You're claiming Japanese in America are representative of Japanese in Japan which is not the case. It's straight up racist too.
Though I do not use the example myself, preferring the murder rate example, the parent's suggestion is a reasonable way to control for this inter-society difference. I would bet that Japanese in the U.S. evince rates of criminality more similar to those of Japanese in Japan than to the rest of American society. It's an empirical question; whether merely asking the question is racist or not given your particular sentiments about what's racist is beside the point.
What makes you believe that? If true, what might cause that to be the case?
As for why, there are a few reasons. One interesting one is selection effects (immigrants to a new country are not a random sample of the old country's population). There is culture, of course, probably the biggest factor, as well as other inheritances (material goods / wealth, genes, disease burden...).
I agree with you that we can't tell from a single gene whether someone is likely to be a criminal or not. I'd like you to consider this example, where hopefully I only use premises you already believe:
Some people are predisposed to becoming addicted to drugs. We can make better-than-chance bets about who these people are based on their genomes. We also know that people who are addicted to drugs are more likely to be on the wrong side of the law (and, sadly, are often by definition on the wrong side of the law). So, we conclude that, given someone's genome, we can make better-than-chance bets about that person's likelihood of being on the wrong side of the law.
It's a simplistic example, but hopefully that helps get the idea across. Doing this sort of thing isn't super practical right now, but it will be soon! [oh boy.]
Example of flawed reasoning similar to yours: genes have probably an impact on the way we speak (it's impacting the shape of our tongue somehow, etc etc). So maybe the French have their accent due to some genetic factors, and we could predict someone's accent looking at their genes. Well it turns out in practice, if you have "French genes" and are born and raised in a US environment, you'll have an American accent indistinguishable from someone with "US genes"
The same goes with criminal behavior.
I agree culture is the biggie (education is part of culture).
Your example gets at a correlation that can be used to make good predictions. Better-than-chance bets. The example I gave with drug addiction supposes the reader already believes that genes have a causal relationship with drug addiction. That is, a propensity to be, say, an alcoholic can be in your genome and not a result of environmental factors.
That's some heroic effort in defense of America's high incarceration rate.
Someone else wrote wrote " Not trying to handwave away issues with the Japanese system, but scale seems relevant here. Japan: 40 incarcerated per 100k population [1] USA: 655 incarcerated per 100k population [2]"
So I read your post in that context. The conversation seemed to me to be about whether the American or Japanese justice system is "bad". Fairness would seem to be an aspect of "good" or "bad"... if you weren't trying to argue the American justice system was "good" I've lost the context of what you were trying to say.
In either case, whether it's racist or not is irrelevant to the issue at hand.
Yes. That is the claim that is racist.
I appreciate you engaging. It's fun and important to talk about stuff like this.
Criminality might correlate with ethnicity. Assuming this means causality is the thing that is racist.