How Conversations Move

We tend to remember conversations by what was said.

Someone made an argument. Someone told us something important. We disagreed about a decision or finally talked about something we'd been avoiding. When we describe the conversation later, we usually reconstruct it from those pieces of content.

Yet two conversations containing very similar words can go very differently.

A question can invite explanation in one conversation and feel like an accusation in another. A disagreement can make two people more curious about where they differ or send them into increasingly rigid versions of positions they didn't hold quite so strongly when the conversation began.

The words matter.

So does what each exchange makes more likely to happen next.

I've come to think of this as conversational trajectory. I don't mean that conversations contain predetermined arcs or naturally seek a particular destination. The idea is simpler than that. Each response changes the conditions in which the next response occurs, and over several exchanges those changes can begin to accumulate in a recognizable direction.

Suppose I misunderstand something you've said and respond defensively. You now have to decide whether to clarify what you meant, defend yourself against my response, ignore the misunderstanding or end the conversation. If you become defensive too, my original interpretation may suddenly feel confirmed.

Now we have momentum.

Neither of us necessarily intended to create it. Each response may even have made sense locally while the conversation as a whole moved somewhere neither of us wanted to go.

The same thing can happen in a more useful direction. You question an assumption I've made, I become curious instead of defensive, and my curiosity makes it easier for you to explain why the assumption bothered you. Your explanation gives me a more precise question, which gives you a better opportunity to correct my interpretation.

Again, something accumulates.

This is why the direction of a conversation can sometimes tell us something that an isolated sentence can't. A sarcastic remark by itself may mean very little. Five exchanges in which each person becomes slightly more sarcastic tell us more. One interruption might be accidental. A conversation in which one person repeatedly loses the opportunity to finish a thought begins to reveal a pattern.

We should still be careful about interpreting what we notice. A pause doesn't have one meaning. Silence can indicate thought, discomfort, distraction, restraint or simple fatigue. A person asking many questions may be curious, skeptical, confused or deliberately avoiding stating their own position.

Trajectory doesn't solve that ambiguity.

It gives us more observations to compare.

Instead of deciding what a single moment means, we can ask what has been happening across several moments. Are questions becoming more precise or more accusatory? Are both people introducing uncertainty, or is one person increasingly expected to carry all of it? Can either person revise something they said earlier? Are we still investigating the original question, or have we gradually moved into defending ourselves?

Those questions don't require us to know what is happening beneath the surface. They ask what the interaction is actually doing.

Roles can become visible this way too. One person begins explaining while the other asks questions. One reassures while the other expresses doubt. One challenges and the other defends. These positions aren't necessarily identities, and they aren't inherently unhealthy. Conversation needs temporary roles because we can't all perform every function simultaneously.

The interesting part is whether those roles remain movable.

A person who has been explaining may encounter something they don't know and become the learner. Someone who has been asking questions may eventually offer an interpretation of their own. A disagreement may reveal that the person who appeared certain was actually testing an idea rather than defending it.

When those movements remain possible, the conversation has more ways to develop.

Roles can also reinforce themselves. If I respond to every uncertainty with an explanation, you may increasingly occupy the position of the person being instructed. If you resist that position, I may interpret your resistance as evidence that I haven't explained clearly enough and begin explaining even more.

Now the role and the trajectory are interacting.

This is where the role mirror becomes useful. We can ask not only what role each person seems to occupy, but what each response is inviting from the other person next.

The answer may surprise us. What feels like another person's stubbornness may partly be a response to the increasingly narrow set of positions we've left available to them. What feels like our responsibility to keep explaining may partly be a role we've volunteered to occupy.

Or perhaps not.

Changing our response gives us a way to find out.

If I stop explaining and ask a genuine question, what happens? If I acknowledge the part of your criticism that I think is correct before addressing the part I dispute, does the conversation change? If I say that I don't know instead of producing the best answer I can assemble immediately, does uncertainty become easier for both of us to tolerate?

These aren't techniques guaranteed to improve a conversation. Sometimes the other person really does want a fight. Sometimes the disagreement concerns something that shouldn't be softened. Sometimes there are power differences or consequences that make experimentation costly.

Still, a change in response can reveal something about how the previous trajectory was being maintained.

Timing matters here too, although I wouldn't now say that conversations have natural arcs that need to complete. Sometimes we rush toward a conclusion before understanding the question. Other times we continue exploring long after enough information is available to make a decision. A pause can create useful room, or it can become avoidance.

There isn't one correct tempo.

The question is what the timing is doing in this conversation.

That becomes particularly important around uncertainty. We often experience uncertainty as something that needs to be removed, so the person who produces an answer can appear to be moving the conversation forward. Sometimes they are. Other times the answer arrives before the uncertainty has been described well enough to know what we're answering.

The conversation gains resolution and loses information.

The opposite failure is possible too. People can remain indefinitely in exploration because committing to a position would expose it to criticism or require action. Openness can become its own form of avoidance.

So I don't think the goal is to keep conversations open.

The goal is to keep them revisable for as long as revision matters.

That distinction helps explain why agreement doesn't always feel like completion. Two people can agree because one person became tired, because the cost of disagreement grew too high or because they discovered language vague enough to accommodate incompatible interpretations.

They reached the same sentence without necessarily reaching the same understanding.

A disagreement can produce the opposite result. Two people may begin with a vague sense that they see things differently and eventually discover the exact assumption separating them. They still disagree, but the conversation has become more accurate.

That is movement too.

Over longer relationships, these trajectories can repeat often enough to become familiar. A team may learn that disagreement is followed by investigation, or that it is followed by a meeting in which everyone quietly waits for the most senior person to decide. Friends may develop reliable ways to repair misunderstandings. A couple may discover that certain arguments repeatedly move through the same sequence regardless of the subject that started them.

Repeated trajectories become part of the context for later conversations.

Once that happens, people aren't responding only to what is being said now. They're also responding to what they expect usually comes next.

That expectation can be useful. Familiarity allows enormous compression between people who know one another well. It can also make the present conversation difficult to see because we begin responding to the expected next step before it actually occurs.

This is another reason observation matters.

Not because awareness automatically changes the interaction, but because it can separate what is happening from what we expected to happen.

Perhaps we're repeating an old argument.

Perhaps this one is different.

Perhaps the other person just did something that doesn't fit the role we've assigned them.

If we notice that difference, we have at least the possibility of responding to what happened rather than completing the pattern from memory.

That is what I still find useful in thinking about conversational movement. We don't need to imagine an invisible arc beneath the words or a system trying to find balance. We can simply notice that responses have consequences for subsequent responses, and those consequences can accumulate.

A conversation that begins in curiosity can move toward defense.

One that begins awkwardly can become surprisingly open.

A role can harden or loosen.

A misunderstanding can compound or be corrected.

None of those directions is inevitable.

And sometimes the most useful question in the middle of a conversation isn't simply, “What are we talking about?”

It's also:

“Where is the way we're talking about it taking us?”