The British National Corpus

In​  bnchs01 Carlo tells an old Mitch Hedberg joke: ‘Is a hippopotamus a hippopotamus or just a really cool opotamus?’ bnchs01 is an audio file I recorded and submitted as part of a drive to assemble a multimillion-word record of spoken British English. Carlo is a friend, and the joke is part of one of the many conversations I recorded. (Hedberg, 1968-2005, was an American stand-up comedian.) For each hour of audio, I earned £18.

The spoken British National Corpus is being compiled by the Centre for Corpus Approaches to Social Science at Lancaster University and Cambridge University Press. cass was set up to investigate the use and manipulation of language. Other projects include the Discourse of Online Misogyny (or doom) and Changing Climates, which uses corpus-based data to compare the way global warming is talked about in Brazil (where the concept is generally accepted) and Britain (where climate change scepticism is on the increase). One team is assembling a corpus of historical texts made up of roughly ten billion words to examine British attitudes to poverty in print media between 1473 and 1900.

cass started collecting samples of real-life, informal, spoken interactions between speakers of British English in 2014. The project is now in its final stages; the collected recordings are being transcribed and will soon form a publicly available resource. I began submitting conversations in 2015 after I saw an announcement on a university noticeboard calling for recordings of face-to-face conversation between people from all over the country. They could be on any topic so long as the conversations were as natural as possible.

I was prolific: it was like spinning money out of thin air. It beat going to Flu Camp in Whitechapel – something other broke friends of mine were resorting to at the time – to be injected with a vaccine, infected with a strain of flu, quarantined in a sealed room for ten days and pored over by scientists in hazmat suits. Flat-screen TVs and games consoles were provided. There was no natural sunlight. I listened to – and recorded – stories of such procedures as the Nasal Tolerance Test. ‘I’d say it was worth it for the money,’ my friend says. He has been describing how his snot was tubed away for testing several times a day.

I collected conversations for more than a year: everything was on the record. I had to get consent, but only once for each person. It was surprising how quickly people seemed to forget they were being recorded, even those who were prone to paranoia or had complained to me about cctv and snooping. It wasn’t at all sinister, I assured them. It would be used for linguistic research and to help people learn English. That sort of thing – as well as making me some easy cash. I didn’t really think about what might happen to my recordings or how they might be used. Occasionally I imagined linguists in headphones sitting attentively at desks, listening in. Did they have favourite contributors?

Eventually I was asked to stop sending in recordings. Contracts were redrafted to include a limit of 27 hours for new contributors. Preliminary research (during the 2015 general election) found that ‘the British talk about cake 50 times as much as the deficit.’ The Great British Bake Off had hit its stride – but I wondered if the numbers had been affected by my living with a baker.

The first version of the spoken British National Corpus was compiled in the early 1990s; the new iteration provides a means to compare and contrast. Armed with this massive data set, linguists can set to work on a range of questions. One team at the University of Mannheim is assessing the validity of lexical gender clichés. Is it still the case that female speakers use ‘booster’ words (such as ‘really’, ‘totally’, ‘absolutely’) more often than male speakers? An array of sociolinguistic trends can now be identified or posited. Do we talk more or less about education? About family? Politics? Has creeping American influence replaced ‘marvellous’ with ‘awesome’? What has the rise of the internet done to the way we speak to one another?

I have hours and hours of conversation stored on a hard drive. Slices of my life, acoustic snapshots: a 15-minute drive to visit a relative full of throwaway comments (‘look, there used to be a building there’). Or long exchanges between old friends, conversational threads weaving in and out. Rip-roaring anecdotes; boring anecdotes. Heated discussions. Plans in the offing. More chitchat. Killing time in a car or waiting at an airport. ‘I can’t believe this bottle of water cost £3.20.’

Last September a rash of newspaper headlines heralded the demise of British English. A hsbc report, commissioned to mark the launch of a voice-based biometric security system, predicted sweeping changes. In London in 2066, the dental fricative – ‘th’ – will have all but disappeared. And as American influence grows, the phenomenon known as ‘yod-dropping’ (whereby ‘duke’ becomes ‘dook’ and ‘news’ becomes ‘nooze’) is set to become prevalent. There’s nothing surprising about this. You need only listen to a broadcast from 1966 to hear how much British English has changed in the last 50 years.

Some newspapers were quick to blame immigration and multiculturalism for the perceived decline of traditional British English. It is true that non-native speakers of English find dental consonants – articulated with the tongue against the upper teeth – difficult. But there are many other factors. Language is a key way to mark generational difference. Young people often want to sound different from their parents. Language is used to signal belonging, identification with a particular group, and it costs nothing – unlike the clothes and gadgets that might serve the same purpose. It also seems likely that people will spend more time talking to computers, flattening out distinctive accents in the process. ‘The Sound of 2066’ predicts that keyboards will soon be obsolete. The report also suggests that while English has always borrowed words from other languages to plug linguistic gaps, computers will soon invent words and phrases in order to make communication more efficient.

Based on the conversations I recorded with my younger brother, the dental fricative is indeed endangered. I ask if I can borrow his bike. ‘Well, the fing is, Harry,’ he sets off. In conversations with my grandmother (not a native speaker of English) the dental fricative is alive and well. ‘I do think this is going to be a rather good year for the roses,’ she says after I compliment her on her garden. In my own speech, the dental fricative seems to come and go. It’s mostly there for my grandmother, but mostly absent for my brother.

Listening to the recordings, it isn’t long before I lose interest in the voiced dental fricative. I find myself caught up in old conversations. I click through several recordings, skipping back and forth and into the middle of conversations. I try to place them, to guess what’s coming. I launch into the middle of bnchs07. I hear muffled background noise: scraping chairs, music turned down low; a crinkle of plastic bags, perhaps. I am about to skip forward, and then a familiar voice cuts through the hubbub. ‘Oh man, today … I was so hung over.’ It belongs to an American chef with a wonderful laugh. And I don’t have to wait long. Here it comes: a deep, round chuckle right from the middle of the chest, rolling up the oesophagus and out into a high-pitched giggle. I remember this particular evening – we were in Paris and I remember the hard-won warm beers, the long search for a late-night off-licence. The chef is telling us about his lunch: he had it on the street, eating with his hands, hidden behind a dumpster. ‘I just … I couldn’t wait. Juice on my hands, face, everything. Bent over. Like, really gross.’ He was caught in the act. An elderly Frenchman stopped in front of him. The man looked him in the eye, and delivered a good-natured, full-blooded ‘Bon appétit!’ before walking on. ‘Best meal I’ve had in Paris,’ he says. Cue the laugh. He had spent most of the week in Paris’s best restaurants.

I choose a different file. The cursor lands on bnchs24, and I glance at the short information sheet I submitted along with the recording. It is late February. Some friends and I are eating dinner at home. The conversation is boring, a bit stilted. Being hit by lightning is the topic. Electric shocks. Lots of near misses – my friend this, my aunt that. ‘Have you ever been Tasered?’ someone asks. No one has.

It was an uneventful evening but as I listen to a wandering discussion on Tasers, I remember where I sat at the unsteady table and the plate I had – the one with a blue rim and a thin crack down its middle. I don’t remember what we ate. No one mentions the food. I fast forward half an hour. More of the same – if a little looser now. I hear the gurgle of wine being poured into a glass. I notice how much my friends swear. Someone is talking about Kanye West, and someone else admits: ‘I don’t fully … know who he is.’ I pause the recording. Even before I listened, I had remembered this remark, and the response: ‘I literally have to talk about him three times a day.’ But it had lodged itself in my memory as part of another more lively evening. It is surprising to have a memory proved so emphatically wrong. Did I really have the cracked plate with the blue rim? What else have I misremembered?

I press play. ‘How many times have you talked about him today, then?’ I ask. ‘At least five or six,’ comes the answer, quite serious. I suppose it is a common thing, to feel unnerved by the sound of your own voice on record. Do I really sound like that? There’s a reason for this. In a recording, sound reaches the ear only through air conduction. But when I talk, I hear the sound of my voice not only through air conduction but through bone conduction. As I talk my vocal chords set off vibrations that travel through the body to the inner ear. Flesh and bone conduct deeper tones better than air, and the acoustics of a human skull give a lower, richer, more mellifluous sound. But no one else hears what I hear when I talk – they hear what’s on the record, what’s vibrating through the air. So yes: I really do sound like that.

When I hear myself on the recordings, it is almost like listening to another person. I can revisit these conversations – but only for a moment. For a moment, I am back in a small room in Paris late at night or at an unsteady table, slightly bored. Details re-emerge. Facial expressions, furniture; the time of day or the weather. But as soon as I begin to talk, the spell is broken. In Paris, I ask the chef: ‘Whereabouts were you?’ The question flops out. In my memory of that evening I had been funny and sharp, but I am neither of those things. I am tired and drunk. I had remembered that evening fondly – now I’m ambivalent.

I have an audio diary of my life at that time, just as I intended. I know in great detail where I was and who I was talking to for the best part of that year. I also know what people were talking about – not Trump, not Brexit. At the time, I imagined the recordings would be useful. I could comb them for stories or jokes that I could use in writing perhaps or in later conversations. In truth, it’s hardly worth the trouble. There is no filter on these recordings, no interpretation, no emphasis, no elision. The conversations are largely inane, often boring. It is fair to say that conversation flowed – but I understand the metaphor now. If something was memorable, I have remembered it. The rest flowed away. I don’t need a recording of the American chef’s laugh to remember it, or a two-hour recording of a dull dinner to remember one amusing exchange about Kanye West. Memory is a great organisational tool, editing the torrential minutiae of experience.

I sat listening to these recordings for a long time. In the end, I began to think of Beckett’s Krapp. I saw him hunched over his tape player in the half-light, crooning the word ‘spool’. His whole life, there on record. He selects an old tape. ‘Hard to believe I was ever that young whelp. The voice! Jesus! And the aspirations!’ He listens to his younger self describe a moment of spiritual realisation – ‘that memorable night in March at the end of the jetty, in the howling wind, never to be forgotten’ – without a flicker of recognition. Krapp then files a new recording in a rasping voice. ‘Just been listening to that stupid bastard I took myself for thirty years ago, hard to believe I was ever as bad as that!’

I took my headphones off and looked at the recordings I had collected. Hours and hours of them. Let the sociolinguists have them.