Tip:
Highlight text to annotate it
X
Today's lecture is about the problem of design. So the obvious place to begin is with things
that are clearly not designed. This is just a plain, ordinary stone.
The laws of physics, left on their own, will produce something like that.
They can also produce something like this, which looks superficially like a boot, but
the resemblance is purely accidental, it means absolutely nothing.
Nor does the resemblance of this to a fish mean anything,
nor the resemblance of that to a duck's head. This is slightly more interesting but it is
still purely fortuitous. It looks like an egg and inside is a little
dummy embryo, but again, it is pure luck that is just produced
by physics alone. That is also true of these rather beautiful
looking crystals. But this is rather more interesting because
crystals are what you get when atoms, all of the same kind, are allowed
to stack up together in a way that they "want to do".
This is a different kind of crystal, a desert rose.
Almost looks as if it might have been made by a jeweler.
But all these objects are fortuitous. None of them is designed.
All the stones belong in the category I am going to call "simple."
The same is true of clouds and stars. Nobody designed them,
they came to be the way they are by the simple consequences of unaided laws of physics.
They're examples of the way things just happen to be.
Now we are ready to look at some objects that really are designed.
This microscope, nobody could possibly mistake for an object which just happens to be the
way it is. Everything about it has "design" written all
over it. It has a long tube to look down, a lens this
end, another lens that end, a mirror to reflect light up through the tube,
knobs to change the focus, other knobs to move the slide from side to
side and back and forth. Even the knobs themselves are roughened to
make them easier to grip. A designed object.
The same is true of this calculating machine, the same is true of this watch.
There are some slightly more difficult cases. These flint arrowheads.
There is not much doubt that they are designed. They are shaped in a way you would'nt normally
expect a stone on the beach to be shaped. This one is a bit more doubtful.
Experts tell us, archaeologists tell us, that that is a designed object,
that some primitive people did indeed shape that for a purpose.
I'm prepared to believe them. But that is a slightly more difficult case.
Never mind about them. There are plenty of objects which are absolutely,
obviously designed. What do they all have in common, these designed
objects? They are all good for some purpose,
and they couldn't have come to be the way they are by luck.
The microscope is obviously very good at its purpose of greatly magnifying objects
and most certainly it couldn't have come about by luck.
If you take a lot of atoms and shake them up at random then you may get a crystal
but you will not in a billion billion billion years get a microscope.
This is a guzunder, so called because it goes under,
and it clearly has a more humble purpose that the microscope,
but it works very well for its purpose, it's clearly designed.
Once we realize what its purpose is, which is to hold water,
we can come up with crude measure of how good it is.
We can measure, we can say the cost of the pot is the amount of clay that goes into it
and the benefit of it is amount of water that it holds.
And so its efficiency is the ratio of the weight of water that it holds
to the weight of clay that goes into making it.
We compare it with this pot which is not made by man but is a natural stone.
It also would hold water but it would not hold very much water for
the amount of stone that goes into it. Its efficiency ratio is not very high and
indeed it is not a designed object. It is a simple object which just happened
to be. So we have divided these objects into those
that are designed and those that I'm calling "simple."
But now I want to introduce a large and very important category of objects
that are certainly not simple and I shall argue that they are not designed.
But they look overwhelmingly and compellingly as though they were designed.
And I'm going to call them "designoid" objects. Designoid objects look designed but they actually
got their designed look from a very different process which we will
come onto later. You may find it hard at this stage to believe
that designoid objects are not designed but just wait.
So, let's have our first designoid object. This is Andrew, isn't it? And what's the snake
called? Squeeze. Squeeze. The snake is called Squeeze.
It's a boa constrictor and it is a magnificent example of a designoid object.
It looks as though it has been beautifully designed for a purpose.
One of those purposes is swallowing prey which look very much too large for it to swallow.
And one of the ways in which it achieves this is by the head -
The bones of the head, the bones of the skull are capable of detaching,
coming apart, under the skin, of course, so that the head swells to a huge size relatively
to what it starts. And there is a great gaping maw
which is capable then of swallowing very much larger prey than you'd think.
The skin here is a beautiful mottled color which you can imagine would be very, very
highly camouflaged in a forest. The snake has lost its legs. Losing legs is
a very common thing among reptiles. There are many lines of evolution of reptiles
which have lost their legs. What the boa constrictor is best at is throttling
its prey and I think both Andrew and Squeeze deserve
a round of applause. Thank you very much.
Just looking at the outside of Squeeze gives us no real idea at all
of what an extraordinarily complicated structure he and all other living things are.
A living thing like Squeeze is not just more complicated than the microscope.
It is billions of times more complicated than the microscope.
Let's come back to pots. We have seen a designed pot and we have seen
a simple, accidental pot. Now, here is a designoid pot.
This is a pitcher plant. Here you see the individual pitchers. There
is an enlargement. They are filled with water and they are traps
for insects. Insects fall into the trap and drown.
This is how a pitcher begins, this is the next stage in development.
It begins as a leaf. You see the hole just beginning to form.
This is a young pitcher and then we are going to see the full pitcher.
There is a fly climbing up it. There is a slippery area at the top.
The fly is down, into the water. It's now going to drown and in due course
its products will be digested by the plant. Here is a live pitcher plant. Here is one
of its pitchers. You can see the little lid over it.
There is the slip zone there, and inside is the water.
It appears to be a well designed object, a well designed pot.
If we were to do our cost/benefit ratio, our measure of the amount of water that it
holds over the weight of the plant material itself,
we would find that it is a very efficient pot indeed.
But we would also find that if we looked inside, cut sections of it,
it would have a very complicated structure. This is the inside of a single cell of a pitcher
plant through an electron microscope. And you can see the complexity of it.
What is more interesting is that this internal structure is very well fitted
to make a lot of oxygen and secrete it into the water of the pitcher.
And this has a very useful effect because in the water of the pitcher
are living a motley crew of little maggots and other insects.
Now, what are they doing? Well, it is all very well eating insects,
like a pitcher plant does, but a plant doesn't have any teeth and it
is difficult to eat insects if you haven't got teeth.
So what the pitcher plant in effect does is to borrow the teeth of these maggots in its
pitcher. For what the maggots do is they eat the prey
that fall into the pitcher and then their excretory products are what
are finally absorbed by the pitcher plant. So really the pitcher plant is just getting
the same thing as any other plant gets when it eats manure, in effect.
But what the pitcher plant is doing is sequestering for itself a private supply of manure
by luring insects and by supplying the maggots that live in the pot
with a nice, oxygen rich, fresh water atmosphere which otherwise they wouldn't like to live
in. Here is another sort of pot, a designoid pot.
This is made by a trapdoor spider, you see the trapdoor at the top.
The spider lives inside. And another one here. This is the pot of a
potter wasp. That's a solitary wasp, not like the social
wasps which build combs, more like this honey comb here.
The potter wasp, female, builds a pot like that out of mud and lays her eggs in it
and then the larvae grows up inside the pot. Look how beautifully it resembles this designed
pot. This is a truly designed pot, made by somebody
in Mexico. See how similar it is to the pot made by the
potter wasp. Yet another animal pot.
This was made by a mason bee. Exquisite little thing.
It is used for the same purpose as a potter wasp's pot
but it has a different structure. It is just like a house built by humans.
These are little individual stones which the bee, the female bee, has gathered
and has cemented together to build up this delightful pot structure.
And the story doesn't end there because we can only see one pot,
but underneath here are four more pots. And they've been carefully covered up by the
bee who has gone to the river and gathered clay
to cover over her pots. The clay is exactly the same color and texture
as the rock on which the pot is placed, so that bee's predators, the predators who
might come and eat larvae out of the pot, would never know in a million years that there
were any pots under there. Here is another beautiful example of designoid
architecture. These gigantic megaliths, like Avebury or
Stonehenge, are built by the compass termite of Australia.
They are huge structures, like blocks of flats. Certainly, on termite's scale, they are like
blocks of flats. They are all pointing exactly north-south,
which is another very cunning feature because that means that they get the morning
sun on one and they get the evening sun on the other,
so they get nicely warmed up in the cool parts of the day.
But in the hot part of the day, the midday sun hits them end on,
and so they don't heat up too much, which is why all these termite nests, they're
called compass termites, when you are in the desert, you can always
tell which direction is north-south by looking
for a compass termites' nest. Even larger is this other kind of termites'
nest. You can see the scale there. This is a most
colossal structure. The Austrian ethologist, Karl von Frisch,
remarked that if humans built structures on the same
scale as termites do the structures would be four times as high
as the Empire State Building. So termites are very, very impressive architects.
These designoid objects are very impressive indeed.
We are switching now from objects which are apparently designed by animals
to the design of animals themselves. The apparent design of animals themselves.
And I am beginning with camouflage. If you are walking through the desert you
will probably think, to casual look, that that was a stone. But it's not a stone,
it's a grasshopper. It just looks like a stone and it gets protection
from looking like a stone. And the next example. This looks to me exactly
like seaweed. It is one of my favorite of all designoid
objects. It is, in fact, a fish. It's a seahorse. There's its head, there is
its neck, there's its body. And these objects sticking out here are parts
of the fish's body. But anybody would think that they were parts
of a seaweed. They look exactly like parts of a seaweed,
and the seahorse hides among seaweed of just the right type.
It is almost perfectly camouflaged. And we have a few more examples.
This is a film of a leaf insect. There's the shield over the thorax, there's
the head. Here come the wings, and when it isn't moving,
perhaps even when it is moving, it looks exactly like leaves. It's just flown
off. Here is another. The thing looks just like
a plant. Turns out to be a green snake. And this you would think was a plant, with
a bud on the end, a long, green stem. More buds.
Only when we get to the front end do we just about spot that it has an eye,
antennae and legs. It is, in fact, a stick insect.
Look at these leaves here. Autumn leaves. Look at the vein up the middle of the leaves.
Look at the veins on either side. Look at the little splotches of dark colored
mold on the leaves. But those are not leaves. Those are butterflies.
Look, you can just see the body, there, there, there, there.
That's what these butterflies look like when they open their wings.
This is what they look like on the underside of the wings.
And they normally sit with the wings folded so that you only see the underside of the
wings. And you are very hard put to it to see that
they are not dead leaves. Only when they open their wings you get this
flash of brilliant coloration. Camouflaged animals resemble inedible objects.
Designoid objects sometimes resemble other designoid objects for other reasons.
Because they are doing the same job. And this is called "convergent evolution."
This is an ordinary hedgehog. That is nothing whatever to do with a hedgehog
but superficially it looks like it. This is a spiny anteater. It is a mammal but
you might say, only just a mammal. It's an egg laying mammal, a very primitive
mammal from Australia and New Guinea. As a matter of fact its way of life is not
that close to a hedgehog's way of life. This is an ant eater whereas hedgehogs eat
more general things: insects and worms and things.
But both of them gain protection from having spiny skins.
And so they both superficially look very alike. An example of convergent evolution.
An even better example of convergent evolution is the so called marsupial wolf.
Now, if you saw that going along on a lead down the street you would think it was a dog.
A slightly odd sort of dog, perhaps. There are not many dogs that have quite that
structure at the tail end. But you would think that that would not really
be out of place at Crofts. But this is not a dog. It has nothing whatever
to do with dogs. This is a marsupial. It's much more closely related to kangaroos
and wombats and koalas. It's now, most unfortunately, extinct. Only
fairly recently extinct. It went extinct this century in Tasmania.
It went extinct some thousands of years ago on the mainland of Australia.
And the reason it looks so like a dog is that it does the same job as a dog
or did the same job as a dog. It ran and hunted prey in the same sort of
way as a dog does. And that is why it has the same shape as a
dog. The structure inside also resembles that of
a dog. This is a dog skull. And this is the skull of the Tasmanian wolf.
It's a bit larger but the size of a skull would depend on how big the dog is.
And we could easily get a bigger dog's skull than this.
The only reliable way to tell that this is a marsupial and not a real dog is if you look
underneath. Those two holes in the palate there give the
game away. Those are the telltale clues that tell us
that this is a marsupial and not a real dog. Real dog doesn't have the same kind of holes
there. Well, that's convergent evolution among designoid
objects. Designed objects, too, sometimes resemble
each other because they are doing the same job.
Two airplanes closely resemble one another not because of industrial espionage
not because of imitation but because the wind tunnel is a great leveler
of differences. These planes have both been designed for the
same purpose and when you design planes for the same purpose
of flying very fast through air and carrying a large payload of passengers,
those planes are going to come out looking pretty similar,
in just the same way as the dog and the marsupial wolf come out looking similar.
So we have seen convergence between two designoid objects
and we have seen convergence between two designed objects.
How about convergence between designed and designoid objects?
Well, here is a camera, which is a designed object,
and here is an eye, a designoid object. They both do something very similar.
They both have a lens at the front which focuses an image on a light sensitive surface at the
back. In the eye it is called the retina: in the
camera it is a film. There are detailed resemblances as well. Both
of them have an iris diaphragm. It opens and closes to regulate the amount
of light that goes in. In an automatic camera the amount of light
that goes in and out is automatically regulated by a light meter, which when it gets brighter
closes down the hole, when it gets darker, opens up the hole.
And the human eye also has an automatic light meter.
I wonder whether we could have a volunteer? Right. In the front row there. Yes. Right.
What is your name? Gillian.
Would you like to take your glasses off, Gillian. Thank you.
Come and sit down here, please. Now, what you do is look into the camera there
and they are going to take a picture of your eye, of your iris.
What I'm going to do is I'm going to shine this light into the other eye.
And what we hope to see is that your iris will contract when I shine the light in.
So, look into the camera and I'm going to shine the light.
Did you see it contract? Look into the camera. There, it contracts.
Did you see it? I think it's best perhaps to look into the
eye that I'm actually shining the light into. No, not you.
Look into the camera. Is it too bright for you? No, okay, look into the camera.
There it goes. Did you see it? Thank you very much, Gillian.
And there are numerous other examples of living things
being exactly the way a human engineer would have designed them.
I hope that's enough to convince you that there's something special about living objects.
They look designed, they look overwhelmingly as though they're designed.
I call them "designoid" and I ask you to accept this different title.
But it is terribly, terribly tempting to use the word "designed".
Time and again I have to bite my tongue and stop myself saying, for example,
that this swift is designed for rapid, high speed, highly maneuverable flight.
And, as a matter of fact, when talking to other biologists
we none of us bother to bite our tongues, we just use the word "designed".
But I've told you that they're actually not designed and coined a special word "designoid"
and I said that there is a special process that brings designoid objects into existence
and gives them their apparently designed look. What is that special process?
The answer to this question was discovered surprisingly recently,
in the middle of the last century. One of the greatest discoveries of all time,
made by one of the greatest scientists of all time - Charles Robert Darwin.
A surprisingly long time after he discovered his principle of evolution by natural selection,
he wrote this famous book, The Origin of Species. This is an original first edition, inscribed
by the author. Very valuable. Darwin began his argument on natural selection,
he introduced it, in terms of another process called artificial
selection or selective breeding. All these vegetables have been bred by human
breeders for different food purposes. There is an ordinary cabbage, cauliflower,
red cabbage, broccoli, Brussels sprouts, kohlrabi. Each of these different sorts of plants has
emphasized a different aspect of the original wild ancestor
- the wild cabbage. So kohlrabi for example has a greatly swollen
stem. Cauliflower has a greatly enlarged flower.
So does broccoli but in a different way. They're all descended over the last couple
of thousand years from the same wild ancestor, the wild cabbage.
Now that's the wild cabbage as grown by Bryson. Bryson has many virtues but green fingers
are not among them. Nevertheless, if you were to take this home
and grow it for a little while, water it properly, look after it, it would
grow up into a wild cabbage. It wouldn't look very like any of those cabbages.
That's the point I'm making: they've all come from the wild cabbage but
they are all very different from the wild cabbage.
All the breeds of domestic dogs have been bred from the same common ancestor, namely
a wolf. Those dogs look terribly different.
You would never think they were the members of the same species but they all,
in fact, come from the same species, a wolf. Now, what is this artificial selection, this
selective breeding that enables you to go from a wolf to something
like that or that or that? Well, I'll tell you very, very briefly what
it is. You start with the ancestor, the wolf. And I am going to suppose for simplicity that
everybody on this side of the room is going to imagine breeding for smaller and
smaller wolves, and everybody that side is going to imagine
breeding for larger and larger wolves. So in every litter of wolves that we get,
if you are on the small side - what you do is to look out for those individual
puppies that are a bit smaller than the average. Those are the ones that you breed from,
and on this side you breed from the larger ones.
Now, it is going to take a long time, generation after generation.
You mate together relatively small dogs, wolves. And on this side you mate together relatively
large wolves. And after many generations, perhaps hundreds,
perhaps thousands of generations, perhaps a couple of thousand years of this
selective breeding, because there are genes involved in controlling
the differences between the cubs, eventually you may end up with something like
...what I hope is now going to come in. That would be the end product of breeding
for larger and larger sized wolves. That would be something like the original
ancestor that you started from, and that would be the end product of breeding
for smaller and smaller wolves. What are their names? Jemima and Wilf
Jemima and Wilf are Chihuahuas. Jemima is a smooth coat Chihuahua. Wilf is
a rough coat Chihuahua and both of them comparatively recently are
descended from a wild wolf. What's her name? Sequin
This is Sequin, a German shepherd and I think we could say that Sequin is the
one who most resembles the ancestral wolf. And this is Archie, that's a fine name, Archie,
a great Dane, and he is what you get from breeding for larger
and larger size but they all descended from a wolf, they're
all cousins of one another and Sequin is showing great interest.
Thank you very much indeed. Thank you. Charles Darwin was very interested in dogs.
He was also very interested in pigeons. You're on my notes, pigeon.
This is a Marchenero cropper. It's a pigeon which has been bred by artificial
selection from the wild rock dove and in this case it has been bred for the
thickness of feathers and for size of crop. You see the great big crop at the front which
is blowing up like that and see how big the feathers are.
In the cage here - we didn't quite trust these ones to be let out. We trust this one.
This is a domestic flight pigeon and you see how it has been bred for this
curious little ruff round the back of the neck. And also the red ring around the eye.
The other one is an English short-faced tumbler pigeon.
You see the extraordinary short face and the tiny small beak.
That again is a product of artificial selection, just like in the case of dogs and cabbages.
The beak is so short in the case of English short-faced tumbler
this breed is no longer capable of feeding its own young.
And so the only way that breed can be reared is by its babies being reared by pigeons of
another breed. That sort of thing often happens, by the way,
with artificial selection. It's true of bulldogs. You probably know that the bulldog breed of
dog has a head that is so big that it can't be born and the only way a bulldog
can be born is by Cesarean section. So the entire breed depends upon humanity
to keep it going. If we went extinct, bulldog would go extinct.
Artificial selection, the process that produces these dogs, these pigeons, and these cabbages,
is too slow for us to see during the course of one lecture.
But we can imitate it on a computer. And I'm going to do it with a program called
"arthromorphs." These are arthromorphs.
This is the parent arthromorph and round the edge of it are eight child arthromorphs.
They resemble the parent very closely but there may be a genetic change, a mutation,
a random genetic change, as you go from parent to child.
So that one, for example, has longer legs, that one has got legs up instead of down.
The way you breed arthromorphs - Could I have a volunteer -
My goodness. Right. Yes. Yes. Have you ever used a computer with a mouse?
Yes. What you do is you choose the one you want
to breed from and just click it once. So you are going for the long legged one,
I think. Click it. It goes to the center and becomes the parent
of the next generation. Now you see, all in the next generation have
longer legs. What's your name, by the way? Lawrence.
Lawrence seems to be breeding for longer legs. I wonder if he is going to continue that.
That's right. Keep going. Don't wait for me. Just keep on breeding.
Okay, Lawrence likes long legs and they are getting longer and longer.
What I said was that these creatures have genes which are going from parent to child.
What would I mean by talking about genes in a computer?
In a computer, of course, genes would just be numbers.
They are not real genes, they are not made of DNA, but nevertheless
they are genes in the sense that they are what go through from generation to generation.
There is no sex in these creatures, by the way,
these are all reproducing asexually, like stick insects and like aphids.
Carry on. Breed as fast as you can to get through a
lot of generations. So what we are seeing now, what Lawrence is
doing, is artificial selection, just like our ancestors did with dogs and
pigeons. But he is managing to achieve in a couple
of minutes what would have taken several centuries for
our ancestors to have achieved. What are you going for, Lawrence?
You're trying to get a lot of zigzags in the legs, are you?
All right. Perhaps we better proceed now, so thank you very much indeed, Lawrence.
I think that we are all convinced by now that artificial selection works.
We've seen the results of it in dogs, cabbages and pigeons,
and we have seen it happening before our eyes in the computer arthromorphs.
But this is just artificial selection. We began talking about artificial selection
because we're really interested in natural selection.
Natural selection is like artificial selection except that instead of humans doing the choosing,
nature does the choosing. Of all the puppies in a litter, wolf cubs
in a litter, instead of our choosing which one shall breed,
what happens is that nature chooses which one shall breed.
The ones that have what it takes to survive will be the ones that breed, automatically
chosen. The ones that are good at running fast,
the ones whose legs are not too short and not too long.
The ones whose teeth are not too blunt and not too sharp,
because if they are too sharp they might break easily.
Natural selection, nature, is constantly choosing which individual shall live,
which individual shall breed. And the result, after many generations of
natural selection, is much the same as the result after many
generations of artificial selection. So what would it take to change the arthromorph
program so that it simulated natural selection instead
of artificial selection? Because at present the arthromorphs are just
being chosen by the eyes of a human. Could we somehow make the computer do its
own choosing? Choosing on the basis of quality of arthromorphs.
The trouble is, it's not easy to judge what quality in a arthromorph might mean,
because these arthromorphs are living in a very strange environment:
a two dimensional computer screen. They don't have a real world in which to live,
they don't have predators, they don't have prey, they don't have food
they've got to catch. Perhaps we'd do better if we made a computer
model of a two dimensional designoid object. Like a spider web.
Now if we could have the lights down, I think we might be able to see -
There is a spider in the middle of its web and that, I think, shows quite nicely.
You know what the spider web is for: it is for catching flies and other prey.
It is a net and it works in two dimensions. We would've liked to actually shown you a
spider building its web but this one seems to be pretty satisfied
with the web it has already got. So what I'm going to do instead is to show
you a computer reconstruction of a spider's movements while it builds a
web. Do watch carefully. This is rather speeded
up. Can we have that more slowly now, Peter?
What the spider is now doing is the radii of the web.
Now it's doing the structural spiral, which is a kind of scaffolding.
And now it is doing the sticky spiral which is the bit that actually does the business
of catching the flies. Let's have it once more, slowly.
There are the radii. Now it's doing the scaffolding.
And now it is doing the real sticky spiral. What we've seen there is not actually a picture
of a web itself. That's a picture of the movements of a spider
that were recorded on a particular day. That's a particular spider on a particular
day. Its movement were all fed into the computer
and now the computer is playing it back to us.
But that's just a recording of the web of a real, particular spider.
In order to do our trick of making an arthromorph-like program out of spider webs
we've got to make the computer behave like a spider.
And this is the program written by Peter Fuchs who is next door, controlling the computer.
And what his program does is to make the computer build the web as if it was a spider.
So the computer is holding in its little head the rules, that we know something about, of
how a spider builds a web. So the computer does the radii like that.
It does the spiral like that. Just as in the case of the arthromorphs, what
Peter has done is to make the building rules of the computer
spider, under genetic control. There are genes in the computer, just as they
were for the arthromorphs, and just as they were for the arthromorphs,
the genes are simply numbers. That is the parent web, these are the daughter
webs more strictly, that's the web that was build
by the parent spider, these are the webs that were built by the
daughter spiders. To begin with we can treat this just as if
it was an arthromorph program. So we want another volunteer, let's have a
girl this time. What's your name? Ursula.
Ursula, have you ever used a computer with a mouse? Yes, good.
So this time is just like the other one only you have to click twice instead of once
to choose which one you think is the best web.
Okay, now that web has gone up to the top there. That's now become the parent.
And here are the daughter webs that are being drawn.
And now you can choose another one, another generation.
So you see we're doing just the same as we did for the arthromorphs
but now we have got spider webs coming. But the point we were going to go for was
not artificial selection. The whole point of doing this with spider's
web is to do something like natural selection. And to do that we simply make the computer
work out how good each web would be at catching flies.
And we can do that because, unlike the case of the arthromorphs,
the webs are two dimensional structures and we know what they are for.
They're for catching flies. So the benefit is simply going to be the number
of flies caught and the cost we can calculate as the amount
of silk used, because the webs are made of silk.
So the cost of a web is the amount of silk used and the benefit is the flies.
If you would like to stop now, Ursula. Thank you very much.
Now, we no longer have a human selector, we now have the flies doing the selecting.
So the flies are going to hit the web and Peter gets to start it up again.
Now we've got a new generation of webs being built
and we are now going to see the flies hitting the webs.
There come the flies. Flies again. Now the computer is going to calculate which
of the webs is best at catching flies and it's that one that has gone dark.
So that one will now become the parent of a next generation.
And now, once again. The webs are being built, the child webs are being built. Once again.
the flies will come, the computer will measure which one of them is the best.
There it is. And that becomes the parent of the next generation.
It wouldn't take very long for us to see the evolution starting from nowhere at all
and going to a nice web that works very well, but we haven't quite got time for it.
So instead, what we did, was to let the computer run all night, all last night,
and we've got a "fossil record" of all the webs that were built during that time.
This was the starting web, the thing that started at the beginning of the night's run
and then, every 20 generations, we have a printout of the shape of the web.
So you see, we start with almost no spiral at all,
and you could imagine the fly just whizzing straight through and not getting caught.
But then natural selection in the computer led to a gradual improvement in the web,
more and more spiral, more and more flies caught,
and so evolution went in the direction of a nice full web
with a nice full spiral like that, catching lots of flies.
That all went on in the computer last night, very fast, telescoping into one night
what would have taken thousands of years, perhaps millions of years in nature.
In nature, the successful and the unsuccessful webs would not, of course,
be judged by the computer doing a calculation, about how many flies would have been caught
- would be expected to be caught -
the webs would be judged automatically and without any thought by the flies themselves.
The flies themselves that fly into webs, thereby choosing webs for breeding.
The flies don't know they're choosing the webs for breeding.
They don't particularly want to fly into the webs,
but nevertheless the consequence of their inadvertently flying into webs
is, that the spider that built the successful web,
is a spider that is more likely to breed and therefore more likely to pass on the genes
for building that sort of web. So, as the generations go by, webs get better
and better and better, just as they did in the computer in our overnight
run. That's natural selection in the case of spider
webs and exactly the same principle works for every
living creature, for the bodies of every living creature.
Every lion and tiger, every camel, every dog, every human, every giraffe.
They all have evolved by the same process of evolution by natural selection.
So the Darwinian view is that designoid objects are not designed at all.
They have evolved by natural selection. The most popular alternative to the Darwinian
view is called "creationism." Creationists believe that designoid objects
are really designed objects. The only difference is that whereas these
designed objects were designed by humans, these designed objects were designed by a
divine creator. And the favorite argument of creationists
is the so called "argument from design," which was most famously expressed by
Archdeacon William Paley, in 1802, in the book, Natural Theology.
Paley begins his book: "In crossing a heath, suppose I pitched my foot against a stone,
and were to ask how the stone came to be there; I might possibly answer
that, for anything I knew to the contrary, it had lain there forever."
In other words, the stone is the kind of object which had always been there
and doesn't need any special kind of explanation. But, Paley goes on: What if I accidentally
kicked a watch? The watch, I open it up, I see the mechanism,
I see the cogwheels, and the springs, everything about it looks
designed. It had to have a designer. It had to have
a watchmaker. And if the watch had to have a watchmaker,
then how much more - Paley argued - must these objects, these living objects,
including ourselves, have had a "Divine watchmaker." For Paley it follows, as clearly as the night
follows the day, that just as the watch had to have a designer,
so do we have to have a designer. But, of course, just to show that animals
and plants look as though they have got a designer begs
the question. I've spent much of this lecture trying to
persuade you that animals and plants look as though they've got a designer.
But I've spent the other half of the lecture showing you
that there's another very good alternative explanation
for why they look as though they had a designer, namely, natural selection.
Well, Paley, of course, lived before Darwin, so he couldn't be expected to know about the
alternative. Nevertheless, it was even without knowing
about Darwin, it was possible back in the 18th century
to know that Paley's argument was a pretty bad argument.
And this was pointed out by David Hume, one of the greatest philosophers of all time.
Hume made the point that the argument from design, which was Paley's argument,
is that things like elephants and humans are too complicated to have come about by chance.
They have many parts, just like a watch, too complicated to come about by chance.
A designer, a watchmaker, an engineer is certainly one way in which these objects
could come about. But a watchmaker or a designer or an engineer,
if he is to be any good as a watchmaker or an engineer,
must be pretty complicated object himself. It's no good just postulating a designer because
a designer is just the very kind of thing, just the very kind of complicated, ordered
thing, that seems to need the very kind of explanation
that we are searching for. If a human is too complicated to have come
about by luck or if a swift is too complicated to have come
about by luck then a thousand times more so any being capable
of creating humans must be too complicated to have come about
by luck. The argument from design certainly proves
that living things couldn't have come about by chance,
but by the same token, and even more strongly, it shows that the divine creator couldn't
have just come about either. The creator would have needed an even bigger
creator, and so on. The argument from design is a powerful seeming
argument and it powerfully shoots itself in the foot.
The Darwinian argument of evolution by natural selection, of course,
doesn't suffer from this problem. The Darwinian argument does not explain things
as due to chance. Chance, in form of random genetic mutations,
comes into it, but by far the most important part of the
Darwinian explanation is the non-random process of natural selection.
There's another, rather interesting, little curiosity which is that natural objects, designoid
objects, have imperfections which you wouldn't expect
to get in objects which were designed by a real designer.
This is a flatfish, a halibut. Its ancestors once swam normally in the water
like a normal fish does, like that. But the ancestors of the halibut settled down
on the bottom of the sea, one side down. They lie on the bottom of the sea, and now
a modern flatfish moves along like that. You have probably seen them doing it.
But when it did that the ancestor found that one of its eyes
was looking straight into the sand. Only the other one was looking up.
And so, gradually in evolution, the other eye, the one that was looking into the sand,
migrated round the side of the head and came up to the top.
With the result that the skull of the halibut is now an extremely distorted object.
It's like a sort of Picasso painting of a fish. It's got its two eyes on one side.
Anybody who was going to design a flatfish wouldn't do it that way.
You would do it like a skate, which is another kind of shark, which is also a flatfish.
And it flattened itself, its ancestors flattened itself,
by going on to its belly so that both its eyes were looking upwards
and it had no need to do any kind of distortion. But by some kind of historical accident the
ancestors of the halibut and the sole and the plaice all did it by
lying on their side and that meant that they had this distortion.
So this is an imperfection in design. Which is just the kind of thing you'd expect
to see if these creatures had evolved, but is very much not the kind of thing you'd
expect to see if these creatures had been created.
Evolution starts from simple beginnings. The starting point of evolution is the kind
of thing we see here. Something like crystals, something at least
as simple as crystals. And it builds upon simplicity to get towards
complexity. We start with a simple foundation, simple
things are easy to understand. We do not have to start with a complicated
thing, like a creator. On this simple foundation are built designoid
objects by natural selection. And only when we have designoid objects with
brains as big as human brains, does design finally emerge.
But why do I say just humans for design? Isn't it rather unfair to the wasps that built
these pots, and bees and spiders and things, rather unfair to this ovenbird that built
that mud nest or that nest of social wasps,
which is very similar to that convergently, also built of mud?
Why do I use the word "designed" only for human creations
and not for the manufactures of these animals? The difference is that human designs
get their goodness and efficiency from conscious human foresight.
Wasp pots and ovenbird nests get their efficiency directly
from natural selection by a kind of hindsight rather than foresight.
Gene are selected which influence the bodies of the ovenbird and the wasps
and particularly the nervous systems which influence the building behavior.
The birds and the wasps have no idea of why they are doing what they do.
Natural selection simply favors those that build good nests.
Humans, on the other hand, do build with foresight. At least they do usually.
This is an engineer called Ingo Rechenberg from Germany who designs windmills
and he claims that he designs his windmills by a kind of natural selection.
He does it by putting his windmills in a wind tunnel, measuring how good they are
and then - as he calls it - breeding from those windmills
that are good at spinning around in a wind tunnel.
The windmills have genes, and again, not real genes, but they are numerical attributes,
they're numbers that are used to make other windmills that resemble the parent windmill.
And in every generation of windmills he breeds from the ones that do best in the
wind tunnel. And after many generations of testing them
and breeding, testing and breeding, he ends up with a windmill, he claims, is
better than you would get by the ordinary processes
of engineering design. But you could say that all engineering design
and even all art has a certain Darwinian component. I want to illustrate this with another computer
program called "biomorphs". So, can we have a volunteer to run the biomorphs?
What's your name? Rachel Rachel. Have you ever used the mouse before?
Yes. Good. Now, there you have some biomorphs.
Try one of those now. What she is doing is guiding the evolution
of biomorphs. The biomorphs are controlled by genes,
just like the arthromorphs and just like the spider webs.
And they are coming by random mutation but the direction of the evolution is being
guided by the human eye. Just like the direction of breeding cabbages
or dogs. In this case we are just looking for pure
aesthetic appeal. Just looking for the prettiest ones.
I think you might imagine breeding wall paper or bathroom tiles or something like this.
Okay, thank you very much indeed, Rachel. But in any case, all creation, all design,
all machines, houses and paintings and computers and airplanes,
everything designed and made by us, everything made by other creatures, is only
made possible because there are already brains put together
as designoid objects. And designoid objects come about only through
gradual evolution. Creation, when it does occur in the universe,
is an afterthought. When creation appeared on this planet, it
came locally and it came late. Creation does not belong in any account of
the fundamentals of the universe. Creation is something that, rather late in
the day, grows up in the Universe Thank you very much.