---
title: "The Copenhagen Trap"
subtitle: "How legal systems made passivity the safer strategy"
author: Elias Kunnas
description: "The Copenhagen Trap appears when touching a problem creates far more liability than leaving it untouched, even though the outcome difference is small. Under that attribution rule, passivity becomes the safer strategy."
canonical: https://kunnas.com/articles/copenhagen-trap
url: https://kunnas.com/articles/copenhagen-trap.md
date_published: 2025-11-30
date_modified: 2026-08-31
corpus_frame_url: https://kunnas.com/articles/how-to-read-this.md
---
## How to read this corpus

The corpus applies one lens to many domains: what mechanisms produce the outcome? It shares four methodological commitments and one explicit directional commitment. Each linked page argues for its part; the links are derivations and disputes, not evidence inherited by every page. The directional commitment does not by itself settle system boundary, distribution, sacrifice, or institutional authority.

1. **Mechanisms are what act.** Incentive gradients, selection pressures, feedback loops, and capital stocks produce the distribution of outcomes. Intentions, labels, official categories, and stated values are evidence about mechanisms, or are themselves coordination mechanisms. They are not causal substitutes. — [Mechanism Realism](https://kunnas.com/articles/mechanism-realism.md) · [Only Selection](https://kunnas.com/articles/only-selection.md)
2. **The reference telos is sustained flourishing.** The broadest achievable adaptive safety margin over deep time — not the continuity of any incumbent state, coalition, institution, or doctrine. A mechanism's own stated goal can still serve as a local proof obligation — showing that its incentives defeat even the purpose it claims is a bounded finding — but meeting that goal establishes nothing about the margin. — [Flourishing Is Maximum Safety Margin](https://kunnas.com/articles/flourishing-is-maximum-safety-margin.md)
3. **Law, rights, legitimacy, democracy, markets, and sovereignty are mechanisms under evaluation.** They are constraints, carriers, or proxies inside the analysis. None is a terminal value or a boundary of what is real. Treating one as terminal ends the mechanism search before it starts. Evaluation carries current function, replacement cost, path dependence, uncertainty, capture risk, reversibility, and who bears model error into the ledger. — [The Stack](https://kunnas.com/articles/the-stack.md) · [Mechanism Space](https://kunnas.com/articles/mechanism-space.md)
4. **Optimization is a system function.** A civilization has to build, exercise, and revise metamechanisms that search mechanism-space, discard dominated options, install, observe effects, and repair under uncertainty. Not running that loop leaves margin unrealized, and that is itself the failure. No single component — analyst, model, or institution — is presumed to contain a global optimum; the capacity is a property of the system. — [From Telos to Policy](https://kunnas.com/articles/from-telos-to-policy.md) · [The Three-Layer Architecture](https://kunnas.com/articles/three-layer-architecture.md)
5. **Uncertainty is preserved, not spent.** Partial orders, binding constraints, unknowns, and residuals stay explicit. An unmeasured effect is not a favorable default. — [The Compression Paradox](https://kunnas.com/articles/compression-paradox.md) · [Cargo Cult Epistemology](https://kunnas.com/articles/cargo-cult-epistemology.md)

*Each essay bears its own evidence. Links carry definitions, derivations, applications, and disputes; they do not transfer proof. Criticism is answered on its substance.*

Canonical: <https://kunnas.com/articles/how-to-read-this.md>

---

# The Copenhagen Trap

*How legal systems made passivity the safer strategy*

Elias Kunnas

## Thesis {#thesis}

The Copenhagen Trap appears when touching a problem creates far more liability than leaving it untouched, even when the outcome difference is small. Passivity becomes the safer strategy, and careers select for the Unstained Incompetent: leaders skilled at avoiding attributable decisions.

## Standard objections addressed in this essay

- “Duty-to-rescue and Good Samaritan laws already address this.” — [§X](#x-the-geography-of-the-trap) (They cover bounded rescue cases, not the wider attribution architecture.)
- “Actions and omissions really are causally different.” — [§IV](#iv-why-it-feels-like-moral-truth) (The trap begins where liability differences exceed outcome-relevant differences.)
- “Penalizing inaction creates coercion and defensive overreaction.” — [§XII](#xii-why-it-wont-fix-itself) (The essay diagnoses the asymmetry; repairs require domain-specific limits.)
- “Tort law, self-defense, regulation, and social media are unrelated.” — [§VII](#vii-the-internet-weaponization) [§X](#x-the-geography-of-the-trap) [§XI](#xi-the-thermodynamic-frame)(The shared claim is asymmetric attribution, not doctrinal identity.)
- “The selection effect on leaders is asserted, not demonstrated.” — [§IX](#ix-the-selection-effect) (It is a testable population-selection hypothesis, not an established prevalence estimate.)

---

## I. The Paradox {#i-the-paradox}

A company discovers a potential safety flaw in its product. It commissions a voluntary audit. The audit misses something. A customer is injured. The company is now liable for negligent audit—liability it would not have faced had it never looked. The lesson: don't look.

A physician considers withdrawing life support from a brain-dead patient. Withdrawing feels like killing—an action. Not starting support in the first place would have been omission. Same outcome, different liability exposure. The lesson: never start what you can't continue.

A regulator considers approving an experimental drug. If approved and someone dies: scandal, hearings, career destruction. If delayed and people die waiting: no attribution, no scandal, invisible deaths. The lesson: delay is always safer.

This is the **Copenhagen Trap**—named after the "[Copenhagen Interpretation of Ethics](https://forum.effectivealtruism.org/posts/QXpxioWSQcNuNnNTy/the-copenhagen-interpretation-of-ethics)" (Jai Dhyani, 2015): like observing a particle collapses its wavefunction in quantum mechanics, *interacting* with a problem makes you responsible for it. Ignore the problem, retain immunity.

The trap begins where an action–inaction consequence gap is materially smaller than the resulting liability or attribution gap. Legal liability, organizational blame, and social-media attention are distinct substrates, each with its own rules and remedies; the shared diagnostic is asymmetric attribution. This essay traces those separate manifestations and considers a selection hypothesis about their institutional consequences.

## II. The Ancient Baseline {#ii-the-ancient-baseline}

The action/inaction asymmetry is ancient—Roman law struggled to assign causation to non-events. But it was held in check by countervailing forces: honor cultures punished inaction as social death, feudal obligations created webs of positive duties, canon law held omission sinful, noblesse oblige made passivity shameful. The asymmetry existed in embryonic form, but passivity was costly.

## III. The Common Law Crystallization {#iii-the-common-law-crystallization}

English Common Law hardened the asymmetry into rigid doctrine.

The key case: **Hurley v. Eddingfield** (1901). A physician refused to travel to treat a dying patient, despite being the family doctor and the only available aid. The court held: no liability. The ruling underscored the common law's commitment to individual liberty—the law forbids you from hurting your neighbor, but does not force you to be his savior.

This became the "no duty to rescue" rule. The logic: negative rights (not to be killed) are stringent; positive rights (to be saved) are optional. The non-rescuer doesn't *cause* the death; the drowning water does.

Good Samaritan laws exist in most US states—but they provide *immunity* for rescuers, not a *duty* to rescue. They reduce the penalty for helping without creating any penalty for not helping. The asymmetry remains.

## IV. Why It Feels Like Moral Truth {#iv-why-it-feels-like-moral-truth}

The Utilitarians challenged this. Bentham and Mill introduced a consequentialist framework where outcomes matter regardless of method. If morality is determined by consequences, then failing to prevent a death is mathematically equivalent to causing a death, provided the cost to the agent is low. Mill wrote explicitly: "A person may cause evil to others not only by his actions but by his inaction, and in either case he is justly accountable."

But the **Deontological backlash** reasserted the asymmetry. For the Kantian, the moral quality of an agent is determined by adherence to rules—primarily negative prohibitions. The consequences of inaction are tragic, but they are not violations of the moral law in the same way as active harm. Requiring men to be "good" (positive duty) is religion; requiring them not to be "bad" (negative duty) is law.

The **Trolley Problem** (Foot 1967, Thomson 1985) formalized the intuition: pulling the lever to divert the trolley (action) feels viscerally different from standing by while five die (inaction), even with identical body counts. Philosophy tutored by law; law reinforced by philosophy. A feedback loop that made the asymmetry feel like moral truth rather than contingent legal convention.

Actions and omissions can be causally, informationally, and administratively different. The Copenhagen diagnosis does not deny that difference. It applies where those outcome-relevant differences are small relative to a much larger difference in liability, blame, or attention.

## V. The Institutional Encoding {#v-the-institutional-encoding}

The 20th century embedded the asymmetry into bureaucratic institutions across medicine, corporate compliance, drug regulation, and self-defense law. In each domain below, the same attribution structure recurs: liability tracks the act, not the outcome.

### Medical Ethics: The Withdrawal Trap

Withholding life support (not starting it) is omission—the disease kills, not the doctor. Withdrawing life support (stopping it) is action—the doctor kills. Same outcome, radically different liability. Physicians are measurably less likely to withdraw than to withhold. Result: overtreatment, prolonged dying, because stopping feels like killing.

### Corporate Liability: Willful Blindness as Strategy

Under Restatement (Second) of Torts § 324A, voluntarily undertaking a safety audit creates a duty of care. Negligent audit = liability you wouldn't have faced without looking. A poor compliance program is worse than no compliance program. The rational strategy: don't look for problems you'll be blamed for not fixing.

### Regulatory Paralysis: The Invisible Graveyard

Approve a drug that kills one person: scandal, hearings, career destruction. Delay a drug and a thousand people die waiting: invisible deaths, no attribution, no story. Commission Errors are punished; Omission Errors are invisible. A regulator facing that attribution structure has a strong structural incentive toward delay. The people who died waiting have no lobby.

### Self-Defense: The Imminence Trap

The "imminence" requirement forces defenders to absorb the first blow before acting. Strike preemptively against a certain threat and you're the aggressor. Wait until you're actually being attacked and maybe you can defend yourself—if you survive long enough to exercise the right. (See Section X for country comparisons, and [The Thermodynamics of Power](thermodynamics-of-power.md) for the full treatment.)

## VI. The Liability Singularity {#vi-the-liability-singularity}

As systems become more complex, interconnectivity increases: any action has more side effects to attach liability to. If the attribution rule from [§I](#i-the-paradox) holds — liability tracks the act, not the outcome — then complexity raises the odds that some negative side effect of any given action becomes attributable, even as the odds that the action itself was net-harmful stay flat or fall.

Push that trend far enough and you approach a limit worth naming: **the Liability Singularity**, where the marginal cost of acting exceeds the marginal cost of not acting in nearly every case, and rational actors default to inaction. No institution examined in this essay has been shown to sit exactly at that limit; the claim is the direction of the pressure, not a measured value for P(liability \| action).

The felt version of this pressure is familiar inside modern institutions: the operative question shifts from "what should we do?" to "what can we do without creating liability?" [§V](#v-the-institutional-encoding) and [§X](#x-the-geography-of-the-trap) supply the specific domains — medicine, compliance, drug approval, self-defense — where that shift has been documented rather than merely felt.

## VII. The Internet Weaponization {#vii-the-internet-weaponization}

Social media didn't create the asymmetry—it amplified it. Every interaction is now observed by billions of potential critics. The Copenhagen dynamic scales: interact and become responsible; ignore and retain immunity.

**The mechanics:**

1.  **Interaction implies liability:** Help a homeless person imperfectly → criticized for the imperfection
2.  **Profit implies guilt:** Sell cheap water in a drought → "profiteer," "monster"
3.  **Ignorance implies innocence:** Ignore the problem entirely → zero criticism

**Case studies:**

Kabul, 2021. Private contractors offered evacuation flights at high prices. They were labeled "profiteers" and "monsters." The action (providing a flight at a price) was strictly better than no flight at any price. Yet the actors who saved lives for profit were vilified; those who offered nothing faced no criticism. The profiteer is hated more than the bystander who lets people die for free.

Newark, 2010. Zuckerberg donated \$100 million to Newark schools. A decade later: years of scrutiny over consultant fees, governance failures, conditions attached. Other tech billionaires who donated nothing to education? No scrutiny. No responsibility for Newark's continued problems. The donor becomes responsible for the school system; the non-donor remains innocent.

The pattern: **marginal improvement is punished more harshly than total neglect.**

## VIII. The Nuclear Proof {#viii-the-nuclear-proof}

The Copenhagen Trap explains one of the great paradoxes of modern energy policy.

**Coal** kills millions through particulate matter, respiratory disease, and climate effects. These deaths are statistical, diffuse, invisible. No single death can be attributed to a specific decision. No one is liable.

**Nuclear** risks meltdown. Meltdowns are concentrated, visible, events. They generate headlines. They create liability. Someone approved the plant. Someone is responsible.

The result: we burn coal instead of splitting atoms. We prefer **statistical death** (millions of invisible casualties) over **liable risk** (possible visible casualties). The decision-maker who approves nuclear and faces a meltdown is destroyed. The decision-maker who continues coal and presides over diffuse death is safe.

This is not irrational given the incentive structure. It is perfectly rational given the asymmetry. The asymmetry is what's irrational—but no individual actor can change it.

## IX. The Selection Effect {#ix-the-selection-effect}

The Copenhagen Trap doesn't just affect decisions. It affects **who makes decisions**. This is not about individual choices. It is about civilizational selection pressure.

**The Ombudsmen vs. Operators problem:**

People who actually fix things—operators—accumulate "liability mud." They've made decisions. Some decisions had negative side effects. Their record includes visible failures, even if the failures were necessary costs of larger successes.

People who manage process without touching outcomes—ombudsmen—have clean hands. They've never approved anything that failed, because they've never approved anything. They've never been blamed for a negative side effect, because they've never caused a side effect.

This yields a testable population-selection hypothesis: as the liability environment intensifies, career paths may favor low-exposure administrators over operators with visible decision records. Establishing it requires evidence on promotion, hiring, attrition, and decision histories; this essay does not supply a prevalence estimate.

“Unstained Incompetent” names that hypothesized selection outcome, not an established diagnosis of all leaders. If the mechanism operates, it would make avoiding attributable decisions disproportionately valuable and make reform less likely from the resulting career pool.

### The Psychological Relief Function

Why do people accept this? Because agency is terrifying.

The Copenhagen Trap is a **relief**. It tells the risk-averse: "You don't have to act. In fact, it is *moral* not to act. You are being 'responsible' by ignoring the drowning man, because you aren't trained. You are being 'prudent' by not intervening, because you might make it worse."

The Trap transforms passivity into virtue—the comforting assurance that doing nothing is the ethical choice.

## X. The Geography of the Trap {#x-the-geography-of-the-trap}

The bystander rescue asymmetry is not a human universal—it is a **Common Law peculiarity**. Continental Civil Law systems take the opposite approach:

- **France:** Article 223-6 of the Penal Code—failure to assist a person in peril carries up to 5 years imprisonment and €75,000 fine
- **Germany:** Section 323c StGB—up to 1 year imprisonment for failing to render assistance
- **Denmark:** Penal Code §253—up to 2 years imprisonment (increased from 3 months in 2004)
- **Poland:** Article 162—up to 3 years imprisonment
- **Spain, Italy, Portugal, Switzerland:** All criminalize failure to rescue

The Civil Law tradition, rooted in concepts of *solidarité sociale*, treats the failure to help as a crime against the social fabric. The Common Law tradition, rooted in negative liberty and the misfeasance/nonfeasance distinction, protects the right to do nothing. For bystander rescue, this is a clear Common Law vs. Civil Law split—but the trap manifests differently in different legal domains.

The **Peng Yu Case** (China, 2006) demonstrates the trap's social effects. A man helped an elderly woman who had fallen. She sued him. The court ruled against him, reasoning that "according to common sense," no one would help a stranger unless they felt guilty. The ruling sparked national outrage and created a chilling effect: a decade of bystanders refusing to help accident victims, captured in viral videos of injured people ignored on sidewalks. China finally passed Good Samaritan laws in 2017—eleven years later—to undo the damage.

When helping implies guilt, no one helps. Social trust collapses. (Peng Yu later confessed to police that he had in fact caused the fall—but this is irrelevant to the mechanism: the court's reasoning established a general precedent where helping = evidence of guilt, and the chilling effect persisted regardless.)

### The Self-Defense Trap: A Different Domain

Duty-to-rescue laws address one manifestation of the Copenhagen Trap. But the trap can manifest in a completely different domain: **self-defense law**. Here the pattern cuts across the Common Law / Civil Law divide entirely.

**Nordic countries (Finland, Sweden):** Legal advice: "always run away and call the police." Using a knife against multiple unarmed attackers will almost certainly result in prosecution for excessive defense. Carrying weapons for self-defense is illegal. Defending others with "disproportionate" force? Same prosecution. The law tells you: *call for help, but don't provide it effectively.*

**France:** The "Bijoutier de Nice" case (2013). A 67-year-old jeweler was robbed. As the robbers fled on a scooter, he followed them outside, knelt, and shot one in the back. Self-defense was rejected—the threat had ended when they fled. He was convicted of "voluntary violence with weapon resulting in death" and sentenced to 5 years fully suspended. The public outcry was immense (1.5 million Facebook supporters), but the law was clear: once the threat retreats, defensive action becomes aggression.

**United Kingdom:** Common Law, but with the same pathology. Duty to retreat, near-total gun prohibition, and regular prosecution of homeowners who defend against burglars. Tony Martin (1999) shot two burglars in his isolated Norfolk farmhouse after years of repeated break-ins; he was convicted of murder (reduced to manslaughter on appeal) and served three years in prison. The surviving burglar later sued Martin for damages. The message: even in your own home, after repeated victimization, passivity is legally safer than defense.

**Contrast with Germany and Poland:** Both Civil Law—but with robust self-defense. German law holds that "Right need not yield to Wrong" (*Das Recht braucht dem Unrecht nicht zu weichen*)—no duty to retreat, and defense of property as well as person is explicitly protected. Poland is similarly defense-friendly: all knives are legal to carry (not classified as weapons), there is no duty to retreat, and the law recognizes that defenders under psychological stress may exceed strict proportionality without criminal liability.

The pattern: **the Copenhagen Trap is not simply "Common Law vs. Civil Law."** It is a syndrome that can manifest in different legal domains. A jurisdiction may reject the trap for bystander rescue (mandating help) while embracing it for self-defense (criminalizing defensive action). The Nordic countries are perhaps the purest expression: you are legally required to help strangers, but legally forbidden from helping yourself with effective force. The trap operates wherever the law makes passivity safer than action.

## XI. The Thermodynamic Frame {#xi-the-thermodynamic-frame}

Entropy increases in closed systems. Fighting entropy requires *work*—coordinated action against the gradient.

A legal and social system that penalizes action and rewards inaction is a system that **penalizes the fight against entropy**. It makes the thermodynamically expensive choice (coordinated action, building, defending, creating) *also* the legally and socially expensive choice.

Push the incentive far enough and the system stops fighting entropy: preemptive defense becomes unavailable (must wait for imminent attack), rapid innovation stalls (must prove safety before acting), crisis response slows (acting creates liability, inaction doesn't), and altruism gets penalized (helping creates liability, ignoring doesn't). [§V](#v-the-institutional-encoding) and [§X](#x-the-geography-of-the-trap) document each of these in the specific institutions examined — US tort liability, UK/Nordic/French self-defense law — not as a claim about the West as a whole.

The action/inaction asymmetry is the **mechanical linkage** that converts safetyism into institutional reality.

## XII. Why It Won't Fix Itself {#xii-why-it-wont-fix-itself}

The Copenhagen Trap is not a bug. It is a feature of a civilization that has chosen comfort over agency, safety over adaptation, clean hands over effective action.

It is the legal expression of a deeper axiological shift: a system that uses its own abundance to fund the mechanisms of its paralysis. When survival was hard, inaction was punished by reality. When survival became easy, inaction became safe—and the law followed.

If the proposed selection effect is present, the trap becomes self-reinforcing: career pools favoring low exposure may be less likely to reform the conditions that rewarded them. This remains a hypothesis about incentives and career paths, not a claim that every administrator is incapable of reform.

This is a diagnostic, not a general warrant to penalize inaction. Any repair must be domain-specific: it must define duties, evidentiary standards, safe harbors, and limits that avoid coercion and defensive overreaction.

Selection explains the passive resistance to reform. But the Trap also persists through active maintenance: state actors *benefit* from the monopoly on legitimate action. More dependency means more power. More intermediation means more budget. More cases mean more jobs. Police unions, regulatory agencies, legal professions—all have concentrated interest in preserving the architecture. The losers (crime victims, the over-regulated, those who die waiting for approvals) are diffuse and unorganized. Classic Olsonian collective action failure: those who gain from the Trap coordinate to preserve it; those who lose cannot coordinate to dismantle it. This is [Moloch](physics-of-moloch.md)—each actor optimizing locally, the collective outcome anti-civilizational, no one "in charge" of the result.

A civilization that cannot act cannot survive. Among the jurisdictions examined here, the Anglosphere and Nordic Europe have each built, through different mechanisms, an architecture that makes passivity the safer strategy across most of the domains this essay traces. Germany and Poland are the documented exceptions, retaining both duty-to-rescue and robust self-defense rights. Where the architecture holds, it produces the leaders suited to it: people who have mastered the art of not deciding, not acting, not touching.

The Unstained Incompetent inherit the earth. Until the earth is inherited by someone else.

---

**Related reading:**

- [The Axiological Malthusian Trap](axiological-malthusian-trap.md) — The broader framework on how abundant civilizations fund their own decay
- [The Thermodynamics of Power](thermodynamics-of-power.md) — The criminal law parallel: self-defense, state monopoly on violence, and anarcho-tyranny
- [The Tyranny of the Present](tyranny-of-the-present.md) — The 1966-1976 ideological roots of the safety-first worldview
- [Simulated Metamorphosis](simulated-metamorphosis.md) — Where the system channels remaining agency-drive: protests, voting, posting as stability mechanisms
- [The Physics of Moloch](physics-of-moloch.md) — Multi-polar selection and joint-move traps in the compositional model
- [You Should've Known](you-shouldve-known.md) — a neighboring attribution error: treating the visible cost of inquiry as proof that the inquiry should not have been made

## Sources and Notes

**The Concept:**

- Dhyani J. ["The Copenhagen Interpretation of Ethics."](https://forum.effectivealtruism.org/posts/QXpxioWSQcNuNnNTy/the-copenhagen-interpretation-of-ethics) *Effective Altruism Forum*, 2015. — Originating formulation: interacting with a problem makes you responsible for it.

**Legal Foundations:**

- *Hurley v. Eddingfield*, 156 Ind. 416, 59 N.E. 1058 (1901). — No common-law duty to rescue; physician not liable for refusing to treat.
- Restatement (Second) of Torts § 324A. — Voluntarily undertaking a safety inspection creates liability for negligent performance. The "don't look" incentive.

**Moral Philosophy:**

- Foot P. "The Problem of Abortion and the Doctrine of the Double Effect." *Oxford Review* 5, 1967, 5–15. — The original Trolley Problem formulation.
- Thomson JJ. "The Trolley Problem." *Yale Law Journal* 94(6), 1985, 1395–1415. — The Fat Man variant and the action/inaction asymmetry.
- Mill JS. *On Liberty*, 1859, Ch. I. — "A person may cause evil to others not only by his actions but by his inaction, and in either case he is justly accountable."

---

***Acknowledgements:** Thanks to [Hacker News commenters](https://news.ycombinator.com/item?id=46094465) for corrections on geographic scope and duty-to-rescue laws. The essay has been revised to show the Copenhagen Trap as a syndrome cutting across legal traditions: tort liability (US), self-defense restrictions (UK, Nordic, France), or both. Germany and Poland escape in both domains.*

## Synthesis {#synthesis}

**The argument in four sentences:** Derived from the common-law "no duty to rescue" rule and amplified by modern tort liability, this asymmetry punishes marginal improvement more harshly than total neglect. As system complexity grows, any action carries more side-effect liability exposure, pushing rational actors toward inaction wherever the attribution rule holds. This dynamic pressures out effective operators who accumulate "liability mud" and favors those whose records are clean because they never did anything. Where the pressure wins, the result is institutional paralysis: civilization stops fighting entropy because building and creating is legally penalized.