WEBVTT

00:00:06.833 --> 00:00:08.725
"Some are born great,

00:00:08.725 --> 00:00:10.620
some achieve greatness,

00:00:10.620 --> 00:00:15.711
and others have greatness thrust
upon them", quoth William Shakespeare.

00:00:15.711 --> 00:00:17.301
Or did he?

00:00:17.301 --> 00:00:21.969
Some people question whether Shakespeare
really wrote the works that bear his name,

00:00:21.969 --> 00:00:24.889
or whether he even existed at all.

00:00:24.889 --> 00:00:28.809
They speculate that Shakespeare
was a pseudonym for another writer,

00:00:28.809 --> 00:00:30.236
or a group of writers.

00:00:30.236 --> 00:00:32.348
Proposed candidates 
for the real Shakespeare

00:00:32.348 --> 00:00:37.945
include other famous playwrights, 
politicians and even some prominent women.

00:00:37.945 --> 00:00:41.436
Could it be true that the greatest writer
in the English language

00:00:41.436 --> 00:00:44.941
was as fictional as his plays?

00:00:44.941 --> 00:00:47.867
Most Shakespeare scholars 
dismiss these theories

00:00:47.867 --> 00:00:51.439
based on historical 
and biographical evidence.

00:00:51.439 --> 00:00:55.514
But there is another way to test
whether Shakespeare's famous lines

00:00:55.514 --> 00:00:58.500
were actually written by someone else.

00:00:58.500 --> 00:01:00.689
Linguistics, the study of language,

00:01:00.689 --> 00:01:04.090
can tell us a great deal about the way
we speak and write

00:01:04.090 --> 00:01:09.585
by examining syntax, grammar, 
semantics and vocabulary.

00:01:09.585 --> 00:01:11.416
And in the late 1800s,

00:01:11.416 --> 00:01:15.447
a Polish philosopher 
named Wincenty Lutosławski

00:01:15.447 --> 00:01:18.226
formalized a method known as stylometry,

00:01:18.226 --> 00:01:23.428
applying this knowledge to investigate 
questions of literary authorship.

00:01:23.428 --> 00:01:25.395
So how does stylometry work?

00:01:25.395 --> 00:01:29.279
The idea is that each writer's style
has certain characteristics

00:01:29.279 --> 00:01:33.613
that remain fairly uniform 
among individual works.

00:01:33.613 --> 00:01:37.094
Examples of characteristics include
average sentence length,

00:01:37.094 --> 00:01:38.953
the arrangement of words,

00:01:38.953 --> 00:01:42.487
and even the number of occurrences
of a particular word.

00:01:42.487 --> 00:01:47.566
Let's look at use of the word thee
and visualize it as a dimension, or axis.

00:01:47.566 --> 00:01:50.554
Each of Shakespeare's works
can be placed on that axis,

00:01:50.554 --> 00:01:54.668
like a data point, based on the number
of occurrences of that word.

00:01:54.668 --> 00:01:57.235
In statistics, the tightness 
of these points

00:01:57.235 --> 00:02:02.498
gives us what is known as the variance,
an expected range for our data.

00:02:02.498 --> 00:02:07.995
But, this is only a single characteristic
in a very high-dimensional space.

00:02:07.995 --> 00:02:11.340
With a clustering tool
called Principal Component Analysis,

00:02:11.340 --> 00:02:16.131
we can reduce the multidimensional space
into simple principal components

00:02:16.131 --> 00:02:19.905
that collectively measure the variance
in Shakespeare's works.

00:02:19.905 --> 00:02:22.396
We can then test the works
of our candidates

00:02:22.396 --> 00:02:24.867
against those principal components.

00:02:24.867 --> 00:02:26.055
For example,

00:02:26.055 --> 00:02:30.394
if enough works of Francis Bacon
fall within the Shakespearean variance,

00:02:30.394 --> 00:02:32.263
that would be pretty strong evidence

00:02:32.263 --> 00:02:37.045
that Francis Bacon and Shakespeare
are actually the same person.

00:02:37.045 --> 00:02:39.161
What did the results show?

00:02:39.161 --> 00:02:42.477
Well, the stylometrists who carried 
this out have concluded

00:02:42.477 --> 00:02:46.557
that Shakespeare is none other 
than Shakespeare.

00:02:46.557 --> 00:02:49.191
The Bard is the Bard.

00:02:49.191 --> 00:02:54.370
The pretender's works just don't match up
with Shakespeare's signature style.

00:02:54.370 --> 00:02:57.642
However, our intrepid 
statisticians did find

00:02:57.642 --> 00:03:00.884
some compelling evidence 
of collaborations.

00:03:00.884 --> 00:03:03.138
For instance, one recent study concluded

00:03:03.138 --> 00:03:08.216
that Shakespeare worked with playwright
Christopher Marlowe on "Henry VI,"

00:03:08.216 --> 00:03:10.624
parts one and two.

00:03:10.624 --> 00:03:15.642
Shakespeare's identity is only one of 
the many problems stylometry can resolve.

00:03:15.642 --> 00:03:18.308
It can help us determine
when a work was written,

00:03:18.308 --> 00:03:21.040
whether an ancient text is a forgery,

00:03:21.040 --> 00:03:23.685
whether a student has committed plagiarism,

00:03:23.685 --> 00:03:29.020
or if that email you just received 
is of a high priority or spam.

00:03:29.020 --> 00:03:31.551
And does the timeless poetry 
of Shakespeare's lines

00:03:31.551 --> 00:03:34.475
just boil down to numbers and statistics?

00:03:34.475 --> 00:03:35.885
Not quite.

00:03:35.885 --> 00:03:40.900
Stylometric analysis may reveal what makes
Shakespeare's works structurally distinct,

00:03:40.900 --> 00:03:45.525
but it cannot capture the beauty of
the sentiments and emotions they express,

00:03:45.525 --> 00:03:48.509
or why they affect us the way they do.

00:03:48.509 --> 00:03:50.826
At least, not yet.