What if the probability is tiny but the possible harm is catastrophic?

Probability is not the only factor. Consider irreversibility, evidence quality, the number exposed and narrower precautions. The answer depends on the actual case.

Why is a fair competition allowed when somebody must lose?

Mill treats some conflicts of interest as unavoidable. Society ordinarily protects a fair process rather than guaranteeing freedom from disappointment. Fraud, treachery or force changes the judgement.

Does the harm principle oppose all regulation before harm occurs?

No. Mill considers preventative measures. The questions are whether evidence supports the risk, whether the response is proportionate and whether a less restrictive measure could protect people.

On Liberty

Does a risk of harm automatically justify a ban?

Possible harm to others does not automatically justify prohibition. Mill asks us to examine evidence, means, consent, proportionality and narrower alternatives.

Damage, or probability of damage, to the interests of others does not therefore always justify interference.On Liberty, Chapter V: Applications

The direct answer: risk opens the argument; it does not decide it

Not necessarily. Mill explicitly warns that damage, or its probability, may be necessary before social interference can be justified without always being sufficient. Applications, examinations, trade and competition predictably disappoint people and cost them opportunities. If every foreseeable loss authorised prohibition, little independent action would remain.

Chapter V of *On Liberty* therefore moves from “is there a loss?” to “which interest, by what means, and at what cost should society intervene?” Mill's example is competition. A successful candidate benefits while another loses, yet a fair contest is generally allowed. Fraud, treachery and force change the case.

Separate the outcome, the risk and the means

Similar bad outcomes can arise through very different processes. Losing a fair contest, accepting an informed risk and suffering after deception should not receive the same response merely because each ends painfully. Assess the probability and severity of harm, the quality of the evidence, whether affected people can consent, and whether costs are imposed on people who had no reasonable choice.

Freedom cannot require zero risk. Transport, medicine, employment and innovation all contain uncertainty that cannot be removed. The practical question is whether this risk exceeds what comparable activities ordinarily allow, whether information, design, insurance or narrower rules can reduce it, and what a ban would itself destroy.

A five-step proportionality check

Name the likely victim and loss instead of saying only that something “might be harmful”. Estimate probability and severity while marking uncertainty. Check for deception, coercion, impaired capacity or inability to leave. List responses ranging from warning and disclosure to venue rules, licensing and prohibition. Finally, compare their effectiveness, side effects and enforcement burden.

This does not manufacture one answer for every case. It blocks the most common leap: “the risk is not zero, therefore impose a total ban”.

Prevention needs limits too

Waiting until damage occurs can be too late. A plausible risk of irreversible, catastrophic or widely distributed harm may justify action in advance, especially when potential victims cannot protect themselves. Prevention nevertheless does not cancel the duty to give reasons. The more intrusive the measure, the stronger the case that narrower alternatives would fail, and the greater the need for review, time limits and correction.

Under uncertainty, honesty matters. Neither pretend that the risk is known nor treat missing information as proof of safety. State assumptions, begin with reversible measures where possible, and update the response as evidence changes.

Mill is not a risk calculator

The harm principle constrains coercion; it is not a formula in which probability multiplied by consequence automatically produces policy. Rights, distribution, fairness, institutional trust and historical power all shape a decision. Its lasting value is to make proponents of restriction explain harm and proportionality while making defenders of liberty face real external costs.

Begin with what the harm principle actually means, then browse research across all six classics for other accounts of responsibility and public judgement.