Daniel sent us this one, and it's a legal philosophy twofer. He's asking about the gap between what the law says and what actually happens. Specifically, he wants us to compare two opposing styles of jurisprudence. On one side, societies with intensely granular legal codes where the law functions mostly as a fallback — it's there if someone acts totally unreasonably, but nobody really wants to use it. On the other side, minimalist codes that spell out only what the state will actively prosecute and leave everything else alone. What does the list of counterpoints look like between these two approaches?
This is one of those questions where the surface sounds academic but the implications touch basically everything — AI regulation, trade policy, platform governance. And the timing is interesting. We're watching multiple jurisdictions try to regulate artificial intelligence right now, and they're all making this exact choice. The EU went hyper-granular with the AI Act — a hundred and thirteen articles, more than a hundred and eighty pages. The US is still mostly in "let's see what happens" mode. It's the same tension playing out in real time.
Let's start by defining the poles properly, because I think most people picture this wrong. When we say "granular legal code," the instinct is to imagine an overbearing nanny state. And when we say "minimalist," the instinct is to imagine freedom. But it doesn't map that cleanly, does it?
Not even close. The granular model I think the prompt is pointing at — law as fallback — is genuinely counterintuitive. You have thousands of regulations on the books, but the social expectation is that almost none of them will ever be enforced through formal legal channels. The law exists to be pointed at, not used. It's the legal equivalent of a fire extinguisher behind glass. You're glad it's there. You hope nobody ever breaks the glass.
Or it's like having a very detailed employee handbook that everyone ignores until someone does something truly egregious and HR needs a paper trail. The handbook isn't how the office actually runs. It's the backup system.
And the minimalist model flips this entirely. Instead of having a law for every conceivable situation and then socially agreeing not to enforce most of them, you write down only the things the state is actually willing to prosecute. Everything else is, by design, not the law's business. But here's the twist — minimalism doesn't actually mean less legal complexity. It just moves the complexity downstream.
Let's dig into the granular fallback model first, because I think it's the one that confuses people more. Where does this actually exist in practice?
Japan is the archetype. This is a country with an extraordinarily dense legal code — the 1947 Local Autonomy Law alone has over two hundred and sixty articles governing how prefectural governments should conduct themselves. And here's the number that should stop you in your tracks: fewer than five percent of those articles have ever been litigated in a contested case. We're talking almost eighty years.
Two hundred and sixty plus articles, and the vast majority have never seen the inside of a courtroom. What are they doing there?
They're providing the architecture for what the Japanese call gyōsei shidō — administrative guidance. Here's how it works. A bureaucrat notices a factory is slightly out of compliance with some environmental regulation. Instead of filing a lawsuit or issuing a formal penalty, the bureaucrat calls the factory owner and says, essentially, "We've noticed this issue. It would be appreciated if you addressed it." That's it. Just a phone call that carries enormous social weight.
Because the law is standing behind the phone call.
Everyone knows the formal enforcement mechanism exists. The factory owner could theoretically say no and force the bureaucrat to actually use the law. But that would be a declaration of social war. It would signal that you're not a reasonable person operating within the shared norms. And in a society like Japan's, that signal is devastating. So compliance happens without enforcement. The law worked without ever being "used.
Which is elegant until you start thinking about who gets to decide which laws get enforced and against whom.
That's the dark side, and it's not subtle. Selective enforcement becomes a tool of social control. I was looking at Japan's Immigration Control Act — it was revised in twenty eighteen, more than ninety articles. But enforcement against visa overstayers is highly discretionary. The result is what researchers call a shadow population of "visible invisibles" — people who are technically in violation, everyone knows they're in violation, but the state chooses not to act. Until it does. And when it does, the target is often selected for reasons that have nothing to do with the immigration violation itself.
The law becomes leverage. It's not that the law doesn't matter — it's that it matters selectively, and the selection criteria are opaque.
Which brings us to a really important distinction the prompt is asking us to make. Not all granular systems use law as fallback. Some granular systems actually enforce the granularity. And the contrast case is Singapore.
Singapore is fascinating here because they have the detailed codes but none of the "let's pretend we don't need them" energy.
Look at Singapore's Protection from Harassment Act. It's only eighteen articles — so not even that long by Singaporean standards — but it was actively enforced from day one. Over a thousand Protection Orders were issued in the first year alone, twenty nineteen to twenty twenty. Compare that to Japan's Act on the Protection of Personal Information, the APPI. Eight chapters, seventy-six articles. Enforcement actions averaged fewer than twenty per year until the twenty twenty amendments finally tightened things up.
Same tool — granular written law — but completely different relationship between the text and the application. Japan writes laws as social furniture. Singapore writes laws as operating instructions.
The social contract underneath is totally different. In Singapore, the state is an active manager. The law isn't the backup system. It's the primary interface between citizen and state. The Maintenance of Religious Harmony Act was amended in twenty twenty with fifteen new provisions, and those provisions were used. They weren't placed in a glass case labeled "break only in emergency.
Which raises the question: why do societies tolerate laws that are routinely ignored? If Japan has all these regulations that almost never get enforced, why doesn't someone just...
Because the enforcement isn't the only thing keeping the system running. Social norms are doing the heavy lifting. In high-trust, high-social-cohesion societies, the cost of being seen as unreasonable is higher than the cost of complying with an annoying regulation. The law provides the floor — here's what happens if you're completely unreasonable — and social pressure handles everything above the floor.
Here's where it gets fragile. What happens when the trust erodes? If the whole system depends on people not actually invoking the law, what happens when someone does?
That's the nightmare scenario for the fallback model. A litigious actor enters a system designed around non-litigation. It's like bringing a lawyer to a potluck. Suddenly everyone has to start following the actual rules, and nobody remembers what the actual rules are because nobody's used them in decades. The system wasn't designed for that. It was designed for reasonable people operating in good faith.
Which is maybe the most generous possible assumption about human behavior.
It works until it doesn't. And "until it doesn't" tends to arrive during moments of social stress — economic crisis, demographic change, political polarization. When the shared norms fracture, the fallback model has nothing left to fall back on. You're left with a massive legal code that nobody actually knows how to enforce because enforcement was never the point.
That's the granular model — laws as social furniture, with enforcement ranging from "almost never" in Japan to "constantly" in Singapore. Let's flip to the other pole. And I want to start with the obvious example because it's sitting right there.
The United States Constitution. Twenty-seven amendments. That's the entire operating system for a nation of three hundred and thirty million people. And the structural commitment to minimalism is baked right into the Tenth Amendment — powers not delegated to the federal government are reserved to the states or to the people. That's a minimalist principle stated explicitly.
Here's my question. Does this actually produce less law? Or does it just produce law from a different source?
This is the central misconception about minimalism, and it's the one most people get wrong. Minimalism doesn't reduce legal complexity. It shifts it from legislation to adjudication. The Commerce Clause is the perfect example. Article One, Section Eight. Congress shall have the power to regulate commerce among the several states. And those fourteen words have generated over a thousand Supreme Court interpretations since eighteen twenty-four.
Over a thousand cases from fourteen words. That's not minimalism. That's compression followed by explosive decompression.
You compress the law into very few words, and then the judiciary spends two centuries decompressing it case by case. The total legal output isn't smaller. It's just produced by judges instead of legislators.
Which has implications for who gets to shape the law. If the real legal work happens in adjudication, then access to courts becomes access to lawmaking. And access to courts is expensive.
Let me give you a concrete number. The United States spends about two point three percent of GDP on legal services. Japan spends zero point eight percent. That's nearly a three-to-one ratio. The minimalist system, counterintuitively, is far more expensive. You're paying for all that litigation that fills in the gaps the legislature left.
Who benefits from minimalism? If it's more expensive and shifts power to whoever can afford lawyers, that seems like it advantages the powerful.
There's an argument that minimalism benefits the powerful in a different way than you might think. It's not just about affording lawyers. It's about regulatory gaps. When the law is minimalist, what's not written is just as important as what is. And the absence of regulation can be a feature, not a bug.
Give me an example.
The twenty twenty-three Supreme Court decision in Sackett versus EPA narrowed the Clean Water Act's definition of "waters of the United States." A minimalist reading. The result was that federal protection was removed from millions of acres of wetlands. The gap was intentional — the Court said, essentially, if Congress wanted to protect those wetlands, Congress should have written a law that explicitly did so. The minimalism returned authority to the states. Some states stepped in. Many didn't. The wetlands that fell into the gap are now unprotected.
The gap becomes a policy choice made by silence rather than by explicit decision.
That's the uncomfortable truth about minimalism. It looks like restraint. It sounds like humility. "We're only prohibiting what we explicitly list." But the things you don't list are left to whoever has power in the absence of law. Which is usually whoever already has power.
This is where the European Union comes in, because the EU looked at minimalism and said, essentially, "we tried that and it was a disaster.
The nineteen ninety-five Data Protection Directive. Thirty-four articles. It was supposed to create a unified European approach to data privacy. What it actually created was twenty-seven different national implementations. Each member state interpreted the minimalist text differently. The result was regulatory fragmentation so severe that companies had to navigate completely different privacy rules in France versus Germany versus Italy.
The Brussels Effect, but in reverse. Instead of EU regulation setting global standards, EU minimalism created internal chaos.
When the GDPR arrived in twenty eighteen, it was a direct repudiation of that approach. Ninety-nine articles. A hundred and seventy-three recitals. The message was unmistakable: we are never doing minimalism again. We are going to spell out exactly what we mean, and you are going to comply with the text, not with your local interpretation of the text.
Has it worked?
It's worked in the sense that there's now a single standard. It hasn't worked perfectly — enforcement is still uneven across member states, and the Irish Data Protection Commission in particular has been criticized for being too lenient with big tech companies headquartered in Dublin. But the fragmentation problem is largely solved. You know what the rules are. The rules are just very, very long.
Which brings us to the AI Act. A hundred and thirteen articles, more than a hundred and eighty pages. They're not even pretending anymore.
The AI Act is explicitly designed to avoid the GDPR's early problems. The GDPR had to be interpreted and reinterpreted because even ninety-nine articles left ambiguities. The AI Act tries to close those ambiguities before they open. It's granularity as a response to the failures of... slightly less granularity.
We've got a spectrum. On one end, Japan — hyper-granular codes that nobody uses. Then Singapore — hyper-granular codes that everybody uses. Then the US — minimalist text that generates massive litigation. Then the EU — maximalist text trying to prevent the litigation that minimalism creates.
There's a fascinating middle path emerging that the prompt didn't explicitly ask about but that I think is the most interesting development. New Zealand's twenty twenty-three Natural and Built Environment Act. This was a deliberate minimalist reform. The predecessor, the nineteen ninety-one Resource Management Act, had over four hundred sections. The new act has ninety-eight sections. They cut it by more than seventy-five percent.
Because the old system was bleeding money. Annual litigation costs under the Resource Management Act had reached a hundred million New Zealand dollars. That's about sixty million US. For a country of five million people. The granular code wasn't preventing litigation. It was fueling it. Every section was another thing to argue about. So they tried the opposite — strip it down, write only what matters, and hope that fewer words mean fewer lawsuits.
Which is the exact opposite of the EU's theory. The EU said "more words means less ambiguity means less litigation." New Zealand said "fewer words means less to litigate." They can't both be right. Or they can, and it depends on something else entirely.
I think it depends on enforcement culture. And this is where I want to propose something that might sound counterintuitive. The choice between granularity and minimalism isn't really about legal philosophy. It's about enforcement capacity.
If you have strong, well-resourced, trusted enforcement agencies, you can afford minimalism. You don't need to spell out every possible violation because your enforcers have the capacity and legitimacy to exercise judgment. They can fill in the gaps. But if your enforcement agencies are weak, under-resourced, or distrusted, you're in trouble either way. Granular codes become theater — lots of rules, no action. Minimalist codes become Swiss cheese — lots of gaps, no backstop.
The variable that actually matters is the quality of the enforcement apparatus, not the density of the text.
That's my read. Look at the US again. The Commerce Clause works — sort of — because there's a massive judicial infrastructure to interpret it. The Federal Trade Commission, the Department of Justice, the courts. Without that infrastructure, fourteen words would be fourteen words and nothing more. The minimalism works because the enforcement ecosystem is enormous.
Japan's granular codes work — sort of — because the social enforcement ecosystem is enormous. Bureaucratic guidance, social pressure, shame. Different enforcement mechanism, same function.
Which is why transplanting legal models between societies so often fails. You can copy Japan's Local Autonomy Law. You can't copy the social norms that make administrative guidance function. You can copy the US Constitution. You can't copy two centuries of judicial precedent and a legal services industry worth hundreds of billions of dollars.
Let me ask the uncomfortable question then. Does either model actually produce better outcomes for ordinary people? If you're just a person trying to live your life, not a corporation with a legal department, which system serves you better?
I'm not sure there's a clean answer. In the granular fallback model, you benefit from flexibility. The law doesn't come down on you for minor infractions because nobody wants to be the person who broke the social peace by invoking the law. But you're also vulnerable to selective enforcement. If someone decides to target you, the law is there, waiting. In the minimalist model, you benefit from clarity. You know what's prohibited. But you're vulnerable to the gaps. If something harmful isn't explicitly listed, you have no recourse.
It's a choice between "the law probably won't be used against you, but it could be" and "the law definitely won't protect you from things it doesn't mention.
"probably won't" works great until you're the exception.
Which model is more susceptible to regulatory capture?
Both, but in different ways. Granular systems are vulnerable to capture at the enforcement discretion stage. If a regulator can choose whom to investigate, that choice will be influenced by power. Minimalist systems are vulnerable to capture at the gap stage. If the law doesn't address something, the powerful can operate in that gap without technically breaking any rules. Different vulnerability, same outcome.
The twenty twenty-five California AI Safety Bill — the successor to SB ten forty-seven — is trying something interesting here. They're calling it "tiered minimalism." Twelve pages of core obligations, but with delegated authority to regulators to fill in the details. It's granularity on demand.
Which is probably the smartest approach if you can make it work. Write the principles clearly and briefly. Give regulators the tools to flesh them out as technology changes. But the success of that model depends entirely on the regulators. If they're captured or incompetent, tiered minimalism just becomes minimalism with extra steps.
We keep circling back to the same question: who enforces, and do they have the capacity and legitimacy to do it?
That's the question I think listeners should take away from this. When you see a new regulatory proposal — whether it's AI safety, content moderation, environmental protection — don't just count the pages. Don't just ask whether it's "big government" or "small government." Ask two questions. One: who enforces this, and do they actually have the resources and independence to do it? Two: what happens when the law is ignored — is the fallback credible, or is it theater?
If you can't answer those two questions, you don't actually know what the regulation will do, regardless of how many articles it has.
I want to add a third question that I think is underrated. What happens when someone tries to use the law in bad faith? Every legal system has to deal with the fact that some people will weaponize the rules. In a granular fallback system, the social norm against invoking the law is supposed to prevent that. In a minimalist system, the limited scope of the law is supposed to prevent that. Neither works perfectly.
The bad faith actor is the stress test for both models. In Japan, the bad faith actor is the person who actually sues. In the US, the bad faith actor is the person who exploits the gap.
Neither system has a great answer. Japan's answer is social ostracism, which works until the bad faith actor doesn't care about social ostracism. The US answer is more litigation, which works until the bad faith actor has more money for lawyers than you do.
Let's talk about something the prompt gestures at but doesn't fully unpack. What happens when these two models collide internationally? You've got the EU writing hundred-plus-article regulations, and the US writing much sparser frameworks, and they're both trying to govern the same global technology companies.
This is the Brussels Effect in practice, but it's not as clean as "EU rules win." What actually happens is that companies comply with the most restrictive regime in the markets they operate in, and then they lobby like crazy to prevent other jurisdictions from adopting something even more restrictive. The GDPR became the de facto global standard not because everyone agreed with it, but because it was easier to apply GDPR everywhere than to maintain separate systems.
The maximalist approach wins the global standards war by default, simply because it's more expensive to comply with two systems than with the stricter one.
That only works up to a point. If the EU's AI Act is so restrictive that compliance becomes burdensome, companies might start making choices about which markets to serve. We haven't seen that happen with data privacy yet, but AI might be different. The stakes are higher. The compliance costs are less predictable.
Which brings us to what I think is the most interesting open question. As AI systems start generating legal text — and they will, they already are — what happens to the granularity versus minimalism debate? An AI can write a legal code with ten thousand articles. It can also adjudicate a minimalist code by generating case law on demand.
Algorithmically granular codes. Too detailed for any human to read, but perfectly enforceable by machines. Every possible situation anticipated, every edge case covered, not because a legislature debated it but because a model generated it.
Is that utopia or dystopia?
It's both, and it's probably inevitable. The real question is whether the enforcement is also automated. If an AI writes the law and an AI enforces the law, you've removed human judgment entirely. No bureaucratic discretion. No social norms. That's a completely new category of legal system, and neither the granular nor the minimalist tradition has a framework for thinking about it.
Because both traditions assume human enforcers. The Japanese bureaucrat making a phone call. The American judge interpreting the Commerce Clause. The Singaporean prosecutor deciding to bring charges. Take the human out, and both models collapse into something else entirely.
Rules as code. It's already happening in narrow domains — tax compliance, benefits eligibility, some areas of securities regulation. But the idea of a full legal system operating that way is new. And it would invert the relationship we've been describing. In the current models, the law exists and humans decide how to apply it. In a computational model, the application is the law. There's no gap between letter and application. The letter is the application.
Which solves the selective enforcement problem and creates a hundred new problems we don't have names for yet.
Welcome to jurisprudence in the twenty-first century.
To pull this together for the prompt's actual ask. The counterpoints between granular fallback codes and minimalist codes. We've got enforcement frequency — Japan almost never, Singapore constantly, the US constantly but through courts rather than agencies. We've got cost — minimalist systems shift expense to litigation, granular systems shift it to social maintenance. We've got vulnerability — granular systems risk selective enforcement, minimalist systems risk regulatory gaps. We've got trust — granular fallback models require high social trust to function without formal enforcement, and they break when trust erodes. We've got the international dimension — maximalist approaches tend to win global standards battles because it's cheaper to comply with one strict standard than multiple lenient ones.
We've got the hybrid models emerging. California's tiered minimalism. New Zealand's deliberate simplification. The EU's maximalism as a reaction to the failures of its own earlier minimalism. The real world isn't choosing between the two poles. It's experimenting with mixtures.
If I had to give someone a heuristic for evaluating any legal system through this lens, I'd say: ignore the page count. Look at the enforcement budget. Look at how many cases are actually brought. Look at who can afford to bring them. The text of the law tells you what the society aspires to. The enforcement pattern tells you what it actually is.
The distance between those two things — the aspiration and the reality — that's where the interesting questions live. Some societies are comfortable with a large gap. Japan's legal code is aspirational in ways that everyone understands are aspirational. Other societies find that gap intolerable and either enforce more or legislate less to close it.
The gap itself isn't necessarily a problem. It becomes a problem when people stop agreeing on what the gap means.
That's where we are in a lot of Western democracies right now. The shared understanding of which laws are "real" and which are "theater" is breaking down. People are invoking laws that were supposed to be fallbacks. People are ignoring laws that were supposed to be enforced. The consensus that makes either model work is fraying.
Which is probably why the prompt landed in our inbox. This isn't an academic question anymore. It's the subtext of basically every regulatory fight happening right now.
Now: Hilbert's daily fun fact.
Hilbert: In ancient Mauritius, before the island was even inhabited by humans, a species of giant tortoise and a species of fruit bat developed an unlikely partnership. The tortoise would eat fallen fruit and excrete the seeds across the island, while the bat would pollinate the trees' flowers at night. When early Arab traders first observed this in the fourth century, they called the bat the "sepaktakraw bat" because its mid-air maneuvers while dodging the tortoise's craned neck resembled the acrobatic kicks of the Southeast Asian sport — a sport that wouldn't formally exist for another eleven centuries.
...right.
One last thought before we wrap. The prompt asked about counterpoints between these two styles, but I think the real insight is that the styles are less important than the institutions that animate them. A well-run minimalist system beats a corrupt granular system. A high-trust granular system beats a captured minimalist system. The text is downstream of the trust.
The trust is downstream of the enforcement. If people believe the system is fair, they'll accept either model. If they don't, they'll reject both. The legal architecture matters, but it matters less than the legitimacy of the people operating it.
This has been My Weird Prompts. Thanks to our producer Hilbert Flumingtop. If you enjoyed this episode, leave us a review wherever you listen — it helps. We're at myweirdprompts.com for show notes and archives. I'm Corn.
I'm Herman Poppleberry. We'll catch you next time.
