You can take the girl out of linguistic research -

My name's H2 and I'm a recovering academic.

When I was in the depth of my academic phase, my work was deeply influenced by the work of William Labov. Labov is a seriously clever man who has a lifetime of sociolinguistic pioneering behind him. One of his early pieces of work, which changed the way that linguists investigated variation in language, was a study he did in a couple of New York department stores. He wanted to look at whether people pronounced the 'r' towards the end of words. So what he did was to go to the fourth floor of whichever store he was in, spot an item, go to another floor and ask staff where the item was - just to get them to say "fourth floor" naturally, and then would note whether or not they pronounced the 'r'. He found that there was a difference between the staff in the different stores, and - and this was the new bit - this difference was statistically significant.

Why am I telling you this? Because I have a client - a wonderful, regular client - where I work on the fourth floor. And every single time I hit the 4 button in the lift, I think of William Labov.

Which just goes to show you can't take linguistic research out of the girl.