Organizational Dignity As Terminal Value
When the dignity of an organization out-weighs the dignity of its people.
Dignity by Design | Perceptual Ethics Series
There’s a specific moment that keeps recurring in my thinking. I’ve watched it happen in churches, nonprofits, healthcare systems, graduate programs. Someone gets hurt, genuinely and measurably hurt, by the way a system operates. The injury is real. The pain is real. And then the organization responds.
It responds well, even. It refers them to counseling. It creates space for them to process. It follows its protocols. Sometimes it even apologizes, genuinely and warmly. And when the person eventually returns to functioning, when they’re able to show up, to contribute, to continue, there’s a quiet exhale. A sense that things have been handled. That the system worked.
What I kept not being able to name was why that exhale bothered me.
I’ve been writing about moral injury for the past year as part of the Dignity by Design project, and I’ve spent sustained time with the moral injury intervention literature: the research on how organizations and practitioners respond when people are harmed by systems they trusted. The literature is sophisticated in important ways. But it has a structural problem that took me a long time to articulate precisely.
The intervention literature’s endpoint is restored functioning. It measures whether the injured person can return to effective participation in the organization. This sounds like a therapeutic goal. It is, in fact, an organizational goal. The difference matters enormously.
Organizations, like people, are prediction machines.
This is not a metaphor. Karl Friston’s work on active inference describes how biological systems continuously generate predictive models of their environment and minimize the gap between what they predict and what they encounter. Organizations have no cognition, no behavior, and no activity apart from the humans who staff and reason within them. This means the predictive processing that Friston describes at the neurological level, the same processing that drives heuristic formation, pattern recognition, and the emotional logic of threat and reward, scales directly into organizational behavior through the humans who model the institution, process its feedback, and enforce its responses. When we describe an organization as a prediction machine, we are describing what happens when individual human predictive systems operate within a shared structure that shapes which predictions get made, which evidence gets admitted, and which signals get routed where.
When prediction error is high, when reality diverges from the model, the system has two choices: update the model, or act on the environment to suppress the signal producing the error. Organizations preposition resources to respond to their predictions. They accept evidence. They update. Or they don’t.
A closed epistemic system, one that accepts evidence only from within its own domain and adjudicates that evidence by its own criteria, has structurally eliminated the first option. It cannot genuinely update its model because the evidence it admits is evidence it generates, and the standards by which it evaluates that evidence are standards it set. When a high-prediction-error signal arrives, someone being genuinely and measurably harmed by how the system operates, the only available responses are suppression or elimination.
Individual intervention is how the suppression happens. Termination is how elimination happens.
This is not a fringe pathology. It is the operating condition of every institution that decides what counts as valid evidence about itself. The social sciences have a name for the pattern at the individual level: groupthink, the condition in which internal consensus substitutes for independent verification. But groupthink understates the problem at the organizational level, because the issue isn’t merely that everyone is thinking alike. It’s that the system has engineered which thoughts are permitted to arrive. Miranda Fricker calls a related failure hermeneutical injustice: the condition in which a system lacks the conceptual vocabulary to recognize a harm it is producing, and so cannot be held accountable for it even by people of good faith operating within it. The moral injury intervention literature reproduces this failure structurally. It cannot name organizational culpability because its entire framework is organized around individual restoration. And if you press on why that framework is so stable, why the vocabulary for naming organizational harm develops so slowly against such obvious need, you arrive at a question the intervention literature cannot answer: what produced the cultural conditions in which organizational dignity is treated as a terminal value in the first place?
The answer is functional, not conspiratorial. It operates through legal logic repeated across decades until it becomes social fact.
In 1971, Lewis Powell, then a corporate attorney and soon to be a Supreme Court Justice, wrote a confidential memo to the U.S. Chamber of Commerce. Its argument was strategic: American corporations were under assault from regulatory, legislative, and cultural forces, and the remedy was organized, sustained corporate entry into the political, legal, and academic arenas where the rules of public life were being written. The memo prescribes the intervention: corporations must participate, fund, litigate, and shape, or cede the epistemic environment to their critics.
Seven years later, Powell wrote the majority opinion in First National Bank of Massachusetts v. Bellotti. The Court held that First Amendment protections attach to speech itself, not to the nature of the speaker. The identity of the entity producing political expression is constitutionally irrelevant. What matters is whether the expression contributes to public discourse. Powell did not argue explicitly that corporations are persons with dignity equivalent to human persons. He didn’t need to. The functional logic of the opinion produces that outcome without stating it: if an organization’s speech is constitutionally protected regardless of what the organization is, the organization begins to accumulate the cultural standing of a rights-bearing subject. Personhood follows function. An entity that can speak, litigate, fund, and participate in democratic deliberation with constitutional protection will be treated, behaviorally and culturally, as though it has the standing of a person. Including the standing to have its dignity protected.
Citizens United v. Federal Election Commission (2010) extended that logic to its limit. If corporate speech is constitutionally protected and the speaker’s identity is irrelevant, then limits on corporate political expenditure are limits on protected speech. The last significant structural constraint on corporate participation in democratic deliberation was removed. What the memo recommended as strategy, Bellotti enabled as right, and Citizens United ratified as absolute.
What matters for this argument is not Powell’s intent. Intent is beside the point. What matters is repeatability. Forty years of consistent legal application, extending constitutional speech protections to organizations, treating organizational participation in public life as equivalent to human participation, produced cultural norms that now operate below the threshold of conscious reasoning. People don’t consciously think: the organization has rights equivalent to mine, its continuity deserves the same protection I extend to persons. They simply behave that way. Because the legal framework has been telling them so, functionally, across the entire span of their adult lives. The cultural acceptance of organizational dignity as a terminal value is not propaganda. It is the sedimented residue of a legal logic applied consistently enough to become invisible.
This is what the intervention literature cannot see from inside its own framework: the reason individual restoration feels like sufficient repair is that the surrounding legal and cultural architecture has normalized organizational dignity as the value against which individual harm is measured. Of course the person should be restored to functioning. Of course the organization’s continuity matters. Of course the harm signal should terminate when the individual recovers. The system is working exactly as the culture it inhabits expects it to work.
Which is why the people doing the actual work of responding to injury are in an impossible position that most accountability frameworks refuse to name honestly.
I’m not accusing anyone of bad faith. The clinicians providing moral injury intervention are often doing genuinely excellent work. The HR professionals who refer employees to EAPs are often doing what they believe is right. The pastoral counselors sitting with people in the aftermath of institutional harm are often the most compassionate people in the room.
But compassion operating inside a constitutionally protected architecture of epistemic closure is not the same as repair. The structure of individual intervention routes the harm signal through a process that terminates before it reaches the organizational model. “This person has been helped” completes a circuit. The injury has been addressed. The organization has fulfilled its obligations.
The latent error, the structural condition that produced the injury, remains invisible. The process designed to address harm also terminates the harm signal before it becomes data the organization has to reckon with.
This reveals a moral hierarchy the intervention literature leaves unnamed.
When organizational dignity, reputation, functioning, continuity, self-concept, is the terminal value, individual restoration becomes instrumental. It serves the organization’s sense of itself as a place that cares. The intervention program becomes evidence of organizational moral health. We have resources. We provided care. We took this seriously.
This isn’t cynical calculation, usually. It’s the natural output of a system whose model has organizational dignity at the top of its value hierarchy. The model predicts that good organizations address injury. When injury occurs and is addressed through individual intervention, the prediction is confirmed. The model doesn’t update because, from inside the system, everything worked exactly as intended.
There’s a further layer that took me longest to see.
The individual’s recovery, their demonstrated resilience, their ability to return to functioning, is recirculated into the system as reputation capital.
The organization that harmed someone, provided intervention, and watched that person recover and return to contribution can point to this sequence as evidence of its own moral adequacy. Their survival becomes the organization’s proof.
The cost-shifting runs deeper than financial. It’s not only that the burden of healing falls on the individual who couldn’t refuse to be harmed. It’s that the moral labor of surviving the harm, the hard work of recovery, rebuilding, returning, is extracted from the individual and deposited into the organization’s account. Their resilience underwrites the institution’s innocence.
Perceptual Ethics, the framework I’ve been developing as part of Dignity by Design, begins from the observation that moral perception is conditioned prior to moral reasoning. We don’t first reason clearly and then perceive accurately. We perceive through frameworks we didn’t choose, formed by conditions we were embedded in before we had the tools to question them. This is why moral failures in organizations so often look, from the inside, like moral successes: the perceptual framework has been shaped by the same closed evidence loop that shaped the intervention model. People who are genuinely trying to do right look at the sequence (harm, response, recovery) and see a system that works.
Organizations have moral perception too. A system whose perception has been shaped by vested-interest evidence will perceive individual intervention as genuine restoration, not because the people within it are lying, but because their perception has been formed by a framework whose terminal value is organizational dignity.
This is why external input has to precede assessment. Not because outside voices are infallible or epistemically privileged, but because of what a closed system does to evidence that arrives in its own language: it routes it. Every institution that has operated long enough to develop a coherent self-concept has also developed, without design and often without awareness, a set of absorption pathways: conceptual channels through which potentially disconfirming information is received, processed, and neutralized before it registers as genuine prediction error. A grievance filed in the institution’s own procedural language travels those channels automatically. An internal review conducted by people whose standing depends on the institution’s reputation travels them automatically. Even sincere internal critique, offered in good faith by people who want the institution to improve, tends to arrive pre-formatted for absorption, because the people offering it learned to think about the institution in the institution’s own terms. The evidence arrives. The model doesn’t update. And everyone involved can point to a process that was followed.
External input interrupts this not because it is correct, but because it arrives without those pathways already in place. The closed system cannot pre-adjudicate criteria it did not generate. It cannot route testimony through filters it hasn’t built yet. This is why external assessment is not a supplement to existing accountability structures or a stronger version of internal review: it is a categorically different kind of intervention, the only kind that forces the model to reckon with evidence before it has been metabolized into confirmation.
The scale of what this pattern explains is not abstract. The same closed epistemic architecture operates across every institutional domain: healthcare systems that track patient satisfaction scores while the structural conditions producing preventable harm remain untouched; universities that route faculty grievances through administrative processes designed to confirm the institution’s existing self-assessment; religious organizations that respond to abuse disclosures with pastoral care for the survivor and internal review panels composed entirely of people whose standing depends on the institution’s continued reputation. In each case, the intervention infrastructure is genuine. In each case, it terminates the harm signal before it reaches the model.
What the institutional examples show operating across domains, the disparity argument shows operating across the population of potential harm-witnesses: the same filtering logic that excludes certain evidence from the organizational model also excludes certain people from being treated as credible sources of evidence about their own injury.
The pattern also operates across the categories of persons the organization admits as full subjects of moral concern, and here the argument becomes harder to look at directly. Organizations that have historically distributed dignity unequally along racial, ethnic, religious, sexual, and gender lines do not correct that distribution by providing individual intervention to the people harmed by it. But the relationship between historical disparity and epistemic closure is not merely correlational. It is causal, and the causality runs in both directions. A closed epistemic system that treats organizational dignity as the terminal value will suppress harm signals structurally; and the same filtering architecture that determines which evidence reaches the organizational model also determines whose testimony about harm is admitted as evidence in the first place. The people whose dignity has been treated as instrumental rather than terminal are also the people whose accounts of injury are most likely to be routed through skepticism, procedural delay, or the burden of extraordinary proof before they register as data the organization must reckon with. The structural condition doesn’t merely fail to correct the disparity. It actively reproduces it, because the mechanism that produces disparity and the mechanism that suppresses evidence of disparity are the same mechanism. Each cohort inherits the disparity the previous cohort’s individual recoveries left intact, not as a residue but as the predictable output of an architecture that has not been asked to update.
Individual intervention, internal review, formal grievance processes: each gets routed through processes the system controls, evaluated by standards the system sets, and terminated at conclusions the system can absorb without updating. The harm signal never reaches the organizational model. The exhale happens. And the next cohort enters.
So what would genuine restoration look like?
Not restored functioning. Not a return to extractable contribution. Genuine restoration would be evidenced by organizational model updating: demonstrable change in the predictive apparatus, revision of the organization’s working model of what it can do to people and still call itself a dignity-affirming system.
That metric is different from anything the intervention literature currently uses. It measures organizational revision against what dignity requires, rather than measuring individual recovery against the organizational baseline. And the instrument of that revision cannot be internal, because internal assessment is subject to the same closed epistemic conditions that produced the harm. Outside assessment whose explicit goal is organizational model updating is not a supplement to existing accountability structures. It is the only mechanism that imports criteria the system cannot suppress before they arrive, which makes it, structurally, the only kind of intervention that can force the model to update rather than forcing the environment to conform.
We’re not trying to restore individuals to a system that will harm the next cohort. We’re trying to produce conditions under which the system’s deepest operating assumptions, about what it can do to people, about what counts as evidence, about who has standing to say so, must update in response to criteria it did not generate and cannot adjudicate alone.
That’s what outside assessment is for. That’s what genuine intervention targets.
That’s what organizational dignity, taken seriously, actually costs.