Word by Word: The Secret Life of Dictionaries

by Kory Stamper

It is unfortunate that the entries that take up most of the lexicographer’s time are often the entries that no one looks at. We used to be able to kid ourselves while tromping through “get” that someone, somewhere, at some point in time, was going to look up the word, read sense 11c (“hear”), and say to themselves, “Yes, finally, now I understand what ‘Did you get that?’ means. Thanks, Merriam-Webster!” Sometimes, in the delirium that sets in at the end of a project when you are proofreading pronunciations in six-point type for eight hours a day, a little corner of your mind wanders off to daydream about how perhaps your careful revision of “get” will somehow end with your winning the lottery, bringing about world peace, and finally becoming the best dancer in the room. But nowadays, thanks to the marvels of the Internet, we know exactly what sorts of words people look up regularly. They generally don’t look up long, hard-to-spell words—no “rhadamanthine” or “vecturist” unless the National Spelling Bee is on TV.*3 They tend to look up words in the middle of the road. Some of the all-time top lookups at Merriam-Webster are “paradigm,” “disposition,” “ubiquitous,” and “esoteric,” words that are used fairly regularly but also in contexts that don’t tell the reader much about what they mean.

"You shall know a word by the company it keeps" (Firth, J. R. 1957:11)Apparently, not always an easy task. Also, makes me wonder whether rare words in language models could benefit from use of their dictionary definitions in a way. Maybe also there's a way to recognize that not all context are equally beneficial and some context could somehow be given more weight than others, in case it's not already done.