- Spitfire List - https://spitfirelist.com -

FTR #1075 Surveillance Valley, Part 1: Eugenics, Racism and High Tech

Dave Emory’s entire life­time of work is avail­able on a flash dri­ve that can be obtained HERE [1]. The new dri­ve is a 32-giga­byte dri­ve that is cur­rent as of the pro­grams and arti­cles post­ed by the fall of 2017. The new dri­ve (avail­able for a tax-deductible con­tri­bu­tion of $65.00 or more.)

WFMU-FM is pod­cast­ing For The Record–You can sub­scribe to the pod­cast HERE [2].

You can sub­scribe to e‑mail alerts from Spitfirelist.com HERE [3].

You can sub­scribe to RSS feed from Spitfirelist.com HERE [3].

You can sub­scribe to the com­ments made on pro­grams and posts–an excel­lent source of infor­ma­tion in, and of, itself, HERE [4].

Please con­sid­er sup­port­ing THE WORK DAVE EMORY DOES [5].

This broad­cast was record­ed in one, 60-minute seg­ment [6].

Intro­duc­tion: Begin­ning a crit­i­cal­ly impor­tant series on a vital­ly impor­tant book titled Sur­veil­lance Val­ley: The Secret Mil­i­tary His­to­ry of the Inter­net, this pro­gram explores the gen­e­sis of high tech and data pro­cess­ing, an ori­gin that is inex­tri­ca­bly linked with eugen­ics, anti-immi­grant doc­trine [7] and–as is char­ac­ter­is­tic of fas­cism [8], the fear of the ubiq­ui­tous, malev­o­lent “oth­er.”

High­lights of dis­cus­sion and analy­sis include:

  1. The gen­e­sis of high tech was Her­man Hol­lerith’s tab­u­lat­ing machine. ” . . . . A few years ear­li­er, work­ing for the U.S. Cen­sus Bureau, Hol­lerith had devel­oped the world’s first func­tion­al mass-pro­duced com­put­er: the Hol­lerith tab­u­la­tor. An electro­mechan­i­cal device about the size of large desk and dress­er, it used punch cards and a clever arrange­ment of gears, sorters, elec­tri­cal con­tacts, and dials to process data with blaz­ing speed and accu­ra­cy. What had tak­en years by hand could be done in a mat­ter of months. As one U.S. news­pa­per described it, ‘with [the device’s] aid some 15 young ladies can count accu­rate­ly half a mil­lion of names in a day.’ . . .’
  2. Hol­lerith’s machine found its (arguably) great­est appli­ca­tion with the com­pi­la­tion of the cen­sus and the appli­ca­tion of the pseu­do-sci­ence of eugen­ics to it: ” . . . . Grasp­ing about for solu­tions, many set­tled on var­i­ous strains of race sci­ence quack­ery [9]. So-called social Dar­win­ists relied on a twist­ed ver­sion of the the­o­ry of evo­lu­tion to explain why the poor and mar­gin­al­ized should remain that way while the wealthy and suc­cess­ful deserved to rule unchal­lenged. Tak­ing this notion a step fur­ther, adher­ents of eugen­ics fer­vent­ly believed that nat­u­ral­ly supe­ri­or Anglo-Amer­i­cans were on the verge of being wiped out due to the high birth rates of ‘degen­er­ate’ and immi­grant stock. To head off this threat, they advo­cat­ed strict con­trols on repro­duc­tion — breed­ing humans for qual­i­ty in the same way that farm­ers did cows and hors­es. . . .”
  3. Hol­lerith’s machine was seen as the per­fect vehi­cle for real­iz­ing eugenic prac­tice through refin­ing the cen­sus: ” . . . . The cen­sus had been a racial instru­ment from its incep­tion, begin­ning with the orig­i­nal con­sti­tu­tion­al clause that instruct­ed cen­sus offi­cials to count black slaves sep­a­rate­ly from whites and to assign them a val­ue of only three-fifths of a per­son. With each decade, new ‘racial’ cat­e­gories [10] were invent­ed and added to the mix: ‘free col­ored males and females’ and ‘mulat­to’ were count­ed, includ­ing sub­di­vi­sions like includ­ing ‘quadroon’ and ‘octoroon.’ Cat­e­gories for Chi­nese, ‘Hin­doo,’ and Japan­ese were added, as were ‘for­eign’ and ‘native born’ des­ig­na­tions for whites. The cen­sus slow­ly expand­ed to col­lect oth­er demo­graph­ic data, includ­ing lit­er­a­cy lev­els, unem­ploy­ment sta­tis­tics, and med­ical ail­ments, such as those who were ‘deaf, dumb, and blind’ and the ‘insane and idi­ot­ic.’ All of it was bro­ken down by race. . . .The cen­sus need­ed to improve dras­ti­cal­ly. What it need­ed was a tal­ent­ed inven­tor, some­one young and ambi­tious who would be able to come up with a method to auto­mate tab­u­la­tion and data analy­sis. Some­one like Her­man Hol­lerith. . . .”
  4. [11]

    Her­man Hol­lerith

    Hol­lerith’s technology–when applied to the cen­sus, antic­i­pat­ed the mass sur­veil­lance tech­nol­o­gy of the inter­net: ” . . . . Overnight, Hollerith’s tab­u­la­tor tech­nol­o­gy had trans­formed cen­sus tak­ing from a sim­ple head count into some­thing that looked very much like a crude form of mass sur­veil­lance. To the race-obsessed polit­i­cal class, it was a rev­o­lu­tion­ary devel­op­ment. They could final­ly put the nation’s eth­nic make­up under the micro­scope. The data seemed to con­firm the nativists’ worst fears: Poor, illit­er­ate immi­grants were swarm­ing America’s cities, breed­ing like rab­bits, and out­strip­ping native Anglo-Amer­i­can birth rates. Imme­di­ate­ly fol­low­ing the cen­sus, the states and the fed­er­al gov­ern­ment passed a flur­ry of laws [12] that heav­i­ly restrict­ed immi­gra­tion. . . .”

  5. As dis­cussed in FTR #279 [13], IBM’s Hol­lerith machines (acquired when Thomas J. Wat­son bought out Hol­lerith) were fun­da­men­tal to the oper­a­tions of the Third Reich: ” . . . . ‘Indeed, the Third Reich would open star­tling sta­tis­ti­cal venues for Hol­lerith machines nev­er before insti­tut­ed — per­haps nev­er before even imag­ined,’ wrote Edwin Black in IBM and the Holo­caust [14], his pio­neer­ing 2001 exposé of the for­got­ten busi­ness ties between IBM and Nazi Ger­many. ‘In Hitler’s Ger­many, the sta­tis­ti­cal and cen­sus com­mu­ni­ty, over­run with doc­tri­naire Nazis, pub­licly boast­ed about the new demo­graph­ic break­throughs their equip­ment would achieve.’ . . . Demand for Hol­lerith tab­u­la­tors was so robust that IBM was forced to open a new fac­to­ry in Berlin to crank out all the new machines. At the facility’s chris­ten­ing cer­e­mo­ny, which was attend­ed by a top U.S. IBM exec­u­tive and the elite of the Nazi Par­ty, the head of IBM’s Ger­man sub­sidiary gave a rous­ing speech about the impor­tant role that Hol­lerith tab­u­la­tors played in Hitler’s dri­ve to puri­fy Ger­many and cleanse it of infe­ri­or racial stock. . . .”
  6. The Trump admin­is­tra­tion’s fram­ing of ques­tions for the 2020 cen­sus appear aimed at cre­at­ing a “nation­al registry”–a con­cept rem­i­nis­cent of the Third Reich’s use of IBM’s Hol­lerith-col­lect­ed data: ” . . . . Based on a close read­ing of inter­nal Depart­ment of Com­merce doc­u­ments tied to the cen­sus cit­i­zen ques­tion pro­pos­al, it appears [15] the Trump admin­is­tra­tion wants to use the cen­sus to con­struct a first-of-its-kind cit­i­zen­ship reg­istry for the entire U.S. pop­u­la­tion — a deci­sion that arguably exceeds the legal author­i­ty of the cen­sus. ‘It was deep in the doc­u­men­ta­tion that was released,’ Robert Groves, a for­mer Cen­sus Bureau direc­tor who head­ed the Nation­al Acad­e­mies com­mit­tee con­vened to inves­ti­gate the 2020 cen­sus, told me by tele­phone. ‘No one picked up on it much. But the term ‘reg­istry’ in our world means not a col­lec­tion of data for sta­tis­ti­cal pur­pos­es but rather to know the iden­ti­ty of par­tic­u­lar peo­ple in order to use that knowl­edge to affect their lives.’ Giv­en the administration’s pos­ture toward immi­gra­tion, the fact that it wants to build a com­pre­hen­sive cit­i­zen­ship data­base is high­ly con­cern­ing. To Groves, it clear­ly sig­nals [16] ‘a bright line being crossed.’ . . .”

1. This pro­gram con­sists of a read­ing of an arti­cle by Yasha Levine, the author of Sur­veil­lance Val­ley: The Secret Mil­i­tary His­to­ry of the Inter­net. The pro­gram ana­lyzes the gen­e­sis of high tech and data pro­cess­ing, an ori­gin that is inex­tri­ca­bly linked with eugen­ics, anti-immi­grant doc­trine and–as is char­ac­ter­is­tic of fas­cism, the fear of the ubiq­ui­tous, malev­o­lent “oth­er.”

CORRECTION: The date of the Yasha Levine arti­cle is  ” . . . . 2019,” NOT 2018, as read in the pro­gram.

“The Racist Ori­gins of America’s Tech Indus­try” by Yasha Levine; Zero One; 4/28/2019. [17]

On a freez­ing day in Decem­ber 1896, an Amer­i­can inven­tor by the name of Her­man Hol­lerith rushed to catch a train out of the Russ­ian city of St. Peters­burg. He wore a fur cap and a thick fur-lined coat with a huge col­lar but­toned up all the way over his ears. It cov­ered his mouth as well as his big droopy mus­tache — leav­ing just a bit of pink flesh peek­ing out at the world.

Hol­lerith was a hypochon­dri­ac who pre­ferred stay­ing at home with his wife and moth­er-in-law, tin­ker­ing with inven­tions. He hat­ed trav­el, and he hat­ed trav­el­ing in Europe most of all. Like a 19th-cen­tu­ry ver­sion of a tech bro, he was obsessed with effi­cien­cy and mocked the locals for being bogged down by time-wast­ing tra­di­tions. “They are all liv­ing in what hap­pened thou­sands of years ago,” he wrote [18] to his wife from Italy. “I saw them cut­ting lum­ber on the road from Naples to Pom­peii, and, when I got to Pom­peii, I found paint­ings on walls show­ing exact­ly the same way of cut­ting lum­ber.”

For all his grum­bling about trav­el, the inven­tor had come far in his own life. Hol­lerith was only 36 years old and had been raised in a mod­est home by a wid­owed moth­er in New York, yet he had just spent weeks rub­bing shoul­ders with aris­to­crats from one the most exot­ic roy­al dynas­ties in the world. And now he was on his way back home with a fat and juicy con­tract for his new busi­ness ven­ture.

A few years ear­li­er, Russ­ian Czar Nicholas II issued an impe­r­i­al decree order­ing his min­is­ters to car­ry out the Russ­ian Empire’s first coun­try­wide cen­sus. With the 1897 dead­line loom­ing, they were scram­bling to com­ply. They knew it was going to be a mon­u­men­tal task — and per­haps an impos­si­ble one.

The peo­ple in charge of the cen­sus knew the only way to fin­ish the job in a rea­son­able amount of time would be to use the most advanced tech­nol­o­gy on the mar­ket. And that’s where the 36-year-old Hol­lerith came in.

The Russ­ian Empire had a pop­u­la­tion esti­mat­ed between 100 mil­lion and 200 mil­lion peo­ple, a range that might tell you why the czar need­ed to car­ry out a cen­sus. It stretched from Europe through the full length of Asia, a land­mass almost three times the size of the Unit­ed States. To count all these peo­ple, cen­sus tak­ers would have to trav­el to extreme­ly iso­lat­ed regions and poll peo­ple in dozens of dif­fer­ent lan­guages. And there was trou­ble brew­ing [19] already. Tatar Mus­lim com­mu­ni­ties in South­ern Rus­sia saw the planned count as a secret czarist plot to con­vert them to Chris­tian­i­ty, while some Russ­ian Ortho­dox sects saw the cen­sus as a sign of the Antichrist and vowed they would soon­er burn them­selves alive than sub­mit to such blas­phe­my.

Mak­ing the count was just the half of it. The data also need­ed to be tal­lied and ana­lyzed. The czar want­ed the cen­sus to be as mod­ern as pos­si­ble — includ­ing infor­ma­tion on age, lit­er­a­cy, gen­der, nation­al­i­ty, place of birth, res­i­den­cy, and occu­pa­tion. It was a bureaucrat’s worst night­mare.

The peo­ple in charge of the cen­sus knew the only way to fin­ish the job in a rea­son­able amount of time would be to use the most advanced tech­nol­o­gy on the mar­ket. And that’s where the 36-year-old Hol­lerith came in.

[11]

Her­man Hol­lerith

A few years ear­li­er, work­ing for the U.S. Cen­sus Bureau, Hol­lerith had devel­oped the world’s first func­tion­al mass-pro­duced com­put­er: the Hol­lerith tab­u­la­tor. An electro­mechan­i­cal device about the size of large desk and dress­er, it used punch cards and a clever arrange­ment of gears, sorters, elec­tri­cal con­tacts, and dials to process data with blaz­ing speed and accu­ra­cy. What had tak­en years by hand could be done in a mat­ter of months. As one U.S. news­pa­per described it, “with [the device’s] aid some 15 young ladies can count accu­rate­ly half a mil­lion of names in a day.”

Rus­sia was not the only coun­try inter­est­ed in Hollerith’s com­put­er tech­nol­o­gy. He had set up shop in New York just a few years ear­li­er but was already known in rar­i­fied bureau­crat­ic cir­cles around the globe. Cana­da, Ger­many, and Nor­way were eager to lease his machines. A com­pa­ny in Aus­tria was try­ing to pirate his designs and offer them to Euro­pean gov­ern­ments at a low­er cost.

Hollerith’s tab­u­la­tors could work with any kind of data and adapt to any large infor­ma­tion-inten­sive cor­po­rate enter­prise. Rail­road and insur­ance com­pa­nies were lin­ing up out­side Hollerith’s door for their own cus­tom-built data solu­tions.

Hollerith’s inven­tion caught the zeit­geist of the Sec­ond Indus­tri­al Rev­o­lu­tion — a time of rapid automa­tion and mech­a­niza­tion. It was an era of rail­roads, giant steam­er ships, telegraphs, radio, elec­tric­i­ty, mas­sive fac­to­ries, and unprece­dent­ed real-time inter­na­tion­al com­mu­ni­ca­tion. Ship­ping sched­ules, train timeta­bles, com­plex bank­ing data, actu­ary tables, social wel­fare pro­grams, and gov­ern­ment bud­gets were all pro­lif­er­at­ing faster than humans could keep track of them. Infor­ma­tion was king, and data pro­cess­ing was in con­stant demand.

With­in a few years, Hol­lerith was the mul­ti­mil­lion­aire own­er of a com­pa­ny that would even­tu­al­ly launch the U.S. com­put­er indus­try. A few decades after return­ing from Rus­sia, his tech­nol­o­gy would form the back­bone of Inter­na­tion­al Busi­ness Machines, or IBM — a glob­al con­glom­er­ate that for almost a cen­tu­ry would be syn­ony­mous with infor­ma­tion pro­cess­ing and com­put­er tech­nol­o­gy. Sold under IBM’s brand name, Hollerith’s tech­nol­o­gy would pow­er civil­ian gov­ern­ments, mil­i­taries, and cor­po­ra­tions around the world, crunch­ing num­bers into the Cold War and all the way to the dawn of the inter­net in the 1960s. The world did not sim­ply use Hollerith’s tab­u­la­tors; it became addict­ed to them and was shaped by them.

Hol­lerith was hailed as a genius. Many believed his inven­tion was part of a larg­er data-dri­ven tech­nol­o­gy rev­o­lu­tion that would lead to a bet­ter, more effi­cient, and more har­mo­nious world. One lead­ing Amer­i­can sta­tis­ti­cian pre­dict­ed it would ush­er in an age of “uni­ver­sal jus­tice” and “make inter­na­tion­al wars impos­si­ble.”

But for all the utopi­an talk about Hollerith’s com­put­ers, the tech­nol­o­gy had dark roots.

Dream­ing of MAGA

The U.S. Cen­sus — specif­i­cal­ly man­dat­ed by the Con­sti­tu­tion to take place every 10 years — is back in the news not only because the next count kicks off in 2020 but because, as it often has in the past, the cen­sus is a polit­i­cal flash­point with inevitable racial under­tones.

The cur­rent con­tro­ver­sy revolves around a plan devised [20] by Don­ald Trump’s for­mer advi­sor Steve Ban­non to add a cit­i­zen­ship ques­tion [21] to the 2020 cen­sus form. On the sur­face, it seems like an incon­se­quen­tial detail. But there is wide agree­ment that adding it will have pro­found polit­i­cal impli­ca­tions for a decade to come.

Aggre­gat­ed pop­u­la­tion data pro­vid­ed by the cen­sus is a crit­i­cal com­po­nent in our demo­c­ra­t­ic sys­tem. Its most impor­tant func­tion is to appor­tion Con­gres­sion­al rep­re­sen­ta­tion for the com­ing decade, but it also deter­mines the struc­ture of the Elec­toral Col­lege and guides the dis­tri­b­u­tion of hun­dreds of bil­lions in fed­er­al spend­ing. Objec­tions to the addi­tion of a cit­i­zen­ship ques­tion are based on con­cerns about under­count­ing. The fear, wide­ly shared by for­mer cen­sus offi­cials [22], is that ask­ing peo­ple for their cit­i­zen­ship sta­tus will push some immi­grants and Lati­nos to avoid tak­ing part in the cen­sus alto­geth­er. A large enough under­count of a spe­cif­ic minor­i­ty or socio-eco­nom­ic group will skew how seats are appor­tioned in the House of Rep­re­sen­ta­tives, shift­ing polit­i­cal pow­er and fed­er­al resources away from dis­tricts where these groups reside.

The Trump admin­is­tra­tion has claimed [23] that the cit­i­zen­ship ques­tion is being added for a good cause: to help the fed­er­al gov­ern­ment enforce the Vot­ing Rights Act and pro­tect minori­ties from vot­er dis­crim­i­na­tion. But many immi­grant advo­cates don’t buy this log­ic. They see the cit­i­zen­ship ques­tion as part and par­cel of Trump’s larg­er anti-immi­grant agen­da.

For most of its his­to­ry, the cen­sus — and the con­sti­tu­tion­al­ly man­dat­ed gov­ern­ment bureau­cra­cy that car­ries it out — has been inter­twined with nativism, big­otry, and fear of “the oth­er.”

Arturo Var­gas, exec­u­tive direc­tor of the Nation­al Asso­ci­a­tion of Lati­no Elect­ed and Appoint­ed Offi­cials, put it this way: “The deci­sion by Sec­re­tary of Com­merce Wilbur Ross to force the last-minute addi­tion of an untest­ed ques­tion on cit­i­zen­ship will result in an under­count of Lati­nos. While we do not know the true moti­va­tion behind these actions, we know the impact: as a con­se­quence of these actions, Cen­sus 2020 is on track to sig­nif­i­cant­ly under­count the Lati­no pop­u­la­tion in the Unit­ed States.”

Oth­ers are more direct.

“Our pres­i­dent, the face of our fed­er­al gov­ern­ment, which over­sees the cen­sus, has based his can­di­da­cy on a deeply anti-immi­grant plat­form,” says Bet­sy Plum, vice pres­i­dent of pol­i­cy at the New York Immi­gra­tion Coali­tion, one of the orga­ni­za­tions try­ing to stop the Trump admin­is­tra­tion from adding the ques­tion. “What the cit­i­zen­ship ques­tion did was it took a much broad­er fear and focused it right onto the cen­sus. The risk to a place like New York’s con­gres­sion­al rep­re­sen­ta­tion can­not be under­stat­ed.”

Demo­graph­ic shifts are already expect­ed [24] to result in North­east­ern states los­ing sev­er­al con­gres­sion­al seats, and the added ques­tion may make the sit­u­a­tion worse. “I think this is absolute­ly the intent of real­ly weaponiz­ing the cen­sus against immi­grant com­mu­ni­ties. Places like New York and places like Cal­i­for­nia are very much the tar­get.”

But there is anoth­er, less dis­cussed dimen­sion to the issue. Based on a close read­ing of inter­nal Depart­ment of Com­merce doc­u­ments tied to the cen­sus cit­i­zen ques­tion pro­pos­al, it appears [15] the Trump admin­is­tra­tion wants to use the cen­sus to con­struct a first-of-its-kind cit­i­zen­ship reg­istry for the entire U.S. pop­u­la­tion — a deci­sion that arguably exceeds the legal author­i­ty of the cen­sus.

“It was deep in the doc­u­men­ta­tion that was released,” Robert Groves, a for­mer Cen­sus Bureau direc­tor who head­ed the Nation­al Acad­e­mies com­mit­tee con­vened to inves­ti­gate the 2020 cen­sus, told me by tele­phone. “No one picked up on it much. But the term ‘reg­istry’ in our world means not a col­lec­tion of data for sta­tis­ti­cal pur­pos­es but rather to know the iden­ti­ty of par­tic­u­lar peo­ple in order to use that knowl­edge to affect their lives.”

Giv­en the administration’s pos­ture toward immi­gra­tion, the fact that it wants to build a com­pre­hen­sive cit­i­zen­ship data­base is high­ly con­cern­ing. To Groves, it clear­ly sig­nals [16] “a bright line being crossed.”

Mul­ti­ple states have chal­lenged the Trump administration’s plans, and their law­suits are head­ed to the Supreme Court, which is sched­uled to hear the case in April. Mean­while, at an over­sight hear­ing [25] in March, Demo­c­ra­t­ic U.S. Rep. Alexan­dria Oca­sio-Cortez dug into Wilbur Ross, Trump’s mul­ti-mil­lion­aire Sec­re­tary of Com­merce, who over­sees the U.S Cen­sus Bureau. She accused him of con­spir­ing with nativists and white suprema­cists [26] to add the cit­i­zen­ship ques­tion — and of over­step­ping his author­i­ty. “Why are we vio­lat­ing law to include this ques­tion?” she demand­ed [27].

What­ev­er the courts ulti­mate­ly decide, the lat­est debate over the cen­sus is hard­ly new. For most of its his­to­ry, the cen­sus — and the con­sti­tu­tion­al­ly man­dat­ed gov­ern­ment bureau­cra­cy that car­ries it out — has been inter­twined with nativism, big­otry, and fear of “the oth­er.”

The dark and ugly his­to­ry of the cen­sus makes it a unique­ly telling weath­er­vane of race pol­i­tics in Amer­i­ca. That the cen­sus simul­ta­ne­ous­ly played a cen­tral role in the devel­op­ment of the com­put­er age more than 130 years ago makes it dou­bly rel­e­vant, offer­ing a glimpse into how com­put­ers, sur­veil­lance, and racist gov­ern­ment poli­cies have been linked from the very begin­ning.

Count­ing for democ­ra­cy

Gov­ern­ments have been count­ing their peo­ple since the begin­ning of record­ed his­to­ry. You can find descrip­tions of cen­sus­es in the Old Tes­ta­ment, on Sumer­ian cuneiform tablets, and in the writ­ings of the ancient Greeks. There were cen­sus­es in pre-mod­ern Europe. Most Amer­i­can colonies kept pop­u­la­tion records, too. Gov­ern­ments count­ed peo­ple for two main rea­sons: rais­ing state rev­enue and wag­ing war. They need­ed to know who and what to tax, and they need­ed to know how many fight­ing-age men could be mobi­lized. It was the U.S. Con­sti­tu­tion that added a third and nov­el rea­son for count­ing peo­ple: rep­re­sen­ta­tion­al democ­ra­cy.

When the drafters of the U.S. government’s found­ing set of prin­ci­ples met in Philadel­phia in 1787, one of the first things they ham­mered out was a clause man­dat­ing that the pop­u­la­tion be count­ed every 10 years. This direc­tive for a decen­ni­al cen­sus appears up at the top of the Con­sti­tu­tion, long before the doc­u­ment gets around to lay­ing out the struc­ture of the gov­ern­ment. To the framers of the Con­sti­tu­tion, the cen­sus came first [28] because it deter­mined tax­a­tion and the bal­ance of con­gres­sion­al polit­i­cal pow­er.

Under the Con­sti­tu­tion, the num­ber of seats in the House of Rep­re­sen­ta­tives appor­tioned to each state would be based on pop­u­la­tion, which meant the gov­ern­ment need­ed to know the pre­cise num­ber of peo­ple liv­ing in each state.

By the end of the 1800s, the bureau­crat­ic prob­lem had become unten­able: The cen­sus was tak­ing near­ly 10 years to com­plete, mean­ing the results were out­dat­ed even before they came in.

The first cen­sus took place in 1790 and was over­seen by Thomas Jef­fer­son, who was then serv­ing as Sec­re­tary of State. It was most­ly a straight­for­ward head count [29] designed to meet the con­sti­tu­tion­al man­date. The whole enter­prise was expect­ed to take no more than nine months to com­plete. But despite its sim­plic­i­ty and our nation’s tiny pop­u­la­tion, it took near­ly two years to ful­ly tab­u­late. And it only got worse from there.

With every pass­ing decade, the cen­sus took longer to com­plete. It was filled with errors and under­counts, which led to nasty scan­dals and accu­sa­tions that the data was being manip­u­lat­ed for polit­i­cal ends. By the end of the 1800s, the bureau­crat­ic prob­lem had become unten­able: The cen­sus was tak­ing near­ly 10 years to com­plete, mean­ing the results were out­dat­ed even before they came in.

When the first cen­sus was car­ried out, there were 3.9 mil­lion peo­ple liv­ing in 13 states. By 1890, the U.S. encom­passed 42 states and had a pop­u­la­tion of 63 mil­lion — increas­ing 16 times over in the span of a cen­tu­ry. Nev­er before had a coun­try swelled so much so quick­ly. Still doing their work the old-fash­ioned way — with pen and paper — cen­sus work­ers strug­gled to keep up. They were drown­ing in data.

Mean­while, on top of hav­ing to enu­mer­ate a rapid­ly grow­ing pop­u­la­tion, gov­ern­ment offi­cials began to cram the cen­sus [30] with more and more ques­tions: data on occu­pa­tions, lit­er­a­cy lev­els, crim­i­nal his­to­ries, med­ical con­di­tions, home own­er­ship, eco­nom­ic trends, and a whole lot of prob­ing about people’s race and immi­gra­tion sta­tus.

As the 19th cen­tu­ry drew to a close, cen­sus offi­cials had start­ed trans­form­ing what should have been a sim­ple head count into a sys­tem of racial sur­veil­lance.

Anglo-Amer­i­can supe­ri­or­i­ty

It was a dif­fer­ent U.S. back then: small­er and most­ly rur­al and rapid­ly expand­ing along the west­ern fron­tier. The Civ­il War had come to an end, and with it, the U.S. Army shift­ed its resources to fight­ing and exter­mi­nat­ing Native Amer­i­cans west of the Mis­sis­sip­pi. Transcon­ti­nen­tal rail­roads were con­nect­ing huge swaths of the coun­try — shrink­ing time and space and shift­ing eco­nom­ic pow­er to a new set of finan­cial and rail­road busi­ness elites.

The country’s demo­graph­ics and race pol­i­tics were rapid­ly chang­ing, too.

Slav­ery had been abol­ished, allow­ing mil­lions of blacks to freely move, attempt to take charge of their own des­tinies, and play a role in the country’s polit­i­cal life. Immi­gra­tion was mak­ing itself felt. Well into the 19th cen­tu­ry, free immi­gra­tion into the U.S. had been large­ly dom­i­nat­ed by Eng­lish set­tlers. But start­ing in the 1850s, that pat­tern began chang­ing dras­ti­cal­ly. Mil­lions of Irish peas­ants streamed into the coun­try to escape the pota­to famine, which killed over one mil­lion peo­ple. Mil­lions more were flee­ing the crush­ing pover­ty of South­ern Italy and the east­ern ter­ri­to­ries of the Russ­ian Empire. Chi­nese labor­ers were arriv­ing on the West Coast en masse to build U.S. rail­roads.

This influx was a boon to an emerg­ing indus­tri­al oli­garchy, a source of nev­er-end­ing cheap labor. But it was also a source of polit­i­cal insta­bil­i­ty. Wide­spread inequal­i­ty and exploita­tion led to mas­sive­ly pop­u­lar move­ments for change. There were labor protests and strikes, the emer­gence of the pop­ulist move­ment [31], and a nation­wide self-help orga­ni­za­tion cre­at­ed by dirt-poor farm­ers. Social­ist and anar­chist ideas achieved broad adher­ence. Black civ­il rights activism emerged.

America’s polit­i­cal estab­lish­ment looked on this insta­bil­i­ty, social unrest, and change with hor­ror. They saw the mass­es of free blacks and Chi­nese, Jew­ish, Irish, and Ital­ian immi­grants — and their tat­tered clothes, alien lan­guages, unnat­ur­al reli­gions, and demands for bet­ter treat­ment and polit­i­cal rights — as a threat.

Grasp­ing about for solu­tions, many set­tled on var­i­ous strains of race sci­ence quack­ery [9]. So-called social Dar­win­ists relied on a twist­ed ver­sion of the the­o­ry of evo­lu­tion to explain why the poor and mar­gin­al­ized should remain that way while the wealthy and suc­cess­ful deserved to rule unchal­lenged. Tak­ing this notion a step fur­ther, adher­ents of eugen­ics fer­vent­ly believed that nat­u­ral­ly supe­ri­or Anglo-Amer­i­cans were on the verge of being wiped out due to the high birth rates of “degen­er­ate” and immi­grant stock. To head off this threat, they advo­cat­ed strict con­trols on repro­duc­tion — breed­ing humans for qual­i­ty in the same way that farm­ers did cows and hors­es.

These were not fringe ideas but were firm­ly embraced by the Amer­i­can cul­tur­al and polit­i­cal main­stream [32]. From future pres­i­dents like Theodore Roo­sevelt [33], Her­bert Hoover, and Calvin Coolidge to rob­ber barons like J.P. Mor­gan and Leland Stan­ford to writ­ers like H.G. Wells and pro­gres­sive activists like Mar­garet Sanger, eugen­ics was all the rage.

In the first decades of the 20th cen­tu­ry, 32 states passed ster­il­iza­tion laws [34] to deal with the threat of genet­ic degra­da­tion — laws that were upheld by the Supreme Court. And few wor­ried more about the threat of genet­ic degra­da­tion than the offi­cials at the U.S. Cen­sus Bureau.

Puri­ty police

Born into a wealthy Boston fam­i­ly, Fran­cis A. Walk­er served in the Civ­il War as a gen­er­al, dab­bled in jour­nal­ism, and ulti­mate­ly made a name for him­self as an influ­en­tial Pro­gres­sive Era econ­o­mist and sta­tis­ti­cian who would lat­er become pres­i­dent of the Mass­a­chu­setts Insti­tute of Tech­nol­o­gy.

As a pro­fes­sion­al econ­o­mist, Walk­er had a keen inter­est in the nation’s chang­ing demo­graph­ics — and he was hor­ri­fied by what he saw. Like most upper-class Amer­i­cans at the time, Walk­er believed [35] that the country’s orig­i­nal Eng­lish colonists had evolved to be the most supe­ri­or race on the plan­et — supe­ri­or even to the orig­i­nal Eng­lish race from which they sprang.

To him, Anglo-Amer­i­cans stood on the pin­na­cle of the world’s race pyra­mid. He and his peo­ple were “as far ahead of the Eng­lish as the Eng­lish were ahead of any oth­er branch of the Teu­ton­ic race, which was in turn far ahead of the Slavs or the Celts,” he wrote.

He believed that the influx of poor immi­grants from Ire­land and Italy as well as Jews and Slavs from East­ern Europe was dilut­ing the Unit­ed States’ supe­ri­or racial stock and threat­en­ing to drag Amer­i­can genet­ic supe­ri­or­i­ty back into a cesspool of degra­da­tion and decline. He blamed these immi­grants — “vast hordes of bru­tal­ized peas­ants” — for the social and polit­i­cal unrest that was hap­pen­ing around him.

“They are beat­en men from beat­en races; rep­re­sent­ing the worst fail­ures in the strug­gle for exis­tence,” he declared [36]. “They have none of the ideas and apti­tudes which fit men to take up read­i­ly and eas­i­ly the prob­lem of self-care and self-gov­ern­ment.”

He not only pushed to restrict immi­gra­tion in order to pre­vent what he viewed as Anglo-Amer­i­can “race sui­cide,” but also advo­cat­ed forced ster­il­iza­tion. “We must strain out of the blood of the race more of the taint inher­it­ed from a bad and vicious past,” he wrote. “The sci­en­tif­ic treat­ment which is applied to phys­i­cal dis­eases must be extend­ed to men­tal and moral dis­ease, and a whole­some surgery and cautery must be enforced by the whole pow­er of the state for the good of all.”

In addi­tion to his oth­er con­tri­bu­tions to U.S. life, Walk­er served as super­in­ten­dent of both the 1870 and 1880 U.S. Cen­sus.

Lim­its of tech­nol­o­gy

The cen­sus had been a racial instru­ment from its incep­tion, begin­ning with the orig­i­nal con­sti­tu­tion­al clause that instruct­ed cen­sus offi­cials to count black slaves sep­a­rate­ly from whites and to assign them a val­ue of only three-fifths of a per­son.

With each decade, new “racial” cat­e­gories [10] were invent­ed and added to the mix: “free col­ored males and females” and “mulat­to” were count­ed, includ­ing sub­di­vi­sions like includ­ing “quadroon” and “octoroon.” Cat­e­gories for Chi­nese, “Hin­doo,” and Japan­ese were added, as were “for­eign” and “native born” des­ig­na­tions for whites. The cen­sus slow­ly expand­ed to col­lect oth­er demo­graph­ic data, includ­ing lit­er­a­cy lev­els, unem­ploy­ment sta­tis­tics, and med­ical ail­ments, such as those who were “deaf, dumb, and blind” and the “insane and idi­ot­ic.” All of it was bro­ken down by race.

Most of these ques­tions were includ­ed in a hap­haz­ard fash­ion. They were overt­ly polit­i­cal, added in response to what­ev­er par­tic­u­lar racial fear gripped the nation­al rul­ing elite at the time.

The cen­sus need­ed to improve dras­ti­cal­ly. What it need­ed was a tal­ent­ed inven­tor, some­one young and ambi­tious who would be able to come up with a method to auto­mate tab­u­la­tion and data analy­sis. Some­one like Her­man Hol­lerith.

A racial cat­e­go­ry for Chi­nese was added after rail­road com­pa­nies began import­ing cheap, exploitable labor­ers from Chi­na. Cat­e­gories for “mulat­to” came after the abo­li­tion of slav­ery caused a pan­ic about the dan­gers of racial mix­ing. Ques­tions about men­tal health and race were first added at the behest of a South­ern sen­a­tor right before the out­break of Civ­il War. The results seemed to show that free blacks liv­ing in North­ern states were on aver­age 11 times more like­ly [37] to be insane than South­ern blacks liv­ing in slav­ery. Such ques­tion­able sta­tis­tics were tak­en up by South­ern politi­cians to bol­ster racist the­o­ries and argue against abo­li­tion.

To Walk­er, these ear­ly efforts didn’t go near­ly far enough. As an econ­o­mist and sta­tis­ti­cian, he want­ed to col­lect and process more data [38] and to pro­fes­sion­al­ize and stan­dard­ize the effort. He want­ed it to be a prop­er, sci­en­tif­ic “nation­al inven­to­ry [39]” — not a hap­haz­ard col­lec­tion of facts.

But his dreams kept run­ning up against a hard lim­it: tech­nol­o­gy. The cen­sus was still count­ed and ana­lyzed by hand. The work was slow and lim­it­ed. Sophis­ti­cat­ed analy­sis was next to impos­si­ble.

The cen­sus need­ed to improve dras­ti­cal­ly. What it need­ed was a tal­ent­ed inven­tor, some­one young and ambi­tious who would be able to come up with a method to auto­mate tab­u­la­tion and data analy­sis.

Some­one like Her­man Hol­lerith.

The inven­tor

Hol­lerith was born in Buf­fa­lo, New York, in 1860. His father, a clas­sics teacher, died when he was a child, and he was raised by his moth­er. In 1879, when he grad­u­at­ed from the Colum­bia School of Mines with a degree in engi­neer­ing, he was imme­di­ate­ly recruit­ed to help com­pile eco­nom­ic sta­tis­tics for the 1880 cen­sus, which was being run by Walk­er.

At his new job, Hol­lerith, who had devel­oped a rep­u­ta­tion as an inven­tive engi­neer in col­lege, was encour­aged by senior cen­sus offi­cials [40] to study the enu­mer­a­tion process and come up with a solu­tion to speed it up. After the 1880 cen­sus was com­plete, he quit his job for a teach­ing posi­tion at MIT —fol­low­ing Walk­er, who had recent­ly been appoint­ed pres­i­dent.

Hol­lerith kept tin­ker­ing with his inven­tion, and before long, he came up with a design [41] that sep­a­rat­ed the enu­mer­a­tion process into parts. The first involved con­vert­ing data into a for­mat that could be read by a machine. This he accom­plished by punch­ing holes on a piece of paper. The sec­ond step involved pro­cess­ing the data. This was accom­plished by feed­ing the paper through a machine that, through a com­bi­na­tion of pins and dials, read the num­ber and posi­tion of the holes. At first, Hol­lerith exper­i­ment­ed with using a con­tin­u­ous strip of paper — like the recent inven­tion of tick­er tape, which was wide­ly used to trans­mit stock prices via tele­graph. But he wasn’t hap­py with the results.

“The trou­ble was that if, for exam­ple, you want­ed any sta­tis­tics regard­ing Chi­na­men, you would have to run miles of paper to count a few Chi­na­men,” Hol­lerith lat­er explained [42] in a let­ter. Race was nev­er far from his mind when work­ing on his con­trap­tion.

He even­tu­al­ly hit upon a much bet­ter idea: Each per­son would be rep­re­sent­ed by their own punch card — an idea he picked up while tak­ing a train. “I was trav­el­ing in the West and I had a tick­et with what I think was called a punch pho­to­graph… the con­duc­tor… punched out a descrip­tion of the indi­vid­ual, as light hair, dark eyes, large nose, etc.,” he explained [43], not­ing that he’d sim­ply done the same thing.

The dawn of data

In March 1890, Hollerith’s machines [44] were installed at the Inter-Ocean Build­ing on Ninth Street in Wash­ing­ton, D.C., not far from the White House. He over­saw the instal­la­tion him­self, run­ning around and bark­ing orders to work­men who were hoist­ing creaky wood­en crates from the street to the third floor.

Soon the prop­er­ty was trans­formed from a non­de­script office space into the bustling head­quar­ters of the 11th cen­sus. Hun­dreds of clerks worked around the clock in shifts, tak­ing raw cen­sus data col­lect­ed in the field and trans­fer­ring it onto cards using spe­cial­ly designed hole punch machines and then pass­ing the cards to anoth­er set of clerks who worked the tab­u­la­tors and sorters. Hollerith’s machines clanked away all day and all night, with clerks crammed togeth­er like sweat­shops work­ers.

News­pa­pers sent their cor­re­spon­dents to gawk [45] at these futur­is­tic con­trap­tions. Because of the mis­er­able track record of ear­li­er cen­sus­es, the press was awash with pre­dic­tions of incom­pe­tence and fail­ure. They were wrong.

It would take a full four years to fin­ish and release the reports. It was an amaz­ing improve­ment over the pre­vi­ous cen­sus, which took near­ly a decade.

The 1890 cen­sus — the nation’s 11th — was the most ambi­tious yet. It con­tained 35 ques­tions [46], 10 more than the pre­vi­ous cen­sus, on a whole range of data: lit­er­a­cy lev­els, sizes of house­hold, pro­fes­sions, the val­ue of a family’s prop­er­ty, and whether they rent­ed or owned. Per­haps most impor­tant was the racial dimen­sion. The cen­sus col­lect­ed stats on native and for­eign-born Amer­i­cans and broke them into mul­ti­ple racial cat­e­gories: white, col­ored, Chi­nese, Japan­ese and “civ­i­lized Indi­an” (i.e., a Native Amer­i­can no longer liv­ing in a trib­al soci­ety). It was the first cen­sus to include a com­plete count of Native Amer­i­cans liv­ing on trib­al lands. It asked for data on unem­ploy­ment his­to­ry, fer­til­i­ty rates, cit­i­zen­ship sta­tus, crim­i­nal his­to­ry, lit­er­a­cy, and Eng­lish lan­guage pro­fi­cien­cy.

Despite the long list of ques­tions and require­ments for cal­cu­lat­ing a whole slew of new stats, includ­ing birth, unem­ploy­ment, and caus­es of death divid­ed up by race, the basic pop­u­la­tion count was com­plet­ed in just six weeks. It would take a full four years to fin­ish tab­u­lat­ing and edit­ing [47] all oth­er cat­e­gories of data and release the reports. It was an amaz­ing improve­ment over the pre­vi­ous cen­sus, which took near­ly a decade [48].

It wasn’t just the speed that set Hollerith’s inven­tion apart. It was its abil­i­ty to mine and sift through data and even com­bine mul­ti­ple data points. Such fine-grained analy­sis on a mass scale was com­plete­ly unprece­dent­ed [49], and it made Hollerith’s machines an imme­di­ate hit with the Unit­ed States’ race-obsessed polit­i­cal class.

Robert Porter, head of the 1890 cen­sus, who had over­seen the adop­tion of Hollerith’s tab­u­la­tor machines, was deeply impressed [50] by their pow­er to sort immi­grant and non-white pop­u­la­tions based on numer­ous demo­graph­ic vari­ables. He was par­tic­u­lar­ly pleased about being able to ana­lyze the three things most feared by the “race sui­cide” crowd: immi­gra­tion rates, immi­grant fer­til­i­ty rates, and mixed race mar­riages (or what the cen­sus called the “con­ju­gal con­di­tion”), all of which could be bro­ken down by age, race, lit­er­a­cy lev­els, and nat­u­ral­iza­tion sta­tus.

Simon New­ton Dex­ter North, a long­time wool indus­try lob­by­ist who would head the 1900 cen­sus, was also daz­zled by the pow­er of Hollerith’s tab­u­la­tors. Like Walk­er and oth­er cen­sus col­leagues, he was obsessed with immi­gra­tion and cross-breed­ing. He believed they were dilut­ing the country’s supe­ri­or Anglo-Amer­i­can stock and desta­bi­liz­ing soci­ety.

“This immi­gra­tion is pro­found­ly affect­ing our civ­i­liza­tion, our insti­tu­tions, our habits and our ideals,” he warned in 1914 [51]. “It has trans­plant­ed here alien tongues, alien reli­gions, and alien the­o­ries of gov­ern­ment; it has been a pow­er­ful influ­ence in the rapid dis­ap­pear­ance of the Puri­tan­i­cal out­look upon life.”

North believed that bureau­crats and sta­tis­ti­cians like him were fight­ing a new kind of war — a war for America’s genet­ic puri­ty. And Hollerith’s tab­u­la­tor tech­nol­o­gy was a vital weapon — an “epoch-mak­ing” inven­tion — with­out which this fight would be lost.

Feed­ing the nativist beast

Overnight, Hollerith’s tab­u­la­tor tech­nol­o­gy had trans­formed cen­sus tak­ing from a sim­ple head count into some­thing that looked very much like a crude form of mass sur­veil­lance. To the race-obsessed polit­i­cal class, it was a rev­o­lu­tion­ary devel­op­ment. They could final­ly put the nation’s eth­nic make­up under the micro­scope. The data seemed to con­firm the nativists’ worst fears: Poor, illit­er­ate immi­grants were swarm­ing America’s cities, breed­ing like rab­bits, and out­strip­ping native Anglo-Amer­i­can birth rates.

Imme­di­ate­ly fol­low­ing the cen­sus, the states and the fed­er­al gov­ern­ment passed a flur­ry of laws [12] that heav­i­ly restrict­ed immi­gra­tion.

It start­ed with the Immi­gra­tion Act of 1891, which set up the first fed­er­al agency to over­see immi­gra­tion and bor­der con­trol and turned an unused island on the south­ern tip of Man­hat­tan into an elab­o­rate screen­ing cen­ter for immi­grants. It con­tin­ued through the pas­sage of a half-dozen major immi­gra­tion bills, includ­ing one that stripped women of U.S. cit­i­zen­ship if they mar­ried non-nat­u­ral­ized for­eign­ers, cul­mi­nat­ing in the Immi­gra­tion Act of 1924 — a land­mark piece of leg­is­la­tion that intro­duced race immi­gra­tion quo­tas.

This suite of laws gave immi­gra­tion offi­cials the pow­er to ban just about any­one, includ­ing “idiots, imbe­ciles, and fee­ble-mind­ed per­sons” or those who exhib­it­ed “con­sti­tu­tion­al psy­cho­path­ic infe­ri­or­i­ty” or were “men­tal­ly or phys­i­cal­ly defec­tive.” Anar­chists and social­ists were banned out­right as was any­one from the “Asi­at­ic Barred Zone,” which includ­ed most of Asia, the sub-con­ti­nent, the Mid­dle East, and parts of east­ern Rus­sia. Mean­while, immi­gra­tion from Euro­pean coun­tries was con­strained by hard lim­its based on the 1890 cen­sus — the first cen­sus processed by Hol­lerith tech­nol­o­gy. Com­bined with the anti-Chi­nese bills passed in the late 19th cen­tu­ry, these new laws cre­at­ed a vir­tu­al wall around the U.S. Immi­gra­tion rates plunged.

North dreamed of the day when detailed racial data could be col­lect­ed and ana­lyzed for the whole world and be used to guide human genet­ic devel­op­ment. His dream would soon be real­ized in Europe.

The data pro­vid­ed by Hollerith’s inven­tion did not cause the racism, nativism, and eugen­ics that saw class and pover­ty through the lens of breed­ing rather than pol­i­tics and eco­nom­ic pol­i­cy. But it gave those fears con­crete shape — and it pro­vid­ed data to which those fears could be hitched.

To some U.S. bureau­crats, this data-dri­ven eugen­ics sys­tem was just the begin­ning. North, who direct­ed the U.S. Cen­sus Bureau from 1903 to 1909, dreamed of the day when detailed racial data could be col­lect­ed and ana­lyzed for the whole world and be used to guide human genet­ic devel­op­ment.

“The need for restrain­ing the genet­i­cal­ly defi­cient class­es and fam­i­lies from the func­tion of repro­duc­tion, is rec­og­nized as imper­a­tive,” he wrote in 1918 [52] from his perch at the Carnegie Endow­ment for Inter­na­tion­al Peace as World War I was com­ing to an end. “It is the dream of the true sta­tis­ti­cian that the day will some time arrive when the facts of demog­ra­phy will be avail­able on iden­ti­cal bases for the entire globe. When that dream is real­ized, when com­pa­ra­ble inter­na­tion­al sta­tis­tics actu­al­ly and every­where exist, then we shall know the laws which deter­mine human progress and can effec­tive­ly apply them.”

His dream would soon be real­ized in Europe.

Hol­lerith goes glob­al

The imme­di­ate suc­cess of his inven­tion made Hol­lerith wealthy and famous. But that was just the begin­ning. In 1911, he sold his Tab­u­lat­ing Machine Com­pa­ny for a $2.3 mil­lion to Charles Flint, an infa­mous ven­ture cap­i­tal­ist known in his day as the “Father of Trusts.”

Flint bought out Hol­lerith, com­bined his com­pa­ny with sev­er­al oth­er busi­ness­es that made pre­ci­sion mechan­i­cal con­trap­tions — clocks, cash reg­is­ters, cof­fee grinders, and butch­er scales — to cre­ate a com­pu­ta­tion­al monop­oly and hand­ed this new con­glom­er­ate over to an ambi­tious young exec­u­tive by the name of Thomas J. Wat­son.

As Hol­lerith slow­ly went senile in retire­ment, Wat­son ruth­less­ly lever­aged the aging inventor’s com­put­er tech­nol­o­gy to crush com­pe­ti­tion and estab­lish a glob­al monop­oly in the ear­ly com­pu­ta­tion mar­ket. The result was Inter­na­tion­al Busi­ness Machines, the com­pa­ny we now know as IBM, found­ed in 1911.

Installed in fac­to­ries, cor­po­rate offices, and city and mil­i­tary bureau­cra­cies, his tab­u­la­tor com­put­ers not only sped up account­ing but great­ly reduced labor costs. Busi­ness­es and local and fed­er­al gov­ern­ment agen­cies ordered Hol­lerith machines by the truck­load. Insur­ance com­pa­nies relied on them for account­ing and cal­cu­lat­ing actu­ary tables. Rail­roads used them to route freight and work out sched­ules. At one rail­road com­pa­ny, a sin­gle Hol­lerith machine oper­at­ed by two peo­ple replaced the full time work of 20 clerks. They per­son­i­fied the blaz­ing effi­cien­cy and automa­tion of the tech­no­log­i­cal rev­o­lu­tion sweep­ing the Pro­gres­sive Era.

Nowhere was this as obvi­ous as the Social Secu­ri­ty Admin­is­tra­tion, one of the sig­na­ture pro­grams of the New Deal.

Pres­i­dent Franklin D. Roo­sevelt signed the Social Secu­ri­ty Act into law on Aug. 14, 1935, cre­at­ing America’s first old-age pen­sion pro­gram. The Social Secu­ri­ty Act brought about a mas­sive need for account­ing and data pro­cess­ing for both busi­ness­es and the fed­er­al gov­ern­ment. Busi­ness­es sud­den­ly had to keep metic­u­lous records on their employ­ees. They need­ed to track salaries and Social Secu­ri­ty con­tri­bu­tions and file that infor­ma­tion with the fed­er­al gov­ern­ment. The gov­ern­ment, in turn, had to process all that data. It need­ed to mon­i­tor con­tri­bu­tions to each indi­vid­ual Social Secu­ri­ty account over the life­time of each indi­vid­ual. And then, when they hit retire­ment age, it had to cut month­ly checks to mil­lions of Amer­i­cans.

As soon as the leg­is­la­tion passed, busi­ness­es queued up at IBM to get the prop­er tab­u­la­tor pay­roll sys­tems to meet fed­er­al account­ing require­ments. Phones at IBM’s sales offices rang off the hook. A Wool­worth exec­u­tive com­plained to IBM that hand­ing the paper­work to com­ply with the Social Secu­ri­ty Act alone would cost the com­pa­ny a quar­ter of a mil­lion dol­lars a year — $4.5 mil­lion today.

IBM won the con­tract to over­see account­ing for the Social Secu­ri­ty Admin­is­tra­tion, beat­ing out com­peti­tors like Rem­ing­ton Rand. It was the only com­put­er com­pa­ny at the time that had the expe­ri­ence and pro­duc­tion capac­i­ty to under­take a project of that size. As one offi­cial IBM his­to­ry put it, “the Social Secu­ri­ty project cat­a­pult­ed IBM from a mid­size cor­po­ra­tion to the glob­al leader in infor­ma­tion tech­nol­o­gy.”

Nat­u­ral­ly, the mil­i­tary was a big fan of the tech­nol­o­gy. In peace­time, the Depart­ment of War used the machines to keep track of enlist­ment data and track mil­i­tary pen­sions. When the U.S. entered the war, IBM’s Hol­lerith tech became a vital part of the Allied mil­i­tary effort.

Hol­lerith machines were involved in almost every part of the war, from design­ing the atom­ic bomb to man­ag­ing troop deploy­ment. Spe­cial “portable” IBM machines installed on trucks [53] land­ed with U.S. troops in Nor­mandy, Tunisia, Sici­ly, and Italy. They were used on the home front as well.

Hol­lerith tab­u­la­tors were a big hit all over the world. But one coun­try was par­tic­u­lar­ly enam­ored with them: Nazi Ger­many.

Fol­low­ing the attack on Pearl Har­bor, the U.S Cen­sus Bureau hauled out the punch cards from the 1940 cen­sus and reprocessed them to pro­duce block-by-block [54] pop­u­la­tion lists on Japan­ese-Amer­i­cans in a half-dozen states, includ­ing Cal­i­for­nia. Ulti­mate­ly 130,000 Japan­ese-Amer­i­cans were forced to move to con­cen­tra­tion camps.

The head of the Cen­sus Bureau at the time, James Clyde Capt, was ecsta­t­ic with the data they were able to gen­er­ate. For instance, he wrote to a sub­or­di­nate [55], if the data showed “there were 801 Japs in a com­mu­ni­ty and [author­i­ties] only found 800 of them, then they have some­thing to check up on.”

Nazis and num­bers

Hol­lerith tab­u­la­tors were a big hit all over the world. But one coun­try was par­tic­u­lar­ly enam­ored with them: Nazi Ger­many.

Adolf Hitler came to pow­er on the back of the eco­nom­ic dev­as­ta­tion that fol­lowed Germany’s defeat in World War I. To Hitler, how­ev­er, the prob­lem plagu­ing Ger­many was not eco­nom­ic or polit­i­cal. It was racial. As he put it in Mein Kampf [56]: “The state is a racial organ­ism and not an eco­nom­ic orga­ni­za­tion.” The rea­son Ger­many had fall­en so far, he argued, was its fail­ure to tend to its racial puri­ty. There were only about a half-mil­lion Jews in Ger­many in 1933 — less than 1% of the pop­u­la­tion — but he sin­gled them out as the root cause of all of the nation’s prob­lems.

Hitler and the Nazis drew much of their inspi­ra­tion [57] from the U.S. eugen­ics move­ment and the sys­tem of insti­tu­tion­al racism that had arisen in slavery’s wake. Their solu­tion was to iso­late the so-called mon­grels, then con­tin­u­ous­ly mon­i­tor the racial puri­ty of the Ger­man peo­ple to keep the volk free of fur­ther con­t­a­m­i­na­tion.

The only prob­lem: How to tell who is real­ly pure and who is not?

[58]

Third Reich IBM Data Entry Work­ers

The U.S. had a ready solu­tion. IBM’s Ger­man sub­sidiary land­ed its first major con­tract the same year Hitler became chan­cel­lor. The 1933 Nazi cen­sus was pushed through by Hitler as an emer­gency genet­ic stock-tak­ing of the Ger­man peo­ple. Along with numer­ous oth­er data points, the cen­sus focused on col­lect­ing fer­til­i­ty data for Ger­man women — par­tic­u­lar­ly women of good Aryan stock. Also includ­ed in the cen­sus was a spe­cial count of reli­gious­ly obser­vant Jews, or Glauben­sju­den.

Nazi offi­cials want­ed the entire count, esti­mat­ed to be about 65 mil­lion peo­ple, to be done in just four months. It was a mon­u­men­tal task, and Ger­man IBM offi­cials worked around the clock to fin­ish it. So impor­tant was the suc­cess of the con­tract to IBM that CEO Thomas J. Wat­son per­son­al­ly toured the giant Berlin ware­house where hun­dreds of female clerks worked in rotat­ing sev­en-hour shifts 24 hours a day.

Wat­son came away great­ly impressed with the work of his Ger­man man­agers. They had pulled off a seem­ing­ly impos­si­ble assign­ment, one that was com­pli­cat­ed by a cus­tom-enlarged punch card for­mat nec­es­sary for “polit­i­cal con­sid­er­a­tions [59]” — IBM’s cod­ed expla­na­tion for the extra data demands the Nazi regime required.

As Hitler’s Nazi Par­ty tight­ened its grip on Ger­many, it launched all sorts of addi­tion­al data-gath­er­ing pro­grams to puri­fy the Ger­man nation. And IBM helped them do it.

“[T]he pre­con­di­tion for every depor­ta­tion was accu­rate knowl­edge of how many Jews in a par­tic­u­lar dis­trict fit­ted the racial and demo­graph­ic descrip­tions in Berlin’s quo­tas,” write David Mar­tin Lue­bke and Sybil Mil­ton in “Locat­ing the Vic­tim [60],” a study into Nazi use of the tab­u­la­tor machines. “Armed with these data,” they said, “the Gestapo often proved able to antic­i­pate with remark­able accu­ra­cy the total num­ber of depor­tees for each racial, sta­tus, and age cat­e­go­ry.”

Germany’s vast state bureau­cra­cy and its mil­i­tary and rear­ma­ment pro­grams, includ­ing the country’s grow­ing con­cen­tra­tion camp/slave labor sys­tem, also required data pro­cess­ing ser­vices. By the time the U.S. offi­cial­ly entered the war in 1941, IBM’s Ger­man sub­sidiary had grown to employ 10,000 peo­ple and served 300 dif­fer­ent Ger­man gov­ern­ment agen­cies [60]. The Nazi Par­ty Trea­sury; the SS; the War Min­istry; the Reichs­bank; the Reich­spost; the Arma­ments Min­istry; the Navy, Army and Air Force; and the Reich Sta­tis­ti­cal Office — the list of IBM’s clients went on and on.

This his­to­ry reveals an uncom­fort­able and fun­da­men­tal truth about com­put­er tech­nol­o­gy.

“Indeed, the Third Reich would open star­tling sta­tis­ti­cal venues for Hol­lerith machines nev­er before insti­tut­ed — per­haps nev­er before even imag­ined,” wrote Edwin Black in IBM and the Holo­caust [14], his pio­neer­ing 2001 exposé of the for­got­ten busi­ness ties between IBM and Nazi Ger­many. “In Hitler’s Ger­many, the sta­tis­ti­cal and cen­sus com­mu­ni­ty, over­run with doc­tri­naire Nazis, pub­licly boast­ed about the new demo­graph­ic break­throughs their equip­ment would achieve.” (IBM has crit­i­cized [61] Black’s report­ing meth­ods, and has said that its Ger­man sub­sidiary large­ly came under Nazi con­trol before and dur­ing the war.)

Demand for Hol­lerith tab­u­la­tors was so robust that IBM was forced to open a new fac­to­ry in Berlin to crank out all the new machines. At the facility’s chris­ten­ing cer­e­mo­ny, which was attend­ed by a top U.S. IBM exec­u­tive and the elite of the Nazi Par­ty, the head of IBM’s Ger­man sub­sidiary gave a rous­ing speech about the impor­tant role that Hol­lerith tab­u­la­tors played in Hitler’s dri­ve to puri­fy Ger­many and cleanse it of infe­ri­or racial stock.

“We are very much like the physi­cian, in that we dis­sect, cell by cell, the Ger­man cul­tur­al body,” he said [62]. “We report every indi­vid­ual characteristic…on a lit­tle card. These are not dead cards, quite to the con­trary, they prove lat­er on that they come to life when the cards are sort­ed at a rate of 25,000 per hour accord­ing to cer­tain char­ac­ter­is­tics. These char­ac­ter­is­tics are grouped like the organs of our cul­tur­al body, and they will be cal­cu­lat­ed and deter­mined with the help of our tab­u­lat­ing machine.”

On the sur­face, it may seem like the sto­ry of Her­man Hol­lerith and the U.S. cen­sus are his­tor­i­cal relics, an echo from a bygone era. But this his­to­ry reveals an uncom­fort­able and fun­da­men­tal truth about com­put­er tech­nol­o­gy. We can thank nativism and the cen­sus for help­ing to bring the com­put­er age into exis­tence. And as the bat­tle over the 2020 cen­sus makes clear, the dri­ve to tal­ly up our neigh­bors, to sort them into cat­e­gories and turn them into sta­tis­tics, still car­ries the seed of our own dehu­man­iza­tion.