Smithsonian Magazine - - The Dispossessed -

ar­ranged for a 600-pound high-res­o­lu­tion CT scan­ner to be sent by truck from Bel­gium, and he made in­tri­cately de­tailed scans of the scrolls. But af­ter months of an­a­lyz­ing the data, Seales was dis­heart­ened to find that the ink in­side the scrolls, de­spite the traces of lead, was in­vis­i­ble.

What was worse, the scans showed the lay­ers in­side the scrolls to be so car­bonized that in many places there was no de­tectable sep­a­ra­tion be­tween them. “It was just too com­pli­cated for our al­go­rithms,” Seales ad­mits. He played me a video of the CT scan data, show­ing one of the scrolls in cross-section. The whorls of pa­pyrus glowed white against a dark back­ground, like closely wound strands of silk. “Just take a look at that,” said Seales. “This is when we knew we were doomed for the present time.”

What makes vir­tual un­wrap­ping such a com­plex chal­lenge is that, even if you im­aged the in­side of a rolled-up scroll writ­ten in ink that glowed brightly in scans, you would still only see a dizzy­ing mess of tightly packed let­ters float­ing in space, like a three-di­men­sional jig­saw puz­zle—but with­out a fi­nal pic­ture to use as a guide. To de­ci­pher that jum­ble of let­ters, Seales’ key in­no­va­tion was to de­velop soft­ware to lo­cate and model the sur­face layer within a wound-up scroll, which an­a­lyzes each point in as many as 12,000 cross-sec­tions. Then he looks for den­sity changes that cor­re­spond to the ink, and ap­plies fil­ters or other tech­niques to in­crease the con­trast of the let­ters as much as pos­si­ble. The fi­nal step is to fig­u­ra­tively “un­roll” the im­age for read­ing.

Seales spent 2012 and 2013 as a vis­it­ing sci­en­tist at the Google Cul­tural In­sti­tute in Paris, amp­ing up his al­go­rithms to cope with the com­plex struc­tures the CT scans had re­vealed. He got the chance to try his new ap­proach soon af­ter­ward, when Pn­ina Shor, at the Israel An­tiq­ui­ties Au­thor­ity, or IAA, in Jerusalem, con­tacted him about a car­bonized roll of parch­ment found in the an­cient town of Ein Gedi, on the west­ern shore of the Dead Sea. The scroll was ex­ca­vated from the re­mains of a syn­a­gogue, which was de­stroyed by fire in the sixth cen­tury A.D. The charred, cigar-shaped lump was far too frag­ile to open, but Is­raeli re­searchers had re­cently CT-scanned it. Would Seales take a look at the data? Shor handed over a hard drive, and Seales and his col­leagues went to work.

In the mean­time, Seales was chas­ing a new idea for read­ing car­bon-based ink: X-ray phase-con­trast to­mog­ra­phy, a highly sensi- tive form of imag­ing that can de­tect sub­tle den­sity changes in a ma­te­rial—the kind that might re­sult from ap­ply­ing ink to pa­pyrus—by mea­sur­ing the chang­ing in­ten­sity of the beam as it passes through an ob­ject. Only a large par­ti­cle ac­cel­er­a­tor, though, can pro­duce such a beam. One of the near­est was in Greno­ble, in south­east­ern France. Seales’ ini­tial re­quests for “beam time” were re­jected, but he was sub­se­quently ap­proached by an Ital­ian physi­cist named Vito Mo­cella, who had close ties to the fa­cil­ity, and in De­cem­ber 2013 De­lat­tre took Ba­nana Boy and an­other scroll to Greno­ble.

Seales waited ea­gerly for the promised data, but the files did not ar­rive. Then, in Jan­uary 2015, Mo­cella’s group pub­lished the re­sults with­out him. It was, Seales says, an “ex­cru­ci­at­ingly frus­trat­ing” ex­pe­ri­ence. “I be­lieved we were col­lab­o­rat­ing, un­til I re­al­ized that the feel­ing was not mu­tual.”

News sto­ries around the world re­ported that Her­cu­la­neum scrolls had been de­ci­phered at last. But, in fact, Mo­cella had claimed to read only let­ters, and some schol­ars are cau­tious about even those, not least be­cause the group did not pub­lish enough in­for­ma­tion for oth­ers to repli­cate the anal­y­sis. Mo-

It was an “ex­cru­ci­at­ingly frus­trat­ing” ex­pe­ri­ence. “I be­lieved we were col­lab­o­rat­ing, un­til I re­al­ized that the feel­ing was not mu­tual.”

cella fi­nally shared his data with Seales and oth­ers af­ter pub­li­ca­tion. Af­ter re­view­ing it, Seales con­cluded that the find­ings were a bust. “The dataset did not pro­duce any con­trast at the ink,” he told me. Seales thinks the re­searchers, who were with­out soft­ware to model the sur­faces within the scrolls, were see­ing “ghosts”—ran­dom pat­terns in the pa­pyrus’ fiber struc­ture that just hap­pen to look like let­ters. He is now con­vinced that phase-con­trast to­mog­ra­phy alone is not suf­fi­cient to read the Her­cu­la­neum scrolls in any mean­ing­ful way. (Mo­cella in­sists the let­ters he saw were real, and he took is­sue with Seales’ ver­sion of the in­ci­dent. “From my point of view, I and my team are still work­ing with Brent, since we’ve given him, as with other spe­cial­ists like him, most of the scans,” Mo­cella said.)

By that point Seales had fin­ished a pre­lim­i­nary anal­y­sis of the Ein Gedi scroll, and in July 2015 he and the IAA an­nounced their re­sults. “We ab­so­lutely hit a home run,” Seales says.

Un­like the au­thors of the Her­cu­la­neum scrolls, the He­brew scribes had mixed me­tals into their ink. Seales’ soft­ware cor­rectly mapped the let­ters to the rolled-up parch­ment, then vir­tu­ally un­furled it, re­veal­ing all of the sur­viv­ing text, in per­fect se­quence, on each of the five wraps of the scroll. There were 35 lines of text in two col­umns, com­posed of He­brew let­ters just two mil- lime­ters tall. Is­raeli re­searchers iden­ti­fied the text as the first two chap­ters of the Book of Leviti­cus, dat­ing to the third or fourth cen­tury A.D. It was a hugely sig­nif­i­cant find for bib­li­cal schol­ars: the old­est ex­tant copy of the He­brew Bi­ble out­side of the Dead Sea Scrolls, and a glimpse into the his­tory of the Bi­ble dur­ing a pe­riod from which hardly any texts sur­vive.

And it was proof that Seales’ method worked. Fol­low­ing Mo­cella’s pub­li­ca­tion, how­ever, the In­sti­tut de France re­fused fur­ther ac­cess to its Her­cu­la­neum scrolls. Which is why Seales turned his at­ten­tion to Ox­ford.

THE BODLEIAN LI­BRARIES, at Ox­ford Univer­sity, pos­sess four Her­cu­la­neum scrolls, which ar­rived in 1810, af­ter they were pre­sented to the Prince of Wales. They are kept deep in­side the build­ing, in a lo­ca­tion so se­cret that even David How­ell, the Bodleian’s head of her­itage science, says he doesn’t know where it is.

Seales wasn’t per­mit­ted to see the in­tact pa­pyri, never mind scan them. But one of the four, known as “P.Herc. 118,” was sent to Naples in 1883, to be un­rolled us­ing Pi­ag­gio’s ma­chine. It came back as a mo­saic of crumbs, which were glued onto tis­sue pa­per and mounted be­hind glass in 12 wood frames. The text ap­pears to be a his­tory of Epi­curean phi­los­o­phy, prob­a­bly by Philode­mus, but it has been par­tic­u­larly chal­leng­ing for schol­ars to in­ter­pret. A frag­ment might seem cov­ered with con­tin­u­ous lines of writ­ing, says Ob­bink, “but re­ally ev­ery inch you’re jump­ing up or down a layer.”

To prove the value of his ap­proach, Seales asked the Bodleian to let him an­a­lyze P.Herc. 118. If all went well, he hoped, he might get a shot at scan­ning the in­tact scrolls later. “We wouldn’t nec­es­sar­ily have cho­sen to get in­volved, ex­cept for Brent’s en­thu­si­asm,” says How­ell. So in July 2017, the 12 frames were re­moved from stor­age and taken to How­ell’s third-floor of­fice—some­thing of a coup for Seales, given their in­valu­able na­ture. Cheer­ful and ruddy-faced, How­ell has worked in con­ser­va­tion for close to 35 years, and even he felt daunted as the pro­tec­tive glass frames were re­moved, ex­pos­ing the frag­ile pa­pyrus be­neath. “These are the most ter­ri­fy­ing ob­jects I’ve ever han­dled,” he says. “If you sneeze, they’d blow away.”

Seales and an­other col­league scanned these scroll frag­ments us­ing a hand-held 3-D scan­ner called an Artec Space Spi­der. Mean­while, How­ell car­ried out hy­per­spec­tral imag­ing, which uses hun­dreds of wave­lengths of light. How­ell lis­tened to Pink Floyd through noise-can­cel­ing head­phones to es­cape the grind­ing noise of the scan­ner, he says, plus the knowl­edge that if any­thing went wrong, “I might as well pack my bags and go home and not come back.”

Af­ter Seales re­turned to Ken­tucky, he and his col­leagues spent months map­ping all of the avail­able 2-D im­ages onto the 3-D tem­plate pro­duced by the Artec Space Spi­der. This past March, they re­turned to Ox­ford to present the re­sults on a big screen to a packed con­fer­ence room. At such a high res­o­lu­tion, the charred pa­pyrus re­sem­bled a dark-brown moun­tain range as seen from above, with lines of text snaking over the ridges and peaks. There was a gasp from the au­di­ence as Seales’ stu­dent Han­nah Hatch ro­tated the im­age, then zoomed into creases and peeked over folds, flip­ping seam­lessly be­tween high-res­o­lu­tion pho­to­graphs, in­frared im­ages and even the dis­egni draw­ings—all matched up to the 3-D tem­plate.

Shortly af­ter­ward, James Brusue­las, an Ox­ford pa­py­rol­o­gist work­ing with Seales, re­vealed sev­eral new de­tails vis­i­ble in the scans, such as the name Pytho­cles, who was a young fol­lower of Epi­cu­rus. More im­por­tant, Brusue­las was able to de­ci­pher the col­umn struc­ture of the text—17 char­ac­ters per line—which will be cru­cial for read­ing the rest of the roll, par­tic­u­larly when try­ing to join dif­fer­ent frag­ments to­gether. “We have the ba­sic in­for­ma­tion we need to put Humpty Dumpty back to­gether again,” he said.

The au­di­ence buzzed with ques­tions and ap­plause. It was the re­ac­tion Seales was hop­ing for, and a step to­ward his real goal—gain­ing ac­cess to in­tact scrolls.

He’d saved his own pre­sen­ta­tion un­til last. It wasn’t about P.Herc. 118, but rather one tiny let­ter: the lu­nate sigma.

DRIV­ING SOUTH from the stone arch­ways and quad­ran­gles of Ox­ford, the road soon cuts through flat green fields reach­ing to the hori­zon. On the day I vis­ited, fork-tailed red kites hov­ered high in the blue July sky. Af­ter 15 or so miles a sprawl­ing cam­pus of low gray build­ings came into view. At first, it re­sem­bled an or­di­nary in­dus­trial park, un­til I no­ticed the names of the roads: Fermi, Ruther­ford, Bec­querel, all gi­ants of 19th- and 20th-cen­tury physics. Be­hind a wire fence a huge, sil­ver dome, more than a quar­ter-mile in cir­cum­fer­ence, rose from the grass like a gi­ant fly­ing saucer. This was Di­a­mond Light Source, and Seales was wait­ing in­side.

He’d brought a speck of charred papy-

rus from one of the Her­cu­la­neum scrolls he stud­ied a decade ear­lier. The ink on it, he had found, con­tained a trace of lead. In Greno­ble, di­rect X-ray imag­ing of the scrolls had not been enough to de­tect the ink. But when you fire hugely pow­er­ful X-rays through lead, the metal emits elec­tro­mag­netic ra­di­a­tion, or “flu­o­resces,” at a char­ac­ter­is­tic fre­quency. Seales hoped to pick up that sig­nal with a de­tec­tor placed be­side the frag­ment, which was spe­cially cal­i­brated to cap­ture pho­tons at lead’s char­ac­ter­is­tic fre­quency.

It was a long shot. The mi­nus­cule flu­o­res­cence of the let­ter would be swamped by ra­di­a­tion from the pro­tec­tive lead lin­ing the room—like look­ing for a flick­er­ing can­dle from miles away on a rainy night, Seales said, as we stood in the crowded hutch. But af­ter sev­eral days of in­tense work—op­ti­miz­ing the an­gle of the de­tec­tor, shield­ing the main X-ray beam with tung­sten “flight tubes”—the team fi­nally got what it was look­ing for: a grainy, but clearly rec­og­niz­able, “c.”

“We’ve proven it,” Seales said in tri­umph as he dis­played the leg­i­ble im­age to the Ox­ford au­di­ence in March. It is, Seales hopes, the last piece of the puz­zle he needs to read the ink in­side a Her­cu­la­neum scroll.

The re­sults have schol­ars ex­cit­edly re-eval­u­at­ing what they might now be able to achieve. “I think it’s ac­tu­ally very close to be­ing cracked,” says Ob­bink, the Ox­ford pa­py­rol­o­gist. He es­ti­mates that at least 500 Her­cu­la­neum scrolls haven’t been opened. More­over, ex­ca­va­tions at Her­cu­la­neum in the 1990s re­vealed two un­ex­plored lay­ers of the villa, which some schol­ars be­lieve may con­tain hun­dreds or even thou­sands more scrolls.

Many schol­ars are con­vinced that Piso’s great library must have con­tained a range of lit­er­a­ture far wider than what has been doc­u­mented so far. Ob­bink says he wouldn’t be sur­prised to find more Latin lit­er­a­ture, or a once-unimag­in­able trea­sure of lost po­ems by Sap­pho, the revered sev­enth-cen­tury B.C. poet known to­day only through the briefest of frag­ments.

Michael Phelps, of the Early Manuscripts Elec­tronic Library, in Cal­i­for­nia, who re­cently used mul­ti­spec­tral imag­ing to re­veal dozens of hid­den texts on re-

used parch­ment at St. Cather­ine’s Monastery, in Egypt, calls Seales’ meth­ods “revo­lu­tion­ary.” Schol­ars have long faced a choice be­tween at­tempt­ing to read con­cealed texts (and po­ten­tially de­stroy­ing them in the process) or con­serv­ing them un­read. “Brent Seales’ tech­nol­ogy is re­mov­ing that dilemma,” Phelps says.

Suc­cess­fully read­ing Her­cu­la­neum scrolls could trig­ger a new “re­nais­sance of clas­si­cal an­tiq­uity,” says Gregory Hey­worth, a me­dieval­ist at the Univer­sity of Rochester in New York. He points out that vir­tual un­wrap­ping could be ap­plied to count­less other texts. In West­ern Europe alone, he es­ti­mates, there are tens of thou­sands of manuscripts dat­ing from be­fore A.D. 1500—from car­bonized scrolls to book cov­ers made from older, glued-to­gether pages—that could ben­e­fit from such imag­ing.

“We’d change the canon,” Hey­worth says. “I think the next gen­er­a­tion is go­ing to have a very dif­fer­ent pic­ture of an­tiq­uity.”

SEALES HAS L ATELY BEEN en­hanc­ing his tech­nique, by us­ing ar­ti­fi­cial in­tel­li­gence to train his soft­ware to rec­og­nize sub­tle dif­fer­ences in tex­ture be­tween pa­pyrus and ink. He plans to com­bine such ma­chine learn­ing and X-ray flu­o­res­cence to pro­duce the clear­est pos­si­ble text. In the fu­ture, “it’ll all be au­to­mated,” he pre­dicts. “Put it in the scan­ner and it will all just un­furl.”

Seales is still ne­go­ti­at­ing with cu­ra­tors in Ox­ford, Naples and Paris for ac­cess to in­tact scrolls. He has sur­mounted huge tech­ni­cal hur­dles, but the com­plex po­lit­i­cal chal­lenge of nav­i­gat­ing the gate­keep­ers, win­ning beam time at par­ti­cle ac­cel­er­a­tors and lin­ing up fund­ing can, very oc­ca­sion­ally, punc­ture his op­ti­mism. “How does a guy like me make all that stuff hap­pen all at once?” he said in one such mo­ment. He shrugged and looked around him. “It’s more than a com­puter sci­en­tist is re­ally ca­pa­ble of do­ing.”

Then be­lief re­turned to his wide, hazel eyes. “I refuse to ac­cept that it’s not pos­si­ble,” he said. “At ev­ery turn, there has been some­thing that opened up.” Read­ing a com­plete in­tact scroll at last, he went on, would be “like re­turn­ing home to your fam­ily, who have been wait­ing all along for you to do the thing you started.”

In the 18th cen­tury, scrolls were un­wrapped at the rate of a cen­time­ter an hour, us­ing a ma­chine de­signed by Vat­i­can con­ser­va­tor An­to­nio Pi­ag­gio.

Among the many thou­sands killed by Ve­su­vius’ erup­tion was Pliny the El­der, the an­cient world’s great­est nat­u­ral­ist, whose death is de­picted in an 1813 paint­ing by Pierre Henri de Va­len­ci­ennes.

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.