If you spend a few months around people who talk differently than you, you might start sounding a little like them, researchers report Friday in the journal Language. Case in point: reality television. Using the confessional segments recorded by Big Brother UK contestants, the international team examined how much people’s accents change over three months while living together in a confined environment, and it turns out they can vary quite a bit, and in some interesting ways.

Previous research has explored how people’s accents change over the short term (hours and seconds) and long term (years), but changes in the medium term (months) have been trickier to study because confounding environmental variables are hard to control in that time span.

But reality TV — in particular Big Brother — provides ideal conditions for studying linguistic changes over a medium-length period in a closed environment. After all, the contestants interact all the time for months on end, and they can’t leave the house.

To study how accents change over time, the international team of researchers transcribed 14.5 hours of the Big Brother contestants’ “diary room” segments, which are recorded solo, and examined them for five linguistic variables: “voice onset time (VOT), coronal stop deletion (CSD) rate, and the quality of three vowels (goose, strut, and trap),” write the authors. Broadly speaking, the first two features describe the way a person pronounces vowels and consonants in the context of a word or phrase. In short, they describe what we call a person’s accent.

Using a combination of manual coding, machine learning, and speech recognition software, the team analyzed the accents of the 12 contestants. They found that contestants’ accents fluctuated from day to day and that even though some systematic changes occurred, all the contestants didn’t end up sounding alike.

Even with just 12 people in their sample, the researchers had their work cut out for them, says Molly Babel, Ph.D., an associate professor of phonetics and psycholinguistics at University of British Columbia who was not involved in the research but provided the authors with pre-publication comments.

“The amount of data from each speaker is impressive and the amount of hand-correcting that’s necessary in order to trust acoustic phonetic measurements is time intensive,” she tells Inverse.

2005 Royal Visit: Queen Elizabeth II
Even the Queen's accent shifts.

That data, it turns out, is very much in line with existing research on shifting accents. “The time trends they find for this kind of medium-term convergence fall nicely between what we seem to know for short-term flexibility (more variable) and long-term flexibility (less variable),” said Babel.

Previous researchers have found that, in the short term, language plasticity is normal for adults. Many studies have shown that, over the course of a conversation or language task, adults will subtly change the way they pronounce words to mimic that of their conversational partner.

“This plasticity is in line with the ubiquity of style shifting: shifts in a speaker’s linguistic usage as a function of the addressee, topic, and so forth, possibly many times over the course of a day,” the authors of the Language paper write.

In the long term, the general trend is that most adults’ accents remain relatively stable. Sometimes, they shift, usually in a way that reflects linguistic patterns observed in their community. This has been observed in the case of Queen Elizabeth II, whose well-documented public addresses have changed somewhat over the years, reflecting broader changes in England.

A more controversial example of an accent over time is seen in Madonna, who received a lot of ridicule over the years for her shifting accent. More recently, Lindsay Lohan experienced similar ridicule. And while their changes don’t fit in with our commonly accepted views of accents, they’re not total bullshit.

While the mechanisms of accent change merit further exploration, the researchers propose that this study gives some insight into the difference between short-term variability and long-term stability:

“The combination of ubiquitous daily variability with more sporadic longer-term change mirrors the different extents of accent plasticity previously observed in conversation versus over the lifespan, suggesting that speakers show progressively less accent flexibility over progressively longer timescales.”

This paper begins to explore an otherwise lightly-tread area of linguistic research, but Babel cautions that it’s just a beginning, and it’s just a narrow slice of the possible body of language.

“When we study spoken language we must only study a particular style or context because we know that speech and language vary by style and context,” says Babel.

“But, that necessary method always gives us a limited view of the whole linguistic system. So, maybe I shouldn’t call it a limitation of their study, but it’s a caveat that we must always remember when digesting a new piece of research.”