Chapter 7 — The Fence · The World We Make Together

The Fence

A fence becomes most interesting where it fails to be complete.

One board has warped away from the next. A gate hangs slightly open. Grass grows beneath the lowest rail. A child can see through the gap even when an adult standing several feet away sees only a boundary.

The fence still works.

It tells one household where to mow and another where to stop. It keeps a dog from wandering. It gives privacy to people who did not choose their neighbors and may prefer not to know everything happening on the other side.

Then a ball rolls through.

Someone has to decide whether to call out, reach through, walk around, or wait for the neighbor to notice. The boundary that had quietly organized separation becomes the place where a relationship must occur.

Most fences do not reveal whether the people they divide trust one another.

A high fence may protect a warm friendship from constant intrusion. A low fence may separate people who have not spoken in years. Two neighbors may exchange tools across a barrier they have no desire to remove. Another pair may share an open yard until one dispute makes every missing boundary feel dangerous.

The fence tells us that separation has been organized. It does not tell us what the separation means.

The same holds for many boundaries.

A border can protect a political community or divide a people whose lives cross it daily. A locked door can preserve safety or conceal abuse. A professional role can clarify responsibility or prevent knowledge from traveling. A cultural tradition can sustain belonging while teaching members to distrust those who stand beyond it.

Boundaries are neither inherently just nor inherently hostile. They become moral through what they protect, what they exclude, who controls the crossing, and how much risk falls upon the person asking to pass.

Shared power does not require a world without fences. It requires learning which boundaries can hold difference without turning difference into permanent estrangement.

That work begins with trust.

Trust is often described as though it were a feeling we develop after enough evidence has accumulated.

We meet someone. We observe their behavior. They keep promises. We become confident that they will continue doing so.

That describes part of trust, but only the safest portion of it.

By the time another person’s reliability has been completely established, little trust may be required. We possess evidence. We have routines, contracts, references, records, and consequences. The uncertainty has narrowed.

Trust matters because certainty never becomes complete.

The person may disappoint us. The institution may change. The agreement may be interpreted differently under pressure. The neighbor may return the tool damaged. The group that cooperated yesterday may decide that today’s conflict matters more.

To trust is to place something of value within another person’s range of action: information, time, safety, reputation, a shared plan, the future of a relationship.

Trust does not mean believing the other person is incapable of harm. It means acting without total control over what they will do next.

Trust among similar people often feels easier for this reason. Similarity offers shortcuts. The other person speaks like us, worships like us, votes like us, understands the same references, or belongs to a network in which misconduct will carry consequences. We can predict their behavior because we believe we know what kind of person they are.

Sometimes the prediction is accurate. Sometimes similarity hides danger. People inside familiar groups can exploit confidence precisely because no one expects betrayal. A shared identity can become evidence accepted in place of conduct.

Trust across difference removes some of these shortcuts. We cannot assume the other person sees the situation as we do. We may not understand which promises carry weight in their community, what they fear, what they consider respectful, or what history enters the room with them.

The uncertainty becomes visible. This can make cross-group trust feel less natural. It may also make it more honest.

In Belfast, fences became part of the geography of conflict.

Walls, gates, barriers, and defensive structures divided some predominantly Protestant and Catholic neighborhoods. They were often called peace walls, a name that captured the contradiction built into them. The barriers could reduce immediate contact and provide residents with a sense of protection from violence. They also made division physical.

A wall could prevent a stone or bottle from crossing. It could not produce a shared city.

The barriers belonged to a conflict that involved competing national identities, political discrimination, state power, paramilitary violence, policing, territory, memory, and the constitutional status of Northern Ireland. More than a disagreement separated the communities. People had been killed. Families carried losses. Streets and institutions signaled who belonged where.

Trust could not be summoned by asking everyone to notice their common humanity. The humanity had never been the only question.

People also needed to know whether the police would protect them, whether political institutions would represent them, whether armed groups would stop using violence, whether agreements would be honored, and whether the identity of one community could be recognized without erasing the identity of another.

The fence had a history. Anyone asking people to cross it had to take that history seriously.

On April 10, 1998, the British and Irish governments and most of Northern Ireland’s major political parties reached what became known as the Belfast or Good Friday Agreement.

The agreement followed decades of conflict and years of negotiation. It created a power-sharing framework for governing Northern Ireland, established institutions linking Northern Ireland and Ireland, created additional structures for cooperation between the British and Irish governments, and addressed questions involving rights, policing, security, prisoners, weapons, and constitutional status.¹

The agreement did not resolve the conflict by deciding which national identity was correct. It did something more difficult: it built political institutions around the continued existence of incompatible aspirations.

Unionists could remain committed to Northern Ireland’s place within the United Kingdom. Nationalists could remain committed to a united Ireland. Neither community had to surrender its identity as the price of entering the political arrangement.

The agreement recognized that Northern Ireland’s constitutional status would rest upon consent. It also affirmed the right of people born there to identify as British, Irish, or both, and to hold British and Irish citizenship accordingly.²

This was not agreement through sameness. It was agreement about how difference would be carried.

The parties did not discover that their earlier conflict had been based upon a misunderstanding. They did not merge their histories into one account. They did not begin trusting every rival leader, institution, or armed organization. They created rules through which people who remained deeply divided could take political risks without giving the other side complete control over the future.

Trust entered through structure.

This can sound less inspiring than reconciliation. A handshake suggests that hostility has softened. An institutional design suggests committees, voting rules, jurisdictions, guarantees, and procedures. But where trust is weak, procedure can carry relationships farther than goodwill alone.

The Good Friday Agreement established an Assembly and a power-sharing Executive for Northern Ireland. It also created North-South institutions linking Northern Ireland and Ireland and East-West institutions linking the British and Irish governments and the wider islands.³

These structures acknowledged that no single center contained the whole political reality. Northern Ireland had its own institutions; its relationship with Ireland mattered; its relationship with Britain mattered. The communities within it could not be governed sustainably by pretending one of those relationships was imaginary.

The design did not eliminate power. It attempted to prevent one political identity from treating victory as permanent ownership.

This is one of the central tasks of shared power.

A group must be able to act without converting temporary authority into domination. It must be able to lose without interpreting the loss as political extinction.

Trust becomes possible when people know that cooperation does not require surrendering every protection against the other side.

The fence may remain. But a gate appears, and neither community alone holds the only key.

On May 22, 1998, voters in Northern Ireland approved the agreement in a referendum. More than 71 percent of those voting supported it. Voters in the Republic of Ireland approved related constitutional changes on the same day.⁴

The result mattered.

It gave democratic authorization to an agreement negotiated by leaders, parties, and governments. Ordinary people were not merely asked to endure a settlement produced above them. They were asked to mark whether the new arrangement could claim public consent.

Yet the ballot could not reveal the same meaning in every vote. One person may have believed deeply in reconciliation. Another may have disliked important parts of the agreement but feared the return of violence more. Someone may have trusted a party leader’s judgment. Someone else may have voted for the possibility of an ordinary life in which children reached adulthood without the same conflict.

The result compressed these motives into Yes and No. The agreement passed. Trust did not become universal the following morning.

Political institutions still had to be formed. Weapons had to be addressed. Policing had to change. Prisoner releases produced pain and controversy. Parties had to sit within arrangements that included people associated with movements they regarded as responsible for violence.

The referendum authorized a structure. People still had to inhabit it.

This is where trust differs from optimism. Optimism expects the future to improve. Trust gives another person or institution room to affect the future despite the possibility that it will not.

A unionist entering a power-sharing institution did not need to believe nationalism would disappear. A nationalist did not need to believe unionism had become temporary or insincere. Each side needed enough confidence that the political process could carry disagreement without returning immediately to force.

That confidence could not rest only upon the moral character of the people across the table. Leaders change. Parties divide. Public moods shift. A durable peace cannot depend upon everyone continually liking the people required to sustain it.

Trust needed law, verification, institutions, electoral support, external guarantees, and consequences.

This does not make it less human.

Promises become more trustworthy when people have built ways to survive their failure. A contract identifies expectations so the relationship does not depend upon identical memories. A shared process does not eliminate conflict; it prevents every conflict from becoming a test of whether the relationship still exists.

The phrase building trust can imply a wall constructed one brick at a time—a promise kept, a conversation held—until the wall is high enough to hold the relationship.

But trust may look more like a gate. It opens and closes. It may be wide enough for one kind of exchange and too narrow for another. A person may trust a neighbor with a spare key but not with private financial information. A community may trust another group on emergency services while remaining divided over schools, symbols, or constitutional identity.

Trust is usually specific before it is general: we trust someone to do something under certain conditions. This specificity matters because demands for trust can otherwise become demands for vulnerability without evidence.

A powerful institution says the public should trust it. A majority asks a minority to trust the fairness of the process. A leader asks employees to trust that difficult changes are necessary.

But who is being asked to accept the risk? Who controls the information? Who can recover if the promise fails? Who possesses an appeal?

A person with little power may appear distrustful because they insist upon rules, witnesses, written commitments, or outside review. Those protections may be the conditions that make cooperation possible.

The asymmetry becomes clearer when groups carry different histories into the same encounter.

Suppose a public institution invites a marginalized community into a conversation.

Officials say that everyone should enter with openness and assume good faith. The invitation appears equal. Each person is expected to listen, speak honestly, and avoid prejudging the other side. But the risks may not be equal.

Officials retain their salaries, titles, records, legal authority, and control over implementation. Community members may be asked to describe injuries, reveal vulnerabilities, or identify themselves publicly without knowing whether anything will change. One side risks an uncomfortable meeting; the other may risk retaliation, exposure, disappointment, or the reopening of harm.

Calling upon both sides to trust equally can conceal how unequally trust has been made necessary.

Research on intergroup contact reflects part of this problem. Large reviews have found that contact between groups often reduces prejudice, particularly when it includes conditions such as cooperation, common goals, institutional support, and equal status within the encounter. But the relationship between contact and reduced prejudice has not always operated identically for majority- and minority-status groups. For people whose group has less power, an encounter may carry threats and meanings that the more powerful group does not experience in the same way.⁵

Contact is not a solvent poured onto difference. People can meet and grow more suspicious, work near one another while remaining separated by rank or fear, or share a room while one group retains authority over the consequences.

Trust grows not simply from proximity but from what people are permitted to do together.

Shared work can matter for this reason.

Two groups may begin without agreement about identity, history, or motive. But they may need to maintain the same road, protect the same water supply, operate the same school, or sustain the same political institution. The practical task creates a limited field of interdependence. Each side knows something the other does not. Neither can complete the work alone.

Cooperation does not guarantee affection. It can still produce a new kind of knowledge. The person across the boundary becomes more than a category: they arrive late, misunderstand, keep one promise and resist another, reveal expertise, become frustrated for reasons that begin to make sense.

This ordinary accumulation matters because stereotypes are efficient. They convert a large group into one predictable person. Direct relationship makes that compression harder to sustain—not impossible. A person can treat every generous action as an exception and every failure as proof of the group.

But repeated cooperation can create something arguments alone rarely produce: a record. We know how this person behaved when the plan failed, whether they returned after the disagreement, which promises survived pressure.

Trust begins to move from identity toward conduct.

The Good Friday Agreement built institutions for precisely this kind of repeated political encounter. It did not ask the parties to settle the ultimate constitutional question forever, but to govern that question through democratic consent rather than force, across internal, North-South, and British-Irish forums.⁶

This did not prevent breakdown. Institutions were suspended. Disputes over weapons, policing, identity, rights, language, flags, and governance continued. Political relationships repeatedly strained and sometimes failed to function.

The existence of conflict after an agreement does not by itself prove that trust failed. Trust includes a route back after rupture: an institution to reconvene, a rule to govern the dispute, outside parties who can assist, and earlier commitments that remain available when confidence collapses.

The shared framework becomes a memory larger than the immediate crisis. Institutions matter even when they appear slow, procedural, or disappointing: they preserve a place where disagreement can return without beginning from nothing.

The gate may close. The path toward reopening it remains visible.

Physical barriers reveal the distance between political trust and lived trust.

Northern Ireland’s Department of Justice has maintained a program concerned with interface barriers and has described their reduction or removal as dependent upon confidence and relationship building within affected communities. Its published information continues to list gates at several interfaces that open and close on daily schedules.⁷

This is an extraordinary ordinary detail.

A gate opens in the morning and closes at night. People cross to work, shop, visit, or move through the city. Then the boundary hardens again according to the clock.

The schedule carries two truths: life has become shared enough for passage; fear remains strong enough for closure.

An outsider may look at the wall and ask why people do not simply remove it. Continued division can appear to be an irrational refusal to move forward.

Residents may see something else. The barrier may be ugly, limiting, and resented. It may also stand between a home and a place from which violence once came. A government promise that the wall is no longer necessary cannot erase the memory through which the wall became reasonable.

Removing a barrier without the confidence of the people living beside it can turn reconciliation into another act done to them. Keeping every barrier forever can allow fear to become inheritance. The work lies between these failures.

Trust cannot be demanded from the people who bear the greatest risk. It must be made possible. That usually involves small changes before dramatic ones: a gate stays open longer; a shared route becomes ordinary; representatives establish a way to respond when an incident occurs; rumors can be checked before they gather into retaliation; young people encounter one another through a task that gives each group real standing; residents influence what happens to the barrier rather than receiving a completed plan.

None of these acts resolves the whole conflict. They change the evidence available for the next act. Trust often grows at this scale—not as faith in humanity, but as a revised estimate of what may happen if we take one more step.

The language of trust can become morally dangerous when it forgets betrayal.

Someone says the relationship cannot heal unless the injured person lets go. A leader says employees must stop dwelling on the past. A government asks communities to trust reformed institutions without making records available or accepting responsibility for earlier harm.

The future is used to silence the evidence required to enter it safely.

Trust cannot be separated from truth.

A person who conceals what happened and asks to be trusted is not building reconciliation. They are asking another person to carry the uncertainty created by their concealment.

Shared power requires that information move. Who knew? Who decided? Who benefited? Who was harmed? Which commitments failed? What has changed beyond the language describing the change?

These questions can feel accusatory because they resist the emotional relief of beginning again. But trust built upon amnesia is fragile. The unspoken past returns whenever a new event resembles it.

Truth does not guarantee reconciliation. It prevents reconciliation from depending entirely upon denial.

Trust also requires the possibility of refusal.

This may sound contradictory. A relationship appears strongest when neither person threatens to leave. A political arrangement appears stable when every participant accepts it.

But trust without meaningful refusal can become submission. The person who cannot say no has not necessarily entrusted anything. They may simply lack alternatives.

A worker who shares information because the employer demands it is complying. A citizen who cooperates because the state can punish refusal may be prudent. A community that enters negotiations because every resource has been conditioned upon participation may have little genuine choice. These actions can still contribute to shared work. They should not be romanticized as trust.

Trust becomes more visible when a person could withhold cooperation and chooses not to.

Power differences matter here. The more powerful side can often exit with less cost: postpone a meeting, replace a supplier, continue governing, hire private services. The less powerful side may depend upon the relationship for access to work, safety, recognition, infrastructure, or rights.

Both parties may need one another. They do not necessarily need one another equally. A trustworthy arrangement recognizes this asymmetry and builds protections around it. It does not call dependence mutuality merely because both sides signed.

The fence between two yards may belong legally to one household. Both households live with it.

One side may have paid for the materials and possess the right to replace it. The other sees the structure from the kitchen window every day. A decision that belongs formally to one person enters the life of another.

Shared power often begins in this kind of overlap. The rights are real. So is the relationship.

The owner may say, “I am allowed to build it.”

The neighbor may say, “I have to live beside it.”

Neither statement fully answers what should happen.

Trust grows when formal authority makes room for affected knowledge without pretending the authority has disappeared. The owner asks what the neighbor sees. The neighbor recognizes that consultation does not create a veto over everything. They discuss height, placement, cost, maintenance, privacy, sunlight, pets, and the tree near the boundary. Perhaps they still disagree. But the disagreement now contains more of the situation.

This is not a grand act of reconciliation. It is a small practice of shared power: the person with the right to decide treats the other person’s experience as morally relevant; the person without the final decision enters as more than an obstacle. Trust is formed not because they agree but because neither uses their available power to make the other disappear.

Trust beyond similarity does not ask us to believe that difference is always enriching.

Some differences are painful. Some beliefs deny another person’s dignity. Some goals cannot be pursued together. Some people use the language of pluralism to protect conduct that places others in danger.

A boundary may need to be enforced more strongly, not removed.

The challenge is distinguishing protection from avoidance. Does the fence prevent a real harm? Does it preserve privacy necessary for relationship? Or does it allow us to treat people on the other side as permanently unknowable? Does the gate have a process? Can anyone cross? Who decides when it closes? Are the criteria clear? Can the person excluded appeal?

A just boundary explains itself. An unjust one often presents itself as nature.

The Good Friday Agreement did not create peace by discovering a position halfway between British and Irish identity. It created a political structure in which neither identity had to vanish for shared institutions to exist.

That is a different model of cooperation. Compromise often means each side surrendering part of what it wants until a tolerable middle appears. Sometimes that is necessary. The deeper achievement may be to redesign the relationship so that incompatible commitments no longer require constant domination.

One person may remain British, another Irish, a third both. The shared political order must make room for those identities without allowing any one of them to determine everyone else’s civic worth.

Shared power does not promise an end to disagreement. It promises that disagreement need not end the shared world.

The ball remains on the other side of the fence.

Someone eventually notices.

The neighbor picks it up and looks toward the gap. They could leave it there, throw it back, carry it around, or wait until the child comes to ask.

The decision is small enough that no one will remember it as history. Yet the action enters the relationship.

The ball is returned carefully. The gate is opened. A name is learned. The next time something crosses the boundary, neither person begins from the same uncertainty.

Trust has not removed the fence. It has changed what the fence means.

The boards still mark a difference. The gate still belongs to someone. The possibility of disappointment remains. But the boundary no longer tells the whole story of the people standing on either side.

A shared world begins when we stop treating the fence as proof that nothing valuable can pass through it—not when every fence comes down.

See Belfast Agreement (Good Friday Agreement), 10 April 1998, including the multi-party agreement and British-Irish agreement on power-sharing, North-South and East-West relations, rights, security, policing, prisoners, and weapons. ↩
See the Belfast Agreement’s consent principle regarding Northern Ireland’s constitutional status and recognition of the right to identify and be accepted as British, Irish, or both. ↩
See the Belfast Agreement: Strand One (Northern Ireland Assembly and Executive); Strand Two (North-South institutions); Strand Three (East-West institutions). ↩
See Northern Ireland Office / Electoral Office for Northern Ireland results for the 22 May 1998 referendum on the Belfast Agreement, in which more than 71 percent of Northern Ireland voters supported the agreement. ↩
See Thomas F. Pettigrew and Linda R. Tropp, “A Meta-Analytic Test of Intergroup Contact Theory,” Journal of Personality and Social Psychology 90, no. 5 (2006): 751–83. ↩
See the Belfast Agreement’s continuing institutions for internal governance, North-South cooperation, and British-Irish relations under the consent principle. ↩
See Northern Ireland Department of Justice, “Department of Justice Interface Programme,” on physical barriers, confidence-building, and operational gate schedules. ↩

The Fence

Footnotes