IBM and the Holocaust

by Edwin Black

  No less a Vichy commander than Admiral Francois Darlan, vice prime minister of Vichy with direct oversight of the anti-Jewish bureaus, was quoted as telling his Cabinet: “The stateless Jews who, for the past 15 years have invaded our country do not interest me. But the others, the good old French Jews, are entitled to all the protection that we can give them: I even have some of them in my family.” Hence, a long list of special exemptions crept into the official French enforcement of anti-Semitic statutes, on either side of the Vichy line.96

  Those in both zones quickly learned that their anti-Jewish collaboration would be dealt with when France was liberated. As early as November 11, 1940, the leader of Free French forces, Charles DeGaulle, issued Nazi surrogates in Vichy a warning: “Be assured… the cruel decrees directed against French Jews can and will have no validity in Free France. These blows are no less a blow against the honor of France than they are an injustice against her Jewish citizens…. the wrongs done in France itself [will] be righted.” DeGaulle broadcast his remarks from recaptured French Equatorial Africa, and then asked that they be proclaimed to a meeting of the American Jewish Congress at Carnegie Hall.97

  Oppressive Nazi rule could have dictated its iron will to all reluctant French authorities, and conquered the demographic uncertainties of a French Jewry in two zones if only the Holleriths could be deployed. That is precisely what Holleriths brought to any problem—organization where there was disorder and tabular certainty where there was confusion. The Nazis could have punch-carded the Jews of France into the same genocidal scenario in force elsewhere, including Holland. But in the aftermath of the MB’s technologic ravages, France’s punch card infrastructure was simply incapable of supporting the massive series of programs Berlin required. Even if the machines could have been gathered, transferred, or built—CEC just didn’t have the punch cards.

  CEC had openly reported to IBM NY about its inability throughout 1941 and 1942 to locate a dependable paper supply. Even with its reduced customer load, CEC clients required 50 tons of paper monthly. But the French subsidiary’s deliveries were generally rationed to as little as 15 tons per month—a mere ten-day supply. The Reich was diverting the bulk of the cellulose needed for pulping into nitro-cellulose at explosive plants throughout France. On January 1, 1942, CEC’s punch card paper stocks totaled 318 tons. One year later, CEC informed IBM NY that the on-hand inventory had dwindled to just 222 tons, of which only 71 tons were from a paper mill that could reliably produce the high-tech stock tabulators required; the remaining 151 tons had come from a new and untested vendor. Bad paper only jammed Holleriths, worsening the situation.98

  No wonder CEC bluntly informed IBM NY, “The outlook was very dark… we tried to obtain paper by all available means.” Only by scraping the barrels of substitute suppliers could CEC “live under the most drastic [paper] restrictions, which we imposed on our customers.”99

  In September 1940, just after German forces transferred hundreds of French Holleriths to the Greater Reich, the first census was ordered in Occupied France.100

  During the months and war years to come, exactly which censuses were taken, by whom, and by what method, would constitute a maze of mysteries, befuddling all in France and Germany. Registrations and census efforts would be announced and scheduled, and then delayed and rescheduled. They would take place over a period of months, but often the deadlines would be extended because the results were so incomplete. Some counting efforts were undertaken just in Vichy, some just in Occupied France, some in both. Many campaigns counted Jews; many counted the entire French population. Most were executed by a myriad of ad hoc census styles from the inept to the diligent. Incomplete, inaccurate, and inconsistent data ruled the entire enterprise. Heads of households were often counted instead of entire families. Children were often not included. Addresses were not infrequently omitted. Confused officials were on occasion forced to admit they simply had no idea where many of the Jews of France were.101

  No wonder the Prefect of Tarn complained in a December 22, 1941, letter to the General Commission for Jewish Questions, or GCJQ, that the results of the census ordered just months earlier were already obsolete since the method did not systematically record changes of address.102

  A later census inspector’s report summarizing a major inquiry into the head-counts of eight provinces concluded: “If it was needed, this work is proof that the census of Jews prescribed by the June 2, 1941 law was poorly done… a new census is necessary.”103

  Germany’s mushrooming labor needs only intensified the counting havoc. By 1942, Berlin had demanded as many as 600,000 French conscript laborers under a strict recruitment schedule as the requirements of the Reich’s war work changed. In October 1942, the Reich demanded France deliver 35,000 of its railway workers for assignments in Germany. Earlier, France had suggested the release of 50,000 French prisoners of war in exchange for 150,000 skilled workers. Recruitment projects required waves of age-specific worker registrations. One labor census ordered all French men and women between the ages of 18 and 50, plus anyone working less than thirty hours per week, to register. A second survey sought unmarried women between the ages of 21 and 35 by professional category. Yet another sought to immediately register all young men in a major youth movement over the age of 21. All three census and registration programs were ordered within weeks during fall 1942.104 Dozens more, equally disorganized, were undertaken during the occupation years.

  A new census program would be launched before the previous one was completed. Germany was of course accustomed to multiple, overlapping census and registration programs, but only because it could organize them with battalions of card punchers and fleets of Holleriths fed by endless IBM cards. Those simply did not exist in France.

  Cascading chaos as Germany sought to count Jews and others in France ensured that the multifarious census and registration efforts would not only be misreported, misunderstood, and mishandled throughout the war years, but would be misinterpreted for decades after liberation as well.105

  A fundamental cause of France’s profound counting disarray arose from its decentralized, almost anarchic, registration infrastructure. Registration was not implemented by professional statisticians or experienced census offices. With no one to do the job properly, Germany assigned it to the nation’s police departments, the prefectures. That stood to reason, since police departments for years were accustomed to registering refugee Jews who entered their jurisdiction. Each prefecture executed its own count in its own way, employing its own interpretations, and not always using the same forms as the next prefecture. They did not use punch cards, but small colored pieces of paper and index cards. The machines they utilized were not IBM Holleriths whirring at great speed, but Remington typewriters with sticking keys that constantly broke down. Pen and pencil were readily used when typing ribbon was not available.106

  The first real effort to systematically count French Jews had been rumored since the summer, but was finally announced in late September 1940. Jews in the Occupied Zone were ordered to register with information about their businesses at police stations on specified days, according to an alphabetical sequence. The entire process was to take eighteen days.107

  The numerical tally for the Seine province, including Paris, was reported as 149,734; of these, 85,664 were categorized as French nationals and 64,070 foreign born. The northern Jewish group registered approximately 11,000 Jewish businesses as well. Outside the Paris area, an additional 20,000 Jews were counted. These numbers were for the Occupied Zone only.108

  It was one thing to count the Jews numerically even if the count was approximate. It was quite another to track them month in and month out, and organize them centrally for either ghettoization or deportation. With a Hollerith, that diverse information would be sorted, tallied, and summarized to yield the desired results. But rather than compile information in succinct automated tabulated results, French information was segmented into a series of traditional paper forms: yellow, beige, white, and
red—often as many as five forms to complete a single personal file. The forms, which came into use in January 1941, were known as Tulard files, named for the Vichy police bureaucrat Andre Tulard, stationed in Paris. Quickly, the Tulard file became famous within French authoritarian circles, which, under the circumstances, seemed the best means of tracking Jews. Soon, the Tulard system, which had first been implemented in the Occupied Zone, was adopted in Vichy as well.109

  Unlike a Hollerith, which proofed and verified all its information to avoid errors, the Tulard system had no method of automatically deleting duplications. After the October 1940 counts for the Occupied Zone, some French newspapers began extrapolating the raw numbers for all of France. On March 4, 1941, Bunle, head of the General Statistics Office of France, advocated another national census. “Certain newspapers have recently published fanciful evaluations of the numbers of Jews in France,” Bunle wrote. “These very exaggerated evaluations equal several times the actual number of Jews in the territory.”110

  So a second major census was launched on June 14, 1941. This one covered both zones, counting 287,962 Jews in 87 of France’s 90 geographic “departments.” German forces oversaw the count in Occupied France. Vichy performed the function in the south.111

  The Jewish catalog that emerged required six pages of instructions sent to the police prefectures. Each multi-card file would be marked with various letters: J for Jewish; NJ for non-Jewish, or for someone cleared of Jewish ancestry; N for nationality; D for domicile information; and P for their professional details. Multiple manual catalogs were set up for foreign Jews and French-born Jews. One whole card was designated just to record a woman’s maiden name.112

  Confusion and lack of preparation were everywhere. A May 15 letter from the General Commission on Jewish Questions to the Attorney General tautologically clarified, “If a Jew has not declared equipment for use in the practice of his profession, this constitutes a non-declaration of goods.” Another explanatory letter from the GCJQ, on June 24, 1941, confirmed to an inquiring government official that the census laws simply did not apply to certain colonial territories. On July 9, the GCJQ informed the President of the Council that it extended the June 14 census deadline “because the prefectures have not had time to prepare.” Ads were placed in the newspapers publishing the new deadline. That same day, July 9, a census staffer jotted a notation that the census forms had been received, and they “will be done as soon as the Department of the Interior has the necessary personnel.”113

  A December 1, 1941, letter from a provincial administrator to the Prefect of Police explained that only Jews who had a residence in France were required to register. That left out the many dislocated Jews residing in hotels or living with another family, including many refugees. One survey of French provinces returning census reports itemized a long list of data missing from their forms: many left out gender, profession, nationality, and in many cases, all three. Many lists bore numerous typos and overtyped sections, so some names and addresses were illegible.114

  When the Interior Ministry dispatched its long list of instructions to all prefects in France, it included a caution: “Since the number of these cards is limited due to their high price, please let me know, after creating the cards, the categories in which you have extra cards. This way I can distribute them to other Prefectures who have a shortage, as they will also be letting me know which cards they are short on.”115

  As France began its census activities, none of it was being recorded on punch card for automated retrieval. The operating budget reports for the GCJQ listed every expenditure for rentals, ribbons, and repairs on seven Remington and Underwood type writers, but not a franc on punch card services.116

  Even as French prefects and the GCJQ struggled to keep up with the waves of registrations, the Germans besieged their offices with unending requests for specific data that stretched their manual capabilities to the maximum. The experience of a single prefect was typical. On January 29, 1942, a German official asked the Prefect of Eure for the personal declarations made by district Jews, as well as inventories of their Jewish goods and enterprises. That same day, he also requested information from the provisional administrators of Jewish housing. On March 14 and again on April 2, the Prefect of Eure furnished the Police for Jewish Concerns with selected demographic information. On June 5, 1942, the Eure Prefect sent the Police for Jewish Concerns various tables listing Jews by age, gender, nationality, and profession—all in duplicate. On July 3, 1942, the Police for Jewish Concerns needed the number of Jews between ages 16 and 45 wearing the yellow star, all Jews between 16 and 45 required to wear the yellow star even though married to an Aryan, Jews older than 45 wearing the yellow star, and those Jews who because of their nationality did not wear the yellow star at all. All names were to be typed by gender listing all names and addresses where available.117

  In a move to centralize all the information into a single, easily accessible catalog, 100,000 forms were sent to a special police unit headquartered in the Hotel de Russie in Vichy. In December 1941, the Police for Jewish Concerns could finally visualize the massive quantity of census and registration forms to be sorted. The police bureau told CGJQ director Xavier Vallat that the task was impossible. On December 19, 1941, Vallat acknowledged the crisis. “The file would normally be established by your police service,” asserted Vallat, “but it has become apparent from several conversations with your offices that it has neither the personnel required nor the equipment necessary for this operation. Due to this situation we have thought to ask [for] help.”118

  Vallat unexpectedly found the help he needed. Rene Carmille, comptroller general of the French Army, had for years been an ardent advocate of punch cards. More than that, he had machines in good working order at his government’s Demographic Service. Carmille came forward and offered to end the census chaos. He promised that his tabulators could deliver the Jews of France.119

  * * *

  WHEN GERMAN requisition teams began pilfering Hollerith machines in August 1940, Rene Carmille, a mysterious French military technocrat, rescued his tabulators. Under cover of night, he moved the precious devices from his military finance office to a hiding place in a garage. In so doing, the French military’s punch card capability was preserved. A few months later, on November 14, 1940, Carmille single-handedly created France’s Demographic Service in Vichy. It was just a few weeks after the first Jewish census in the Occupied Zone, and from the Nazi viewpoint, the new agency was wholly compatible with Berlin’s labor recruitment and racial agenda. The Vichy government entrusted his agency with the responsibility for all social statistics projects. It operated at least twenty offices on both sides of the zone boundary.120

  Once established, Carmille carefully began strengthening his small arsenal of tabulators. Not only did he work with Holleriths, but he also had access to a Bull and even a Powers machine. In 1941, Carmille signed a 36-million-francs contract with Bull for new machines, even though the chance of delivery from their new plant in Lyon was in great doubt. He also signed a multi-million-franc contract with CEC calling for a series of powerful machines to be delivered on a strict schedule; millions of francs in fines and penalties were to be paid by CEC if they could not maintain the delivery regimen.121

  When Carmille began the Demographic Service, he was not allied with the General Commission on Jewish Questions, the anti-Semitic Vichy agency coordinating Jewish registrations. But he was certainly hoping to attract the agency’s attention. In March 1941, he made a point of informing GCJQ Director Vallat that his statistics for Jewish students were incorrect.122

  Carmille had been working for months on a national Personal Identification Number, a number that would not only be sequential, but descriptive. The thirteen-digit PIN number would be a manual “bar code” of sorts describing an individual’s complete personal profile as well as professional skills in great detail. For example, one number would be assigned for metal workers, with a second modifying number for brass, and then a third
modifying number for curtain rods. Tabulators could then be set to whisk through millions of cards until it located French metal workers, specializing in brass with experience in curtain rods. Those metal workers could also be pinpointed in any district. The system mimicked a concurrent Reich codification system that assigned a descriptive bar code-like number to every product and component in Germany. Carmille’s number would ultimately evolve into France’s social security number.123

  In spring 1941, Carmille was readying the professional July 1941 census of all French citizens aged 14 to 65. Question 11 asked Jews to identify themselves not only by their professed religion but also by their grand-parents. The program was highly publicized. An article in the New York Times several weeks later reported, “Special registering machines are being used in this census. By a system of perforations—every citizen will receive a number composed of thirteen digits—information will be obtainable… county by county.” An interesting emphasis in the questionnaire was on agricultural skills. A second New York Times article on the census actually identified the program as tied to an agricultural drive.124 Any Vichy official or Nazi representative in France would surely approve of such a census. It would not only identify the workers, including farm laborers, but the Jews among them.

  In mid-June 1941, Vichy launched the second major census of Jews, this one in both zones. The question was how would the questionnaires be processed, through manual methods with paper files, or with the dynamic capabilities of Holleriths?

  LYON, JUNE 18, 1941

  To Xavier Vallat

  General Commissioner for Jewish Questions