A note before we begin: the story that runs through this article is a composite. The people, the room and the clause are invented to show how confidential dialogue works. They describe no real person, organization or event.
The talks had been stuck for three days on a single clause.
Both delegations had read it so many times that the words had stopped meaning anything. One side could not accept it as written. The other could not change it without looking weak. So each morning they sat down, repeated the positions they had stated the day before, and adjourned for lunch.
On the third afternoon, during a short break, a senior delegate found a junior delegate from the other side beside the coffee table. Nobody else was close enough to hear. The senior delegate asked a quiet question: “If the clause named the outcome instead of the method, would that change anything for you?”
The junior delegate thought about it. “Possibly,” came the answer. Then, lowering his voice: “Strictly between us.”
That small phrase is the subject of this article. Pay attention to what it asks for, because the whole story turns on it.
The Sentence Nobody Can Quote
Ask experienced negotiators when a deal really turned, and you will rarely hear a signing date. You will hear about a doorway, a long walk between sessions, a car waiting at the curb. The most useful sentence in diplomacy is often one that nobody can quote.
Notice what the senior delegate did at that coffee table. She did not make an offer. She asked a question, and the question cost her nothing. If the answer had been no, nobody would ever know she had asked. If the answer was maybe, a new road had just opened.
Asked in the formal session, the same question would have been a concession. Every delegate in the room would have heard it, written it down and reported it. Her side would have been seen offering to move, and the other side would have pocketed the offer and asked for more. The words are identical in both settings. Only the room is different, and the room changes what the words mean.
This is the first idea to hold on to. Private rooms exist so that people can explore a position without owning it.
What the Research Calls It
Scholars have studied this kind of conversation closely. Anthony Wanis-St. John calls it back channel negotiation and describes it as officially sanctioned talks, held in secret between the parties to a dispute, that run beside the visible process or replace it [1].
One word in that definition matters a great deal: sanctioned. A back channel is not a rogue conversation held by someone acting alone. Someone with authority stands behind it, and that is what gives the words spoken there real weight.
He adds a useful observation about who tends to speak in these rooms. Back channel negotiators, he notes, tend to stand closer to the top decision makers than their front channel counterparts. That closeness lets them explore a wider range of options with real authority [1]. It also explains why a quiet question at a coffee table can matter more than a page of formal text. The person asking may be able to say yes.
The setting changes how people listen as well. In public, every listener is partly an audience member, working out how a remark will look to everyone else. In private, a listener can simply listen. And a position that shifts in private looks like learning, whereas the same shift in public looks like retreat.
Why Public Words Cost So Much
Back to our story. Why did the two delegations spend three days repeating themselves? Because everything said in the formal room was said to two audiences at once.
Robert Putnam described this clearly. In his account, a negotiator plays two games at the same time: one across the table, and one with the people back home whose approval, formal or informal, any agreement must finally win [2]. A public statement speaks into both games in the same breath. The audience across the table wants to hear flexibility. The audience at home wants to hear firmness. Rarely do they want the same sentence.
Public statements also set expectations. A position announced loudly becomes the yardstick for every compromise that follows, so the louder the announcement, the bigger the apparent loss when a compromise finally arrives.
David Stasavage took this question seriously by comparing two ways of bargaining. In open bargaining, the public can observe every offer. In closed bargaining, the public sees only the final outcome [3]. He weighs the value of public scrutiny against its possible costs, and he treats the question of when transparency is desirable as one for careful analysis, not slogans [3]. That is a fair way to think about it. Openness is not good or bad in itself. The useful question is what each setting makes possible.
There is one more effect, and our two delegations were living it. A position stated publicly invites a public reply, and public replies are written for audiences, not for counterparts. Two people who could have found a middle path end up trading statements aimed past each other. Neither wanted to harden. The setting did it for them.
The Three Promises Hidden in One Phrase
Now return to the coffee table and the phrase “strictly between us.”
What did the junior delegate promise? Perhaps he meant that the senior delegate could repeat the idea so long as she never said where it came from. Perhaps he meant that she could not repeat it at all. Perhaps he meant she could act on it but never mention it. He said four words, and the listener could reasonably hear three different promises.
Most breaches of confidence begin in exactly this gap. Two people leave the room carrying different promises, and neither thought to check.
It helps to separate the promises. Think of them as three levels.
On the record is the simplest. Everything said can be repeated, and it can be attributed to the speaker by name.
Under the Rule refers to the best known framework in the field. The Chatham House Rule lets participants use what they learn, but it forbids them to disclose who spoke, whom the speaker represents, or who else attended [4]. This is not secrecy. The information can travel. Only the source stays protected. Any group can adopt the Rule for its own meetings, and its publisher says it helps people discuss sensitive issues and judge an idea on its merits, whatever its origin or the identity of the speaker [4].
Off the record is the strictest of the three. The same publisher notes that when the Rule feels too loose, a meeting can be held off the record instead [4]. Here is the catch. The phrase has no single fixed meaning, so it is the promise that most needs spelling out.
| Promise | What may travel | What stays protected |
| On the record | Everything, with the speaker named | Nothing |
| Under the Rule | The substance of what was said | The identity and affiliation of every speaker and participant |
| Off the record | Only what the organizer and speaker agree | Everything else |
In our story, nobody spelled it out. Hold that thought, because it comes back.
Two Rooms, One Negotiation
Serious negotiations often run in two rooms at once. The first is the acknowledged table, with delegations, agendas and statements for the press. The second is smaller and harder to see. Scholars call them the front channel and the back channel.
Wanis-St. John observes that back channel negotiators seem to reach agreement often in cases where their front channel colleagues fail. He argues that theorists and practitioners should look harder at how the two rooms differ [1]. He also notes that sometimes an unofficial third party carries messages between the sides [1].
Dean Pruitt adds a helpful map of the roles a back channel can play. It can be used to explore whether formal talks are even feasible. It can help when formal talks become deadlocked. It can also stand in for formal talks altogether, and in his account intermediaries, sometimes whole chains of them, carry the messages [5].
Here is a simple way to hold the picture. The front room produces the record. The back room produces movement. A record without movement is theater, and movement without a record is rumor. Each room needs the other.
The craft lies in the passage between the two rooms, and that passage has one practical rule: never use the back room to promise what the front room is not authorized to deliver. A private understanding that outruns authority does not speed up the formal process. It damages it. The first time the formal table cannot honor what the private table seemed to promise, both rooms lose their value at once.
Back to our story. The senior delegate carried the “maybe” back to her own team. She did the right thing: she told her sender, the people who had given her the mandate to negotiate, exactly what she had asked and exactly what she had heard. She made no promises and she offered no concessions. She had only learned that a door existed.
The Door Before the Door
Some of the most valuable private conversations happen before a negotiation has even begun. Picture two parties who cannot yet be seen speaking to each other, or who have not agreed on what to call their talks. They still need a way to find out whether talking is worth attempting.
Pruitt describes this early stage in practical terms. In his account, a secret back channel brings flexibility, provides political cover, costs little, does not require formal recognition of the other side, and allows contact with parties who do not yet meet the preconditions set for negotiation [5].
Aroop Mukharji and Richard Zeckhauser make a point that is easy to miss. They observe that comparatively few studies look at how hostile parties begin negotiating at all, or at the dangers that lurk during the waiting period before talks start. They argue that back channels give adversaries a critical route toward cooperative solutions [6].
The reasoning is plain once you see it. A public approach is costly, because a refusal becomes a public rejection. A private approach is cheap, because a refusal stays a private answer. Nobody loses standing. So the real work of the early stage is to lower the cost of asking.
The Day the Promise Broke
Now we reach the part of the story where “strictly between us” does its damage.
Two days after the coffee table conversation, the junior delegate’s team circulated a summary of the week for their own leadership. It was meant to be internal. In it, a careful writer, trying to be thorough, recorded that “a senior delegate on the other side had floated rewording the clause around outcomes, and our own delegate had signaled openness.”
Notice what happened. The idea survived. The attribution did not. The summary named a side and an individual, and it went to a wider group than the junior delegate had imagined.
When the summary reached the other side, the senior delegate learned that her quiet question was now a documented position. Her own team asked why she had offered a concession. She had not. She had asked a question. But the written record made the question look like an offer, and the junior delegate suddenly looked like someone who had talked too freely.
Nobody in this story acted with bad intent. The senior delegate asked honestly. The junior delegate answered honestly. The summary writer was trying to be useful. Yet trust took a blow, and the two delegates stopped speaking privately for the rest of the week. A single sentence naming the promise would have prevented all of it.
This is how confidence breaks in real life: through a small, well meant lapse, not through villainy. One person assumes a promise that was never made. Someone writes a line with a name attached. A sentence is repeated to someone who was not meant to hear it.
What Each Person in the Room Owes
If you take one practical lesson from this article, take this one. Confidential dialogue works only when everyone accepts a duty. Most breaches come from forgetting one of those duties, not from bad intent.
The speaker owes accuracy. A private remark is not a license to say what one would never defend. Speakers who use closed rooms for claims they could not stand behind soon find the doors closing.
The listener owes care. What is heard in confidence may inform a decision, shape a question or change a view. It may not be quoted or traced to its source, and that is exactly the line the Rule draws [4].
The convener owes clarity. Someone has to say, before anyone speaks, which promise applies. If the organizer stays silent, each person will quietly assume the version that suits them best.
The sender owes a mandate. The sender is the person or office standing behind a negotiator. Without a clear mandate, a negotiator can only offer opinions. With one, a negotiator can offer options, and options are what private rooms are for.
Everyone owes a plan for the day after. Sooner or later someone will have to explain a private conversation to a public audience. The explanation must be truthful, and it must reveal nothing given in confidence. If both cannot be done, the conversation should not have happened in that form.
Here is the cleanest way to say it. Confidentiality withholds. Deception misleads. The first is a normal part of serious work. The second ends it.
Where Off the Record Reaches Its Limits
A fair account of confidential dialogue has to name its costs, and there are real ones.
Stasavage starts from the widely accepted benefit of openness, which is public oversight, and asks what openness might cost [3]. The reverse question applies to closed rooms. A private channel that never returns to the public process becomes a place where responsibility goes missing.
The remedy is sequence. Early offers, drafts and exploratory questions tend to do better in a room that closes. Outcomes, commitments and reasons belong in the open. A process that explores privately and then explains publicly can defend itself, because the explanation is genuine. A process that announces its hopes loudly and settles its terms quietly will be accused, fairly or not, of hiding the wrong things.
Whatever was learned privately must also become a commitment that can stand in public. Otherwise it remains a pleasant conversation without consequence. Putnam’s two games explain why: an understanding that cannot win approval from the audience behind the negotiator is not yet an agreement [2].
Pruitt adds a warning that deserves attention. Heavy reliance on back channel communication, he writes, can produce agreements that are weak because they are too narrowly based or fail to address major issues. He also notes that the problem can be avoided if enough time is spent assembling a broad central coalition [5]. An understanding is only as strong as the number of people who will stand behind it.
Legitimacy is the second limit. The people outside the room did not hear the reasoning, and they still have to live with the result. A simple test helps. Could the participants repeat in public the substance of what they offered in private? Would the people affected call the privacy reasonable once they learned how the matter was handled? If either answer is no, the room has become a habit instead of a tool.
How the Story Ends
Let us go back one last time to the two delegations and the clause that would not move.
The senior delegate, embarrassed by the leaked summary, did something sensible. She asked for a short private meeting with the junior delegate and the leader of his team. Before saying anything about the clause, she spoke about the promise. She said that she had understood the coffee table conversation to be one where ideas could travel but names could not. She asked whether he had understood it the same way.
He had not. He had understood it as completely private. They talked honestly about how the misunderstanding had happened, and they agreed on terms in plain words. From that point on, each private exchange would begin with one sentence naming the promise, and each side would report back to its sender only the substance, never the source.
Then they returned to the clause. Working from the outcome based wording that had first come up beside the coffee table, the two teams reached a revised draft within a day. The formal record shows only the final text. It does not show a coffee table, a quiet question, a leaked summary or a hard conversation about trust. Nobody was asked to retreat, and nobody did.
That is the real work of confidential diplomacy. It is rarely dramatic. It is careful, human and mostly unseen.
A Practitioner Standard for Confidential Diplomacy
Three habits keep confidential dialogue honest.
- Name the promise before the first word, and keep to it. State whether the meeting is on the record, under the Rule, or off the record.
- Carry a mandate. Speak only to what you are authorized to explore, so that listeners know the weight of what they hear.
- Convert, then close. Turn what private rooms prepare into positions that can be defended openly, and report back to those who gave you the mandate.
None of this asks for secrecy as a style. It asks for clear terms, a clear mandate and a clear return to the table. Followed every time, these habits turn privacy into a service to the process instead of a shelter from it.
A confidence kept is rarely noticed. A confidence broken is rarely forgotten.
The record tells the world what was decided, and the conversation before it decides whether the decision holds.
Sources
- Anthony Wanis-St. John, “Back-Channel Negotiation: International Bargaining in the Shadows,” Negotiation Journal 22, no. 2 (2006): 119 to 144. https://doi.org/10.1111/j.1571-9979.2006.00091.x
- Robert D. Putnam, “Diplomacy and Domestic Politics: The Logic of Two-Level Games,” International Organization 42, no. 3 (1988): 427 to 460. https://www.jstor.org/stable/2706785
- David Stasavage, “Open-Door or Closed-Door? Transparency in Domestic and International Bargaining,” International Organization 58, no. 4 (2004): 667 to 703. https://doi.org/10.1017/S0020818304040214
- “The Chatham House Rule,” Chatham House. https://www.chathamhouse.org/about-us/chatham-house-rule
- Dean G. Pruitt, “Back-channel Communication in the Settlement of Conflict,” International Negotiation 13, no. 1 (2008): 37 to 54. https://doi.org/10.1163/138234008X297922
- Aroop Mukharji and Richard J. Zeckhauser, “Back Channel Negotiations and Dangerous Waiting,” Negotiation Journal 34, no. 3 (2018): 297 to 307. https://doi.org/10.1111/nejo.12232









