What is a source of law? To be a source is to be the originator. The source of a stream comes before the stream. The source of a leak is a necessary condition of a leak. The source of an idea is anterior to the idea.
An Act of Parliament is a source of law. An Act is constituted where a Bill passes both the House of Commons and the House of Lords and receives the Royal Assent from the sovereign. But how do we know this is so? What tells us that this is a source of law?
A beginner might naively think that there ought to be a yet further more foundational document that tells us what the sources of law in a State are. In the United States, Article 1 of the Constitution stipulates that legislative powers are vested in the Congress, which consists of the Senate and the House of Representatives. Copies of this document are on reverential display in the National Archive in Washington.
But why is the US Constitution a source of law? Is there a still yet higher document, kept in an even more secure glass box, that tells us that? If there were we would hit an infinite regress problem.
HLA Hart in The Concept of Law postulated that any system of positive law must have a Rule of Recognition at its base. This rule has as its primary function the provision of a test for the validity of a law. Examples of such rules are that a Bill that passes both Houses and receives Royal Assent is good law, or that the Constitution of the United States is valid. Such rules as they are a necessary condition of what constitutes a positive law cannot, as a matter of logic, themselves be posited. Where then do they come from?
Hart suggested, plausibly, that they arise by Convention among officials (usually judges) whereby they accept such rules. Conventions are a matter of fact and can change overtime. It is a Convention that men do not wear a hat in church. This Convention has not been posited anywhere. Although Hart was the greatest exponent of legal positivism, his account is dependent upon a rule that is not posited by authority at all.
This account works less well in relation to international law. Often there is no official or judge available to determine what international law is as no formal court has been set up to adjudicate on the matter. How did we know in, say, 1648 that the Treaty of Westphalia was legally binding? Because States by Convention treated, and treat, agreements entered into between them as binding.
Where there are multiple sources of law, the Rule of Recognition should (must?) also have rules for determining a hierarchy between them in cases of conflict. The common law, judge made law, is a source of law in the United Kingdom. It is not dependent for its validity upon any Act of Parliament. What happens when the common law says X and an Act of Parliament says not-X? The answer is that the legislation prevails. As a matter of logic, the legislature itself could not have alone validly stipulated that its Acts were to prevail. Further fine tuning is possible. So, the judges now seem to have recognised a special category of ‘Constitutional Statutes” that did not exist when I was young, which seem to have some kind of interpretive priority over other more humdrum Acts.
In the Miller decision the majority seek to meet the interpretive objections of the minority, by postulating that EU law is a source of law, and as such cannot be removed by ministerial action alone. This suggestion first came from Lord Sumption (see transcript at page 40). This seems plausible. After all, UK legislation is without doubt a source of law, and EU law has priority over it within our domestic legal order. EU law can be directly effective as soon as it is posited and is not dependent upon new legislation for its validity. It would be, to say the least, startling if ministers could, by their decisions, remove legislation or the common law as a source of law, so how could they do so with any other source?
Further, describing EU law as a source of law seems to circumvent Lord Reed’s nitpicking legalistic objections based upon the words of the European Communities Act. If EU law is a legal source it is anterior to the positive law itself, and is not determined by it.
Unfortunately, this characterisation of EU law is wrong. If the legislature passed an Act stipulating that the rules of the game Monopoly as they exist “from time to time” were henceforth to be legally enforceable, and to take priority over other rules, would that make those rules a ‘source of law’? They would not. The source would be the rule that Acts of Parliament are valid. The incorporation of the rules of the game would be wholly dependent on that rule.
Precisely the same is true of EU law within UK domestic law. It is only valid because of the European Communities Act. That Act in turn is dependent upon our Rule of Recognition in the UK. EU law is directly effective because the Act says so, and has priority over UK legislation because that is what the Act says and the judges have given effect to that. If EU law had its own freestanding validity it could not be removed even by an Act of Parliament overturning that Act.
At an international level, EU law is not dependent upon any UK legislation. But on that plane it takes its validity from the agreement between member states. It is not a new sui generis form of law.
Lord Sumption’s clever move of describing EU law as a ‘source of law’ has a strong rhetorical attraction. He was the best advocate of his generation. It is a sleight of hand.
spinninghugo 
There can be no controversy that the words in a statute are a legitimate source of law.
The Oxford English Dictionary is a useful source of what the ordinary meaning of words are,
“Referendum”, for example, is defined as:
“1. The process or principle of referring an important political question (e.g. a proposed constitutional change) to be decided by a general vote of the entire electorate; a vote taken by referendum. Cf. plebiscite n.
“In early use chiefly associated with the Swiss constitution.
“2. A written request for instructions from a diplomat to a government. Obs. rare.”
The OED will need rewriting after the Miller Supreme Court judgment.
Amidst all my regrets about the UK withdrawing from the EU, and my disdain for the arguments that took us out, I have to admit that I am glad about it now on one point at least: the special nature that our courts have ascribed to it whenever it suits them will be of no further consequence.
What might the consequences of the Miller ruling be were Brexit to be called off, given this special status beyond coming from parliamentary statute?
Over on my own blog I speculated, before the decision, what the position would be were the UK to conclude a treaty with the imaginary country of Timbuktoo giving the citizens of that country the right to fish on the Thames. Parliament, to avoid the fuss of having to amend the statute every time rules about when/where the people of Timbuktoo might fish on the Thames, chooses to give this effect by a simple statute declaring that the “All rights arising from time-to-time as a result of the Anglo-Timbuktoo agreement shall be available in the UK” or something similar. Of course, the government tires of this treaty for various reasons and then decides to leave it by invoking an exit clause of the agreement.
The problem with this hypothetical, I now see, is that the EU is not seen in the majority opinion as a mere collection of treaties and agreements, but as something beyond that. A simple bilateral treaty would not (I hope) be deemed a new source of law imbued with all sorts of mystical properties, but the EU seems beyond this.