WEBVTT

00:00:01.193 --> 00:00:02.127
When you marry,

00:00:02.127 --> 00:00:06.811
usually you take on your partner's surname
or your parter takes on yours.

00:00:06.811 --> 00:00:09.057
Two people with different surnames become

00:00:09.057 --> 00:00:10.867
two people with the same.

00:00:10.867 --> 00:00:13.566
One surname spreads, the other one goes...

00:00:13.566 --> 00:00:15.266
...extinct? Usually not.

00:00:15.266 --> 00:00:17.350
There might be siblings, cousins,

00:00:17.350 --> 00:00:19.510
strangers who happen to share the surname,

00:00:19.511 --> 00:00:22.271
to carry it on for the one who lost it.

00:00:22.271 --> 00:00:25.199
But if one person fails to pass on the surname,

00:00:25.199 --> 00:00:27.327
so might the others.

00:00:27.327 --> 00:00:29.200
In fact, every now and then

00:00:29.200 --> 00:00:31.932
entire surnames do go extinct

00:00:31.932 --> 00:00:33.982
when its last bearer passes away

00:00:33.995 --> 00:00:35.835
without passing it on.

00:00:36.011 --> 00:00:37.681
According to the Daily Mail,

00:00:37.681 --> 00:00:40.949
in England and Wales, 200,000 surnames

00:00:40.949 --> 00:00:43.109
were lost since 1901.

00:00:43.213 --> 00:00:45.143
You can find lists of endangered surnames

00:00:45.148 --> 00:00:48.688
on websites such as Ancestry.com and myheritage.com

00:00:48.868 --> 00:00:51.178
Ancestry counts surnames with less than

00:00:51.184 --> 00:00:53.884
50 carriers left as endangered,

00:00:53.884 --> 00:00:55.043
which in England and Wales,

00:00:55.043 --> 00:00:56.876
would currently be names such as

00:00:56.876 --> 00:00:59.976
Pober, Mirren, Febland (heh, Febland),

00:01:00.082 --> 00:01:02.122
Nighy - N-Nighy?

00:01:02.480 --> 00:01:04.860
While some of these names might be 
more of a loss than others,

00:01:04.875 --> 00:01:10.255
it's sad to think that they might all 
cease to exist within a few generations.

00:01:10.394 --> 00:01:13.904
Back in the days, new surnames 
were created as well

00:01:13.933 --> 00:01:18.603
based on someone's job or father's given name or where they came from.

00:01:18.641 --> 00:01:20.941
But that doesn't really happen anymore,

00:01:20.947 --> 00:01:22.887
not on a large scale, anyway.

00:01:22.887 --> 00:01:25.885
So more surnames are lost than
new ones are being born.

00:01:25.885 --> 00:01:28.508
Keep this experiment going long enough,


00:01:28.515 --> 00:01:33.505
and we will all end up with
the same surname eventually, won't we?

00:01:33.847 --> 00:01:38.440
When we look at Earth's more ancient
civilizations - even more ancient-

00:01:38.440 --> 00:01:43.310
intensive research reveals that most 
Chinese surnames in use today

00:01:43.331 --> 00:01:46.191
were handed down from thousands of years ago.

00:01:46.365 --> 00:01:50.591
While historically about 12,000 
Chinese surnames have been recorded,

00:01:50.591 --> 00:01:53.971
only a bit over 3,000 are currently in use,

00:01:53.971 --> 00:01:56.975
a reduction of 75 percent!

00:01:56.987 --> 00:02:01.897
And only a fraction of those are taking 
over a majority of the entire population.

00:02:01.901 --> 00:02:07.503
The three most common surnames in mainland
China are Li, Wang and Zhang,

00:02:07.503 --> 00:02:12.303
which make up for more than 7 percent 
of the Chinese population each.

00:02:12.313 --> 00:02:15.980
Together they belong to close to
300 million people

00:02:15.980 --> 00:02:19.559
and are easily the most common surnames
in the world.

00:02:19.559 --> 00:02:24.429
In China, the phrase "three Zhang, four Li"
is used to say just "anybody."

00:02:24.441 --> 00:02:28.051
So after thousands of years, 
the Chinese people aren't down to

00:02:28.051 --> 00:02:31.527
one all-dominating surname, but several.

00:02:31.527 --> 00:02:32.687
What's going on here?

00:02:32.702 --> 00:02:35.392
This effect can be shown in a simulation.

00:02:35.736 --> 00:02:39.126
What you see here is the result of a 
Galton-Watson process,

00:02:39.126 --> 00:02:43.800
which maps out how the distribution of
family names changes over time.

00:02:43.800 --> 00:02:47.116
It starts out with a very large number
of unique family names


00:02:47.116 --> 00:02:51.126
each represented by a different color, 
and after 40 generations,

00:02:51.126 --> 00:02:54.896
or around a thousand years,
ends with the ones which are left.

00:02:54.896 --> 00:02:56.472
When you look at the very end there,

00:02:56.472 --> 00:03:00.262
what you see is very similar to the Chinese situation.

00:03:00.276 --> 00:03:03.896
The top three names take over 20 percent
of the cake.

00:03:04.173 --> 00:03:07.705
But the question is: 
if we keep the simulation going,

00:03:07.705 --> 00:03:10.565
will we end up with only one surname?

00:03:10.872 --> 00:03:14.562
Mathematically, the entire population does
converge to only one surname.

00:03:14.733 --> 00:03:19.741
But in real life, if we start out with,
say, 10,000 surnames,

00:03:19.741 --> 00:03:21.814
(and there are actually
much more than that)

00:03:21.814 --> 00:03:26.276
after 40 generations we'd still be left
with over 400.

00:03:26.276 --> 00:03:29.229
Okay, how about 200 generations?

00:03:29.229 --> 00:03:31.639
Still 93 left.

00:03:31.892 --> 00:03:35.006
While the less frequent names
are dying out quickly,

00:03:35.006 --> 00:03:37.882
the more frequent ones
become so widely spread

00:03:37.882 --> 00:03:41.919
that humans will probably cease to exist
before they do.

00:03:41.919 --> 00:03:43.502
The probability of extinction of

00:03:43.502 --> 00:03:47.450
a unique family name that is carried
by only one young couple

00:03:47.450 --> 00:03:50.687
is 45 percent, at least in the West.

00:03:50.687 --> 00:03:53.773
That's the average likelihood of them
having no children

00:03:53.773 --> 00:03:57.843
or only children who won't pass
on the family name.

00:03:57.843 --> 00:04:01.307
But the likelihood of a family name
which is held by multiple couples

00:04:01.307 --> 00:04:07.567
going extinct within one generation
is 45 percent to the power of the number of couples.

00:04:07.572 --> 00:04:13.121
So with a few more people sharing a surname
it becomes very unlikely very quickly

00:04:13.121 --> 00:04:15.581
that this surname should disappear soon.

00:04:15.581 --> 00:04:19.027
If you want to know how often your
family name is currently in use,

00:04:19.027 --> 00:04:22.312
you can find that out on websites such as
Forebears.

00:04:22.312 --> 00:04:24.552
According to the US Census Bureau,

00:04:24.552 --> 00:04:27.261
the most common family names
in the US are currently

00:04:27.261 --> 00:04:29.901
Smith, Johnson, and Williams,

00:04:29.901 --> 00:04:34.836
which together make up for around
2 percent of the entire US population.

00:04:34.836 --> 00:04:37.841
That's, of course, not very impressive
to China.

00:04:37.841 --> 00:04:41.998
In a way, you could say that 
on this timeline, the US is somewhere here

00:04:41.998 --> 00:04:44.845
while China is already over there.

00:04:44.845 --> 00:04:48.572
As fewer family names
become more widely spread,

00:04:48.572 --> 00:04:53.082
we might follow the Chinese feat and
become more creative about given names.

00:04:53.082 --> 00:04:54.900
So instead of Tim Smith, 


00:04:54.900 --> 00:04:57.950
you might be called the
TalentedPeaceful Smith.

00:04:57.959 --> 00:05:01.669
And, instead of Tom, I might be called
The Rest of Us.