Richard Albert has a new article at SSRN called "Nonconstitutional Amendments" talks about an interesting subject, the deviations that grow between the written text of a document called a constitutions, and the living constitution that operates in practice. But, Albert also misses some steps in my view.
Unconstitutional Constitutional Amendments and Revolutions
Quite a bit of the paper is devoted to discussing the quite narrow concept that not all constitutional amendments that are approved with the proper process are valid that is found in constitutional law in Germany, India and South Africa (by judicial fiat). Canada and other newly formed states have similar hierarchies of especially difficult to amend provisions expressly included in constitutional texts (as does the U.S. Constitution in Article V, a point not mentioned in the article).
The evolution of the doctrine in Germany, which was the source of the innovation in India, was that constitutional amendments must be made by parliament and parliament is not entitled to enact unconstitutional law -- a provision that even a proposed constitutional amendment must adhere to in practice. (California has a similar doctrine distinguishing constitutional revisions from mere constitutional amendments.)
Likewise, the paper makes cursory note of the fact that new regimes (like the current U.S. Constitution relative to the Articles of Confederation) often arise extra-legally as judged by the old regime.
The American Living Constitution
More interesting, but not well developed for my taste, are the discussions of extraconstitutional amendment through practice via constitutional interpretation (superprecedents), ordinary laws of special importance (superstatutes), and historical practice and precedent. This is a rich area, not well enough explored for my druthers in the paper. While a couple of examples are vaguely stated, like the New Deal, the creation of judicial review, the Civil Rights Act of 1964, and the Administrative Procedure Act, the National Environmental Policy Act, many better examples are not, and Albert doesn't do nearly enough to elucidate what distinguishes an judicial, legislative or executive branch act of constitutional significance from those that amount to the day to day business of government. For example, he suggests that almost every U.S. Supreme Court constitutional case amounts to a constitutional amendment and uses that proposition as a reason to disregard non-Article V constitutional amendments via the judiciary almost entirely.
Yet, one really can't begin the inquiry without a working definition of what does and does not constitute the complete living constitution. For example, the living constitution might include the central constitutional provisions, statutes, court cases, regulations and executive branch precedents that establish the core distribution of powers, checks and balances and protections of individual rights vis-a-vis the government in a country.
One of the big ongoing dialogs of U.S. Constitutional law is the extent to which individual rights expressed there regulate action by sate government. The Slaughter House cases said that privileges and immunities clause do not have that effect, and together with the Dred Scott separate but equal doctrine effectively nullified the constitutional civil rights gains secured in the Reconstruction Era after the U.S. Civil War. Later, the selective incorporation doctrine, which used the 14th Amendment due process clause to protect some individual rights with the federal constitution against violation by state and local actors, reversed much of this setback, and cases like Bivens circumvented immunity doctrines that had limited the availability of a remedy for violation of constitutional rights. At the statutory level, the habeas corpus statute, Section 1983, and the Federal Tort Claims Act serve similar "superstatutory" roles. At the state level, Colorado's Governmental Immunity Act has a similar superstatutory role.
The U.S. Civil War itself is widely viewed as having extraconstitutionally settled the question, "Do states have a right to secede?" in the negative. Executive branch actions that are viewed as precedential, the War Powers Act, budgetary politics and the "political question" doctrine have conspired to create the U.S. Constitutional laws of war, with the Courts mostly stepping in to address the constitutionality of detention, discipline and relocation of individuals.
Another group of governmental policies with something close to constitutional character is the decentralized way that the U.S. has structured control of law enforcement. The Posse Comitatus Act has removed the U.S. military from law enforcement. A less specific, but critically important decision has intensely decentralized law enforcement employment. The vast majority of law enforcement officers are employed by local county and municipal governments where they exist, and hence do not report to anyone in the state or federal government directly. State police departments tend to be a very small percentage of all police officers and like federal law enforcement officers often operate primarily in "extrajurisdictional" places like interstate highways, state parks and state buildings. The number of federal law enforcement officers is not great, and federal law enforcement officers are overwhelmingly concentrated in places where no state or substate local government has jurisdiction like federal property, border crossings, and the coastal waters of the United States. Many states popularly elect their criminal prosecutors at a local level. Even U.S. attorneys in the federal system are generally nominated only with the consent of both of a state's U.S. Senators and operate with little guidance from the central government. Similarly, the director of the FBI serves for a long fixed year term, rather than reporting to the President as a purely political appointee. U.S. marshals, in practice, report to the judiciary more than the executive branch, and many states have similar arrangements. This arrangement practically limits the ability of any one person to use force under the color of state law on a large scale in defiance of legitimate civilian authority.
The judicial determination that the one man, one vote principal was mandatory for all sub-federal governments surely has a constitutional character, as imposes strict boundaries on how state constitutions can be written.
A final quibble I have with Albert's analysis is his inclination to describe the United States as a nation with "popular sovereignty," and some other countries as having "judicial sovereignty." The evidence for American popular sovereignty beyond symbolism (where "We the People" substitutes for references to the Crown in prior British political symbolism), simply are not there. The United States is not a nation like modern France or Germany or Russia or Pakistan that has witnessed political change or governmental discipline in the face of pivotal mass protests.
One of the main themes of American constitutional law is that the courts act as a countermajoritarian force, and usually win when they choose to press their case. The President's willingness to call out the National Guard to enforce a court desegregation order in Little Rock, Arkansas is the singular non-judicial precedent establishing judicial supremacy in our constitutional system. Open defiance of court orders by public officials in the United States is rare and almost never prevails for long. Indeed, the American system, unlike that of many countries, not only allows the U.S. Supreme Court (our proxy for a constitutional court) to declare a statute or executive branch action unconstitutional, it affords that power to essentially every local, state and federal judge, subject to appeals, of course.
In contrast, efforts to act out of pure popular sovereignty in the United States usually fail. Shay's Rebellion, not long after the Revolutionary War, was put down. Colorado's Governor just commemorated the Ludlow Massacre in which the government used violent force to put down a budding popular uprising organized by a union against management misconduct; in France this still happens and the union members get away with it, in the United States, the union members get slaughtered when push comes to shove. Every Confederate state conducted elaborate and unequivocal democratic processes amongst their existing franchises at the state level before leaving the Union. American popular poltical culture has vehemently disavowed our popular sovereigntist history of lynchings, with at least a couple of anti-lynching/anti-mob rule literary works included in every school child's curriculum (e.g., The Crucible, To Kill A Mockingbird). Robert Bolt's 1960 British Play, "A Man for All Seasons" has entered the U.S. political pyche as a mascot for the importance of the rule of law, even when it is unpopular or wrong. The U.S. does not permit and has never held national initiatives, referrenda, or even truly national elections (Presidential elections are conducted by state elections officials and aggregated at the national level though the electoral college, not through a tabulation of the national popular vote). For the first half of the Republic's history, even U.S. Senators were not directly elected.
Also, while many parliamentary systems permit or require new elections sooner than would be otherwise required when the current national leadership "loses confidence," as a referendum on the current group of elected officials, the American system of government holds elections strictly according to the calendar without regard to the political reality. In the U.S. system, even a crisis like a failure to adopt a budget causing a government shutdown, or a failure to adopt a plan for new legislative districts following a census, does not provide a basis for ousting the existing political leadership prematurely.