This is part of my constitutional design series.
In this piece, I am going to lay out a conception of government and law so I have a solid foundation on which to build a constitution. It’s not going to be the most exciting piece in the world, but I think it’s important to define things so it’s clear what I’m even talking about.
What is a government?
The government is an entity with sovereignty. That’s all it is. Like any other entity it has resources (money, property), employees, and internal structure (law, legislature, bureaucratic hierarchy, etc). Unlike every other entity however, it is sovereign, meaning nobody can physically stop the government from doing what it wants to do. This is attributed to the government’s “monopoly on violence”, it having way more force than private actors.
In the west, there’s another thing that’s said to differentiate the government from other entities: the government’s punishment and coercion power. The government can use the threat of violence (usually taking your resources or locking you in prison) to force people to do things, while all private action must be voluntary (or mandated by contracts one voluntarily enters into).
While this tends to be true, it’s not something I would consider super fundamental. Plenty of exceptions can be found: abusive spouses, quasi-monopolies on effectively necessary services, externalities of private action, and a bunch of factors influencing preferences that make you question the concept of “voluntary” to begin with.
So that’s government. At the rate we’re progressing through the title, the piece will be over in no time, right? Well…
Law
What is it?
Most people define laws as rules. I would instead argue that laws are instructions, instructions specifying how the government works. “The Secretary of Health and Human Services, not later than 6 months after the date of enactment of this Act, shall review all infectious and communicable diseases which may be transmitted through handling the food supply…” (from the American with Disabilities Act, just to pick random example) is not really a rule; it’s an instruction.
Some laws are also rules. This definition is not super precise, but generally I would classify any law that applies in many different situations - as opposed to occurring once - as a rule. So “murder is a crime” is a rule because it actually means “when person A accuses person B of murder, the judiciary shall try them in a court using xyz procedure. When someone is found guilty by this procedure, the police shall take them to prison and then guards shall make sure they don’t leave for the specified sentence.”
Writing non-rule laws - what I will call directives - in higher-level law is generally discouraged because the government is supposed to be act consistently, in fair ways that people can work around, rather than arbitrarily. This norm is actually enforced (in part) explicitly by the US Constitution with the provision against bills of attainder, as well as indirectly in the due process clause, equal protection clause, and provision against post facto laws.
Statutory law - law passed by the legislature, what I refer to in most other pieces simply as law - is only one level of law. In the US, above statutes are constitutional provisions, and below statutes are court rulings and then below them executive laws. Everything down to the level of “orders from a government manager to their employee” could technically be considered law, though at lower levels you rarely get rules outside of mandated procedures. Legislative procedure - procedure outlining how the legislature itself should function - is also law. Most law interacts with the levels below it by inclusion, where lower-level law is allowed only as a more concrete implementation of higher-level law. Exclusion is possible as well though, where lower-level laws are allowed unless they violate higher-level laws.1
That being said, to clash less with the conventional usage of terms, I am going to try to avoid the word “law” in this piece except in the general sense. When talking about specific instances, I will stick to the words “provision”, “rule” and “directive”. Here is a list of all words I might use:
So that’s the current system. But why is written law the chosen method for deciding government action? There have to be other ways of doing it, right?
Why Law?
One justification for law, as I mentioned above, is it allows government action to be used consistently and thus worked around. But there is a more fundamental explanation for why law is used - or rather, there are two competing ones.
One explanation is that, since there is no one actor with total control over the government, the only neutral method for determining government action is with written law. Law is the final determinant of government action - I will call this the “final determinant” model of law. This model of law applies well to statutes in presidential democracies like the US. Different branches may get their say in what statutes are - Congress with the ability to propose and vote on statutes, the president with veto power, and the Supreme Court with judicial review - but every government entity must then do what the statute says.
The other explanation is that one actor has full control over the government, and law is merely a convenient tool for stating that actor’s intentions. I’ll call this the “convenient tool” model of law. This fits well for decrees in a dictatorship, and you could even argue it fits for statutory law in a parliamentary democracy. In this model, everyone in government below the actor - the dictator in the case of a dictatorship, or the legislature in the case of parliamentary democracy - behaves to serve the actor. Those who act in ways the actor disapproves of - whether or not they are explicitly violating law - can be overridden or replaced. In parliamentary democracy, this is called “parliamentary supremacy” over the executive and judiciary. Limits to the actor’s power from above, like constitutional review for statutory law, are still consistent with this model even if they’re not the most natural; you could interpret statutes as “the legislature’s will restricted to constitutional actions”.
I think the final determinant model more accurately reflects how statutory law works even in parliamentary democracies. The convenient tool model cannot account for the differing opinions among different people in the legislature2. The convenient tool model also doesn’t make sense in a system where the people in parliament change - a newly elected parliament cannot be expected to change every single law they find suboptimal. That said, the convenient tool model always applies somewhat, as no matter how specific the statute is there will always be some wiggle-room for the person implementing it to make choices, which can lead to changes in statutory law or personnel as a result.
Executive Law
The executive is the institution responsible for, well, executing statutes; literally taking the actions laid out in statutes. But because statutes are rarely at the granularity of “put OJ Simpson in US Federal Penitentiary XYZ”, the executive also creates directives and rules as concrete implementation of statutory law. Alongside execution of law, I think the creation of directives to implement rules - “arrest OJ” to implement “anyone credibly accused of committing murder should be arrested” - corresponds to the usual definition of "enforcement” of law. So I’ll use enforcement to refer both to literal execution and to the creation of directives that lead to execution. Note that by this definition the legislature does some of its own enforcement, though most of it is up to the executive and judiciary.
The specifics of the executive process vary significantly by country, so instead of talking in generalities I’ll comment on the US. I would consider American law coming from the executive to fit with the final determinant model because of civil service protections. Civil service protections mean that government employees can’t be removed just because their views differ from a higher-up’s; if the head executive wants to change policy, they have to issue orders. This means that the executive law itself, rather than the will of the head executive, forms the basis for executive decisions. Executive law is the final determinant of actions of the executive.
Judicial Law
The judiciary’s roles are telling the executive when they are doing something unlawful (“interpreting law”) and signing off every time the executive wants to use punishment power (ie, trials). Neither of these roles are strictly necessary - we could just trust the executive to interpret law and apply punishment power faithfully just as we currently trust the judiciary - but they seem to be a good system we’ve developed over time.
The judiciary’s directives fit with the convenient tool model of law, as the judge will follow up and issue further directives if they feel that a directive isn’t implemented correctly. But - in theory, especially in common law - the judiciary also has non-enforcement legal power in the form of writing case (also called “common”) law, what I will call judicial rules. These are rules for how cases should be decided in the future, often creating legal tests and principles to evaluate statutes. The existence of these rules is based entirely on precedent: if the court doesn’t follow precedent in cases where it disagrees with the precedent, the whole thing is fake. This should follow the final determinant model, where the words the judge wrote are then the official ones put in place, rather than just a tool for expressing the intent of the judge. (Though I think how much this actually holds true in practice is debatable; courts often issue follow-up cases if they feel one case was misinterpreted).
Note that there is no fundamental difference between inclusion and exclusion relations - any exclusion relation could be considered an inclusion relation with the broad clause “any law not disallowed here is permissible”, and vice versa with the clause “every lower-level low must implement something written here”.
Under a system where laws are passed by majority vote, these differing preferences don’t even result in a definite set of preferences for all of parliament - ie, a condorcet winning law. Imagine three factions in parliament - A, B and C - with preferences over three states 1,2 and 3. A prefers 1>2>3, B prefers 2>3>1, and C prefers 3>1>2. The initial state is 1. Then C proposes to pass a law shift to state 2, and both B and C vote in favor so it passes. B then proposes to shift to state 2, and that passes with A and B votes. Finally A proposes to shift policy to state 1, which passes and you are back to where you started. The only solution to this is having sensible people somewhere in the process who don’t do stupid things like this.