WEBVTT

00:00:06.914 --> 00:00:09.594
Language is endlessly variable.

00:00:09.594 --> 00:00:12.744
Each of us can come up with 
an infinite number of sentences

00:00:12.744 --> 00:00:14.134
in our native language,

00:00:14.134 --> 00:00:16.924
and we’re able to do so from an early age—

00:00:16.924 --> 00:00:20.384
almost as soon as we start 
to communicate in sentences.

00:00:20.384 --> 00:00:22.164
How is this possible?

NOTE Paragraph

00:00:22.164 --> 00:00:26.164
In the early 1950s, Noam Chomsky 
proposed a theory

00:00:26.164 --> 00:00:31.463
based on the observation that the key 
to this versatility seems to be grammar:

00:00:31.463 --> 00:00:35.053
the familiar grammatical structure 
of an unfamiliar sentence

00:00:35.053 --> 00:00:37.363
points us toward its meaning.

00:00:37.363 --> 00:00:39.683
He suggested that there are 
grammatical rules

00:00:39.683 --> 00:00:44.603
that apply to all languages,
and that the rules are innate—

00:00:44.603 --> 00:00:49.788
the human brain is hardwired to process
language according to these rules.

NOTE Paragraph

00:00:49.788 --> 00:00:53.168
He labelled this faculty 
universal grammar,

00:00:53.168 --> 00:00:57.433
and it launched lines of inquiry 
that shaped both the field of linguistics

00:00:57.433 --> 00:01:01.735
and the emerging field 
of cognitive science for decades to come.

00:01:01.735 --> 00:01:04.705
Chomsky and other researchers 
set out to investigate

00:01:04.705 --> 00:01:07.675
the two main components 
of universal grammar:

00:01:07.675 --> 00:01:10.725
first, whether there are, in fact, 
grammar rules

00:01:10.725 --> 00:01:13.285
that are universal to all languages,

00:01:13.285 --> 00:01:17.955
and, second, whether these rules 
are hardwired in the brain.

NOTE Paragraph

00:01:17.955 --> 00:01:21.105
In attempts to establish 
the universal rules of grammar,

00:01:21.105 --> 00:01:26.200
Chomsky developed an analytical tool
known as generative syntax,

00:01:26.200 --> 00:01:31.564
which represents the order of words 
in a sentence in hierarchical syntax trees

00:01:31.564 --> 00:01:34.294
that show what structures are possible.

NOTE Paragraph

00:01:34.294 --> 00:01:37.934
Based on this tree, we could suggest 
a grammar rule

00:01:37.934 --> 00:01:41.034
that adverbs must occur in verb phrases.

00:01:41.034 --> 00:01:43.914
But with more data, 
it quickly becomes clear

00:01:43.914 --> 00:01:47.224
that adverbs can appear 
outside of verb phrases.

00:01:47.224 --> 00:01:50.854
This simplified example illustrates 
a major problem:

00:01:50.854 --> 00:01:54.544
it takes a lot of data 
from each individual language

00:01:54.544 --> 00:01:56.904
to establish the rules for that language,

00:01:56.904 --> 00:01:59.274
before we can even begin to determine

00:01:59.274 --> 00:02:02.964
which rules all languages 
might have in common.

00:02:02.964 --> 00:02:05.644
When Chomsky proposed universal grammar,

00:02:05.644 --> 00:02:08.887
many languages lacked the volume 
of recorded samples

00:02:08.887 --> 00:02:12.417
necessary to analyze them 
using generative syntax.

00:02:12.417 --> 00:02:14.020
Even with lots of data,

00:02:14.020 --> 00:02:17.670
mapping the structure of a language
is incredibly complex.

00:02:17.670 --> 00:02:23.659
After 50 years of analysis, we still
haven’t completely figured out English.

NOTE Paragraph

00:02:23.659 --> 00:02:26.809
As more linguist data 
was gathered and analyzed,

00:02:26.809 --> 00:02:31.266
it became clear that languages 
around the world differ widely,

00:02:31.266 --> 00:02:35.421
challenging the theory that there were 
universal grammar rules.

00:02:35.421 --> 00:02:38.545
In the 1980s, Chomsky revised his theory

00:02:38.545 --> 00:02:41.245
in an attempt to accommodate 
this variation.

00:02:41.245 --> 00:02:45.872
According to his new hypothesis 
of principles and parameters,

00:02:45.872 --> 00:02:48.892
all languages shared certain 
grammatical principles,

00:02:48.892 --> 00:02:53.410
but could vary in their parameters, 
or the application of these principles.

00:02:53.410 --> 00:02:57.581
For example, a principle is 
“every sentence must have a subject,"

00:02:57.581 --> 00:03:01.925
but the parameter of whether the subject
must be explicitly stated

00:03:01.925 --> 00:03:03.835
could vary between languages.

NOTE Paragraph

00:03:03.835 --> 00:03:06.365
The hypothesis of principles 
and parameters

00:03:06.365 --> 00:03:11.133
still didn’t answer the question of which
grammatical principles are universal.

00:03:11.133 --> 00:03:16.404
In the early 2000s, Chomsky suggested 
that there’s just one shared principle,

00:03:16.404 --> 00:03:21.571
called recursion, which means structures
can be nested inside each other.

00:03:21.571 --> 00:03:23.124
Take this sentence,

00:03:23.124 --> 00:03:26.764
which embeds a sentence within a sentence 
within a sentence.

00:03:26.764 --> 00:03:30.958
Or this sentence, which embeds 
a noun phrase in a noun phrase

00:03:30.958 --> 00:03:32.408
in a noun phrase.

00:03:32.408 --> 00:03:35.661
Recursion was a good candidate 
for a universal grammar rule

00:03:35.661 --> 00:03:38.431
because it can take many forms.

00:03:38.431 --> 00:03:42.351
However, in 2005 linguists 
published findings

00:03:42.351 --> 00:03:45.181
on an Amazonian language called Piraha,

00:03:45.181 --> 00:03:49.181
which doesn’t appear to have 
any recursive structures.

NOTE Paragraph

00:03:49.181 --> 00:03:51.991
So what about the other part 
of Chomsky’s theory,

00:03:51.991 --> 00:03:55.171
that our language faculty is innate?

00:03:55.171 --> 00:03:57.701
When he first proposed universal grammar,

00:03:57.701 --> 00:04:02.419
the idea that there was a genetically 
determined aspect of language acquisition

00:04:02.419 --> 00:04:05.379
had a profound, revolutionary impact.

00:04:05.379 --> 00:04:09.679
It challenged the dominant paradigm, 
called behaviorism.

00:04:09.679 --> 00:04:15.174
Behaviorists argued that all animal 
and human behaviors, including language,

00:04:15.174 --> 00:04:18.357
were acquired from the outside 
by the mind,

00:04:18.357 --> 00:04:20.927
which starts out as a blank slate.

00:04:20.927 --> 00:04:24.544
Today, scientists agree that behaviorism 
was wrong,

00:04:24.544 --> 00:04:28.541
and there is underlying, 
genetically encoded biological machinery

00:04:28.541 --> 00:04:29.930
for language learning.

00:04:29.930 --> 00:04:32.870
Many think the same biology 
responsible for language

00:04:32.870 --> 00:04:37.011
is also responsible for other 
aspects of cognition.

00:04:37.011 --> 00:04:39.505
So they disagree with Chomsky’s idea

00:04:39.505 --> 00:04:45.015
that there is a specific, isolated,
innate language faculty in the brain.

NOTE Paragraph

00:04:45.015 --> 00:04:49.135
The theory of universal grammar
prompted the documentation and study

00:04:49.135 --> 00:04:52.235
of many languages 
that hadn’t been studied before.

00:04:52.235 --> 00:04:57.151
It also caused an old idea to be 
reevaluated and eventually overthrown

00:04:57.151 --> 00:05:01.113
to make room for our growing 
understanding of the human brain.