The Improbability of God
by Richard Dawkins
The following article is from Free Inquiry MagazineVolume 18, Number 3.
Much of what people do is done in the name of God. Irishmen blow each other up in his
name. Arabs blow themselves up in his name. Imams and ayatollahs oppress women in
his name. Celibate popes and priests mess up people's sex lives in his name. Jewish
shohets cut live animals' throats in his name. The achievements of religion in past
history - bloody crusades, torturing inquisitions, mass-murdering conquistadors, culture-
destroying missionaries, legally enforced resistance to each new piece of scientific truth
until the last possible moment - are even more impressive. And what has it all been in
aid of? I believe it is becoming increasingly clear that the answer is absolutely nothing at
all. There is no reason for believing that any sort of gods exist and quite good reason for
believing that they do not exist and never have. It has all been a gigantic waste of time
and a waste of life. It would be a joke of cosmic proportions if it weren't so tragic.
Why do people believe in God? For most people the answer is still some version of the
ancient Argument from Design. We look about us at the beauty and intricacy of the world
- at the aerodynamic sweep of a swallow's wing, at the delicacy of flowers and of the
butterflies that fertilize them, through a microscope at the teeming life in every drop of
pond water, through a telescope at the crown of a giant redwood tree. We reflect on the
electronic complexity and optical perfection of our own eyes that do the looking. If we
have any imagination, these things drive us to a sense of awe and reverence. Moreover,
we cannot fail to be struck by the obvious resemblance of living organs to the carefully
planned designs of human engineers. The argument was most famously expressed in
the watchmaker analogy of the eighteenth-century priest William Paley. Even if you
didn't know what a watch was, the obviously designed character of its cogs and springs
and of how they mesh together for a purpose would force you to conclude "that the
watch must have had a maker: that there must have existed, at some time, and at some
place or other, an artificer or artificers, who formed it for the purpose which we find it
actually to answer; who comprehended its construction, and designed its use." If this is
true of a comparatively simple watch, how much the more so is it true of the eye, ear,
kidney, elbow joint, brain? These beautiful, complex, intricate, and obviously purpose-
built structures must have had their own designer, their own watchmaker - God.
So ran Paley's argument, and it is an argument that nearly all thoughtful and sensitive
people discover for themselves at some stage in their childhood. Throughout most of
history it must have seemed utterly convincing, self-evidently true. And yet, as the result
of one of the most astonishing intellectual revolutions in history, we now know that it is
wrong, or at least superfluous. We now know that the order and apparent
purposefulness of the living world has come about through an entirely different process,
a process that works without the need for any designer and one that is a consequence
of basically very simple laws of physics. This is the process of evolution by natural
selection, discovered by Charles Darwin and, independently, by Alfred Russel Wallace.
What do all objects that look as if they must have had a designer have in common? The
answer is statistical improbability. If we find a transparent pebble washed into the shape
of a crude lens by the sea, we do not conclude that it must have been designed by an
optician: the unaided laws of physics are capable of achieving this result; it is not too
improbable to have just "happened." But if we find an elaborate compound lens,
carefully corrected against spherical and chromatic aberration, coated against glare, and
with "Carl Zeiss" engraved on the rim, we know that it could not have just happened by
chance. If you take all the atoms of such a compound lens and throw them together at
random under the jostling influence of the ordinary laws of physics in nature, it is
theoretically possible that, by sheer luck, the atoms would just happen to fall into the
pattern of a Zeiss compound lens, and even that the atoms round the rim should happen
to fall in such a way that the name Carl Zeiss is etched out. But the number of other
ways in which the atoms could, with equal likelihood, have fallen, is so hugely, vastly,
immeasurably greater that we can completely discount the chance hypothesis. Chance
is out of the question as an explanation.
This is not a circular argument, by the way. It might seem to be circular because, it could
be said, any particular arrangement of atoms is, with hindsight, very improbable. As has
been said before, when a ball lands on a particular blade of grass on the golf course, it
would be foolish to exclaim: "Out of all the billions of blades of grass that it could have
fallen on, the ball actually fell on this one. How amazingly, miraculously improbable!"
The fallacy here, of course, is that the ball had to land somewhere. We can only stand
amazed at the improbability of the actual event if we specify it a priori: for example, if a
blindfolded man spins himself round on the tee, hits the ball at random, and achieves a
hole in one. That would be truly amazing, because the target destination of the ball is
specified in advance.
Of all the trillions of different ways of putting together the atoms of a telescope, only a
minority would actually work in some useful way. Only a tiny minority would have Carl
Zeiss engraved on them, or, indeed, any recognizable words of any human language.
The same goes for the parts of a watch: of all the billions of possible ways of putting
them together, only a tiny minority will tell the time or do anything useful. And of course
the same goes, a fortiori, for the parts of a living body. Of all the trillions of trillions of
ways of putting together the parts of a body, only an infinitesimal minority would live,
seek food, eat, and reproduce. True, there are many different ways of being alive - at
least ten million different ways if we count the number of distinct species alive today -
but, however many ways there may be of being alive, it is certain that there are vastly
more ways of being dead!
We can safely conclude that living bodies are billions of times too complicated - too
statistically improbable - to have come into being by sheer chance. How, then, did they
come into being? The answer is that chance enters into the story, but not a single,
monolithic act of chance. Instead, a whole series of tiny chance steps, each one small
enough to be a believable product of its predecessor, occurred one after the other in
sequence. These small steps of chance are caused by genetic mutations, random
changes - mistakes really - in the genetic material. They give rise to changes in the
existing bodily structure. Most of these changes are deleterious and lead to death. A
minority of them turn out to be slight improvements, leading to increased survival and
reproduction. By this process of natural selection, those random changes that turn out to
be beneficial eventually spread through the species and become the norm. The stage is
now set for the next small change in the evolutionary process. After, say, a thousand of
these small changes in series, each change providing the basis for the next, the end
result has become, by a process of accumulation, far too complex to have come about in
a single act of chance.
For instance, it is theoretically possible for an eye to spring into being, in a single lucky
step, from nothing: from bare skin, let's say. It is theoretically possible in the sense that a
recipe could be written out in the form of a large number of mutations. If all these
mutations happened simultaneously, a complete eye could, indeed, spring from nothing.
But although it is theoretically possible, it is in practice inconceivable. The quantity of
luck involved is much too large. The "correct" recipe involves changes in a huge number
of genes simultaneously. The correct recipe is one particular combination of changes out
of trillions of equally probable combinations of chances. We can certainly rule out such a
miraculous coincidence. But it is perfectly plausible that the modern eye could have
sprung from something almost the same as the modern eye but not quite: a very slightly
less elaborate eye. By the same argument, this slightly less elaborate eye sprang from a
slightly less elaborate eye still, and so on. If you assume a sufficiently large number of
sufficiently small differences between each evolutionary stage and its predecessor, you
are bound to be able to derive a full, complex, working eye from bare skin. How many
intermediate stages are we allowed to postulate? That depends on how much time we
have to play with. Has there been enough time for eyes to evolve by little steps from
The fossils tell us that life has been evolving on Earth for more than 3,000 million years.
It is almost impossible for the human mind to grasp such an immensity of time. We,
naturally and mercifully, tend to see our own expected lifetime as a fairly long time, but
we can't expect to live even one century. It is 2,000 years since Jesus lived, a time span
long enough to blur the distinction between history and myth. Can you imagine a million
such periods laid end to end? Suppose we wanted to write the whole history on a single
long scroll. If we crammed all of Common Era history into one metre of scroll, how long
would the pre-Common Era part of the scroll, back to the start of evolution, be? The
answer is that the pre-Common Era part of the scroll would stretch from Milan to
Moscow. Think of the implications of this for the quantity of evolutionary change that can
be accommodated. All the domestic breeds of dogs - Pekingeses, poodles, spaniels,
Saint Bernards, and Chihuahuas - have come from wolves in a time span measured in
hundreds or at the most thousands of years: no more than two meters along the road
from Milan to Moscow. Think of the quantity of change involved in going from a wolf to a
Pekingese; now multiply that quantity of change by a million. When you look at it like
that, it becomes easy to believe that an eye could have evolved from no eye by small
It remains necessary to satisfy ourselves that every one of the intermediates on the
evolutionary route, say from bare skin to a modern eye, would have been favored by
natural selection; would have been an improvement over its predecessor in the
sequence or at least would have survived. It is no good proving to ourselves that there is
theoretically a chain of almost perceptibly different intermediates leading to an eye if
many of those intermediates would have died. It is sometimes argued that the parts of
an eye have to be all there together or the eye won't work at all. Half an eye, the
argument runs, is no better than no eye at all. You can't fly with half a wing; you can't
hear with half an ear. Therefore there can't have been a series of step-by-step
intermediates leading up to a modern eye, wing, or ear.
This type of argument is so naive that one can only wonder at the subconscious motives
for wanting to believe it. It is obviously not true that half an eye is useless. Cataract
sufferers who have had their lenses surgically removed cannot see very well without
glasses, but they are still much better off than people with no eyes at all. Without a lens
you can't focus a detailed image, but you can avoid bumping into obstacles and you
could detect the looming shadow of a predator.
As for the argument that you can't fly with only half a wing, it is disproved by large
numbers of very successful gliding animals, including mammals of many different kinds,
lizards, frogs, snakes, and squids. Many different kinds of tree-dwelling animals have
flaps of skin between their joints that really are fractional wings. If you fall out of a tree,
any skin flap or flattening of the body that increases your surface area can save your life.
And, however small or large your flaps may be, there must always be a critical height
such that, if you fall from a tree of that height, your life would have been saved by just a
little bit more surface area. Then, when your descendants have evolved that extra
surface area, their lives would be saved by just a bit more still if they fell from trees of a
slightly greater height. And so on by insensibly graded steps until, hundreds of
generations later, we arrive at full wings.
Eyes and wings cannot spring into existence in a single step. That would be like having
the almost infinite luck to hit upon the combination number that opens a large bank vault.
But if you spun the dials of the lock at random, and every time you got a little bit closer to
the lucky number the vault door creaked open another chink, you would soon have the
door open! Essentially, that is the secret of how evolution by natural selection achieves
what once seemed impossible. Things that cannot plausibly be derived from very
different predecessors can plausibly be derived from only slightly different predecessors.
Provided only that there is a sufficiently long series of such slightly different
predecessors, you can derive anything from anything else.
Evolution, then, is theoretically capable of doing the job that, once upon a time, seemed
to be the prerogative of God. But is there any evidence that evolution actually has
happened? The answer is yes; the evidence is overwhelming. Millions of fossils are
found in exactly the places and at exactly the depths that we should expect if evolution
had happened. Not a single fossil has ever been found in any place where the evolution
theory would not have expected it, although this could very easily have happened: a
fossil mammal in rocks so old that fishes have not yet arrived, for instance, would be
enough to disprove the evolution theory.
The patterns of distribution of living animals and plants on the continents and islands of
the world is exactly what would be expected if they had evolved from common ancestors
by slow, gradual degrees. The patterns of resemblance among animals and plants is
exactly what we should expect if some were close cousins, and others more distant
cousins to each other. The fact that the genetic code is the same in all living creatures
overwhelmingly suggests that all are descended from one single ancestor. The evidence
for evolution is so compelling that the only way to save the creation theory is to assume
that God deliberately planted enormous quantities of evidence to make it look as if
evolution had happened. In other words, the fossils, the geographical distribution of
animals, and so on, are all one gigantic confidence trick. Does anybody want to worship
a God capable of such trickery? It is surely far more reverent, as well as more
scientifically sensible, to take the evidence at face value. All living creatures are cousins
of one another, descended from one remote ancestor that lived more than 3,000 million
The Argument from Design, then, has been destroyed as a reason for believing in a
God. Are there any other arguments? Some people believe in God because of what
appears to them to be an inner revelation. Such revelations are not always edifying but
they undoubtedly feel real to the individual concerned. Many inhabitants of lunatic
asylums have an unshakable inner faith that they are Napoleon or, indeed, God himself.
There is no doubting the power of such convictions for those that have them, but this is
no reason for the rest of us to believe them. Indeed, since such beliefs are mutually
contradictory, we can't believe them all.
There is a little more that needs to be said. Evolution by natural selection explains a lot,
but it couldn't start from nothing. It couldn't have started until there was some kind of
rudimentary reproduction and heredity. Modern heredity is based on the DNA code,
which is itself too complicated to have sprung spontaneously into being by a single act of
chance. This seems to mean that there must have been some earlier hereditary system,
now disappeared, which was simple enough to have arisen by chance and the laws of
chemistry and which provided the medium in which a primitive form of cumulative natural
selection could get started. DNA was a later product of this earlier cumulative selection.
Before this original kind of natural selection, there was a period when complex chemical
compounds were built up from simpler ones and before that a period when the chemical
elements were built up from simpler elements, following the well-understood laws of
physics. Before that, everything was ultimately built up from pure hydrogen in the
immediate aftermath of the big bang, which initiated the universe.
There is a temptation to argue that, although God may not be needed to explain the
evolution of complex order once the universe, with its fundamental laws of physics, had
begun, we do need a God to explain the origin of all things. This idea doesn't leave God
with very much to do: just set off the big bang, then sit back and wait for everything to
happen. The physical chemist Peter Atkins, in his beautifully written book The Creation,
postulates a lazy God who strove to do as little as possible in order to initiate everything.
Atkins explains how each step in the history of the universe followed, by simple physical
law, from its predecessor. He thus pares down the amount of work that the lazy creator
would need to do and eventually concludes that he would in fact have needed to do
nothing at all!
The details of the early phase of the universe belong to the realm of physics, whereas I
am a biologist, more concerned with the later phases of the evolution of complexity. For
me, the important point is that, even if the physicist needs to postulate an irreducible
minimum that had to be present in the beginning, in order for the universe to get started,
that irreducible minimum is certainly extremely simple. By definition, explanations that
build on simple premises are more plausible and more satisfying than explanations that
have to postulate complex and statistically improbable beginnings. And you can't get
much more complex than an Almighty God!