detail.)
The same principle applies at the most basic level, where the function is
specification itself, the function on which all other functions depend. When
we wonder, with Monod, how the three-dimensional shape of the proteins
gets fixed, given that the information in the genome must underspecify them,
we see that only a pruning of the nonfunctional (or less functional) could
explain it. So the acquisition of a particular shape by a molecule involves a
mixture of historical accident on the one hand and the "discovery" of
important truths on the other.
From the outset, the process of the design of molecular "machines" ex-
hibits these two features of human engineering. Eigen (1992, p. 34) provides
a good instance of this in his reflections on the structure of the DNA code.
"One might well ask why Nature has used four symbols, when she might just
as well have made do with two." Why indeed? Notice how naturally and
inevitably a "why" question arises at this point, and notice that it calls for an
"engineering" answer. Either the answer is that there is no reason—it is
historical accident, pure and simple—or there is a reason: a condition was or
is present that makes this the right way or best way for the coding system to
get designed, given the conditions that obtained.4
All the deepest features of molecular design may be considered from the
engineering perspective. On the one hand, consider the fact that macro-
molecules come in two basic shape categories: symmetrical and chiral (with
left-handed and right-handed versions). There is a reason why so many
should be symmetrical:
The selective advantage in a symmetrical complex is enjoyed by all the
subunits, while in an asymmetric complex the advantage is only effective
in the subunit in which the mutation arises. It is for this reason that we find
so many symmetric structures in biology, "because they were able to make
the most effective use of their advantage, and thus—a posteriori—won the
selection competition; this was not, however, because symmetry is—a priori—an
indispensable requirement for the fulfillment of a functional purpose." [Küppers
1990, p. 119, incorporating a quotation from Eigen and Winkler-Oswatitsch
1975.]
But what about the asymmetric or chiral shapes? Is there a reason why
they should be one way—left-handed, say—rather than the other? No, prob-
ably not, but: "Even if there is no a priori physical explanation for the
decision, even if it was just a brief fluctuation that gave one or the other
equivalent possibility a momentary advantage, the self-reinforcing character
of selection would turn the random decision into a major and permanent
breach of symmetry. The cause would be a purely 'historical' one" (Eigen
1992, p. 35 ).5
The shared chirality of organic molecules (in our part of the universe ) was
thus probably another pure QWERTY phenomenon, or what Crick (1968) has
called a "frozen accident." But even in the case of such a QWERTY
phenomenon, if the conditions are just right and the opportunities and hence
pressures are great enough, the tables might be turned and a new standard
established. This is apparently just what happened when the DNA language
displaced the RNA language as the lingua franca of encoding for complex
organisms. The reasons for its preferability are clear: by being double-
stranded, the DNA language permitted a system of error-correcting or
proofreading enzymes, which could repair copying errors in one strand by
reference to its mate. This made the creation of longer, more complicated
genomes feasible (Eigen 1992, p. 36).
Note that this reasoning does not yield the conclusion that double-stranded
DNA must develop, for Mother Nature had no advance intention to create
multicellular life. It just reveals that // double-stranded DNA happens to
begin to develop, it opens up opportunities that are dependent on it. Hence it
becomes a necessity for those exemplars in the space of all possible life
forms that avail themselves of it, and if those life forms prevail

5. Danny Hillis, the creator of the Connection Machine, once told me a story about some
computer scientists who designed an electronic component for a military application (I
think it was part of a guidance system in airplanes). Their prototype had two circuit
boards, and the top one kept sagging, so, casting about for a quick fix, they spotted a brass
doorknob in the lab which had just the right thickness. They took it off its door and
jammed it into place between the two circuit boards on the prototype. Sometime later,
one of these engineers was called in to look at a problem the military was having with the
actual manufactured systems, and found to his amazement that between the circuit
boards in each unit was a very precisely milled brass duplicate of the original doorknob.
This is an Ur-story that has many well-known variations in engineering circles and among
evolutionary biologists. For instance, see Primo Levi's amusing account of the mystery of
the varnish additive in The Periodic Table ( 1984).

4. Eigen suggests that there is a reason why there are four letters, not two, but I am not
going to pass it on. Perhaps you can figure out for yourself what it might be before seeing
what Eigen says. You already have at your fingertips the relevant principles of engineering
to give it a good shot.