Tip:
Highlight text to annotate it
X
PETER NORVIG: Welcome everybody.
It's our pleasure to have with us at Google
today, Daniel Dennett.
He has a long and distinguished career.
I won't try to summarize all of it.
You can Google it.
[LAUGHTER]
PETER NORVIG: Instead, I'll just try to introduce very
briefly the book that he's written most recently and is
talking about today, which is a collection or a
tool kit for thinking.
And I think the reason I think it's such a great book is
because a lot of it is our tool kit.
It's these computational tools for thinking that we take for
granted, but Dan is trying to introduce to the rest of the
world who sometimes aren't aware of how
to think like that.
And so what do I mean by that?
Well, in his 1995 book on evolution, he wrote that,
"Life on Earth has been generated over billions of
years by an algorithmic process." And that's kind of a
natural way for us to think about it, but it's hard way
for other people to think about.
And some people don't quite understand that by taking lots
and lots of small steps that you could arrange and come up
with this result.
And other philosophers, who we won't name, have taken the
approach, well look at these incredible results.
And if I simulate the steps for three or four steps I
don't get there.
So it must be magic.
Rather than saying, well, maybe a billion or 100 billion
is actually qualitatively different than a couple steps.
So Dan reminds us how to think that way.
He also reminds us that there can be emergent properties
that aren't implicit in any one piece of an ensemble and
yet they come out in the end.
We're all familiar with that.
Our programs have emergent properties
that we didn't expect.
And we *** our heads trying to fix them.
So we're aware of it.
But the rest of the world isn't.
So that's why they need a book like this.
I just want to talk about one more of the tools.
And this is the using lay audience as a decoy.
So this is a trick if you have a philosophy professor who's
not explaining all the details, rather than ask them
to explain a few you say, oh, well why did you give a
lecture to this introductory class.
And then it will become more clear because the professor
will have to work harder to get it to that level rather
than trying to keep it at the professor to professor level.
One of our Google employees Rob Pike wrote a book where he
talked about a similar approach.
Instead of calling it the lay audience as decoy approach, he
called it the teddy bear approach.
Which is, there was a introductory class in computer
programming for which the TAs were overstocked.
And there was always a long waiting line of students
waiting to get in to talk with the TAs about why the program
didn't work.
So what they instituted was a chair that held a teddy bear
outside their office.
Before you were allowed to talk to the TAs you had to
talk to the teddy bear.
And so the students would come up and say, oh teddy bear, my
program doesn't work.
And it must work because this loop, see, this could never be
more than one.
Oh, except, oh, never mind.
And it turns out the teddy bear successfully solved about
half the problems.
And people didn't have to go on to the TA.
So I think that's sort of in the same toolbox.
And Dan will tell you about the rest of the tools.
DANIEL DENNETT: Thanks, Peter.
That's a great introduction.
I'm going to add the teddy bear.
It occurred to me as you were telling us about that, that
that probably explains the success rate of psychoanalysts
right there.
You can replace them with a teddy bear and
get the same results.
AUDIENCE: [INAUDIBLE].
DANIEL DENNETT: Yeah, Teddy bears are a lot cheaper.
So I want to talk about better tools for thinking.
One of my best students, a Swede name Bo Dahlbom once
said, "You can't do much carpentry with your bare
hands, and you can't do much thinking with your bare
brain." He hits the nail on the head right there.
It's obvious, in a way, when you think of it, even
something as simple as a word is a tool.
It's a thinking tool.
And these thinking tools have been around for a long time.
And recently we've learned that maybe these thinking
tools are having a remarkable effect.
How many of you know what the Flynn effect is?
Oh, a few of you do.
And the rest of you, all right, now let me tell you
about the Flynn effect.
Jim Flynn, it's named after him, but he's not really the
one that did the first research.
He's the one who drew people's attention to it.
It turns out that IQ worldwide is up.
And it's up a significant amount.
If you use today's test, if you gave that test to a person
who got 100 back in the '30s they get probably an 80.
And we'd be viewed as almost retarded.
I mean, there's a big jump.
It's like 20 or 30 IQ points.
Of course, IQ tests are always scaled with 100 as average.
But if you use the same test items, at the very least we've
gotten a lot better at taking IQ tests.
And it has nothing directly to do with how much tutoring or
schooling or your genes.
It's a very large effect.
And it can't be due to improvements in our genes as
we're talking about such a short time periods.
Jim Flynn thinks-- and I hope he's right.
That doesn't mean he's right, but I think there's a lot to
be said for it-- that what's happened is that the thinking
tools that are first developed and tested and debugged by
scientists and engineers and other professional thinkers
have been filtering down into every day people's thinking.
Even as simple an idea as percentage.
Which kids, you just pick it up very fast.
And 50 years ago, 100 years ago, that was a pretty
complicated concept for a lot of people.
They just weren't ready for that.
And so probably it's all these thinking tools that we install
without even realizing we're doing it.
We install in on our own brains and in other brains.
Probably that is a main source of our improved thinking.
And certainly that is the thesis that I want to explore,
in any case.
Because what I'm doing in this book, as a philosopher, I
discovered that a lot of my colleagues are remarkably
unselfconscious about the tools of the trade.
Weird because philosophers are famously self conscious and
reflective and navel-gazing.
But a lot of them don't ever ask wait a
minute, how do I do that.
And why do I think that's a good way of proceeding?
What tools am I suing?
And I wanted them to become more self conscious.
I want everybody to become more self conscious about the
tools that they use.
So what kind of thinking tools are there?
There's words, as I say.
And numbers, of course.
And diagrams and maps and methods.
And intuition pumps.
Which is a term I coined more than 30
years ago to describe--
well in the first instance it was John Searle's notorious
Chinese room.
And since I was beating up on that intuition pump, a lot of
people thought I thought that intuition pumps were bad.
No, I think they're wonderful.
It's just got to have good ones.
There's defective pumps.
And there's good ones.
And the way you tell the difference is, as you know, by
reverse engineering.
Doug Hofstadter came up with the lovely phrase, you want to
turn all the knobs on an intuition pump to see what
makes it do its work.
You want to try out the variation.
Nothing mysterious about why that works.
Is just Mill's method of differences.
You're just empirically testing to see what variations
make what differences.
See how robust the intuition pump is under deformation.
And sometimes they're remarkably sensitive to tiny
variations in the way you tell a story.
Then you know it's probably a defective pump.
Not doing what you think it's doing.
So what intuition pumps are are persuaders.
And it's important that they're persuaders.
So much good thinking is accomplished in a friendly but
competitive environment where I am trying to persuade an
interlocutor, a target, that P and the target is trying to
persuade me that not P. And this opponent process is
actually a good way, very often, of
getting at the truth.
So they're also discoverers.
They're attention-holders.
And they're anchors for a fixed points.
In my work in philosophy there are no fixed points.
That's the point a philosophy.
Philosophy is what you're doing when you're not quite
sure what the right questions to ask are.
And everything is up for reconsideration always.
But how do you proceed then?
You say, well, maybe this as a fixed point for this
discussion.
And then you sort of put a pin down here.
And here's one here.
And here's one here.
And then you say, well, suppose we
take those on board.
What can we do with this idea with a few of these candidate
fixed points?
One thing you may do with them is throw all three of them out
or throw out one against the other two.
It's in this intellectually informal.
There's no algorithms for doing this, certainly not yet.
This is an informal exploratory process for which
there are really no rules.
But intuition pumps are persuaders to use in those
harrowing circumstances.
Here's one that isn't in the book, but I've talked about it
a lot recently and I think it's important to understand
the context.
The MacCready explosion.
This is Paul MacCready.
How many of you know Paul MacCready?
Oh well good.
I can educate you about that.
You will know him, not by name, as the inventor of the
"Gossamer Albatrosss." A visionary engineer.
The greenest of green engineers.
Wonderful guy.
Died a few years ago.
And in a Ted talk a few years back he came up with the
following startling observation.
Let's go back 10,000 years to the dawn of agriculture.
At that point, human beings all around the world we're
beginning to domesticate animals and plants, of course.
And at that point, our species, *** sapiens, and
these are, of course, entirely modern *** sapiens.
At that point 10,000 years ago, if you put them on one
side of the balance scale together with their pets and
livestock, all the domesticated vertebrates on
one side, including us.
And you put all the rest of the terrestrial vertebrates on
the other side.
Animals, not the insects, not the worms, not the fish.
He calculated that at that time, 10,000 years ago, we and
our livestock accounted for a fraction of 1% by weight of
the terrestrial vertebrate biomass.
What do you think the percentage is today?
Any guesses?
AUDIENCE: 10%.
DANIEL DENNETT: 10%.
AUDIENCE: Over 50%.
DANIEL DENNETT: Over 50%.
AUDIENCE: 99%.
DANIEL DENNETT: 99%, are you kidding?
No it's 98%.
It is 98%.
We human beings have engulfed the planet.
This is one of the most astonishing biological
phenomena that has ever happened.
And here's what Paul said about this.
"Over billions of years on a unique sphere, chance has
painted a thin covering of life--
complex, improbable, wonderful, and fragile.
Suddenly we humans have grown in population, technology, and
intelligence to a position of terrible power: we now wield
the paintbrush."
I've highlighted the technology and intelligence
because those are the thinking tools.
The only difference between us and our ancestors of 100,000
years ago is the thinking tools.
We have the same genes.
We have pretty much the same justice systems.
The same muscles.
The same brains.
The thinking tools have changed us and made all of
this power possible.
It is the source of our power.
Which raises a sort of chicken egg problem.
Did evolved tools make us smarter?
Or did we evolve to become smart enough to make tools?
And like all good chicken and egg questions,
the answer is yes.
It's a coevolutionary, bootstrapping process where
you have a little evolution makes us smarter.
And then we can be smarter about making tools.
We became can become self conscious about
the tools we make.
Our ancestors weren't self conscious.
They didn't invent words.
They didn't coin words.
They just found themselves using words.
And it made them smarter.
Now we are very deliberate.
We have become intelligent designers of tools.
But initially we weren't intelligent
designers of tools.
We were simply the beneficiaries of tools.
Well how did they get there then?
Not by the genes.
They evolved, culturally.
My next book is going to be on cultural
evolution and that story.
But I'm not going to say more about it now because I'm going
to be looking today at recent tools intelligently designed,
supposedly intelligent designed thinking tools.
Everybody knows reductio ad absurdum where this is the
sort of crowbar of logical argument.
It's what you use.
You take your opponent, your target's premise.
You take your target's thesis on board,
for the sake of argument.
And then you use the crowbar.
You try to deduce a contradiction from that.
And if you succeed you win.
And we use it all the time very unselfconsciously.
Very few people, probably you've used to sort of
disguise reductio three or four times in the last day.
You've said something like well if that's a bear then
bears have antlers.
That's a reductio.
You'll find yourself making them.
Now why do I bother mentioning that since it's a very
familiar tool?
Because I want to point out that rhetorical questions, the
questions you're not supposed to answer.
Many of them, not all of them, but very many of them are, in
fact, disguised reductio ad absurdum arguments.
They imply a reductio ad absurdum argument that the
writer doesn't bother presenting.
It's too obvious.
It would be embarrassing.
So you just get the rhetorical question and nudge,
nudge you go on.
So what I want to do is to create in you a little alarm.
Sort of like a Google alert.
So that whenever you see a rhetorical question in an
argument, ding.
You think, ah rhetorical question.
You look to see, maybe I can just answer it.
Maybe there is a perfectly unembarrassing answer.
I remember years ago seeing a "Peanuts" cartoon and then the
next to last frame Charlie Brown is saying, "Who's to say
what the difference is between right and wrong." Lucy in the
last panel says, "I will." and think about answering a
rhetorical question but just answering it.
See if you can do that.
A similar thing is what I call the surely alarm.
Whenever you see the word surely a bell
should ring, ding.
Let me just give you one.
Oh, yes, you could make a computer they could sort fruit
by its color, but surely no computer, ding.
Surely no computer could ever have a favorite color or
appreciate color
When you hear that surely you are being--
now maybe you think, I go along with that one.
That's got to be right.
Actually in this audience I bet a lot of people say, I
don't think that's so obvious.
[LAUGHTER]
DANIEL DENNETT: Why surely?
Well the author wants to get this proposition, the
proposition that follows it into the discussion.
Thinks it's true.
Thinks it's obvious.
But it's not quite obvious enough so that it
goes without saying.
If it did you wouldn't have to say it.
It would go without saying.
It goes with saying, but he doesn't bother arguing for it.
So he gives you just a little nudge and says surely.
So be on the lookout.
I tested my hunch about this by going
through, thanks to Google.
Thanks to string search.
Going through some dozens of philosophy of mind papers that
were online.
Counting up the surelys.
Most papers don't have the word surely in them.
But I found several dozen instances of surely.
And about 1/3 of them were I thought clearly cases of this
was the weakest place on the whole argument.
This is the place you want to look first.
So now if I've alerted you on rhetorical
questions and on surely.
Another one is rathering, rather.
As in, it's not the case of blah blah blah.
Rather it's ta da da da.
Rathering signals a dichotomy.
And very often it singles a false dichotomy and what you
should do is, again, have a little
alarm ring for rathering.
When you see the word rather take a good hard look and see
if you can't say, well maybe both blah blah
blah and da da da.
That hasn't been asserted.
It's just been presupposed and glossed over
with the word rather.
So if I get you to adopt these habits, look what I've done.
I've installed an app on your necktop.
[LAUGHTER]
DANIEL DENNETT: And I mean that literally.
The thing between your ears is a kind of computer.
And I don't have to tell you folks, but the power of a
computer is the apps and the levels.
And you can build and build and build and build and build
apps on top of apps.
That's what the thinking tools do.
And that's what makes us smarter.
A CPU without an operating system, you can't do
very much with it.
You start building up all the software and that's where the
power comes from.
I don't have to tell you that computers are thinking tools
par excellence.
And in the book I put an interlude on
computers in the book.
And then Peter, I'm happy to say, he mentioned that I want
everybody who reads the book to come away with a little bit
more basic understanding of just where
computers get their power.
One of the things I do, and this is so eccentric.
I think that's probably the only trade book in history to
have this feature.
In this chapter called, "The Seven Secrets of Computer
Power revealed," I teach everybody how to program a
register machine.
A Hao *** simple register machine that can just do
increment and decrement and branch.
And they teach it how to do arithmetic.
And then they teach it how to do some other things.
And there's problem sets.
I've actually got some simple problem sets in this book.
And I say, well you don't have to do them, of course, but if
you really want to understand.
If you want to be a computer for a while and understand
what it is.
And, of course, they learn about loops.
They learn about counters.
They learn about all kinds of things.
That they hear these words in other contexts and they think
they know what they mean.
Now they're going to know what they mean.
The reason I think that's so important, and again, I know,
I'm preaching to the choir here, is one of the most
important epistemological facts about computers is that
if you can make a computer simulation of
something that works.
You know to a moral certainty, there's no magic.
There's no morphic resonance.
There's no psionic forces or ectoplasm.
It's good old-fashioned, every day,
hardened, honest causation.
And you just got a proof of it because of
the computer itself.
That's what it is.
There's no place for any mystery forces to play a role.
But people shouldn't just take that on faith.
They should know in their bones why that's true
and how it's true.
And that's what that section is supposed to show them.
Some intuition pumps don't even have words,
they're just images.
I love this one.
This is from my friend Matt Ridley, wonderful evolutionary
biologist and journalist.
And it just startles for a moment.
But what points can you make with it?
I'll just make one then move on.
On the left you see an Acheulean hand-axe.
On the right, I don't have to tell you what that is.
The Acheulean hand-axe was made by our ancestors without
any discernible stylistic change.
Without any discernible design improvement for
over a million years.
Absolute stasis.
Amazing.
Some people have even thought maybe it wasn't to be used axe
at all, maybe it was like the peacock's tail.
Maybe it was, hey, you want to see my hand-axe collection?
I am so good that I can make hand-axes all day long and
still get enough to eat, take care of myself.
Bragging.
Sort of like having a Lamborghini or
something like that.
But notice at the one on the right, familiar as is it is to
us, it probably has a rather short lifetime ahead of it.
How many bet that there will still be mice
in a hundred years?
Right, point made.
Now here's a similar comparison.
On the left we have a termite castle.
Australian termite castle.
And on the right we have Sagrada Familia, the great
Gaudi church in Barcelona.
As I think you can see, they're stunningly similar
structures.
And even internally, if you look at the cutaways, there's
a lot of design features that they have in common.
These are two remarkable artifacts made by animals.
And yet, for all their similarities, you can see at a
glance that they are the products of profoundly
different histories of design and construction.
The one on the left is, this is bottom up Darwinian design
and construction.
The termite, the individual termites, could
hardly be more clueless.
They don't know what they're doing.
There's no boss termite.
There's no architect.
There's no blueprint.
There's no leadership roles.
There's no hierarchy.
This is bottom up activity.
And it produces that.
Whereas Gaudi, it's almost comical.
He was like a cartoon caricature of
the inspired genius.
Autocratic, dictatorial, laying down the law.
He had blueprints and manifestos.
He had accounts of everything.
And he bossed people around.
And he told the bosses how to boss the people.
Classic case of hierarchical design and construction.
Here is one way of looking at the problem, I think, that
evolution and cultural evolution has to explain.
How did our planet get from the left to the right?
How did we get from termite type R&D to Gaudi-type R&D?
I think the answer has to be, well, Gaudi did not have a
bare brain.
He had all those thinking tools.
And those thinking tools are the key to opening up top
down, as it were, intelligent design type design.
What Darwin showed us is we don't need the intelligent
designer that lots of bottom up, mindless grubbing around,
trial and error, will eventually produce all the
wonders of the biosphere.
But then the process keeps right on going
with cultural evolution.
And pretty soon it starts to reverse itself.
And we get more and more intelligent design.
To where we now have people like touring, better example
than Gaudi, because he really had the goods.
Who notice, when he built his computer, he already had proof
of concept.
He had the blueprint.
He had the design.
He proved it all.
Nobody would have ever given him the money to make that
computer if he hadn't proved it in advance.
This is where the idea really is leading the way.
That's intelligent design.
And we've got to explain how it has happened to be one of
the fruits of evolution.
Because it wasn't the cause.
The Bible says in the beginning was the Word.
No, no, the word, very recent development.
The word is the beginning of intelligent design.
The Bible is right about that.
Another thinking tool which is just a picture, but what a
lovely picture.
And, by the way, this is Leonard Eisenberg's version of
the Tree of Life.
And you can get it on t-shirts, big shiny posters
that you put up in your office.
There's all sorts of ways you can get it if
you go to his website.
And what I love about it, there's many different ways of
representing the tree of life.
This has a few nice features.
I just want to sort of point them out to you.
I ran out of battery at some point.
OK.
There's no other pointer up here, is there?
I'll just do it like this.
Here's a pointer.
A thinking tool!
Right here on the end of my finger.
How wonderful.
So time goes out radially.
Here's the birth of the earth.
Here's the birth of life, the origin of life.
And as you can see, the bacteria and the archea are
the first single-celled organisms.
And then we have a period of single-celled eukaryotes.
This is a great, great, great moment in the history of life.
It's the eukaryotic revolution.
When the endosymbiotic event put two of these types
together they joined forces and became a
more powerful thing.
This is technology transfer.
This is what happens when you have two independent lineages.
They've been evolving independently
for a billion years.
And now they join forces.
They don't all have to reinvent the wheel.
They get the benefit of each current, and
those eukaryotic cells.
All the rest of life is eukaryotes.
You've heard of the Cambrion explosion.
There it is.
In a very short period of time tremendous flourishing of
different kinds of life.
A period of fantastic fecund R&D.
How was it explained?
That's a story I'd love to tell, but for another day.
You can see mass extinctions.
And then way over here, this last little fork here, that's
about 7 million years on that scale.
Which is about the length of time since we shared an
ancestor with our nearest relative, the chimpanzee.
And, of course, of that seven million years, only a million
years at most has any language.
Some people will push it back a little further.
The Acheulean hand-axe comes before language.
Some people think that's a controversial issue.
But so now notice, Cambrion explosion.
Fabulous.
The MacCready explosion happens there.
And transforms the planet.
All thanks to thinking tools.
So now, a few recent intuition pumps that are in the book.
Because I haven't really talked about them.
These are these little stories.
So here's an intuition pump I recently devised to discomfort
neuroscientists and cognitive scientists who have been going
around saying I think ill-considered things about
what neuroscience teaches us about free will.
So first a little science fact.
Damiaan Denys, a very brilliant researcher in
Amsterdam, has made a little microchip which controls
obsessive-compulsive disorder.
You implant it in a person's brain.
Really does a very good job.
Science fiction.
So one day this guy with OCD goes to see his neurosurgeon.
And she implants one of in these Denys
microchips in his brain.
Sews him up.
The anaesthesia wears off.
And as he's about to leave she debriefs him a little bit.
She says, oh and by the way, my staff will be monitoring
you 24 hours a day, seven days a week from our
control room here.
And you're going to think you have free will.
You're going to think you're making your own decisions.
But, in fact, we're going to be making all the
decisions for you.
Free will is just an illusion for you.
Have a nice life, goodbye.
And he believes her.
Well, shiny lab, lab coat, all the rest.
And as you might imagine he becomes a little negligent in
his habits.
And a little bit aggressive.
And indulges in some of his worst behaviors.
And soon enough he gets in trouble with the law.
Gets hauled into court, charged with a
fairly serious crime.
He says, but Your Honor.
I don't have free will.
How do know that?
Well my neurosurgeon told me, free will is just an illusion.
It feels like I'm making these decisions, but I'm not really.
So they call the neurosurgeon to the stand
and ask her to testify.
And under oath she says, Yeah, that's what I told him.
But it was just a joke.
I was just messing with his head.
Now how many of you agree that she did, not just a mean
thing, a socially irresponsible thing.
But she did very serious harm to this guy.
With her language, she did nonsurgically she claimed to
do surgically.
She disabled him as a moral agent.
By just convincing him that he didn't have free will.
Now I ask my neuroscientist friends, aren't you just doing
the same thing wholesale?
When you write these articles saying, we have learned in
neuroscience that we don't have free will.
The reason they say that turns out to be very bad reasons.
Having to do with, well everything in your brain is
determined by what happens in your brain
a little bit before.
Yeah, of course.
So what?
That doesn't show you don't have free will.
They have a very simplistic idea of what free will is.
But it's the idea that would actually ground moral
responsibility.
If you didn't have it you wouldn't have moral
responsibility.
But they're not showing that at all.
And I think that's irresponsible.
Now you might say, well, if your brain is determined, and
then you don't have free will.
How many of you are inclined to think, if every event in
your brain is determined then you don't have free will?
The general run of people think that's sort of, even
obviously true.
So here's a little intuition pump to
dissuade you from that.
Two lotteries.
In lottery A the winning ticket is chosen after the
tickets are sold.
Most lotteries are like that.
In Lottery B the winning ticket is chosen before the
tickets are sold.
And the stub is secretly put in a safe for safekeeping
until the tickets are sold.
Now how many of you think that you're a fool to buy a ticket
in Lottery B?
That only lottery A gives you any chance of winning?
How many think that they're even?
Yeah.
And so does most of the rest of the world.
When you put in these terms.
They recognize that, no, you've got a
chance either way.
And Publishers Clearing House actually says, you may
already have won.
People see that they still have a chance if the winning
tickets are chosen in advance.
But now think, if determinism is true then all your lottery
tickets were chosen before you were born.
At the Big ***, if you like.
And sort of put in on a little secret safe in your body.
And as life goes on, whenever you need a coin flip, whenever
you need a little chaos, you just take the next one out of
the envelope and you win some you lose some.
That's just as free as if there's a quantum miracle that
happens right then and there.
So what if they're chosen first.
Well somebody might say, well you see, if all your lottery
tickets are chosen before you're born, you're determined
to have good luck on some occasions
and bad luck on others.
You're determined.
Well that's true even if indeterminism is true.
Even if indeterminism is true, you're determined on some
occasions to win and some occasions to lose, you just
don't know which.
Fairness isn't always winning.
You've got to accept the fact that there's luck.
And that luck is consistent with free will.
But a lot of philosophers haven't seen it.
A very good one, David Wiggins, once wrote some years
ago about what he called the cosmic unfairness of
determinism.
And you can see, or many people think, yeah right, it's
a bummer if determinism is true.
But I want to ask what about the cosmic unfairness of
indeterminism?
Just as bad.
You know.
Life isn't always a picnic.
That doesn't mean that we can't be responsible agents.
Being a responsible agent, having free will in the sense
that matters, has zip to do with whether physics is
deterministic or indeterministic.
Or whether your brain is deterministic or
indeterministic.
Has to do with the software you've got in your brain.
It has to do with the competence that you have had
installed in your brain since you were a little kid.
If you were trained with a good, moral upbringing then
you make the right decisions at the right
time most the time.
And you are, in general, competent enough so when you
screw up, you are held responsible.
Why?
Because you're good enough to be responsible.
That's a long story.
I'm not going to say more about that.
Time to quit this talk.
So the book is sort of a tapas book.
There are 77 chapters.
And some of them are just a page.
Some of them there are maybe as much as 10 pages long.
And they each introduce a tool making it autonomous.
I tried to make them all portable and accessible.
So my recommendation is to try them out.
Take them for a spin.
And above all, turn all the knobs to see how they work.
Thanks for coming.
[APPLAUSE]
DANIEL DENNETT: Questions?
AUDIENCE: Hello.
So intelligent design is better in some
domains than others.
We can design a very nice mobile phone or computer.
When it comes to an economy, well, some may argue, but so
far we've yet to show we can
intelligently design an economy.
Do you have, well two questions.
First of all, what you think the profit is of things which
are not very tractable to intelligent design?
And do you think eventually all problems we'll have the
right tools in order to design everything?
DANIEL DENNETT: No, I think there's a trade-off which
other people have described pretty well.
Natural selection is a great engine.
And it feeds on chaos and so-called randomness.
But it works, remember, by trial and error.
And there's an awful lot of errors.
And it's a very wasteful product.
Very wasteful way of doing business.
You've got to have a lot of failures.
And so that's what happens, let's say, with an economy.
And to that extent, that grubby old social Darwinist
line about how the survival of the fittest in the economy has
some truth to it.
But that doesn't mean that we should simply honor that and
make that a moral principle.
Because I think we have very good reasons to think that an
unfettered, completely chaotic, laissez-faire,
dog-eat-dog economy is not a good way to go.
Too much harm can accrue.
You can get these winner take all situations
which can be very bad.
You know about winner take all algorithms.
You know about opponent processes in the brain.
In the work that I've been doing on how brains are
organized, I think, first of all, they are computers.
But they're very, very different from laptops.
Precisely because you've got 100 million
or 200 million neurons.
Each one is like a little termite.
They actually have their own agendas.
There's no two alike.
And the way software works in the brain is by creating
coalitions and getting them to work together.
And unity is an achievement, it's possible.
But it isn't built in from the top down.
And once we think about the brain being organized that
way, we should think that's probably economies are best
built in that way too.
Without too much top down central control, but with some
policing to protect us from the excesses that otherwise
would accrue.
AUDIENCE: Could you give your opinion on Newcomb's paradox?
DANIEL DENNETT: Yeah.
Do you know what Newcomb's paradox is?
This is the one boxer or two boxer.
OK, so you've got.
First I'm going to have to explain it.
See if I can get it right.
So you're on a sort of quiz show and you're given the
following weird choice.
You could be a one boxer, or you can be a two boxer.
In one box there's a $1,000.
In the other there's $1 million versus $0.
It would be hard enough if you had to choose either box A or
box B. Would you take a chance on $1 million, or would you go
for the sure $1,000?
But that's not your choice.
Your choice is either to go for a box of B which has
either got the $1 million or nothing.
Or to go from boxes.
Got it?
Well one line of research, one line of thinking, says well,
dummy, of course you go for both boxes.
That way at least you get $1,000 and maybe get a $1
million and $1,000.
That dominates decision tree, draw it out.
That's the one you should do.
And so it would be, except for one thing.
The MC of the program tells you, oh, by the way, you're
the 5,000th contestant on the show.
And we're very good at predicting whether you're
going to be a one boxer or a two boxer.
And if we predict that you're going to make the two boxer
choice, then there's only the $1,000 in box A. We never put
the $1 million.
We only put the $1 million in when we predict that you're
going to go for one box.
And do you want to know our track record?
It's almost perfect.
They might add, well you know, one poor devil went for box A
and it was empty, or box B and it was empty.
And one lucky devil out of 500 or whatever, went for both and
got both rewards.
But in general it's a very strong statistical
correlation.
So now I ask you.
How many of you go for both boxes?
And how many of you go for just the one box?
What's nice about Newcomb's paradox is that it ain't
obvious at all.
And an awful lot of ink has been spilled trying to get
clear about this.
And I have, myself, not spilled any ink because I
still don't have a firm view on it.
I tend to think there must be something a little bit screwy
about the setting up of this thought experiment.
But it's very hard to show that.
And I highly recommend it to you.
There's a lot of good things have been written about it.
So I'm sorry I can't answer the question.
I go back and forth.
Some days I'm a one boxer, some days I'm a two boxer.
AUDIENCE: I wanted to ask you about
determinism versus free will.
Since you were talking about earlier.
And, in particular, I think Hofstadter would talk about a
lover's disparity there.
That determinism is kind of a totally different question
then free will.
I wondered if you could talk about that.
DANIEL DENNETT: Yeah Doug and I have talked about
this over the years.
And Doug disagrees with me.
It's one of the few points where he and I have a serious
disagreement.
And he's actually quite pessimistic about this and
thinks that the sense of free will that really matters is
one that we can't have because of determinism.
And I've done my level best in several books and many hours
to convince him otherwise without success yet.
Unlike the fellow in my tale, it doesn't lead him to behave
in a irresponsible way.
So there's at least to could be said for it.
I think that Doug--
here's what I think the problem is.
That the term free will.
Where does it come from?
It comes from every day.
It comes from the law and from tradition.
It's not a scientific term.
It's a term you get from what's called the manifest
image, in philosophy.
The world of colors and dollars and
baseball and so forth.
Not the world of neurons and proteins and atoms, or for
that matter, scientific laws.
When you look at it's home territory, free will, you can
see it does two jobs.
And they both seem to be definitive or constitutive.
On the one hand, it seems definitive that well free will
means uncaused, it means it's free.
There's no sufficient causal antecedent.
For many people I think that's just obvious that's what free
will means.
But the other thing it means is it's the standard condition
for being morally competent.
For instance, how many of you believe yourself qualified,
eligible, to buy a car to take out a loan?
Yeah.
And you sign the paperwork of your own free will.
Now there are people who don't have that competence.
They're senile or they're too young.
There's something you've got that they don't have.
And it is your very condition of your living a free life as
a citizen and conducting business, making promises that
count and so forth.
I think that's the best notion of free will.
That's the one that matters.
And I think that notion has nothing to do with
determinism.
For instance, I don't think your competence to sign a
contract would be disparaged in the slightest if
determinism were true.
So I have argued for years there's just no reason to
prefer that your brain be deterministic over
indeterministic.
No reason at all.
No good reason.
I've recently had to back off that a little bit by a very
persistent lover of indeterminism and free will,
who finally showed me that there was, by my own lights, a
condition where I would agree with them.
I would wish for indeterminism in that case.
And I'll tell you what it is.
You all know the game rock-paper-scissors?
Well the best way to play that game, of
course, is to play random.
It's very hard to fake randomness or to put together
a random series.
If you're random you give you give your opponent no pattern
to track at all.
So they can't do any better than play random against you.
You'll break even.
So since that's the case, you might just as well memorize a
few hundred entries in a table of random numbers and use that
to guide your choices.
You'll do fine as long as you don't let the other
guy see your list.
Which would be terrible.
So now I can tell you, here's the condition under which I
would hope that indeterminism is true.
If I were playing rock-paper-scissors for high
stakes with God.
I don't really anticipate that happening in the future, so I
am quite content to live out my life in a deterministic
world and I'll take responsibility for my acts,
even though they are, for all I know or care, determined by
the kazillion things that happened both before
and after my birth.
Yes.
AUDIENCE: Is consciousness an epiphenomenon of advanced
biological systems and tends to disappear as you look back
across the Tree of Life?
Or is it a fundamental dimension to nature and
reality like the dimensions of space and time?
DANIEL DENNETT: OK you said or.
This is sort of a rathering.
You didn't say rather, but there was a sort
of rathering there.
But I'll take it on because I do not think that it is a
fundamental category that splits the universe in two.
I think that the idea that it is deeply rooted in culture.
And I suppose it might be right.
But I think that consciousness is like reproduction,
metabolism, and self repair, healing, blood clotting.
It's a very interestingly complex and deeply recursive
biological phenomenon.
It is not in the interesting sense, and there's many senses
of consciousness.
In the sense that we in this room are conscious, which is
the anchoring that's home base from our own experience and
our reading of literature.
We read novels and things like that.
We have the sense of what a conscious life is like.
I think we're the only creatures that have that kind
of conscious life.
I don't think dogs have it.
I don't think dolphins have it.
They're sentient.
They're awake.
They feel pain.
But they don't have most of what makes human consciousness
so important.
And I think that phenomena is, in the end, software.
And I think it has to be installed.
So it is an implication, in my view, and
don't try this at home.
Don't test this.
That if a human baby could be raised entirely without any
interaction with other human beings and
cut off from language.
If it could grow to adulthood without benefit of all the
thinking tools that would normally be installed in the
course of just growing up in a social community.
I think that human being would have a mind so profoundly
different from ours that we would all pretty well agree he
or she wasn't really conscious.
Now that may be a very eccentric notion of
consciousness.
But if you're wanting to know, do I think that, if you look
for a better notion of consciousness or one that
isn't so eccentric.
And you start thinking about what it is.
The capacity to experience color and flavor and to have
qualia, as philosophers say.
I think that's all pretty badly confused.
And I've tried to say how and why over the years.
But part of what makes it confusing is it people want
consciousness to be, to use a technical term,
supercalifragili sticexpialidocious.
And if you're not talking about something that is
supercalifragili sticexpialidocious then you're
not talking about consciousness.
And there's a quote my friend Lee Siegel, he's a philsopher
and a magician.
A very fine stage magician.
Wrote a wonderful book on Indian street magic called,
"Net of Magic."
And there's a passage in there I love to quote.
He says, I tell my friends I'm writing a book on magic and
they ask, do you mean real magic?
And by real magic they mean magical.
They mean thaumaturgical acts and so forth.
And he says, no.
Not real magic.
Sleight of hand.
Stage magic.
And then he goes on, he says, in other words, the kind of
magic you can actually do is not real magic.
Real magic is a kind that doesn't exist.
Well what I've come to realize, my whole career, I've
been confronted with people.
Because what I want to deal with are topics that a lot of
people think of as real magic.
Free will and consciousness in particular.
And if you don't mean something that as one
proponent of it says, every free
decision is a little miracle.
If you don't think that that's what free will is then you're
not talking about free will.
Not real free will.
You're talking about some cheesy substitute.
And I'm all for this cheesy substitutes, I'm afraid.
I think they're quite good enough.
And science can explain why we have varieties of free will
that are worth wanting, even if they aren't real free will.
They'll secure our lives as meaningful and
responsible and so forth.
And we can have consciousness even if it isn't something
that sunders the universe in twain in this
very special way.
Sorry it's such a long answer, but you got me going.
AUDIENCE: What do you think about
utilitarianism and moral reasoning?
DANIEL DENNETT: I'm not a utilitarian, or at least not
an ordinary utilitarian.
I don't that any.
I'm a consequentialist.
That is to say, I think the consequences matter.
I'm not one who takes a strictly
deontological view of ethics.
Using some technical terms.
I want to try to get away from them.
This is an issue that's very close to the service in work
that I'm doing right now on free will and punishment.
Because I say, no, we don't.
I notoriously said in my last book on free will, I don't
want to live, no in an article on free will.
I don't want to live in a world without punishment.
And these people jumped all over me and said oh, you dirty
retributivist, you.
I said, no, I'm not a retributivist.
Kant was the great retributivist.
And if you want to know what bullet he bit, he said, if we
know the world was going to end tomorrow, one of our
highest duties would be to go into the prisons and execute
the people on death row.
By that he wanted he wanted us to think that punishment was
in itself good.
That a crime unpunished is a an offense to
nature or God or something.
And that's what a retributivist is one who
thinks that punishment is a good in itself.
I don't.
But I think it's a very foundational good for a very
simple reason.
I want to live in a world in which there are promises.
And you can't really have promises unless there's
consequences for not keeping your promise.
And those consequences are not rehabilitation.
They're not reform.
They're punishment.
And it's OK.
And we should all be willing to say, no, I'd rather take my
punishment.
I'd rather be a free citizen in a free society.
I accept that if I do something really wrong I am
eligible for punishment and that's OK.
I think the world we live in, when we have a system of law
and order with punishment, I think our current system of
punishment is obscene and should be completely reformed.
But I don't think we should eliminate punishment as some
of the current crop of cognitive scientists and
neuroscientists are saying.
I don't think they've look closely at the last time this
was considered a really wise move.
And that was back in the Stalinist era when people
weren't punished, they were treated.
I don't think we want to be treated.
I think we want to be punished.
Because then there's rules.
And you can serve your time and get out and
be a citizen again.
An that dire note, I'll stop.
AUDIENCE: There was a decision this week about patents that
touched on some deeply philosophical issues.
The question was whether software should be patentable.
And one of the opinions, which unfortunately, didn't gather
majority support was that basically things you do in
software are like natural laws.
They're tools that everybody can use.
And some particular arrangements with these tools
should not be patentable.
Other judges felt that if you have a computer and it's
programmed with software, that's a new device.
And devices are clearly patentable.
Do you have some wisdom to offer us on that?
DANIEL DENNETT: Well I thought about it in the past.
And I've thought about the difference between patent and
copyright and trade secret.
And actually I'm trying to think about this in quite
fundamental terms.
Because I think that what we need to develop is the sense
of information, which is distinct from Shannon-Weaver
information.
Which depends on Shannon-Weaver information,
but is semantic information.
It's cognitive information.
And, to me, the best way of thinking about this is that
it's useful information.
It is information defined as that which permits the
transmission of adaptations for one purpose, one utility
or another.
And then under that view then, of course, software, I think
intuitively when we're measuring whether software
what counts as theft, we have to look at the utility.
And if it's the utility that's been stolen, then we should
have penalties for that.
That should be discouraged.
The trouble is that copyright law also involves this idea.
It's a good question.
And I'm not surprised that the Justice--
Is this at the Supreme Court?
AUDIENCE: It's the Court of Appeals for the Federal
Circuit that handles patent cases.
DANIEL DENNETT: I want to know more about it.
I'm not surprised that they are divided.
I think it's a tough issue.
But obviously when I talk about apps, as tools for
thinking, I am thinking of these as mechanisms that have
been designed.
And so on that ground I would tend to come down on the side
that you can patent it.
Thank you.
[APPLAUSE]
[MUSIC PLAYING]