“Why Wasn’t It ‘Grapes of Glee’?” asked The New York Times last week, practically demanding my attention. The article was on a study using big-data techniques to document the correlation between the so-called economic-misery index and what the researchers called the literary-misery index. According to the report, published in Plos One, it takes about 11 years for words in the economic-misery index to surface in books. Authors absorb words from the culture slowly: “We do have a collective memory that conditions the way we write,” said one of the authors, “and … economics is a very important driver of that.”
The Times reporter took this information to two of America’s reigning contemporary literary figures, Chang-Rae Lee and Jane Smiley, who seemed bemused but not about to argue with hard data. Your correspondent, however was a bit more skeptical.
I wondered what the Times, and by extension the authors of the study, meant by the term literary. They had relied on the Google N-gram Viewer, which, as we all know, searches lots of books.
If something is written in a book, by one definition, it’s literature. Lee and Smiley are novelists. Most contemporary readers, asked about literature, will refer to fiction, or perhaps to poetry.
But plenty of published books are not self-consciously literary in the way that fiction and poetry often are, and one might think that, even with a lag effect, the correlation between economics and idiom might be less robust in such self-consciously literary genres.
Of course, there’s the question of the market as well—we’re not talking here just about what language authors are using, but what books publishers are choosing to produce and distribute to the book-buying public.
I perused the original article. Now, I am close to illiterate when it comes to social-science equations. Give me a formula like
and I skip ahead to the words I can read. But I did read the article through to the section on methods, where I found this bit of information from the authors: “We considered English-language books, in four distinct corpora (all books in English, fiction books in English, American English books, and British English books).”
Wait a moment. The Times hadn’t said anything about those corpora. It had referred simply to literature and then gone quoting Lee and Smiley. I backed up to the tables proving the study’s results and found this:
For purposes of this post, we’ll leave aside my discovery here of the delightful new word hedonometer. Let’s focus, instead, on data showing that the highest degree of correlation lies between literary misery expressed in English books as a whole and economic misery from 11 years earlier. But in the notes, we find that one of the scores omitted, for lack of statistically significant correlation with economic misery, was the literary misery in fiction books.
In other words, the correlation among books in general is so strong that it overcomes what is essentially a noncorrelation in books by people like Chang-Rae Lee and Jane Smiley.
Well, ta-da. What I think of as literature, and perhaps you think of as literature, contains language that does not have all that direct a correlation with economic ups and downs. It does change over time, certainly—Lee and Smiley and just about any other writer will acknowledge such change during their career trajectories. Moreover, however self-conscious we may be about our use of language, we live in the same ever-changing stream as everyone else and are bound to feel its idioms.
But if there is a basic qualitative difference between the way language is used in Lee’s On Such A Full Sea and the way it’s used in Sheryl Sandberg’s Lean In, then Plos One’s “Books Average Previous Decade of Economic Misery” may, inadvertently, suggest one way of getting a handle on that difference.