How to pronounce "sledgehammer"

Every day we face issues like climate change

or the safety of vaccines

where we have to answer questions whose answers

rely heavily on scientific information.

Scientists tell us that the world is warming.

Scientists tell us that vaccines are safe.

But how do we know if they are right?

Why should be believe the science?

The fact is, many of us actually don't believe the science.

Public opinion polls consistently show

that significant proportions of the American people

don't believe the climate is warming due to human activities,

don't think that there is evolution by natural selection,

and aren't persuaded by the safety of vaccines.

So why should we believe the science?

Well, scientists don't like talking about science as a matter of belief.

In fact, they would contrast science with faith,

and they would say belief is the domain of faith.

And faith is a separate thing apart and distinct from science.

Indeed they would say religion is based on faith

or maybe the calculus of Pascal's wager.

Blaise Pascal was a 17th-century mathematician

who tried to bring scientific reasoning to the question of

whether or not he should believe in God,

and his wager went like this:

Well, if God doesn't exist

but I decide to believe in him

nothing much is really lost.

Maybe a few hours on Sunday.

(Laughter)

But if he does exist and I don't believe in him,

then I'm in deep trouble.

And so Pascal said, we'd better believe in God.

Or as one of my college professors said,

"He clutched for the handrail of faith."

He made that leap of faith

leaving science and rationalism behind.

Now the fact is though, for most of us,

most scientific claims are a leap of faith.

We can't really judge scientific claims for ourselves in most cases.

And indeed this is actually true for most scientists as well

outside of their own specialties.

So if you think about it, a geologist can't tell you

whether a vaccine is safe.

Most chemists are not experts in evolutionary theory.

A physicist cannot tell you,

despite the claims of some of them,

whether or not tobacco causes cancer.

So, if even scientists themselves

have to make a leap of faith

outside their own fields,

then why do they accept the claims of other scientists?

Why do they believe each other's claims?

And should we believe those claims?

So what I'd like to argue is yes, we should,

but not for the reason that most of us think.

Most of us were taught in school that the reason we should

believe in science is because of the scientific method.

We were taught that scientists follow a method

and that this method guarantees

the truth of their claims.

The method that most of us were taught in school,

we can call it the textbook method,

is the hypothetical deductive method.

According to the standard model, the textbook model,

scientists develop hypotheses, they deduce

the consequences of those hypotheses,

and then they go out into the world and they say,

"Okay, well are those consequences true?"

Can we observe them taking place in the natural world?

And if they are true, then the scientists say,

"Great, we know the hypothesis is correct."

So there are many famous examples in the history

of science of scientists doing exactly this.

One of the most famous examples

comes from the work of Albert Einstein.

When Einstein developed the theory of general relativity,

one of the consequences of his theory

was that space-time wasn't just an empty void

but that it actually had a fabric.

And that that fabric was bent

in the presence of massive objects like the sun.

So if this theory were true then it meant that light

as it passed the sun

should actually be bent around it.

That was a pretty startling prediction

and it took a few years before scientists

were able to test it

but they did test it in 1919,

and lo and behold it turned out to be true.

Starlight actually does bend as it travels around the sun.

This was a huge confirmation of the theory.

It was considered proof of the truth

of this radical new idea,

and it was written up in many newspapers

around the globe.

Now, sometimes this theory or this model

is referred to as the deductive-nomological model,

mainly because academics like to make things complicated.

But also because in the ideal case, it's about laws.

So nomological means having to do with laws.

And in the ideal case, the hypothesis isn't just an idea:

ideally, it is a law of nature.

Why does it matter that it is a law of nature?

Because if it is a law, it can't be broken.

If it's a law then it will always be true

in all times and all places

no matter what the circumstances are.

And all of you know of at least one example of a famous law:

Einstein's famous equation, E=MC2,

which tells us what the relationship is

between energy and mass.

And that relationship is true no matter what.

Now, it turns out, though, that there are several problems with this model.

The main problem is that it's wrong.

It's just not true. (Laughter)

And I'm going to talk about three reasons why it's wrong.

So the first reason is a logical reason.

It's the problem of the fallacy of affirming the consequent.

So that's another fancy, academic way of saying

that false theories can make true predictions.

So just because the prediction comes true

doesn't actually logically prove that the theory is correct.

And I have a good example of that too, again from the history of science.

This is a picture of the Ptolemaic universe

with the Earth at the center of the universe

and the sun and the planets going around it.

The Ptolemaic model was believed

by many very smart people for many centuries.

Well, why?

Well the answer is because it made lots of predictions that came true.

The Ptolemaic system enabled astronomers

to make accurate predictions of the motions of the planet,

in fact more accurate predictions at first

than the Copernican theory which we now would say is true.

So that's one problem with the textbook model.

A second problem is a practical problem,

and it's the problem of auxiliary hypotheses.

Auxiliary hypotheses are assumptions

that scientists are making

that they may or may not even be aware that they're making.

So an important example of this

comes from the Copernican model,

which ultimately replaced the Ptolemaic system.

So when Nicolaus Copernicus said,

actually the Earth is not the center of the universe,

the sun is the center of the solar system,

the Earth moves around the sun.

Scientists said, well okay, Nicolaus, if that's true

we ought to be able to detect the motion

of the Earth around the sun.

And so this slide here illustrates a concept

known as stellar parallax.

And astronomers said, if the Earth is moving

and we look at a prominent star, let's say, Sirius --

well I know I'm in Manhattan so you guys can't see the stars,

but imagine you're out in the country, imagine you chose that rural life —

and we look at a star in December, we see that star

against the backdrop of distant stars.

If we now make the same observation six months later

when the Earth has moved to this position in June,

we look at that same star and we see it against a different backdrop.

That difference, that angular difference, is the stellar parallax.

So this is a prediction that the Copernican model makes.

Astronomers looked for the stellar parallax

and they found nothing, nothing at all.

And many people argued that this proved that the Copernican model was false.

So what happened?

Well, in hindsight we can say that astronomers were making

two auxiliary hypotheses, both of which

we would now say were incorrect.

The first was an assumption about the size of the Earth's orbit.

Astronomers were assuming that the Earth's orbit was large

relative to the distance to the stars.

Today we would draw the picture more like this,

this comes from NASA,

and you see the Earth's orbit is actually quite small.

In fact, it's actually much smaller even than shown here.

The stellar parallax therefore,

is very small and actually very hard to detect.

And that leads to the second reason

why the prediction didn't work,

because scientists were also assuming

that the telescopes they had were sensitive enough

to detect the parallax.

And that turned out not to be true.

It wasn't until the 19th century

that scientists were able to detect

the stellar parallax.

So, there's a third problem as well.

The third problem is simply a factual problem,

that a lot of science doesn't fit the textbook model.

A lot of science isn't deductive at all,

it's actually inductive.

And by that we mean that scientists don't necessarily

start with theories and hypotheses,

often they just start with observations

of stuff going on in the world.

And the most famous example of that is one of the most

famous scientists who ever lived, Charles Darwin.

When Darwin went out as a young man on the voyage of the Beagle,

he didn't have a hypothesis, he didn't have a theory.

He just knew that he wanted to have a career as a scientist

and he started to collect data.

Mainly he knew that he hated medicine

because the sight of blood made him sick so

he had to have an alternative career path.

So he started collecting data.

And he collected many things, including his famous finches.

When he collected these finches, he threw them in a bag

and he had no idea what they meant.

Many years later back in London,

Darwin looked at his data again and began

to develop an explanation,

and that explanation was the theory of natural selection.

Besides inductive science,

scientists also often participate in modeling.

One of the things scientists want to do in life

is to explain the causes of things.

And how do we do that?

Well, one way you can do it is to build a model

that tests an idea.

So this is a picture of Henry Cadell,

who was a Scottish geologist in the 19th century.

You can tell he's Scottish because he's wearing

a deerstalker cap and Wellington boots.

(Laughter)

And Cadell wanted to answer the question,

how are mountains formed?

And one of the things he had observed

is that if you look at mountains like the Appalachians,

you often find that the rocks in them

are folded,

and they're folded in a particular way,

which suggested to him

that they were actually being compressed from the side.

And this idea would later play a major role

in discussions of continental drift.

So he built this model, this crazy contraption

with levers and wood, and here's his wheelbarrow,

buckets, a big sledgehammer.

I don't know why he's got the Wellington boots.

Maybe it's going to rain.

And he created this physical model in order

to demonstrate that you could, in fact, create

patterns in rocks, or at least, in this case, in mud,

that looked a lot like mountains

if you compressed them from the side.

So it was an argument about the cause of mountains.

Nowadays, most scientists prefer to work inside,

so they don't build physical models so much

as to make computer simulations.

But a computer simulation is a kind of a model.

It's a model that's made with mathematics,

and like the physical models of the 19th century,

it's very important for thinking about causes.

So one of the big questions to do with climate change,

we have tremendous amounts of evidence

that the Earth is warming up.

This slide here, the black line shows

the measurements that scientists have taken

for the last 150 years

showing that the Earth's temperature

has steadily increased,

and you can see in particular that in the last 50 years

there's been this dramatic increase

of nearly one degree centigrade,

or almost two degrees Fahrenheit.

So what, though, is driving that change?

How can we know what's causing

the observed warming?

Well, scientists can model it

using a computer simulation.

So this diagram illustrates a computer simulation

that has looked at all the different factors

that we know can influence the Earth's climate,

so sulfate particles from air pollution,

volcanic dust from volcanic eruptions,

changes in solar radiation,

and, of course, greenhouse gases.

And they asked the question,

what set of variables put into a model

will reproduce what we actually see in real life?

So here is the real life in black.

Here's the model in this light gray,

and the answer is

a model that includes, it's the answer E on that SAT,

all of the above.

The only way you can reproduce

the observed temperature measurements

is with all of these things put together,

including greenhouse gases,

and in particular you can see that the increase

in greenhouse gases tracks

this very dramatic increase in temperature

over the last 50 years.

And so this is why climate scientists say

it's not just that we know that climate change is happening,

we know that greenhouse gases are a major part

of the reason why.

So now because there all these different things

that scientists do,

the philosopher Paul Feyerabend famously said,

"The only principle in science

that doesn't inhibit progress is: anything goes."

Now this quotation has often been taken out of context,

because Feyerabend was not actually saying

that in science anything goes.

What he was saying was,

actually the full quotation is,

"If you press me to say

what is the method of science,

I would have to say: anything goes."

What he was trying to say

is that scientists do a lot of different things.

Scientists are creative.

But then this pushes the question back:

If scientists don't use a single method,

then how do they decide

what's right and what's wrong?

And who judges?

And the answer is, scientists judge,

and they judge by judging evidence.

Scientists collect evidence in many different ways,

but however they collect it,

they have to subject it to scrutiny.

And this led the sociologist Robert Merton

to focus on this question of how scientists

scrutinize data and evidence,

and he said they do it in a way he called

"organized skepticism."

And by that he meant it's organized

because they do it collectively,

they do it as a group,

and skepticism, because they do it from a position

of distrust.

That is to say, the burden of proof

is on the person with a novel claim.

And in this sense, science is intrinsically conservative.

It's quite hard to persuade the scientific community

to say, "Yes, we know something, this is true."

So despite the popularity of the concept

of paradigm shifts,

what we find is that actually,

really major changes in scientific thinking

are relatively rare in the history of science.

So finally that brings us to one more idea:

If scientists judge evidence collectively,

this has led historians to focus on the question

of consensus,

and to say that at the end of the day,

what science is,

what scientific knowledge is,

is the consensus of the scientific experts

who through this process of organized scrutiny,

collective scrutiny,

have judged the evidence

and come to a conclusion about it,

either yea or nay.

So we can think of scientific knowledge

as a consensus of experts.

We can also think of science as being

a kind of a jury,

except it's a very special kind of jury.

It's not a jury of your peers,

it's a jury of geeks.

It's a jury of men and women with Ph.D.s,

and unlike a conventional jury,

which has only two choices,

guilty or not guilty,

the scientific jury actually has a number of choices.

Scientists can say yes, something's true.

Scientists can say no, it's false.

Or, they can say, well it might be true

but we need to work more and collect more evidence.

Or, they can say it might be true,

but we don't know how to answer the question

and we're going to put it aside

and maybe we'll come back to it later.

That's what scientists call "intractable."

But this leads us to one final problem:

If science is what scientists say it is,

then isn't that just an appeal to authority?

And weren't we all taught in school

that the appeal to authority is a logical fallacy?

Well, here's the paradox of modern science,

the paradox of the conclusion I think historians

and philosophers and sociologists have come to,

that actually science is the appeal to authority,

but it's not the authority of the individual,

no matter how smart that individual is,

like Plato or Socrates or Einstein.

It's the authority of the collective community.

You can think of it is a kind of wisdom of the crowd,

but a very special kind of crowd.

Science does appeal to authority,

but it's not based on any individual,

no matter how smart that individual may be.

It's based on the collective wisdom,

the collective knowledge, the collective work,

of all of the scientists who have worked

on a particular problem.

Scientists have a kind of culture of collective distrust,

this "show me" culture,

illustrated by this nice woman here

showing her colleagues her evidence.

Of course, these people don't really look like scientists,

because they're much too happy.

(Laughter)

Okay, so that brings me to my final point.

Most of us get up in the morning.

Most of us trust our cars.

Well, see, now I'm thinking, I'm in Manhattan,

this is a bad analogy,

but most Americans who don't live in Manhattan

get up in the morning and get in their cars

and turn on that ignition, and their cars work,

and they work incredibly well.

The modern automobile hardly ever breaks down.

So why is that? Why do cars work so well?

It's not because of the genius of Henry Ford

or Karl Benz or even Elon Musk.

It's because the modern automobile

is the product of more than 100 years of work

by hundreds and thousands

and tens of thousands of people.

The modern automobile is the product

of the collected work and wisdom and experience

of every man and woman who has ever worked

on a car,

and the reliability of the technology is the result

of that accumulated effort.

We benefit not just from the genius of Benz

and Ford and Musk

but from the collective intelligence and hard work

of all of the people who have worked

on the modern car.

And the same is true of science,

only science is even older.

Our basis for trust in science is actually the same

as our basis in trust in technology,

and the same as our basis for trust in anything,

namely, experience.

But it shouldn't be blind trust

any more than we would have blind trust in anything.

Our trust in science, like science itself,

should be based on evidence,

and that means that scientists

have to become better communicators.

They have to explain to us not just what they know

but how they know it,

and it means that we have to become better listeners.

Thank you very much.

(Applause)