Acquiring Language

How do you think kids learn to talk?

>> By imitating their mothers and
the people around them.

>> I guess their parents teach them.

People around them.

>> Children learn language from their
parents or whosever rearing them or

raising them.

>> Yes, I think that a baby would
imitate what he hears, yes.

[MUSIC]

>> This is part two of a series

on the human language.

[MUSIC]

>> Where's Cookie Monster?

Hi, Cookie Monster.

>> How do children acquire language
without seeming to learn it?

>> Bye, Big Bird.

>> In advance of experience,
the child is already

equipped with an understanding of
the basic structure of any human language.

>> [SOUND]

[MUSIC]

[SOUND]
>> Show me your [FOREIGN].

>> At the age of three,
there are many things a child can't do.

They can't perform mathematical
operations like multiplication and

division, or even subtraction.

>> But it only takes a short
amount of time to get a kid

to be a member of a language community,
such that he can infer

reliably from the noises that people
make to their states of mind.

>> Papa.

>> How on earth does he do it?

>> Papa, papa.

>> Either it's there at birth or
he has to learn it.

Now do birds teach their young to fly?

Do mothers teach their
children how to speak?

Well, I really don't know whether
birds teach their young to fly or

not, though I suspect not.

But I'm certain that mothers don't
teach their children to talk.

[MUSIC]

[NOISE]
>> This part of our series is about

a great mystery.

How do children acquire language
without seeming to learn it?

How do they note so many things with so
little life experience to go on,

a problem posed by Plato, 2000 years ago.

[MUSIC]

How do children know how to walk and
run around, how to pick up things with

their fingers, how to use grammar
effortlessly without lessons?

How do they do it?

Approaches to Plato's problem
have changed in recent years.

[MUSIC]

And linguists credit one man
with newly raising the issue and

profoundly influencing their thoughts.

Noam Chomsky.

>> Well, there is a general view
that learning language is just like

solving any other problem, that we have
certain general ways of solving problems,

call them mechanisms or
general intelligence.

We apply them to this task,
to that task, to the other task.

It's the same mechanisms.

One of the problems is acquiring language.

There's another approach,
which suggests that, in fact,

the brain is like every other
system in the biological world.

That is, it's highly differentiated
into subsystems of special design and

structure, and that one of these
happens to have a special design for

acquiring language.

>> Chomsky has argued,
to my mind convincingly,

that many of the processes of mental
computation involved in the normal

use of language are inherited, you're born
with them, they're in your genetic makeup.

Coded in the DNA the way it's
coded in the DNA of a human being

that is going to develop
arms instead of wings.

>> And then if it's, as it were,

coded in the genes it's not surprising
that we're real good at it.

It's like we're all good at having hair or
something, having two arms,

as Chomsky likes to say.

>> Walking is presumably
encoded in your DNA.

It's part of the innate program
of human beings in their

development that they're going to
learn to walk and not learn to fly.

[MUSIC]

Perhaps not learn to climb very well.

[MUSIC]

But they'll learn to
walk terrifically well.

[MUSIC]

>> We're designed to walk.

It's actually an open question
to what extent the child has to

be in an environment of walking
people in order to walk.

But that we are designed
to walk is certain and

that we are taught to walk is impossible.

And pretty much the same with language.

Nobody is taught language.

In fact you can't prevent
a child from learning it.

It's probably a mistake to
even use the word learning

in connection with language acquisition.

As far as we understand it,

it has very much the properties
of normal physical growth.

>> So much of the,
what is called controversy,

about Chomsky's approach appears to arise

from people's belief that they have to man
this barricade on one side or the other.

Either language is built in as birdsong,
let us say, is built in, or

it is wholly learned from exposure to
specific properties of the environment.

Why can't it be both?

In fact, I take that this is
the question of modern linguistics.

How much of language does a child have
to learn and how much is built in?

Jill de Villiers,

at Smith College, can show that children
know more than we think they do.

>> Here's a story about a boy who
loved to climb trees in the forest.

But one afternoon, he slipped and
fell to the ground.

That night, when he had a bath,
he found a big bruise on his arm.

He said to his dad, I must have hurt
myself when I fell this afternoon.

When did the boy say he hurt himself?

[MUSIC]

[MUSIC]

>> Climbed up a tree.

>> Right, he did.

>> When he was taking a bath.

>> He was taking a bath.

[MUSIC]

>> In the bathtub.

>> In the bathtub, mm-hm.

>> When he fell down from the tree.

>> That's right.

When did the boy say he hurt himself?

>> When he fell.

>> When he fell this afternoon.

>> [LAUGH].
>> Any other possibilities?

>> In the bathtub.

>> In the bathtub,
there are two possibilities.

Well, he said it in the bathtub, and
he hurt himself in the afternoon.

That night when he had a bath,
he found a big bruise on his arm.

He said to his dad, I must have hurt
myself when I fell this afternoon.

When did the boy say how he hurt himself?

>> In the bathtub.

>> When he was bathed.

>> When he was in the bath, you're right.

[MUSIC]

>> So where's the other answer?

When did the boy say how he hurt himself?

Why isn't it when he fell
out the tree that afternoon?

>> Give us the two events side by side?

>> When did the boy say he hurt himself?

When did the boy say how he hurt himself?

The presence of that middle question
how [SOUND] seems to block one of

the interpretations.

This is not the kind of sentence
anybody has ever taught them about.

They haven't had lessons sitting down
with their parents saying no, no,

no you misinterpreted what I meant.

I meant this.

I would think that the child would have
to already have some kind of knowledge of

grammar and syntactical structure.

>> This is the claim.

>> When did the boy say
how he hurt himself?

>> When he was taking a bath.

>> Mm-hm.

>> It's the imitation theory
versus the innateness theory.

>> [SOUND] [SOUND] Now, listen,

I am going to teach you to
say something very important.

It might be the most important
word you'll ever learn, okay?

And that word is Ernie, okay,
go ahead, say it, say Ernie.

>> [SOUND]
>> People often ask,

what's the big problem about
a child learning grammar?

Doesn't the child just imitate what he or
she hears?

Get reinforced one way or another and
end up knowing the language?

>> Okay, listen, here's Er-nie,

Er-nie, Er-nie, Er-nie, Er-nie,

[SOUND], now you just say, Ernie, okay?

>> [SOUND]
>> Our common sense

theory about how children learn to talk
is that they listen to their parents and

they imitate their parents.

>> But every child, like every adult,
can produce brand new sentences that child

has never heard before and
never produced before.

>> I hate you, mommy.

Now, come on,
you didn't learn this from your mother.

>> You just have to listen to
a three-year-old for a few minutes to

realize that they're not simply imitating
what they hear from their parents.

I've heard children say things like,
stop giggling me, or

my teacher holded the baby rabbits.

Or my nose is crying,
when a child's nose is running, or

I'm barefoot all over,
it's a very funny sort of imitation.

>> Now, Ernestine, look, watch.

Er-nie, Ernie.

>> [SOUND]
>> Well.

[SOUND]
>> Hi, Ernie.

>> Hi, Bert.

>> Bert!

[MUSIC]

>> If we don't learn by imitation,
how do we learn?

[MUSIC]

And it's one of the linguists' strongest
arguments that acquiring language is

different from learning.

Because we don't seem to learn
language the same way we learn other

difficult things.

[MUSIC]

>> With learning how to ride a bicycle, we
all remember being told how to get on it,

falling off, we even had training wheels.

But we never give training wheels
when we're learning language.

In fact,
what we know is that when mothers try and

help their children learn
language by making it simpler.

The children systematically ignore the
information that's being given to them.

>> I don't know why he says pajamas,
I mean, jamamas,

because I say pajamas, right?

>> Jamamas.

>> Jamamas?

What about pajamas?

Can you say pajamas?

>> Jamamas.

>> We come to cherish
our children's errors.

Rare as they are, they're so cute.

>> When my little girl would say she
wanted to make wee-wee, I would say,

the word is urinate.

Urinate, I said, the word is urinate.

Urinate, dear.

Right, I'm a Nate.

[MUSIC]

>> We know that some mothers
correct their children, but

we also know that not all mothers do.

>> [LAUGH]
>> And

we know that all children learn to speak.

>> It doesn't seem to matter how many
times I correct him, he still says,

jamamas, he likes that word.

>> Jamamas.

[MUSIC]

>> Mothers are important,
because they provide a lot of the data.

>> [SOUND]
>> Kid makes a noise, and

the mother expands on it,
and that sort of thing, but

children don't copy
what's done around them.

They acquire language by being surrounded
by it, immersed in it all the time.

The philosopher, Wittgenstein,

said that children acquire speech
by playing the language game.

A game in which mothers often
seem to imitate their children.

>> [SOUND]
>> [SOUND]

>> [SOUND]

>> [SOUND]

>> [SOUND]

>> Ready!

>> Okay.

>> Sammy is three and a half years old.

How much grammar does he know?

>> We know the Cookie Monster
eats cookies and cakes, right?

Ask the rat what he thinks.

[SOUND]
>> What do you think Cookie Monster eats?

>> What?

>> What do you think Cookie Monster eats?

>> I think Cookie Monster eats,
maybe pizza?

Goodness.

>> Try again.

>> So maybe he likes to eat cookies.

>> Cookies and what else?

>> Ice cream?

Well, that's a tough one, Sammy.

>> I'll give you a guess or
I'll give you a hint.

>> Okay.
>> It starts with a K.

It starts with a K.

>> What do you think Cookie Monster eats?

>> With a K, maybe cake.

>> Right!

>> Okay.

>> Cookie.

>> What do you think Cookie Monster eats?

It's really rather remarkable that such a
young child can produce such a complicated

sentence.

[MUSIC]
This is a complex sentence that has

one sentence inside another.

Cookie Monster eats something,
inside the larger sentence,

you think Cookie Monster eats something.

Then it's been changed into a question,
and

the way it's changed into a question is,
something has changed into what.

And then what is displaced from
the very end of this long sentence,

to the very beginning.

[MUSIC]

The child was able to do it unerringly.

[MUSIC]

Now it has been thought for
a while that children take five,

eight, 10, 12 years to learn their syntax.

To learn how to produce
such complicated sentences.

But experiments like these seem to
indicate that child was able to produce

this very complicated sentence.

At an age when he has
difficulty tying his shoes.

>> Just do the bow part for
me, would you, Sammy?

>> One more?

>> Yeah, one more bow.

>> Any four-year-old can pull it
off with almost the facility that

any 34-year-old can.

It's quite a remarkable achievement, and

one which comes intuitively,
naturally, unconsciously.

Nothing troubles us about it.

>> No matter how young
you investigate them,

they seem to get all of these
fundamental syntactic things right.

>> And nobody can teach that to the child.

I think that's one of Chomsky's
most powerful arguments.

[MUSIC]
This is something the child has to figure

on his own, or on her own.

[MUSIC]

>> Chomsky is often misunderstood as
claiming that all of language is innate.

Now I would go so far as to say
that most of language is innate.

That is, much of what you know
when you know a language,

you couldn't possibly have learned.

Because you never had evidence for
it, you never had training in it,

but surely you don't inherit all of it.

>> Certainly nobody is saying
that French is innate or

Spanish is innate, I mean,
there is no argument of the sort.

There is a sense in which
language is obviously learned

from specific facts in
the surrounding environment.

>> The environment
certainly has an effect.

So, for example, I'm talking English,
I'm not talking Japanese.

And that's because I grew up
in Philadelphia, Pennsylvania,

not in Northeastern Tokyo.

>> So experience is certainly relevant.

What Chomsky has shown is that
>> A handsome prince came,

and a little girl came out,

and laid on the couch.

A big girl came out, and
she had a thing around here, and

she took it off and laid it on a girl.

>> The child is able to say sentences
that he or she has never heard before.

>> And then she-
>> Well, how could this be done?

>> And then she walked around and

she showed a handsome prince and
then snow came down.

>> You could not debate that,
being so many sentences.

That the child has just heard all of
these things, and memorized them.

>> And then we went in and
watched a movie.

>> There is a traditional semi-answer to
this, and that is, we do it by analogy.

New sentences are like the ones
we've heard before, and

that's how we understand them.

But let's have a look at that.

>> So suppose we assume the child
has heard the sentence,

I painted the red barn.

So now by analogy, the child can say,
I painted a blue barn.

That's exactly the kind of theory we want,
you hear a sample,

and you extend it to all of
the new cases by similarity.

That's the right theory,
in addition to, I painted a red barn,

you may also hear the sentence,
I painted a barn red.

So it looks as if you can
take those last two words and

switch them around in their order.

That sounds good, so now you want to
extend this to the case of seeing.

Because now you want to look at barns,
instead of painting them.

So you have heard I saw a red barn, so

now you try a new sentence.

I saw a barn red.

I saw a red barn.

I saw a barn red.

[SOUND]
>> Alarm, something's gone wrong.

This is an analogy, but
the analogy didn't work.

It's not a sentence event.

[NOISE]
>> Even from these simple examples, we can

see it warrants the concepts like analogy
you're not going to do very much work.

In fact, there is no, no one has
ever proposed a concept of analogy.

That it doesn't break down at once,
under investigation.

In fact, these examples show us that some
kind of mental computation is going on,

that it's providing rather
surprising results.

[SOUND]
>> Suppose that a child has

learned to understand the word, eat.

He knows that somebody does the eating,
something gets eaten.

He understands, John ate an apple.

He's now faced with John ate.

Well, the child understands that, he or

she know that John ate,
means John ate something, or other.

>> Taro ate his sandwich.

But suppose that he hadn't eaten his
sandwich, suppose that he ate his shoe.

Taro ate his shoe, Taro ate his hat.

It's a funny thing to do, but

a perfectly normal thing to say,
if that's what Taro did.

But now, you can't say Taro ate.

Taro ate, means he ate his sandwich,
or ate breakfast, or lunch, or

something normal.

Not his shoe, or his hat, or his words.

Taro ate, but he didn't eat his shoe.

How do I know that,
how does any speaker of English know that?

That Taro ate,
doesn't mean that he ate his shoe.

>> Now let's extend that,
by analogy to new utterances.

>> Now suppose Taro's a farmer.

>> Suppose the same child hears
the sentence, John grows tomatoes,

and knows that growing,
is something that a person can do.

And something gets grown, so
that's just like, John eats an apple.

And now supposed tomatoes is dropped so
we have, John grows.

Well, now the sentence does not mean,
John grows something or other, but

I don't know what.

In fact, John grows means
something totally different.

>> Doesn't mean the same thing at all.

>> John is undergoing some
sort of a development.

[NOISE] And I hear,
the analogy is wildly broken.

But yet we all do this instantaneously,
without training, without experience.

And in a way,

which is based on principles that
are quite common to the human species.

And underly our very
understanding of language.

>> Pretend you're sick, okay?

>> [COUGH] I gotta call the doctor.

>> I'll call the doctor.

[MUSIC]
>> [SOUND]

>> We might ask how old the child

has to be,

before it begins to appreciate something
of the grammar of his native tongue.

>> Okay.
>> Told you to give him some medicine.

>> Come on.
>> When does he know about ideas like,

subjects of the sentence and
objects of a sentence?

When does he know the difference between,

the horse kicked the cow and
the cow kicked the horse?

But the subject is the one
who does the kicking, and

the object is the one that got kicked?

>> Experiments at Temple University,
try to learn how early the child knows.

>> Monster?

>> Where's Big Bird?

Find Big Bird.

Where's Cookie Monster?

Find Cookie Monster.

Where's Big Bird?

>> We present two films, one on the left
screen, and one on the right screen.

>> Mama.

>> Then we simply ask the child
through a centralized speaker.

>> Where's Cookie Monster
washing Big Bird?

Find Cookie Monster washing Big Bird.

>> So, the question behind all our
studies is, will the child look more

at the screen that matches
the language that they're hearing.

>> Look,
Big Bird's feeding Cookie Monster.

Find Big Bird feeding Cookie Monster.

[SOUND] Ma.

Ma.

Ma. Ma. >> There's Cookie Monster.

>> The remarkable thing,
is that some of these children.

Who are only 16 months old, and

who have only two words in
their productive vocabularies.

Nonetheless, are understanding
the order of the information,

as it comes into our sentences.

That is, that Cookie Monster
is doing the action, and

he's in the first
position in the sentence.

And Big Bird is receiving the action, and

he's in the second
position in the sentence.

>> Here's Big Bird,
tickling Cookie Monster.

>> Word order is one of the two devices,

that languages all over
the world seem to use.

To map the objects and events,
that are going on in the world.

Which is a very important part of grammar,
and

may in fact be telling us,
that these young children have syntax.

>> [SOUND]
>> I think it

makes very good sense to think of language
as essentially, an organ of the mind.

And bear in mind,
when we use the word mind,

we're simply talking about
the brain at some abstract level.

Now remember,
when a child learns a language,

the child is basically
creating the language.

The language is growing in a child's mind.

>> Does this apply to words?

Surely, words don't exist
in the child's mind.

>> [INAUDIBLE]
>> Yet, we learn words so easily.

>> [INAUDIBLE]
>> You might wonder if our

brains give us some special help there,
too.

>> [INAUDIBLE]
>> Problem is how the child learns

the meanings of words.

>> Circle.

[FOREIGN]
>> Seems like a simple problem, and

I suppose it is a simple problem
from the point of view of a child.

The mother points to
something in the world,

let's say the car that's going by there,
and says, car.

So the child says, aha!

The word car in English must mean car.

The travelers that can't
be the whole story.

[MUSIC]

>> When my son was little,
we had a dog called NuNu.

This is his successor Freud,
and NuNu was a wonderful dog.

Very different from Freud.

She had hair all over her face, and

two black shiny nostrils stuck
out from the middle of it.

Nicholas loved NuNu, he liked to poke
his fingers in her nostrils, and

as linguists, we were convinced
that NuNu would be his first word.

It had the right structure to it,
two reduplicate it's syllables, and

sure enough,
NuNu was Nicholas' first word.

But on the very same day
he pointed to a picture

of a golden cocker spaniel in
a photograph, and he said, Nunu.

Very different from our own dog.

So perhaps the word meant dog in general.

But no, the very next day he pointed to
a black and white cow and said, NuNu.

So we thought maybe it
means animal in general.

But then we were in a shop and he saw
some pink furry slippers like these ones,

and he felt the slipper,
felt how soft it was, and he said NuNu.

Now what does the word mean?

But then we were in a restaurant,
and our salad was served and

Nicholas took one look at it and
said, NuNu.

Then he actually ate one of the olives.

>> [SOUND]
>> The question was when he said NuNu,

what did he mean?

[MUSIC]

>> The trick in learning
a word meaning is not so

much using it to apply to that
which you saw it applied to, but

using it in the future
to apply to new things.

So, the child who was introduced
to Fido and told this is a dog,

not only has to use the word
dog to refer to Fido, but

he has to know it applies to Rex.

It applies to Spot.

It doesn't apply to Felix.

>> How does the dog know
that's another dog?

>> Words after all have meaning,
they're signs.

So maybe what we want to say is that
a word is something that stands for

a concept, but then we have another
question, what's a concept?

Take a word like house,
that seems pretty simple.

You ask a five year old to draw a picture
of the house, and most of them give you

the same square thing with a chimney, and
a couple of doors, and that kind of stuff.

So you say, okay that's a house, but then
if you think the concept, if you start

worrying about yourself, all right,
what did somebody have in their head?

Is the meaning of a house in my head
different from the meaning of a house in

that Chinese persons
head who lives in a cave?

>> [SOUND] A simple, homely item.

A clothes pin.

[MUSIC]

What else does a clothes pin,
after you learn this word?

Well,

[MUSIC]

How about that?

[MUSIC]

Is that a clothes pin?

In some ways it's very much more
like this object than that one is.

[MUSIC]

To understand what the problem
of word meaning is,

you have to understand how
the child picks out a category.

A category of things which
are relevantly alike.

[MUSIC]

>> A dog is alive.

Is the concept of alive the same for
everyone?

Do our concepts change as we grow up?

>> What does the word alive mean?

Are dogs alive?

>> Yes.

>> How come?

>> Because there's eyes, and
mouth, and teeth, and tongue.

>> And there's wagging tail.

>> Wagging tail, okay,
>> Bark.

>> How about a worm?

Is a worm alive?

>> Yes.

>> Eww.

>> Does a worm have legs?

>> No.
>> No.

>> Does it have teeth?

>> No.
>> No.

>> Well then how come a worm is alive?

>> Because it moves.

>> Okay, well then how about a car?

Is a car alive?

>> No.
>> Yes, because it moving.

>> What do you think?

>> I think it's alive too.

>> What is a person referring to when he
or she says a word, and how do we know?

At Harvard University
philosopher WVO Quine

posed the question known
as the gavagai problem.

Suppose you find yourself someplace
you don't know anything about.

Next to you there's a man, and
you don't know what language he speaks.

At that moment a rabbit
crosses the horizon and

the man says to you gavagai.

Gavagai.

What does he mean?

>> What do you think I mean?

>> Gavagai.

Blue rabbit.

[LAUGH]
>> It's a rabbit.

>> If you pretend that you can't
understand my language at all and

I point up there and I say, gavagai,
what do you think I mean by that?

>> A rabbit.

>> Look at the rabbit.

>> How do you know I don't mean fur?

>> It could be animal,
a physical object, a hippity hop.

>> Gavagai.

>> Rabbit.

>> The only evidence is that it's
appropriate to announce gavagai in

the presence of a rabbit, but
the rabbit is present only when

rabbit parts are present,
provided they're undetached.

Rabbit parts.

Furthermore, a rabbit is
present only when the abstract

attribute of rabbithood is manifested, but

the abstract property of rabbithood
is quite another thing for rabbits.

>> How do you know that
I didn't mean ears?

>> All of these ideas apparently
don't come to child's mind.

What comes to his mind is rabbit, and

the question is how this could be?

>> For a child to learn a word meaning,

it would help to have certain
inherited assumptions.

What assumptions?

>> Have you ever seen one of these before?

>> What might an inborn assumption be?

>> Well, this here is called a flemic.

>> How you do that?

>> It's called a flemic.

>> Coud I try one of them?

>> Now, what do you think this is called?

>> Flemic.

Wait this one's open and
that one's closed.

So that can't be a Flemic.

What do you think it is?

>> Closed.

>> [LAUGH] Yeah.

>> Children are biased learners.

They're not open-mindly considering
all the possible hypothesis about

what a word could mean and
waiting for the evidence to come in.

One of the assumptions that they make
about what words could mean is they

start out expecting object labels
to refer to the whole object.

[MUSIC]

So when someone points to an object and
says, see the dog.

Can you put the dog in there?

They start out expecting that a label
will refer to the whole object,

not to its part, not to its substance, not
to its color, shape, size, and so forth.

Where's the ball?

Put the ball in there.

>> They expect the label that they hear
to refer to the object as a whole.

>> Okay.

>> What other assumptions could they have?

>> Hand me the spud.

>> This one.

>> Why do you think that's the spider?

>> Well I think because it looks like it.

>> It looks like a spider,
that's a good answer.

>> Children expect objects to have one and
only one label.

Children expect an object to have one and
only one name.

[SOUND]
>> If a child already knows that that's

called a flimmick and he's asked to point
to the spud you can easily pick this

object out because he knows that can't
be a spud because it's a flimmick.

And if it's a flimmick, it's not a spud.

>> Flimmicks.

>> Flimmicks.

>> Flimmicks.

>> Words seem to be learned one by one.

Just what the common sense idea might be.

>> Animal.

>> You learn language by learning an item.

Perhaps by explicit instruction and
in effect memorize them.

>> Animal.

>> To a large extent, I would say
words might well be learned that way.

>> Animal.

>> Sentences, surely,
aren't learned that way.

>> My birthday.

>> When you want to use a sentence,
you're surely creating something new.

And when you understand the sentence
that somebody else utters,

you are really participating
in an act of creation.

>> Each summer, I go to Puerto Rico.

And my grandma lives there.

But before she used to live in Miami and

she had a dog and it was named Pepper.

>> You've got a finite number of words.

>> And it ran away.

>> You've got a very
small number of rules and

together you can use them to make
up an infinite number of sentences.

That's the system.

[MUSIC]

>> And it's a toy store and
an ice cream store.

>> The richness of the concepts
that we employ, and

the minimal character,
the evidence on which we've derived them.

Essentially leaves no alternative but to
believe that these concepts are available

prior to experience and we're simply
selecting them out of a store.

And that means that in effect
we're sort of born with it.

[MUSIC]

>> For the claim to be right that the
human species has a common human language,

wouldn't we have to find it being used,
even in far off places?

[MUSIC]

Like Papua New Guinea?

>> [FOREIGN]
>> Here on this island,

many tribes with many cultures have lived
apart from the rest of the world for eons.

>> [FOREIGN]
>> Are their languages anything like ours?

[MUSIC]

>> This region, Melanesia,
contains 0.1% of the world's

population and
about 20% of the world's languages.

There's nothing like this country
in the world linguistically.

This country has 750 different languages
spoken by three million people.

Some of them as different
as English from French.

Some of them as different
as French from Chinese.

>> [FOREIGN]

>> [FOREIGN]

>> The Manu people have never written

their language.

>> They've never invented the wheel,

but they speak a language that
has as complicated a grammar and

discourse structure as any
other language in the world.

>> Carl Whitehead of Manchester England
has lived here with his family for

ten years.

To complete his study of
the Manga language he thinks

may take him another ten years.

>> [FOREIGN]
>> The single

most remarkable feature of the Manga
language is the verb system.

Most verbs can have anywhere between
2,000 and 3,000 different forms,

but compared to English where
a verb can have up to five forms.

It has a vague complex model system,
but as where English talks of I may go,

I can go, I should go, I will go, Manga
collapses all that into a single word.

It has about 14 different ways of
referring to an event that hasn't

happened yet, but that could or that will
or should happen sometime in the future.

>> [SOUND]
>> That's the genius of Manga.

>> [SOUND]
>> When

we start looking across linguistically,
we find great universals.

We don't find dozens or

hundreds of different ways of
building the human language.

We find a very small set of possibilities.

One possibility has to do with
the position of the verb.

You can put the verb at the beginning,
the middle, or the end.

>> [FOREIGN]

>> The Manga verb [FOREIGN] can have many

different meanings,
depending on its context.

It doesn't occur by itself.

So by itself, it is meaningless.

It's the noun in combination with the verb
[FOREIGN] that has specific meaning.

So, for example,
[FOREIGN] means she is washing.

[FOREIGN] He is crying.

[FOREIGN]
He's rubbing mud on his face.

[FOREIGN]
He's building a fence.

[FOREIGN] It is raining.

[MUSIC]

And there's a couple of poetic
ones that I really like.

The first one is referring
to an aura around the moon.

[FOREIGN], which literally means
the moon is building a fence.

[FOREIGN]
>> [FOREIGN]

>> And the other one refers to a special

kind of sunset when the whole ground
lights up with a bright orange glow.

And the expression is, [FOREIGN].

Which literally means,
the sun is painting the ground red.

[SOUND]

>> [FOREIGN]

>> The reasonable first guess from all

observers is that languages
are utterly different from each other.

>> There are 5,000, maybe more than
5,000 languages spoken in the world.

And if you're studying a foreign language,

what you notice is how different that
language is from the one that you speak.

Of course,
what you notice are the differences.

But in fact,
what Chomsky noticed is that these 5,000

languages of the world are really very,
very similar.

>> [FOREIGN]
>> [FOREIGN]

>> [FOREIGN]

>> You could even consider the 5,000 so

called languages that
are spoken on this planet,

to be all dialects of one language,
human language.

[FOREIGN]
>> There

was a tremendous change in
the research program for

linguistics right around
the turn of the 20th Century.

And the leading figure in this change

was a Swiss linguist named
Ferdinand de Saussure.

Saussure thought of languages
very much like chess games.

That each language has a set of pieces,
namely,

the building blocks of the language,
the words, the sounds.

And they were put together
by different rules.

>> [FOREIGN]
>> Different languages, of course,

have different rules.

>> [FOREIGN]
>> Interestingly,

Saussure discovered every
language has rules.

There are no languages that are so
primitive, for example,

that they don't have rules at all.

>> There's no such thing as a primitive
language or primitive people, for

that matter.

>> And that was an important discovery.

Because before the beginning of
the 20th century people talked

freely about primitive languages.

What Saussure and people who worked around
him said was, every langauage has rules.

[MUSIC]

>> Every language?

Everywhere?

[MUSIC]

In the Bering Sea on St. Lawrence island,

is one of the last hold outs of
a language called Siberian Yupik Eskimo,

once thought to have complex rules
completely different from other languages.

Yet, examined carefully, even Eskimo turns
out to be less different than it sounds.

One scholar of the language is
a native of the island, Darlene Orr.

>> The people here on St Lawrence Island
speak Siberian Yupik.

And it is also spoken on the other side,
and that's in the Siberian part.

It's the only indigenous language

to the old world and to the new,

spoken by non colonial peoples.

>> [FOREIGN]
>> [SOUND]

>> [FOREIGN]

>> The language,

the verb might go in one place.

In another language this sound
might pattern this way, but

another language in a different way.

But every language has rules.

>> [FOREIGN]
>> [FOREIGN]

>> [FOREIGN].

And that means,
were you trying to call him?

>> There are two major ground plans,
two great options in building a language.

Relying on the order of words to convey
the meaning of your overall thought, or

changing the endings of the words one
by one, and then shuffling them around.

>> [FOREIGN]
>> The reason this word looks so long,

is because Siberian Yupik Eskimo
relies heavily

on suffixation using suffixes in a word.

There are no prefixes or
infixes in Siberian Yupik.

And linguists speculate that this is
probably the most inflected language in

the world, which means that it could
have multitudes of suffixes for endings.

This part means being acted upon.

This much here means without.

And the last part means us.

So the whole word means,

without us being sprayed

upon by water when you're

traveling in a boat.

[MUSIC]

>> Children seem to love rules.

And it's this kind of phenomenal.

The fact that children seem to love
rule making that has led many people to

propose that there's something
special about the learning of grammar.

The children essentially
come pre programmed,

perhaps with some innate ideas about
the forms that languages will take.

[MUSIC]

>> In advance of experience,
the child is already equipped with

an understanding of the basic
structure of any human language.

So we know, even in advance of inquiry,

that there's going to be a fundamental
invariant core to language.

>> [FOREIGN]
>> Linguists have come

to use the term Universal Grammar.

It's a notion that there's some
underlying set of characteristics

that are true of all languages,
all over the world.

>> All human languages have
something that is sort of nouny, and

something that is sort of verby.

They all do.
All human languages have a way to make

things negative.

>> No way, Jose.

>> No way, Jose.

>> All human languages have
a way to ask a question.

>> If I'm asking a question,
I'll lower my eye brows.

Where are you from?

What's your name?

Who is that person?

>> All human languages have a way to
indicate the difference between just one

and more than one.

>> In the Eskimo languages you have
not only the singular, for example,

[FOREIGN], one rabbit, [FOREIGN], two
rabbits, [FOREIGN], three or more rabbits.

And so it goes, each language has a list
of obligatory distinctions, male, female,

definite, indefinite, singular,
plural, past, present.

This is the stock of categories that the
human mind uses to schematize experience.

>> And Chomsky asked the question,
why is this?

Why are languages so similar?

Why are they all cut from the same mold?

And his answer was-
>> There are fixed invariant principles,

fixed invariant structural principles,
which are simply part of the human

biological endowment, and that determine
what counts as a human language.

>> It's because the human
brain is pre wired to accept

only certain kinds of languages.

And that the grammatical properties
of the languages of the world

have those properties that they do because
the human mind has those properties.

>> Those things, which are true of
all languages, are the candidates for

what the child comes
into the world knowing

about the nature of the language
to which he is being exposed.

>> The child might very
well have a plan for

what is a possible rule
in a human language.

So languages can have verbs then objects,
or objects then verbs, but

those are two possibilities that every
language has one or the other of.

And the child can simply worry about which
of those two versions his language has.

What he's got to pick up are particular
versions of the rules that everyone else

in the community is using.

[MUSIC]

>> You have this one, you have this one,
you have this one, you have this one.

[MUSIC]
Everybody got their pictures?

What did your person do
before they went to school?

>> He drive to school.

>> He did what?

>> They drive.

>> Drive to school?

>> Children are designed in such a way to
look for rules in the data around them.

They're very good at finding
rules that are there, and

they're very good at even over
generalizing these rules.

[MUSIC]

>> What's going overhead?

>> Goose.

Geeses.

>> We find everywhere that children take
the irregular patterns of their language

and try to make them as
regular as possible.

So just like English speaking
children will say things like,

two foots instead of two feet, or
it breaked instead of it broke.

Children in every language will fill in
the missing gaps, fill in the errors and

try to make the language follow a system.

They have a clear sense of system.

>> We find that deaf children of deaf
parents begin to learn first words and

then begin to learn the grammar.

At the same age as hearing
children learn spoken language.

And in so doing, we find they make
the same kinds of mistakes or

over-generalizations.

Showing that, in fact, they're extracting
out the rules of the language.

So if you want to sign, this is duck,
if you want to sign two ducks,

you'd probably say, two ducks.

[MUSIC]

But, a deaf baby instead will
over-generalize and sign two ducks,

like this.

>> [SOUND] So they are looking for
some deep principles,

they follow these deep principles.

If the language chooses to violate
those principles now and then,

the children seem to say,
so much for the language.

>> Ruth says that they're foots,
I say that they're feet.

What do you say they are?

>> I say they're foots.

>> But there's certain kinds of mistakes
that children never seem to make.

Even when it seems very reasonable
that they might make a mistake.

>> Children will be able to ask questions
like, what did you eat your eggs with?

But no child has ever asked the question,
what did you eat your eggs and?

Even though that's a straightforward
extention of, I ate ham and eggs.

Children will hear, I baked a cake for
Mary, I baked Mary a cake.

They'll hear, I painted the house for
six hours, they'll never say,

I painted six hours the house.

Why don't children make these errors?

They're perfectly logical.

It's very hard to figure out on the basis
of what the various sentences look like

why the child wouldn't
make these obvious leaps.

If you can say,
what did you eat eggs with?

Why can't you say,
what did you eat eggs and?

>> When we imagine a reasonable sort
of mistake for a child to make, but

never find a child making it.

We assume that the mistake
would violate some principle or

rule of universal grammar.

Universal grammar is what the child
already knew and doesn't have to learn.

>> What's in here?

>> Ask him what he thinks.

>> What do you think what's in here?

[MUSIC]

>> Well,
that sounded to me like monster heads.

>> No!

>> [LAUGH] Silly, silly.

I know, M&M's.

>> What do you think what's in here?

>> Great!

>> Great.

>> An adult would say,
what do you think is in here?

But Sam said.

>> What do you think what's in here?

>> It's not just a random error, it's not
putting the words together in any old way.

Rather it's a mistake that's actually
a rule of a number of other languages.

For example,
in certain dialects of German,

exactly the way you would ask this
question is by repeating the what.

It just doesn't happen to be
the syntactic rule of English.

>> What do you think was in here?

>> He's not born knowing
the grammar of English, but

what he knows is that languages
will fall into these classes.

And he won't step outside
those classes and

create some novel language that
couldn't be a human language.

[MUSIC]

>> [SOUND] I'm still holding me.

Stop, stop!

Let go of me!

No!

>> No, no, no.

>> No!

>> If we weren't so familiar with them,
we'd realize that three-year-olds

are among the most exotic
creatures on the planet.

We usually don't think of them as
terribly competent intellectually.

We don't let them drive or vote,
it's hard to teach them long division.

I think about what it feels as
them when they learn how to talk.

He hasn't gotten any lessons, all he's
done is listen to other people talk.

He makes very few errors and unlearns them
without having to have been corrected.

There are countless, logical,
tempting errors that he never makes.

And in less than two years,
he's developed the ability to

express an infinite variety of
brand new thoughts into words.

In a way that other people know
exactly what he's thinking.

[MUSIC]

>> There are a number of
quite striking conclusions.

General conclusions that we've
arrived at as a result of this study,

even with our present partial knowledge.

One is a kind of awe at the intricacy and

complexity of the mind and its resources.

The structures of language that
the child automatically constructs on

the basis of extremely sparse evidence.

A second conclusion is that,
these intricate,

subtle resources must be
shared across the species.

Since the systems of thought and
expression that the child's mind

develops are only barely hinted
at by the experience available.

They must be deriving
from the child's mind.

But the child's mind is the same,

whether the child is going to be
exposed to one culture or another.

>> And correspondingly,
it must be that at their core.

The cultural products,
including language, are very similar,

and, in fact, rooted,
ultimately, in human biology.

>> [SOUND]
>> From what we've observed in children

learning different
languages around the world.

It seems to us that
the capacity to learn language

is deeply ingrained in us as a species.

Just as the capacity to walk,
to grasp objects, to recognize faces.

We don't find any serious differences in
children growing up in congested urban

slums.

In isolated mountain villages, or
in privileged suburban villas.

[MUSIC]

>> Children who are abused, and
mistreated, and unmotivated,

if they can't hear, language
filters out through their fingers.

Almost no matter what the circumstance is,
the language bubbles up.

[MUSIC]

>> It seems inescapable.

Children aren't learning language
the way they learn most other things.

They're not taught, they're not corrected,
they don't even think about it,

they just do it.

>> [SOUND]

[MUSIC]

>> Next time in our series
on the human language,

if language is biological,
how did it evolve?

Program three, the human language evolves,
with and without words.

[MUSIC]