Twitter is full of regional 'accents,' study finds
Published: Tuesday, January 11, 2011 at 3:19 p.m.
Last Modified: Tuesday, January 11, 2011 at 3:19 p.m.
PITTSBURGH (AP) — Tweeting about what club "y'all" are going to tonight? Must be from the South. Looking forward to "suttin" special? Then you probably live in New York. Think that new movie was "koo?" Northern California.
The words you write on Twitter can tell people more than just the status of your relationship or how you like the latest Bon Jovi CD. It may just indicate not only how you're living, but where you're living in the U.S.
Researchers at Carnegie Mellon University examined 380,000 messages from Twitter during one week in March 2010 and found that the social networking site is full of its own kinds of geographical dialects.
Take the word cool. Southern Californians tend to write the shorthand "coo," while their neighbors up north use the phonetic shorthand "koo."
The 4.5 million words the researchers examined were full of similar examples. Some were obvious — like "y'all" in the South or "yinz" in Pittsburgh — and some more mysterious. The word "suttin" was found over and over in New York City, a shorthand for "something."
Jacob Eisenstein, a post-doctoral fellow of computer science in Carnegie Mellon's Machine Learning Department, and his colleagues were able to analyze the geotags attached to Twitter messages sent from mobile phones for the study. In all, they looked at 4.5 million words.
"Some of what we found really just confirms previous intuitions, but some things were much more specific for social media," said Eisenstein, noting the phrase "very tired."
Northern Californians tend to substitute "hella" for very, whereas New Yorkers opt for "deadass" tired; those in Los Angeles would be more likely to follow the word tired with the abbreviation "af" — short for "as (expletive.)"
Some of the differences across Twitter can be explained by the need to write concisely to fit the site's 140-character limit. But others, not so much.
While using "u'' in place of the two-character longer "you" is pretty common, a lot of New Yorkers do the opposite and lengthen the word to "youu." Or even emphasize "I'' by writing two of them — as in "II."
Scott Kiesling, associate professor of linguistics at the University of Pittsburgh, said social media provides researchers lots of easily obtainable data in which they can explore and examine how people are speaking. He said the next step is examining whether these phrases spread like "pancake batter hitting a pan or hop from city to city" — if they spread at all.
"That's sort of the big question," Kiesling said. He said there's a burgeoning interest among linguists to study online speech more closely, and noted a conference this year at Georgetown University that will be examining language and new media.
Eisenstein said some of the online "accents" mirror those in the spoken language, but not all. For example, many people in the Great Lakes region tend to have similar accents when speaking, but that wasn't necessarily found to be true in the study, he said.
"One thing I think that it shows is that people really have a need to communicate their identity — their cultural identity and their geographic identity in social media," he said.
Reader comments posted to this article may be published in our print edition. All rights reserved. This copyrighted material may not be re-published without permission. Links are encouraged.