Back in the 19th century, our marketing folks decided to play up the refined usage angle because prescriptivism was very popular: our dictionary is where you go to learn anything about anything. That really set the tone in North America for how people responded to dictionaries.

There's a measure of prescriptivism and descriptivism in every dictionary. Prescriptivism believes that the language should mirror the best practices of English, and editors are prescriptivist in so far as they don't want to let things they consider to be inelegant or ungrammatical into print.

It's difficult to sit down and write a letter back saying, "you know what, even if we remove the word from the dictionary, people will still continue to use it." That's the tightrope that we walk - "gay marriage" is another example, or the word "nude."

Whenever you try and simplify how people speak, it's just hard to squish them into a simple rule. Language doesn't work that way.

We want to emulate the educated class, even if we don't think of educated people as a class these days.

What makes autocorrect really interesting for a lexicographer is we read in a very weird way. A lot of times we'll catch things that are so clearly autocorrect typos, and you learn to tell the difference between an honest-to-goodness, ham-fisted typo and an autocorrect typo because the latter will not necessarily be a difference of one letter apart on the keyboard, which is usually what typos are, or it won't be the wrong its or the wrong there.

The history of English is full of that, lots of things done with good intentions that 200 years down the road have resulted in a giant mess, where someone's pet peeves - like John Dryden and his hatred of terminal prepositions - could become real standards.