In July, 2013, Will Shortz, the New York Times’ longtime puzzle editor, asked me to be his assistant. I had just graduated from college, and, to my mind, the invitation had little rationale. It arrived on the heels of minimal correspondence: two e-mails in which Shortz had accepted two of my puzzles, with minor revisions. I doubted his motives for hiring me as much as my qualifications for the job. Surely, there were many more prolific and talented crossword constructors who could have assisted him. The only thing that distinguished me, I thought, was my gender: I was a young woman, and this was a field rife with men. Shortz was clear that I would work with him between September and the following May, when his longtime summer assistant, Joel Fagliano, would graduate and take up the job full time. I assumed, in other words, that I was not only a pinch hitter but a diversity hire.
This account has since been humbled by the memory of my in-box, which holds a much more extensive correspondence between me and Shortz. I had selectively remembered the two acceptances but forgotten three rejections. I had also repressed a long chain of e-mails about a potential Bloomsday puzzle, commemorating the publication of Joyce’s “Ulysses,” which he vetoed. (“For solvers who aren’t familiar with “Ulysses”—which would be a large number of people—I don’t think this puzzle would be very satisfying.”) Apparently, I had been courting Shortz’s attention and approval for three years. And, once I got it, I negated it. My revisionist history protected my ego (only two e-mails, two acceptances), but it also exposed me to more ego-withering insecurity (Shortz barely knew me; I had no track record). If my gender hadn’t informed Shortz’s offer, it had deformed my self-image as a worker: my ambition was offset by suspicion, and my efforts were overshadowed by a large chip on my shoulder.
I’m not sure how I knew that most crossword constructors were men. I had never met a constructor, nor scrutinized puzzle bylines. If anything, the puzzles that I enjoyed most were those in New York magazine by Maura Jacobson, who, for thirty-one years, wrote weekly puzzles that were loaded with puns and consistently avoided the most common pitfalls of crossword construction, dull repetition and excruciating corniness. But despite Jacobson, and despite the fact that I had been making crosswords since I was a teen-ager, somehow I knew that to think of crossword constructors today—if one thinks of them at all—is to think of men. And not just any men, but a particular kind of American man: white, STEM-educated, and charmingly (or brazenly) undersocialized.
If my understanding of Shortz’s motives for hiring me was a paranoid misread—ungenerous to both of us—my premonitions about the demographics and ethos of puzzle-making were eventually confirmed. Although data on the gender of crossword constructors before the so-called Shortz Era is incomplete—many puzzles were published without a byline or under pseudonyms, for example, and the numbers largely fail to account for trans and nonbinary constructors—several efforts have been made to trace these trends. One constructor determined that under the two editors before Shortz, Will Weng (1969-77) and Eugene Maleska (1977-93), women constructed approximately thirty-five per cent of all published puzzles. Between 1993 and 2013, women accounted for only nineteen per cent. There is reason to think that the share of women crossword constructors over all, not just in the Times, also declined during this period. According to Antony Lewis, the developer of the crossword-construction software Crossword Compiler, women accounted for roughly forty-five per cent of the company’s U.S. purchases in 2000, but by 2020 that number had fallen to about thirty-five per cent.
Lewis’s numbers still seem to outpace those of women published in wide-circulation newspapers. Looking at the statistics together, one is tempted to use them as a basis for speculation. Are women downloading constructing software but just not using it? Are they using it, but not very well? Are they using it well but not being published? In other words, is the problem time, facility, or bias? In an e-mail to me, Lewis implied another theory. His data, he wrote, include “people like teachers making educational puzzles, who may have distinct demographics from the newspaper folk.” His suggestion, as I understood it, was that women may be downloading his software but using it only in feminized professions, or jobs that women are more prone to do, such as teaching, whereas men are using it in professions where they hold leadership positions, including publishing.
Although I didn’t know it when I started working for Shortz, the early life of the crossword was indelibly shaped by women. The founding editor of the New York Times crossword, who oversaw the section from 1942 to 1968, was Margaret Petherbridge Farrar. Many of the most significant contributions to crossword culture—the first crossword contest, the rules for the grid’s symmetry and design—were pioneered by women who, like Farrar, found in the puzzle an intellectual outlet and an escape from the doldrums of housework. For decades, crossword constructing was a pastime conflated with first-wave feminists and bored housewives. How did it become an industry so male-dominated that its gender gap is now tracked by men who are prominent in the field?
One set of potential explanations lies in the digitization of crossword construction, which began in the nineteen-nineties and flourished in the early two-thousands with programs like Crossword Compiler. Although the rise of this kind of software doesn’t fully explain the masculinization of crossword labor and culture, it does offer some clues. The connection between computing and the demographics of constructing was first brought to my attention at the American Crossword Puzzle Tournament in 2014, when the topic of the field’s gender trouble was broached by an unlikely messenger: the seventeen-year-old puzzle wunderkind David Steinberg.
In a presentation he delivered based on data collected for his high-school science-research course, Steinberg proposed that as puzzle-making became increasingly informed by computer programming it began to replicate the imbalance of tech culture. Digitally assisted crossword-making, he argued, had changed the way the pursuit was perceived, altering its image from one associated with—or at least neutral toward—women to one unwelcoming of them. He invoked “numerous studies” suggesting that “females are far less likely than males to enter computer and other technology-related fields,” and hypothesized that as the practice of puzzle-making moved from graph paper and dictionaries to word databases and algorithms, it became less of a “literary exercise” and more of a “mathematical” one. Women, he posited, may have been “left behind” as a result.
Still constructing puzzles by hand, a woman alienated by the field’s digital turn, I considered his findings troublingly reductive, even as I confirmed them. One doesn’t need to know how to code to use constructing software, but, by 2014, the sensibility of crossword culture seemed to have merged with that of Silicon Valley. Aimee Lucido and Zoe Wheeler, two young women constructors who published their first puzzles in the Times around when I did, went on to work at Facebook and complete a master’s degree in computer science, respectively.
The history of computer programming itself is a useful paradigm for understanding the lack of women making puzzles today. It, too, is an industry in which women originally flourished but which, in the second half of the twentieth century, became a field led by men, its labor coded as masculine. During the Second World War, the realm of programming was regarded as meticulous and repetitive, and therefore ideal for women workers. Programming was considered clerical work, an extension of the secretarial pool, itself an extension of women’s work in the home—straightening, organizing, and managing the operations of the family unit. As a result, many of the earliest computer specialists were women—employees of corporations or wartime government agencies whose job was to crunch numbers by hand and to input data into the earliest electronic computer prototypes.
These women, who were referred to as “computers,” have resurfaced in popular culture with the success of the 2016 film “Hidden Figures,” which depicted the largely Black team of women whose work was essential to NASA’s operations during the space race. Although they are much less widely known than the women who took on manual jobs during wartime (crystallized in the figure of Rosie the Riveter), women with college degrees—especially degrees in math and engineering—were also newly encouraged to participate in the workforce. The government explicitly enlisted these highly educated women, who might otherwise have sought jobs as teachers, perhaps working only until they were married, to work with early computers, performing calculations underpinning ballistics operations and decrypting Japanese and German codes. They became experts at using the earliest general-purpose electronic computer, the ENIAC, encoding information so it could be legible to the machine. Women such as Adele Goldstine, Grace Hopper, Kay McNulty, and Betty Snyder developed early computing languages and debugging methods, becoming some of the nation’s earliest and most adept programmers, underrated as that label (and their work) was at the time.
Although crossword constructors were not explicitly sex-typed by employers, as computer programmers were, the requirements for being a good programmer aligned closely with those of crossword enthusiasts. Or at least the U.S. government thought so. As Liza Mundy wrote in “Code Girls,” her comprehensive history of female code breakers during the Second World War, women graduating from élite private colleges were asked two questions before being recruited for computer-programming jobs: (1) Do you like crossword puzzles? and (2) Are you engaged? In the middle of the twentieth century, before the advent of the computer-nerd and tech-bro stereotypes, crosswords and computing were not seen as markers of intelligence per se. But they were evidence of a woman’s shrewdness, work ethic, and fastidiousness—all highly compatible with the mid-century workforce, so long as she wasn’t yet committed to building a home.
Computer programming began to be perceived as a job suited for men only after technical innovations allowed its grunt work (encoding, debugging) to be done by the computer itself. That women eventually stopped being as involved, or involved at high levels, in the field isn’t because they stopped being good at the job once more aspects of it were computerized, nor because the job required a higher or different kind of intelligence. It is, rather, because the work of the computer began to overlap neatly with the kind of work that women had been encouraged to do. Once computers took on that support work, women either had to fight for jobs that had historically gone to men, or were once more relegated to the narrow set of opportunities that had previously been available to them.
This white-collar warfare between woman and machine is acutely dramatized in “Desk Set,” a 1957 romantic comedy starring Katharine Hepburn and Spencer Tracy. In the film, Hepburn plays Miss Watson, the head of the Reference Department of the fictional Federal Broadcasting Company, while Tracy plays Richard Sumner, the inventor of an electronic computer known as the EMERAC (or Emmy). Watson’s job is to be on hand for television broadcast announcers, answering questions about topics as various as which poisons leave no trace and which baseball player has the highest lifetime batting average. (Apparently, she knows facts like these from memory—like Alexa or Siri, she’s a data repository personified.) An efficiency expert, Sumner designs Emmy to answer any question about as fast as Miss Watson can; the all-women staff of her department worry that, as a result, they’ll be replaced. By the end of the film, the twinned threats of Sumner’s invention and Watson’s intelligence have been efficiently neutralized, and Miss Watson is no longer a “miss” at all: the couple get engaged; they embrace; and the mainframe computer spells out the words “THE END.”
But is that really the end? We imagine that Watson has left her job to become a wife, but what is she to do with all that free-floating knowledge—the origin of every Bible quotation, the words to every Longfellow poem—that she’s flaunted throughout the film? Maybe she’ll start writing crossword puzzles.
There is no reason to believe that men are better crossword constructors or solvers than women. Crossword culture has, however, been shaped by technological, economic, and social forces that have led women to feel unwelcome, underutilized, and underestimated in its already niche community. If Mrs. Sumner (née Watson) had turned to puzzle-making, in other words, the very forces that would have displaced her from her first job might have edged her out of constructing, too.
But it is not just the née Watsons among us who have been affected: the masculinization of puzzle labor leaves its trace all over the contents of today’s crosswords, not just their conditions of production. Editors in wide-circulation newspapers want their puzzles to be solvable by a large portion of their readership. This policy of inclusion—which has some editors imagining a Platonic crossword solver, unmarked by gender or race—can lead to a tacit policy of exclusion. All crossword editing involves judgment calls about language and its use that could hardly be called politically neutral, and, ultimately, the question of a word or phrase’s “puzzle-worthiness” is a negotiation between the constructor’s voice and that of a publication’s house style—a style usually set to the sensibility of its puzzle section’s editor. A list of terms rejected by the white, male, longtime puzzle editors of major newspapers in the past ten years, compiled by myself and by fellow-constructors, includes:
Looking at entries like these together, we begin to see a pattern: they refer to women, woman-produced culture, and phenomena and objects with origins in nonwhite cultures.
The question of which entries—and whose culture—can be deemed “mainstream” or “relevant” has high stakes for the fate of this lively art. It is intuitively but not necessarily true that a woman constructor will include in her grid references to “women’s culture” and “women’s language,” imprecise as those concepts may be. The constructor Elizabeth Gorski, for example, has added a staggering fifteen hundred and fourteen unique terms to the Times crossword lexicon, including ANNE SEXTON, CHICK FLICK, TUBE TOPS, and STATESWOMEN. Soleil Saint-Cyr and Portia Lundie, who are the first and second Black women Shortz knows of who have published crosswords in the print edition of the Times, introduced five new phrases in their 2021 débuts, including LET YOUR HAIR DOWN and ESSENTIAL WORKER.
Since 2014, the puzzle industry has changed dramatically, in part because members of the community are more willing to discuss the effects of implicit bias. In 2020, I was part of a group of constructors and avid solvers who wrote an open letter encouraging the Times to diversify its editorial staff. Since then, it has made several new hires and changed its submission process to be more transparent. New outlets have also proliferated, including a host of indie crossword sites, some of which explicitly publish puzzles by women for women. The incessant gender-tracking among crossword constructors publishing in these and other outlets has been accompanied by diversity initiatives meant to mentor women constructors and foster their talents. The result of these various efforts may be to rediscover the woman in the crossword puzzle—the housewife, the feminist, the computer. She was, after all, always there. ♦
This is drawn from “The Riddles of the Sphinx: Inheriting the Feminist History of the Crossword Puzzle.”
