Civics · Law · Geopolitics: The Logic of Constitutions

A constitution isn't simply "the longest, highest-ranked law." It differs from ordinary law in function, not just status. It does at least three things. First, it constitutes — it creates the organs of state (legislature, executive, courts) and sets how they form and operate; it draws the state's "org chart." Second, it limits and grants power — marking the boundaries of authority while listing rights that cannot be taken from citizens. Third, it sets the rule for making rules — defining what makes any other law valid, serving as the yardstick against which all legislation is measured. An engineering analogy: ordinary laws are like apps running on a system; a constitution is the boot firmware plus the permissions protocol — it brings the machine up and defines the interfaces before any app can legitimately run. Because it is the "mother rule" of all rules, it must sit above ordinary legislation and be especially hard to change.

• Toward a "limits list": the U.S. Constitution. A spare text, full of "Congress shall make no law…" — negative restraints on government, centered on shielding the individual. This is the classic "negative rights" approach (freedoms secured by government not acting). The cost: it says almost nothing about positive entitlements like education or health care.
• Toward a "value declaration": Germany's Basic Law. Article 1 opens with "Human dignity shall be inviolable," placing a core value as the bedrock of the whole text, and it builds in a "militant democracy" to stop anyone using freedom to destroy freedom. The cost: the more value-laden the text, the wider the room for interpretation and dispute.
• Toward a "social blueprint": many later-developing and socialist constitutions. Beyond structure and rights, they write in development goals, economic models, and state tasks, so the constitution doubles as a "program of action." The cost: writing goals into the constitution invites a gap between text and reality when those goals go unmet.

Should a constitution be "thin" or "thick"? The thin view holds it should govern only the most basic architecture and floor of rights — the shorter the better: fewer clauses mean more stability, less dispute, slower obsolescence, leaving concrete policy to elected legislatures. The thick view holds it should carry a society's core values and social commitments (social rights, environmental duties), or those fundamentals can be swept away by a single government. The thin view sacrifices guarantees on substantive goals, handing more to politics; the thick view "constitutionalizes" matters that are arguably political, potentially letting courts make value choices in place of legislatures.

"A constitution is just the biggest law in the land." Higher rank is only the surface. The real difference is function: ordinary laws exist by the constitution's authorization and are tested against it; the constitution is the source of that authorization and testing. Treating it as a beefed-up ordinary law misses its special status as the "rule for making rules."

Most countries' constitutions are a single, written, supreme document — a codified (written) constitution. But a few — most famously the United Kingdom — have no such document; their constitution is scattered across statutes, court precedents, historical texts, and political conventions, an uncodified (unwritten) constitution. The key point: "unwritten" is a misleading label — most of the UK constitution is written (statutes and case law are text); it just isn't gathered into one book called "the Constitution." What is truly "unwritten" are the constitutional conventions — rules with no legal force that are nonetheless strictly followed in practice. Codification is like "packaging scattered rules into one read-only file": the upside is clear boundaries and a single anchor; the cost is greater rigidity. The uncodified version is like "a protocol maintained by community custom": it evolves flexibly, but its binding force depends on everyone's self-restraint.

Britain is the textbook "unwritten constitution," yet what has run it for centuries isn't disorder but a thick culture of conventions. The vulnerability lies there too: conventions have no legal teeth, so once a key actor chooses not to follow them, with no single codified document to invoke, disputes are harder to resolve.

Pro-codification: putting the rules into one supreme document lets citizens clearly know where the lines are and what rights they hold, gives a definite basis when disputes arise, and makes judicial review easier — certainty is its strongest selling point. Pro-uncodified: a constitution should grow slowly with society; freezing it into a code means freezing one generation's judgment, whereas a convention-based constitution can keep fine-tuning without upheaval, making it more resilient. One trades "rigidity" for "certainty," the other trades "ambiguity" for "flexibility" — which is exactly why attempts to codify an uncodified constitution tend to stall on "which conventions to write down, and who decides."

"An uncodified constitution means there's no constitution, or just oral agreements." Quite wrong. Britain has a great deal of written constitutional law (such as historic acts defining the relationship between Parliament and the Crown) and case law — it simply isn't bound into one volume. What it lacks isn't "text" but "a single, unified, supreme document."

For a constitution to serve as the "fundamental rule," it can't be amendable by a bare majority like ordinary law, or a momentary majority could rewrite the founding bargain at will. So nearly every codified constitution sets a higher amendment threshold, ranging — by difficulty — from rigid to flexible constitutions. Designers have several "knobs" to tighten or loosen: requiring a supermajority in the legislature (say two-thirds), requiring a second body to ratify (a majority of states/provinces), submitting it to a referendum, even building in a time delay (passage by two successive legislatures). The higher the bar, the more stable the fundamental rules — but the more likely they get stuck in the past. The extreme is an "eternity clause": locking certain core principles so that even the amendment procedure cannot touch them — designed to close the most dangerous loophole of all: "legally using the legal procedure to destroy the constitution itself."

• Highly rigid: the U.S. An amendment needs two-thirds of both houses of Congress to propose and three-fourths of the states to ratify — an extremely high bar; in over two centuries only 27 amendments have passed. The upside is rock-stable fundamentals; the cost is that even clearly outdated clauses are nearly impossible to update.
• Relatively easy: India. Most provisions can be amended by a special majority in Parliament; since the constitution took effect in 1950 it has been amended over a hundred times (more than 100 amendments as of 2026), letting it respond fast to social change. The cost is that a sitting majority might rewrite it frequently — so India's courts developed the "basic structure" doctrine (Kesavananda Bharati, 1973), holding that even Parliament's amending power cannot destroy the constitution's fundamental framework.
• Partly locked: Germany. Most of the Basic Law can be amended by a two-thirds vote of both houses, but Article 79(3) declares that core principles — human dignity, federalism, democracy and the rule of law — are unamendable forever. This was a deliberate design drawn from the lesson of the Weimar Republic's "legal path to dictatorship."

Should there be "unamendable" clauses? For (anti-self-destruction): history shows regimes that hollowed out their constitutions and slid into dictatorship through entirely legal amendments; locking down floors like democracy and human rights installs a "fuse even a majority cannot remove." Against (popular sovereignty): by what right does the founding generation get to settle things permanently for every generation after, declaring some things "never to be changed"? That itself defies the democratic logic that "each generation should decide its own fundamental rules" — and who decides which parts count as "core," and might that be abused? One side sacrifices future generations' power to amend in order to guard against the worst case; the other accepts the "self-destruction" risk to protect each generation's sovereignty.

"The harder a constitution is to amend, the better and more stable it is." Not really. Too high a bar and society moves on while the constitution stays put, pushing people to bypass formal amendment and quietly change it through judicial "interpretation" instead — making constitutional change less transparent. Stability and adaptability trade off against each other; there's no "the harder the better" free lunch.

The constitution says "Congress shall make no law infringing a certain right" — but if Congress does exactly that, who stops it? A constitution with no enforcer is just a declaration. Constitutional review is that enforcement mechanism — a body that judges whether ordinary law violates the constitution and voids the offending law. Globally there are roughly two models. Decentralized (American): ordinary courts review incidentally while deciding a concrete case; the power traces to Marbury v. Madison (1803), where Chief Justice Marshall held that courts may declare an unconstitutional law void. Centralized (European / Kelsenian): a dedicated constitutional court monopolizes the power; the origin is the Austrian Constitutional Court designed by Hans Kelsen in 1920, which can also review laws "in the abstract" without any concrete case. The two models hand the key of "final interpreter of the constitution" to different hands.

• Decentralized: the U.S. Any court can review a law's constitutionality within a case; it needs a real dispute to trigger. The upside is review embedded in actual conflict — restrained and reactive; the cost is that it's slow, requires someone to be harmed enough to sue, and different courts may read it differently, needing the Supreme Court to unify.
• Centralized: Germany, Austria, South Korea, etc. A dedicated constitutional court handles review and can hear abstract challenges, with unified, authoritative rulings. The upside is speed, expertise, and consistency; the cost is that the court is more easily drawn into the center of highly political disputes.
• Weak review / legislature-led: the Netherlands, the UK tradition. The Dutch constitution expressly forbids courts from reviewing the constitutionality of parliamentary statutes — the legislature polices itself; Britain traditionally holds parliamentary sovereignty, so courts cannot strike down an Act of Parliament. The upside is leaving fundamental value choices to elected bodies; the cost is one fewer judicial backstop for minority rights.

Constitutional review touches a deep problem: the counter-majoritarian difficulty — by what right can a handful of unelected judges overturn a law passed by an elected legislature? For: precisely because they are unelected, judges stand apart from momentary public mood and can hold the constitutional floor, protecting minorities who might be steamrolled by a majority; constitutional rights with no court to enforce them are a blank check. Against (political constitutionalism): handing a society's most fundamental value choices to a few judges is "judicial oligarchy" deciding for everyone; such judgments should rest with parliament and the voters. Many countries build buffers: Canada's "notwithstanding clause" lets the legislature declare, for a limited time, that a law operates despite certain rights review, keeping part of the final say with politics. The more complete the independence, the weaker the democratic accountability; reclaim accountability and you may weaken the capacity to guard the floor.

"Having constitutional review means courts override parliament and judges have the last word." Designs vary enormously across countries. Some bar courts from touching parliamentary statutes; some add a "notwithstanding" override that lets the legislature reverse a ruling for a time. The strength of constitutional review is a spectrum, not a binary of "judicial dictatorship" versus "no one minding the store."