The morning after a 2005 videotape of Donald Trump remarking to Access Hollywood host Billy Bush that fame gave him license to “grab [women] by the pussy,” the shame-resistant Republican presidential nominee was finally suffering consequences for his outrageous statements. But for the media organizations covering his comments, this was a big moment on another level.
After what one New York Times editor described as “extensive discussion,” the Times broke one of its cardinal style rules: The words fuck, pussy, bitch, and tits appeared in print on the front page.
This had never happened before.
According to the most recent New York Times Manual of Style and Usage, “the Times very rarely publishes obscene words, and it maintains a steep threshold for vulgar ones.” (“Very rarely” is itself a relaxation of standards; previous editions of the manual said the Times “virtually never“ publishes obscenities.) The Times’ main argument for treating such words as unprintable is the need to maintain standards of civility. Yet article after article will still describe uncensored acts of war, torture, child abuse, and other uncivil behavior.
But the boundaries between polite society and popular culture had been blurring long before Donald Trump came on the scene. These changes in public discourse have made profanity more prevalent and therefore less shocking—which only makes the Times’ position seem increasingly outdated. In fact, profanity has become so commonplace that the Times has had to find ways to report obscenities without actually using the obscenities themselves. Even when the subject of the story is a profanely named product, the Times will coyly write around it: The best-selling picture book Go the Fuck to Sleep, for instance, was the subject of three Times articles that failed to mention the title.
After reading all three, I wrote a blog post noting the Times’ timidity. Every so often I noticed another writer, peeved by a missing expletive, would commit several hours and hundreds of words to making arguments about hypocrisy, changing social mores, and the porous media ecosystem—and nothing would change. I didn’t expect my post would make a difference, either.
But writing an article is not the only way to make a point. This was in 2011, when data journalism was on the rise. Nate Silver had licensed his website FiveThirtyEight to the Times, and Quartz would launch the following year. Every time I noticed the Times avoiding an obscenity, I began to document the instance on Twitter using the hashtag #fittoprint and invited others to join me. Eventually, I thought, I would have a long list with a large and surprising number of examples on it, all public and verifiable.
In the first 100 examples of Timesian expletive avoidance that were collected, I started to notice patterns. For example, articles about HBO shows kept showing up. (What is HBO, after all, if not foul-mouthed entertainment for Times-reading sophisticates?) I also began to recognize signals that an expletive was being written around: Words like expletive, profanity, four-letter, and unprintable were frequent indicators, and less common ones like saltier.
This knowledge paved the way for a more systematic approach to data collection. In early 2014, I moved the Fit to Print project to Tumblr, since it allowed more than 115 characters (after a link and the hashtag) for context. I also registered for a Times API key, which let me search all new articles for indicator words.
Around lunchtime, when other people would be looking for new cat GIFs, I’d check the Times Search API instead to see whether any expletives were being avoided that day. Saltier returned false positives from the food section—but not as often as you’d expect.
Over 2014 and 2015, I collected 663 examples of articles that avoided expletives. (If an article avoids multiple expletives, I counted it only once.) Some friends who knew about the Fit to Print project shared links with and I stumbled across others while reading Times stories on social media, but most of these examples came through the Times Search API; these search results had become my main way into the Times.
I would not claim that this data set is complete; evidence of avoidance can be almost as hard to find as evidence of absence. Nevertheless, rudimentary data analysis yields some interesting insights.
Unsurprisingly, the big day for expletive-avoidant articles, with 36% of all examples, is Sunday. The Sunday paper is the fattest and includes the most articles. It also has sections dedicated to topics where expletives are most often avoided, such as the arts and sports, plus the magazine, which avoided expletives in 43 articles in 2015—almost once per issue.
The “arts” section in the chart above includes music and television. If you count books, movies, and theater as well, there were 200 instances of expletive avoidances, or 30% of all articles I found. (And the number of arts-related stories with expletive avoidance is surely higher than I found.)
Of course the frequent use of expletives is not limited to cultural activities: Expletive avoidance runs throughout the paper, from business to international affairs to domestic politics. Donald Trump’s name comes up in no fewer than nine articles, as either curser or cursed. (If nine sounds low, remember that this data set ends on December 31, 2015.)
So when the Times avoids an expletive, what euphemism does it use to replace it?
Expletive is the most common euphemism, used 40% of the time. 39% of those instances—107 of 273—have the word in brackets inserted into a quotation, e.g. “I mean, this [expletive] hurt.” (The Times used these bracketed insertions 60% more often in 2015.) The remaining 61% are paraphrases: “‘It was just,’ he added, using an expletive, ‘bad judgment.’”; ”He used an expletive in referring to Keselowski, who took third.”
After expletive, the most common euphemism is unprintable. I find this term particularly irksome since it suggests physical impossibility, that their typesetting software can’t produce that sequence of glyphs. As the Trump tape revealed, the Times is perfectly capable of printing the word fuck. Another instance of its printing was in an ambitious 2013 series on a homeless Brooklyn girl. Margaret Sullivan, who in her tenure as Times public editor kept raising questions about expletive avoidance, looked into why fuck stayed in that 2013 story, and was told that editors “had a very thorough discussion” and decided to make an exception.
The “other” category includes faux-folksy formulations such as “a word more pungent than ‘slop,’” and “a stronger version of the phrase ‘gol darn,’” as well as the straightforward, ”He swore.” When I began the Fit to Print project, I could enjoy the cleverness of some of these contortions. But after reading through hundreds of examples over several years, expletive avoidance no longer strikes me as an interesting puzzle for a writer to solve. The policy just seems prissy, arbitrary, and delusional. All this information is on the internet and, increasingly, the Times links out to the words it will not print.
It is also inconsistently applied. Two weeks before the Trump tape surfaced, the widow of Keith Scott, an African-American man who was shot by police in North Carolina, gave the Times a video showing his final moments. The Times printed a verbatim transcript, but then removed the expletives from the accompanying article. And a few days after describing the tape, the Times ran Jennifer Senior’s review of a wave of books that are critical of Hillary Clinton and claim she “that Mrs. Clinton is a potty mouth” without giving any concrete examples.
At times, I identify with the defeatist tone of New York Times Magazine contributor Dan Brooks, who wrote, ”It was too crude to reproduce here, and anyway, I’ve learned not to try.” But reporters keep trying. Some (and I’ve heard from members of the newsroom who oppose this policy) may practice expletive avoidance as a form of protest. In other cases, it may simply be that, on a given beat, expletives are not avoidable. The reigning king of expletive avoidance is pop music writer Jon Caramanica, with 16 articles (one with a shared byline) over the course of two years, followed by book reviewer Dwight Garner (11) and reporter Matt Flegenheimer (11, two of them shared).
Between the limitations of my data collection and my statistical acumen, there are limits to what conclusions can be drawn from the Fit to Print data set at this point. I haven’t compiled a record of the expletives that were avoided, but many of those can be tracked down online. There isn’t a control set of articles that don’t avoid expletives that it can be compared with, and there’s no way to know when expletives have been avoided altogether by not being acknowledged at all. You can also quibble with my approach to classification or with what I decided to include or exclude.
In fact, I invite you to do so. Take a look at the spreadsheet. Maybe one of you can take the list of 663 sentences and train a bot that generates Timesian, expletive-avoidant sentences, or can search the Times archives for them better than my primitive API calls.
Or maybe this amassing of information can help put an end to the thorough extensive discussions that take up Times editors’ valuable time and get between readers and the information they seek. Remember: If readers can find it elsewhere, they will.