Where did our words go?

Science has a rather good history of people finding things they weren’t looking for. Alexander Fleming famously returned from holiday to some unpromising-looking Petri dishes and ended up discovering penicillin. Percy Spencer was working on radar technology when the chocolate bar in his pocket melted, which eventually gave us the microwave oven.

My favourite might be Arno Penzias and Robert Wilson, two radio astronomers who spent a long time trying to work out why their antenna was picking up an irritating background hiss. They investigated possible interference and even evicted the pigeons that had taken up residence inside the antenna. After clearing out what Penzias called “white dielectric material”, the hiss was still there.

It turned out they were listening to the afterglow of the Big Bang.

I love stories like this because research doesn’t always politely answer the question you asked. Sometimes it taps you on the shoulder with a much more interesting one.

Something similar happened to psychologists Matthias Mehl and Valeria Pfeifer. They were revisiting the much-debated question of whether women talk more than men when another number began bothering them.

Mehl’s earlier research, published in 2007, had estimated that the average person spoke around 15,900 words a day. When Pfeifer analysed their newer data, the figure was closer to 12,700. Mehl assumed something had gone wrong and asked her to check it. She did, and the number remained stubbornly where it was.

So they followed it through data from 22 studies involving 2,197 people who had been recorded going about their everyday lives between 2005 and 2019. Those studies had originally been conducted for completely different reasons. Nobody involved knew that years later someone would be looking at the recordings to find out how many words they had spoken.

When Mehl and Pfeifer plotted those word counts against time, they found a steady decline. Their estimate suggests that with every successive year, people were speaking around 338 fewer words per day.

Which leaves us with an intriguing question. Where did all those words go?

The disappearing conversation

Mehl doesn’t think we’ve abandoned some great daily fireside conversation. He suspects the loss is spread across smaller exchanges, the sort we barely notice having. It might be asking someone for directions or exchanging a few words with a neighbour. Perhaps it’s the conversation that used to happen at the checkout.

There is something mundane about that explanation, which is why I find it interesting. Technology has gradually made many of these conversations unnecessary.

I can’t remember the last time I asked someone for directions. My phone tells me where to go and, usually, where I’ve gone wrong. Self-checkouts allow us to move through shops with minimal interaction, while screens take more of our orders in restaurants and cafés.

I’m not nostalgic for inconvenience. Google Maps and I are very happy together.

But perhaps inconvenience was occasionally doing something useful without us noticing.

Those little exchanges brought us into contact with people outside our immediate social circles, and there is some fascinating research suggesting they matter.

Psychologists Gillian Sandstrom and Elizabeth Dunn have studied what are known as “weak ties”, the acquaintances and familiar strangers who sit at the edges of our social worlds. They found that people reported a greater sense of belonging on days when they interacted with more of these weaker connections.

In another experiment, people buying coffee were asked either to make their transaction as efficient as possible or to have a social interaction with the barista. The people who interacted came away feeling better about the experience.

The conversation you barely remember having may still have done something for you.

Why we avoid talking to strangers

There’s another piece of research I like because it involves commuters.

Nicholas Epley and Juliana Schroeder asked people travelling on public transport to strike up a conversation with a stranger. Others were asked to keep to themselves.

If you’ve ever sat on a silent commuter train, you can probably guess which option people expected to enjoy more. Yet the people who had a conversation reported a more positive journey.

People had predicted the experience incorrectly. They underestimated how receptive the stranger would be and how much they themselves would enjoy the interaction.

We seem to be rather good at talking ourselves out of talking to people.

And technology has become rather good at helping us.

The practice we don’t realise we’re getting

This is where the research becomes interesting to me because so much of my work involves helping people become more confident communicators.

Conversation contains an enormous amount of improvisation. You can prepare carefully for an important meeting and think through what you want to say, only for another human being to ask the one question you hadn’t considered.

Networking is good at exposing this. You can rehearse a perfectly sensible introduction, walk over to someone and discover within 20 seconds that the conversation is heading somewhere else. Occasionally you’re halfway through a story when you realise the other person’s eyes have developed a sudden fascination with the buffet.

Learning to handle those moments comes partly from experience.

Sandstrom and her colleagues explored this by encouraging people to approach strangers and start conversations over the course of a week. As people accumulated experience, they became less concerned about rejection and more confident in their conversational ability.

That makes those seemingly insignificant exchanges more interesting. A conversation while you’re waiting for an event to begin may go nowhere at all, but you’ve still had to find your way through an interaction you couldn’t script in advance.

If everyday life gives us fewer of those opportunities, we may be getting less practice without ever deciding to practise less.

Then came 2020

There is an important detail in Mehl and Pfeifer’s study. Their data ends in 2019.

Generative AI didn’t cause the decline they observed, and the researchers don’t claim to know precisely what did. Mehl has pointed towards digital communication and technology as possible contributors, while also suggesting there may have been broader changes in how our society “congregates and communicates”.

That word “congregates” is interesting, because their study stops just before we conducted a rather enormous experiment in what happens when people stop congregating.

COVID closed offices and emptied many of the places where we casually encountered other people. The conversation before a meeting disappeared when the meeting became a Zoom link. Cafés, pubs and other social spaces closed too, taking many chance encounters with them.

Sociologists have a name for some of these spaces: “third places”. Home is your first place and work your second; third places are the informal spaces where we spend time around people we haven’t necessarily arranged to meet.

Research into remote working gives us a glimpse of what happens when people no longer occupy the same physical spaces. One large study of more than 61,000 Microsoft employees found that remote working made collaboration networks more siloed. Another study, found that thousands of weak connections between people disappeared when they were no longer physically in the same place. When some co-location returned, some of those connections returned too.

Most of us have experienced a less scientific version of this. A scheduled video call usually involves the people who were invited to it. Spend a day in an office or attend an event and you don’t quite know who you’ll end up talking to.

That’s part of the value.

You might have an entirely forgettable conversation. Or you might meet someone who gives you an idea you hadn’t considered and the conversation carries on long after the event has finished. You couldn’t have scheduled it because you didn’t know it was going to happen.

COVID didn’t create the decline Mehl and Pfeifer found. Their data tells us it was already underway. But the pandemic disrupted many of the places where those small, unplanned interactions happened, and some of the habits we developed during that period have stayed with us.

Then our relationship with technology changed again.

What happens when AI finds the words for us?

Generative AI adds a different dimension to all of this.

I use AI and find it useful. It can help us prepare for an important conversation by playing devil’s advocate or exposing something we haven’t thought through. For someone staring at a blank screen with no idea how to begin, it can get things moving.

AI can now formulate the communication itself.

If I can’t quite work out how to phrase something difficult and ask AI to do it for me, I’ve saved myself time and possibly produced a clearer message. That’s useful.

But what happens if I do that every time finding the words feels uncomfortable?

We don’t yet know what generative AI will do to our conversational skills over the long term.

What Mehl and Pfeifer’s research tells us is that we were already talking less before ChatGPT arrived. Since their data ended, we’ve experienced a pandemic that disrupted face-to-face interaction and we’re now adopting technology capable of communicating on our behalf.

It feels like a good moment to pay attention to the conversations we’re still having.

Could networking become more valuable?

This research has also made me think differently about networking.

Networking has a persistent image problem. Even the word can conjure up rooms full of name badges and people attempting to make small talk while wondering how soon they can politely leave.

Yet maybe some of the awkwardness we try so hard to eliminate is part of its value.

When you walk into a room and start talking to somebody you don’t know, you can’t predict what will happen next. Sometimes you find common ground within minutes. On other occasions you discover considerably more about somebody’s train journey than you anticipated.

Either way, you’re practising conversation.

There’s another reason I think this may become more valuable in the age of AI. We can now generate a polished email or LinkedIn message in seconds. As that becomes commonplace, it becomes harder to know how much of the person we’re encountering through their digital communication.

Spend ten minutes talking to someone and something different happens. The conversation wanders. You notice what sparks their interest or discover a connection neither of you knew existed when you started talking.

That’s often how relationships begin.

And perhaps that’s why networking and other opportunities to meet people may become more important as AI becomes more capable. They put us in the way of encounters we haven’t engineered in advance.

The value of finding what you weren’t looking for

Which brings us back to those scientists at the beginning.

Penzias and Wilson weren’t looking for evidence of the Big Bang. Mehl and Pfeifer weren’t looking for a decline in conversation. They noticed something unexpected and followed it.

There’s a version of that in human interaction too.

Some of the most useful conversations we have aren’t the ones in the diary. We meet someone we weren’t expecting to meet, the conversation wanders somewhere neither person intended and an idea appears.

There is a lot to be said for technology that makes our lives more efficient. I’m keeping Google Maps, and I’m certainly not going to start deliberately joining the longest supermarket queue in the interests of conversational self-improvement.

But efficiency isn’t the only thing worth designing our lives around.

Mehl and Pfeifer point out that 300-odd words isn’t actually very much. It could amount to little more than “a brief hallway conversation with a neighbor, a colleague, or even a stranger.”

Perhaps we don’t need to become dramatically more sociable. We could simply become more conscious of the moments when conversation is possible.

Stay for the networking drinks occasionally. Talk to the person sitting next to you before everyone reaches for their phone. Allow a conversation to wander when you have nowhere urgent to be.

Because every now and then, you find something you weren’t looking for.

Sometimes, it’s even a person.