0:00:00.471,0:00:02.955 The great texts of the ancient world 0:00:02.955,0:00:06.270 don't survive to us in their original form. 0:00:06.270,0:00:09.789 They survive because medieval scribes copied them 0:00:09.789,0:00:11.788 and copied them and copied them. 0:00:11.788,0:00:14.438 And so it is with Archimedes, 0:00:14.438,0:00:16.204 the great Greek mathematician. 0:00:16.204,0:00:19.206 Everything we know about Archimedes as a mathematician 0:00:19.206,0:00:21.904 we know about because of just three books, 0:00:21.904,0:00:23.820 and they're called A, B and C. 0:00:23.820,0:00:28.653 And A was lost by an Italian humanist in 1564. 0:00:28.653,0:00:31.456 And B was last heard of in the Pope's Library 0:00:31.456,0:00:37.288 about a hundred miles north of Rome in Viterbo in 1311. 0:00:37.288,0:00:41.585 Now Codex C was only discovered in 1906, 0:00:41.585,0:00:44.134 and it landed on my desk in Baltimore 0:00:44.134,0:00:48.260 on the 19th of January, 1999. 0:00:48.260,0:00:50.678 And this is Codex C here. 0:00:50.678,0:00:55.847 Now Codex C is actually buried in this book. 0:00:55.847,0:00:57.297 It's buried treasure. 0:00:57.297,0:00:59.632 Because this book is actually a prayer book. 0:00:59.632,0:01:02.596 It was finished by a guy called Johannes Myrones 0:01:02.596,0:01:06.213 on the 14th of April, 1229. 0:01:06.213,0:01:08.698 And to make his prayer book he used parchment. 0:01:08.698,0:01:11.113 But he didn't use new parchment, 0:01:11.113,0:01:14.427 he used parchment recycled from earlier manuscripts, 0:01:14.427,0:01:16.164 and there were seven of them. 0:01:16.164,0:01:19.647 And Archimedes Codex C was just one of those seven. 0:01:19.647,0:01:25.114 He took apart the Archimedes manuscript and the other seven manuscripts. 0:01:25.114,0:01:28.497 He erased all of their texts, 0:01:28.497,0:01:30.915 and then he cut the sheets down in the middle, 0:01:30.915,0:01:33.159 he shuffled them up, 0:01:33.159,0:01:35.077 and he rotated them 90 degrees, 0:01:35.077,0:01:37.129 and he wrote prayers on top of these books. 0:01:37.129,0:01:38.826 And essentially these seven manuscripts 0:01:38.826,0:01:43.174 disappeared for 700 years, and we have a prayer book. 0:01:43.174,0:01:46.423 The prayer book was discovered by this guy, 0:01:46.423,0:01:49.150 Johan Ludvig Heiberg, in 1906. 0:01:49.150,0:01:50.591 And with just a magnifying glass, 0:01:50.591,0:01:53.475 he transcribed as much of the text as he could. 0:01:53.475,0:01:56.657 And the thing is that he found two texts in this manuscript 0:01:56.657,0:01:58.425 that were unique texts. 0:01:58.425,0:02:00.025 They weren't in A and B at all; 0:02:00.025,0:02:01.723 they were completely new texts by Archimedes, 0:02:01.723,0:02:04.958 and they were called "The Method" and "The Stomachion." 0:02:04.958,0:02:07.227 And it became a world famous manuscript. 0:02:07.227,0:02:09.042 Now it should be clear by now 0:02:09.042,0:02:11.756 that this book is in bad condition. 0:02:11.756,0:02:15.226 It got in worse condition in the 20th century 0:02:15.226,0:02:16.958 after Heiberg saw it. 0:02:16.958,0:02:18.475 Forgeries were painted over it, 0:02:18.475,0:02:21.558 and it suffered very badly from mold. 0:02:21.558,0:02:25.308 This book is the definition of a write-off. 0:02:25.308,0:02:27.042 It's the sort of book 0:02:27.042,0:02:29.707 that you thought would be in an institution. 0:02:29.707,0:02:31.893 But it's not in an institution, 0:02:31.893,0:02:36.525 it was bought by a private owner in 1998. 0:02:36.525,0:02:38.275 Why did he buy this book? 0:02:38.275,0:02:41.812 Because he wanted to make that which was fragile safe. 0:02:41.812,0:02:45.058 He wanted to make that which was unique ubiquitous. 0:02:45.058,0:02:49.724 He wanted to make that which was expensive free. 0:02:49.724,0:02:52.859 And he wanted to do this as a matter of principle. 0:02:52.859,0:02:57.391 Because not many people are really going to read Archimedes in ancient Greek, 0:02:57.391,0:03:00.075 but they should have the chance to do it. 0:03:00.075,0:03:03.374 So he gathered around himself the friends of Archimedes, 0:03:03.374,0:03:05.510 and he promised to pay for all the work. 0:03:05.510,0:03:07.208 And it was an expensive job, 0:03:07.208,0:03:10.678 but actually it wouldn't be as much as you think 0:03:10.678,0:03:12.843 because these people, they didn't come from money, 0:03:12.843,0:03:14.662 they came from Archimedes. 0:03:14.662,0:03:16.378 And they came from all sorts of different backgrounds. 0:03:16.378,0:03:18.561 They came from particle physics, 0:03:18.561,0:03:20.127 they came from classical philology, 0:03:20.127,0:03:22.095 they came from book conservation, 0:03:22.095,0:03:24.677 they came from ancient mathematics, 0:03:24.677,0:03:26.594 they came from data management, 0:03:26.594,0:03:29.511 they came from scientific imaging and program management. 0:03:29.511,0:03:32.344 And they got together to work on this manuscript. 0:03:32.344,0:03:36.878 The first problem was a conservation problem. 0:03:36.878,0:03:39.311 And this is the sort of thing that we had to deal with: 0:03:39.311,0:03:41.511 There was glue on the spine of the book. 0:03:41.511,0:03:43.712 And if you look at this photograph carefully, 0:03:43.712,0:03:45.760 the bottom half of this is rather brown. 0:03:45.760,0:03:47.128 And that glue is hide glue. 0:03:47.128,0:03:48.513 Now if you're a conservator, 0:03:48.513,0:03:50.645 you can take off this glue reasonably easily. 0:03:50.645,0:03:54.010 The top half is Elmer's wood glue. 0:03:54.010,0:03:56.127 It's polyvinyl acetate emulsion 0:03:56.127,0:03:58.794 that doesn't dissolve in water once it's dry. 0:03:58.794,0:04:01.960 And it's much tougher than the parchment that it was written on. 0:04:01.960,0:04:05.311 And so before we could start imaging Archimedes, 0:04:05.311,0:04:06.760 we had to take this book apart. 0:04:06.760,0:04:09.611 So it took four years to take apart. 0:04:09.611,0:04:13.144 And this is a rare action shot, ladies and gentlemen. 0:04:13.144,0:04:15.915 (Laughter) 0:04:15.915,0:04:20.181 Another thing is that we had to get rid of all the wax, 0:04:20.181,0:04:22.416 because this was used in the liturgical services 0:04:22.416,0:04:24.157 of the Greek Orthodox Church 0:04:24.157,0:04:25.468 and they'd used candle wax. 0:04:25.468,0:04:27.331 And the candle wax was dirty, 0:04:27.331,0:04:28.564 and we couldn't image through the wax. 0:04:28.564,0:04:31.899 So very carefully we had to mechanically scrape off all the wax. 0:04:31.899,0:04:33.801 It's hard to tell you exactly 0:04:33.801,0:04:36.515 how bad this condition of this book is, 0:04:36.515,0:04:39.281 but it came out in little bits very often. 0:04:39.281,0:04:41.631 And normally in a book, you wouldn't worry about the little bits, 0:04:41.631,0:04:44.631 but these little bits might contain unique Archimedes text. 0:04:44.631,0:04:46.930 So tiny fragments 0:04:46.930,0:04:50.717 we actually managed to put back in the right place. 0:04:50.717,0:04:55.381 Then, having done that, we started to image the manuscript. 0:04:55.381,0:04:56.881 And we imaged the manuscript 0:04:56.881,0:04:58.647 in 14 different wavebands of light. 0:04:58.647,0:05:02.815 Because if you look at something in different wavebands of light, 0:05:02.815,0:05:04.031 you see different things. 0:05:04.031,0:05:05.598 And here is an image of a page 0:05:05.598,0:05:08.115 imaged in 14 different wavebands of light. 0:05:08.115,0:05:10.248 But none of them worked. 0:05:10.248,0:05:14.665 So what we did was we processed the images together, 0:05:14.665,0:05:17.965 and we put two images into one blank screen. 0:05:17.965,0:05:20.798 And here are two different images of the Archimedes manuscript. 0:05:20.798,0:05:22.614 And the image on the left 0:05:22.614,0:05:23.965 is the normal red image. 0:05:23.965,0:05:25.877 And the image on the right is an ultraviolet image. 0:05:25.877,0:05:27.017 And in the image on the right 0:05:27.017,0:05:29.465 you might be able to see some of the Archimedes writing. 0:05:29.465,0:05:32.447 If you merge them together into one digital canvas, 0:05:32.447,0:05:35.265 the parchment is bright in both images 0:05:35.265,0:05:37.140 and it comes out bright. 0:05:37.140,0:05:40.156 The prayer book is dark in both images 0:05:40.156,0:05:41.853 and it comes out dark. 0:05:41.853,0:05:45.204 The Archimedes text is dark in one image and bright in another. 0:05:45.204,0:05:47.604 And it'll come out dark but red, 0:05:47.604,0:05:49.871 and then you can start to read it rather clearly. 0:05:49.871,0:05:52.021 And that's what it looks like. 0:05:52.021,0:05:56.055 Now that's a before and after image, 0:05:56.055,0:05:59.037 but you don't read the image on the screen like that. 0:05:59.037,0:06:01.656 You zoom in and you zoom in 0:06:01.656,0:06:05.204 and you zoom in and you zoom in, 0:06:05.204,0:06:07.272 and you can just read it now. 0:06:07.272,0:06:13.620 (Applause) 0:06:13.620,0:06:17.200 If you process the same two images in a different way, 0:06:17.200,0:06:19.770 you can actually get rid of the prayer book text. 0:06:19.770,0:06:20.788 And this is terribly important, 0:06:20.788,0:06:23.771 because the diagrams in the manuscript 0:06:23.771,0:06:26.238 are the unique source for the diagrams 0:06:26.238,0:06:29.055 that Archimedes drew in the sand in the fourth century B.C. 0:06:29.055,0:06:31.556 And there we are, I can give them to you. 0:06:31.556,0:06:34.438 With this kind of imaging -- 0:06:34.438,0:06:37.104 this kind of infrared, ultraviolet, invisible light imaging -- 0:06:37.104,0:06:39.704 we were never going to image through the gold ground forgeries. 0:06:39.704,0:06:41.557 How were we going to do that? 0:06:41.557,0:06:43.520 Well we took the manuscript, 0:06:43.520,0:06:47.807 and we decided to image it in X-ray fluorescence imaging. 0:06:47.807,0:06:50.821 So an X-ray comes in in the diagram on the left 0:06:50.821,0:06:54.673 and it knocks out an electron from the inner shell of an atom. 0:06:54.673,0:06:56.671 And that electron disappears. 0:06:56.671,0:07:00.306 And as it disappears, an electron from a shell farther out 0:07:00.306,0:07:02.788 jumps in and takes its place. 0:07:02.788,0:07:04.221 And when it takes its place, 0:07:04.221,0:07:07.055 it sheds electromagnetic radiation. 0:07:07.055,0:07:08.487 It sheds an X-ray. 0:07:08.487,0:07:11.138 And this X-ray is specific in its wavelength 0:07:11.138,0:07:13.020 to the atom that it hits. 0:07:13.020,0:07:15.188 And what we wanted to get 0:07:15.188,0:07:16.921 was the iron. 0:07:16.921,0:07:18.788 Because the ink was written in iron. 0:07:18.788,0:07:20.471 And if we can map 0:07:20.471,0:07:23.055 where this X-ray that comes out, where it comes from, 0:07:23.055,0:07:25.121 we can map all the iron on the page, 0:07:25.121,0:07:27.720 and theoretically we can read the image. 0:07:27.720,0:07:32.055 The thing is that you need a very powerful light source to do this. 0:07:32.055,0:07:34.737 So we took it to the Stanford Synchrotron Radiation Laboratory 0:07:34.737,0:07:36.871 in California, 0:07:36.871,0:07:38.439 which is a particle accelerator. 0:07:38.439,0:07:40.171 Electrons go around one way, 0:07:40.171,0:07:41.788 positrons go around the other. 0:07:41.788,0:07:43.104 They meet in the middle, 0:07:43.104,0:07:45.540 and they create subatomic particles 0:07:45.540,0:07:47.770 like the charm quark and the tau lepton. 0:07:47.770,0:07:50.792 Now we weren't actually going to put Archimedes in that beam. 0:07:50.792,0:07:54.340 But as the electrons go round at the speed of light, 0:07:54.340,0:07:55.607 they shed X-rays. 0:07:55.607,0:07:58.257 And this is the most powerful light source in the solar system. 0:07:58.257,0:08:00.206 This is called synchrotron radiation, 0:08:00.206,0:08:02.024 and it's normally used to look at things 0:08:02.024,0:08:03.691 like proteins and that sort of thing. 0:08:03.691,0:08:07.688 But we wanted it to look at atoms, at iron atoms, 0:08:07.688,0:08:10.557 so that we could read the page from before and after. 0:08:10.557,0:08:12.855 And lo and behold, we found that we could do it. 0:08:12.855,0:08:15.626 It took about 17 minutes to do a single page. 0:08:15.626,0:08:18.958 So what did we discover? 0:08:18.958,0:08:21.140 Well one of the unique texts in Archimedes 0:08:21.140,0:08:22.943 is called "The Stomachion." 0:08:22.943,0:08:25.357 And this didn't exist in Codices A and B. 0:08:25.357,0:08:27.776 And we knew that it involved this square. 0:08:27.776,0:08:29.569 And this is a perfect square, 0:08:29.569,0:08:31.590 and it's divided into 14 bits. 0:08:31.590,0:08:34.373 But no one knew what Archimedes was doing with these 14 bits. 0:08:34.373,0:08:36.557 And now we think we know. 0:08:36.557,0:08:38.324 He was trying to work out 0:08:38.324,0:08:40.857 how many ways you can recombine those 14 bits 0:08:40.857,0:08:43.184 and still make a perfect square. 0:08:43.184,0:08:46.640 Anyone want to guess the answer? 0:08:46.671,0:08:52.048 It's 17,152 divided into 536 families. 0:08:52.048,0:08:54.415 And the important point about this 0:08:54.415,0:08:58.129 is that it's the earliest study in combinatorics in mathematics. 0:08:58.129,0:09:01.893 And combinatorics is a wonderful and interesting branch of mathematics. 0:09:01.893,0:09:05.128 The really astonishing thing though about this manuscript 0:09:05.128,0:09:07.497 is that we looked at the other manuscripts 0:09:07.497,0:09:08.915 that the palimpsester had made, 0:09:08.915,0:09:11.481 the scribe had made his book out of, 0:09:11.481,0:09:14.413 and one of them was a manuscript containing text by Hyperides. 0:09:14.413,0:09:18.781 Now Hyperides was an Athenian orator from the fourth century B.C. 0:09:18.781,0:09:21.198 He was an exact contemporary of Demosthenes. 0:09:21.198,0:09:25.597 And in 338 B.C. he and Demosthenes together 0:09:25.597,0:09:27.281 decided that they wanted to stand up 0:09:27.281,0:09:29.281 to the military might of Philip of Macedon. 0:09:29.281,0:09:32.930 So Athens and Thebes went out to fight Philip of Macedon. 0:09:32.930,0:09:34.198 This was a bad idea, 0:09:34.198,0:09:37.631 because Philip of Macedon had a son called Alexander the Great, 0:09:37.631,0:09:39.914 and they lost the battle of Chaeronea. 0:09:39.914,0:09:42.632 Alexander the Great went on to conquer the known world; 0:09:42.632,0:09:45.382 Hyperides found himself on trial for treason. 0:09:45.382,0:09:48.982 And this is the speech that he gave when he was on trial -- 0:09:48.982,0:09:50.130 and it's a great speech: 0:09:50.130,0:09:52.398 "Best of all," he says, "is to win. 0:09:52.398,0:09:54.132 But if you can't win, 0:09:54.132,0:09:56.173 then you should fight for a noble cause, 0:09:56.173,0:09:57.748 because then you'll be remembered." 0:09:57.748,0:09:59.346 Consider the Spartans. 0:09:59.346,0:10:01.214 They won enumerable victories, 0:10:01.214,0:10:03.046 but no one remembers what they are 0:10:03.046,0:10:05.181 because they were all fought for selfish ends. 0:10:05.181,0:10:08.747 The one battle that the Spartans fought that everybody remembers 0:10:08.747,0:10:10.463 is the the battle of Thermopylae 0:10:10.463,0:10:12.466 where they were butchered to a man, 0:10:12.466,0:10:14.398 but fought for the freedom of Greece. 0:10:14.398,0:10:16.663 It was such a great speech 0:10:16.663,0:10:20.398 that the Athenian law courts let him off. 0:10:20.398,0:10:22.464 He lived for another 10 years, 0:10:22.464,0:10:25.115 then the Macedonian faction caught up with him. 0:10:25.115,0:10:27.699 They cut out his tongue in mockery of his oratory, 0:10:27.699,0:10:31.130 and no one knows what they did with his body. 0:10:31.130,0:10:34.281 So this is the discovery of a lost voice from antiquity, 0:10:34.281,0:10:36.263 speaking to us, not from the grave, 0:10:36.263,0:10:37.741 because his grave doesn't exist, 0:10:37.741,0:10:39.514 but from the Athenian law courts. 0:10:39.514,0:10:41.163 Now I should say at this point 0:10:41.163,0:10:43.930 that normally when you're looking 0:10:43.930,0:10:45.926 at medieval manuscripts that have been scraped off, 0:10:45.926,0:10:47.412 you don't find unique texts. 0:10:47.412,0:10:50.514 And to find two in one manuscript is really something. 0:10:50.514,0:10:53.731 To find three is completely weird. 0:10:53.731,0:10:55.412 And we found three. 0:10:55.412,0:10:56.678 Aristotle's "Categories" 0:10:56.678,0:10:59.131 is one of the foundational texts of Western philosophy. 0:10:59.131,0:11:03.547 And we found a third century A.D. commentary on it, 0:11:03.547,0:11:06.731 possibly by Galen and probably by Porphyry. 0:11:06.731,0:11:08.929 Now all this data that we collected, 0:11:08.929,0:11:11.464 all the images, all the raw images, 0:11:11.464,0:11:13.814 all the transcriptions that we made and that sort of thing 0:11:13.814,0:11:17.300 have been put online under a Creative Commons license 0:11:17.300,0:11:20.214 for anyone to use for any commercial purpose. 0:11:20.214,0:11:26.747 (Applause) 0:11:26.747,0:11:29.513 Why did the owner of the manuscript do this? 0:11:29.513,0:11:33.797 He did this because he understands data as well as books. 0:11:33.797,0:11:35.132 Now the thing to do with books, 0:11:35.132,0:11:37.263 if you want to ensure their long-term utility, 0:11:37.263,0:11:39.479 is to hide them away in closets 0:11:39.479,0:11:41.279 and let very few people look at them. 0:11:41.279,0:11:43.698 The thing to do with data, if you want it to survive, 0:11:43.698,0:11:46.999 is to let it out and have everybody have it 0:11:46.999,0:11:50.098 with as little control on that data as possible. 0:11:50.098,0:11:51.262 And that's what he did. 0:11:51.262,0:11:54.690 And institutions can learn from this. 0:11:54.690,0:11:57.242 Because institutions at the moment 0:11:57.242,0:12:00.299 confine their data with copyright restrictions and that sort of thing. 0:12:00.299,0:12:02.491 And if you want to look at medieval manuscripts on the Web, 0:12:02.491,0:12:05.962 at the moment you have to go to the National Library of Y's site 0:12:05.962,0:12:08.630 or the University Library of X's site, 0:12:08.630,0:12:10.529 which is about the most boring way 0:12:10.529,0:12:12.306 in which you can deal with digital data. 0:12:12.306,0:12:15.299 What you want to do is to aggregate it all together. 0:12:15.299,0:12:18.396 Because the Web of the ancient manuscripts of the future 0:12:18.396,0:12:20.682 isn't going to be built by institutions. 0:12:20.682,0:12:23.790 It's going to be built by users, 0:12:23.790,0:12:25.564 by people who get this data together, 0:12:25.564,0:12:29.099 by people who want to aggregate all sorts of maps 0:12:29.099,0:12:30.731 from wherever they come from, 0:12:30.731,0:12:32.531 all sorts of medieval romances 0:12:32.531,0:12:34.080 from wherever they come from, 0:12:34.080,0:12:37.863 people who just want to curate their own glorious selection 0:12:37.863,0:12:39.380 of beautiful things. 0:12:39.380,0:12:40.997 And that is the future of the Web. 0:12:40.997,0:12:44.180 And it's an attractive and beautiful future, 0:12:44.180,0:12:45.782 if only we can make it happen. 0:12:45.782,0:12:48.963 Now we at the Walters Art Museum have followed this example, 0:12:48.963,0:12:52.313 and we have put up all our manuscripts on the Web 0:12:52.313,0:12:53.996 for people to enjoy -- 0:12:53.996,0:12:56.580 all the raw data, all the descriptions, all the metadata. 0:12:56.580,0:12:59.463 under a Creative Commons license. 0:12:59.463,0:13:00.981 Now the Walters Art Museum is a small museum 0:13:00.981,0:13:02.529 and it has beautiful manuscripts, 0:13:02.529,0:13:04.547 but the data is fantastic. 0:13:04.547,0:13:06.330 And the result of this 0:13:06.330,0:13:09.348 is that if you do a Google search on images right now 0:13:09.348,0:13:12.680 and you type in "Illuminated manuscript Koran" for example, 0:13:12.680,0:13:17.097 24 of the 28 images you'll find come from my institution. 0:13:17.097,0:13:22.647 (Applause) 0:13:22.647,0:13:27.680 Now, let's think about this for a minute. 0:13:27.680,0:13:29.663 What's in it for the institution? 0:13:29.663,0:13:32.259 There are all sorts of things that are in it for the institution. 0:13:32.259,0:13:34.382 You can talk about the Humanities and that sort of thing, 0:13:34.382,0:13:36.263 but let's talk about selfish things. 0:13:36.263,0:13:40.131 Because what's really in it for the institution is this: 0:13:40.131,0:13:42.664 Now why do people go to the Louvre? 0:13:42.664,0:13:45.863 They go to see the Mona Lisa. 0:13:45.863,0:13:49.230 Why do they go to see the Mona Lisa? 0:13:49.230,0:13:51.612 Because they already know what she looks like. 0:13:51.612,0:13:53.946 And they know what she looks like 0:13:53.946,0:13:58.648 because they've seen pictures of her absolutely everywhere. 0:13:58.648,0:14:03.097 Now, there is no need 0:14:03.097,0:14:05.376 for these restrictions at all. 0:14:05.376,0:14:07.463 And I think that institutions should stand up 0:14:07.463,0:14:10.597 and release all their data under unrestricted licenses, 0:14:10.597,0:14:12.813 and it would be a great benefit to everybody. 0:14:12.813,0:14:15.979 Why don't we just let everybody have access to this data 0:14:15.979,0:14:18.315 and curate their own collection 0:14:18.315,0:14:20.480 of ancient knowledge and wonderful and beautiful things 0:14:20.480,0:14:23.568 and increase the beauty and the cultural significance 0:14:23.568,0:14:24.796 of the Internet. 0:14:24.796,0:14:26.313 Thank you very much indeed. 0:14:26.313,0:14:30.746 (Applause)