There's a Word for It (or Soon Will Be)

Here's a little thought experiment: let's say you've been invited as a dinner guest to my house and you are particularly impressed with one dish I have served, one you've never tasted before. You ask me what it is and I say it's called ghallipkorz. That's an approximate spelling — what I actually say is something like /ɣɘɬɪpkȏʀz/. You're so taken with the dish (and by the way, you're a food writer for a national newspaper) that you write about ghallipkorz in your column the next day, and by the end of the week, /ɣɘɬɪpkȏʀz/ (the word, not the dish) is on the lips of celebrities, politicians, restaurateurs, and chat show hosts. After the weekend, major dictionaries update their online versions with the new word, complete with its proper pronunciation.

What's wrong with this picture? Quite a lot, and you can say that with certainty even if you don't have a PhD in linguistics. For one, we all know that words don't catch on this fast. Number two: words that require phonemes that are not part of a language's standard set don't find acceptance in that language. At a minimum, they must be shoehorned into a pronunciation that native speakers would consider normal, and if they don't lend themselves to that treatment, chances of finding a foothold in English are minimal. Think about Eyjafjallajökull, the volcano in Iceland that paralyzed European air traffic in 2010. For most people, it never became more than “that volcano in Iceland” and today probably less than one percent of people could reproduce one syllable of it. Number three, dictionary editors adhere to a standard, somewhat conservative process in deciding whether a new word can join the ranks of the anointed ones that enjoy the status of headword in their publications.

Long ago in the Lounge we briefly explored the dignified and stately process through which new words, when they achieve a respectable degree of circulation, receive the imprimatur of a dictionary definition. Let's call that the front end of lexicon development and expansion. The back end of that process — in which speakers introduce or innovate additions to the language that then gain circulation — is a little more difficult to document: even today in the Internet age, tracing the origins of linguistic innovation is a sleuth's game and it's a subject that intrigues linguists.

Today researchers are trying to bring more light to the process by which people create, learn and use new words. Wordovators, a collaborative project of Northwestern University in the U.S. and the University of Canterbury in New Zealand, is looking into the subject. Janet Pierrehumbert, professor of linguistics at Northwestern, says "We are taking a fresh look at the whole idea of a word. We think about words as things that are alive and that reproduce when people learn them and use them." The professor's organic-lexical metaphor is not new and the parallels that exist between language and reproduction biology have supplied, and continue to supply, many helpful and illuminating analogies.

Recent research suggests that this fertile field of analogy is not coincidental: it is now claimed that the emergence of human language is the most recent of a small number of highly significant evolutionary transitions in the history of life on earth. Why? Because language enables an entirely new system for information transmission: human culture. Language uniquely supports heredity of cultural information, allowing our species to develop a unique kind of open-ended adaptability. The nuts and bolts of how this happens, however, are only now beginning to be explored.

The Wordovators Project draws specifically on analogies between biodiversity and language diversity. People know an enormous number of words. New words continually arise as people modify and recombine parts of existing words — much as new biological species arise through evolution.

One approach to understanding the process that leads to a novel word being accepted in the lexicon has been explored and reported here before — in this interview with Dr. Allan Metcalf, author of Predicting New Words. The Wordovators Project will draw much more heavily on large-scale experiments. How is that going to work? It would be unwieldy and uncertain at best to attempt to create the conditions for a new word to emerge and then see if you can get early-adopting speakers to cooperate, so Wordovators will use mathematical modeling and other Internet-age resources to try to pin down how the word innovation process works. The Wordovators website currently has two GWAPs available for people to play, as well as an experiment that is open anyone registered as a mechanical Turk worker.

We've talked about GWAPs (Games With A Purpose) in the Lounge before, specifically about Google Image Labeler (here) and Wordrobe (here). A well-designed GWAP is a win-win: fun for the players, and valuable for the data it provides for the creators. GWAPs are a natural development arising from the convergence of several phenomena: the penchant that people have for amusement and diversion, the broadly-based interconnectivity that the Internet makes possible, and the need of researchers to acquire useful data that is not prohibitively expensive. What do the designers of Wordovators hope to learn from the data that their games will provide? Their ultimate goal is unabashedly lofty: to discover the fundamental mechanisms that support the complexity of the lexicon in human languages.

To return for a moment to ghallipkorz, we all probably have an instinctive notion, even if we've never articulated it before, about how a putative new word might succeed in a language: it has to be pronounceable and consistent with the language's sound patterns and prosody; it has to somehow “sound” right — in other words, it has to match the thing it stands for in a way that other words already in the language match the similar things that they stand for; it can't encroach too obviously or fully in form or in meaning on a word already in the language, but at the same time it helps a new word to have some pieces that are found in other words; and there has to be a genuine and enduring need for the new word. These many criteria make it fairly easy for all of us to spot from a mile away a word that would stand little chance of acceptance, but the subtle ways by which new words do find acceptance has not been well explored empirically, and that's what Wordovators sets out to do.

At the same time that the Internet makes possible a large-scale project like Wordovators, we probably shouldn't lose sight of the ways in which the Internet changes the way we use and view language and new words. The notion that words even need “the imprimatur of a dictionary definition” to gain acceptance today needs reexamining. Those who think of a dictionary as an authoritative book are ever decreasing in number; more who will know it mainly as a helpful but not necessarily authoritative Internet-based service are born every minute. So the dynamics of word acceptance today are more fluid and volatile than in the past, and they will probably remain that way.