You are engaged in conversation with a few friends. Arabella is currently speaking, and you have something to say in response. When do you speak?

Probably a few seconds after her last word. You don’t want to wait too long, otherwise someone might beat you to the punch. And you don’t want to wait too short, just in case Arabella hasn’t finished her thought. And you definitely don’t want to start speaking before her last word—that’s an interruption and just rude.

How many seconds should you wait? What is your “waiting threshold”?

Let’s say your rule is to wait three seconds. You’ve refined that number through trial and error as you grew up in your culture.

But perhaps Brenna was raised in a different culture, a more impatient culture, and only waits two seconds. Then when you both have something to say, Brenna will always speak first. You’ll have to wait until Brenna’s thought is done (remember, interruptions are rude!) and then you can wait your three seconds and speak (provided Arabella doesn’t speak first). Over the course of the conversation, Brenna will dominate you, in the sense that she’ll contribute more to the conversation and be a Social Butterfly while you smile and nod helplessly in the background, collecting your esprits d’escalier, your repository of almost-verbalised thoughts.

Or perhaps Carolyn has learned to wait four seconds. And then (in cases Brenna doesn’t have a response) you’ll dominate her, you’ll always beat her to the punch while she silently resents how she can never get a word in.

This toy model would predict that a conversation among many friends will naturally tend towards being dominated by the two people with the shortest waiting threshold. They can, of course, make a concerted effort to include the others, either by deliberately staying silent to allow the next threshold to kick in, or addressing a wallflower directly. (“So Carolyn, what are your thoughts on the price of tea in China?”) But this is entirely their prerogative. They hold the rest of the conversationalists at their mercy.

In what ways is this model too simplistic? Its most salient limitation is the fixed waiting threshold—in real life people can dynamically adjust their waiting thresholds for the situation. If Carolyn has waited too long without getting a word in, she might steel herself and shorten her threshold. This takes effort, though. She may as well leave the conversation and enjoy herself in another group with higher waiting thresholds. Or she’ll lean against the wall, sipping her gin and tonic, alone in her thoughts.

Or she can bottle her pent-up social frustration until she explodes.

But, simplistic as it is, I think that this toy model has some useful things to say about ways in which conversations can suck. Namely:

A participant’s ability to contribute is determined not by the depth of her idea wellspring, but by the length of her waiting threshold. Carolyn may have the most brilliant ideas, but we’ll never hear them if she’s too polite to interject. Brenna may be speaking far beyond her competence, but as long as her waiting threshold is the shortest, she commands the floor. All this happens without any malice on Brenna’s part. Don’t blame Brenna for making the conversation worse; blame whatever God was in charge when they were handing out waiting thresholds.

And chats among friends are a fairly benign case. In situations where you signal competence by dominating the conversation, this dynamic can turn much more malignant. Group interviews. Politics. Employees angling for a promotion. These are downright toxic.

I never liked when grading was based on class participation. Not only do you need something to say, you also have to speak before the hundred other students in the lecture hall with their own distribution of waiting thresholds. That is to say, immediately.

I like one-on-one conversations. You don’t have to compete with anyone to respond to your interlocutor’s remark. The competition dynamic only kicks in when a third person joins the conversation. The larger the group, the more stifling the dynamic, which may explain why large conversations will tend to fracture into multiple smaller ones. Unless you’re on Zoom, of course, where you can’t do that. Zoom group-calls are awful.

Also, because all this nonsense only applies to real-time conversations, I think intellectual debates are best off conducted in writing.