Daniel sent us this one — he's been thinking about something that came up in an earlier conversation, about how laws become this unwieldy mess of amended statutes. And he's asking two things. First, who actually sits down at a computer and writes the first draft of a law? Because in practice it's often someone invisible — a parliamentary assistant, a legislative counsel — and that means there's a whole class of anonymous people shaping the rules we all live by. Second, he's asking whether the default style of legislation — that dense, convoluted, borderline unreadable language — has ever been seriously challenged outside of New Zealand. Because if the point of law in a democracy is to have public rules everyone can understand and follow, writing them in a way only specialists can parse seems to directly frustrate that purpose. So where do we even start?
With the people. The invisible ones. And I love this question because it's one of those things where once you see it, you can't unsee it. In the UK, the Office of the Parliamentary Counsel — the OPC — is the body that actually drafts most government bills. These are lawyers, about fifty or so at any given time, working inside the Cabinet Office. They're not elected. They're not on television. But the words they choose — where a comma goes, whether a clause uses "must" or "may" — that's the stuff that ends up binding millions of people.
Fifty people, give or take, writing the first draft of British law. That's a surprisingly small dinner party.
And the thing is, they're extremely good at what they do, but what they do is a very specific kind of craft. It's not about clarity for the average person. It's about precision that can survive judicial scrutiny. Those are not the same thing.
The invisible hand of the legislative draftsperson is actually just trying to make sure the thing doesn't fall apart in court.
And the UK model has been exported all over the Commonwealth. Canada has its own version, Australia has the Office of Parliamentary Counsel, and in the United States, the House and Senate each have an Office of the Legislative Counsel — nonpartisan lawyers who do the actual drafting. The member of Congress says "I want a bill that does X," and these lawyers turn that into statutory language. The elected official might tweak it, but the architecture, the scaffolding — that's built by someone whose name you will never know.
Which makes them the ghostwriters of democracy. The credited author is the politician, but the actual prose is someone else's.
And here's where it gets really interesting. In most parliamentary systems, there's also a whole tier below the formal legislative counsel. Parliamentary assistants, researchers, policy staffers — they're often the ones producing the first rough draft, the "instructions to counsel" document that says "here's what we want to achieve." And those instructions can be more or less detailed. Sometimes the assistant basically writes a draft in plain language and the legislative counsel translates it into legalese. Other times the assistant just hands over a one-page policy brief and the counsel builds the whole structure from scratch.
The authorship is genuinely distributed. It's not one person, it's a chain.
A chain with a lot of weak links, potentially. I was reading about this recently — there was a study out of the University of London a few years back that looked at how drafting errors happen. And a surprising number trace back to the handoff between the policy team and the drafters. The policy people think they've communicated the intent clearly, the drafters interpret it differently, and by the time anyone notices, the bill's already in committee.
That's the legislative equivalent of the telephone game, except the garbled message at the end carries the force of law.
Nobody's really accountable for the garbling because accountability diffuses across the chain.
Let me ask the obvious question. If we've got this whole system built on the premise that legal language needs to be hyper-precise and therefore necessarily complex, is that premise actually true? Or is it just an assumption the profession has never seriously questioned?
That's the second part of the prompt, and it's the one where New Zealand really does stand out. So in the late 2010s and early 2020s, New Zealand's Parliamentary Counsel Office started a pretty ambitious plain language initiative. They didn't just issue a style guide and call it a day. They fundamentally rethought how legislation gets drafted. The goal was to make laws that a reasonably literate person could actually read and understand without a law degree.
What did that look like in practice?
Shorter sentences, for one. Active voice instead of passive. Getting rid of archaic terms like "heretofore" and "notwithstanding anything to the contrary." Using "you" and "we" where it made sense. Breaking up those infamous single-sentence sections that run for half a page. The New Zealanders basically asked: what if we wrote laws the way we write anything else that's meant to be understood?
The radical notion that the governed should understand the rules governing them.
They had some measurable success. Their plain language versions of existing laws tested significantly higher on comprehension metrics. People could actually answer questions about what the law required after reading the revised version. With the traditional version, comprehension rates were abysmal.
I want to pause on that word "abysmal" because I think most people assume legal language is hard but that they'd get the gist. Are we talking about actual tested failure to understand?
There have been comprehension studies going back decades on this. One famous one from the 1980s tested jury instructions — the judge explaining the law to the jury before deliberation. In the standard legal formulation, comprehension rates among mock jurors were around fifty percent. They literally did not understand the instructions they were supposed to apply. When researchers rewrote the same instructions in plain language, comprehension jumped above eighty percent.
Half the jury is guessing. That's not a justice system, that's a lottery with nicer furniture.
Jury instructions are supposed to be the accessible part. That's the law being explained to citizens by a judge in a courtroom. If that's failing, imagine what's happening with the tax code.
I'd rather not. But okay, New Zealand tried this. Did it stick?
It's still ongoing, and it's been a mixed picture. The plain language drafting principles are now embedded in their Parliamentary Counsel Office's standard practice. New bills are drafted with these principles in mind. But there's been pushback. Some in the legal profession argue that plain language can introduce ambiguity — that the old formulations are battle-tested, that they've been interpreted by courts and have settled meanings. Change the wording and you might change the meaning in ways nobody intended.
Which is a fair point, actually. If a phrase like "reasonable person" has two centuries of case law behind it, swapping it for something friendlier could create uncertainty.
And that's the tension at the heart of this whole debate. The legal profession isn't just being obstinate — well, not entirely. There's a real argument that the complexity serves a function. Legal language is a technical vocabulary, like engineering or medicine. You wouldn't ask a doctor to write a prescription in plain language that a patient could self-administer.
No, but I would ask a doctor to explain to me what the prescription does and why I'm taking it. And that's the distinction. The law isn't a private communication between a professional and a client. It's a public document that I'm presumed to know. Ignorance of the law is no excuse, but the law is written so that ignorance is practically guaranteed.
That's the democratic critique in a nutshell. And New Zealand isn't the only country that's tried to address it. Sweden has a long tradition of drafting what they call "frame laws" — broad, principle-based legislation that's meant to be readable. France has had plain language movements within parts of the civil service. Canada's Department of Justice has been running a legislative plain language program since the 1990s.
Canada in the nineties. So this isn't new.
Not new at all. But it's never really broken through to become the default. And I think there's a structural reason for that. The people who benefit from complex legal language — the legal profession, the compliance industry, the lobbyists who navigate the maze for their clients — those are the people with the most influence over how laws get written. The average citizen who just wants to know whether their business complies with the regulations has no seat at that table.
The beneficiaries of opacity have a lobby. The beneficiaries of clarity are everyone, which means nobody in particular.
That's the Mancur Olson collective action problem applied to legislative drafting. Diffuse benefits, concentrated costs. The lawyers who would lose billable hours if laws became self-explanatory — they know who they are and they're organized. The small business owner who would save five hours a month on compliance — she doesn't even know this is a fight she could be having.
Let's talk about what actually makes legislative language the way it is. You mentioned precision, but is that really what's driving it? Or is there something else going on?
There are a few things. One is that legislative drafters are trained to anticipate every possible interpretation and close it off. So you get these incredibly elaborate definitions sections, and you get sentences that try to cover every edge case. The result is text that reads like a computer program written by someone who's terrified of bugs.
Which, to be fair, is not an unreasonable fear. A bug in the tax code can cost billions.
But the second thing is less noble. There's an element of guild preservation. Legal language signals membership. If you can write in that register, you're inside the club. If you can't, you need to hire someone who can. And the profession has a vested interest in maintaining that barrier.
The shibboleth function. It's not just about precision, it's about proving you belong.
The third thing, which I think is underappreciated, is that legislative language is often bad because it's written under terrible conditions. Bills are drafted under immense time pressure, often late at night, with political compromises being hammered out in real time. Someone runs down the hall with a new amendment that has to be inserted, and nobody has time to make it elegant. The result is text that's been stitched together from a dozen different hands, none of whom had the luxury of stepping back and asking whether the whole thing made sense.
The legislative equivalent of a group project where nobody read the final draft before submitting.
The group project becomes law. So you end up with what the plain language movement calls "legislative fog" — a combination of unnecessary complexity, poor organization, and archaic conventions that nobody would choose if they were starting from scratch.
I want to go back to something you mentioned earlier about Sweden and frame laws. What does that actually look like on the page?
The Swedish approach is different from the common law tradition. They tend to write laws at a higher level of generality, stating principles rather than exhaustively listing every application. A Swedish statute might say something like "employers shall ensure a safe working environment," and then leave the detailed specifications to regulations and agency guidance. The statute itself is readable. The detail lives elsewhere.
Which sounds elegant, but doesn't that just shift the complexity from the statute to the regulations? The citizen still has to find and understand the detailed rules.
It does shift the complexity. But it also creates a clearer hierarchy. The statute tells you what the law is trying to achieve. The regulations tell you how to comply. And in theory, you can read the statute and understand your obligations at a high level without wading through all the technical detail.
I suspect in practice, the regulations end up just as impenetrable.
They often do. But the Swedish model at least starts from the premise that the law should be understandable. That's a philosophical difference from the common law tradition, which starts from the premise that the law should be precise and let understanding sort itself out.
The common law approach is basically "we'll write it for judges, and if citizens want to know what it means, they can ask a lawyer who'll ask a judge.
That's not entirely unreasonable in a system where courts are the primary interpreters. But it does mean we've built a legal system that's structurally dependent on intermediaries. You can't directly access the law. You need a translator.
Which brings us back to Daniel's point about democracy. If the rules are public in name only — if they're published but not comprehensible — is the democratic principle actually being served?
I think the honest answer is: partially. The rules are public in the sense that anyone can read them and, with enough effort and expertise, understand them. That's not nothing. A system where laws were secret would be much worse. But "not secret" is a pretty low bar for democratic legitimacy.
"The law is not secret." A ringing endorsement for the statute books.
It's the legal equivalent of "the food is technically edible.
What would a accessible legal system look like? If we took the New Zealand experiment and pushed it further?
I think you'd need a few things. First, plain language drafting as the default, with a requirement that any departure from plain language has to be justified. Second, you'd need official plain language summaries published alongside the authoritative text — not replacing it, but sitting next to it. A lot of countries do this for specific audiences, like small business guides to tax law, but making it systematic would be a real change.
The authoritative version and the human-readable version, like a software license with a plain English summary that actually means something.
And third, you'd need to test legislation for comprehensibility before it passes. The same way we do environmental impact assessments, we could do readability impact assessments. Give the draft to a sample of citizens, ask them what they think it means, and see if they're right.
That would be transformative. And politically impossible, I assume.
The pushback would be enormous. You'd have legislators arguing that it slows down the process, that it introduces another veto point, that the test subjects weren't representative. But the core idea — that we should verify that laws are understandable before we pass them — seems almost obvious once you say it out loud.
Which makes the fact that we don't do it all the more interesting.
There's an assumption baked into the process that if the text is legally precise, it's good enough. The question of whether anyone can actually read it is treated as secondary, almost aesthetic. Like we're talking about font choices rather than the fundamental accessibility of the state's commands.
The musical equivalent of beige wallpaper. Nobody's supposed to notice the prose, they're just supposed to comply with it.
And that works fine if you have a lawyer on retainer. It works less well if you're trying to figure out whether your home business needs a particular license by reading the actual statute.
Which most people don't even attempt. They Google it, find some summary on a law firm's website, and hope it's accurate.
That's the hidden cost of this system. It's not just that laws are hard to read. It's that the difficulty creates a market for intermediaries, and those intermediaries are expensive. Access to law becomes a function of resources. If you're rich, you hire a good lawyer who explains everything. If you're not, you're navigating on your own and hoping you don't make an expensive mistake.
The opacity of legal language is, in effect, a regressive tax on legal compliance.
I think that's exactly right. And it's a tax that nobody voted for. It's just an emergent property of how we've chosen to write laws.
Let me ask you about a counterexample. Are there areas of law where the language has actually gotten simpler over time, either because of a deliberate reform or just because the subject matter forced it?
Consumer protection law is probably the best example. A lot of jurisdictions now require consumer contracts to be in plain language — credit card agreements, mortgage documents, terms of service. There's been real progress there. The European Union's consumer rights directive has plain language requirements. The US has state-level plain language laws for consumer contracts going back to the 1970s.
We've proven we can do it for contracts. Why not statutes?
Because consumer contracts are between private parties, and one of those parties — the consumer — is presumed to need protection. Statutes are between the state and everyone, and the state has historically been less interested in whether everyone understands.
The state is the one party in the transaction that doesn't need to make itself understood to maintain the relationship. It has a monopoly.
If a bank writes incomprehensible mortgage documents, customers might go to a different bank. If the government writes incomprehensible tax laws, you still have to pay your taxes. The feedback mechanism doesn't exist.
The incentive to be clear is entirely political — it depends on citizens demanding clarity and legislators caring enough to provide it.
On legislators believing that clarity is in their interest. Which it often isn't. Opaque laws give politicians flexibility. If nobody's quite sure what the law requires, you can claim it means whatever's convenient at the time.
Strategic ambiguity as a feature, not a bug.
And that's not always malicious. Sometimes you need flexibility to deal with unforeseen circumstances. But the line between necessary flexibility and convenient obscurity is not always clear.
Where does that leave us? New Zealand tried. Sweden has a different model. Consumer law shows it's possible. But the core of the legal system — the statutes that govern everything from criminal justice to environmental regulation — remains largely impenetrable to the people governed by them.
I think there's a deeper question here that the prompt is gesturing toward. It's not just about drafting technique. It's about who the law is for. Is the primary audience for legislation the courts that will interpret it? The lawyers who will advise on it? Or the citizens who will live under it?
The answer, historically, has been the courts. Everything else is secondary.
And that made sense in a world where most people didn't read and the law was administered by a small elite. It makes less sense in a democracy where everyone is presumed to know the law and can theoretically read it online.
The law is published on government websites now. It's technically accessible. But it's published in a form that assumes the reader is a judge.
Which creates this strange situation where transparency has increased — you can find the law more easily than ever before — but comprehensibility hasn't budged. We've solved the distribution problem and left the readability problem untouched.
We've built a high-speed rail network for documents nobody can read.
That's the show. High-speed rail for unreadable documents.
Let's talk about the authorship question a bit more, because I think there's something specific about parliamentary assistants that's worth pulling out. These are often young people, right? Recent graduates, early in their careers. And they're writing the first drafts of legislation that will affect millions.
It's one of the great underappreciated facts of democratic governance. In the UK Parliament, in the US Congress, in legislatures around the world, the typical parliamentary assistant or legislative aide is in their twenties, a couple of years out of university, working brutal hours for not much money. And they're the ones turning policy ideas into statutory language.
The intern who brings you coffee is also bringing you the first draft of the criminal justice reform bill.
I'm exaggerating slightly — the most consequential bills will get more senior attention. But for the vast majority of legislation, the first draft really is produced by someone quite junior, working from instructions that may or may not be clear, under a deadline that is almost certainly unreasonable.
That draft then goes to the legislative counsel, who may be more experienced but is also working under pressure and may not fully understand the policy intent.
The whole system is optimized for throughput, not for quality. And especially not for readability. The metric of success is whether the bill passes and survives judicial review, not whether anyone can understand it.
It's the legislative equivalent of "ship it and we'll patch it later," except the patches are also written by overworked twentysomethings.
The patches are called amendments, and they make the original even harder to read because now you're reading a text that's been modified six times by six different people, none of whom had the authority to rewrite the whole thing cleanly.
Which is how you get those statutes where section 14 subsection 3 paragraph B clause ii references section 7A which was repealed in 2019 but still applies in certain circumstances described in schedule 4.
That's not a parody. That's a real thing that happens.
That's the depressing part.
The citizen who encounters that — the small business owner, the landlord, the parent trying to understand their rights — they just give up. They either hire a lawyer or they take their chances.
Or they rely on conventional wisdom, which is often wrong. "Everyone knows" the law says X, but the law actually says Y, and the gap between X and Y is where the trouble happens.
That gap is partially a function of how the law is written. If the law were clearer, conventional wisdom would be more accurate.
The drafting problem cascades into a compliance problem, which cascades into an enforcement problem. The unreadable law doesn't just fail to inform; it actively misinforms.
It erodes trust in the legal system more broadly. When people feel like the law is a maze designed to trap them, they don't just avoid the maze — they start to question whether the whole system is legitimate.
Which brings us back, in a roundabout way, to Daniel's point about democracy. The rule of law depends on the law being knowable. If it's not knowable, you don't have rule of law. You have rule by lawyers.
That's a strong way to put it, but I don't think it's wrong. There's a genuine tension between the complexity of modern governance and the democratic ideal of laws that ordinary citizens can understand. We're not going to solve that tension in a podcast episode, but naming it matters.
Let's try to pull out some concrete takeaways. If someone's listening and thinking, okay, what would actually improve this? What are the levers?
I'd point to three things. One, make plain language drafting the default in legislative drafting offices, with training and resources to support it. The New Zealand model shows it's possible. Two, require plain language summaries to be published alongside every new law — not as a replacement for the authoritative text, but as an official translation into accessible language. And three, invest in the people who do the drafting. Better pay, better training, more time. Treat legislative drafting as the high-skill profession it is rather than a bottleneck in a political production line.
None of which is politically glamorous. "We will adequately staff the Office of the Parliamentary Counsel" is not a winning campaign slogan.
No, but "you should be able to understand the laws you have to follow" might be.
That's actually not bad.
It's the kind of thing that sounds obvious until you realize how far we are from it.
The opposition writes itself. "Soft on crime." "Vague laws that let criminals off on technicalities." "Bureaucrats rewriting the will of Parliament.
All of which have been leveled against plain language movements at various points. The fear of ambiguity is real, and it's the strongest weapon the defenders of the status quo have.
Even though the status quo is already ambiguous. It's just ambiguous in ways that benefit people who can afford lawyers.
The ambiguity is priced in, as they say.
There's one more thread I want to pull before we wrap. You mentioned that the UK's Office of the Parliamentary Counsel has about fifty drafters. That's for an entire country. I assume other jurisdictions are similarly lean.
The US Office of the Legislative Counsel — which serves the House of Representatives — has about seventy attorneys. The Senate has its own office of similar size. Canada's legislative counsel is maybe forty or fifty people. These are tiny offices relative to the volume of legislation they produce.
We're trusting the literal wording of our laws to a group of people you could fit in a single restaurant.
A modest restaurant. Not even a banquet hall.
They're not just drafting new laws. They're also handling amendments, reviewing private members' bills, providing legal advice to committees. The workload is staggering.
Which is why the quality suffers. Not because the drafters aren't skilled — they're exceptionally skilled — but because they're being asked to do an impossible amount of work in an impossible amount of time. The miracle isn't that laws are hard to read. The miracle is that they're coherent at all.
The legislative draftsperson as the unsung hero of functional government. Working late into the night to make sure the new environmental regulations don't accidentally legalize something horrifying.
I think that's true. These are the people who catch the catastrophic drafting errors before they become law. And they get basically no recognition for it.
Because if they do their job perfectly, nobody notices. The law just works.
It's the most thankless form of legal practice imaginable.
Now: Hilbert's daily fun fact.
Hilbert: In the 1900s, a remote valley in Papua New Guinea was reported to host an annual buzkashi-style competition using coconuts instead of a goat carcass — a practice thought to have gone extinct by the 1930s, until a missionary's journal from 1987 described what appeared to be the same ritual, now performed with a hollowed log painted to resemble a pig.
Hollowed log painted to resemble a pig.
The coconut-to-log transition is really the detail that sells it.
This has been My Weird Prompts. Thanks to our producer, Hilbert Flumingtop. You can find us at myweirdprompts dot com. If you've got thoughts on legislative drafting or anything else we talked about, that's where you'll find us. We'll be back soon.
