The Age of Female Computers David Skinner

2 2

oday, mathematics and com- problems had to be broken up into puter science often appear discrete, simple parts and done by Tas the province of geniuses hand. Where scads of numbers need- working at the very edge of human ed computing—for astronomical ability and imagination. Even as purposes at the Royal Observatory in American high schools struggle to Greenwich, England, or to establish employ qualified math and science the metric system at the Bureau du teachers, American popular culture Cadastre in Paris—such work was has embraced math, science, and com- accomplished factory-style. In his puters as a mystic realm of extraordi- book When Computers Were Human, nary intellectual power, even verging a history of the pre-machine era in on madness. Movies like A Beautiful computing, David Alan Grier quotes Mind, Good Will Hunting, and Pi all Charles Dickens’s Hard Times to cap- present human intelligence in the ture the atmosphere of such work- esoteric symbolism places: “a stern of long, indecipher- When Computers Were Human room with a deadly able, but visually by David Alan Grier statistical clock in Princeton ~ 2005 ~ 424 pp. captivating equa- it, which measured $35 (cloth) tions. One has to every second with think of such prosaic activities as a beat like a rap upon a coffin-lid.” paying the mortgage and grocery The most famous modern exam- shopping to be reminded of the quiet ple of such work is probably Los and non-revelatory quality of rudi- Alamos, where scientists’ wives were mentary arithmetic. Which is not recruited in the early stages to com- to put such labor down. Adding the pute long math problems for the price of milk and eggs in one’s head Manhattan Project. is also brain work, and we should The social history of pre-machine never forget the central place of computing is also interesting in light mere calculation in the development of contemporary debates about gen- of more sophisticated areas of human der and scientific achievement, and knowledge. here Grier’s reconsideration of the Long before the dawn of calcula- past sheds useful light on the pres- tors and inexpensive desktop com- ent. Resigned Harvard president puters, the grinding work of large Lawrence Summers became an aca-

96 ~ THE NEW ATLANTIS

Copyright 2006. All rights reserved. See www.TheNewAtlantis.com for more information. THE AGE OF FEMALE COMPUTERS demic outcast after speculating that understand the desire to correct past there might be an “intrinsic” basis orthodoxies about the female mind for the unequal numbers of men and with new ones. But even as we right- women engaged in science and engi- ly decry a past when even the most neering at the university level. The talented women were prevented from idea that men and women are dif- pursuing math and science in the ferent creatures, with distinct drives most prestigious posts, we should and ways of thinking, is apparently remember—and honor—the crucial so radical that even to raise it leads to role of women in advancing mathe- the academic guillotine. And yet only matical and scientific knowledge one a few decades ago, it was assumed by detailed calculation at a time. even the most civilized societies that women were not fit for serious intel- t the beginning of the long line lectual pursuits, especially scientific Aof women who made their marks ones. The occasional female endowed as human computers was Nicole- with truly extraordinary talent occu- Reine Lepaute. Like many women pied the unfortunate position of the featured in Grier’s book, Lepaute George Eliot character who tells enjoyed a personal connection to her son: “You may try—but you can the intellectual world, allowing her never imagine what it is to have a to gain experience with scientific man’s force of genius in you, and yet matters in spite of conventions that to suffer the slavery of being a girl.” warned women away from science. Note that even this extraordinary She owed her education to the for- character, created by an intellectually bearance of understanding parents; accomplished, great female novelist, her freedom to pursue an intellectual refers to genius as something par- career to an obliging husband; and ticularly male. her professional position to Joseph- In the history of computing, the Jérôme de Lalande, her longtime humbler levels of scientific work were scientific collaborator. open, even welcoming, to women. In a book published in 1705, using Indeed, by the early twentieth cen- Isaac Newton’s new calculus, the tury computing was thought of as English gentleman-astronomer women’s work and computers were Edmond Halley identified and pre- assumed to be female. Respected dicted the return of the comet even- mathematicians would blithely tually named after him. But it was approximate the problem-solving the French mathematician Alexis- horsepower of computing machines Claude Clairaut, along with Lalande in “girl-years” and describe a unit of and Lepaute, who first computed the machine labor as equal to one “kilo- date of the comet’s perihelion with girl.” In this light, one can surely any precision in 1757, predicting

SPRING 2006 ~ 97

Copyright 2006. All rights reserved. See www.TheNewAtlantis.com for more information. DAVID SKINNER it would occur in the spring of the be done in parallel by different indi- following year. Sitting “at a com- viduals.” mon table in the Palais Luxembourg Mme. Lepaute was central to this using goose-quill pens and heavy effort, if largely unrewarded with linen paper,” writes Grier, the three professional position and prestige. friends slowly computed the course Lalande hired her as his assis- of Halley’s Comet along a parabola- tant when he became the editor of shaped orbit, reducing the math to an Connaissance des Temps, an astronomi- extraordinary series of baby steps. cal almanac, where together they Lalande and Lepaute focused on prepared tables predicting the posi- the orbits and gravitational pulls of tions of various celestial bodies. She Jupiter and Saturn (the three-body performed valuable but largely unap- problem), while Clairaut focused on preciated work. the comet’s orbit. “With the perspec- tive of modern astronomy,” Grier alf a century later and an ocean writes, “we know that Clairaut did Haway, Maria Mitchell would not account for the influences of play the next part of the willing Uranus and Neptune, two large plan- female computer supporting the bold ets that were unknown in 1757.” designs of male scientists. In the Still, the result of their number- 1840s, as American manufacturing crunching was a tenfold improve- swelled to claim some 25 percent of ment in accuracy over Halley’s pre- the economy and American pride vis- diction, if still not perfect. When the à-vis Europe launched a new era of comet reached its perihelion just a economic and political competition, a couple of days shy of the two-month movement took hold to establish an window in which Clairaut and col- American nautical almanac. Lacking leagues said it would, Clairaut’s com- such a publication, claimed one sup- puting method was ridiculed by one porter, “our absent ships could not of the great intellectuals of the day, find their way home nor those in our Jean d’Alembert, one of the editors ports grope to sea with any certainty of the Encyclopédie and himself an of finding their way back again.” The astronomer, who called the calcula- almanac’s chief mathematician was tions more “laborious than deep.” But Harvard professor Benjamin Pierce, this has not been the verdict of his- while the computing staff consisted tory. “Beyond the simple accuracy of of several students and amateurs. his result,” writes Grier, “Clairaut’s Mitchell was the only woman in the more important innovation was the group. The daughter of a banker and division of mathematical labor, the amateur astronomer, she was not recognition that a long computation some anonymous savant: her discov- could be split into pieces that could ery of a new comet in 1847 brought

98 ~ THE NEW ATLANTIS

Copyright 2006. All rights reserved. See www.TheNewAtlantis.com for more information. THE AGE OF FEMALE COMPUTERS her fame and a medal from the king social inquiry. Computing followed of Denmark. Mitchell herself felt the growth of the social sciences: the no need to announce her discovery, effort to move away from broad ideas mentioning it only to her father, who and conceptual investigations toward quickly checked to see if the comet empirically-based methodologies in had been claimed by anyone else pursuit of a scientific knowledge of and then insisted on publicizing her human affairs. accomplishment. Francis Galton looked to mathe- Mitchell proved an able computer, matics to help prove Darwin’s theory not out of place among the gentle- of natural selection. In one inves- men who filled this minute trade. tigation, he gathered crude data on She went on to become the first African women “endowed,” he wrote female professor of astronomy at to his brother, “with that shape which Vassar College, gaining some of the European milliners so vainly attempt recognition and opportunities that to imitate.” Returning to England, Lepaute never did. The tide was Galton worked for the “Committee indeed slowly turning in women’s for Conducting Statistical Inquiries favor, though far from decisively. In Into the Measurable Characteristics the two decades following the Civil of Plants and Animals,” where such War, Grier reports, women went efforts to support Darwinism came from holding one out of six hun- under the powerful influence of Karl dred office jobs to one in fifty. The Pearson, who introduced a break- Harvard Observatory in particular through formula for correlation. found women to be especially desir- Pearson was also an unusual char- able computers, since they accepted acter, a man of far-flung intellec- payment equal to half the going rate tual interests and progressive social for men. opinions. Grier, who passes up few opportunities to enliven his history, y the end of the nineteenth cen- describes Pearson’s Hampden Farm Btury, astronomy was no lon- House project, where women and ger driving the science of comput- men worked together in an egalitar- ing. New scientific interests—from ian atmosphere studying plants. On Darwinist anthropological investiga- Fridays, the workers would break tion to modern mathematical eco- for what were called “biometric nomics to war production—would teas,” while calculation and number- come to require and ultimately redi- crunching took place on weekends. rect the aims of computing. In this One of Pearson’s larger projects col- period, the discipline of statistics lated data on some 4,000 children and as we know it was born, reshap- parents in an attempt to demonstrate ing the character of all kinds of that “moral qualities” of character

SPRING 2006 ~ 99

Copyright 2006. All rights reserved. See www.TheNewAtlantis.com for more information. DAVID SKINNER and intelligence were hereditary. It American mathematicians was to be was a fine example of how rigorous found among those who had gone calculation in service to misguided through the discipline of the Proving theories is error masquerading as a Ground.” But Aberdeen was not the thousand facts—a problem that obvi- United States’s only wartime home ously has not gone away. to computers. Nearby in Washington at the Experimental Ballistics Office, orld War I shifted the focus Elizabeth Webb Wilson, the top Wof computing to two kinds mathematics student in her class of questions: military problems con- at George Washington University, cerning artillery trajectories and found employment with several other atmospheric drag, and economic women converting the raw data from problems concerning production, Aberdeen into tables usable at the as the United States strove to out- warfront. After the war, she looked fit, feed, and arm the American in vain for another computing job, Expeditionary Force. England’s eventually becoming a high school Ministry of Munitions relied heavily mathematics teacher instead. on Pearson’s Biometrics Laboratory Wilson’s story confronts us with for help calculating ballistics for anti- a paradox of social progress. In a aircraft munitions. In the United post-feminist world, a distinguished States, such work was handled at young talent like Wilson would eas- the Aberdeen Proving Ground in ily find employment working with Maryland. The main task on both numbers. Meanwhile, high schools sides of the Atlantic was revising go begging for anyone of Wilson’s Francesco Siacci’s theory of ballis- ability—male or female—to teach tics trajectory, which worked well mathematics. That the old system enough for the artillery of the nine- was unjust is indisputable; that the teenth century but needed significant new system is better at raising up the revision in the age of aerial warfare. next generation of mathematicians is Human computers struggled to cal- a complicated question. culate trajectories and end points for Economists also used computing aerial bombs, anti-aircraft artillery, to track domestic productivity. The and the weaponry of aerial combat. punched-card tabulator, which the The Aberdeen Proving Ground Census Bureau first used for the was the Manhattan Project of its 1890 census, became an increasingly day. “For many years after the First important tool for tracking retail World War,” said the mathemati- pricing data mailed in to the Food cian Norbert Wiener, voicing con- Administration by thousands of cor- ventional wisdom on this point, “the respondents scattered nationwide. overwhelming majority of significant Washington was not exactly convert-

100 ~ THE NEW ATLANTIS

Copyright 2006. All rights reserved. See www.TheNewAtlantis.com for more information. THE AGE OF FEMALE COMPUTERS ed overnight to such numerical rep- series of unreliable contraptions to resentations of American economic more sophisticated adding machines life. As Grier describes: “The notion and cash registers. Both science and that the sprawling agricultural econ- business came to rely on the rapid omy could be described with dif- numerical calculations that machines ferential equations or probed with alone could efficiently produce. statistics calculations was not widely The last hurrah for pre-machine accepted in 1917-18.” The work of computing was a product of the Harry C. Wallace (an editor and New Deal, the Mathematical Tables future Secretary of Agriculture) and Project. Its task was to produce his son Henry A. Wallace (a writ- mathematical tables for use, “not only er and future Vice President) dur- by mathematicians and astronomers, ing this period foreshadowed the but also by surveyors, engineers, future use of statistics to calculate chemists, physicists, biometricians, everything from consumer confi- statisticians, etc.” The Work Projects dence to inflation to the produc- Administration (W.P.A.) required tivity of American manufacturing. this large-scale computing operation Using carefully crunched numbers, to use labor-intensive methods and the Wallaces tried to convince U.S. limit the number of female hires to Food Administrator Herbert Hoover 20 percent of staff. Few of the human to guarantee a price for corn in order computers they hired had completed to shore up the related price of swine, high school. As one of the early but to no avail. Hoover feared med- computers recounted, “arrested TB dling in the private sector—a stance cases, epileptics, malnourished per- that has of course become harder for sons abounded.” American leaders to maintain, in part Arnold Lowan, a physicist who had because mathematical models of the fled anti-Semitic pogroms in Europe economy have grown increasingly but could not find a regular teaching sophisticated and thus ever more position in the United States, was the inviting to political intervention. director of the Mathematical Tables Project. His first lieutenant was y the early twentieth century, the Gertrude Blanch, another Eastern Bmachine was catching up with European immigrant who could not the human computer, as suggested find academic employment, despite by the presence of those punched- a doctorate in mathematics. Blanch card tabulators in Washington. proved to be a true leader. While While computing went from merely the regulations of the W.P.A. seemed supporting astronomy to an essen- well-designed for a make-work proj- tial tool of social science, the tech- ect of endless mediocrity, she and nology of computing went from a Lowan worked overtime to check

SPRING 2006 ~ 101

Copyright 2006. All rights reserved. See www.TheNewAtlantis.com for more information. DAVID SKINNER calculations and ensure high-quality promptly hung a “women only” sign products free of errors. Blanch even on the door of their lab. organized a lunch-hour math cur- At the time, there were almost riculum for willing workers that no researchers whose primary inter- took them from elementary arith- est was computing, still seen as a metic through high school algebra, mere handmaiden to other, more trigonometry, all the way to college substantial scientific interests. But calculus and, finally, matrix calcula- this was changing fast as machines tions, the theory of differences, and began to outperform human comput- special functions. It was the most ers. Up until World War II, human successful mathematical tables proj- computers had the advantage. As ect in history. Grier writes: “A punched-card tabu- The arrival of World War II sound- lator could work much faster than a ed the death knell for work-relief human being, but this advantage was projects, but the Mathematics Tables lost if the operator had to spend days Project was certified as an urgent war- preparing the machine.” Richard time program, granting it a reprieve Feynman, then a junior staff member and a degree of respect Lowan had at Los Alamos, arranged a showdown otherwise sought in vain. Grier notes between man and machine, pitting a an interesting moment of contact group of human computers against between Lowan and John Brainerd the Los Alamos IBM facility with at the University of Pennsylvania, both performing a calculation for where a team was struggling to build the plutonium bomb. For two days, what would become ENIAC, an elec- the human computers were able to trical analyzer that was being devel- keep up with the machines. “But on oped to calculate ballistics for the the third day,” recalled one observer, Aberdeen Proving Ground. Brainerd “the punched-card machine opera- was looking for highly skilled human tion began to move decisively ahead, computers, but Lowan’s group was as the people performing the hand not what he had in mind. Lowan used computing could not sustain their machines to facilitate the work of initial fast pace, while the machines human computers; Brainerd wanted did not tire.” Shortly after the war, human computers to aid the work of the machines took over; their human his machine. Brainerd then met his accompanists were now “operators” own Nicole-Reine Lepaute figure, and “programmers.” Adele Goldstine, the wife of a ballistics officer who had done graduate hen Computers Were Human work in mathematics. Goldstine set Wtells an important story. up a classroom program to educate Interesting for its insights into sci- their own team of computers and ence and computing, Grier’s book

102 ~ THE NEW ATLANTIS

Copyright 2006. All rights reserved. See www.TheNewAtlantis.com for more information. THE AGE OF FEMALE COMPUTERS is also an impressive work of eco- minds who could not stake claim to nomic and social history. With the better-paid academic positions was discovery of binary logic, the sim- an important boost to many serious plest parts of long problems became intellectual enterprises. That women both too voluminous and too simple of the capacities of Elizabeth Webb for human hands. Yet in slightly Wilson or Gertrude Blanch are now more complicated form, the simple much freer to pursue their own inter- number-crunching of long problems ests is an even greater boost to made ideal work for the attentive the sciences, though not without its and moderately educated, and it was costs. sometimes the only work available to well-educated women. That scien- David Skinner is an assistant manag- tists often had the benefit of highly ing editor at The Weekly Standard and talented and under-rewarded female the editor of Doublethink magazine.

SPRING 2006 ~ 103