The Dictionary as Data

Looking up a word in the dictionary is an intimate act. Usually, no one sees us do it. We choose the moment, the word, and the reference work ourselves. It is an intellectually active moment with a physical expression. We reach for the book, we click to the site, we open the app. Our minds are seldom as ready to receive new information as when we look up a word in a dictionary. The moment may be viewed as reflective of a faith both in the possibility of self-improvement and in the authority of the dictionary: two sides of the same intellectual coin that constitute a kind of social contract that binds a dictionary to its user.

If no one sees us look up a word, then similarly we don’t see how and when others do so. Indeed, unlike members of a knitting circle, or music lovers who gather at a concert venue, or Red Sox fans who can be identified by the cap they wear, dictionary users would seem to be participating in what amounts to the opposite of a collective experience.

Or are they?

Anecdotally, I’ve heard evidence to the contrary for years. Many people confide in me, upon learning that I work as a lexicographer, that they consider themselves unusual because they read the dictionary. The tone might be a conspiratorial whisper or the loud pride of the marcher to a different drummer. I hear “My family thinks I’m crazy because I read the dictionary” as often as I hear “In my family we were always reading the dictionary.” I have encountered this kind of thing often enough to make me conclude that, in fact, many people do this; we just don’t always know who they are. Most people who “read” the dictionary seem to think they are unique. It turns out that they aren’t.

But when they look up particular words, which words are they looking up? The privacy of the act has meant that, for nearly all of the history of published dictionaries, only the users have known. Lexicographers and publishers could never have known whether their labors on any given word were read often—or never. This might make for a grim perspective on one’s life’s work (“harmless drudge,” indeed), but it is obviously understood by all dictionary makers that in order for a dictionary to be generally useful, it must contain all the specific information about words that is likely to be needed. This is the true pact between the user and the dictionary: whenever you have questions, here are answers.

We know of some instances of words that were looked up in a dictionary precisely because the answers were not there. Dr. Johnson was supposedly congratulated by a lady for having omitted the “naughty words” from his dictionary, to which he is said to have replied, “Madam, I find that you have been looking them up.”¹ During the early work on the Oxford English Dictionary, editor James Murray encountered an obscure medical term and wondered whether to include it. He posed the question to the Oxford medical faculty, who assured the great lexicographer that the term in question had no business in a general dictionary, as it was used only by specialists. A few years later, upon the death of the queen, her successor’s coronation was delayed by a sudden and serious illness. The future Edward VII’s health became the national obsession, but the name for his condition turned out to be the very word that had been omitted from the dictionary: appendicitis.²

Such unsatisfied curiosity about a word makes for a vivid predigital example of how dictionaries are used: when we reach for them, and why. But once dictionaries went online—once dictionaries became data—a new story could be told.

The Merriam-Webster Online Dictionary went live in 1996. (For context: eBay and Amazon began in 1995, Google was founded in 1996, and Dictionary.com went online in 1998.) From the very beginning, it was a digitized version of the Collegiate Dictionary. As the most successful American “desk...