You might be interested in the "anti-Blackstonian" movement in legal theory. I have in mind in particular Larry Laudan's work (eg, early chapters of "Truth, Error, and Criminal Law"), but he's been influential. While he doesn't model the relevant tradeoffs exactly as you do, what he does is pretty similar in spirit, and he also concludes that the idea that we should very strongly prefer acquitting the guilty to convicting the innocent turns out to be a lot harder to justify than normally thought.
Fundamental failure to understand the operative frame of reference.
To punish the guilty is a matter of proportion: does the punishment fit the crime? To punish the innocent who committed no crime is therefore like trying to divide by 0: the result is effectively infinitely disproportionate.
Worse, they aren't merely like values incremented in positive and negative directions. To punish the innocent is categorically different from punishing the guilty, it is itself another moral wrong of commission. Commission and omission are not directly comparable. Justice and injustice can't be traded like a fungible unit of exchange.
Then you finish with an entirely unsupported bit of bad faith mind reading. This entire article is an illogical ad hominem against people who disagree with your preferred policy. Given the title of this space, it doesn't belong here.
I worked as a lawyer for a state prison at one point. I’m going for a little vagueness to stay anonymous. The system had an intake process that could not (or at least never was) be used against the new inmate. The point of it was to protect the inmate from people who might want to do him harm. So they asked questions about gang membership, who might have it out for you, and what crimes you committed. Because of the situation, most inmates copped to everything they ever did. While I won’t say that my interactions are a statistical sampling, I can say that I never came across someone who should not have been in prison based on prevailing norms of the time—none of them were free of fairly serious felonies. Note, though, that most of the inmates I encountered were in higher security units. The inmates in minimum security units struck me as a completely different population.
I suspect that the actually innocent are concentrated in (a) homicides, where the victim is not a witness and (b) misdemeanants, who would be in jail rather than prison, and (c) minimum security prisons, which I visited less. That said, the prisoners who were in for some form of homicide had copped to some form of homicide in the interviews. So I might be wrong about that. But death-row inmates also did not show up in the population that I worked with. And that 1 in 25 statistic is about death-row inmates. So I don’t know that it can be generalized to all prisoners. Death-row inmates tend to have been in for a long time, so they might have been in for decades. A lot of the Innocence Project’s successes have been where new techniques (eg DNA) have proven innocence. But those techniques had become more routine in the underlying prosecutions by the time that I was working with prisoners. And as I noted above, the victim can’t usually testify in a homicide case, so the best possible evidence is missing.
When the interview process was first implemented (along with a number of other measures), the amount of inmate-on-inmate violence fell precipitously. It’s obviously hard to tease out causality when multiple things happened at once. And we shouldn’t really be surprised that prisons are violent places—it is where society puts violent people for long periods of time. I don’t think that impugns the validity of the data that the process collected.
Finally, I’d say that the number does not reach the thousands. And it isn’t a statistical sample. So it’s just lots of anecdote.
To a utilitarian, all moral actions are directly comparable. That's a feature, not a bug. If it were really true that justice and injustice can't be traded like a fungible unit of exchange, then we should have no justice system, since we would have literally NO IDEA how to make that trade-off.
Shrug. I am not a utilitarian, AFAIK most people are not utilitarians, and I'm unaware of ANY country or legal jurisdiction that has explicitly based their legal system on utilitarian ethics, yet we have generally had little trouble designing and implementing justice systems throughout history without any need for utilitarianism.
So I'm inclined to say you're clearly attempting to apply the wrong framework to the question and demonstrably wrong that justice and injustice must be fungible for anyone to design a working justice system. No matter how good you think your hammer is, that doesn't actually make everything else a nail.
My point holds only for non-utilitarians, not for utilitarians. If you don't think the justice of punishing the guilty can be traded off against the injustice of punishing the innocent, how do you decide how much certainty is required to convict someone? The 2 things are apples and oranges under your point of view, so you can't say anything at all, other than to retreat into the fantasy that your judicial system can be 100% accurate.
Your point holds for utilitarians because no utilitarian has EVER come up with anything even REMOTELY close to an objective, convincing measurement procedure for this fantastical object they call “utility”. Talk of “utility” is no better than New Age talk of “energies and vibrations, man”. The only difference is that it attracts dupes of a slightly different cultural bent.
While you can’t measure a “util,” you can measure preferences for a single person based on what they do. This works well for repeated market transactions. It works much less well for (hopefully) one-time non-market events like involvement in the justice system. But even then, we can say something because a person who committed the offense had a utility function for the conduct that resulted in its expected utility to be higher than the risk-weighted punishment. (Or they were drunk or high or crazy or or or.)
Interpersonal ranking of utility are much harder, and almost impossible outside of market transactions. If Joe buys an apple for $1 and Fred refuses to buy an orange for $1, we know that Joe found more utility in that apple than Fred found in that orange. Repeated transactions can help wash out edge cases. But that really doesn’t work very well for non-market events where money is the universal medium of exchange.
Multiple postulated "utility functions" are mathematically compatible with the same observed behavior (and not just by scaling), so this doesn't show that the person "had a utility function" in any objective or canonical sense.
How do we decide how much certainty is required to convict someone?
I'll digress slightly for illustration: I happen to be allergic to oranges. OTOH, I enjoy eating apples. If someone were to give me a basket of mixed fruit containing both, do you really think that I would have "NO IDEA" how to recognize and remove the oranges to be given away or to recognize and eat the apples? Even if, for some party game, I was blindfolded and offered fruits, the only fact I would need to determine to eat safely isn't some direct comparison between apples and oranges, it's just "Is this NOT an orange?", so if I can reasonably identify the fruit as something other than an orange, I can eat it, but if I can't, I should reject it. Categories don't need to be directly comparable for categorization itself to be determined, sorting to take place, and actions taken on each category to differ.
Now, back to certainty: to avoid Fallacy of Equivocation, let's define exactly what kind of certainty we're talking about. The U.S. Supreme Court has described "beyond a reasonable doubt" (the standard used for criminal convictions) as establishing a "moral certainty" rather than an absolute or mathematical certainty.
"Beyond a reasonable doubt" is the conclusion of a number of presumably reasonable people (the jury), unbiased to the extent practical (by disqualifying those obviously biased), selected to be representative of the community (peers randomly chosen), informed on the facts of the case to the extent necessary and practical to reach a conclusion whether or not there is no other reasonable explanation for the facts presented at trial other than that the accused is guilty.
There's no trading, balancing, or comparing guilt and innocence or justice and injustice anywhere in this process; the reasoning IS categorical, not mathematical. You won't find formulas in criminal trial jury instructions, you find categorical definitions and rules. "Beyond a reasonable doubt" is never defined as some arbitrary percentage of mathematical probability. Defendants are sorted by the jury into the the categories of guilty or not guilty by simple process of following the Definitions in Law and Rules of Evidence according to their own reasoning and common sense regarding what level of certainly they themselves consider reasonable to be applied by and to themselves.
For example:
Pattern Jury Instructions, Criminal Cases, U.S. Fifth Circuit,
District Judges Association (1990).
PRESUMPTION OF INNOCENCE BURDEN OF PROOF, REASONABLE DOUBT
THE INDICTMENT OR FORMAL CHARGE AGAINST A DEFENDANT IS NOT EVIDENCE OF GUILT. INDEED, THE DEFENDANT IS PRESUMED BY THE LAW TO BE INNOCENT. THE LAW DOES NOT REQUIRE A DEFENDANT TO PROVE HIS INNOCENCE OR PRODUCE ANY EVIDENCE AT ALL [AND NO INFERENCE WHATEVER MAY BE DRAWN FROM THE ELECTION OF A DEFENDANT NOT TO TESTIFY]. THE GOVERNMENT HAS THE BURDEN OF PROVING THE DEFENDANT GUILTY BEYOND A REASONABLE DOUBT, AND IF IT FAILS TO DO SO, YOU MUST ACQUIT THE DEFENDANT.
WHILE THE GOVERNMENT'S BURDEN OF PROOF IS A STRICT OR HEAVY BURDEN, IT IS NOT NECESSARY THAT THE DEFENDANT'S GUILT BE PROVED BEYOND ALL POSSIBLE DOUBT. IT IS ONLY REQUIRED THAT THE GOVERNMENT'S PROOF EXCLUDE ANY "REASONABLE DOUBT" CONCERNING THE DEFENDANT'S GUILT.
A "REASONABLE DOUBT" IS A DOUBT BASED UPON REASON AND COMMON SENSE AFTER CAREFUL AND IMPARTIAL CONSIDERATION OF ALL THE EVIDENCE IN THE CASE. PROOF BEYOND A REASONABLE DOUBT, THEREFORE, IS PROOF OF SUCH A CONVINCING CHARACTER THAT YOU WOULD BE WILLING TO RELY AND ACT UPON IT WITHOUT HESITATION IN THE MOST IMPORTANT OF YOUR OWN AFFAIRS. IF YOU ARE CONVINCED THAT THE ACCUSED HAS BEEN PROVED GUILTY BEYOND A REASONABLE DOUBT, SAY SO. IF YOU ARE NOT CONVINCED, SAY SO.
This framing and Robin's both ignore the true purpose and value of a justice system: deterring vigilantism. Allowing vigilantism creates cycles of retribution that make society unlivable, especially as society scales up, so the primary goal of the justice system is to offer sufficient probability of sufficient punishment that the victim and their family are willing to let the state handle it, but with low enough probability of false conviction and punishment that the people don't avoid interacting with the system or simply vote to eliminate it.
This makes the question of the tradeoff a purely empirical one, though I suppose that if we ever develop Hari Seldon's psychohistory we might be able to do it analytically. You pick a process and a standard that seem reasonable and see how the people respond to it. This dovetails nicely with a democratic form of government, because if the people are unhappy with how the system is working, they'll presumably vote to move the pendulum in the other direction and the result will be that the system will oscillate within a window of acceptability.
This seems like a good place for my favorite utilitarian joke: a utilitarian is a person who is generally thought to be endowed with the predisposition to ignore his or her own drowning child in order to push buttons that produce mild sexual gratification and warehouses full of rabbits. A Benthamite is a person who actually would ignore his or her own drowning child to push the rabbit gratification button.
The alleged incomparability of commission and omission is the whole problem of what some people call "Copenhagen approach to ethics" which basically compels inaction in most cases.
A general factor you haven't included, that seems to me to be relevant, is something like "perception of legitimacy of the justice system, especially among the currently law-abiding." I suspect that as this falls, you will see things like:
- an increase in crime, among people currently disinclined towards it
- a decrease in cooperation with the justice system among people who are currently inclined to cooperate with it (e.g. as witnesses, jurors, etc.)
- perhaps a decrease in perception that convicted people are most likely to be guilty
I wonder, even if you increased the sensitivity of the justice system by loosening the specificity (i.e. convicting more guilty people by allowing more innocent people to be convicted), if this could paradoxically cause an increase in vigilantism past a certain point, as general trust in the system falls. (After all, most people won't really get the concept of type 1 and type 2 errors; I expect that if they observe that the system is worse in one way, they will tend to infer it is also worse in other ways, versus trading them off.) I suspect that "personally knowing someone who was falsely convicted" is likely to destroy someone's trust in the system very effectively, in a way even more significant than seeing a criminal get off. So a relevant quantity might be "number of people who personally know a falsely-convicted person", which you might estimate as something like 100x the number of such people.
the things you lose when you're wrongfully-convicted of a hate-crime--friends and spouse (reputation), job, rights (second amendment)--are far worse than the things you lose by being the victim of any crime.
The government can take far more away from you than any criminal, so it's better to make sure there are fewer victims of the government/courts than fewer victims of your neighbors
would you rather see your robber go unpunished or lose your friends/family cause you were wrongfully-convicted of a crime?
would you rather see your robber go unpunished or lose your fundamental-rights now that you have to go through a background-check to exercise them?
would you rather see your robber go unpunished or lose your job cause you were wrongfully-convicted of a hate-crime and they think you're a racist?
would you rather see your robber go unpunished or be expelled from your college cause you were wrongfully-convicted of a hate-crime and they think you're a racist?
Part of the mechanism which is generally implicit (or explicitly assumed away) is that the criminal justice system is run by people, and those people are not entirely trustworthy, and might be tempted to use the law to target people who would otherwise be considered innocent.
A higher standard of avoiding innocent convictions is partly a cost paid to make it harder to use the law to perpetrate crimes.
let's say i've just been wrongfully found-guilty of a crime for a 2-year sentence, but they'll shorten my sentence to 1-year if i bear-false-witness against you for a hate-crime you've been falsely-arrested for that'll get you permanently fired from your job.
Of course i'll do it, and now you've been convicted of a crime you didn't do, because i was convicted of one i didn't do.
" I've seen real numbers that show like half of fine only convictions are actually innocent as are nearly a forth of all no time probation only convictions if a plea deal was involved"
I strongly encourage you to try to publish such numbers.
Even in the context of the civil system, courts enforce rules arbitrarily. Or perhaps it's more accurate to say the enforcement decisions often have a corrupted motivation, since the decisions ARE often predictable.
Most obvious example of this in practice that comes to mind is civil forfeiture. There are no "good cops" or government officials who participate in it or work somewhere that does. It's state-sanctioned organized crime.
When I said we have "rule of law", I was using it in a multiple-meaning way. But it's not such a great joke when I have to explain it, nor is a laughing matter for the people who actually experience what more or less amounts to armed robbery.
Your problem might be that your model is incapable of representing what I suspect is a somewhat common view. One might have the view that, for purposes of determining criminal sentences, we should pretend that the suffering of a guilty person going jail is 0, we should not treat it as a negative thing. If one holds this view, and also thinks that the suffering of innocent people going to jail is bad, then your model just can't represent that view. That might partially account for this 10,000 factor.
Most obviously for the reason we are considering - that some of the people we convict will be innocent, and their suffering (and that of their loved ones) surely counts for something.
Also because, as our system is currently implemented, executing criminals is actually more expensive than locking them up for life.
And in terms of incentives, we want to disincentivize people who commit lesser crimes from escalating to greater crimes, and that only works if greater crimes get greater sentences. If we executed everyone convicted of any crime, or even just everyone convicted of a serious felony, then why wouldn't a rational rapist escalate to murdering their victim? If convicted they get the death penalty either way, and they are less likely to be convicted if the victim is not alive to testify.
The function you need is S(crime, punishment), where 'crime' is the badness of the crime, and 'punishment' is the badness of the punishment. S can't be a linear function, because both too little and too much punishment are bad, while just the right amount of punishment is (in a JUSTICE system, which values justice, which I here take to mean punishing wrongdoers but not the innocent) a positive good.
If you have a beef with the concept of justice, don't use the word "justice". Say something like "niceness" instead.
(The concept of justice is confused in the West, owing to its root in a Greek word which meant "punishment of the guilty and reward to the good", "righteousness", and "the traditional thing to do." Also owing to a 2500-year history of people deliberately abusing the word to impose their politically-preferred interpretation on it.)
Or... you are asking a lot of really hard to asses questions in your polls and everyone is erroring on the side of safety and you've ended up with a really, really large margin of error
I, for instance, cannot figure out how to answer a couple of them, and therefore I haven't
if it's wrong to believe rumors about the covid-vaccine without evidence, isn't it even worse to believe rumors/false-accusations about you or your children without evidence?
Don't people deserve as much respect as the covid-vaccine?
if you tell people they're guilty-til-innocent, they'll assume people are more likely to abuse their freedom (violate others) rather than use their freedom (invent and innovate), and they'll be suspicious of their neighbors---leading to the censorship, disarmament, and short-term projects (copy existing technology), like you see in China
if you tell people they're innocent-til-guilty, they'll assume people are more likely to invent-and-innovate rather than take-advantage-of-others, and they'll trust their neighbors---leading to free-speech, free-trade, new technology and long-term projects (futarchy), like you see in Jewish/english-speaking countries
Your math assumes that both H and S are positive. For people who consider punishing the guilty to be a positive good, S is negative, and hence so is H. Your equation is then wrong, because it counts the harm done to the innocent as good, because HF is < 0.
Actually, for a "justice" system, S will be a nonlinear harm function which is positive when the crime outweighs the punishment or the punishment outweighs the crime, and negative when the punishment fits the crime.
if professors pay union-dues so they're innocent-until-proven-guilty from student-accusations, then don't citizens pay taxes so they're innocent-until-proven-guilty from police-accusations?
shouldn't citizens get the same due-process that govt-employees get?
Wouldn't the risk compensation be totally deleterious to productive activity if you majorly decreased the barrier to innocent being convicted? Me Too would have been a lot more extreme, and public figures generally would be very vulnerable
You might be interested in the "anti-Blackstonian" movement in legal theory. I have in mind in particular Larry Laudan's work (eg, early chapters of "Truth, Error, and Criminal Law"), but he's been influential. While he doesn't model the relevant tradeoffs exactly as you do, what he does is pretty similar in spirit, and he also concludes that the idea that we should very strongly prefer acquitting the guilty to convicting the innocent turns out to be a lot harder to justify than normally thought.
I didn't know about Laudan, thanks for the reference.
Fundamental failure to understand the operative frame of reference.
To punish the guilty is a matter of proportion: does the punishment fit the crime? To punish the innocent who committed no crime is therefore like trying to divide by 0: the result is effectively infinitely disproportionate.
Worse, they aren't merely like values incremented in positive and negative directions. To punish the innocent is categorically different from punishing the guilty, it is itself another moral wrong of commission. Commission and omission are not directly comparable. Justice and injustice can't be traded like a fungible unit of exchange.
Then you finish with an entirely unsupported bit of bad faith mind reading. This entire article is an illogical ad hominem against people who disagree with your preferred policy. Given the title of this space, it doesn't belong here.
Everyone is guilty of SOMETHING.
I worked as a lawyer for a state prison at one point. I’m going for a little vagueness to stay anonymous. The system had an intake process that could not (or at least never was) be used against the new inmate. The point of it was to protect the inmate from people who might want to do him harm. So they asked questions about gang membership, who might have it out for you, and what crimes you committed. Because of the situation, most inmates copped to everything they ever did. While I won’t say that my interactions are a statistical sampling, I can say that I never came across someone who should not have been in prison based on prevailing norms of the time—none of them were free of fairly serious felonies. Note, though, that most of the inmates I encountered were in higher security units. The inmates in minimum security units struck me as a completely different population.
I suspect that the actually innocent are concentrated in (a) homicides, where the victim is not a witness and (b) misdemeanants, who would be in jail rather than prison, and (c) minimum security prisons, which I visited less. That said, the prisoners who were in for some form of homicide had copped to some form of homicide in the interviews. So I might be wrong about that. But death-row inmates also did not show up in the population that I worked with. And that 1 in 25 statistic is about death-row inmates. So I don’t know that it can be generalized to all prisoners. Death-row inmates tend to have been in for a long time, so they might have been in for decades. A lot of the Innocence Project’s successes have been where new techniques (eg DNA) have proven innocence. But those techniques had become more routine in the underlying prosecutions by the time that I was working with prisoners. And as I noted above, the victim can’t usually testify in a homicide case, so the best possible evidence is missing.
When the interview process was first implemented (along with a number of other measures), the amount of inmate-on-inmate violence fell precipitously. It’s obviously hard to tease out causality when multiple things happened at once. And we shouldn’t really be surprised that prisons are violent places—it is where society puts violent people for long periods of time. I don’t think that impugns the validity of the data that the process collected.
Finally, I’d say that the number does not reach the thousands. And it isn’t a statistical sample. So it’s just lots of anecdote.
Then prove it in court.
Or just do whatever the hell you want to whoever you want and call it punishment for whatever crimes they've committed but not convicted for.
One of the above options sustains a civilized society. The other leads to anarchy and mutual destruction. Pick one.
The logical end point of the assumption of guilt is that everyone deserves death. After all, we can't prove a negative, so we might as well all die.
To a utilitarian, all moral actions are directly comparable. That's a feature, not a bug. If it were really true that justice and injustice can't be traded like a fungible unit of exchange, then we should have no justice system, since we would have literally NO IDEA how to make that trade-off.
Shrug. I am not a utilitarian, AFAIK most people are not utilitarians, and I'm unaware of ANY country or legal jurisdiction that has explicitly based their legal system on utilitarian ethics, yet we have generally had little trouble designing and implementing justice systems throughout history without any need for utilitarianism.
So I'm inclined to say you're clearly attempting to apply the wrong framework to the question and demonstrably wrong that justice and injustice must be fungible for anyone to design a working justice system. No matter how good you think your hammer is, that doesn't actually make everything else a nail.
My point holds only for non-utilitarians, not for utilitarians. If you don't think the justice of punishing the guilty can be traded off against the injustice of punishing the innocent, how do you decide how much certainty is required to convict someone? The 2 things are apples and oranges under your point of view, so you can't say anything at all, other than to retreat into the fantasy that your judicial system can be 100% accurate.
Your point holds for utilitarians because no utilitarian has EVER come up with anything even REMOTELY close to an objective, convincing measurement procedure for this fantastical object they call “utility”. Talk of “utility” is no better than New Age talk of “energies and vibrations, man”. The only difference is that it attracts dupes of a slightly different cultural bent.
While you can’t measure a “util,” you can measure preferences for a single person based on what they do. This works well for repeated market transactions. It works much less well for (hopefully) one-time non-market events like involvement in the justice system. But even then, we can say something because a person who committed the offense had a utility function for the conduct that resulted in its expected utility to be higher than the risk-weighted punishment. (Or they were drunk or high or crazy or or or.)
Interpersonal ranking of utility are much harder, and almost impossible outside of market transactions. If Joe buys an apple for $1 and Fred refuses to buy an orange for $1, we know that Joe found more utility in that apple than Fred found in that orange. Repeated transactions can help wash out edge cases. But that really doesn’t work very well for non-market events where money is the universal medium of exchange.
Multiple postulated "utility functions" are mathematically compatible with the same observed behavior (and not just by scaling), so this doesn't show that the person "had a utility function" in any objective or canonical sense.
How do we decide how much certainty is required to convict someone?
I'll digress slightly for illustration: I happen to be allergic to oranges. OTOH, I enjoy eating apples. If someone were to give me a basket of mixed fruit containing both, do you really think that I would have "NO IDEA" how to recognize and remove the oranges to be given away or to recognize and eat the apples? Even if, for some party game, I was blindfolded and offered fruits, the only fact I would need to determine to eat safely isn't some direct comparison between apples and oranges, it's just "Is this NOT an orange?", so if I can reasonably identify the fruit as something other than an orange, I can eat it, but if I can't, I should reject it. Categories don't need to be directly comparable for categorization itself to be determined, sorting to take place, and actions taken on each category to differ.
Now, back to certainty: to avoid Fallacy of Equivocation, let's define exactly what kind of certainty we're talking about. The U.S. Supreme Court has described "beyond a reasonable doubt" (the standard used for criminal convictions) as establishing a "moral certainty" rather than an absolute or mathematical certainty.
"Beyond a reasonable doubt" is the conclusion of a number of presumably reasonable people (the jury), unbiased to the extent practical (by disqualifying those obviously biased), selected to be representative of the community (peers randomly chosen), informed on the facts of the case to the extent necessary and practical to reach a conclusion whether or not there is no other reasonable explanation for the facts presented at trial other than that the accused is guilty.
There's no trading, balancing, or comparing guilt and innocence or justice and injustice anywhere in this process; the reasoning IS categorical, not mathematical. You won't find formulas in criminal trial jury instructions, you find categorical definitions and rules. "Beyond a reasonable doubt" is never defined as some arbitrary percentage of mathematical probability. Defendants are sorted by the jury into the the categories of guilty or not guilty by simple process of following the Definitions in Law and Rules of Evidence according to their own reasoning and common sense regarding what level of certainly they themselves consider reasonable to be applied by and to themselves.
For example:
Pattern Jury Instructions, Criminal Cases, U.S. Fifth Circuit,
District Judges Association (1990).
PRESUMPTION OF INNOCENCE BURDEN OF PROOF, REASONABLE DOUBT
THE INDICTMENT OR FORMAL CHARGE AGAINST A DEFENDANT IS NOT EVIDENCE OF GUILT. INDEED, THE DEFENDANT IS PRESUMED BY THE LAW TO BE INNOCENT. THE LAW DOES NOT REQUIRE A DEFENDANT TO PROVE HIS INNOCENCE OR PRODUCE ANY EVIDENCE AT ALL [AND NO INFERENCE WHATEVER MAY BE DRAWN FROM THE ELECTION OF A DEFENDANT NOT TO TESTIFY]. THE GOVERNMENT HAS THE BURDEN OF PROVING THE DEFENDANT GUILTY BEYOND A REASONABLE DOUBT, AND IF IT FAILS TO DO SO, YOU MUST ACQUIT THE DEFENDANT.
WHILE THE GOVERNMENT'S BURDEN OF PROOF IS A STRICT OR HEAVY BURDEN, IT IS NOT NECESSARY THAT THE DEFENDANT'S GUILT BE PROVED BEYOND ALL POSSIBLE DOUBT. IT IS ONLY REQUIRED THAT THE GOVERNMENT'S PROOF EXCLUDE ANY "REASONABLE DOUBT" CONCERNING THE DEFENDANT'S GUILT.
A "REASONABLE DOUBT" IS A DOUBT BASED UPON REASON AND COMMON SENSE AFTER CAREFUL AND IMPARTIAL CONSIDERATION OF ALL THE EVIDENCE IN THE CASE. PROOF BEYOND A REASONABLE DOUBT, THEREFORE, IS PROOF OF SUCH A CONVINCING CHARACTER THAT YOU WOULD BE WILLING TO RELY AND ACT UPON IT WITHOUT HESITATION IN THE MOST IMPORTANT OF YOUR OWN AFFAIRS. IF YOU ARE CONVINCED THAT THE ACCUSED HAS BEEN PROVED GUILTY BEYOND A REASONABLE DOUBT, SAY SO. IF YOU ARE NOT CONVINCED, SAY SO.
This framing and Robin's both ignore the true purpose and value of a justice system: deterring vigilantism. Allowing vigilantism creates cycles of retribution that make society unlivable, especially as society scales up, so the primary goal of the justice system is to offer sufficient probability of sufficient punishment that the victim and their family are willing to let the state handle it, but with low enough probability of false conviction and punishment that the people don't avoid interacting with the system or simply vote to eliminate it.
This makes the question of the tradeoff a purely empirical one, though I suppose that if we ever develop Hari Seldon's psychohistory we might be able to do it analytically. You pick a process and a standard that seem reasonable and see how the people respond to it. This dovetails nicely with a democratic form of government, because if the people are unhappy with how the system is working, they'll presumably vote to move the pendulum in the other direction and the result will be that the system will oscillate within a window of acceptability.
This seems like a good place for my favorite utilitarian joke: a utilitarian is a person who is generally thought to be endowed with the predisposition to ignore his or her own drowning child in order to push buttons that produce mild sexual gratification and warehouses full of rabbits. A Benthamite is a person who actually would ignore his or her own drowning child to push the rabbit gratification button.
The alleged incomparability of commission and omission is the whole problem of what some people call "Copenhagen approach to ethics" which basically compels inaction in most cases.
A general factor you haven't included, that seems to me to be relevant, is something like "perception of legitimacy of the justice system, especially among the currently law-abiding." I suspect that as this falls, you will see things like:
- an increase in crime, among people currently disinclined towards it
- a decrease in cooperation with the justice system among people who are currently inclined to cooperate with it (e.g. as witnesses, jurors, etc.)
- perhaps a decrease in perception that convicted people are most likely to be guilty
I wonder, even if you increased the sensitivity of the justice system by loosening the specificity (i.e. convicting more guilty people by allowing more innocent people to be convicted), if this could paradoxically cause an increase in vigilantism past a certain point, as general trust in the system falls. (After all, most people won't really get the concept of type 1 and type 2 errors; I expect that if they observe that the system is worse in one way, they will tend to infer it is also worse in other ways, versus trading them off.) I suspect that "personally knowing someone who was falsely convicted" is likely to destroy someone's trust in the system very effectively, in a way even more significant than seeing a criminal get off. So a relevant quantity might be "number of people who personally know a falsely-convicted person", which you might estimate as something like 100x the number of such people.
the things you lose when you're wrongfully-convicted of a hate-crime--friends and spouse (reputation), job, rights (second amendment)--are far worse than the things you lose by being the victim of any crime.
The government can take far more away from you than any criminal, so it's better to make sure there are fewer victims of the government/courts than fewer victims of your neighbors
would you rather see your robber go unpunished or lose your friends/family cause you were wrongfully-convicted of a crime?
would you rather see your robber go unpunished or lose your fundamental-rights now that you have to go through a background-check to exercise them?
would you rather see your robber go unpunished or lose your job cause you were wrongfully-convicted of a hate-crime and they think you're a racist?
would you rather see your robber go unpunished or be expelled from your college cause you were wrongfully-convicted of a hate-crime and they think you're a racist?
Part of the mechanism which is generally implicit (or explicitly assumed away) is that the criminal justice system is run by people, and those people are not entirely trustworthy, and might be tempted to use the law to target people who would otherwise be considered innocent.
A higher standard of avoiding innocent convictions is partly a cost paid to make it harder to use the law to perpetrate crimes.
let's say i've just been wrongfully found-guilty of a crime for a 2-year sentence, but they'll shorten my sentence to 1-year if i bear-false-witness against you for a hate-crime you've been falsely-arrested for that'll get you permanently fired from your job.
Of course i'll do it, and now you've been convicted of a crime you didn't do, because i was convicted of one i didn't do.
Does your math above add up?
" I've seen real numbers that show like half of fine only convictions are actually innocent as are nearly a forth of all no time probation only convictions if a plea deal was involved"
I strongly encourage you to try to publish such numbers.
We most certainly have rule of law. The other two not so much, unfortunately.
Even in the context of the civil system, courts enforce rules arbitrarily. Or perhaps it's more accurate to say the enforcement decisions often have a corrupted motivation, since the decisions ARE often predictable.
Most obvious example of this in practice that comes to mind is civil forfeiture. There are no "good cops" or government officials who participate in it or work somewhere that does. It's state-sanctioned organized crime.
When I said we have "rule of law", I was using it in a multiple-meaning way. But it's not such a great joke when I have to explain it, nor is a laughing matter for the people who actually experience what more or less amounts to armed robbery.
Your problem might be that your model is incapable of representing what I suspect is a somewhat common view. One might have the view that, for purposes of determining criminal sentences, we should pretend that the suffering of a guilty person going jail is 0, we should not treat it as a negative thing. If one holds this view, and also thinks that the suffering of innocent people going to jail is bad, then your model just can't represent that view. That might partially account for this 10,000 factor.
If the suffering of the guilty counts for nothing, why not just kill them all? Much cheaper than jail.
Most obviously for the reason we are considering - that some of the people we convict will be innocent, and their suffering (and that of their loved ones) surely counts for something.
Also because, as our system is currently implemented, executing criminals is actually more expensive than locking them up for life.
And in terms of incentives, we want to disincentivize people who commit lesser crimes from escalating to greater crimes, and that only works if greater crimes get greater sentences. If we executed everyone convicted of any crime, or even just everyone convicted of a serious felony, then why wouldn't a rational rapist escalate to murdering their victim? If convicted they get the death penalty either way, and they are less likely to be convicted if the victim is not alive to testify.
The function you need is S(crime, punishment), where 'crime' is the badness of the crime, and 'punishment' is the badness of the punishment. S can't be a linear function, because both too little and too much punishment are bad, while just the right amount of punishment is (in a JUSTICE system, which values justice, which I here take to mean punishing wrongdoers but not the innocent) a positive good.
If you have a beef with the concept of justice, don't use the word "justice". Say something like "niceness" instead.
(The concept of justice is confused in the West, owing to its root in a Greek word which meant "punishment of the guilty and reward to the good", "righteousness", and "the traditional thing to do." Also owing to a 2500-year history of people deliberately abusing the word to impose their politically-preferred interpretation on it.)
Or... you are asking a lot of really hard to asses questions in your polls and everyone is erroring on the side of safety and you've ended up with a really, really large margin of error
I, for instance, cannot figure out how to answer a couple of them, and therefore I haven't
innocent-until-proven-guilty means
-your action wasn't unlawful until proven not-protected-by-the-constitution (roe v. wade, ninth amendment)
-your action wasn't within the law unless the law says so (fair notice)
-you didn't even intend to threaten or harm others until proven actual-malice
-you didn't know or expect your action to result in harm to others until proven criminal-mind
-you didn't even do the action until proven-guilty-beyond-a-reasonable-doubt
which one do you want to get rid of to reduce crime?
if laws are innocent of violating the constitution until proven guilty, shouldn't people be innocent of violating the law until proven guilty?
shouldn't people be given the same benefit of the doubt that the law-itself gets? shouldn't the law treat people the way the people treat the law?
if it's wrong to believe rumors about the covid-vaccine without evidence, isn't it even worse to believe rumors/false-accusations about you or your children without evidence?
Don't people deserve as much respect as the covid-vaccine?
if you tell people they're guilty-til-innocent, they'll assume people are more likely to abuse their freedom (violate others) rather than use their freedom (invent and innovate), and they'll be suspicious of their neighbors---leading to the censorship, disarmament, and short-term projects (copy existing technology), like you see in China
if you tell people they're innocent-til-guilty, they'll assume people are more likely to invent-and-innovate rather than take-advantage-of-others, and they'll trust their neighbors---leading to free-speech, free-trade, new technology and long-term projects (futarchy), like you see in Jewish/english-speaking countries
Your math assumes that both H and S are positive. For people who consider punishing the guilty to be a positive good, S is negative, and hence so is H. Your equation is then wrong, because it counts the harm done to the innocent as good, because HF is < 0.
Actually, for a "justice" system, S will be a nonlinear harm function which is positive when the crime outweighs the punishment or the punishment outweighs the crime, and negative when the punishment fits the crime.
if professors pay union-dues so they're innocent-until-proven-guilty from student-accusations, then don't citizens pay taxes so they're innocent-until-proven-guilty from police-accusations?
shouldn't citizens get the same due-process that govt-employees get?
Wouldn't the risk compensation be totally deleterious to productive activity if you majorly decreased the barrier to innocent being convicted? Me Too would have been a lot more extreme, and public figures generally would be very vulnerable