Jonathan Franzen’s fifth novel, Purity, which published last week, came as no surprise to fans of the author’s elephantine narratives. “Magisterial sweep is now just what Franzen does,”Radhika Jones wrote in TIME’s review.
Close readers might also notice that Franzen goes back and back to the well of a few dozen signature words. His novels are peppered with motels, carpeting, dorm rooms, and the color beige–all seasoned with regular volleys of profanity.
But don’t take our word for it. To determine Franzen’s 50 favorite words, TIME Labs compared the frequency of each word in Franzen’s five novels to their frequency across all English language novels, using data that Google provides. (The exciting methodology is at the bottom.)
| RANK | WORD | USES | NOVELS | PER FRANZEN NOVEL | PER ALL NOVELS |
| 1 | f–k | 147 | 5 | 29.4 | 2.3 |
| 2 | so-called | 20 | 5 | 4.0 | 0.00015 |
| 3 | french | 68 | 5 | 13.6 | 0.23 |
| 4 | f–king | 111 | 5 | 22.2 | 2.5 |
| 5 | long-term | 13 | 4 | 3.3 | 0.00013 |
| 6 | d–k | 54 | 5 | 10.8 | 0.63 |
| 7 | s–t | 118 | 5 | 23.6 | 3.8 |
| 8 | f–ked | 45 | 4 | 11.3 | 0.67 |
| 9 | a–hole | 42 | 5 | 8.4 | 0.53 |
| 10 | tad | 39 | 3 | 13.0 | 0.26 |
| 11 | carpeting | 38 | 4 | 9.5 | 0.27 |
| 12 | bulls–t | 46 | 4 | 11.5 | 0.65 |
| 13 | underpants | 28 | 5 | 5.6 | 0.17 |
| 14 | marxist | 14 | 5 | 2.8 | 0.023 |
| 15 | boyfriend | 110 | 5 | 22.0 | 2.4 |
| 16 | merger | 65 | 4 | 16.3 | 3 |
| 17 | sofa | 132 | 5 | 26.4 | 2.8 |
| 18 | motel | 43 | 5 | 8.6 | 0.94 |
| 19 | whoa | 19 | 3 | 6.3 | 0.077 |
| 20 | knapsack | 27 | 3 | 9.0 | 0.31 |
| 21 | dorm | 32 | 5 | 6.4 | 0.56 |
| 22 | midwestern | 23 | 4 | 5.8 | 0.2 |
| 23 | goddamned | 19 | 4 | 4.8 | 0.2 |
| 24 | pajamas | 35 | 5 | 7.0 | 0.49 |
| 25 | cruiser | 28 | 4 | 7.0 | 0.8 |
| 26 | stadium | 42 | 3 | 14.0 | 1.3 |
| 27 | blah | 21 | 5 | 4.2 | 0.38 |
| 28 | niceness | 15 | 5 | 3.0 | 0.061 |
| 29 | nightstand | 23 | 5 | 4.6 | 0.26 |
| 30 | grownup | 17 | 4 | 4.3 | 0.08 |
| 31 | carton | 28 | 4 | 7.0 | 0.48 |
| 32 | driveway | 59 | 5 | 11.8 | 1.8 |
| 33 | daddy | 43 | 4 | 10.8 | 1.3 |
| 34 | dude | 25 | 5 | 5.0 | 0.57 |
| 35 | wow | 27 | 4 | 6.8 | 0.34 |
| 36 | wolf | 93 | 4 | 23.3 | 2.7 |
| 37 | bunny | 20 | 3 | 6.7 | 0.25 |
| 38 | beige | 28 | 5 | 5.6 | 0.44 |
| 39 | Coke | 24 | 5 | 4.8 | 0.84 |
| 40 | girlfriend | 95 | 4 | 23.8 | 2.6 |
| 41 | dialed | 27 | 5 | 5.4 | 0.67 |
| 42 | smelled | 81 | 5 | 16.2 | 1.9 |
| 43 | raincoat | 18 | 4 | 4.5 | 0.21 |
| 44 | martini | 16 | 3 | 5.3 | 0.24 |
| 45 | gonna | 50 | 5 | 10.0 | 3.7 |
| 46 | cartons | 20 | 5 | 4.0 | 0.31 |
| 47 | bedspread | 16 | 4 | 4.0 | 0.16 |
| 48 | two-dimensional | 4 | 3 | 1.3 | 0.0001 |
| 49 | turds | 10 | 3 | 3.3 | 0.035 |
| 50 | jeans | 74 | 5 | 14.8 | 2.7 |
Methodology
Words are ranked by a score known as “term frequency-inverse document frequency,” a well-established method of determining which words in a text are particularly important.
Our algorithm considered only lowercase words that appear in at least three of Franzen’s five novels. The only alteration to the original texts that TIME made was to convert uppercase words at the beginning of sentences to lowercase if they appeared elsewhere in the novels as a lowercase word. (This method, while imperfect, tends to sort out proper nouns while not undercounting words that commonly start sentences.)
For a baseline comparison, we used Google’s English Fiction files, using the 2007 levels of each word that Franzen uses.