Let them eat cake — and drink vodka

It's amazing what a little frequency analysis can tell you. A joint research team of the Universities of Cambridge and Lancaster has compared recordings from the 1990s with spoken British English today to see how culture has changed. By looking at the frequency with which certain words are used, they've found some interesting results, especially concerning food, drink and family.

Champagne and cake

Two of the nation's favourites.

It appears that long-standing festive treats, such as sherry and brandy, are declining in popularity, with champagne, vodka and gin now constituting our favourite festive tipples. Based on recent recordings, traditional British favourites and holiday dinner staples, such as Yorkshire puddings and custard, have been overtaken by takeaways, notably pizza and curry, which have risen in frequency by four and five times respectively.

Despite the rise in convenience food, talk of over-indulgence, especially around holiday seasons like Christmas, is also more prevalent today (if you're worried about your own intake, read about the maths of obesity). Compared to the 1990s, we talk about calories twice as much as we used to. But our words don't necessarily translate to actions; our most talked about food item at Christmas – as well as the rest of the year — is cake.

The initial findings also suggest that alcohol is more important to British English speakers than coffee. While tea remains the nation's favourite drink, spoken an average 255 times per million words, hot drinks are dwarfed when compared to booze.

While our tastes may have changed, our desire to spend time with our families at Christmas seems unrelenting and, when we aren't with our families, we are more likely to talk about them. We talk about family almost twice as much as we used to, as it seems that somehow the concept of family is more salient in British English speech today than it was two decades ago.

Claire Dembry, Senior Language Research Manager at Cambridge University Press, said: "This analysis presents an interesting insight into how our use of language has changed over time. Our tastes in food and drink have certainly changed, but our interest in the family seems to be ever increasing. Coupled with the finding that we are also significantly less self-centred in our speech than we used to be, with a reduction in frequency of words associated with self — we say "I" or "me" 47 % less than we used to — this suggests Christmas may be even happier this year. But perhaps this can be attributed to all the talk of champagne!"

Vodka

Forget the sherry, bring on the vodka.

These are only the initial findings from a small pilot of the project, named the Spoken British National Corpus 2014, which is now under way. The Corpus is a very large collection of recordings of real-life informal, spoken interactions between speakers of British English from across the United Kingdom.

Tony McEnery, from the ESRC Centre for Corpus Approaches to Social Sciences (CASS) at Lancaster University, explained: "We need to gather hundreds, if not thousands, of conversations to create a full spoken corpus, so we can continue to analyse the way language has changed over the last 20 years. This is an ambitious project and we are calling for people to send us MP3 files of their everyday, informal conversations in exchange for a small payment to help our team to delve deeper into spoken language and to shed more light on the way our spoken language changes over time."

If you would like to submit recordings to the research team, visit this website.

The academic community will benefit from access to a large corpus of British English speech that is balanced according to a selection of useful demographic criteria, including gender, age, region and social class.

As we have seen before on Plus, the study of word frequencies doesn't just tell us about our culture, it also reveals some interesting maths. It turns out that those frequencies always seem to follow a certain distribution known as Zipf's law — see here to find out more. It would be interesting to know if this is the case for the Spoken British National Corpus 2014 too. If yes, then this would add to the mystery, as nobody seems to understand why the frequencies should follow that distribution. It's a puzzle for linguists to mull over once they've finished their champagne and cake.

This story is based on a news release from the University of Cambridge.