Professor Paul Bloom:
This class today is about
language.
And language is,
to a large extent,
where the action is.
The study of human language has
been the battleground over
different theories of human
nature.
So, every philosopher or
psychologist or humanist or
neuroscientist who has ever
thought about people has had to
make some claim about the nature
of language and how it works.
I'm including here people like
Aristotle and Plato,
Hume, Locke,
Freud and Skinner.
I'm also including modern-day
approaches to computational
theory, cognitive neuroscience,
evolutionary theory and
cultural psychology.
If you hope to make it with a
theory of what people are and
how people work,
you have to explain and talk
about language.
In fact, language is
sufficiently interesting that,
unlike most other things I'll
talk about in this class,
there is an entire field
devoted to its study,
the field of linguistics that
is entirely devoted to studying
the nuances and structures of
different languages.
Now, I'll first,
before getting into details,
make a definitional point.
When I'm talking about language
I'm meaning systems like English
and Dutch and Warlpiri and
Italian and Turkish and Urdu and
what we've seen and heard right
now in class in the
demonstration that preceded the
formal lecture.
Now, you could use language in
a different sense.
You could use the term
"language" to describe what dogs
do, or what chimpanzees do,
or birds.
You could use language to
describe music,
talk about the--a musical
language or art,
or any communicative system,
and there's actually nothing
wrong with that.
There's no rule about how
you're supposed to use the word
"language."
But the problem is if you use
the word "language" impossibly,
incredibly broadly,
then from a scientific point of
view it becomes useless to ask
interesting questions about it.
If language can refer to just
about everything from English to
traffic signals,
then we're not going to be able
to find interesting
generalizations or do good
science about it.
So, what I want to do is,
I want to discuss the
scientific notion of language,
at first restricting myself to
systems like English and Dutch
and American sign language and
Navajo and so on.
Once we've made some
generalizations about language
in this narrow sense,
we could then ask,
and we will ask,
to what extent do other systems
such as animal communication
systems relate to this narrower
definition.
So we could ask,
in this narrow sense,
what properties do languages
have and then go on to ask,
in a broader sense,
what other communicative
systems also possess those
properties.
Well, some things are obvious
about language so here are some;
here are the questions we will
ask.
This will frame our discussion
today.
We'll first go over some basic
facts about language.
We'll talk about what languages
share, we'll talk about how
language develops,
and we'll talk about language
and communication in nonhumans.
I began this class with a
demonstration of--that
illustrates two very important
facts about language.
One is that languages all share
some deep and intricate
universals.
In particular,
all languages,
at minimum, are powerful enough
to convey an abstract notion
like this;
abstract in the sense that it
talks about thoughts and it
talks about a proposition and
spatial relations in objects.
There's no language in the
world that you just cannot talk
about abstract things with.
Every language can do this.
But the demonstration also
illustrated another fact about
language, which is how different
languages are.
They sound different.
If you know one language,
you don't necessarily know
another.
It's not merely that you can't
understand it.
It could sound strange or look
unusual in the case of a sign
language.
And so, any adequate theory of
language has to allow for both
the commonalities and the
differences across languages.
And this is the puzzle faced by
the psychology and cognitive
science of language.
Well, let's start with an
interesting claim about language
made by Charles Darwin.
So, Darwin writes,
"Man has an instinctive
tendency to speak,
as we see in the babble of our
young children,
while no child has an
instinctive tendency to bake,
brew or write."
And what Darwin is claiming
here, and it's a controversial
and interesting claim,
is that language is special in
that there's some sort of
propensity or capacity or
instinct for language unlike the
other examples he gives.
Not everything comes natural to
us but Darwin suggests that
language does.
Well, why should we believe
this?
Well, there are some basic
facts that support Darwin's
claim.
For one thing,
every normal--every human
society has language.
In the course of traveling,
cultures encounter other
cultures and they often
encounter cultures that are very
different from their own.
But through the course of human
history, nobody has ever
encountered another group of
humans that did not have a
language.
Does this show that it's built
in?
Well, not necessarily.
It could be a cultural
innovation.
It could be,
for instance,
that language is such a good
idea that every culture comes
across it and develops it.
Just about every culture uses
some sort of utensils to eat
food with, a knife and a fork,
chopsticks, a spoon.
This probably is not because
use of eating utensils is human
nature, but rather,
it's because it's just a very
useful thing that cultures
discover over and over again.
Well, we know that this
probably is not true with regard
to language.
And one reason we know this is
because of the demonstrated case
studies where a language is
created within a single
generation.
And these case studies have
happened over history.
The standard example is people
involved in the slave trade.
The slave trade revolving
around tobacco or cotton or
coffee or sugar would tend to
mix slaves and laborers from
different language backgrounds,
in part deliberately,
so as to avoid the possibility
of revolt.
What would happen is these
people who were enslaved from
different cultures would develop
a makeshift communication system
so they could talk to one
another.
And this is called a "pidgin,"
p-i-d-g-i-n, a pidgin.
And this pidgin was how they
would talk.
And this pidgin was not a
language.
It was strings of words
borrowed from the different
languages around them and put
together in sort of haphazard
ways.
The question is what happens to
the children who are raised in
this society.
And you might expect it that
they would come to speak a
pidgin, but they don't.
What happens is,
in the course of a single
generation, they develop their
own language.
They create a language with
rich syntax and morphology and
phonology, terms that we'll
understand in a few minutes.
And this language that they
create is called a "creole."
And languages that we know now
as creoles, the word refers back
to their history.
That means that they were
developed from pidgins.
And this is interesting because
this suggests that to some
extent the ability to use and
understand and learn language is
part of human nature.
It doesn't require an extensive
cultural history.
Rather, just about any normal
child, even when not exposed to
a full-fledged language,
can create a language.
And more recently,
there's been case studies of
children who acquire sign
language.
There's a wonderful case in
Nicaragua in sign language where
they acquire sign language from
adults who themselves are not
versed in sign language.
They're sort of second-language
learners struggling along.
What you might have expected
would be the children would then
use whatever system their adults
use, but they don't.
They "creolized" it.
They take this makeshift
communication system developed
by adults and,
again,
they turn it into a full-blown
language, suggesting that to
some extent it's part of our
human nature to create
languages.
Also, every normal human has
language.
Not everybody in this room can
ride a bicycle.
Not everybody in this room can
play chess.
But everybody possesses at
least one language.
And everybody started to
possess at least one language
when they were a child.
There are exceptions,
but the exceptions come about
due to some sort of brain
damage.
Any neurologically normal human
will come to possess a language.
What else do we know?
Well, the claim that language
is part of human nature is
supported by neurological
studies,
some of which were referred to
in the chapters on the brain
that you read earlier that talk
about dedicated parts of the
brain that work for language.
And if parts of these
brains--if parts--if these parts
of the brain are damaged you get
language deficits or aphasias
where you might lose the ability
to understand or create
language.
More speculatively,
there has been some fairly
recent work studying the genetic
basis of language,
looking at the genes that are
directly responsible for the
capacity to learn and use
language.
And one bit of evidence that
these genes are implicated is
that some unfortunate people
have point mutations in these
genes.
And such people are unable to
learn and use language.
So, in general,
there is some support,
at least at a very broad level,
for the claim that language is
in some sense part of human
nature.
Well, what do we mean by
language?
What are we talking about when
we talk about language?
We don't want to restrict
ourselves, for instance,
to English or French.
What do all languages share?
Well, all languages are
creative and this means a couple
of things.
One meaning is the meaning
emphasized by Rene Descartes.
When Rene Descartes argued that
we are more than merely
machines, his best piece of
evidence for him was the human
capacity for language.
No machine could do this
because our capacity for
language is unbounded and free.
We could say anything we choose
to say.
We have free will.
And in fact,
language allows us to produce a
virtual infinity of sentences.
So, we could create and
understand sentences that we
never heard before.
And there are a lot of
sentences.
So, if you want to estimate how
many grammatical sentences under
twenty words in English,
the answer is,
"a lot."
And what this means is that any
theory of language use and
language comprehension cannot
simply appeal to a list.
When you understand a sentence
I said you have to have the
capacity to understand a
sentence even if you've never
heard it before.
And this is because we could
effortlessly produce and
understand sentences that no
human has ever said before on
earth.
Would anybody volunteer to say
a sentence, non obscene,
non derogatory,
that has never been spoken
before on earth,
ever?
Here.
I'll start.
"It's surprisingly easy to get
a purple tie on eBay if you
don't care much about quality."
I could imagine no one else in
the world has said this before.
"I am upset that one cannot
easily download 'Buffy the
Vampire Slayer' through iTunes."
Now, it's possible somebody
said both these sentences
before, but you probably have
not heard them.
But you understand them
immediately.
So, how do you do it?
Well, you have rules in your
head.
You've learnt what the words
mean, but you have abstract and
unconscious rules that take
these words,
figure out the order,
and in a fraction of a second,
give rise to understanding.
And that's the sort of thing
linguists study.
So, take some standard examples
from the linguistic study of
English.
And bear in mind the rules
we're talking about here are not
rules you explicitly know.
They're automatic rules of the
same sort we're going to talk
about in the context of visual
perception in that they're
implicit and unconscious and not
accessible to explicit
understanding.
So for instance,
immediately you read "The pig
is eager to eat" versus "The pig
is easy to eat" and in a
fraction of a second you know
there's an important difference.
"The pig is eager to eat" means
the state of affairs that we're
talking about is when the pig
does the eating.
"The pig is easy to eat" is
when the pig is being eaten.
You would see a sentence like
"Bill knew that John liked him"
and you know,
without even knowing how you
know,
that this could mean that Bill
knew that John liked Bill or it
could mean that Bill knew that
John liked Fred.
But it can't mean that Bill
knew that John liked John.
The natural interpretation,
in fact, is that Bill knew that
John liked Bill.
The two words co-refer.
Contrast that with "Bill knew
that John liked himself," which
only has the meaning Bill knew
that John liked John.
And this is what linguists do
for a living so if you hear me
talking about this and say,
"I want to spend the next forty
years of my life studying that,"
you should become a linguist.
But that's the sort of--those
are the sort of phenomena that
we're interested in.
Now, it gets more complicated.
Those are examples from syntax,
but language has many
structures.
Language has structures going
from the bottom to the top.
All languages--All human
languages have phonology,
which is the system of sounds
or signs;
morphology, which is the system
of words or morphemes,
basic units of meaning;
and syntax, which refer to
rules and principles that put
together words and phrases into
meaningful utterances.
And I want to talk briefly
about each of these three parts
of language before looking at
some other issues.
I'm indebted here to Steven
Pinker's excellent book The
Language Instinct which
provides, I think,
a superb discussion of these
phenomena.
And I'm going to steal some of
my examples from Pinker.
So, phonology.
Phonology is the system of
sounds that languages have.
There's a subset.
There's a list,
a finite list,
of possible sounds that
language can use.
I'm going to put aside for the
moment the question of sign
languages and how they work.
I'm going to talk about them in
a little bit.
The idea is that English has
about forty of these phonemes.
So, if you're a native
monolingual speaker of English
you hear speech and each sound
you hear is categorized as
falling into one of those forty
morphemes--sorry,
phonemes.
So, for example,
English has a phoneme of "lu,"
"l," and a phoneme of "r."
And so, an English speaker can
hear the difference between
"lip" and "rip" and that
corresponds to two different
words in English.
Other languages don't have that
distinction and so those
distinctions are very difficult
for non-native English speakers
to learn.
So, part of what goes on when
you learn, is you have to learn
the language--the phonemes that
your language has.
Another part of the problem of
learning language is you have to
figure out what the boundaries
are between the words.
You have to use sound signals
to figure out the boundaries
between the words.
Now that--If the only language
you've ever heard is English,
that's going to seem like a
really weird example of a
problem because you're listening
to me speak and in between each
of my words you're hearing a
pause.
You don't have to be very smart
to figure out where one word
begins and one word ends.
But the pause is a
psychological illusion.
If you were to just talk into
an oscilloscope that measured
your sound vibrations,
there are no pauses between the
words.
Rather, the pauses are inserted
by your mind as you already know
where one word begins and
another one ends.
And you insert a pause at that
point.
You could see this when you
hear a language you don't
already know.
So, for those of you who have
never heard French before,
when you hear somebody say,
"Je ne sais pas" you could say,
"Remarkable!
French has no pauses between
words."
And you-- And now a French
speaker, of course,
hears "Je ne sais pas."
For Hebrew, I know one sentence
in Hebrew: "Sleecha,
eypho ha-sheeruteem" which I
think is a request for the
bathroom.
But if you don't know Hebrew
there's no pauses.
And the truth is,
when you each gave your
demonstrations,
nobody spoke properly because
nobody spoke--Here's the
sentence: "Glorp fendel smug
wuggle."
Rather, you all sounded like,
"blublublublublub" without any
pauses because I don't know your
languages.
Children come into the world
without knowing any specific
language and so they have to
learn pauses.
They have to learn to interpret
sounds in context and sometimes
they make mistakes.
They get problems of
segmentation.
And there are some
illustrations.
You could see their mistakes if
they're trying to repeat back
something that's already known
within a society.
So, songs are a good example.
These are excerpts from
children.
"I'll never be your pizza
burnin'."
Anybody know--figure out what
that corresponds to?
Student: Beast of burden?
Professor Paul Bloom:
"Beast of burden."
Very good.
"A girl with colitis goes by."
Somebody?
Student: "A girl with
kaleidoscope eyes."
Professor Paul Bloom:
"The ants are my friends;
they're blowin' in the wind."
And [laughter]
this is a religious one.
"Our father with Bart in heaven;
Harold be they name… Lead us
not into Penn Station…" Now,
phonological understanding
illustrates all sorts of aspects
of language processing and,
in fact, of consciousness.
Because remember I said that,
typically, when you hear a
sentence you make--you
manufacture in your mind gaps
between the words.
Typically, when there's
something which is unclear
you'll fill in the gap and
figure out what the word is.
And you'll hear it that way.
So, the few examples--The best
examples, again,
are for when it goes wrong.
So, a classic example is from
the song "Super Freak" by Rick
James.
I got a big lecture about
copyright laws and this is going
to violate most of them.
Rick James is going to be
sitting on the--at--staring at
the web two years from now
saying, "Hey.
That's my thing."
Okay.
So, I want you to listen to
this line.
I'm sure most of you have heard
this before but I want you to
listen closely.
What was that last line?
[laughter]
"The kind of girl you read
about--" Well,
it turns out that nobody really
knows.
And it sounds to many people
who do top-down interpretation
as--to me as well,
that "she's the kind of girl
you read about in
Newsweek magazine."
But that makes no sense at all
given that you don't want to
"bring home to Mama."
And she's--and it's not
the--and in fact,
if you check the notes on the
song, she's in fact,
"the kind of girl you read
about in new wave magazines."
Now, when you listen to it
then, again, knowing that,
you hear it that way.
Now, this top-down--This is
known as "top-down" processing.
Top-down processing is an
example of when you know what
something is you hear it that
way.
And this is extremely useful
when it comes to filling in gaps
in sounds.
In normal conversation,
if I'm to say "s-- entence" you
won't hear that as "s--
entence."
Rather, you hear "sentence."
You fill in the gap.
This can lead to problems.
The problem it's led to in my
life revolves around the song
"Get Crunk" [laughter]
because I've heard "Get Crunk"
and my children asked me if I
would buy them "Get Crunk" from
iTunes.
My children are eight and ten.
And now "Get Crunk," as I was
aware from having heard it
before, involves the consistent
refrain of "get crunk" extremely
bad word,
"get crunk" extremely bad word,
and so I said "no."
And then they said,
"Well, there's a clean version
of it."
So, I downloaded the clean
version.
Unfortunately,
knowing what the clean
version--knowing what the word
is means to me the clean version
is not very clean.
Now, I will add,
[laughter]
before people write letters and
stuff, this is the clean
version.
[laughter]
Thank goodness they took away
that obscene word.
[laughter] Okay.
So, top-down processing affects
how we hear things,
usually, almost always,
for the better.
And in fact,
this is a theme we're going to
return to next class when we
talk about vision because the
same thing is going to happen
there.
How we see the world is often
confusing and befuddled but what
we know can clear things up.
Same with sound.
Morphology is the next level up.
Phonology is sounds.
Morphology is words.
And human language uses this
amazing trick described by
Ferdinand de Saussure,
the great linguist,
as "the arbitrariness of the
sign."
And what this means is we can
use--take any arbitrary idea in
the world, the idea of a chair
or a story or a country,
and make a sound or a sign to
connect to it.
And the link is arbitrary.
You might choose to use a word
for "dog" as "woof woof" because
it sounds like a dog but you
can't use a word for "country"
that sounds like a country.
You could use a sign language
thing for "drink" that looks
sort of like the act of drinking
but you can't use a sign
language word for "country" that
looks like a country,
or for "idea" that looks like
an idea.
So, the way languages work is
it allows for arbitrary naming.
It allows for this map between
a symbol, say a spoken word,
and any sort of thought we want
to use.
And those arbitrary mappings,
as we come to learn them,
make up the vocabulary of a
language.
I'm talking about words but the
more technical term is
"morpheme."
And what a morpheme is is the
smallest meaningful unit in a
language.
Now often, this is the same
thing as a word.
So, "dog" is a word.
And "dog" is also a morpheme,
but not always because there
are single morphemes and then
there are words that are
composed of many morphemes.
So, "dogs" and "complained" are
one word, but two morphemes and
what this means is that you make
the word by putting together two
morphemes.
To put it differently,
in order to know what "dogs"
means, you never had to learn
the word "dogs."
All you had to know is the word
"dog" and the plural morpheme
's' and you could put them
together to create a word.
How many morphemes does the
average speaker know?
The answer is fairly startling.
The average speaker knows,
as a low-ball estimate,
about 60,000 words.
I think the proper estimate is
closer to 80,000 or 100,000.
What this means,
if you average it out,
is that since children start
learning their first words at
about their first year of life,
they learn about nine new words
a day.
And it's not a continuous nine
words every day.
It goes up and down depending
on the age.
But still, the amount of words
we know is staggering.
How many of you know more than
one language pretty fluently?
Those of you who know other
languages might have in your
heads 200,000 words or 300,000
words and you're accessing them
in a fraction of a second.
It is--could legitimately be
seen as one of the most
astonishing things that people
do.
Finally, syntax.
So, we have the sound system of
a language, the phonology.
We have the words of a
language, the morphology,
but all that gives you is
"dog," "cup," "chair," "house,"
"story," "idea."
That won't allow us to
communicate complicated ideas.
So, the final step in the story
is syntax.
And syntax refers to those
rules and principles that allow
us to combine words into phrases
and phrases into sentences.
And syntax uses another neat
trick and this is defined by
Wilhelm von Humboldt as the
"infinite use of finite media."
So, here's the question.
Your vocabulary is finite.
There are just so many words.
You have to learn them one by
one, but you could produce a
virtual infinity of sentences.
How can you do that?
How can you go from a finite
list of symbols to an infinite
number of sentences?
And the answer is you have a
combinatorial system.
Now, language is not the only
thing in culture or nature that
has this sort of combinatorial
system.
Music also has a combinatorial
system.
There's a finite number of
notes but a limitless number of
musical compositions.
DNA also has this sort of
combinatorial system where you
have a finite number of,
I guess, bases or amino acids
that could combine to a possible
infinity of strings,
of DNA strings.
So, how does this happen?
Well, the infinity mechanism,
and many of you will be
familiar with this from
mathematics or computer science,
is recursion.
And there's a lot to be said
about this but it could be
pretty simply illustrated in
language.
So, here's an example of a
simple language.
It's not--It's actually close
to how linguists describe normal
languages, but it's very simple.
It has three nouns,
"Fred," "Barney" and "Wilma,"
and two verbs,
"thinks" and "likes."
A very simple language.
And one rule.
And the way to read this rule
is you make a sentence by taking
a noun, any noun,
putting a verb after it,
and then following that verb
with a noun.
Now, when you do this,
how many--And then so,
for instance,
you get the sentence "Fred
likes Wilma."
When you do this,
how many possible sentences are
there?
Let me just take a second.
Okay.
Any guesses?
Eighteen.
The sentences are "Fred likes
Fred," "Fred likes Barney,"
"Fred likes Wilma,"
"Fred thinks Fred," "Fred
thinks Barney," "Fred thinks
Wilma," and so on.
The three nouns followed by any
of the two verbs followed by any
of the three nouns.
That is not a very interesting
language.
But now, take a more
complicated language--same
vocabulary, the same three
nouns, the same two verbs,
the same sentence,
but now one other sentence.
This sentence expands to a noun
followed by a verb followed by a
sentence and there you get
recursion.
You have one rule invoking
another rule and then you can
get a sentence like "Fred thinks
Barney likes Wilma."
And here you get a potential
infinity of sentences.
And this is obviously a toy
example but you could see the
use of recursion in everyday
life and in everyday use of
language.
You could say,
"John hates cheese," "My
roommate heard a rumor that John
hates cheese,"
"It disturbed Mary when I told
her that my roommate heard a
rumor that John hates cheese,"
"I was amazed that it disturbed
Mary when I told her that my
roommate heard a rumor that John
hates cheese,"
"Professor Bloom had devoted
way too much of his lecture
talking about how I was amazed
[laughter]
that it disturbed Mary when I
told her that my roommate heard
a rumor that John hates cheese,"
"It really bothered me that--"
and there's no limit.
There's no longest sentence.
You could keep producing a
sentence deeper and deeper
embedded until you die.
And this is part of the power
of language.
Now, the syntactic rules are
complicated.
And one of the puzzles of
syntactic rules,
or one of the issues of them,
is that different rules can
conspire to create the same
sentence.
So, you take a sentence
like--This is a classic line
from Groucho Marx:
"I once shot an elephant in my
pajamas.
How it got into my pajamas I'll
never know."
And the humor,
such that it is,
revolves around the ambiguity
of rules that generate it,
like this versus like this.
Often, to illustrate the issues
of ambiguity,
people have collected poorly
thought-out headlines in
newspaper reports that play
on--that inadvertently have
ambiguity.
"Complaints about NBA referees
growing ugly."
So, that's the beauty of that
structure.
"Kids make nutritious snacks."
"No one was injured in a blast
which was attributed to the
buildup of gas by one town
official."
Last summer I was in Seoul
visiting the--visiting Korea
University and the big headline
there on the front page was
"General arrested for fondling
privates."
[laughter]
Now, there actually is--The
ambiguity is actually quite
difficult to avoid in the
construction and understanding
of sentences.
It's one of the ways in which
it's often difficult to write
clearly, and in fact,
there's a whole sub-field of
the law involving the use of
linguistic theory to
disambiguate sentences both in
the Constitution,
in legislation,
as well as in some criminal
cases.
And there was,
several years ago,
a very serious criminal case
that rested on a sentence.
And here's what happened.
There were two brothers,
one of them retarded,
and they get into a robbery.
And a police officer sees them
and points the gun at them.
And one of the brothers points
a gun at the police officer.
The police officer shouts for
the brother, the non-retarded
brother, to drop the gun.
Actually, he said,
"Give me the gun."
The retarded brother shouted,
"Let him have it," whereupon
the brother shot and killed the
police officer.
Now, the brother who did the
shooting was plainly a murderer.
What about that brother who
shouted, "Let him have it"?
Well, it depends on what he--on
how you interpret that sentence
because the sentence is
beautifully ambiguous.
It could mean "shoot him,
let him have it," or it could
mean "give him the gun,
let him have it."
And in fact,
the trial, which I think
somebody could--If people out
there know about this,
please send me an e-mail.
My understanding was he was
found guilty but a lot to turn
on the ambiguity of a sentence.
I want to shift now and talk
about where does all this
knowledge come from but I'll
stop and answer any questions
about the material so far.
What are your questions?
Yes.
Student: How does syntax
differ from grammar or are they
exactly the same?
Professor Paul Bloom:
Syntax--The question is,
"How is syntax different from
grammar?"
They're exactly the same.
Syntax is a more technical term
but it means the same thing as
grammar.
Yes.
Student: You said that
every normal human being that's
born uses at some point or
another some kind of language.
Aren't there people who weren't
born within a culture and grew
up and who never really spoke a
language though they were
physically normal?
Professor Paul Bloom:
Yes.
I'm glad you actually asked me
about that because,
as I said it,
I realized it wasn't quite
right.
The point that was just raised
here is I had said before that
everybody who's neurologically
normal comes to acquire and
learn a language.
But what about people who are
neurologically normal but they
don't have language around them?
And in fact,
there have been,
historically,
some cases of this.
There's been,
probably apocryphal,
stories about children who are
raised by wolves or by dogs.
There are stories,
horrible stories,
some in the twentieth century,
about children who are locked
away by insane or evil parents
and have never learned to speak.
There are stories of deaf
people who are within certain
societies where nobody signs to
them, and so they're what's
known as linguistic isolates.
And they themselves never learn
to speak.
And those cases are the
dramatic exception and they do
tell you something.
They tell you that it's not
enough to have a brain for
language.
Somebody does have to use it
with you.
Interestingly,
it doesn't have to be that many
people.
So, Susan Goldin-Meadow has
studied deaf children that
nobody signed to but what she
studies is deaf children with
deaf siblings and these children
don't just sit there.
They create their own language.
It's not a full-blown language
like American sign-language or
langue des signes
quebecoise but it's a
language nonetheless,
with words and syntax and
phonology.
It's an interesting question.
Any other questions?
Yes.
Student: Could it be
argued that there are inherent
limits to grammar?
Professor Paul Bloom:
It's a good question.
The question is,
"Are there inherent limits in
our abilities to come up with
grammars?"
And most linguists would argue
"yes," that languages are highly
constrained in how they do
things.
So, for instance,
one example is there's no
language in the world that ever
constructs a question by
switching the order of words
around in a sentence.
There's no language in the
world that has a rule that says
the fifth word has to be a verb.
And linguists have all of these
conditions they say,
"no language in the world works
this way."
Now this is--;So,
these are constraints on
grammar and they're really
interesting because they tell us
what's a humanly natural
language versus what's not a
humanly natural language.
But notice, even if there is
incredible constraints on
grammars, still--we could still
produce an infinite number of
sentences.
It's just like if you restrict
me to only a subset of numbers,
only the odd numbers,
still there's an infinity of
odd numbers.
So, grammar can be restricted
but still give rise to an
infinity of possible sentences.
Well, there's a radical claim
about the origin of language
associated with the guy who we
met when we talked about
behaviorism who wrote A
Review of Verbal Behavior,
the linguist Noam Chomsky.
And Chomsky makes this radical
claim.
And this is that we shouldn't
view language learning as
learning at all.
Instead, we should view it as
something similar to growth.
So he says,
No one would take
seriously the proposal that a
human organism learns through
experience to have arms rather
than wings,
or that the basic structure of
particular organs results from
accidental experience.
[Language]
proves to be no less marvelous
and intricate than these
physical structures.
Why, then, should we not study
the acquisition of a cognitive
structure like language more or
less as we study some complex
bodily organ?
So, you might learn to play
baseball, you might learn about
the American Civil War,
but if Chomsky is right you
didn't learn to speak English.
Rather, what happened is you
heard English and--but the
capacity grew in your head and
something a lot more similar to
the development of arms or legs
or a visual system.
Well, should we believe this?
We know there has to be some
effect of the environment
shaping language,
obviously,
because in order to know
English you have to have heard
English, in order to know Dutch
you have had to heard,
to--had to have learned and
heard Dutch.
And in fact,
languages differ in all the
ways that we were talking about.
Some languages like English has
a--have a distinction between
‘l' and ‘r.'
Other languages do not.
For a language like English,
that creature there is referred
to with the morpheme "dog."
That's a historical accident of
English.
In French it's chien and
in Greek it's something else.
And each of those 6,000
languages and people in the room
who know another language would
say,
"Yeah, in Vietnamese it's
this," "In Urdu it's this," "In
Czech it's that."
Finally, there is syntax.
So, English is what's known as
a subject-verb-object language.
That means if you want to
convey the idea that Bill hit
John, you would say,
"Bill hit John."
But not all languages work that
way.
In fact, the majority of
languages, more languages,
are actually
subject-object-verb languages.
So, you would say,
if you wanted to convey that
Bill was the hitter and John was
hit, "Bill John hit."
All of this has to be learned.
And all of this has to be
learned through exposure to
language users.
On the other hand,
there is considerable evidence
that the development of these
language skills,
in some way,
is similar to growth in the way
that Chomsky suggests.
So, here are some basic facts
about language development.
One is something which I had
mentioned before.
All normal children learn
language.
There can be specific
impairments of language.
Now, again, we spoke about them
before when talking about the
brain.
Some of these impairments could
be due to trauma,
the aphasias.
Trauma, a blow to the head,
a stroke can rid you of your
language.
But, also, there are genetic
disorders, some falling under
the rubric of what's known as
"specific language impairment,"
where children are born without
the same ability as the rest of
us to learn to speak.
And these are interesting in
many ways.
One reason that they're
interesting is that they
illustrate something about human
language.
It is not--It would not be
unreasonable for you to think
before listening to his lecture,
"Look.
All you need to have to learn a
language is to be smart" or "All
you need to have to learn a
language is to want to
communicate" or "All you need to
have to learn a language is to
be a social person wanting
to--having the ability to
understand others and deal with
others."
But the cases of specific
language impairments suggest
that all of that is wrong,
because there are children in
this world right now who are
plenty smart,
who really want to communicate,
and who are entirely social
creatures but they can't learn
language.
And this suggests that the
ability to learn language and
understand language is to some
extent separate from these other
aspects of mental life.
Continuing on this theme,
we also know that language is
learnt without any sort of
feedback or training.
There are many Americans who
believe that they need to teach
their children language.
And there's a huge industry
with DVDs and flash cards and
all sorts of things designed to
teach your children language.
And I think many parents
believe that if they didn't
persist in using these things
their children would never learn
to speak.
But we know that that's not
true.
We know that this isn't true
because there are communities
where they don't speak to their
kids.
They don't speak to their kids
because they don't believe it's
important to speak to their
kids.
Some linguists would
interview--Linguists would
interview adults in these
communities and say,
"Why don't you speak to your
babies?"
And these adults would respond,
"It'd be ridiculous to speak to
a baby.
The baby has nothing to say.
You might as well just speak to
your dog."
And then the American linguist
would say, "Yeah.
We speak to our dogs."
[laughter]
Americans and Europeans speak
to everything and everybody.
Other cultures are more picky
and they don't talk to their
children until their children
themselves are talking.
This doesn't seem to make much
of a difference in language
learning.
Some studies have,
motivated by Chomsky's work in
expressed--sorry,
motivated by Chomsky's critique
of Skinner's Verbal Behavior,
have asked even in-- "What
if we just looked at children
within the United States?
Don't these children get
feedback?"
And the answer is yes and no.
So your average highly educated
Western parent does give their
children feedback--do give their
children feedback based on what
they say.
But they don't typically give
feedback based on the syntax or
grammaticality of what they say.
The example given by Brown and
Hanlon in the classic study in
the 1970s is they did all of
these studies looking at what
children say and how parents
responded,
and it turns out parents
respond not to the grammatical
correctness but to the affect or
cuteness or sociability of the
utterance.
So for instance,
if a child says to his mother,
"I loves you,
Mommy," it's a very unusual
parent who would say,
"Oh, no.
The verb agreement is mistaken.
[laughter]
You've added a redundant ‘s.'
It's not appropriate."
Similarly, if a child is to
say, "I hate your guts,
Mother," it's an unusual
mother, "That's wonderful.
There's a subject, verb, object.
The whole thing's structurally
fine."
We respond to our kids like we
respond to each other based on
the message that's conveyed,
not the grammaticality of the
utterances.
Children make grammatical
mistakes all the time but then
they go away and they go away
without correction.
So those are some basic facts.
What do we know about the time
course of language?
Well, early on children start
off and they prefer the melody
of their own language.
These studies were done in
France with four-day-old babies.
And what they did was they used
a sucking method.
Remember, there's a limited
number of things babies can do.
One of the things they can do
is suck, and these babies would
suck on a pacifier to hear
French.
And they would prefer to hear
French than to hear Russian.
And these investigators claimed
this is because they had been
exposed to French in the first
four days of their lives.
Reviewers, mostly from France,
objected and said,
"No.
Maybe French just sounds better.
Everybody's going to like
French."
So, they re-did the study in
Russia.
Russian kids sucked harder to
hear Russian than they did to
French.
And what they're listening to
isn't the words.
They don't know words yet.
They don't know of syntax yet.
It's the rhythm of the language.
For you, French and Russian
sound different.
Even if you're like me and you
don't know a word of either
language, they still sound
different.
They sound different to babies
too.
And a baby being raised in
France or a baby being raised in
Russia knows enough to tell
what's his language and what
isn't.
Early on, children are
sensitive to every phoneme there
is.
So, English-speaking children,
for instance,
can--English-speaking babies
– babies who are born in the
United States – can
distinguish between English
phonemes like "lip" and "rip"
but they could also distinguish
between phonemic contrasts that
are not exemplified in English,
such as phonemic contrasts in
Czech or Hindi.
Yes.
Student:
I'm wondering if you can say
the wrong things to them--to
infants based on what you were
saying before.
Because I was in France one
summer and I had some neighbors
there.
I hated these neighbors,
I thought they were stupid.
Not because they were French,
but they had a baby and it
would gaggle and coo and they
would respond in similar terms.
Professor Paul Bloom:
They would gaggle and coo back
at the baby.
Student: [inaudible]
And I hate these people.
[inaudible]
So I don't know if it--Does it
matter what you say to babies as
long as you say something.
Professor Paul Bloom:
There's a lot going on in your
question.
[laughter]
Some raising--Well,
there's a lot going on in your
question.
The answer to the question--
The question was,
"your baby's going to coo and
'ga ga, goo goo,' does it matter
if you coo and 'ga ga,
goo goo' back?"
No, it doesn't make a
difference.
Your hatred towards them was
unmotivated.
You can be relieved of that
debt, or now you know you feel
bad now, I guess.
[laughter]
If you speak to your children
in perfect English,
it's very strange.
Nobody speaks to their babies
in, "Hello, Son.
It's time--Oh.
You want to change your diaper
right now so stay still."
That's bad parenting.
It sounds kind of silly.
More--What most people do is,
"Oh.
You're such a cute little baby."
And it probably--One--T
here's--Evolutionary
psychologists debate the
function of why we talk funny to
babies.
And some people have argued
that it does help their language
learning.
And some people have argued
instead that what it does is it
calms them.
They like to hear the music of
a smooth voice and so on.
But whether or not you do so
doesn't seem to make a big
difference.
It is very difficult to find
any effect of how parents talk
to their kids on how their kids
learn language,
particularly when it comes to
babies.
So, early on babies can--are
sensitive to all phonemes and
then that goes away.
Around twelve months of age it
goes away.
This is one thing you were much
better at when you were a baby
than you are now.
When you were a baby you were a
multilingual fool.
You could understand the sound
differences of every language on
earth.
Now, if you're like me,
you could barely understand
English.
[laughter]
You narrow down until you're
sensitive just to the language
you hear.
And this narrowing down is
largely in place by about twelve
months of age.
Around seven months is babbling.
And I want to stop at this
point to go back to the issue--I
promised you I would turn a bit
to sign language and I want to
describe now a very elegant--I
want to show a little film now
of a very elegant series of
experiments looking at the
question of whether babies who
are exposed to a sign language,
babble.
One of the real surprising
findings in my field over the
last ten/twenty years has been
that the acquisition of sign
languages has turned out to be
almost exactly the same;
in fact, as far as we know,
exactly the same as the
acquisition of spoken languages.
It didn't have to be that way.
It could have been just as
reasonable to expect that
there'd be an advantage for
speech over sign.
That sign languages may be
full-blown languages but they
just take--they're just harder
to learn because the brain and
the body have adapted for
speech.
It turns out that this just
isn't the case.
It turns out that sign and--the
developmental milestones of sign
languages and the developmental
milestones of spoken languages
are precisely the same.
They start babbling at the same
point.
They start using first words,
first sentences,
first complicated
constructions.
There seems to be no
interesting difference between
how the brain comes to acquire
and use the spoken language
versus a sign language.
Around twelve months of age,
children start using their
first words.
These are words for objects and
actions like "dog" and "up" and
"milk."
They start showing some
sensitivity to the order of
words.
So they know that "dog bites
cat" is different from "cat
bites dog."
Around eighteen months of age,
they start learning words
faster.
They start producing little,
miniature sentences like "Want
cookie" or "Milk spill" and the
function morphemes,
the little words,
"in," "of," "a," "the," and so
on start to gradually appear.
Then the--Then there's the bad
news.
Around seven years of age going
up through puberty,
the ability to learn language
starts to go away.
The best work on this has been
done by Elissa Newport and Sam
Supalla who have studied people
who have been in the United
States for many,
many years – 30,40 years –
and seeing how well they have
come to speak English.
And it turns out the big
determinant of how well you
speak English as an immigrant
isn't how smart you are.
It's not how many family
members you have when you're
here.
It's not your motivation.
It's how old you were when you
started.
It turns out that if you start
learning a language – a second
language is where most of the
work's been done – within the
first few years of life you're
fine.
You'll speak like a native.
But then it starts getting
worse and worse.
And once you hit puberty,
suddenly there's huge variation
in the abilities you have to
learn language.
It is very rare,
for instance,
for somebody who has learned
English past puberty to speak
without an accent.
An accent is very hard to shake
and it's not just an accent.
It's also other aspects of
phonology, syntax,
and morphology.
It's like the part of the brain
that's responsible for language
learning is only around early in
development and if you don't get
your language by then it'll just
run out.
I want to begin next class with
this question,
the question of animals.
And that will shut down the
language learning part.
But one thing I'll put up here
is your second reading response.
So, I'll also put this up on
Wednesday, and by Wednesday you
might have a bit of a better--be
in a better position to answer
this question.
But I'll continue with language
on Wednesday and then we'll also
talk about vision,
attention, and memory.
I'll see you then.