What does it actually mean to agree?

Two people say: I agree. Do they have the same representation in their heads? Perhaps. But perhaps they understood the words differently. Used different evidence. Applied different rulers. Had different reasons. Or arrived at the same answer by completely different routes. Agreement can be useful. But agreement isn't proof of shared understanding. And disagreement may tell us more than we first think.

Two receivers, two representations

We don't receive the world directly and reproduce it perfectly. Each receiver brings something to the encounter. Different experience. Different expectations. Different models of normal. Different resolution. Different priorities. Different information. Different resources. So two receivers can encounter apparently the same thing and construct different representations of it. That needn't mean either receiver has failed. Different receivers can reasonably see different things. The interesting question becomes: Why do we see this differently?

What exactly do we disagree about?

Imagine: A: That's too risky. B: No, it isn't. It looks like a disagreement about risk. But look more closely. Perhaps they agree about the probability of something going wrong but disagree about the consequence. Perhaps they agree about both but are using different thresholds for what counts as acceptable. Perhaps one is thinking about today and the other about the next ten years. Perhaps one has information the other doesn't. Perhaps “too risky” means something slightly different to each of them. Before trying to resolve the disagreement, it can be useful to ask: Where exactly is the difference?

What unit? What ruler?

Many disagreements that appear to be about facts are partly disagreements about how those facts are being judged. Is this good? Is that fair? Is this expensive? Is that safe? Which option is better? Better according to what? Two receivers can agree on the underlying information and still reach different judgments because they're using different: • units • rulers • thresholds • priorities • periods of time Making those visible may not resolve the disagreement. But it can give us a better representation of what we're actually disagreeing about.

Disagreement gives us difference

Our natural response to disagreement is often to eliminate it. Who is right? Who is wrong? Which representation should survive? Sometimes we need to decide. But before eliminating the difference, it may be worth allocating some resources to it. Because the other receiver may have something you don't. They may have noticed a signal you missed. They may have greater resolution in one part of the problem. They may be using a ruler you hadn't considered. They may simply be wrong. But you don't know that merely because they disagree with you. So disagreement can itself become a signal: Wait a second. Why are we seeing this differently?

Agreement can hide difference

Disagreement is visible. Agreement can be more deceptive. Imagine: Do you understand? Yes. We have agreement. But what exactly has been agreed? The sender may think: They now understand what I meant. The receiver may mean: I have constructed something that makes sense to me. Those aren't necessarily the same thing. This matters particularly in communication. A repeated instruction, a nod, an “OK” or even successful action may provide evidence of shared understanding. But none gives us direct access to another receiver's representation. Agreement can be real without being complete.

Shared understanding doesn't require

identical understanding

Perhaps the goal isn't to make two representations identical. That may be impossible. You remain you. I remain me. We bring different histories, knowledge, experiences and perspectives. For many purposes, we don't need identical representations. We need them to overlap enough for what we're trying to do together. That “enough” matters. Two pilots coordinating an aircraft may require a different degree of shared understanding from two friends deciding where to have lunch. So shared understanding, like GOOD ENOUGH, is relative to purpose and consequence. The useful question may be: Do our representations overlap sufficiently for what we are trying to do?

Another receiver can increase our resolution

Suppose you see: A and I see: A plus something else. If we communicate effectively, you may gain access to a difference that wasn't previously available to you. And the reverse can happen too. This is one reason groups can sometimes do things individuals cannot. Not because several people automatically produce a better answer. But because different receivers can make different differences available to one another. Your representation can challenge mine. Mine can challenge yours. Perhaps neither survives unchanged. We may construct something neither of us brought into the conversation.

But difference isn't automatically valuable

More disagreement isn't necessarily better. Another perspective isn't automatically useful. A receiver can be mistaken. Poorly informed. Dishonest. Confused. Or simply contributing a difference that doesn't matter for the current purpose. Difference still requires judgment. Does this difference deserve more of our limited resources? Sometimes yes. Sometimes no. Sometimes: I understand why we disagree. I still disagree. That can itself be a perfectly legitimate outcome. Understanding does not require agreement.

Agreement can be a decision

At some point a group may need to act. We have explored the differences. Shared information. Questioned assumptions. Perhaps changed our representations. But we still don't completely agree. Now finite resources matter. Continuing the discussion has a cost. Sometimes we need to decide: This is good enough for us to proceed. That doesn't mean every receiver now believes the same thing. It means the group has reached a sufficient level of agreement for the present purpose. And sometimes the responsible decision is the opposite: We don't yet understand one another well enough to proceed. WAIT remains an option.

Could we get better at disagreement?

Perhaps the aim isn't to eliminate disagreement. Nor to celebrate it automatically. Perhaps we can become a little better at using it. When someone disagrees: What are they seeing that I'm not? What am I seeing that they aren't? Are we working from the same information? What models of normal are we bringing? What differences are consequential to each of us? What unit? What ruler? What period of time? And perhaps the most difficult question: What evidence would cause me to revise my representation? Build the model. But remain open to questioning the model. Another receiver may give you the difference that allows you to do exactly that.

Agreement isn't the end

We agree. Good. Perhaps we have sufficient shared understanding to move forward. But agreement shouldn't make the representation immune from further evidence. New information may arrive. The environment may change. Another receiver may see something neither of us saw. What was GOOD ENOUGH may stop being good enough. So perhaps: Don't treat disagreement as failure. Don't treat agreement as proof. Use both as information. Then decide what, if anything, deserves to happen next.