You work up the courage to ask your boss for something over Slack. Or you tap out a 150-word, point-by-point counter to the last point that friend you’re fighting with made over text.
You iterate, edit, and then finally, when you have all the words right, hit “send.”
It might seem like this glassy hell is unique to e-communication, but it actually has an equivalent in the spoken discourse we experience every day. N.J. Enfield, a linguist at the University of Sydney, studies the unspoken rules of conversation. In his book How We Talk, out Nov. 14 from Basic Books, Enfield addresses the strange but necessary silence that falls after one person asks a question and another answers.
On average, writes Enfield, English speakers take 230 milliseconds (ms), about a quarter of a second, to respond to a question. But not every situation yields the same awkward silence. Research on conversational lag time shows that when people respond to a yes/no question, a “no” generally takes twice as long as a “yes.” And, on average, the longer someone takes to respond, the more likely they are to give a non-answer.
Imagine two similar exchanges, with slightly different response times:
A: Can I expense a cab ride to the airport?
B: (200 ms) Sure.A: Can I expense a camel rental for a photoshoot?
B: (650 ms) I’d have to check.
It’s possible that the longer hesitation in the second exchange just means person B needed more time to figure out the answer than she did in the first. But it’s also possible, posits Enfield, that by pausing, person B in the second exchange is nonverbally signaling that she is hesitant. The delayed response could be used as a buffer to soften the blow of an answer the questioner doesn’t want to hear.
Research suggests we intuitively understand the meaning of these delays. In a 2013 study, participants listened to recorded exchanges in which one person asked for a ride, and a second answered “yes.” The listeners were asked to rate how willing the second person seemed to actually give the ride. Response time mattered: as long as the “yes” answer came within half a second, participants didn’t perceive a meaningful difference in how willing the person seemed. But after 600 ms, the listeners started to perceive a drop in willingness, and responses that took between 700 ms and 800s ms (still less than one second) were statistically significantly less likely to be perceived as genuine.
The field of texting-as-conversation is understudied and ripe for more research, says Enfield. But he believes the expectations we have of textual missives are likely to mirror those we have for spoken conversation. “You see people getting kind of calibrated to a certain metric”, he says, “In the sense that you start a conversation and you have a set of expectations about how quickly people will respond.”
We expect someone to respond to a yes/no question in less than half a second; if the respondent takes longer, it starts to feel like it might be a “no”, or, at best, a wishy-washy answer. When you see someone start to type over text or Slack, and then stop, you know the person is there and knows a response is needed, but is delaying. You worry: Maybe she will say no, and if she does say “yes”, but it took her too long, maybe she didn’t really mean it.
Of course there are lots of reasons someone might not respond within a given person’s expected time frame to a chat message or an email. Perhaps she was free when she started to reply, but then became busy. Maybe she was in a meeting, so her attention was split. It’s also possible that we haven’t been texting long enough as a society to agree on a reasonable response rate, and what she thinks is perfectly normal gives you server anxiety.
As we become more finely tuned to the nuances of our electronic missives, we may come to interpret delays there in much the same way we do with the brief pauses we already accept in everyday conversation.