You have a regional dialect - even on Twitter

Tweeting about 'suttin' - These New Yorkers could use different language online compared to others in different parts of the US

With vast, sophisticated social networking sites, you could say the world has never felt smaller.

Many have referred to the internet being very much a "global village".

But even though we are now supposedly part of one huge web community with a shared lexicon, we still maintain characteristics specific to our locality - even when we communicate in just 140 characters.

That is according to a new study by Dr Jacob Eisenstein from Carnegie Mellon University in Pittsburgh, Pennsylvania.

He used geo-tagging to locate 9,500 users of Twitter across the US to see if, when comparing how we chat among our friends, people in the same places speak in similar ways.

His study of over 380,000 tweets revealed that even with Twitter's ultra-succinct limitations, people in Washington, for example, used different words and spellings to people in New York.

"We had a certain set of expectations going in," he told the BBC World Service's Digital Planet programme.

"We expected people would talk about the city that they live in, local sports teams, local celebrities. There were certain words that we know from spoken English that are associated with certain parts of the US.

"What surprised us was that, in addition to the sorts of words and names that we expected to see, there was a whole other class of words that seemed to have a very strong geographical affinity that we had never known about before."

Hella koo

Take "you". Almost everywhere, it is common to see it shortened to a simple "u".

However, Mr Eisenstein spotted that in urban areas of the US, a different shortening of "yu" was common. But when it came to tweeters in rural areas it was rarely seen.

Even more specifically, in New York, "uu" - meaning the same thing - was used.

Facebook is better for picking up dialects, says Professor David Crystal

Other words like "something" get similar treatment. "Somethin" is widespread, but many in the Big Apple prefer "suttin".

In Northern California, something considered "cool" is often described as "koo", but South Californians are more likely to write "coo".

North Californians are also prone to being "hella" tired, rather than just "very".

Most interestingly, said Dr Eisenstein, these are variations that do not simply reflect written versions of existing dialect. Rather, these seem to be online-only accents.

"There are certainly things that we're seeing in Twitter that would be very hard to bring into your speech."

A further example is in acronyms. Popular since the early days of web-speak, phrases like "lol" - laugh out loud - become less popular in places like Washington DC, where 'lls' - the long form of which is unpublishable here - is widely used as well.

"There are a number of other abbreviations like that, most of which involve profanity, unfortunately, but that are very geographically specific."

Internet dialect

In the UK, David Crystal, language expert and honorary professor of Linguistics at Bangor University, brought together his study of the use of language on the internet in a book, Internet Linguistics.

He believes that the type of language used on Twitter relates more to a person's ethnicity rather than geographical location.

"The tweets I've seen aren't very regional distinctive," he said.

"It's a bit early days yet to see dialects starting to appear," said Prof Crystal. "The one thing about internet language, people join it and what quickly evolves is an 'internet dialect' as it were."

"I suppose the first thing you notice is whether it's British or American in the sense that British spelling will be used. Although there is quite a lot of deviant spelling on the internet, I saw, on the whole, a great deal of standard English spelling."

He adds that when Twitter changed its page to ask "What's happening?" rather than "What are you doing?", people began to adapt their language to suit the new question.

"'What's happening?' means it has become much more of a news service over the past few months," he said. "As a result, when you take a collection of tweets, what you get, over half the time these days, is advertisements and news reports which is much more related to the sort of language you'd get in a newspaper or on the BBC or things like that."

Youthful spontaneity

Prof Crystal said that stronger dialect intricacies could be found on Facebook, rather than Twitter. He puts this down to the younger age of the typical Facebook user.

"As a result you don't get that immediate, young, youthful spontaneity that you see so clearly on Facebook."

"You can tell immediately, 'oh here's a group of kids from Liverpool'," he said. "That's not so easy to find out on Twitter."

Dr Eisenstein noted that the dialects he observed in Twitter use did not exist so prominently in blogs. This could be because blogs present a more formal approach to communication, thus making language more homogenised across the entire web.

For this reason, most studies of dialect in the past have relied on face-to-face interviews rather than written communication. Mr Eisenstein said he was excited about the possibilities for researching dialect that technology offers.

"I really think that the availability of data like Twitter is a real game-changer for how people study language.

"There's a lot of richness you can get by doing a personal interview in terms of finding out someone's life story, and exactly how their location effects the way that they speak.

"On the other hand, we're able to do things on a much larger scale. We're going to scale it up to 100,000 to look at the UK as well as the United States.

"It's something that you could really apply throughout the world."