Words and the way we use them offer a rare window on the popular zeitgeist. Kory Stamper and Peter Sokolowski know that better than most.

The Merriam-Webster lexicographers (and must-follow tweeters) are part of a team that edits the dictionary, which means they spend a lot of time poring over our collective textual output — books, blogs, even take-out menus — and studying word-use trends. Some, like the popular appendage of “super-” to pretty much any word, emerged gradually over decades. But some words cycle in and out of the popular consciousness much more quickly, a phenomenon that, thanks to the Internet, we can watch in almost-real time.

Merriam-webster.com is actually a gold mine of these little insights, if you know where to look. The site keeps a running tally of the most popular look-ups over the past day, week and four-month period, as well as a “trend watch” blog that explains spikes in certain words. Popular over the past month: “thug,” “vortex” and “inclement,” all for obvious reasons.

But the way we use words, and what that says about us collectively, goes far beyond mere traffic trends. Sokolowski, Stamper and I recently discussed these issues by e-mail. A lightly edited transcript of our conversation follows.

So maybe we should start with some background — what is a lexicographer?

Peter Sokolowski: A lexicographer is a person who writes or edits a dictionary. We call ourselves “editors” in-house. But “lexicographer” is a pretty fun word to say. The job of a dictionary editor is to prepare and present research about language. One of our former editors in chief said, “tell the truth about words,” and that sums it up for me. It’s about looking at how the word is actually used.

What do you guys do all day?

Kory Stamper: The two main duties of a lexicographer’s day are reading and defining. We read everything: books, trade journals, blogs, web magazines, phone books, cereal boxes, take-out menus, and so on. What we’re looking for are words that catch our eye, whether old, new, or new uses of old words. We take the surrounding context of that word and the bibliographic information for the source and create a citation out of it. Those citations get added to a database, and they’re the raw material we use in writing definitions. (Our citation files are enormous: the paper files go back to the 1800s. It’s a lot of raw material.)

When we’re writing or revising a dictionary, most of our time is spent on defining. We define contextually–that is, based on how all those words are used within those citations–so a good chunk of the day is spent reading, sifting, and sorting through citations, making little mental piles based on whether the use of this word on this citation is covered by sense 1, or 2, or not at all. Once the sorting is done, you look at words that aren’t covered, and then draft a definition.

Drafting definitions requires some training, some experience, and a lot of concentration. It’s very, very quiet on the editorial floor.

How closely do trending words on m-w.com seem to correlate with the news? Do look-up trends ever alert you to news before other sources do?

Sokolowski: When the dictionary first went online back in 1996, we could see for the first time which were the most looked-up words in English: affect, effect, ubiquitous, pragmatic, and integrity were among the top words. It was fascinating for us — dictionary editors spend so much time writing definitions but never could have known if anyone ever read them. Now we had data. There are evergreen words like these that top the list nearly every day.

But then, when Princess Diana died, the words at the top of the lookup list changed: paparazzi, cortege, princess. We could follow what people were thinking about according to what words they were looking up online. We can even sometimes tell what people are doing: on Christmas afternoon, lookups for two-letter words like za and qi spike — people are playing Scrabble. We do indeed see words trending and then research the news story that caused the spike. If it’s something that was said on TV, I ask why a certain word is spiking on Twitter and someone will fill me in.

How do you decide when a new word is more than just a fad — that it’s an actual word? In other words (heh), how do you decide when to add/subtract words from the dictionary?

Stamper: First, the easy one: removing words. In order to remove a word from a dictionary, you have to show that it no longer has current, written use and no significant historical use. This is almost impossible to do, which is why archaic words like “thee” and “thou” still have entries in the Collegiate Dictionary. For a new edition of the Collegiate, we will enter thousands of new words; we may only remove a few dozen words. (Of course, with an entirely electronic dictionary like the new Unabridged, we don’t have to deal with space constraints, so we’re removing fewer and fewer words in that reference.)

Putting a word in requires a little bit more finesse and sprachgefühl. You need to evaluate the evidence to see if it’s met three criteria: substantial use, sustained use, and meaningful use. Is the word used in a fair amount of widely read prose across all subjects and all registers? Has it been in consistent use for a good number of years? Does it, in fact, have a meaning? If so, it’s eligible for entry.

But, as you get at, it’s not always so cut and dried. Lexicography tends to lag behind the language because you need to look at the long view of a word’s use. Lots of new words see enormous use once or twice in a few decades; is that sustained enough? Or a word may be very specialized vocabulary for a long time before being gradually picked up by the general public. When do those gradual adoptions finally enter the lexicon fully? “Because X” and “feels” (as in, “all the feels”) are really common in essays, blogs, and magazines on the Internet, but we tend not to see either construction in printed materials much. So does that mean that “because X” or “feels” has fully entered the lexicon? Maybe, maybe not.

And frankly, we like linguistic fads. We try on new constructions and words like teenage girls try on prom dresses: you may not take any of them home to keep, but damn if it doesn’t make you feel great to try on a bunch and see how they look.

So what is the next great linguistic fad?

Stamper: It’s really hard to predict linguistic trends: English doesn’t like to behave according to our expectations. But there are old trends that will continue: shortening (“whatevs”), blends (“snowmageddon”), and initialisms (“SMH” or “IKR”), all of which we’ve been doing for centuries, will continue to pop up (and no doubt irritate people). Specialized vocabulary will continue to seep into general vocab. Words will undergo functional shift, and drift in and out of use. People will still write us and try to get their pet coinage entered into the dictionary. And so the language goes.