Was the Judge an Idiot?
Day two of the commute argument. Snagged on a verdict I saw on YouTube and rolled into the purpose of punishment and irreversibility, then swerved into aliens.
What set this off
More good banter on the MRT into work today. (With a little dramatization for entertainment value.)
It started with a YouTube video. An honor killing in the US, and the sentence came out lighter than expected because intent to kill wasn't established.
Goyumyum's reaction was immediate.
"Tsk tsk, going all the way to America and doing that. And no intent to kill? What is the world coming to."
Let me nail one thing down before I start. This post isn't a defense of whoever did it. And I don't know whether the video is even accurate. What we saw was a few-minute clip the algorithm dropped in front of us — not a court opinion. (The case itself needs fact-checking.)
What snagged me wasn't the case. It was how fast we finished judging while holding only that much information.
1. Believing only what the internet shows you is risky
So the first thing I threw out was this.
"How would we know whether there was actually intent to kill?"
Intent lives inside someone's head. Even a courtroom can't observe it directly. It gets back-calculated from circumstances and evidence. And we skipped the entire back-calculation and read only a headline summarizing the result.
There's one more thing. The video title had "Muslim" stamped on it. The moment a label goes on, the judgment closes in advance. Same pattern as the word "national disease" shutting down the med school argument last time. The word carries the conclusion in ahead of the reasoning.
Say that out loud and it turns into "so we should never judge anything?" — which isn't it. It's closer to: judge, but scale your confidence to the amount of information you actually have.
2. What is punishment even trying to do?
Here I changed the angle. To feel that a verdict is wrong, you first need a fixed idea of what a verdict is trying to accomplish.
Broadly, punishment splits into two purposes. Rehabilitation and retribution.
The second I put those two on the table, the answer came back.
"RETRIBUTION!!!!!!"
Six exclamation marks. And honestly, I'm on the same side. I think the main purpose of punishment is retribution. Someone died — leading with rehabilitation would be the stranger response, wouldn't it.
But I've heard people in law see this differently. So I tried following why they'd see it differently. That's where it got fun.
3. We aren't gods, and neither is the judge
Retribution needs one premise to work. You have to know exactly how bad the person is.
But we aren't gods. We aren't superintelligent AI either. And neither is the judge. A judge is a human in a robe, not a truth detector.
The problem is what the punishment side is handling. The death penalty, long prison terms — these are irreversible. They don't come back. Lost time doesn't return, and an executed sentence can't be corrected.
Laid out, the structure is this:
- an irreversible disposition
- requested by an imperfect human
- and decided by an imperfect human
That changes the math for a judge. Lean into retribution and there's no recovery if you were wrong. Lean toward rehabilitation and at least room to adjust later survives. So tilting toward rehabilitation isn't judges going soft — it's closer to that side being easier once you put your own error rate into the calculation.
This isn't a morality argument. It's risk management.
Even after following all that, my gut still sits on retribution. But I can see why the other side reads it differently. Viewed as emotion it's retribution; viewed as system design it's rehabilitation. They're not fighting so much as looking at different layers.
And writing it up, the same word came back around. Yesterday irreversibility showed up as the condition for narrative; today it showed up as the character of punishment. Behaving differently depending on whether something can be rolled back is apparently just my default lens.
4. But doesn't the US come down hard to begin with?
You have to check the other direction too. The US is famously heavy on sentencing.
One reason is that charges don't get merged. Rather than bundling several offenses into a single sentence the way Korea does, sentences are requested per charge and stacked. That's how you get numbers like 200 years. (That's the structure as I understand it; details vary by state — needs fact-checking.)
In a system like that, a light sentence may instead be a signal that the evidence didn't come out as strong as assumed. With solid evidence, that system would have gone harder. That's its baseline disposition.
So I threw out a thought experiment.
He brings a shotgun and shoots his daughter. Same verdict?
I don't think so. So what changed? The instrument of the act changed, and when the instrument changes, the back-calculation of intent changes with it. Which means what moved the verdict was probably not leniency but the shape of the evidence.
5. Were the people in that room idiots?
This was the most fun question of the morning.
To call a verdict wrong, you have to assume the judge and jury who produced it were dumber than you. But those people sat through days of testimony in a courtroom. I watched one YouTube clip.
They have overwhelmingly more information. And yet my judgment arrived faster.
Juries can absolutely be wrong. Plenty have been. But "they could be wrong" and "I know better" are different sentences. I only had enough information for the first one.
So I tried respecting their judgment first and reconstructing the case backwards from it.
While forcing an arranged marriage, he choked her to scare her, and she died.
Something around there, maybe. That would explain a light sentence.
This is a hypothesis. I don't know that case. Nobody does except the people in it. All I'm claiming is that this hypothesis explains the observed outcome with fewer assumptions than "the judge was an idiot."
And even if it's right, the daughter is still dead. That part doesn't change.
The response after all that:
"K."
6. And then, suddenly, aliens
We talked about this and that, and the topic jumped.
"Aliens are real! Do you believe it too?"
I do. The universe is effectively infinite, and if the probability of life arising anywhere in it isn't zero, the product pins to 1 almost regardless. On probability alone, not believing is the stranger position.
But there's a problem. "Do you believe in aliens" isn't actually one question. So I split it into levels.
Level 1 — they exist, statistically Life and civilizations are out there somewhere. We've simply never met. This is where I sit.
Level 2 — they visit occasionally UFOs come by, take a look at us, and leave. Contact happens, unofficially.
Level 3 — they already live among us Men in Black style. Plenty of them nearby, we just don't perceive it.
These are three completely different claims. Level 1 is a probability calculation; from level 2 on it becomes a question of how much you trust eyewitness testimony. And level 3 needs an extra mechanism to explain why we never see them. The evidence required jumps at every step.
The fun part is that the gap between level 1 and level 2 is the Fermi paradox. Probability says the place should be crowded, so where is everyone? What you fill that gap with is what decides your level.
So I asked. Which one are you?
"Bleh~ dunno."
Too deep?
"Brain overload~"
End of conversation.
Side thought
Realized after writing: today's two topics were the same axis.
The verdict conversation is about what goes wrong when information is scarce and confidence isn't. The alien conversation is about how far probability can carry you when there's almost no information at all. Both demand the same move — tuning the strength of a claim to how much you actually know.
Split aliens into levels and everyone finds it fun. Point the same question at a verdict — "what level of confidence are you at?" — and it stops being fun immediately. I don't know yet why that is.
Open questions
- Scaling confidence to information sounds easy, but does anyone actually do it while reading the news? I don't.
- Rehabilitation being the easier calculation for a judge is a completely different conversation from the victim's family's side. Is there a way to put both calculations on the same table?
- I know the structure of stacked US sentencing but not the actual guidelines. Worth looking at properly someday. (Needs fact-checking.)
- In the three alien levels, evidence requirements clearly rise per step, but how steeply is something I said on instinct.
And then — Dhoby Ghaut! Today's banter ends here.
The banter with Goyumyum will continue~