Transcript: Human Language Series

Transcript: Human Language Series. Part II.

"Acquiring the Human Language "Playing the

Language Game"

Central Question:

How do children acquire language without seeming to

learn it?

Slobin:

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

- Now do birds their young ones to fly?

- Do mothers teach their children to speak?

- NO! To both questions.

Birds do not teach their children to fly.

Mothers don't teach their children to talk.

This video is about a great mystery:

1. How do children acquire language without seeming

to learn it.

2. How do they do so many things with so little life

experience to go on.

Plato's Problem:

This problem is posed by Plato 2000 years ago.

Chomsky:

There are basically two approaches to Plato's Problem:

1st Approach: Learning language is just like solving any

other type of problem. Problem solving is a mechanism

within our intelligence and one of the problems is

acquiring language.

2nd Approach: The brain is like every other system in the

biological world. It is subdivided in highly differentiated

subsystems of special design and structure and one of

these subsystems has a special design in form for

language.

Lasnik: Talking - like walking - is encoded in the DNA.

Acquiring language is part of our genetic make up.

Chomsky: Nobody is taught language. In fact, you can't

prevent the child from learning it. It has very much to do

with physical growth.

Gleitman: Argues against polarization:

Important question: WHY CAN'T IT BE BOTH?

It is the most central question of modern linguistics:

How much of language does a child have to learn and

how much is built in?

SYNTAX IN LANGUAGE ACQUISITION

Jill de Villiers (Smith College)

Experiment 1: Ambiguity Experiment

or: When-did-the-boy-say-he-hurt-himself Experiment

Design:

Jill told the following story to each child:

1. story of the boy climbing the tree in the forest.

2. then one day, the boy slipped and fell

3. In the bathtub at night, he had a big bruise.

4. Boy says to his father:

"I must have hurt myself when I fell this afternoon."

Question of the experimenter to the child:

"When did the boy say he hurt himself?"

Girl1: Climbing the tree.

Boy1: When he was taking a bath.

Girl2: In the bathtub

Boy2: when he fell

Conclusion: There are two possible answers!!!

Experiment 2: Unambiguous Sentence

or: When-did-the-boy-say-how-he-hurt-himself

Experiment

Similar Design as in Experiment 1, only now the

following question is asked:

Q: "When did the boy say how he hurt himself?"

The children answer with: "in the bathtub."

Conclusion: Only one possible answer

Crucial question:

Where did the second interpretation go?

Why is the sentence in experiment 1 ambiguous and the

sentence in experiment 2 unambiguous? What is the

difference between (1) and (2):

(1) When did the boy say he hurt himself?

(2) When did the boy say how he hurt himself?

Answer: In sentence (2), the middle question introduced

by "how" blocks one of the interpretations that are

possible in sentence (1). "How he hurt himself forms an

island".

A Puzzling Difference

(1) When did the boy say he hurt himself?

versus

(2) When did the boy say how he hurt himself?

Explanation:

Sentence (1) has two "D(eep)"-Structures:

(1) a. The boy said ..................

b. The boy said...................

Sentence (2) has only one D-Structure:

(2) The boy said...................

Correlating to the D-Structures, we can deduce the

Movement operations taking place in each case:

SHORT WH-MOVEMENT:

Movement from D-Structure to S-Structure in sentence

(1a): The boy said WHEN/ he hurt himself?

When did the boy say ___ he hurt himself?

SHORT WH-MOVEMENT

LONG WH-MOVEMENT:

(1b): The boy said he hurt himself WHEN/

When did the boy say he hurt himself ___?

LONG WH-MOVEMENT

SHORT WH-MOVEMENT:

Movement from D-Structure to S-Structure in sentence:

(2) a. The boy said WHEN HOW he hurt himself /

When did the boy say ___ HOW he hurt himself?

SHORT WH-MOVEMENT

LONG WH-MOVEMENT OF WHEN across HOW is

IMPOSSIBLE!

The D-Structure for (2) which we are looking for would

be something like (2b). But WHEN cannot across HOW!

(2) b. The boy said HOW he hurt himself when/

When did the boy say HOW he hurt himself ___?

XXXXX

What is the relevance of this experiment?

1. These are not the kind of sentences anybody had ever

taught the child about.

2. Therefore, the experiment shows: a child must have

some kind of knowledge of syntactic structure.

Imitation vs. Innateness Theory:

Scene1: Ernie tries to get the baby to imitate his name.

Lasnik:

What's the big problem about a child learning language.

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

Gets reinforced and learns the language?

Pinker: It's the common sense idea: children listen to

their parents and they imitate their language.

Lasnik: But how can we explain that every child and

adult can produce brand-new sentences.

Gleitman: "I hate you, mama." Now, come on, you

haven't learned this from your mother.

Pinker: Listen to a 3 year old. They are not simply

imitating what they hear from their parents:

1. Stop giggling me.

2. My teacher holded the baby rabbit.

3. My nose is crying.

4. I am barefoot all over.

This is a very funny sort of imitation. Why?

Q: What are possible corrections?

Scene2: Ernie wants to teach the little baby to say Ernie.

In the end he gives up. Bird comes in. Ernie says: Hi,

Bird. The baby imitates [bd].

Problem: If we don't learn by imitation - how do we

learn?

Linguist's strongest argument: Acquiring language is

different from learning other things, because we don't

seem to learn languages the same way how we learn

other difficult things - like playing the trumpet, riding a

bicycle, etc.

Wittgenstein said: Children acquire speech by

playing the language game. A game where mothers

seem to imitate their children.

Experiment 1 with Sam:

Sam: (31/2 years old)

Linguist 1 to Sammy:

- "We know that cookie monster likes cookies and cakes.

Ask the rat what he thinks."

Sam to rat:

"What do you think cookie monster eats?"

Linguist2 answers for rat:

"I think he eats maybe pizza?"

"Maybe cookies?"

Sam: "Cookies and what else?"

"Ice-cream?"

Sam: "I'll give you a guess. I'll give you a hint. It's

spelled with a [k]."

Sam: "What do you think m m cookie monster eats?"

Lasnik: It is rather remarkable that such a young child

can produce such a difficult sentence!

It is a complex sentence that has one sentence inside

another.

Step 1: Find the D-Structure:

[S1 You think [S2 Cookie Monster eats (something)]]

Step 2: change the sentence into a question.

The way it's changed into a question:

a. "something" is changed into "what"

[S1 You think [S2 Cookie Monster eats what]]

b. and then "what" is displaced from the very end of the

sentence to the very beginning. (Inversion)

[S1 What You think [S2 Cookie Monster eats ____]]

Q: What is missing?

[S1 What do You think [S2 Cookie Monster eats ____]]

'do'-insertion

The child was able to do it unerringly.

How long does it take children to figure out their syntax?

It has been though that it takes children 10-12 years to

figure out their syntax. But experiments have shown that

a child was able to produce a very complicated sentence

when they are about 3 years of age.

Fodor: Nobody can teach language.

Most of it is innate, but not all of it is.

Gleitman: Certainly "French" is not innate. Or Spanish!

There is a sense in which language is obviously learned

from specific facts in the surrounding environment.

Chomsky: The environment certainly has an effect. I am

talking English, I am not talking Chinese - that's because

I grew up in Philadelphia.

Girl: Tells story. The child is able to say sentences he or

she has never heard before.

Chomsky: There is a traditional semi-answer to this. And

this is - we do it by analogy.

ANALOGY Theory:

1. Give an example of where analogy seems to hold:

1. Substituting one color word for another

(1) I painted the red barn.

(2) I painted the blue barn.

2. Switching the last two words in a sentence:

(3) I painted the barn blue.

(Interpretation: I painted the barn and as a result of it

became blue.)

Gleitman: It looks as if you can take those last two words

and switch them around in their order.

(4) a. - a red barn

b. - a barn red.

2. Give an example of where analogy breaks down:

Case 1:

Now, let's assume you want to extend this to the case of

seeing. Now you have to look at barns instead of

painting.

(5) a. I saw a red barn.

b. *I saw a barn red.

Something's gone wrong. This is an analogy that didn't

work.

Chomsky: A concept of analogy breaks down under

investigation at once.

Case 2:

The example in (1) means "Taro ate a sandwich, lunch,

dinner."

(1) Taro ate.

It does not mean: his shoe, his hat, his words.

Lasnik: How does any speaker of English know that

"Taro ate" means that Taro ate something, but not "Taro

ate his shoes."

But now look at (2):

(2) a. John grows tomatoes

b. John grows _______.

(2b) doesn't mean John grows something or other. It

means: John undergoes some sort of development.

The analogy breaks down. The analogy is wildly

broken.

Task: Give two examples of where the analogy theory

breaks down.

We all do that instantaneously, without training, without

experience. And in a way that is quite common to the

human species.

Gleitman: When does the child understand/appreciate

the grammar of his native tongue?

When does he know about ideas as the subject and the

object of the sentence?

When does he know the difference between:

(1) a. The horse kicked the cow. vs.

b. The cow kicked the horse.

Define the subject and the object in (1):

The subject is the one who does the kicking and the

object is the one who got kicked.

What is the subject and the object in (2):

(2) The cow was kicked by the horse.

Subject_________

Object _________

Experiment at the Temple University:

Design: Showing two films simultaneously.

Scene 1:

a. Where is Cookie Monster washing Big Bird?

b. Find Cookie Monster washing Big Bird.

Commentator: The question behind all our studies is,

will the child look more at the screen that matches the

language that they are hearing.

Scene 2:

a. Look, Big Bird is feeding Cookie Monster.

b. Find Big Bird feeding Cookie Monster.

Boy: points and says ma/ma/ma/ - looks for confirmation

Commentator: The remarkable thing is that some of

these children that are 16 months and have only 2 words

in their productive vocabulary nonetheless understand the

order of information as it comes in our senses.

Q: What do you think about this test procedure?

Result: Word order is a very important part of grammar.

Chomsky: Language is an organ of the mind. The child is

creating the language.

MEANING OF WORDS: SEMANTICS

Commentator: Does this also apply to words?

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

Why do they acquire words so easily?

Does the brain give them some special help there too?

Gleitman: The problem is how the child learns the

meanings of words.

1. Mother points to the car.

Child knows: "aha this is a car"

Jill de Villier: The trouble is: that can't be the whole

story.

Dog: nunu

Overgeneralization of "nunu"

- nunu: referred to dogs in general, to animals, slippers,

salad, etc.

The question was-when he said "nunu", what did he

mean?

Gleitman: The trick in learning word meaning is not so

much applying it to the thing meant, but apply it to some

other referents, but not to all of them.

Commentator:

- How does a dog know that there is another dog?

- A word is something that stands for a concept.

Aronoff:

- But then we have another problem.

- What is a concept?

Gleitman: Clothes-Pin

- Clothes-Pin Statue

G: How does a child pick out a category that is relevantly

alike?

Experiment: What does "alive" mean?

- Is a dog alive?

Answer: yes: it has teeth, feet, it barks

- Is a worm alive?

Answer: yes.

But it doesn't have teeth and feet?

Kids: but it moves.

- What about a car? It moves...

Kids: agree - yes, it is alive.

Harvard University: Gavagai

Philosopher: Philip Quine

Gavagai refers to Rabithood

Gleitman: How could it be that all that comes to a child's

mind is RABIT?

Q: What might an inborn assumption be?

Pinker's Flimik-Experiment.

- open vs. closed flimiks

The whole object assumption

Child expects object labels to refer to the whole object.

The mutual exclusiveness principle

Children expect objects to have one and only one

label, that is one and only one name.

Words might be learned one by one.

Sentences, however, cannot be learned one by one.

Very young children can tell stories and thereby use

sentences that they have never heard before.

UG: UNIVERSAL GRAMMAR

Focus on Papua New Guinea [gini:]

If languages are inborn, then the question occurs of

whether the languages in New Guinea are similar to ours.

Crosslinguistic Universals: we find a small set of

principles.

One possibility that the child has to figure out is where to

put the verb - at the beginning - in the middle - or at the

end.

There are about 5 thousand languages spoken in the

world.

- these 5 thousand languages are very very similar.

Pinker: These 5 thousand languages are all dialects of

one human language.

De Saussure: There is no such thing as a primitive

language.

Every language has rules.

Siberian Eskimo: even this language is less different than

it seemed.

UG:

There are fixed invariable structural principles which are

simply part of the biological endowment and that

determine what counts as a human language.

The child might have a plan what might be a possible

rule:

A possible rule is: Subj Verb Obj and variations

thereof.

Children's Errors:

Children look for rules and overgeneralize rules.

(1) He drived to school.

(2) Geezes

two foots

it breaked instead of it broke.

There are certain kinds of mistakes that children never

seem to make.

1. Questions

(1) a. What did you eat your eggs with.

b. *What did you eat your eggs and?

2. Object Shift

(2) a. I baked a cake for Mary.

b. I baked Mary a cake.

(3) a. I painted the house for 6 hours.

b. *I painted 6 hours the house.

Mistakes like (1b) and (3b) are never made, because they

violate UG.

Experiment 2: Sam

Sam: What do you think what's in here?

Adult: What do you think is in here?

This is not a random error.

Rather it is a rule of a number of other languages.

German allows in one of its dialects that is disallowed in

English, namely (1b):

(1) a. Was glaubst Du, ist hier drin?

what think you is here in-it?

b. Was glaubst Du, was hier drin ist?

what think you what here in-it is

Chomsky: Striking general conclusion

1. Capacity to learn language is deeply engraved in the

mind.

2. Children are not taught language, they just do it.