Leadership · Voice and Silence
The Junior Who Saw the Real Fault
The person closest to the failure is usually the one with the least standing to report it. That is a design, not an accident.
After a failure, the inquiry almost always finds somebody who saw it coming. Not a senior somebody. A technician who noticed the reading drift, a nurse who thought the dose looked wrong, an apprentice who lit the lamps the night of the storm and saw water coming through the base of the wall rather than over the top. The finding is rarely that nobody knew. It is that the person who knew did not say, or said it to someone who could not act, or said it once and was not asked again. Organizations are built, without anyone deciding it, so that the person with the clearest view has the least standing to use it.
Research foundation
- Organizational silence and employee voice
- Implicit theories of when it is safe to speak
- Psychological safety and leader inclusiveness
- Status differences and who gets heard
- Dispersed knowledge and the person on the spot
- The Four A's of Organizational Readiness™
Key Takeaways
- Silence upward is a system property, not a character flaw. Two decades of research find that people withhold concerns for two reasons, fear and futility, and that both are produced by the structure around them more than by the people inside it.
- The most useful observation in an organization is often held by the person with the least formal standing, because they are closest to the work. Knowledge and the authority to act on it sit in different people, and the gap between them is where failures grow.
- An open door is not an invitation. People speak up when the leader is seen to ask, when speaking has produced a visible result before, and when status differences in the room have been explicitly set aside by the person at the top of them.
- The practical test is not whether your people could tell you. It is whether, this week, you asked the person closest to the thing you most need to know, and what happened to the last person who told you something you did not want to hear.
Nobody decided to build it this way
Morrison and Milliken named the pattern organizational silence: a collective condition in which people across an organization withhold what they know about problems, and do so in a way that is stable enough to feel like the culture. Their argument was that it is not a sum of individual timidity. It is produced by structures and beliefs, in particular managers’ fear of negative feedback and a set of implicit assumptions that people below are self-interested and that management knows best. Those beliefs shape what gets asked, how upward information is routed, and what happens to the person who volunteers it.
The evidence from the people who stay quiet is consistent. When Milliken, Morrison and Hewlin interviewed employees about issues they had chosen not to raise, most had withheld a concern at least once, and the reasons were not exotic. They feared being seen as a complainer or a troublemaker, feared damaging a relationship, and, most of all, believed that saying it would change nothing. Morrison’s later review of the whole literature settles on the same two motives, fear and futility, and finds that silence is most likely exactly where it costs most: toward people higher in the hierarchy, about issues that touch someone else’s domain.
The finding after a failure is rarely that nobody knew. It is that the person who knew had no standing to be heard.
The rules nobody wrote down
Detert and Edmondson asked what people actually believe about speaking up, and found a set of self-censoring rules that people carry from job to job and apply without checking whether they hold here. Do not bypass your boss. Do not embarrass your boss in public. Do not speak without solid data or a solution ready. Assume the person you are about to challenge will take it as identification with them. Assume speaking up has career consequences. They called these implicit voice theories, and the important word is implicit: the rules operate even in organizations whose leaders would say, sincerely, that they want to hear everything.
Read those rules from the apprentice’s side and the silence is not a mystery. She saw water at the base of the wall. She has no data, only what she saw. Reporting it means going past the keeper she reports to, in a matter that belongs to the warden and the guild. She is seventeen and has never spoken at a council sitting. Every rule she carries says wait to be asked. Nobody asks, because the people with authority to act are looking at the breach at the top of the wall, which is the one they can see.
Knowledge and authority live in different people
Hayek’s argument about dispersed knowledge is eighty years old and still the clearest statement of why this matters. The knowledge that decides outcomes is largely knowledge of particular circumstances of time and place, and it exists only in the person on the spot. It cannot be gathered upward in full, because most of it is not the kind of thing that survives being written into a report. An organization that routes decisions to where authority sits, and information to where it is convenient, has arranged for the two to meet rarely.
The Four A's call this the authority condition, and it is the one this pattern lives inside. The question is not whether the apprentice is competent. It is who may decide, who may be heard before a decision, and whether the distance between them is something the organization has looked at. Nembhard and Edmondson’s study of health care teams is the cleanest demonstration: professional status differences reliably suppressed the willingness of lower-status members to speak and to try improvements, and the thing that reversed it was not a policy but leader inclusiveness, the leader visibly inviting and appreciating contributions from the people the hierarchy would otherwise mute.
An open door is not an invitation
Leaders tend to believe their people could tell them anything. Detert and Burris tested the belief and found that what predicts voice is not the leader’s general style but whether the leader is seen as open: whether raising something has, in the past, been met with interest rather than defense, and whether anything happened afterward. That perception works through psychological safety, the shared belief Edmondson identified that a team is safe for interpersonal risk, and it is built or destroyed one response at a time. The last person who told you something you did not want to hear is, for everyone who watched, the current policy.
It is also a property of the unit and not only of the pair. Tangirala and Ramanujam found that silence about critical work issues varied with the climate of the unit people sat in, and that a unit where procedures were seen as fair softened the tendency of otherwise quiet people to stay quiet. A leader inherits that climate and can change it, but not by announcement.
The Four A's of being told
Attention
Ask the person closest to the work, on purpose. Attention is allocated by who gets asked. If your questions go only to the people who report to you, you have decided, without deciding, whose view you will never have.
Alignment
Make the account of the problem shared before the plan is. A team that agrees on a fix while one member holds a different picture of the fault has aligned on the wrong thing, and will find out when the fix does not hold.
Authority
Give the observation standing, not only the observer. A report from the bottom of the hierarchy should not need a sponsor to be read. Decide in advance who must hear a first-hand observation before a decision that depends on it.
Adaptability
Treat a late-arriving observation as a signal about the system. When the inquiry finds the person who knew, the question is not why they stayed quiet. It is what in the design made quiet the rational choice, and whether that is still true.
What a leader can do this week
| Instead of | Try |
|---|---|
| Saying the door is open | Asking one named person, below the level you usually ask, what they have seen that nobody has asked them about. Then acting on it where you can, visibly. |
| Requiring data before a concern counts | Accepting a first-hand observation as enough to warrant a look. The data is what the look produces. |
| Routing concerns up the chain | Naming, for the decisions that matter most, who must hear from the people closest to the work before the decision is taken, and letting that path skip levels. |
| Asking why they stayed quiet | Asking what happened to the last three people who did not. Their answer is the policy your people have already read. |
Practice it before the wall comes down
The apprentice in the opening is not invented for this paper. She is Ivy Marsh, seventeen, a lamplighter’s apprentice in a flooded ford town called Emberford, in Ardenwake, a persistent world Mission Intelligence Systems publishes so that these conditions can be practiced rather than read about. Take up her life and you hold the observation that will decide whether the wall stands, with no standing to be heard and nobody asking. Take up the warden’s and you have five people wanting you and a wall you have not been down to look at. The realm runs in your browser and sends nothing anywhere.
Walk as Ivy and see what it costs to say it. Then ask who, in your own organization, is lighting the lamps tonight.