Trust is often treated as something that can be installed.

An organization changes leadership, announces a policy, redesigns its public face and explains what people should now expect. When trust doesn't follow, the obvious response is to improve the message. Explain the change more clearly. Demonstrate good intentions. Find better language for whatever the institution is trying to become.

Sometimes communication really is the problem.

Other times we're asking communication to carry something it can't carry by itself.

I can tell you that someone is trustworthy. I can explain why I trust them, introduce you personally and give you examples from twenty years of knowing them. All of that gives you information you didn't have before, and my own relationship with you may give the information additional weight.

What I can't give you is twenty years of having known them.

That distinction seems obvious between people, but it becomes easier to lose when we're talking about organizations. Reputation, credentials, policies, endorsements and branding can all provide evidence about what someone should expect. They can reduce uncertainty enough for a relationship to begin.

They aren't the relationship.

Trust develops through what happens afterward.

Someone says they'll do something and does it. Something goes wrong and we see how they respond. They make a mistake, acknowledge it and repair what they can. Or they avoid responsibility, change the story and leave somebody else carrying the cost.

One event rarely settles the matter. What accumulates is a history from which expectations become increasingly reasonable.

That's why repeated encounters matter so much. In a small town, a workplace, a business relationship or a long collaboration, today's behavior doesn't disappear when the interaction ends. There is likely to be another encounter, and both people may arrive carrying some memory of the previous one.

That changes the environment.

A business owner who mistreats a customer may encounter that customer at the grocery store next week. A newspaper that gets something wrong may hear about it from the person involved and then cover the same institution again six months later. Someone who makes a promise inside a long-running collaboration doesn't begin from zero the next time another promise is made.

The future gives the present consequences.

That doesn't automatically produce trust. Repeated interaction can produce distrust just as efficiently. A long history of broken promises is still a history.

What repetition provides is evidence.

This is one reason I think trust and familiarity are easy to confuse. Seeing someone repeatedly can make them feel familiar without making them reliable. Shared history can create loyalty that survives long after the evidence supporting it has changed. Communities can protect people who shouldn't be protected precisely because relationships are dense enough to discourage contradiction.

Proximity isn't virtue.

It simply makes certain kinds of feedback harder to avoid.

Distance changes the available evidence, but it doesn't eliminate it. I trust plenty of things built or operated by people I will never meet. I don't personally know the engineers who designed an aircraft before I board it, and I don't need a relationship with everyone involved in processing a payment before using a credit card.

Large systems developed other ways to carry evidence across distance.

Standards matter. Audits matter. Contracts matter. Professional reputations matter. Warranties, regulation, independent testing and a history of reliable operation can all make cooperation possible among strangers.

In some cases, the point of a good system is to reduce how much personal trust is required.

I don't need to trust a merchant's memory if I have a receipt. I don't need to trust that somebody will remember what we agreed to if the agreement is recorded. I don't need to know the person operating a remote server if the service has demonstrated sufficient reliability and I have a workable recovery plan if it fails.

That isn't a lesser form of human relationship.

It's a different solution to the problem of uncertainty.

The distinction matters because “trust must be earned” can otherwise become an excuse for closed circles. If only people with long shared histories can be trusted, newcomers begin with a disadvantage they cannot overcome except by waiting. Outsiders become inherently suspect. Familiarity starts masquerading as evidence.

A healthy community needs ways for new trust to begin.

Reputation can help with that. So can an introduction from someone both parties know. Clear rules can reduce the amount of trust required during the early stages of a relationship. Small commitments can create opportunities to produce new evidence without requiring either side to assume enormous risk.

Trust doesn't have to appear all at once.

It can have a beginning.

This also changes how I think about institutional attempts to restore trust. A campaign isn't necessarily useless. If something has genuinely changed, people need some way to learn about it. An institution that quietly fixes a problem while continuing to communicate as though nothing happened may never give anyone reason to reconsider an older judgment.

The message can open the door.

What happens afterward has to support it.

If an organization announces greater transparency and then becomes evasive when the first uncomfortable question arrives, the announcement becomes evidence too. If leadership promises a different relationship with customers and the next dispute is handled exactly as disputes were handled before, people have learned something more useful than the slogan.

The problem wasn't that communication failed to create trust.

Communication created an expectation that subsequent behavior failed to support.

The reverse can happen as well. An organization may communicate badly while behaving reliably. People who interact with it repeatedly may develop considerable trust despite awkward messaging because experience keeps giving them reasons to update their judgment.

That doesn't make communication irrelevant.

It puts communication in its proper place.

Trust also doesn't require vulnerability to be perfectly symmetrical. Relationships rarely distribute risk equally. A patient and physician, employee and employer, customer and bank, source and journalist all enter relationships with different powers and different exposures.

What matters is whether those asymmetries are understood and whether there are meaningful consequences when the more powerful party abuses them. Sometimes that comes through the relationship itself. Sometimes it comes from rules, institutions or outside accountability.

Again, different mechanisms can carry different parts of the problem.

This is why I still think there is something important in saying that trust cannot be imported. The phrase just needs a boundary around it.

Evidence can be imported.

Reputation can arrive before the person does.

A trusted friend can give someone the benefit of an introduction.

Institutions can create rules that make initial cooperation safer.

A history of reliable behavior can become known far beyond the people who directly experienced it.

All of those things can give trust somewhere to begin.

What they can't provide is the history that hasn't happened yet.

That still has to be enacted.

For a newspaper, this matters more to me than any declaration about being trusted. We can describe our standards, publish corrections, explain how something works and make our processes more legible. Those things provide information people can reasonably use when deciding what to expect from us.

Then we have to publish next week.

And the week after that.

We will get things right. We will sometimes get things wrong. People will see what we choose to cover, what we miss, whether we correct mistakes, whether our relationships with advertisers change our reporting, whether old stories remain available and whether we're still around when something that began two years ago finally reaches its conclusion.

That history will say more than the claim ever could.

Trust isn't something an institution possesses and then distributes to everyone who encounters it. It's an expectation people form from whatever evidence is available to them, and different people can reasonably arrive at different conclusions from different histories.

We can make that evidence easier to see.

We can make cooperation less dangerous while trust is still forming.

We can behave in ways we hope will remain defensible when people encounter us again.

What we can't do is announce the conclusion on their behalf.

Trust has to leave room for the other person to decide whether we earned it.