The Patch That Was Never Named

Second in the Yukti series. Same argument: across five very different domains, the surface output is observable but the reasoning behind it has never been written down in a form anyone can build on. This one is about India’s 1976 delimitation freeze and the 2026 debate to unfreeze it. The surface is the freeze. What the freeze was supposed to be replaced by has never been named.

Indian politics is primarily an extraction mechanism. The sentence is uncomfortable to write and uncomfortable to read, but most of the arguments I have listened to about Indian governance make more sense once I accept it.

The framing is that every region and every party extracts to roughly the degree of the capacity it has, and that capacity is determined by which group currently sits closest to public money, public contracts, and public discretionary power. The virtue stories we tell each other, about clean Southern governance, about the Hindi heartland’s welfare politics, about reformist parties versus captured ones, are mostly descriptions of who has the larger extraction apparatus at a given moment rather than descriptions of moral difference.

This is not an Indian pathology. Every complex state has an extraction apparatus; what varies is how much the architecture constrains it.

I care about this framing because it flips the delimitation debate on its head. If politics is extraction, delimitation is a question about who controls the taps and what constraints the Constitution puts on them. The extraction logic is regime-agnostic. The architecture is what changes how it plays out.

What the framers built

The Constituent Assembly met in Delhi between 1946 and 1949. They were designing a union of more than three hundred million people across a country that had just been partitioned along lines of language, religion, caste, and economic geography. The thing they worried about most was majoritarianism. In the debate on safeguards for minorities in May 1949, Z. H. Lari quoted Lord Acton’s line that “the one pervading evil of democracy is the tyranny of the majority”.¹ Alladi Krishnaswami Ayyar built a theory of layered safeguards against it.²

The safeguards they chose were architectural. A federal division of powers across Union, State, and Concurrent Lists. A bicameral Parliament with a directly elected Lok Sabha tempered by a Rajya Sabha that was supposed to speak for states. Fundamental rights backstopped by judicial review. A Constitution that was harder to amend than ordinary law.

Population-based representation was in the design. Article 82 mandated delimitation after every decennial census. The framers chose this because they believed a directly elected chamber should reflect where people actually lived.

What they did not anticipate was what would happen if fertility in one region halved relative to another over fifty years. It was outside the operating envelope of their design. Demographic divergence on the scale India has seen since the 1970s was not theorised because nobody had any reason to theorise it. It was not a failure of imagination so much as a different kind of future than the one they were planning for.

The patch that was never named

In 1976, during the Emergency, the 42nd Amendment³ froze Lok Sabha seat allocations at 1971 census numbers. The 84th Amendment⁴ in 2002 extended the freeze to 2026. Both were sold as family-planning incentives: do not punish states that deliver on population control by shrinking their representation.

I want to read the freeze differently.

Functionally, whether or not the drafters intended it, the 1976 freeze was a blunt anti-majoritarian check. Without it, the Lok Sabha would have drifted over fifty years toward a permanent Hindi-heartland bloc. With it, representation was held in place while every other aspect of Indian politics moved. A reader from 1949 looking at this would probably recognise the move. It is precisely the kind of protection the framers built into the Rajya Sabha design. The freeze did in the Lok Sabha, bluntly, what the framers were trying to do architecturally.

The drafters of the 42nd Amendment almost certainly did not think about it this way. They framed it as an incentive. But incentives and architectural safeguards can look identical from a distance. The freeze worked for the reason the framers’ other safeguards worked. It prevented any single demographic bloc from dominating the directly elected chamber permanently.

The 2026 debate is about whether to unwind this patch. That is a coherent question. It is not the question most of the current discourse is asking.

The argument that isn’t being made

The current debate, on both sides, is framed as should we delimit. People who want delimitation describe it as overdue and fair. People who oppose it describe it as punishing Southern states for success. Both positions assume, quietly, that pure population-weighted representation is the constitutional default and that any deviation from it requires special justification.

That assumption is historically wrong. Pure population weighting was never the framers’ design. Every layer of the Constitution they built pulled against it. Pure population weighting was also not what the 1976 drafters delivered, whether or not they said so out loud. Nobody in this country has lived under pure population-weighted representation in the Lok Sabha for any sustained period since independence.

The real question is what explicit mechanism replaces the implicit one we are removing.

If we unwind the freeze and put nothing in its place, we are not returning to the framers’ vision. We are building something neither the framers nor the 1976 drafters designed. A directly elected chamber in which one region has a permanent demographic lock on legislative power. That is a larger constitutional change than either side is admitting.

The conversation needs to move from delimit-or-not to design-the-replacement.

Four candidate patches

If you accept that the freeze was an unnamed safeguard, you have to design its successor with the same honesty. These four patterns are not prescriptions. They are candidate patches. Mechanism families that other federations have used to do exactly the work the 1976 freeze has been doing quietly, each available to us, each needing adaptation to Indian constitutional realities, and more principled than either pure population weighting or permanent freeze.

Degressive proportionality in an expanded Lok Sabha

The European Parliament allocates seats so larger countries get more total seats, but not in strict proportion to population. Smaller countries get a disproportionately higher voice per capita. The principle is codified as degressive proportionality. In plain language: smaller units get a slightly louder voice per person, to stop the largest ones from drowning out the rest. Minimum floor per state, maximum ceiling, inverse ratio of citizens per representative. Combine this with an expanded total seat count, so no state loses absolute representation in the transition. This is the pattern that has started surfacing in the Indian public debate recently. The mechanism is sound, and an obvious candidate for a larger Lok Sabha in which no state loses seats.

A Rajya Sabha with real federal teeth

This is the most underused lever in the constitutional toolkit and the one I would push hardest on. The current Rajya Sabha is a weak revising chamber whose state-representation claim is mostly symbolic. A reformed version would have equal or degressive state weighting regardless of population, hard supermajority thresholds on federal matters, and something closer to the blocking power a real federal upper house has in other mature federations. If you re-weight the Lok Sabha for population, you have to strengthen the Rajya Sabha in the same breath; otherwise you have built a system in which the federal chamber is a rubber stamp. Nobody in the current debate is making this point loudly.

Variable geometry with performance criteria

Different states implement different policies at different depths. This already happens informally. Some states are ahead of central mandates on certain reforms, behind on others. Formalise it. A state that has delivered sustained primary-school literacy gains could opt into a deeper federal education programme at a higher funding tier. A state that has crossed industrial-employment benchmarks could opt out of certain labour-code defaults. A state that has not yet converged on either could receive targeted transfers with explicit conditions. The mechanism is structural, not ideological: a way to let convergence happen at different speeds without forcing the fastest states to wait or the slowest to pretend. Precedents exist for this inside the Indian Constitution, though the current political climate makes it difficult to lean on the specific precedents. The principle is more durable than the cases.

Rotational, non-dominated coordination bodies

When the European Union needs its member states to negotiate policy together, the chair of the relevant Council rotates on a six-month cycle among countries, the secretariat is independent, voting thresholds are high, and disputes go to an external judicial body. No single country has permanent agenda-setting power. Contrast this with how India actually does interstate coordination, and the contrast is instructive.

What the GST Council actually proves

The GST Council was introduced in 2016 as India’s showpiece of cooperative federalism. States collectively hold two-thirds of the weighted votes. Decisions require a three-fourths supermajority. The architecture reads, on paper, as a federal body designed exactly for the kind of multi-state coordination the country needs.

In practice, the Council runs as a Centre-dominated body. The Union finance minister chairs permanently. Decisions are framed as consensus because the voting arithmetic is engineered so that states cannot pass anything unanimously without the Centre. Small states have equal formal voice, which dilutes the weight of larger or higher-contributing ones. High-contributing states have raised dissent on major decisions and been overridden. What was supposed to be a federal coordination body has become a coordination ritual whose outcomes are effectively set in advance.

It is tempting to read this as a specific failure of the current government. The more useful reading is that it is a failure of design. The voting arithmetic, the permanent chair, the absence of an independent secretariat, the lack of a dispute-resolution mechanism that can actually overrule the Centre, these are all design choices that would produce a Centre-dominated outcome under any government. The lesson is not about a party.

This matters for the delimitation debate because any replacement patch for the 1976 freeze will involve some new federal coordination body. If we design it the way the GST Council is designed, it will not do the work we need it to do, no matter what we call it or who runs it. A bad GST Council design is a preview of how a bad replacement patch will behave.

Rent-seeking in a well-designed federation

Rent-seeking does not disappear under good structural design. It shrinks. Visibility and comparability across states shrink it, because voters and investors can see where the extraction is highest and the delivery is lowest. Performance-linked representation shrinks it further, because states have something to gain by converging upward. Rotational coordination bodies shrink it again, because there is no permanent seat at the table from which to extract. Opt-in mechanisms shrink it from the other direction, because lagging states have a real reason to catch up rather than a real reason to slow everyone down.

None of this makes Indian politics less political. It makes political competition converge on outcomes that are harder to hide. That is different from making politicians virtuous, and it is more durable than hoping for virtuous ones.

The third rail

I cannot write an essay about federal representation and not mention language. Every time I have tried to think about delimitation without thinking about language policy, I have run into the same wall. Any reform to the representation formula that ignores language does not survive first contact with Indian reality.

I will not pretend I have a full answer. What I will say: English works as a link language, local languages need genuine autonomy backed by real funding, and Hindi should be one language among many rather than a unification device. Policies that treat language as a unification project have failed. Policies that treat language as a coordination problem have worked. The replacement patch we design for the 1976 freeze has to assume the second framing, not the first. If we build a federation in which representation reflects performance but language policy is imposed from above, we have built the thing we were trying to avoid.

I have a stake in this. I grew up in Kerala, Orissa, and Punjab, which meant learning whichever language the school insisted on, alongside English and the language my parents spoke at home. At the time it felt like an unreasonable amount of work. Looking back, it is the single most useful inheritance India gave me.

I now run workshops in rooms where the primary language is Brazilian Portuguese, Mandarin, or German and English is the bridge. I cannot always follow the words in those rooms. I can usually follow the room. Languages I do not speak are not foreign to me. The way words attach to emotions is mostly common across languages, and what feels exotic about a language usually turns out to be its grammar rather than its meaning. That is the muscle a childhood of switching between languages gave me: the instinct to treat languages as coordination problems before they become identity tests.

The question

The Indian Constitution was designed by people who were paying close attention to the risks of majoritarianism. It was extended in 1976 by people who stumbled into an anti-majoritarian patch while thinking they were doing something else. We are now being asked to decide whether to remove that patch. The honest debate is about what we build to replace it, and whether we are willing to design consciously what the last generation designed without knowing it.

The framers reached for the instinct in 1949. The 1976 drafters carried it blindly for fifty years. The question now is whether we can do it with our eyes open before the next delimitation lands.

The patch was never named. The replacement has to be.

Sources

1. Z. H. Lari, Constituent Assembly Debates Volume VIII, 25 May 1949, in the debate on the Report of the Advisory Committee on Minorities, quoting Lord Acton: “The one pervading evil of democracy is the tyranny of the majority, that succeeds by force or fraud in carrying elections.” The phrase recurs in the debates from other members, including Kazi Syed Karimuddin on 4 January 1949 and Naziruddin Ahmad on 12 September 1949. The full CAD archive (Lok Sabha Secretariat, twelve volumes) is mirrored at constitutionofindia.net.

2. Alladi Krishnaswami Ayyar, profile and selected speeches, Centre for Law and Policy Research. Ayyar chaired the Advisory Committee on Fundamental Rights, Minorities and Tribal and Excluded Areas, and served on the Drafting Committee. His speeches across the CAD develop the layered-safeguards architecture (federal division of powers, fundamental rights, judicial review, amendment difficulty) cited here. For the standard scholarly treatment, see Granville Austin, The Indian Constitution: Cornerstone of a Nation (Oxford University Press, 1966).

3. The Constitution (Forty-second Amendment) Act, 1976, passed during the Emergency. Among other changes, the amendment froze Lok Sabha and state legislative-assembly constituency boundaries at 1971 census numbers, suspending the readjustment that Article 82 otherwise mandates after every decennial census. Official Act text linked from the Wikipedia page; canonical PDF at India Code (legislative.gov.in).

4. The Constitution (Eighty-fourth Amendment) Act, 2001, which came into force in 2002. Extended the 1976 freeze on Lok Sabha and state-assembly seat allocations until the first census taken after 2026. Official Act text linked from the Wikipedia page.