What is law?
Last September I listened to a podcast about legal theory called Jurisprudence. It is the audio version of a course with the same name, taught by Prof Scott J Shapiro. I found the podcast interesting and learnt many new things. Below is a summary of miscellaneous concepts, based on some notes I took, using Wikipedia to fill gaps. The course uses his own textbook Legality as the primary material, and comes with a lot of additional reading, all of which I skipped.
There is a related subject called normative jurisprudence, which deals with what law ought to be, whether it is meant to enforce morality or to deter wrongful behavior or to rehabilitate wrongdoers or something else. None of that is covered here. Analytical jurisprudence is about one question: what is law? The course addresses that question from a positivist perspective.
Where does law come from?
Suppose we say that laws come from an entity with legal authority. The question is, for that entity to possess that authority, it ought to have been conferred by some law. Where did that law come from? What happens if we follow that chain of thought all the way?
e.g. My building co-op has some rules and regulations. Our board members can create new rules and amend existing ones. How do the board members get this authority? The board rules say so. Which of these came first: the board or the board rules? Let us say that a higher entity, the city council, has a law which formally recognizes co-ops as buildings that can start by forking a set of rules (provided by that city law), and those rules include an authorized process to start a board if one doesn’t already exist. The city law was created by the city council, and can be amended by them along with all other city laws. Which of these came first: the city law or the council? Let us say that a higher entity, the state government, … which gets its legal authority bestowed by rules in the central/federal government. The highest level of government gets its legal authority from the constitution or some equivalent, and the constitution can be amended or extended by the government. Which came first: the government or the constitution?
How should judges decide cases?
There are two schools of thought.
Formalism says that judges can decide cases based on four core principles:
Determinacy: Law is complete. One can always find some law that is applicable to any case. No need for discretion.
Conceptualism: Law’s completeness is not due to some impossible standard of comprehensiveness. Law is organized in such a way that a judge can learn principles on how to derive the right answer, and then look up what the answer is.
Amorality of adjudication: Judges ought not to engage in moral reasoning.
Judicialism: Judges ought to always apply and never override law.
It looks simple, aspirationally appealing (to me), but also too idealistic to be true.
Realism is a criticism of formalism. Realists say that reality disproves formalism. Judges are not machines, and exercise discretion all the time. Judicial opinions are filled with claims about morality, fairness, virtue, and various values to be promoted. There isn’t always precedence, and even when it does, how narrowly or broadly it is applied may vary based on individual judgment.
There is an extreme form of realism called rule skepticism, which claims that rules are window dressing, and judges are not bound by them. This sounds anarchist, and based on faulty assumptions.
There are no pure formalists anymore. SCOTUS Justice Antonin Scalia was the most famous (impure) formalist in recent times. Most people are somewhat realistic.
How is law related to morality?
Natural lawyers believe that law is inseparable from morality. Morality is objective and discoverable by human reasoning, so the good according to law is whatever is moral. Natural law has been around since ancient times, and often has religious roots. Even today, it plays a big role in theocracies and states that aren’t secular. It may be a receding force in modern and secular states, but it is still around, including at the highest levels of law (e.g. a few SCOTUS justices).
In natural law, there is a higher entity (God or nature), so the chicken comes first. Natural lawyers are not formalists because they believe that judges ought to engage in moral reasoning. They believe that if a law is unjust, then it is not law.
Legal positivists believe that law is a social construct made up by humans. They are not pushing for lawyers and judges to be amoral, just saying that law itself is separate from morality. They believe that if a law is unjust, then it is still law, but we shouldn’t want to obey it, and should work on amending it into something just.
Legal positivism first showed up in the 16th century, developed some meat during the age of enlightenment, has steadily been gaining popularity, and appears to be the dominant force today. Positivists have the challenge of addressing the chicken or egg problem. They are automatically ascribed formalism, because they attempt to cut out or at least minimize the role of moral reasoning. But because they emphasize descriptive explanations to everything, they have the challenge of defending themselves as realists.
Prof Shapiro is a positivist.
Command Theory: Might is Right
In the 19th century, British legal philosopher John Austin came up with a theory: law is the set of rules made by a sovereign. A rule is an expression of a wish backed by a threat to inflict a punishment on failure to fulfill that wish, issued to someone who is willing to act on the threat. A sovereign is someone who is obeyed by most people most of the time, and who doesn’t usually obey anyone else.
This concise theory is describing the world as it is, rather the world as it was. According to it, the sovereign comes first. It can back the idea that law is separate from morality. Austin was an influential positivist.
A century later, another British legal philosopher named HLA Hart wrote The Concept of Law. In it he extensively critiqued Austin’s theory, and pointed out its shortcomings and incompleteness.
First, weak criticisms: Austin apparently uses the word ‘habitual’, but when the sovereign changes, citizens continue to obey even though there is no habit of obeying the (new) sovereign to speak of. Commands imply directness, whereas most laws are written by people who aren’t empowered with carrying out those laws (e.g. legislators, who may outsource it to staffers, lobbyists, and think tanks), and signed by a sovereign who may not even be fully aware of what they are signing. Etc. I found these to be superficial quibbles that do not fundamentally refute Austin and could be addressed by tweaking the theory.
Second, not all laws are coercive. Some laws, like those that deal with crimes, do involve threats of punishments. But there are also laws which are facilitative in that they bestow a benefit. For example, a person upon becoming a parent gains various legal rights on the child (that other people don’t have), and may additionally become eligible to claim some benefits like tax credits on behalf of that child. The laws around co-op formation are granting some powers to a group of households to delegate certain responsibilities efficiently. A more complicated example: filing tax returns is a legal requirement that has both coercive and facilitative elements. We are penalized if we don’t file, or file incorrectly. If we do file, we might receive a refund which we otherwise wouldn’t. Trying to think of benefits as a negative of punishment could work when thinking solely in economic terms, but is kludgy.
Third, one can’t command themselves, so according to Austin’s theory, the sovereign is unaffected by law. But that is not true in reality. The sovereign may have some absolute or qualified immunities, which are essentially benefits bestowed to them that aren’t available to anyone else. The sovereign is also bound by some legal constraints that don’t apply to anyone else.
Fourth, Austin’s theory by being too descriptive misses the full picture. It is not only that citizens do as the sovereign says, but most citizens feel that they ought to do what the sovereign says. They are not motivated by fear alone. Internally, they likely differentiate between how they regard the sovereign and the law, relative to how they would regard a thug (whom they would also obey under duress).
Question: Why did it take 120 years — Austin’s book in 1832 to Hart’s lectures starting in 1952 — for someone to point out all the shortcomings in Austin’s theory? If anyone tried to survey, say a 100 laws each from a few different countries, and verified whether they have the characteristics of command theory, shouldn’t they have been able to identify many of its flaws? Was there ever a retrospective on it, and if so were there any changes in the way people study jurisprudence as a result?
The Hart of the Law
In addition to tearing apart Austin’s theory, Hart proposed his own positivist theory of the law.
Imagine a society without a legal system, which is able to function because everyone follows some customs. Even in equilibrium, it would have three problems:
Uncertainty: A disagreement about rules. If two groups in the society disagree on what the rule should be or which rules do or don’t apply in a situation, they need a peaceful way to resolve it.
Inefficiency: A disagreement about whether a rule has been broken. If an individual accuses another of breaking a rule, and the other disagrees with the accusation, they need a peaceful way to resolve it.
Stasis: Inadequacy of existing rules. Say a new epidemic starts spreading. The society would need a way to quickly adopt some new rules in response, and not have people fighting over whether or not it is sufficient to follow the existing rules alone. Groups can take a while to organically figure out what to do under new circumstances. So it should be feasible to efficiently change rules when needed.
These problems become inescapable with time and as the society grows. They need a general process in order to deal with them non-arbitrarily.
Hart says that law is a solution to these problems. In particular he distinguishes between what he calls primary rules and secondary rules. Primary rules roughly correspond to the rules that the pre-legal society already had, about what a citizen ought to or ought not do in general. Secondary rules deal with procedures for resolving the uncertainty, inefficiency, and stasis in the primary rules. Law is a union of primary and secondary rules.
According to him there are three secondary rules:
Rule of recognition: This is the ultimate rule that defines what makes a rule a rule. It applies only to legal officers, who ought to accept it. It doesn’t apply to anyone else, and it doesn’t matter if they don’t accept it, because they are only subjects of rules, not makers or executioners or adjudicators of those rules. This rule can also explain succession (what happens when the sovereign is incapacitated), precedence (who can override whom), and some other things. The rule of recognition comes first.
Rule of change: This is a power-conferring rule that describes who gets to change rules, the scope of those rules, etc. It applies to what we traditionally think of as the legislature.
Rule of adjudication: This is another power-conferring rule that describes who gets to resolve disputes about rules (often specific instances of those rules), and the dispute resolution method. It applies to what we traditionally think of as the judiciary.
Using these secondary rules, primary rules (and presumably secondary rules too) can be determined, created, and amended.
Hart pushes against the idea of rule skepticism. Rule skepticism requires judges to ignore not only primary rules but also eventually secondary rules, but a legal system cannot exist without secondary rules.
He also defends against criticism that the theory is formalist, and counter-claims that it is in fact anti-formalist. Whereas formalism says there is already always an answer, Hart’s theory accepts that problems of vagueness and incompleteness always exist, and the purpose of secondary rules is to deal with those. Hart claims that judges have discretion and they do some reasoning, but it is social (or policy) reasoning and not moral reasoning. Because they are doing social reasoning, their decisions are final, but not infallible, and the secondary rules help with overriding any such wrong precedent if necessary in the future.
I didn’t buy the social vs moral reasoning point, especially because of what follows in the next section. There were other things to the theory, including about social rules vs social practices, which I either don’t remember or didn’t understand.
Law vs Morality
As the challengers of natural law, an important feature of positivists is to argue that law is distinct from morality, and that the two are not necessarily related. Hart does that extensively.
First, an acknowledgment that there are some similarities between law and morality.
Both are concerned with ensuring conditions for social survival.
Both are binding without consent – an individual in a society is subject to a rule whether or not they agree to be the subject.
Both are more focused on what is wrong, on what one ought not to do.
Both are backed by pressure of social conformity.
But there are big differences.
Morality is often more general. It is concerned with right and wrong, but not necessarily the precise consequences. Whereas law often is clearer about the punishment for wrongdoing.
Morality is concerned only with important social values. Whereas law cares about things like the different rates for taxation, the exact speed limits on different roads, and the intricacies of food labeling regulation.
Morality is immutable, whereas law can be changed.
Morality has an element of voluntariness. It is amenable to extenuating circumstances. Whereas in law, even if these are taken into account for sentencing, circumstances don’t change whether the underlying action is considered legal or not.
Morality is backed by social pressure alone, whereas law is backed by financial and physical sanctions.
Hart also gives some practical reasons for why law is separate from morality.
The main duty of judges is to apply rules, to treat like cases alike. To do that they are typically not seeking morality. The rules being applied may be immoral, e.g. the various kinds of legalized discrimination of social groups. But until they are repealed, they are still legal.
Sometimes judges run out of rules to apply, in which case they are allowed some judicial discretion. As part of that they may do moral reasoning. But the reason for doing that in the first place is because they ran out of law, which implies that it is distinct from morality. As part of their reasoning, they make mistakes, but their decisions are still legally binding.
If we deny that laws that are unjust are not laws, then how do we study them? How do we critique them? It doesn’t make sense to say, “Don’t follow this law, because it isn’t a law.”
I somewhat disagree, especially about whether morality is immutable and flexible. Many things that were not considered immoral by our ancestors are considered immoral by us. Future generations will hopefully develop similar views about some things that are not considered immoral by us today.
Naturally some natural lawyers criticized Hart’s theory. Lon Fuller had some criticisms related to the interpretation of rules. I didn’t understand them, and Hart was apparently able to refute or at least tweak his theory to satisfy Fuller. Ronald Dworkin had many detailed criticisms, which I found easier to understand. The following was his main criticism.
Laws have two parts: the text of the law, and its grounds for legal validity. The law applies if the grounds are valid, and otherwise doesn’t. According to Dworkin, there can be two kinds of disagreements about a law. Empirical disagreements about whether the grounds are satisfied, and theoretical disagreements about what the grounds should even be. Theoretical disagreements are important, and happen all the time. There was a controversial and complicated theoretical disagreement in the US a couple of months ago, but let us go back a few decades and consider his fun example of TVA v. Hill.
In the 70s, the Tennessee Valley Authority (TVA) started constructing a water dam that cost more than a hundred million dollars. Law student Hank Hill, and his friends, found a tiny fish of a species called the snail darter in the neighborhood where the dam was being built. They felt that the construction would destroy a critical habitat of an endangered fish, and requested for the construction to be stopped. TVA declined. The two parties couldn’t settle the dispute and went all the way to the US Supreme Court. Hill’s argument was that federal projects are not allowed to jeopardize any endangered species according to the Endangered Species Act of 1973 (ESA). TVA’s argument was: hundred million dollars!
Dworkin calls this a theoretical disagreement. Both parties agree about what the ESA says, including that it allows no exceptions. Both parties agree that the dam construction would jeopardize the snail darter habitat. The disagreement is about whether the ESA should be waived in this particular case. In a positivist legal system, grounds of law are defined by consensus, and therefore the plain-text reading of a law is always true, and therefore theoretical disagreements are impossible. Positivism has no way to deal with such disagreements, and so it is bad. (I didn’t fully follow this chain of logic.)
In the aftermath, the supreme court took the plain-text reading and ruled 6-3 in favor of the environmentalists. The US government then amended the ESA and created a committee empowered to deal with exceptions. The committee unanimously said that the dam was not only environmentally bad but economically impractical, and stopped its construction. A few years later, the project was once again unblocked, and this time the dam was fully constructed. Meanwhile the snail darters remained indifferent to human indecision, multiplied plentifully, and petitioned to be removed from the endangered species list. Today their conservation status is closer to ‘least concern’ than to ‘extinct’. The only lasting damage was to Hart’s theory.
Prof Shapiro agrees that this is a valid and very serious criticism, and says that positivists remain conspicuously silent about it.
Another unresolved criticism is that Hart’s theory runs into the is-ought problem. That the rules exist doesn’t mean legal officers ought to follow them. But in a legal system you gotta have some oughts.
The Planning Theory of Law
In this century, Prof Shapiro proposed his own positivist theory: legal systems are self-certifying compulsory social planning organizations with a moral aim.
The insight it starts with is that human beings are not only rational creatures but also planning creatures. The general characteristics of such planning include deciding ends, incrementally settling on means to achieve those ends, and forming partial (and nested) plans that gradually become clearer, all the while being flexible to changes. e.g. If two people plan on meeting for lunch, they don’t immediately settle on a plan like, “Let us meet at hour H of day D at place P and eat E1, E2, and E3.” Instead they may first agree to meet up sometime, then find out when they are both free to meet and settle on the meeting time, then choose where they want to meet based on some criteria, then optionally make a reservation, and typically don’t wait until the moment they are sitting at the table and turning the pages of the menu for deciding what to eat.
According to Prof Shapiro, legal activity is very much like this, a sophisticated way of making plans.
Social planning: The legal officials – legislature, courts, bureaucrats – are all involved in the planning. Their plans apply to non-officials, the civilians, as well. The plans (laws) are rarely fully fleshed out from the beginning. Many start as an idea, and reach the next stage only if they are sufficiently popular among the people or the planners. They are often still vague at the time of being passed by the legislature. The bureaucrats will concretize the plans at the time of implementing them. Disagreements about certain steps in those plans can result in going to the courts, where the judges use their own understanding of the plan and bring clarity to the disputed areas. If the legislature disagrees with the judges, they may retroactively amend the law with something specific addressing and clarifying the dispute area. (Note that this applies very well to the TVA v. Hill case from the previous section.)
Organization: The legal officials are conducting their planning activity in an office-like setting, where turnover is common and expected. The activity is independent of the individual officials involved. As long as they agree to a shared plan, it doesn’t really matter what the motives of any of them are. Which is why it is said to be in an institutional capacity.
Compulsory: no one can opt out of the legal system.
Moral aim: The fundamental aim of law is to rectify the moral deficiencies in legality. When a community faces serious moral problems whose solutions may be complex, contentious, or arbitrary, the legal system gets involved because it is sophisticated in coordination. Its function will be to efficiently meet the moral demands of those circumstances. Though the aim is moral, law may fail to be, and so the primary criticism that legal systems face is about the immorality of laws.
Self-certifying: This is the main distinction between legal systems like governments (at any level), and private actors like my co-op board. Although city law must comply with state law, and state law with federal law, in each case the higher authority gives the legal system below it the benefit of doubt, by presuming that the laws they pass will be in compliance. The lower governments can pass and enforce the laws they want without explicitly taking permission from the higher level. If someone believes that there has been a violation, they can sue, and such exceptions will be adjudicated by the higher courts. However, my co-op board as a private actor cannot enforce many laws on its own, even if it passed them, and instead will have to involve the city government (the police or a housing-related body) or specialist third-parties (real estate lawyers, eviction management companies, etc).
Planning isn’t a moral activity, and is based solely on social facts. So this theory is positivist. According to it, at the root is the norm of planning. It doesn’t impose any moral obligations or powers on anyone, including legal officials. It instead generates norms associated with planning. It allows for plenty of vagueness, and so it is anti-formalist.
I like this theory more than Austin’s or Hart’s. On the other hand, I don’t know anything about the criticism it may have received, and it is unclear to me how popular this is.
Questions: Is the social planning characteristic too close to democracies, or can it sufficiently explain the way law works in other political systems as well? Is it true that morality is the aim? In the US, the most important problems facing the country according to polls have historically been economic issues that are not directly driven by moral aims. Despite all the coverage that cultural issues get, the primary objective of governments, the majority of time (debating and passing laws) and resources spent, and the primary criticisms (reasons for unpopularity, losing elections) are not moral, but directly relate to the (mostly economic) hardships experienced by citizens.
Inclusive vs Exclusive Legal Positivism
A few weeks before I started listening to this podcast, I had listened to some random thing, where Prof Shapiro spoke passionately about being an exclusive legal positivist and made fun of the inclusive kind. I had understood all the words separately, but not together. The debate sounded like a tabs vs spaces debate which software engineers have managed to recognize as bikeshedding and avoid. I couldn’t understand what the debate was about even after someone explained it to me. I discovered an hour long episode about it in this podcast, and I thought the 12 hours that precede it might be sufficient background. I remain confused.
The dispute is about two theses. Ultimacy thesis says that legal facts are ultimately determined by social facts alone. Exclusivity thesis drops the word ultimately from that sentence. Which of these is valid? The background is that law is incomplete. For some hard cases, judges have to exercise discretion. The way they do it is often by looking towards morality. So would it be right to say that morality is a part of law?
Inclusive positivists say yes. Most laws are exclusively dependent on social facts. If there is a social norm that says that judges should try to be moral, then they should when law runs out. The connection between law and morality is social. As long as a law is derived through the secondary rules, it is fine. The vast majority of positivists, including Austin and Hart, are inclusive.
Exclusive positivists vehemently disagree. They say that all laws are exclusively dependent on social facts alone.
Joseph Raz, an exclusive positivist, argued that just because a judge may be under an obligation to follow some third-party rules does not make those rules part of the legal system. e.g. Many countries recognize religious freedom, and as such judges are obliged to consider certain exceptions based on someone’s religion. Countries like India and USA even allow some religious exemptions. But it doesn’t mean that religious laws are a part of the laws in those countries.
Prof Shapiro has a simpler argument which is that laws can be repealed, but moral principles cannot be, and therefore laws cannot be moral. My disagreement with this is similar to my disagreement with Hart about morality being immutable.
Prof Shapiro sounds very worked up about this, which is why I think I still didn’t understand the disagreement. People often justify what they advocate for based on moral reasoning, even when they are only motivated by self-interest. This is also true when it comes to laws, and done by not only activists and politicians but also other principal actors and groups participating in society. I imagine that is unlikely to change even if they agree that law is exclusively positive.
Thinking of analogies: Should the course and podcast have only been titled Analytical Jurisprudence because it excludes normative jurisprudence? Is Python a purely object-oriented language because everything in it is an object, or is it not because not every object has attributes or has methods or is subclassable? What are the stakes here?
Why do we care?
The podcast doesn’t discuss any applications of analytical jurisprudence. This is purely my speculation and wishful thinking.
Even though I see one as prescriptive and the other as descriptive, both are advocating for certain values.
Natural law theories are usually prescriptive. They start with a model legal system that is based on some core moral principles, and explain how various real world systems fit into it. This makes for somewhat rigid frameworks. I worry that some legal systems in the world would be considered illegitimate according to this. Historically infringement of liberty within a state and sovereignty of other states could be justified in this way. My opinion may be unfair, and is weakly held. The course covers only one natural law theory. A course-length treatment of analytical jurisprudence by a natural lawyer might be interesting, especially if the debate between the two is far from settled.
Legal positivist theories are descriptive. They are less constrained by morality. They can look at legal systems that exist today, and work backwards into an abstraction that fits a greater number of them. Positivist theories may be better in recognizing states with very different legal systems as legitimate, in separating religion from governance, and for governing pluralist societies. They may have the downside of having to tolerate states that we think are immoral, but they have the upside of peace and cooperation.