[Script Info]
Title: 
[Events]
Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
Dialogue: 0,0:00:00.00,0:00:02.00,Default,,0000,0000,0000,,Erez Lieberman Aiden: Everyone knows
Dialogue: 0,0:00:02.00,0:00:05.00,Default,,0000,0000,0000,,that a picture is worth a thousand words.
Dialogue: 0,0:00:07.00,0:00:09.00,Default,,0000,0000,0000,,But we at Harvard
Dialogue: 0,0:00:09.00,0:00:12.00,Default,,0000,0000,0000,,were wondering if this was really true.
Dialogue: 0,0:00:12.00,0:00:14.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:00:14.00,0:00:18.00,Default,,0000,0000,0000,,So we assembled a team of experts,
Dialogue: 0,0:00:18.00,0:00:20.00,Default,,0000,0000,0000,,spanning Harvard, MIT,
Dialogue: 0,0:00:20.00,0:00:23.00,Default,,0000,0000,0000,,The American Heritage Dictionary, The Encyclopedia Britannica
Dialogue: 0,0:00:23.00,0:00:25.00,Default,,0000,0000,0000,,and even our proud sponsors,
Dialogue: 0,0:00:25.00,0:00:28.00,Default,,0000,0000,0000,,the Google.
Dialogue: 0,0:00:28.00,0:00:30.00,Default,,0000,0000,0000,,And we cogitated about this
Dialogue: 0,0:00:30.00,0:00:32.00,Default,,0000,0000,0000,,for about four years.
Dialogue: 0,0:00:32.00,0:00:37.00,Default,,0000,0000,0000,,And we came to a startling conclusion.
Dialogue: 0,0:00:37.00,0:00:40.00,Default,,0000,0000,0000,,Ladies and gentlemen, a picture is not worth a thousand words.
Dialogue: 0,0:00:40.00,0:00:42.00,Default,,0000,0000,0000,,In fact, we found some pictures
Dialogue: 0,0:00:42.00,0:00:47.00,Default,,0000,0000,0000,,that are worth 500 billion words.
Dialogue: 0,0:00:47.00,0:00:49.00,Default,,0000,0000,0000,,Jean-Baptiste Michel: So how did we get to this conclusion?
Dialogue: 0,0:00:49.00,0:00:51.00,Default,,0000,0000,0000,,So Erez and I were thinking about ways
Dialogue: 0,0:00:51.00,0:00:53.00,Default,,0000,0000,0000,,to get a big picture of human culture
Dialogue: 0,0:00:53.00,0:00:56.00,Default,,0000,0000,0000,,and human history: change over time.
Dialogue: 0,0:00:56.00,0:00:58.00,Default,,0000,0000,0000,,So many books actually have been written over the years.
Dialogue: 0,0:00:58.00,0:01:00.00,Default,,0000,0000,0000,,So we were thinking, well the best way to learn from them
Dialogue: 0,0:01:00.00,0:01:02.00,Default,,0000,0000,0000,,is to read all of these millions of books.
Dialogue: 0,0:01:02.00,0:01:05.00,Default,,0000,0000,0000,,Now of course, if there's a scale for how awesome that is,
Dialogue: 0,0:01:05.00,0:01:08.00,Default,,0000,0000,0000,,that has to rank extremely, extremely high.
Dialogue: 0,0:01:08.00,0:01:10.00,Default,,0000,0000,0000,,Now the problem is there's an X-axis for that,
Dialogue: 0,0:01:10.00,0:01:12.00,Default,,0000,0000,0000,,which is the practical axis.
Dialogue: 0,0:01:12.00,0:01:14.00,Default,,0000,0000,0000,,This is very, very low.
Dialogue: 0,0:01:14.00,0:01:17.00,Default,,0000,0000,0000,,(Applause)
Dialogue: 0,0:01:17.00,0:01:20.00,Default,,0000,0000,0000,,Now people tend to use an alternative approach,
Dialogue: 0,0:01:20.00,0:01:22.00,Default,,0000,0000,0000,,which is to take a few sources and read them very carefully.
Dialogue: 0,0:01:22.00,0:01:24.00,Default,,0000,0000,0000,,This is extremely practical, but not so awesome.
Dialogue: 0,0:01:24.00,0:01:27.00,Default,,0000,0000,0000,,What you really want to do
Dialogue: 0,0:01:27.00,0:01:30.00,Default,,0000,0000,0000,,is to get to the awesome yet practical part of this space.
Dialogue: 0,0:01:30.00,0:01:33.00,Default,,0000,0000,0000,,So it turns out there was a company across the river called Google
Dialogue: 0,0:01:33.00,0:01:35.00,Default,,0000,0000,0000,,who had started a digitization project a few years back
Dialogue: 0,0:01:35.00,0:01:37.00,Default,,0000,0000,0000,,that might just enable this approach.
Dialogue: 0,0:01:37.00,0:01:39.00,Default,,0000,0000,0000,,They have digitized millions of books.
Dialogue: 0,0:01:39.00,0:01:42.00,Default,,0000,0000,0000,,So what that means is, one could use computational methods
Dialogue: 0,0:01:42.00,0:01:44.00,Default,,0000,0000,0000,,to read all of the books in a click of a button.
Dialogue: 0,0:01:44.00,0:01:47.00,Default,,0000,0000,0000,,That's very practical and extremely awesome.
Dialogue: 0,0:01:48.00,0:01:50.00,Default,,0000,0000,0000,,ELA: Let me tell you a little bit about where books come from.
Dialogue: 0,0:01:50.00,0:01:53.00,Default,,0000,0000,0000,,Since time immemorial, there have been authors.
Dialogue: 0,0:01:53.00,0:01:56.00,Default,,0000,0000,0000,,These authors have been striving to write books.
Dialogue: 0,0:01:56.00,0:01:58.00,Default,,0000,0000,0000,,And this became considerably easier
Dialogue: 0,0:01:58.00,0:02:00.00,Default,,0000,0000,0000,,with the development of the printing press some centuries ago.
Dialogue: 0,0:02:00.00,0:02:03.00,Default,,0000,0000,0000,,Since then, the authors have won
Dialogue: 0,0:02:03.00,0:02:05.00,Default,,0000,0000,0000,,on 129 million distinct occasions,
Dialogue: 0,0:02:05.00,0:02:07.00,Default,,0000,0000,0000,,publishing books.
Dialogue: 0,0:02:07.00,0:02:09.00,Default,,0000,0000,0000,,Now if those books are not lost to history,
Dialogue: 0,0:02:09.00,0:02:11.00,Default,,0000,0000,0000,,then they are somewhere in a library,
Dialogue: 0,0:02:11.00,0:02:14.00,Default,,0000,0000,0000,,and many of those books have been getting retrieved from the libraries
Dialogue: 0,0:02:14.00,0:02:16.00,Default,,0000,0000,0000,,and digitized by Google,
Dialogue: 0,0:02:16.00,0:02:18.00,Default,,0000,0000,0000,,which has scanned 15 million books to date.
Dialogue: 0,0:02:18.00,0:02:21.00,Default,,0000,0000,0000,,Now when Google digitizes a book, they put it into a really nice format.
Dialogue: 0,0:02:21.00,0:02:23.00,Default,,0000,0000,0000,,Now we've got the data, plus we have metadata.
Dialogue: 0,0:02:23.00,0:02:26.00,Default,,0000,0000,0000,,We have information about things like where was it published,
Dialogue: 0,0:02:26.00,0:02:28.00,Default,,0000,0000,0000,,who was the author, when was it published.
Dialogue: 0,0:02:28.00,0:02:31.00,Default,,0000,0000,0000,,And what we do is go through all of those records
Dialogue: 0,0:02:31.00,0:02:35.00,Default,,0000,0000,0000,,and exclude everything that's not the highest quality data.
Dialogue: 0,0:02:35.00,0:02:37.00,Default,,0000,0000,0000,,What we're left with
Dialogue: 0,0:02:37.00,0:02:40.00,Default,,0000,0000,0000,,is a collection of five million books,
Dialogue: 0,0:02:40.00,0:02:43.00,Default,,0000,0000,0000,,500 billion words,
Dialogue: 0,0:02:43.00,0:02:45.00,Default,,0000,0000,0000,,a string of characters a thousand times longer
Dialogue: 0,0:02:45.00,0:02:48.00,Default,,0000,0000,0000,,than the human genome --
Dialogue: 0,0:02:48.00,0:02:50.00,Default,,0000,0000,0000,,a text which, when written out,
Dialogue: 0,0:02:50.00,0:02:52.00,Default,,0000,0000,0000,,would stretch from here to the Moon and back
Dialogue: 0,0:02:52.00,0:02:54.00,Default,,0000,0000,0000,,10 times over --
Dialogue: 0,0:02:54.00,0:02:58.00,Default,,0000,0000,0000,,a veritable shard of our cultural genome.
Dialogue: 0,0:02:58.00,0:03:00.00,Default,,0000,0000,0000,,Of course what we did
Dialogue: 0,0:03:00.00,0:03:03.00,Default,,0000,0000,0000,,when faced with such outrageous hyperbole ...
Dialogue: 0,0:03:03.00,0:03:05.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:03:05.00,0:03:08.00,Default,,0000,0000,0000,,was what any self-respecting researchers
Dialogue: 0,0:03:08.00,0:03:11.00,Default,,0000,0000,0000,,would have done.
Dialogue: 0,0:03:11.00,0:03:13.00,Default,,0000,0000,0000,,We took a page out of XKCD,
Dialogue: 0,0:03:13.00,0:03:15.00,Default,,0000,0000,0000,,and we said, "Stand back.
Dialogue: 0,0:03:15.00,0:03:17.00,Default,,0000,0000,0000,,We're going to try science."
Dialogue: 0,0:03:17.00,0:03:19.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:03:19.00,0:03:21.00,Default,,0000,0000,0000,,JM: Now of course, we were thinking,
Dialogue: 0,0:03:21.00,0:03:23.00,Default,,0000,0000,0000,,well let's just first put the data out there
Dialogue: 0,0:03:23.00,0:03:25.00,Default,,0000,0000,0000,,for people to do science to it.
Dialogue: 0,0:03:25.00,0:03:27.00,Default,,0000,0000,0000,,Now we're thinking, what data can we release?
Dialogue: 0,0:03:27.00,0:03:29.00,Default,,0000,0000,0000,,Well of course, you want to take the books
Dialogue: 0,0:03:29.00,0:03:31.00,Default,,0000,0000,0000,,and release the full text of these five million books.
Dialogue: 0,0:03:31.00,0:03:33.00,Default,,0000,0000,0000,,Now Google, and Jon Orwant in particular,
Dialogue: 0,0:03:33.00,0:03:35.00,Default,,0000,0000,0000,,told us a little equation that we should learn.
Dialogue: 0,0:03:35.00,0:03:38.00,Default,,0000,0000,0000,,So you have five million, that is, five million authors
Dialogue: 0,0:03:38.00,0:03:41.00,Default,,0000,0000,0000,,and five million plaintiffs is a massive lawsuit.
Dialogue: 0,0:03:41.00,0:03:43.00,Default,,0000,0000,0000,,So, although that would be really, really awesome,
Dialogue: 0,0:03:43.00,0:03:46.00,Default,,0000,0000,0000,,again, that's extremely, extremely impractical.
Dialogue: 0,0:03:46.00,0:03:48.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:03:48.00,0:03:50.00,Default,,0000,0000,0000,,Now again, we kind of caved in,
Dialogue: 0,0:03:50.00,0:03:53.00,Default,,0000,0000,0000,,and we did the very practical approach, which was a bit less awesome.
Dialogue: 0,0:03:53.00,0:03:55.00,Default,,0000,0000,0000,,We said, well instead of releasing the full text,
Dialogue: 0,0:03:55.00,0:03:57.00,Default,,0000,0000,0000,,we're going to release statistics about the books.
Dialogue: 0,0:03:57.00,0:03:59.00,Default,,0000,0000,0000,,So take for instance "A gleam of happiness."
Dialogue: 0,0:03:59.00,0:04:01.00,Default,,0000,0000,0000,,It's four words; we call that a four-gram.
Dialogue: 0,0:04:01.00,0:04:03.00,Default,,0000,0000,0000,,We're going to tell you how many times a particular four-gram
Dialogue: 0,0:04:03.00,0:04:05.00,Default,,0000,0000,0000,,appeared in books in 1801, 1802, 1803,
Dialogue: 0,0:04:05.00,0:04:07.00,Default,,0000,0000,0000,,all the way up to 2008.
Dialogue: 0,0:04:07.00,0:04:09.00,Default,,0000,0000,0000,,That gives us a time series
Dialogue: 0,0:04:09.00,0:04:11.00,Default,,0000,0000,0000,,of how frequently this particular sentence was used over time.
Dialogue: 0,0:04:11.00,0:04:14.00,Default,,0000,0000,0000,,We do that for all the words and phrases that appear in those books,
Dialogue: 0,0:04:14.00,0:04:17.00,Default,,0000,0000,0000,,and that gives us a big table of two billion lines
Dialogue: 0,0:04:17.00,0:04:19.00,Default,,0000,0000,0000,,that tell us about the way culture has been changing.
Dialogue: 0,0:04:19.00,0:04:21.00,Default,,0000,0000,0000,,ELA: So those two billion lines,
Dialogue: 0,0:04:21.00,0:04:23.00,Default,,0000,0000,0000,,we call them two billion n-grams.
Dialogue: 0,0:04:23.00,0:04:25.00,Default,,0000,0000,0000,,What do they tell us?
Dialogue: 0,0:04:25.00,0:04:27.00,Default,,0000,0000,0000,,Well the individual n-grams measure cultural trends.
Dialogue: 0,0:04:27.00,0:04:29.00,Default,,0000,0000,0000,,Let me give you an example.
Dialogue: 0,0:04:29.00,0:04:31.00,Default,,0000,0000,0000,,Let's suppose that I am thriving,
Dialogue: 0,0:04:31.00,0:04:33.00,Default,,0000,0000,0000,,then tomorrow I want to tell you about how well I did.
Dialogue: 0,0:04:33.00,0:04:36.00,Default,,0000,0000,0000,,And so I might say, "Yesterday, I throve."
Dialogue: 0,0:04:36.00,0:04:39.00,Default,,0000,0000,0000,,Alternatively, I could say, "Yesterday, I thrived."
Dialogue: 0,0:04:39.00,0:04:42.00,Default,,0000,0000,0000,,Well which one should I use?
Dialogue: 0,0:04:42.00,0:04:44.00,Default,,0000,0000,0000,,How to know?
Dialogue: 0,0:04:44.00,0:04:46.00,Default,,0000,0000,0000,,As of about six months ago,
Dialogue: 0,0:04:46.00,0:04:48.00,Default,,0000,0000,0000,,the state of the art in this field
Dialogue: 0,0:04:48.00,0:04:50.00,Default,,0000,0000,0000,,is that you would, for instance,
Dialogue: 0,0:04:50.00,0:04:52.00,Default,,0000,0000,0000,,go up to the following psychologist with fabulous hair,
Dialogue: 0,0:04:52.00,0:04:54.00,Default,,0000,0000,0000,,and you'd say,
Dialogue: 0,0:04:54.00,0:04:57.00,Default,,0000,0000,0000,,"Steve, you're an expert on the irregular verbs.
Dialogue: 0,0:04:57.00,0:04:59.00,Default,,0000,0000,0000,,What should I do?"
Dialogue: 0,0:04:59.00,0:05:01.00,Default,,0000,0000,0000,,And he'd tell you, "Well most people say thrived,
Dialogue: 0,0:05:01.00,0:05:04.00,Default,,0000,0000,0000,,but some people say throve."
Dialogue: 0,0:05:04.00,0:05:06.00,Default,,0000,0000,0000,,And you also knew, more or less,
Dialogue: 0,0:05:06.00,0:05:09.00,Default,,0000,0000,0000,,that if you were to go back in time 200 years
Dialogue: 0,0:05:09.00,0:05:12.00,Default,,0000,0000,0000,,and ask the following statesman with equally fabulous hair,
Dialogue: 0,0:05:12.00,0:05:15.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:05:15.00,0:05:17.00,Default,,0000,0000,0000,,"Tom, what should I say?"
Dialogue: 0,0:05:17.00,0:05:19.00,Default,,0000,0000,0000,,He'd say, "Well, in my day, most people throve,
Dialogue: 0,0:05:19.00,0:05:22.00,Default,,0000,0000,0000,,but some thrived."
Dialogue: 0,0:05:22.00,0:05:24.00,Default,,0000,0000,0000,,So now what I'm just going to show you is raw data.
Dialogue: 0,0:05:24.00,0:05:28.00,Default,,0000,0000,0000,,Two rows from this table of two billion entries.
Dialogue: 0,0:05:28.00,0:05:30.00,Default,,0000,0000,0000,,What you're seeing is year by year frequency
Dialogue: 0,0:05:30.00,0:05:33.00,Default,,0000,0000,0000,,of "thrived" and "throve" over time.
Dialogue: 0,0:05:34.00,0:05:36.00,Default,,0000,0000,0000,,Now this is just two
Dialogue: 0,0:05:36.00,0:05:39.00,Default,,0000,0000,0000,,out of two billion rows.
Dialogue: 0,0:05:39.00,0:05:41.00,Default,,0000,0000,0000,,So the entire data set
Dialogue: 0,0:05:41.00,0:05:44.00,Default,,0000,0000,0000,,is a billion times more awesome than this slide.
Dialogue: 0,0:05:44.00,0:05:46.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:05:46.00,0:05:50.00,Default,,0000,0000,0000,,(Applause)
Dialogue: 0,0:05:50.00,0:05:52.00,Default,,0000,0000,0000,,JM: Now there are many other pictures that are worth 500 billion words.
Dialogue: 0,0:05:52.00,0:05:54.00,Default,,0000,0000,0000,,For instance, this one.
Dialogue: 0,0:05:54.00,0:05:56.00,Default,,0000,0000,0000,,If you just take influenza,
Dialogue: 0,0:05:56.00,0:05:58.00,Default,,0000,0000,0000,,you will see peaks at the time where you knew
Dialogue: 0,0:05:58.00,0:06:01.00,Default,,0000,0000,0000,,big flu epidemics were killing people around the globe.
Dialogue: 0,0:06:01.00,0:06:04.00,Default,,0000,0000,0000,,ELA: If you were not yet convinced,
Dialogue: 0,0:06:04.00,0:06:06.00,Default,,0000,0000,0000,,sea levels are rising,
Dialogue: 0,0:06:06.00,0:06:09.00,Default,,0000,0000,0000,,so is atmospheric CO2 and global temperature.
Dialogue: 0,0:06:09.00,0:06:12.00,Default,,0000,0000,0000,,JM: You might also want to have a look at this particular n-gram,
Dialogue: 0,0:06:12.00,0:06:15.00,Default,,0000,0000,0000,,and that's to tell Nietzsche that God is not dead,
Dialogue: 0,0:06:15.00,0:06:18.00,Default,,0000,0000,0000,,although you might agree that he might need a better publicist.
Dialogue: 0,0:06:18.00,0:06:20.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:06:20.00,0:06:23.00,Default,,0000,0000,0000,,ELA: You can get at some pretty abstract concepts with this sort of thing.
Dialogue: 0,0:06:23.00,0:06:25.00,Default,,0000,0000,0000,,For instance, let me tell you the history
Dialogue: 0,0:06:25.00,0:06:27.00,Default,,0000,0000,0000,,of the year 1950.
Dialogue: 0,0:06:27.00,0:06:29.00,Default,,0000,0000,0000,,Pretty much for the vast majority of history,
Dialogue: 0,0:06:29.00,0:06:31.00,Default,,0000,0000,0000,,no one gave a damn about 1950.
Dialogue: 0,0:06:31.00,0:06:33.00,Default,,0000,0000,0000,,In 1700, in 1800, in 1900,
Dialogue: 0,0:06:33.00,0:06:36.00,Default,,0000,0000,0000,,no one cared.
Dialogue: 0,0:06:37.00,0:06:39.00,Default,,0000,0000,0000,,Through the 30s and 40s,
Dialogue: 0,0:06:39.00,0:06:41.00,Default,,0000,0000,0000,,no one cared.
Dialogue: 0,0:06:41.00,0:06:43.00,Default,,0000,0000,0000,,Suddenly, in the mid-40s,
Dialogue: 0,0:06:43.00,0:06:45.00,Default,,0000,0000,0000,,there started to be a buzz.
Dialogue: 0,0:06:45.00,0:06:47.00,Default,,0000,0000,0000,,People realized that 1950 was going to happen,
Dialogue: 0,0:06:47.00,0:06:49.00,Default,,0000,0000,0000,,and it could be big.
Dialogue: 0,0:06:49.00,0:06:52.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:06:52.00,0:06:55.00,Default,,0000,0000,0000,,But nothing got people interested in 1950
Dialogue: 0,0:06:55.00,0:06:58.00,Default,,0000,0000,0000,,like the year 1950.
Dialogue: 0,0:06:58.00,0:07:01.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:07:01.00,0:07:03.00,Default,,0000,0000,0000,,People were walking around obsessed.
Dialogue: 0,0:07:03.00,0:07:05.00,Default,,0000,0000,0000,,They couldn't stop talking
Dialogue: 0,0:07:05.00,0:07:08.00,Default,,0000,0000,0000,,about all the things they did in 1950,
Dialogue: 0,0:07:08.00,0:07:11.00,Default,,0000,0000,0000,,all the things they were planning to do in 1950,
Dialogue: 0,0:07:11.00,0:07:16.00,Default,,0000,0000,0000,,all the dreams of what they wanted to accomplish in 1950.
Dialogue: 0,0:07:16.00,0:07:18.00,Default,,0000,0000,0000,,In fact, 1950 was so fascinating
Dialogue: 0,0:07:18.00,0:07:20.00,Default,,0000,0000,0000,,that for years thereafter,
Dialogue: 0,0:07:20.00,0:07:23.00,Default,,0000,0000,0000,,people just kept talking about all the amazing things that happened,
Dialogue: 0,0:07:23.00,0:07:25.00,Default,,0000,0000,0000,,in '51, '52, '53.
Dialogue: 0,0:07:25.00,0:07:27.00,Default,,0000,0000,0000,,Finally in 1954,
Dialogue: 0,0:07:27.00,0:07:29.00,Default,,0000,0000,0000,,someone woke up and realized
Dialogue: 0,0:07:29.00,0:07:33.00,Default,,0000,0000,0000,,that 1950 had gotten somewhat passé.
Dialogue: 0,0:07:33.00,0:07:35.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:07:35.00,0:07:37.00,Default,,0000,0000,0000,,And just like that, the bubble burst.
Dialogue: 0,0:07:37.00,0:07:39.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:07:39.00,0:07:41.00,Default,,0000,0000,0000,,And the story of 1950
Dialogue: 0,0:07:41.00,0:07:43.00,Default,,0000,0000,0000,,is the story of every year that we have on record,
Dialogue: 0,0:07:43.00,0:07:46.00,Default,,0000,0000,0000,,with a little twist, because now we've got these nice charts.
Dialogue: 0,0:07:46.00,0:07:49.00,Default,,0000,0000,0000,,And because we have these nice charts, we can measure things.
Dialogue: 0,0:07:49.00,0:07:51.00,Default,,0000,0000,0000,,We can say, "Well how fast does the bubble burst?"
Dialogue: 0,0:07:51.00,0:07:54.00,Default,,0000,0000,0000,,And it turns out that we can measure that very precisely.
Dialogue: 0,0:07:54.00,0:07:57.00,Default,,0000,0000,0000,,Equations were derived, graphs were produced,
Dialogue: 0,0:07:57.00,0:07:59.00,Default,,0000,0000,0000,,and the net result
Dialogue: 0,0:07:59.00,0:08:02.00,Default,,0000,0000,0000,,is that we find that the bubble bursts faster and faster
Dialogue: 0,0:08:02.00,0:08:04.00,Default,,0000,0000,0000,,with each passing year.
Dialogue: 0,0:08:04.00,0:08:09.00,Default,,0000,0000,0000,,We are losing interest in the past more rapidly.
Dialogue: 0,0:08:09.00,0:08:11.00,Default,,0000,0000,0000,,JM: Now a little piece of career advice.
Dialogue: 0,0:08:11.00,0:08:13.00,Default,,0000,0000,0000,,So for those of you who seek to be famous,
Dialogue: 0,0:08:13.00,0:08:15.00,Default,,0000,0000,0000,,we can learn from the 25 most famous political figures,
Dialogue: 0,0:08:15.00,0:08:17.00,Default,,0000,0000,0000,,authors, actors and so on.
Dialogue: 0,0:08:17.00,0:08:20.00,Default,,0000,0000,0000,,So if you want to become famous early on, you should be an actor,
Dialogue: 0,0:08:20.00,0:08:22.00,Default,,0000,0000,0000,,because then fame starts rising by the end of your 20s --
Dialogue: 0,0:08:22.00,0:08:24.00,Default,,0000,0000,0000,,you're still young, it's really great.
Dialogue: 0,0:08:24.00,0:08:26.00,Default,,0000,0000,0000,,Now if you can wait a little bit, you should be an author,
Dialogue: 0,0:08:26.00,0:08:28.00,Default,,0000,0000,0000,,because then you rise to very great heights,
Dialogue: 0,0:08:28.00,0:08:30.00,Default,,0000,0000,0000,,like Mark Twain, for instance: extremely famous.
Dialogue: 0,0:08:30.00,0:08:32.00,Default,,0000,0000,0000,,But if you want to reach the very top,
Dialogue: 0,0:08:32.00,0:08:34.00,Default,,0000,0000,0000,,you should delay gratification
Dialogue: 0,0:08:34.00,0:08:36.00,Default,,0000,0000,0000,,and, of course, become a politician.
Dialogue: 0,0:08:36.00,0:08:38.00,Default,,0000,0000,0000,,So here you will become famous by the end of your 50s,
Dialogue: 0,0:08:38.00,0:08:40.00,Default,,0000,0000,0000,,and become very, very famous afterward.
Dialogue: 0,0:08:40.00,0:08:43.00,Default,,0000,0000,0000,,So scientists also tend to get famous when they're much older.
Dialogue: 0,0:08:43.00,0:08:45.00,Default,,0000,0000,0000,,Like for instance, biologists and physics
Dialogue: 0,0:08:45.00,0:08:47.00,Default,,0000,0000,0000,,tend to be almost as famous as actors.
Dialogue: 0,0:08:47.00,0:08:50.00,Default,,0000,0000,0000,,One mistake you should not do is become a mathematician.
Dialogue: 0,0:08:50.00,0:08:52.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:08:52.00,0:08:54.00,Default,,0000,0000,0000,,If you do that,
Dialogue: 0,0:08:54.00,0:08:57.00,Default,,0000,0000,0000,,you might think, "Oh great. I'm going to do my best work when I'm in my 20s."
Dialogue: 0,0:08:57.00,0:08:59.00,Default,,0000,0000,0000,,But guess what, nobody will really care.
Dialogue: 0,0:08:59.00,0:09:02.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:09:02.00,0:09:04.00,Default,,0000,0000,0000,,ELA: There are more sobering notes
Dialogue: 0,0:09:04.00,0:09:06.00,Default,,0000,0000,0000,,among the n-grams.
Dialogue: 0,0:09:06.00,0:09:08.00,Default,,0000,0000,0000,,For instance, here's the trajectory of Marc Chagall,
Dialogue: 0,0:09:08.00,0:09:10.00,Default,,0000,0000,0000,,an artist born in 1887.
Dialogue: 0,0:09:10.00,0:09:13.00,Default,,0000,0000,0000,,And this looks like the normal trajectory of a famous person.
Dialogue: 0,0:09:13.00,0:09:17.00,Default,,0000,0000,0000,,He gets more and more and more famous,
Dialogue: 0,0:09:17.00,0:09:19.00,Default,,0000,0000,0000,,except if you look in German.
Dialogue: 0,0:09:19.00,0:09:21.00,Default,,0000,0000,0000,,If you look in German, you see something completely bizarre,
Dialogue: 0,0:09:21.00,0:09:23.00,Default,,0000,0000,0000,,something you pretty much never see,
Dialogue: 0,0:09:23.00,0:09:25.00,Default,,0000,0000,0000,,which is he becomes extremely famous
Dialogue: 0,0:09:25.00,0:09:27.00,Default,,0000,0000,0000,,and then all of a sudden plummets,
Dialogue: 0,0:09:27.00,0:09:30.00,Default,,0000,0000,0000,,going through a nadir between 1933 and 1945,
Dialogue: 0,0:09:30.00,0:09:33.00,Default,,0000,0000,0000,,before rebounding afterward.
Dialogue: 0,0:09:33.00,0:09:35.00,Default,,0000,0000,0000,,And of course, what we're seeing
Dialogue: 0,0:09:35.00,0:09:38.00,Default,,0000,0000,0000,,is the fact Marc Chagall was a Jewish artist
Dialogue: 0,0:09:38.00,0:09:40.00,Default,,0000,0000,0000,,in Nazi Germany.
Dialogue: 0,0:09:40.00,0:09:42.00,Default,,0000,0000,0000,,Now these signals
Dialogue: 0,0:09:42.00,0:09:44.00,Default,,0000,0000,0000,,are actually so strong
Dialogue: 0,0:09:44.00,0:09:47.00,Default,,0000,0000,0000,,that we don't need to know that someone was censored.
Dialogue: 0,0:09:47.00,0:09:49.00,Default,,0000,0000,0000,,We can actually figure it out
Dialogue: 0,0:09:49.00,0:09:51.00,Default,,0000,0000,0000,,using really basic signal processing.
Dialogue: 0,0:09:51.00,0:09:53.00,Default,,0000,0000,0000,,Here's a simple way to do it.
Dialogue: 0,0:09:53.00,0:09:55.00,Default,,0000,0000,0000,,Well, a reasonable expectation
Dialogue: 0,0:09:55.00,0:09:57.00,Default,,0000,0000,0000,,is that somebody's fame in a given period of time
Dialogue: 0,0:09:57.00,0:09:59.00,Default,,0000,0000,0000,,should be roughly the average of their fame before
Dialogue: 0,0:09:59.00,0:10:01.00,Default,,0000,0000,0000,,and their fame after.
Dialogue: 0,0:10:01.00,0:10:03.00,Default,,0000,0000,0000,,So that's sort of what we expect.
Dialogue: 0,0:10:03.00,0:10:06.00,Default,,0000,0000,0000,,And we compare that to the fame that we observe.
Dialogue: 0,0:10:06.00,0:10:08.00,Default,,0000,0000,0000,,And we just divide one by the other
Dialogue: 0,0:10:08.00,0:10:10.00,Default,,0000,0000,0000,,to produce something we call a suppression index.
Dialogue: 0,0:10:10.00,0:10:13.00,Default,,0000,0000,0000,,If the suppression index is very, very, very small,
Dialogue: 0,0:10:13.00,0:10:15.00,Default,,0000,0000,0000,,then you very well might be being suppressed.
Dialogue: 0,0:10:15.00,0:10:18.00,Default,,0000,0000,0000,,If it's very large, maybe you're benefiting from propaganda.
Dialogue: 0,0:10:19.00,0:10:21.00,Default,,0000,0000,0000,,JM: Now you can actually look at
Dialogue: 0,0:10:21.00,0:10:24.00,Default,,0000,0000,0000,,the distribution of suppression indexes over whole populations.
Dialogue: 0,0:10:24.00,0:10:26.00,Default,,0000,0000,0000,,So for instance, here --
Dialogue: 0,0:10:26.00,0:10:28.00,Default,,0000,0000,0000,,this suppression index is for 5,000 people
Dialogue: 0,0:10:28.00,0:10:30.00,Default,,0000,0000,0000,,picked in English books where there's no known suppression --
Dialogue: 0,0:10:30.00,0:10:32.00,Default,,0000,0000,0000,,it would be like this, basically tightly centered on one.
Dialogue: 0,0:10:32.00,0:10:34.00,Default,,0000,0000,0000,,What you expect is basically what you observe.
Dialogue: 0,0:10:34.00,0:10:36.00,Default,,0000,0000,0000,,This is distribution as seen in Germany --
Dialogue: 0,0:10:36.00,0:10:38.00,Default,,0000,0000,0000,,very different, it's shifted to the left.
Dialogue: 0,0:10:38.00,0:10:41.00,Default,,0000,0000,0000,,People talked about it twice less as it should have been.
Dialogue: 0,0:10:41.00,0:10:43.00,Default,,0000,0000,0000,,But much more importantly, the distribution is much wider.
Dialogue: 0,0:10:43.00,0:10:46.00,Default,,0000,0000,0000,,There are many people who end up on the far left on this distribution
Dialogue: 0,0:10:46.00,0:10:49.00,Default,,0000,0000,0000,,who are talked about 10 times fewer than they should have been.
Dialogue: 0,0:10:49.00,0:10:51.00,Default,,0000,0000,0000,,But then also many people on the far right
Dialogue: 0,0:10:51.00,0:10:53.00,Default,,0000,0000,0000,,who seem to benefit from propaganda.
Dialogue: 0,0:10:53.00,0:10:56.00,Default,,0000,0000,0000,,This picture is the hallmark of censorship in the book record.
Dialogue: 0,0:10:56.00,0:10:58.00,Default,,0000,0000,0000,,ELA: So culturomics
Dialogue: 0,0:10:58.00,0:11:00.00,Default,,0000,0000,0000,,is what we call this method.
Dialogue: 0,0:11:00.00,0:11:02.00,Default,,0000,0000,0000,,It's kind of like genomics.
Dialogue: 0,0:11:02.00,0:11:04.00,Default,,0000,0000,0000,,Except genomics is a lens on biology
Dialogue: 0,0:11:04.00,0:11:07.00,Default,,0000,0000,0000,,through the window of the sequence of bases in the human genome.
Dialogue: 0,0:11:07.00,0:11:09.00,Default,,0000,0000,0000,,Culturomics is similar.
Dialogue: 0,0:11:09.00,0:11:12.00,Default,,0000,0000,0000,,It's the application of massive-scale data collection analysis
Dialogue: 0,0:11:12.00,0:11:14.00,Default,,0000,0000,0000,,to the study of human culture.
Dialogue: 0,0:11:14.00,0:11:16.00,Default,,0000,0000,0000,,Here, instead of through the lens of a genome,
Dialogue: 0,0:11:16.00,0:11:19.00,Default,,0000,0000,0000,,through the lens of digitized pieces of the historical record.
Dialogue: 0,0:11:19.00,0:11:21.00,Default,,0000,0000,0000,,The great thing about culturomics
Dialogue: 0,0:11:21.00,0:11:23.00,Default,,0000,0000,0000,,is that everyone can do it.
Dialogue: 0,0:11:23.00,0:11:25.00,Default,,0000,0000,0000,,Why can everyone do it?
Dialogue: 0,0:11:25.00,0:11:27.00,Default,,0000,0000,0000,,Everyone can do it because three guys,
Dialogue: 0,0:11:27.00,0:11:30.00,Default,,0000,0000,0000,,Jon Orwant, Matt Gray and Will Brockman over at Google,
Dialogue: 0,0:11:30.00,0:11:32.00,Default,,0000,0000,0000,,saw the prototype of the Ngram Viewer,
Dialogue: 0,0:11:32.00,0:11:34.00,Default,,0000,0000,0000,,and they said, "This is so fun.
Dialogue: 0,0:11:34.00,0:11:37.00,Default,,0000,0000,0000,,We have to make this available for people."
Dialogue: 0,0:11:37.00,0:11:39.00,Default,,0000,0000,0000,,So in two weeks flat -- the two weeks before our paper came out --
Dialogue: 0,0:11:39.00,0:11:42.00,Default,,0000,0000,0000,,they coded up a version of the Ngram Viewer for the general public.
Dialogue: 0,0:11:42.00,0:11:45.00,Default,,0000,0000,0000,,And so you too can type in any word or phrase that you're interested in
Dialogue: 0,0:11:45.00,0:11:47.00,Default,,0000,0000,0000,,and see its n-gram immediately --
Dialogue: 0,0:11:47.00,0:11:49.00,Default,,0000,0000,0000,,also browse examples of all the various books
Dialogue: 0,0:11:49.00,0:11:51.00,Default,,0000,0000,0000,,in which your n-gram appears.
Dialogue: 0,0:11:51.00,0:11:53.00,Default,,0000,0000,0000,,JM: Now this was used over a million times on the first day,
Dialogue: 0,0:11:53.00,0:11:55.00,Default,,0000,0000,0000,,and this is really the best of all the queries.
Dialogue: 0,0:11:55.00,0:11:58.00,Default,,0000,0000,0000,,So people want to be their best, put their best foot forward.
Dialogue: 0,0:11:58.00,0:12:01.00,Default,,0000,0000,0000,,But it turns out in the 18th century, people didn't really care about that at all.
Dialogue: 0,0:12:01.00,0:12:04.00,Default,,0000,0000,0000,,They didn't want to be their best, they wanted to be their beft.
Dialogue: 0,0:12:04.00,0:12:07.00,Default,,0000,0000,0000,,So what happened is, of course, this is just a mistake.
Dialogue: 0,0:12:07.00,0:12:09.00,Default,,0000,0000,0000,,It's not that strove for mediocrity,
Dialogue: 0,0:12:09.00,0:12:12.00,Default,,0000,0000,0000,,it's just that the S used to be written differently, kind of like an F.
Dialogue: 0,0:12:12.00,0:12:15.00,Default,,0000,0000,0000,,Now of course, Google didn't pick this up at the time,
Dialogue: 0,0:12:15.00,0:12:18.00,Default,,0000,0000,0000,,so we reported this in the science article that we wrote.
Dialogue: 0,0:12:18.00,0:12:20.00,Default,,0000,0000,0000,,But it turns out this is just a reminder
Dialogue: 0,0:12:20.00,0:12:22.00,Default,,0000,0000,0000,,that, although this is a lot of fun,
Dialogue: 0,0:12:22.00,0:12:24.00,Default,,0000,0000,0000,,when you interpret these graphs, you have to be very careful,
Dialogue: 0,0:12:24.00,0:12:27.00,Default,,0000,0000,0000,,and you have to adopt the base standards in the sciences.
Dialogue: 0,0:12:27.00,0:12:30.00,Default,,0000,0000,0000,,ELA: People have been using this for all kinds of fun purposes.
Dialogue: 0,0:12:30.00,0:12:37.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:12:37.00,0:12:39.00,Default,,0000,0000,0000,,Actually, we're not going to have to talk,
Dialogue: 0,0:12:39.00,0:12:42.00,Default,,0000,0000,0000,,we're just going to show you all the slides and remain silent.
Dialogue: 0,0:12:42.00,0:12:45.00,Default,,0000,0000,0000,,This person was interested in the history of frustration.
Dialogue: 0,0:12:45.00,0:12:48.00,Default,,0000,0000,0000,,There's various types of frustration.
Dialogue: 0,0:12:48.00,0:12:51.00,Default,,0000,0000,0000,,If you stub your toe, that's a one A "argh."
Dialogue: 0,0:12:51.00,0:12:53.00,Default,,0000,0000,0000,,If the planet Earth is annihilated by the Vogons
Dialogue: 0,0:12:53.00,0:12:55.00,Default,,0000,0000,0000,,to make room for an interstellar bypass,
Dialogue: 0,0:12:55.00,0:12:57.00,Default,,0000,0000,0000,,that's an eight A "aaaaaaaargh."
Dialogue: 0,0:12:57.00,0:12:59.00,Default,,0000,0000,0000,,This person studies all the "arghs,"
Dialogue: 0,0:12:59.00,0:13:01.00,Default,,0000,0000,0000,,from one through eight A's.
Dialogue: 0,0:13:01.00,0:13:03.00,Default,,0000,0000,0000,,And it turns out
Dialogue: 0,0:13:03.00,0:13:05.00,Default,,0000,0000,0000,,that the less-frequent "arghs"
Dialogue: 0,0:13:05.00,0:13:08.00,Default,,0000,0000,0000,,are, of course, the ones that correspond to things that are more frustrating --
Dialogue: 0,0:13:08.00,0:13:11.00,Default,,0000,0000,0000,,except, oddly, in the early 80s.
Dialogue: 0,0:13:11.00,0:13:13.00,Default,,0000,0000,0000,,We think that might have something to do with Reagan.
Dialogue: 0,0:13:13.00,0:13:15.00,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:13:15.00,0:13:18.00,Default,,0000,0000,0000,,JM: There are many usages of this data,
Dialogue: 0,0:13:18.00,0:13:21.00,Default,,0000,0000,0000,,but the bottom line is that the historical record is being digitized.
Dialogue: 0,0:13:21.00,0:13:23.00,Default,,0000,0000,0000,,Google has started to digitize 15 million books.
Dialogue: 0,0:13:23.00,0:13:25.00,Default,,0000,0000,0000,,That's 12 percent of all the books that have ever been published.
Dialogue: 0,0:13:25.00,0:13:28.00,Default,,0000,0000,0000,,It's a sizable chunk of human culture.
Dialogue: 0,0:13:28.00,0:13:31.00,Default,,0000,0000,0000,,There's much more in culture: there's manuscripts, there newspapers,
Dialogue: 0,0:13:31.00,0:13:33.00,Default,,0000,0000,0000,,there's things that are not text, like art and paintings.
Dialogue: 0,0:13:33.00,0:13:35.00,Default,,0000,0000,0000,,These all happen to be on our computers,
Dialogue: 0,0:13:35.00,0:13:37.00,Default,,0000,0000,0000,,on computers across the world.
Dialogue: 0,0:13:37.00,0:13:40.00,Default,,0000,0000,0000,,And when that happens, that will transform the way we have
Dialogue: 0,0:13:40.00,0:13:42.00,Default,,0000,0000,0000,,to understand our past, our present and human culture.
Dialogue: 0,0:13:42.00,0:13:44.00,Default,,0000,0000,0000,,Thank you very much.
Dialogue: 0,0:13:44.00,0:13:47.00,Default,,0000,0000,0000,,(Applause)