My former mentor David Deutsch taught me an idea that victims of injustices should be made whole by the person or group who wronged them. If someone violates the non-aggression principle, they should pay for all the harm they caused so that no one innocent is worse off than before. Victims should be restored to their original condition before they were victimized. You can consider "If someone offered me $X to accept Y happening to me, would I accept?" If the answer is no, then damages for Y are above $X. Payouts should be enough that victims are indifferent to the injustice never having happened or receiving the payout.
Deutsch took this principle of restitution much more strictly and literally than society usually does. His view is most similar to libertarians. He's trying to be pure and principled, but it actually conflicts with other principles. I now think Deutsch's strong approach to victim compensation is wrong and dangerous, so I've tried to question, unpack and critique it.
Disclaimer: My understanding of Deutsch's view comes primarily from discussions and posts from over 10 years ago. I don't follow his social media much lately. If he's changed his mind, I'm not aware of it.
Analyzing the Principle
According to Deutsch's position, the legal system frequently under-compensates victims, often by hundreds of thousands of dollars. For example, if someone is mugged at gunpoint, they would be asked, "Would you have been willing to be mugged at gunpoint for a million dollars? Keep in mind it'd be a real mugging with a real gun and an unpredictable outcome. Even though you survived your mugging, take into account the risk of death when answering this question." If the answer is "no" then they should receive at least a million dollars in compensation for being mugged. (How can it be determined if they're being honest or how much compensation is really fair? There's no real way to know. But Deutsch was concerned with the principle more than practical details like methods of accurately determining damages.)
In Deutsch's approach, damages for all sorts of injustices can be very high and also include all costs of lawyers, therapists, courts, etc. Damages include your time dealing with everything, including your time spent helping catch the aggressor, choosing your lawyer, talking with your lawyer, travel time to visit your lawyer, travel time to buy a note pad to take notes while talking with your lawyer, time in court, the gas for your car, etc., all costs.
Today, if you buy a false-advertised product on Amazon which doesn't function as advertised (e.g. you buy something metal but receive a plastic product), you'll get a refund and free return shipping. According to Deutsch's view, you should also be paid for your wasted time, for not having a working product sooner, and potentially for feeling frustrated, unsafe and violated. If someone had offered you $20,000 to find out the world is worse than you thought, so that you become disillusioned and cynical, maybe you would have declined. And if the product was important to your child's fancy birthday party that you said was going to be super amazing and perfect (so your child will be very upset and disappointed if it goes wrong), and you can't get a working product in time, there could be big damages.
Victims are required to make reasonable efforts to minimize the harm they suffer, so at some size of damages the victim ought to have bought backup products to reduce risk or bought the product early enough to still have time to buy another. How big do the damages have to be before the victim should have been more careful? $1k, 10k, $100k or what? It's very hard to know, which is a problem with this approach: it doesn't provide enough predictability or enough ease for people to agree on how large damages are. Ideally, everyone ought to be able to pretty easily know in advance what legal outcomes to expect for various actions. Our current legal system isn't great at providing predictability but Deutsch's alternative would be less predictable. E.g. a negligent Amazon seller today knows they're just going to have to pay for a full refund and either let the customer keep the item or pay for return shipping, but they won't have to pay thousands of additional dollars just for sending the wrong product. They'll have to pay a lot more only if they do something especially bad like contaminate their products with lead. This isn't a perfect system but it's more predictable and has significantly lower amounts of victim compensation than Deutsch's system.
A strong approach to making victims whole focuses too much on figuring out who to blame and assigning all the large costs and harms to be their responsibility. It doesn't adequately address people's fallibility. Mistakes happen. Also there are disagreements about fault. And focusing on who to blame is often the wrong way to approach problem solving.
The allegedly principled approach to making victims whole also allows disproportionate responses, e.g. it can justify using guns and violence to get a million dollars back for one gunpoint mugging that took $20. This sort of escalation can easily lead to escalations by the other side who now thinks they're the victim.
Because a strong making whole doctrine tends to assign blame for all the damages from a conflict to one side and none to the other side, it can embolden both sides to use very aggressive, destructive methods since they believe the other side ought to pay the full bill for everything. Also, the damages are often already high enough that they would be unable to pay if they were determined to be the aggressor, so there's little incentive to avoid additional damages even if they know they might be at fault. And the difficulty of figuring out who is the victim can lead to multiple parties thinking they're the victim and then acting very aggressively to "defend" themselves.
It's really problematic for a huge bill of damages to be accumulating while there's uncertainty, unpredictability or disagreement about who will be liable for the bill. It's also really problematic to accumulate such big damages when it will never be paid because whoever is at fault isn't rich enough.
Deutsch taught me that collateral damage should be blamed on the original aggressor (as long as the defender used minimum necessary force to fully defend himself from harms and be made whole, or less). So you can have two sides shooting at each other, who both think they're the victim, and some third party (who everyone agrees is fully innocent) may be shot. Deutsch's view is that whoever is actually the aggressor is responsible for that, but even if fault is eventually determined in a clear way that everyone agrees on, that won't heal the gunshot wound or make the aggressor rich enough to pay for the damage.
It's also problematic when victims give up on justice because the aggressor doesn't have enough money to pay for their lawsuit (so even if they win and are entitled to be paid for their legal costs, they won't be paid). And it's also problematic when the victim wins in court and the aggressor files for bankruptcy (or otherwise doesn't pay because he can't) and the victim is never paid back for the original injustice and also has now lost hundreds of thousands of dollars in lawyer's fees.
Suppose you're accused of aggression and asked to pay $5,000. If you admit your guilt and pay now it'll be done. But if you deny your guilt and go to court and lose, it could cost you hundreds of thousands of dollars for your lawyers, hundreds of thousands of dollars for the other guy's lawyers, and also hundreds of thousands of dollars in court costs in a libertarian society (today, in the US, the court costs will mostly be paid by tax payers). So to defend your $5,000 you might have to risk losing a million dollars and going bankrupt (and a libertarian society with these principles might not allow bankruptcy). So it's much safer to just pay. But the incentive to pay, even when in the right, encourages false, exaggerated or otherwise problematic accusations.
Should court be cheaper? Yes. But doing a good job of reaching a conclusion about complex cases is hard and expensive. These issues would be mitigated some by more efficient courts but wouldn't fundamentally change.
What if you and someone else both made mistakes and both harmed each other some? So sure maybe you did $5,000 of harm to him, but also he did $8,000 of harm to you. Now he demands $5,000 in compensation, and it's true that you did harm him that much, but it isn't true that you should owe him that much taking into account everything that happened. Then what? If you fight in court to avoid paying $5,000, seek $8,000, or seek $3,000 by having both your and his claims handled in one case, that is all risky. It could cost you over $100,000 in legal fees, which you'll never get back if you win and he's poor. And what if you lose and have to pay his legal fees too? What if the court finds that actually you harmed him $500 more than he harmed you? Your evaluation of the harms was incorrect by $3500 which just barely makes you the more guilty party, so then you have to pay over $100,000 for his lawyers and can't ask him to pay for your lawyers, just because you should have owed him $500 in the first place but figuring that out was expensive. I think this is a bad system that's worse than the status quo.
Trying to be made fully whole also ignores the 80/20 rule. Basically, getting pretty good or good enough outcomes is often cheap but getting perfect outcomes is often super expensive and inefficient. For example, after an injustice, 20 therapy sessions might get you back to being reasonably functional and happy, while 500 therapy sessions still might not get you back to 100% unbothered.
Communication
Sometimes people think someone is an aggressor but they don't tell that person. They think it's obvious that he's violating their rights, and he should know, does know, and did it on purpose, so they don't communicate. Often, people don't know about your complaint, don't know there's a problem without being told, and don't see it the same way as you. If you start "defending" yourself without first telling the person about the problem, that's frequently actually aggression by you.
Sometimes people ghost others and expect them to know the reason and take the hint. Ghosting isn't communication, is inherently ambiguous, and often doesn't work well as a first resort.
There are cases where you shouldn't communicate and should go straight the police, like when you're the victim of a violent crime. But for anything that's more of a petty issue, you should probably communicate before forming a grudge or taking any action that would be aggressive if it wasn't justified defense.
Israel and Palestine
Deutsch often used this idea of making victims fully whole in his analysis of issues. For example, he applied it to Israel and Palestine. Based on many conversations, here's what I think he'd say today, which I disagree with (I've put my own comments in parentheses): Israel is the innocent victim of Hamas' October 7, 2023 attack. Israel is justified in defensively dropping thousands of bombs and killing thousands of people until Israel has used enough violence to be made whole (which will never happen) or Hamas is entirely destroyed or fully surrenders (at which point bombing more is pointless). Excessive defensive violence (more than needed to prevent any future attacks and be made whole) isn't allowed but massively disproportionate violence, including using nuclear weapons, is justified as long as Israel still isn't 100% whole and safe. The only ethical question about using nuclear weapons is whether using them in a specific case will help Israel get closer to being made whole or would be counter-productive. If the Palestinians don't like being collateral damage, Deutsch would tell them to demand Hamas make them whole or to attack and destroy Hamas so Israel stops bombing.
Parenting
Deutsch also brought this attitude to his parenting philosophy, Taking Children Seriously (TCS). I think it's a flaw in TCS (see my Fundamental Philosophical Errors in Taking Children Seriously for more).
TCS looks at things kind of like this: When the parent and child have a conflict and get stuck and fail to resolve the conflict, whose fault is it? Either the parent is being unreasonable or the child is being unreasonable. If the child is being unreasonable, that is due to the parent's past mistakes for being coercive or not helping the child well enough to become reasonable. So either way it's the parent's fault. So the child is the innocent victim and ought to be protected from harm and made whole with compensation. So, while the ideal is to find a solution everyone is happy with, the backup plan is for the parent to at least protect the child, make up for whatever past injustices they can, and avoid victimizing the child further. I disagree with this analysis.
TCS tends to put large burdens on parents and assign them huge obligations to their children. I think this is connected with the strong doctrine of making victims 100% whole, which Deutsch believed and taught me, but which I now think is a bad doctrine.
Justice
What is the correct approach to justice? That's a hard question which I won't try to answer here. I do think a more proportional, limited approach to responding to injustice is better and more compatible with fallibility. Having extremely high stakes hinge on getting a judgment of initial fault right, and extremely high risk if you made a judgment error (or if a judge or jury makes a mistake), is a bad approach from the perspective of taking fallibilism seriously (which TCS and Deutsch claimed to do). Fallibilists should care a lot about having more predictable outcomes that are pretty easy to understand and agree on and which don't bankrupt people when they make mistakes. Peace-loving people should also care about preventing conflicts and total damages from escalating.
Extreme cases with intentional and very serious mistakes, like Bernie Madoff's Ponzi scheme, can be handled more aggressively like with bankruptcy and jail. But I think Deutsch is wrong to take a lot of cases that are handled more leniently today and say those cases should be handled much more harshly with much larger victim compensation. I do think in general our society could compensate victims better, and there's room for improvement, but having huge, unpredictable damages in far more scenarios wouldn't be an improvement. And it's dangerous how Deutsch's principle encourages people to do large escalations and think they're justified. Deutsch doesn't require victims go through the courts or police nearly as much as today's society does. He instead says they would be justified in principle to do various "defensive" actions, including disproportionate escalations, so he's encouraging various types of vigilante "justice".
Deutsch's principles also view inappropriate, unjustified lawsuits or police reports as a form of aggression which justifies defensive actions and compensation to make the victim 100% whole. I think he's right to view using the police or courts as a type of force which must be justified as defensive force or else it's aggression. Non-libertarians sometimes downplay the seriousness of socially-legitimized government force. However, even when it's genuinely defensive, I think it's generally still a bad idea to use the police or courts when it's a disproportionate escalation.
Conclusion
Everyone, especially fallibilists, should try to be robust and resilient. They shouldn't have a rigid sense of justice where they can't tolerate the slightest loss. They should approach life in a way where they can handle some bad luck, some unresolved disputes, and even some injustices or receiving only partial compensation. They should recognize that getting along with other people is hard and other people have different ideas about what behaviors are reasonable, acceptable, defensive or aggressive. Social harmony requires significant tolerance of different perspectives on justice, aggression and rights. Thinking you're a victim doesn't justify massively disproportionate escalations. A better more fallibilist principle is don't escalate much. One or two small escalations can be OK sometimes though it's usually a bad idea, but one large escalation or many small escalations are almost always terrible decisions. People often need to let things go or deescalate, and if the matter is so serious that they consider that impossible, then they should generally go to the police or courts instead of pursuing vigilante justice (but people should be hesitant to do that because those government systems can do a lot of harm and be very expensive, so using them can be a disproportionate escalation).
Messages