What to do when the conversation goes quiet
A pause is not a verdict. We should say up front that nobody has tested that: there is no research on what silences mean or on what best ends one, so this is the most lightly sourced page in the guide. What follows is us reasoning outward from studies about something adjacent, and the parts that are only us are marked as such.
What a silence probably is
The instinct at that point is to reach for something new: a fresh topic, a change of subject, anything to fill it. Two other moves seem better to us, and the second one has real evidence behind it.
Go back one step
Either way, going back for it is cheap: "You said earlier you'd nearly not come — what was that about?" And this is where the guessing stops and the evidence starts. Asking for more about something the other person actually said is the specific move the question-asking research singles out, and people do far less of it than they could. A silence is a good moment to ask the one you skipped.
Or go down instead of across
The other move is the one almost nobody makes: instead of finding a new shallow topic, ask something more personal about the one you already have.
This feels like the riskier option and the evidence says it is not. When researchers had strangers hold either shallow or deep conversations, the deep ones were consistently less awkward than participants had predicted, and people ended up feeling more connected and happier than they expected. People given the choice reliably picked the shallower topics — and were wrong about which would go better.
There is more on why this works, because it is the best-evidenced finding in this whole guide.
In text, quiet means something different
Everything above is about a pause inside a conversation that is still happening. Typed conversation has a second kind of quiet that spoken conversation doesn't: the other person simply stops.
That one is genuinely ambiguous and no amount of reading it will help. They may be bored. They may have been called away mid-sentence. They may have three conversations open and yours is the one they didn't get back to. The information that would tell you which is the same information everybody withholds about when they wanted to stop.
Worth keeping separate from all of that: an opener nobody answered at all is not this problem. On a site where people filter who they even see by gender, age and country, most of that is settled before your message is read. A conversation that started and then stopped is the harder one, because something did happen first.
What is actually established
Less than this page needed, which is why so much of it is marked. Nobody has run an experiment on what pauses mean, so reading one as a turn boundary stays our inference. What is measured is narrower and still worth having: that follow-up questions increase liking, and that people systematically overestimate how badly a deeper question will land. Both were studied in spoken or live conversation rather than in text, and both are set out in full below.
Where this comes from
Picking up something the other person already said is the highest-return move available, and people do far less of it than they could.
What the study actually did
Studies of live conversations found that follow-up questions — those that ask for more about what a partner just said — predicted being better liked. When people were told to ask more questions than usual, the number used to represent a high rate was nine per conversation against a low rate of four.
What it doesn't show. Anything about silences specifically. No one has studied what pauses mean or what best ends them; reading a pause as a dropped thread is our inference from the question-asking work, not a measured result.
Huang, K., Yeomans, M., Brooks, A. W., Minson, J., & Gino, F. (2017). It doesn't hurt to ask: Question-asking increases liking. Journal of Personality and Social Psychology, 113(3), 430–452 · 10.1037/pspi0000097
Going deeper on the topic you already have beats finding a new shallow one — and people consistently expect the opposite.
What the study actually did
Strangers were assigned to discuss either shallow questions or notably personal ones. The deeper conversations were less awkward than participants predicted and left them feeling more connected and happier than they had expected. Given a free choice of topics, people picked the shallower ones.
What it doesn't show. That a deeper question rescues a stalling conversation, which was never tested — the studies assigned topics up front rather than mid-lull. The finding is about mispredicting depth in general. Conversations were face to face or on video. Nobody has run these experiments on anonymous text chat, so we are extending a finding past the setting it was measured in.
Kardas, M., Kumar, A., & Epley, N. (2022). Overly shallow?: Miscalibrated expectations create a barrier to deeper conversation. Journal of Personality and Social Psychology, 122(3), 367–398 · 10.1037/pspa0000281