Preventive detention law in India begins with a paradox that no summary of a single statute can capture. A person may be taken into custody without charge, without trial, and without any offence having been proved, on the satisfaction of an executive authority that the person is likely to act in a manner prejudicial to specified public interests. In most constitutional democracies that power either does not exist in peacetime or survives only as an emergency residue hedged by judicial control. In India it is written into the fundamental rights chapter of the Constitution itself, in Article 22, and it is surrounded there by a set of safeguards whose most protective clauses are expressly withdrawn from the very class of persons the power targets. Understanding this field means understanding that architecture as a system: a constitutional root, a legislative split between the Union and the states, a family of central and state statutes built on top, and a supervision layer made of Advisory Boards and constitutional courts. Any single detention order, passed under any of these statutes in any part of the country, occupies a determinate place in that system, and the reader who can see the system can place the order, identify the safeguards that apply to it, and judge whether they were honoured.

The namable claim of this guide is the three-month clock and the twin exclusion. Every question in this field reduces to two constitutional facts. The first is that Article 22, clause (3), switches off the ordinary arrest protections of clauses (1) and (2) for a person held under a detention law, so the detenu has no constitutional right under that Article to be told the grounds of arrest as an arrested person is, no right under that Article to consult and be defended by a legal practitioner of choice, and no right under that Article to be produced before a magistrate within twenty-four hours. The second is that Article 22, clause (4), sets a clock of three months, beyond which detention may continue only if an Advisory Board, a body of persons of High Court judicial standing, reports that there is sufficient cause, so the authorisation for longer custody comes from a board and not from a court. Get those two facts right and the rest of the field, the statutes, the grounds, the procedure, and the case law, arranges itself around them. Get either of them wrong, as most short accounts do, and everything built on top is wrong as well.

The Problem Preventive Detention Law in India Was Built to Solve

The problem the power answers is a real one, and the strongest accounts of this field begin by conceding it before they criticise the answer. Ordinary criminal law is retrospective. It waits for an act, gathers evidence of that act, charges the person, and tries the charge, and its coercive machinery, arrest on accusation followed by remand and bail, is designed around an act that has already happened. A government confronting a person it believes is organising violence, running a smuggling network, or preparing sabotage faces a gap between what it suspects and what it can prove to the criminal standard, and the gap is widest precisely where the anticipated harm is gravest. Preventive detention is the instrument built for that gap. It authorises custody on a prognosis rather than on a proof, on the detaining authority’s satisfaction about future conduct inferred from past conduct, and it therefore trades the central protection of criminal process, the requirement of a proved act, for an executive judgment about risk.

That trade is the tension inside every provision that follows. A prognosis cannot be tested the way a charge is tested, because the predicted act has not occurred and may be described only through intelligence, informer material, and a pattern of earlier behaviour whose significance the authority asserts and the detenu disputes. The evidence is frequently of a kind that cannot be shown to the detenu without burning the source, and it is never required to reach the standard a criminal court would demand. The Constitution’s response to this tension was not to forbid the trade and not to leave it unregulated either. It was to write the power into Article 22 alongside a bespoke set of procedural safeguards designed for exactly this kind of custody: communication of grounds, an opportunity to make a representation, and an Advisory Board gate at three months. The whole subsequent history of the field is an argument about whether those safeguards, so designed, are adequate to the trade they license.

A second feature of the problem explains why the field cannot be read as ordinary criminal law with harsher edges. Detention under these statutes is not punishment, and courts repeat that proposition because litigants and administrators alike keep forgetting it. A detention order does not record a conviction, it does not depend on a charge sheet, it is not defeated by an acquittal, and it does not become unlawful merely because the same facts could have grounded a prosecution that the state chose not to launch. Its legal character is preventive custody ordered by the executive and supervised through a special constitutional procedure. That character is why the familiar vocabulary of criminal process misleads at every step: there is no bail in preventive detention, because bail answers a charge and there is none; there is no trial to speed up or delay; and the remedy for an unlawful order is not an appeal on the merits but a constitutional challenge to the order’s validity, tested against the statute that authorised it and the procedure Article 22 prescribes.

The scale of the field also flows from the problem’s definition. Because the predicted harms are stated at the level of defence, foreign affairs, the security of India, the security of a state, public order, and essential supplies and services, the power is distributed across the federal system rather than concentrated in one legislature. Parliament legislates for the first group of subjects and state legislatures share the second, and both levels have used the competence freely. The result is not one detention law but a layered code, four standing central statutes and a large and changing family of state Acts, all of them drawing their constitutional authority from the same few clauses and all of them subject to the same safeguards and the same judicial technique. A guide to one statute therefore leaves the reader unable to answer the questions that actually arise, which are comparative and structural: which legislature could make this law, which authority could pass this order, which board reviews it, and which safeguard failed when a court set it aside.

The Law Before the Constitution Took Over

Preventive detention in India is older than the Republic by more than a century, and the colonial pedigree matters because the framers were not inventing a power in the abstract; they were deciding whether to retain one they knew intimately, many of them from the receiving end. The Bengal State Prisoners Regulation of 1818 empowered the colonial executive to hold state prisoners without trial on grounds of state security, and it remained part of the legal landscape deep into the constitutional era. The Defence of India legislation of the two world wars generalised executive detention for the duration of hostilities, and the Anarchical and Revolutionary Crimes Act of 1919, remembered as the Rowlatt Act, attempted to carry detention and curtailed trial into ordinary peacetime governance, provoking the first great all-India agitation against the principle itself. By 1947 the propositions on each side were thoroughly rehearsed. The colonial state had treated custody without trial as a routine instrument against sedition, communal violence, and organised crime alike, and the nationalist movement had made its abolition a standing article of faith.

The Constituent Assembly’s decision, taken against that background, was to constitutionalise the power rather than abolish it, and the record of the Assembly shows a decision taken with visible reluctance rather than enthusiasm. The circumstances pressed hard. Partition had produced mass violence, refugee movement on an enormous scale, a war in Kashmir, and a communist insurgency in Telangana, and the interim government was detaining people under inherited wartime and provincial powers whose legal basis a new fundamental rights chapter would immediately destabilise. Members who spoke to the detention clauses divided between those who regarded the power as a temporary necessity that a responsible government could not surrender at a moment of fragility, and those who warned that a constitution which began by authorising custody without trial had compromised its first principles at birth. Granville Austin, the historian whose account of the Assembly remains the standard scholarly record, treated the detention provisions as among the severest features the framers accepted, a point on which the drafting leadership’s own discomfort is documented rather than inferred. The settlement the Assembly reached is legible in the text it produced: the power was retained, its subjects were enumerated, and a floor of procedure was written into the Constitution so that no legislature could go below it, however permissive its statute might be.

The significance of constitutionalising the power, rather than merely permitting it by silence, is easy to miss and impossible to overstate. Had the Constitution said nothing about preventive detention, the power would have been tested entirely against Article 21’s guarantee that no person shall be deprived of life or personal liberty except according to procedure established by law, and against the arrest safeguards of Article 22, clauses (1) and (2), which speak of persons who are arrested. By writing detention into Article 22 itself, the framers did two things at once. They authorised the category beyond argument, since a constitution does not ordinarily provide safeguards for a power it prohibits, and they insulated the category from the ordinary arrest clauses by expressly disapplying them. Every abolitionist argument since has had to confront that textual fact, which is why the serious case against preventive detention in India is a case for constitutional amendment and not a case about what the law already means.

Article 22, Clause by Clause

Article 22 carries the heading of protection against arrest and detention in certain cases, and its structure is a study in division. Clauses (1) and (2) state the protections of the ordinarily arrested person. A person who is arrested shall not be detained in custody without being informed, as soon as may be, of the grounds of the arrest, and shall have the right to consult and to be defended by a legal practitioner of choice. Every arrested person shall be produced before a magistrate within twenty-four hours of arrest, excluding the time of the journey, and shall not be detained beyond that period without a magistrate’s authority. These clauses constitutionalise the first hours of criminal custody: information, counsel, and a judicial officer between the police and continued detention.

Clause (3) then divides the field. It provides that clauses (1) and (2) shall not apply to an enemy alien, and shall not apply to a person who is arrested or detained under a law providing for preventive detention. The drafting is worth reading slowly, because its effect is routinely misstated in both directions. Clause (3) does not repeal the protections of the detenu; it disapplies two specific clauses and leaves the detenu to the protections of clauses (4) to (7), which are a different and deliberately constructed set. Nor does clause (3) touch Article 21, which protects every person against deprivation of life and personal liberty except by procedure established by law, or Article 19, whose freedoms the Supreme Court would later hold to bear on detention as well. The twin exclusion is therefore precise: what the detenu loses at the threshold is the arrested person’s information right in its clause (1) form, the counsel right in its clause (1) form, and the twenty-four hour magistrate production of clause (2). What replaces them is the subject of the next clauses.

Clause (4) states the three-month rule and the Advisory Board gate. No law providing for preventive detention shall authorise the detention of a person for a longer period than three months unless an Advisory Board, constituted in accordance with the clause’s own composition rule and consisting of persons who are, or have been, or are qualified to be appointed as Judges of a High Court, has reported before the expiration of the three months that there is in its opinion sufficient cause for the detention. The same clause recognises a second route: detention beyond three months is also authorised where the person is detained in accordance with a law made by Parliament that prescribes a maximum period and the classes of case to which the clause (7) exceptions apply. The composition rule deserves its weight. The reviewing body is not a departmental committee; its members must carry, or have carried, or be qualified to carry, High Court judicial office, and the design intention was to place a body of judicial temper between the executive’s satisfaction and any custody longer than a single season.

Clause (5) builds the detenu’s own remedy inside the executive process. The authority making a detention order shall, as soon as may be, communicate to the detenu the grounds on which the order has been made, and shall afford the detenu the earliest opportunity of making a representation against the order. Two disciplines follow. The grounds communicated must be the real grounds, sufficient to make an effective representation possible, and the representation must actually be considered by the appropriate government, which holds the power to revoke its own order. Generations of litigation have turned on this clause more than on any other, because it converts every step, service of grounds, supply of the material relied on, transmission of the representation, and its consideration without unexplained delay, into a condition of the order’s validity.

Clause (6) qualifies clause (5) at its most sensitive point. The authority is not required to disclose facts which it considers to be against the public interest to disclose. The detenu is therefore entitled to the grounds but not necessarily to the whole of the material behind them, and the line between a ground, which must be communicated, and a fact or source, which may be withheld, has been contested throughout the life of these statutes. The structural consequence is plain and is best stated without varnish: the representation that clause (5) guarantees is made by a person who may not know the full case to be answered, and the Advisory Board, which does see the material, sits in a procedure that is not adversarial in the manner of a court. Clause (7) completes the architecture by empowering Parliament to prescribe the classes of case and circumstances in which a person may be detained for longer than three months without the board’s opinion, the maximum period of detention for any class of case, and the procedure Advisory Boards are to follow. The Constitution thus fixes a floor, allocates the rule-making for what stands above the floor, and leaves the operating detail to statute.

The Twin Exclusion and What It Removes

The comparison between preventive detention and ordinary arrest is owned by a companion article in this cluster, and the pillar’s task here is narrower: to fix what the exclusion in clause (3) does to the detenu’s position in the first hours and days of custody, since everything else in the system is compensation for exactly that loss. A person arrested on a charge passes almost immediately into judicial custody discipline. The grounds must be told, a lawyer may be instructed, a magistrate must be faced within a day, and from that hearing onward the criminal court’s calendar, remand limits, and bail jurisdiction structure the detention. A person detained under a preventive detention order passes into none of that. The order is served, the grounds follow under clause (5), and the detenu’s world becomes the representation, the Advisory Board reference, and the writ court. The full working comparison of the two regimes, including bail, the role of a first information report, and the relative ease of challenge, is set out in the comparison of preventive detention with ordinary arrest, and it is that article rather than this one that owns the reader’s which-is-which questions.

What clause (3) removes should be stated with care, because two popular errors grow from careless statements of it. The first error holds that a detenu has no right to a lawyer at all. The accurate position is layered. The constitutional right to consult and be defended by a legal practitioner of choice, as a clause (1) entitlement, does not apply, and a detenu has no right, as a matter of course, to appear through counsel before the Advisory Board, since the board’s proceedings are not a trial and the detaining authority is not represented by counsel either in the ordinary scheme. At the same time, courts have recognised narrower entitlements at the margins, including the assistance of a friend who is not a legal practitioner in board proceedings where fairness requires it, and nothing in clause (3) prevents a detenu from instructing counsel for the purpose that matters most in practice, the drafting of the representation and the preparation of a habeas corpus petition. The second error holds that the detenu is outside the Constitution’s protection altogether. That is false, and the whole case law stands against it. Articles 21 and 32 and 226 remain available, the statute under which the order was passed may be challenged, and the order itself is examinable for compliance with the statute and with clauses (4) to (7) at every point.

The deeper significance of the twin exclusion is institutional. By removing the magistrate from the detenu’s path, clause (3) removes the one actor in the criminal process whose function is to be unimpressed by the executive’s account at first sight. The Advisory Board was designed as the substitute, and the comparison between the two substitutes is where honest assessment has to begin. A magistrate at a production hearing acts in the open, at a fixed early moment, in a proceeding the arrested person attends with counsel. An Advisory Board acts on a reference made by the government, on material that includes what clause (6) withheld from the detenu, in a proceeding that produces a report to the government rather than a judgment to the world. The framers knew the difference and chose the board anyway, on the stated logic that security material could not be ventilated in open court at the threshold of custody. Whether that logic justifies the design remains the live argument of the field, and it is an argument about institutional design, not about whether any particular order was justified.

The Three-Month Clock, the Advisory Board, and the Amendment That Never Commenced

The three-month rule is the most frequently misstated proposition in this field, and the correction must be exact. Three months is not the maximum period of preventive detention in India. It is the threshold at which the Advisory Board gate operates. Within the first three months, a detention order stands on the detaining authority’s satisfaction, the government’s confirmation procedures under the applicable statute, and the detenu’s representation. Beyond three months, the order can support custody only where the Advisory Board has reported sufficient cause before the period expired, or where a parliamentary law within clause (7) provides otherwise for the class of case. Statutes then fix their own outer ceilings above the gate: the National Security Act, for example, fixes a maximum period of twelve months for a detention under it. A statement that a detenu can be held for only three months is therefore wrong, and a statement that the three months is a mere formality is equally wrong, because a reference not made, a board not properly constituted, or a report that arrives after the period has run will each destroy the order in court.

The uncommenced amendment is the sharpest illustration in Indian constitutional law of the distance between enactment and operation. The Constitution (Forty-fourth Amendment) Act, enacted in the post-Emergency settlement and assented to on 30 April 1979, contained in its Section 3 a substituted clause (4) for Article 22. The substitution would have reduced the board threshold from three months to two, and would have reconstituted Advisory Boards on the recommendation of the Chief Justice of the appropriate High Court, with a serving High Court judge as chairman and serving or retired High Court judges as the other members. Most of the Forty-fourth Amendment was brought into force in 1979. Section 3 was not, and it has never been notified. The operative text of Article 22 therefore remains the original text: three months, and boards composed of persons who are, or have been, or are qualified to be appointed as High Court judges. The story of how the restoration amendment rebuilt the Emergency settlement, including the provisions that did commence and those that did not, belongs to the article on the 44th Amendment Act 1978, and the reader tracing the amendment itself should go there.

The legal consequence was tested at the first opportunity. In A. K. Roy v Union of India, decided on 28 December 1981, the Supreme Court upheld the National Security Act against a challenge that included the executive’s failure to commence Section 3. The majority declined to compel the Central Government to bring the section into force, reasoning that the commencement power conferred by the Amendment Act’s own first section was a discretion the court could not convert into a duty on the material before it. Justice A. C. Gupta, dissenting on this point, would have held the failure to commence to be indefensible after the passage of years, and his dissent remains the strongest judicial statement of the opposing view. The episode fixes a doctrine with reach far beyond detention: a constitutional amendment may be perfectly enacted, assented to, and printed in every edition of the Constitution with an editorial footnote, and yet change nothing at all in the law a detenu faces, because commencement was left to notification and notification never came. Students of this field should hold the two propositions together without blurring them. Parliament voted for two months. The law in force says three.

Why does Article 22 still speak of three months when Parliament voted for two?

Because the vote and the commencement are separate legal events. Section 3 of the Forty-fourth Amendment Act, which substitutes two months for three in clause (4) and reconstitutes Advisory Boards, was left to be brought into force by notification, and the notification has never issued, so the original clause (4) remains the operative text.

The practical moral is about where a reader should look for the law. The text of an amendment Act tells the reader what Parliament enacted. The commencement notification tells the reader what of that enactment operates. In the preventive detention field the two diverge on the single most citizen-facing number in the Article, which is why careful accounts of Article 22 state the three-month threshold as the law and the two-month substitution as an enacted and uncommenced reform, in that order, and why any account that announces a two-month rule as the governing position has confused the statute book with the law in force.

The Preventive Detention Architecture Map

The table below is the findable artifact of this article, and it is built to be used rather than admired. Read any detention order down its columns: identify the layer from which the statute draws its authority, the legislature that could enact it, the authority that could pass the order, and the body that supervises custody beyond the threshold. An order that cannot be placed in the table has not been understood, and an order placed in the wrong row will be defended or challenged on the wrong grounds.

Layer	Constitutional or legislative basis	Enacting authority	Statutes at this layer	Detaining authority	Supervising body
Constitutional authorisation	Article 22, clauses (3) to (7)	The Constituent Assembly, as the constitutional text	None; the clauses govern all detention laws	None; the clauses confer no power to detain	Constitutional courts under Articles 32 and 226
Union competence	Entry 9 of the Union List, Seventh Schedule	Parliament alone	Defence, foreign affairs and security of India detention legislation	Central Government and its empowered officers	Advisory Boards and the constitutional courts
Concurrent competence	Entry 3 of the Concurrent List, Seventh Schedule	Parliament and state legislatures	Public order, state security and essential supplies detention legislation	Central Government, state governments and their empowered officers	Advisory Boards and the constitutional courts
Central statutes	The two entries above, as applicable	Parliament	National Security Act 1980; Conservation of Foreign Exchange and Prevention of Smuggling Activities Act 1974; Prevention of Illicit Traffic in Narcotic Drugs and Psychotropic Substances Act 1988; Prevention of Blackmarketing and Maintenance of Supplies of Essential Commodities Act 1980	The governments and specially empowered officers named in each Act	Advisory Boards constituted under each Act
State statutes	Entry 3 of the Concurrent List	State legislatures	State preventive detention Acts, including the Tamil Nadu Act of 1982 directed at specified categories of offenders and the Gujarat Prevention of Anti-Social Activities Act 1985	State governments and specially empowered officers, including district level officers where the Act so provides	Advisory Boards constituted under each Act
Supervision layer	Article 22, clauses (4) and (5), and Articles 21, 32 and 226	The constitutional text and the applicable Act	The representation, board reference and writ procedures themselves	The government considering the representation	Advisory Boards; the High Courts and the Supreme Court on habeas corpus

The map teaches the field’s central structural lesson at a glance. There is one constitutional root and two legislative branches, and every statute in the field hangs from one of the branches. Disputes that look miscellaneous, a board wrongly composed under a state Act, a reference delayed under a central Act, a ground that strays outside the entry that supports the statute, are all failures at an identifiable layer, and the layer determines the argument.

Who Makes These Laws: The Seventh Schedule Split

Legislative competence for preventive detention is allocated by two entries, and the allocation explains the shape of everything built beneath it. Entry 9 of the Union List in the Seventh Schedule gives Parliament exclusive power to legislate for preventive detention for reasons connected with defence, foreign affairs, or the security of India, together with the persons subjected to such detention. Entry 3 of the Concurrent List gives Parliament and the state legislatures shared power to legislate for preventive detention for reasons connected with the security of a state, the maintenance of public order, or the maintenance of supplies and services essential to the community, again together with the persons subjected to such detention. The State List contains no detention entry of its own. A state detention statute must therefore be traced to Entry 3 and to the subjects it names, and a state Act whose operative grounds wander away from state security, public order, and essential supplies invites a competence challenge at its root.

Where does the power to make preventive detention laws come from?

The power comes from two entries working together. Entry 9 of the Union List reserves detention connected with defence, foreign affairs, and the security of India to Parliament alone, while Entry 3 of the Concurrent List lets Parliament and the state legislatures both legislate on detention connected with state security, public order, and essential supplies and services.

The drafting of the two entries repays attention, because each ground in Entry 3 does distinct work. Security of a state addresses threats aimed at the state as a polity. Maintenance of public order addresses the community-facing ground on which the great majority of state detention orders rest, and it is the ground whose boundaries the courts police most strictly, since ordinary breaches of the peace belong to the criminal law and not to detention. Maintenance of supplies and services essential to the community carries the economic ground: hoarding, blackmarketing, and disruption of the systems a community cannot do without. The entries also explain an otherwise puzzling feature of the statute book. Parliament legislated on smuggling and foreign exchange conservation, subjects that might look like state concerns, because the Conservation of Foreign Exchange and Prevention of Smuggling Activities Act rests on the Union side of the allocation, smuggling being treated as a matter engaging the security and economic defences of India as a whole, while a state Act aimed at local hoarding rests on the essential supplies limb of Entry 3. Where both a central and a state law cover neighbouring ground, the ordinary doctrines of the Concurrent List, including the repugnancy rule of Article 254, govern their coexistence, and in practice the two families are administered in parallel by different governments for different scales of the same problems.

The competence split also allocates political responsibility, which matters to how the field is argued about. A detention passed under a state public order Act is the act of a state government answerable to a state legislature and a state electorate, and criticism of the pattern of such orders belongs in state politics. A detention under the National Security Act is the act of whichever government, central or state, passed the order under that central statute, since the National Security Act is administered by both. Confusing the layers produces confused accountability: the Union is blamed for state orders it did not pass, and state governments shelter behind a central statute they themselves invoked. The architecture map is the corrective. Every order has an author, and the author is identifiable from the statute, the entry, and the authority named in the order itself.

The Central Statutes

Four central statutes compose the standing Union code of preventive detention, and each is aimed at a distinct theatre of anticipated harm. The National Security Act, 1980 is the general security statute, and it is the operative central law against which the others are specialists. It authorises detention to prevent a person from acting in a manner prejudicial to the defence of India, the relations of India with foreign powers, or the security of India, and, through its state-facing limbs, to prevent acts prejudicial to the security of a state or the maintenance of public order or of essential supplies and services. It is administered by the Central Government and by state governments and their specially empowered officers, it works through the Advisory Board machinery Article 22 requires, and it fixes a maximum period of twelve months for detention under it. The complete working treatment of that Act, its grounds, its officers, and its litigation record, is given in the guide to the National Security Act 1980, and this pillar confines itself to the Act’s place in the system: it is the statute a reader should reach for when the anticipated harm is stated in the language of security and public order at national or state scale.

The Conservation of Foreign Exchange and Prevention of Smuggling Activities Act, 1974, universally shortened to COFEPOSA, aims detention at the economics of smuggling and foreign exchange offending. Its premise is that the organisers of smuggling networks and the financiers of foreign exchange rackets operate at a distance from the physical acts the criminal law can prove, through carriers, shell accounts, and layered transactions, so that prosecution habitually reaches the instruments and misses the principals. Detention under COFEPOSA is ordered by officers specially empowered for the purpose at the Centre and in the states, and the statute’s life in the courts has been dominated by the representation discipline of Article 22, clause (5): delay in considering a detenu’s representation, and delay in supplying the material relied on, have destroyed COFEPOSA orders with a regularity that makes the Act the standard classroom illustration of procedural strictness. The Prevention of Illicit Traffic in Narcotic Drugs and Psychotropic Substances Act, 1988 applies the same logic to the narcotics trade, authorising detention of persons engaged in, or positioned to engage in, illicit traffic, a field in which the profits are large, the networks are transnational, and the persons at the top are insulated by design. The Prevention of Blackmarketing and Maintenance of Supplies of Essential Commodities Act, 1980 completes the central family on the supplies side, authorising detention to prevent blackmarketing and to secure the maintenance of supplies and services essential to the community, the Entry 3 ground translated into a central statute.

Read together, the four statutes show the central code’s internal logic. One general security Act, and three specialist Acts for the three theatres in which the gap between suspicion and proof was judged widest and most costly: contraband and currency, narcotics, and the supply lines of daily life. Each statute fixes its own grounds, its own empowered officers, its own confirmation and reference procedure, and its own ceiling on the period of detention, one year in the ordinary case under the specialist statutes and twelve months under the National Security Act. The differences matter in litigation, where an order is tested against the statute that made it and not against the field in general, but the constitutional procedure that disciplines all four is the same, and it is supplied by Article 22 rather than by any of them.

The State Family of Acts

Below the central code sits the state family, larger, more varied, and responsible for the great bulk of detention orders passed in the country in ordinary times. State legislatures legislating under Entry 3 have enacted preventive detention Acts aimed at the categories of offending their governments found hardest to reach through prosecution: the Tamil Nadu Act of 1982, directed at persons its scheme describes through categories such as bootleggers, drug offenders, forest offenders, goondas, immoral traffic offenders, land grabbers, and video pirates as its amendments successively widened it; the Gujarat Prevention of Anti-Social Activities Act, 1985, aimed at dangerous persons, drug offenders, and property grabbers among its named categories; and the Jammu and Kashmir Public Safety Act, 1978, a security-facing statute of a different character, long used in the circumstances of that region. Other states maintain Acts of the same family or legislate episodically, and states also pass large numbers of orders under the National Security Act itself, which makes the state layer the system’s centre of gravity in practice whatever the constitutional diagram suggests.

Two features of the state family deserve the reader’s permanent attention. The first is the breadth of the labels. A statute that authorises detention of a person characterised as a goonda or a dangerous person builds its operative test out of a reputation and a record of alleged acts, most of them never tried, and the courts’ insistence that the alleged acts must disturb public order rather than merely break the law is the principal doctrinal barrier between a wide state Act and a general executive power to imprison troublesome people without charge. The second feature is repetition. State detention law is where successive orders against the same person, expiry and fresh orders, and the recycling of stale incidents are most often litigated, because the subjects of state orders are typically local figures known to the district administration, and the temptation the doctrine guards against is the use of detention as a standing supplement to a weak prosecution case. The pillar’s point is structural rather than statistical: the constitutional safeguards were designed at the Centre, in a text, but they are tested overwhelmingly in the states, in district level orders, and the distance between the design and the test explains much of the field’s case law.

The Replacement Pattern: 1950, 1971, 1980

The central history of the field is a pattern of replacement, and the pattern is more instructive than any of its episodes taken separately. The Republic opened with the Preventive Detention Act, 1950, enacted within weeks of the Constitution’s commencement by the Provisional Parliament, piloted by Sardar Patel as Home Minister, defended as a temporary measure for a state facing insurgency and communal violence, and then extended again and again by amending Acts until it was finally allowed to lapse at the end of 1969. The passage story of that statute, its committee stage, its debates, and the Supreme Court’s first encounter with it, is told in full in the legislative history of the Preventive Detention Act 1950. What belongs to this pillar is the shape of the sequence. The 1950 Act lapsed, and within two years the Maintenance of Internal Security Act, 1971 had taken its place as the standing central detention statute, broader in its ambitions and destined to become the detention instrument of the Emergency, before it too disappeared in the repeal that followed the Janata victory. The National Security Act, 1980, arriving first as an ordinance in September 1980 and then as an Act, restored the standing power in a form consciously modelled on its predecessors, and it has remained the general central statute since.

Three readings of the pattern compete, and the stance discipline of this series requires each at its strongest. The first reading, favoured by civil liberties critics, treats replacement as proof of appetite: each lapse or repeal is a democratic verdict against the power, and each successor statute is the executive recovering by fresh drafting what politics had taken away, so the unbroken line from 1950 to 1980 shows that no Indian government, whatever its founding commitments, has been willing to govern without the instrument. The second reading, favoured by the statutes’ defenders across parties, treats replacement as proof of need: the gaps between statutes were brief and troubled, the anticipated harms did not pause for the statute book, and each successor was drafted narrower or with better procedure than its predecessor in response to the courts and to experience, so the line shows a democracy repeatedly reconsidering the power and repeatedly concluding, through elected legislatures, that a regulated version of it is necessary. The third reading is institutional and colder than either. It observes that lapse and repeal have never been accompanied by any change to Article 22 itself, so the constitutional authorisation has stood untouched beneath the statutory churn, and replacement is simply what a permanent constitutional power looks like at the statutory level: statutes are expendable precisely because the power that sustains them is not.

The verdict this pillar defends is that the third reading explains the first two. The appetite and the need are both real, and both operate downstream of a constitutional settlement that has never been reopened. Governments of every complexion have used the power because the settlement makes it available in ordinary legislation, and repeals have remained statutory because no repeal coalition has ever carried its argument to the constitutional level where the power actually resides. A reader who understands the replacement pattern therefore asks the right question of any proposed reform: does it amend a statute, which the next legislature can replace, or does it amend Article 22, which is where the field is finally decided?

The Courts and the Supervision Layer

Judicial supervision of preventive detention has a doctrinal spine that runs through three decisions, and the pillar must state the spine accurately because the specialist articles build on it. A. K. Gopalan v State of Madras, decided in 1950 on a challenge to the first detention statute, upheld preventive detention as a constitutionally contemplated category and read Articles 19, 21, and 22 as separate codes, so that a detention law satisfying Article 22 was not to be tested against the freedoms of Article 19, and Article 21 was satisfied by any procedure the law established. The Court did strike down Section 14 of the 1950 Act, which had barred disclosure of grounds, as inconsistent with clause (5), a detail that matters because it shows the safeguards operating with teeth from the very first case. The silo reading then held the field for two decades. R. C. Cooper v Union of India in 1970 broke the silos in the bank nationalisation context, holding that the fundamental rights operate in combination and that a law depriving a person of property or liberty must answer the relevant articles together. Maneka Gandhi v Union of India in 1978 completed the movement for personal liberty: the procedure contemplated by Article 21 must be just, fair, and reasonable, not arbitrary or oppressive, and a preventive detention law and the orders under it must survive that standard in addition to the letter of Article 22.

What changed between A. K. Gopalan in 1950 and Maneka Gandhi in 1978?

The unit of analysis changed. Gopalan treated Articles 19, 21, and 22 as self-contained codes and asked only whether the detention law matched Article 22. Cooper broke the separation, and Maneka Gandhi then required the procedure under Article 21 to be just, fair, and reasonable, so a detention law must satisfy fairness as well as the text of Article 22.

The mature doctrine that emerged governs the daily work of the writ courts, and its content is best described as exacting proceduralism. The Supreme Court and the High Courts do not, in the ordinary run of cases, substitute their own assessment of the threat for the detaining authority’s satisfaction, which the Constitution commits to the executive subject to the board’s review. What the courts police, with a strictness that surprises readers formed on other fields of administrative law, is the procedure: whether the grounds were communicated in time and in a form the detenu could answer, whether the material relied on was supplied where the law required it, whether the representation was considered by the proper authority without unexplained delay, whether the board reference was made and the board properly constituted, whether the incidents relied on bear a live and proximate link to the order rather than being stale or trivial, and whether the ground invoked is one the statute and the supporting entry actually cover. Detention orders fall in large numbers on these points. The pattern should be described as a pattern and not dressed in an invented percentage: across decades and across statutes, a high proportion of the orders carried to the constitutional courts have been set aside on procedural grounds rather than on any judicial disagreement with the executive’s assessment of danger, and the explanation offered by the courts themselves is that a power exercised without trial must be exercised exactly according to the procedure that substitutes for trial, or not at all.

How do the three concentric circles help a court test a detention order?

The Lohia jurisprudence pictures law and order as the widest circle, public order within it, and the security of the State as the smallest. An act disturbing only individuals sits in the outer circle and belongs to criminal law. Detention requires conduct reaching the middle circle, affecting the community and the even tempo of its life.

The circles doctrine, drawn from the Supreme Court’s Lohia decisions on the meaning of public order, does concrete work in the state Act cases. An alleged act, however disreputable, that affects only its immediate victim and leaves the community’s even tempo undisturbed is a law and order incident, and a detention order built on it fails at the threshold, because Entry 3 and the statutes under it speak of public order and not of crime in general. The same doctrine disciplines the drafting of grounds: the authority must show the community-facing reach of the conduct, not merely its illegality. Alongside the circles, the courts apply the live link requirement, asking whether the incidents relied on are proximate enough in time and character to support a prognosis about the detenu’s future conduct, and the single-act sufficiency rule, accepting that one grave act may ground an order where its community impact is plain. A. K. Roy, discussed above for the uncommenced amendment, is also the authority that consolidated the post-Maneka position on the National Security Act itself, upholding the statute while confirming that the safeguards of Article 22 and the fairness standard of Article 21 govern every order under it. The supervision layer, taken as a whole, is therefore neither a rubber stamp nor a shadow trial. It is a procedural audit with constitutional consequences, and both its defenders and its critics err when they describe it as anything simpler.

The Complaint Route Outside the Courts

Not every grievance about a detention travels to a writ court, and the architecture includes a statutory institution whose relevance is practical rather than doctrinal. The Protection of Human Rights Act, 1993 created the National Human Rights Commission and the state commissions, with functions that include inquiring into complaints of human rights violation, intervening in court proceedings involving such allegations with the court’s approval, and visiting jails and other institutions where persons are detained or lodged, to study the living conditions of the inmates and make recommendations. For a detenu’s family confronting conditions of custody, denial of information, or ill treatment that falls short of a ground for quashing the order itself, the commissions supply a forum whose inquiry does not depend on defeating the order’s validity. The institution, its powers, and its record are treated in their own right in the article on the Protection of Human Rights Act 1993, and the pillar notes the route here for a structural reason: the system’s designers placed a human rights body alongside the boards and the courts, and a complete map of remedies that omits it will mislead the families who need a remedy for the conditions of a detention whose grounds they cannot defeat. The general account in this article is legal information about the structure of the law, and a reader facing an actual detention, with its unforgiving timelines for representation and reference, needs advice on the individual case from a practitioner who can see the order and its papers.

The Case For and Against the Institution Itself

The stance charter of this series requires the contested question at the centre of the field to be argued at full strength on both sides before any verdict, and the contested question here is not the validity of any order but the legitimacy of the institution. The abolitionist case, in its strongest form, runs as follows. Preventive detention is punishment without the protections that make punishment legitimate: no charge, no trial, no proof, no counsel by right at the decisive hearing, and a reviewing body that sits in private on material the detenu cannot fully see. It is directed overwhelmingly at the margins of the state’s coercive attention, at smugglers and local strongmen and political dissidents rather than at the powerful, and its procedural safeguards, being administered by the same executive culture that passes the orders, fail in the steady pattern the courts record. A constitutional democracy that prides itself on Article 21 should not maintain, in peacetime and as a standing system, a power that the colonial state used against the freedom movement and that most comparable democracies confine to war. Since the power is constitutional, abolition requires amendment, and abolitionists accept that burden and argue for deleting or drastically narrowing the detention clauses of Article 22.

The retentionist case, also in its strongest form, answers point for point. The anticipated harms are real, and the criminal law’s retrospective machinery demonstrably fails against organisers who insulate themselves from provable acts; the choice is not between detention and liberty in the abstract but between a regulated power and the unregulated alternatives that states reach for when regulated ones are denied, including serial prosecution on stale charges and informal custody that no board or court ever sees. The Indian design, on this view, is not careless but considered: the power is textually authorised, the grounds are enumerated in the Seventh Schedule rather than left open, the safeguards are constitutional rather than statutory and so cannot be diluted by an ordinary legislature, and the courts have built on them a review practice that quashes orders in substantial numbers and has repeatedly narrowed the power’s reach, from the fairness revolution of Maneka Gandhi to the public order discipline of the circles doctrine. A power that is used, reviewed, and frequently defeated in court is, retentionists conclude, a working constitutional compromise rather than a scandal.

Between those cases lie the reform positions that dominate serious professional discussion, and they should be named rather than blended. The commencement reform would simply notify Section 3 of the Forty-fourth Amendment, halving the board threshold to two months and judicialising the boards, a reform available to the executive by a single notification and blocked, as A. K. Roy established, by executive choice rather than by any legal obstacle. The narrowing reform would confine state detention Acts to genuinely community-facing conduct, writing the circles doctrine into the statutes and stripping out the reputation-based categories that invite routine use against ordinary offenders. The sunset reform would subject every detention statute to periodic parliamentary renewal, converting the accidental pattern of lapse and replacement into a deliberate discipline. Each reform concedes the institution and attacks a different layer of it, commencement, scope, or duration, and the three are compatible in a way the two grand cases are not.

The verdict of this pillar follows the architecture it has described. The abolitionist case is right about the text: the twin exclusion is a genuine withdrawal of the ordinary protections, and describing preventive detention as ordinary law enforcement is a category error whatever one’s politics. The retentionist case is right about the alternative: the record of the gaps between statutes does not support the claim that the underlying executive behaviour disappears when the regulated power does, and the procedural case law is a real discipline that informal custody wholly escapes. The defensible position, on the evidence of the system’s own working, is retention under the discipline the Constitution already prescribes, enforced without relaxation, together with the commencement of the reform Parliament enacted in 1979 and no government has notified. A settlement in which the executive keeps the power and also keeps the stricter safeguard uncommenced is the one position the record does not defend, because it takes the benefit of both sides of the framers’ and the amenders’ bargains and honours neither.

Reading Any Detention Law: A Worked Method

The pillar closes its analytical work with the method its map makes possible, stated as a sequence of questions a reader can apply to any detention statute or order encountered anywhere in the system. Begin with competence: which entry supports the statute, Entry 9 for the Union’s security subjects or Entry 3 for the shared subjects, and are the grounds in the order the grounds the entry names. Move to the statute: which authority may order, on what satisfaction, with what confirmation by the government, and within what ceiling on duration. Move to the Constitution: were the grounds communicated as clause (5) requires, was the representation invited and considered, was the board reference made before three months ran, and was the board composed as clause (4) requires. End with the courts’ doctrines: does the conduct alleged reach public order on the circles test, is the material proximate enough to support a prognosis, and has the procedure, at any step, departed from the statute’s own terms. An order that survives all four stages is, within the system’s own logic, valid; the stages also locate, for the challenger, exactly where a defective order will break.

How does a detention order travel from the detaining authority to the Advisory Board?

The authority passes the order on its recorded satisfaction and serves it with the grounds. The government considers the detenu’s representation and may revoke. Where custody is to continue, the case is referred to the Advisory Board within the statute’s timetable, the board reports on sufficient cause before three months expire, and the government then confirms or revokes.

Teachers and students who handle this material in slide form for classroom revision will find that a deck built around the architecture map holds the whole field in a handful of slides, and a received deck can be checked against the map slide by slide in the ReportMedic PPT viewer, which renders presentation files in the browser without uploading them. The habit worth building is comparative: for every statute the reader meets, complete the map’s row from the statute itself, and the differences between statutes, which are the examinable and litigable content of the field, present themselves without memorisation. The advocate uses the same table defensively, auditing a client’s order layer by layer. The researcher uses it historically, watching statutes appear, lapse, and reappear on the same two branches across three decades. The table is one object, and the three uses are the three reader payloads this pillar was written to serve.

The Grounds and the Satisfaction: Where Orders Are Won and Lost

Every detention order rests on a satisfaction, and the legal character of that satisfaction determines what a court may and may not do with the order. The satisfaction is the detaining authority’s own: that a person is acting, or is about to act, in a manner prejudicial to one of the statutory subjects, or that the person’s being at free will enable such conduct. It is subjective in the technical sense that the Constitution and the statutes commit the assessment to the authority rather than to a court, and a court will not weigh the sufficiency of the material as an appellate body would. It is not, for that reason, unexaminable. The courts test the satisfaction for its existence and its honesty: whether the authority applied its mind to material that exists, whether the material is relevant to the statutory ground invoked, whether the inference from past conduct to future conduct is one a reasonable authority could draw, and whether the order was passed for the statutory purpose rather than for a collateral one, such as keeping a person in custody because a prosecution has failed or bail has been granted in a criminal case. That last abuse, detention used to neutralise bail, is the one the courts name most sternly, because it converts a preventive power into an executive override of the judiciary’s own orders.

The grounds are the documentary heart of the order, and their drafting decides more litigation than any question of high doctrine. A ground must state the incident or material relied on with enough particularity for the detenu to answer it, and the case law distinguishes, with consequences, between the ground itself, which must be communicated, and the sources and intelligence behind it, which clause (6) may protect. Vagueness is fatal where it disables the representation: a ground that recites the statutory language without the conduct that is said to satisfy it gives the detenu nothing to answer, and courts have treated such grounds as no grounds at all. Plurality creates its own doctrine. Where an order rests on several grounds and one is irrelevant or non-existent, the older case law asked whether the authority would have passed the order on the surviving grounds alone, and orders fell where the bad ground could not be separated from the satisfaction. Parliament answered in some statutes with deeming provisions that treat an order made on several grounds as if it had been made separately on each, a legislative repair whose own validity the courts have examined statute by statute. The working rule for the reader is simpler than the doctrine’s history: read the grounds first, because the grounds are where the authority has committed itself, and a challenge that cannot find purchase in the grounds will rarely find it anywhere else.

Two further doctrines complete the satisfaction’s legal frame. The live link requirement, already noticed with the circles, insists that the incidents relied on be proximate to the order in time and character, so that the prognosis is about the person the detenu is rather than the person the detenu was; a long and unexplained gap between the last incident and the order breaks the link and the order with it. The doctrine of the single sufficient ground runs in the opposite direction and protects orders: one incident of sufficient gravity, bearing directly on the statutory subject and affecting the community in the manner the circles require, may sustain a satisfaction without a pattern. Between those two doctrines the courts conduct the real adjudication of this field, and both are applications of a single idea, that a prognosis must be reasoned from material that supports it, even though the reasoning is the executive’s and the standard is not proof.

The Representation: The Remedy That Decides Cases

Clause (5) gives the detenu two linked rights, communication of the grounds as soon as may be, and the earliest opportunity of making a representation against the order, and the second of them is in practice the most consequential procedural right in the field. The representation goes to the appropriate government, the authority in whom the statute vests the power to revoke, and it invites that government to do what no court will do in terms: to reconsider the merits of its own order on the detenu’s answer to the grounds. Its power as a remedy lies in its informality and its speed. It requires no court fee, no limitation period of the writ kind, and no leave, and a government that finds the answer persuasive may end the detention at any moment by revocation. Its weakness is the mirror image of its strength. The detenu writes without knowing the withheld material, writes usually from custody and often without skilled help, and writes to the very executive whose satisfaction is in issue.

The courts have therefore surrounded the representation with disciplines that convert executive lapses into constitutional defects, and three of them recur so steadily that they amount to a code. The grounds and the material the law requires to accompany them must reach the detenu promptly, because a representation invited against grounds not yet supplied is an empty invitation, and delay in supply that is unexplained or attributable to official indifference vitiates the continued detention. The representation, once made, must be considered by the appropriate government itself, expeditiously and on its own application of mind; a file that travels slowly between desks, a consideration postponed while the board process runs in parallel, or a rejection recorded without the competent authority’s engagement has each been held to break the clause (5) promise of the earliest opportunity, the promise being of an opportunity that is real rather than formal. And the detenu’s right is not exhausted by the first representation in the manner of a limitation rule: where fresh material or a change of circumstance arises, the courts have been receptive to the obligation to consider a further representation on its own footing, since the clause speaks of the earliest opportunity against an order that continues to operate day by day.

The representation must also be distinguished from the Advisory Board proceeding, with which it is habitually confused. The representation is the detenu’s address to the executive, made in writing, answered by revocation or by silence in the form of the detention continuing. The board proceeding is the state’s reference to an independent body, made by the government, at which the detenu may appear and be heard in person and to which the government places its material. The two run on separate tracks with separate consequences: a government may reject the representation and the board may still report against sufficient cause, in which event the detention must end; a board may report in favour and the government may still revoke on a later representation, as it may at any time. The detenu who treats the board hearing as the only remedy neglects the earlier and often more fertile one, and the adviser who treats the representation as a formality neglects the ground on which, in the reported cases, orders most often die.

Duration, Confirmation and Revocation: The Life of an Order

A detention order has a life cycle, and each stage of it is a point at which the detention can lawfully end or unlawfully continue. The order is made and served. The grounds are communicated. The government, where the order was made by an empowered officer rather than by the government itself, must approve it within the short period the statute fixes, failing which the order lapses by its own force; this approval stage, routine in appearance, is a genuine filter, and statutes build it in because the officer closest to the ground is also the officer most exposed to local pressure. The representation is made and considered. The reference goes to the Advisory Board where custody is to extend beyond the threshold, and the board reports. The government then confirms the detention for a period within the statute’s ceiling, or revokes it. At every moment of the cycle the government retains the power of revocation, exercisable on a representation, on the board’s report, on fresh assessment, or on its own motion, and a detention that the government no longer believes justified is not merely imprudent to continue but unlawful, since the satisfaction that supports it has ceased to exist.

Expiry and renewal generate the cycle’s hardest doctrine. A detention order spends itself at the end of its confirmed period, and the state that wishes to detain further must pass a fresh order. A fresh order may not simply repeat the grounds of an order that a court has quashed on the merits of those grounds, because the quashing is an adjudication that the material did not support detention, and recycling it would nullify the judgment; the courts state the rule in terms of the same grounds being unavailable once struck down, while fresh grounds arising from later conduct stand on a different footing. Where the earlier order fell on a procedural defect rather than on the insufficiency of the material, the state’s room is wider, and the case law works through the distinction case by case. The pattern to carry forward is the one the architecture predicts: because the power is prognosis, its lawful exercise is perishable, and the system’s disciplines of confirmation, ceiling, and fresh satisfaction exist to prevent a single executive decision from hardening into indefinite custody by inertia. The maximum periods the statutes fix, twelve months under the National Security Act and one year in the ordinary case under the specialist central Acts, are the outer walls of the cycle, and no confirmation or renewal within a single order may cross them.

Detention and the Emergency: What Changes and What Does Not

The relationship between preventive detention and the Emergency provisions is the point at which the system’s ordinary disciplines meet their stress case, and the pillar states the mechanics without trespassing on the habeas corpus article that owns the great case. In ordinary operation, the detenu’s remedies rest on Articles 21 and 22 and on the writ jurisdiction, and those remedies are themselves fundamental rights. A proclamation of Emergency under Article 352 engages Article 359, under which the President may suspend the right to move a court for the enforcement of specified fundamental rights for the period of the proclamation. The Forty-fourth Amendment rebuilt this machinery after the experience of 1975 to 1977: Articles 20 and 21 may not be suspended even during an Emergency, a protection written into Article 359 itself, while the enforcement of Article 22 remains among the rights whose court enforcement a proclamation may reach. The result is a deliberate asymmetry that the reader should understand exactly. The detention clauses’ safeguards are constitutional text in all seasons. Their enforcement by a court can be suspended in an Emergency for Article 22, though after the Forty-fourth Amendment the Article 21 guarantee, with the fairness content Maneka Gandhi gave it, cannot be placed beyond enforcement at all.

The practical consequence for the architecture is that the Advisory Board and representation machinery becomes, in an Emergency in which Article 22’s enforcement is suspended, the only functioning discipline on detention, which is a further reason the framers of the Forty-fourth Amendment sought to judicialise the boards in the very section that was never commenced. The Emergency is therefore not an exception outside this article’s subject; it is the condition the subject’s design repeatedly anticipates, from the Constituent Assembly’s insistence on writing the safeguards into the constitutional text rather than leaving them to statute, to the amenders’ attempt to strengthen the board at exactly the layer an Emergency would expose. A reader who has followed the architecture will see why the habeas corpus litigation of the Emergency years turned on the suspension question rather than on the detention statutes themselves: the statutes were working as designed, and what had been withdrawn was the court’s power to say so.

District-Level Administration and the Empowered Officer

The system’s most consequential design choice after the twin exclusion is the identity of the detaining authority, because preventive detention in India is administered far closer to the ground than its constitutional text suggests. The statutes characteristically vest the power in the Central Government and the state governments, and then extend it to officers specially empowered for the purpose, a class that in the state field means, in practice, District Magistrates and Commissioners of Police. The district officer is the authority who knows the local record, commands the police reporting that generates the grounds, and feels the political cost of a public order failure. The design therefore places the satisfaction in the hands of the officer with the best information and the strongest incentives, which is precisely why the statutes surround the officer’s order with the approval discipline described in the life cycle above: the order must be reported to the government forthwith, together with the grounds and material, and it lapses unless the government approves it within the short statutory window. The filter is meant to catch, at the seat of government and within days, the order that local pressure produced and reflection cannot defend.

The case law on empowered officers adds two refinements that recur across statutes. The satisfaction must be the officer’s own: an order passed on the dictation of a superior, or on a police proposal the officer adopted without applying an independent mind to the material, is not the satisfaction the statute requires, and courts examine the file for evidence of that independent engagement. And the officer’s power is bounded by the government’s continuing supremacy in the scheme: approval, confirmation, and revocation sit with the government, so the district order is best understood as an initiation of detention subject to executive confirmation and constitutional audit, rather than as a completed act of the district administration. For the reader auditing an order, the officer layer is where the file is richest: the proposal, the noting, the approval, and the intervals between them are all recorded, and the disciplines of the system are tested against that record.

One Discipline Across Two Entries

A final structural feature deserves its own statement, because it is the feature that makes a single guide to the whole field possible. Although legislative competence is split between the Union and the states, the constitutional discipline is not split at all. Article 22 binds Parliament and every state legislature identically, since no legislature may authorise detention below the floor the clauses set, and Articles 21 and 19, as the courts have read them since Cooper and Maneka Gandhi, bind every detaining authority in the country in the same terms. A state Act cannot lawfully offer its detenus a lesser procedure than a central Act offers, however local its subject, and a central Act enjoys no immunity from the procedural audit that the state Acts undergo. The federal split allocates the subjects of detention; the fundamental rights chapter unifies its safeguards. That is why the case law moves so freely across the statute book, a COFEPOSA representation case governing an order under a state goondas Act, an NSA board case governing a smuggling detention, and why the practitioner in this field reads the whole line of authority rather than a statute-specific slice. The architecture is federal in its powers and unitary in its discipline, and the combination is characteristically Indian: the same settlement that distributed the subjects also refused to distribute the safeguards.

The Errors This Field Attracts

Because the pillar’s function includes protecting the reader against the standard mistakes, they are collected here in prose, in the order in which they do damage. The gravest is the statement that a detenu may be held for a maximum of three months, which confuses the board threshold with a ceiling and understates the power the statutes confer, the true position being that three months is the point beyond which the board’s report or a clause (7) law is required, with statutory ceilings, twelve months under the National Security Act, standing above it. The second error states that Article 22 guarantees the detenu a lawyer, which inverts clause (3): the clause (1) counsel right is among the protections disapplied, and the detenu’s access to legal help runs through the representation, the limited assistance recognised in board proceedings, and the writ petition, not through a constitutional right to be defended by counsel at the board. The third error treats preventive detention as a branch of criminal law, importing bail, charge, and trial into a system that has none of them, and producing advice that fails at its first step. The fourth treats the Advisory Board as a court and its report as a judgment, when the board is a constitutional reviewing body whose favourable report is a condition of continued detention rather than an adjudication of guilt or innocence, and whose unfavourable report obliges release without deciding any charge. The fifth, and the subtlest, announces the two-month rule of the Forty-fourth Amendment as the law, when Section 3 of that Amendment has never been commenced and the three-month text governs. Each error is a failure to read the record rather than the summary, and the five together are a fair test of any account of this field the reader may meet elsewhere: an account that survives all five may be trusted on the rest.

Inside the Board Proceeding

The Advisory Board is the institution this field invented for itself, and a reader who pictures it as a small court will misunderstand everything it does. The board comes into existence for a case because the government refers the case to it under the applicable statute, placing before it the grounds of detention, the material, and the detenu’s representation where one has been made. The board’s question is the single constitutional question: whether there is sufficient cause for the detention of the person. It is not asked whether an offence was committed, whether the detenu is guilty of anything, or whether the board would itself have passed the order. It reviews the executive’s prognosis through the executive’s own material, enlarged by whatever the detenu places before it, and reports its opinion to the government within the period the statute and the three-month rule allow.

The detenu’s position inside the proceeding is defined by the twin exclusion working itself out. The detenu may appear and be heard in person where the detenu so desires, under the scheme the statutes adopt from the constitutional design, and may place a case against sufficient cause. The detenu has no right to be represented by a legal practitioner before the board, a position the Supreme Court examined and sustained in A. K. Roy on the reasoning that the proceeding is not a trial, that the detaining authority does not appear by counsel either, and that adversarial counsel on one side would unbalance a review designed to be swift and inquisitorial. The same decision recognised the narrower facility that practice had developed: the detenu may take the assistance of a friend, a person who is not a legal practitioner, to help present the case, a facility whose denial in a case where the board’s procedure called for it could bear on the proceeding’s fairness. The board, for its part, is not bound by the detenu’s presence or absence, proceeds on the material, and reports; where it reports that there is no sufficient cause, the government is bound to revoke the order and release the detenu, the one point in the system at which a non-judicial body’s opinion operates as a command. Where it reports sufficient cause, the government may confirm the detention, and the detenu’s remaining route is the writ court on the disciplines already described.

Two features of the board’s design explain most informed criticism of it, and both should be stated plainly. The board sees the material that clause (6) withheld from the detenu, so its review is better informed than the detenu’s representation can be, and simultaneously less tested, since the material is not exposed to adversarial challenge by anyone who knows where its weaknesses lie. And the board’s report is an opinion to the government rather than a judgment published to the world, so the system’s central safeguard produces no body of reasoned precedent by which its own consistency can be judged from outside. Defenders answer that both features are the price of the board’s speed and of the confidentiality the material requires, and that the writ courts supply the public, reasoned layer the board lacks. The structural fact, whichever side the reader takes, is that the framers chose a private review by persons of judicial standing over a public review by a court, and the whole subsequent argument about the institution is an argument about that choice.

Clause (7) and the Outer Limits Parliament Holds

Clause (7) is the least read part of Article 22 and the part that most qualifies the comfort the rest of the Article offers, because it reserves to Parliament three powers that stand above the board discipline. Parliament may prescribe the circumstances and classes of case in which a person may be detained for longer than three months without obtaining the board’s opinion at all. Parliament may prescribe the maximum period for which a person may be detained in any class of case. And Parliament may prescribe the procedure to be followed by an Advisory Board in its inquiry. The first of these is the deepest reserve in the Article: the board gate, which the pillar has presented as the system’s central safeguard, is itself subject to a parliamentary power to disapply it for classes of case Parliament defines. The power has been sparingly used, and no standing central statute builds its ordinary operation on dispensing with the board, which is why accounts of the field that omit clause (7) survive proofreading. But a constitutional guide must record where the floor ends. The safeguards of clauses (4) and (5) bind every legislature; the board requirement within them binds subject to Parliament’s clause (7) competence; and the difference between those two propositions is the difference between the Constitution’s promise and its fine print.

The second and third powers explain the variety the reader finds across the statute book. Maximum periods differ because clause (7) contemplates Parliament fixing them by class, and board procedures differ in detail because the same clause lets Parliament prescribe them. When a reader asks why the National Security Act’s timetable differs from a state Act’s, or why ceilings differ between statutes aimed at different harms, the answer is usually traceable to this clause operating through the statutes: the Constitution built a frame with adjustable members, and legislatures have adjusted them. The discipline that does not adjust is the floor of communication, representation, and, outside the clause (7) classes, the board.

Detention Beside the Criminal Process

Preventive detention does not replace the criminal process, and the two run beside each other in a relationship that generates some of the field’s most counterintuitive doctrine. The same conduct may ground a prosecution and a detention order, and neither proceeding bars the other, because their objects differ: the prosecution punishes a proved past act, the detention prevents an anticipated future one. A detenu may therefore be an undertrial in a pending case, a person discharged or acquitted in one, or a person against whom no case has ever been filed, and the detention order’s validity does not rise or fall with the criminal case’s fortunes as a matter of automatic law. Acquittal does not by itself destroy a satisfaction honestly recorded on material the criminal standard could not carry to proof, a proposition the courts state with care, because the standards and the questions are different. What the courts will not permit is the use of detention to defeat the criminal process’s own outcomes: an order passed to keep a person in custody because bail was granted, on the same material the bail court considered, without fresh grounds engaging the preventive purpose, is the standing example of a collateral purpose, and it fails. The line the doctrine walks is fine and factual: detention may accompany prosecution where the preventive case is genuine, and may not substitute for prosecution where the preventive case is a pretext for custody the criminal courts have refused.

For the family of a detenu the practical lesson of the parallel structure is that two files must be fought, not one. The criminal case, where one exists, proceeds in its courts on its timetable, and success in it, bail or acquittal, does not release the detenu from the detention order, which must be met by the representation, the board, and the writ. Conversely, the quashing of a detention order does not touch a pending prosecution. Advisers who collapse the two files into one misunderstanding cost their clients the remedy that was actually available, and the pillar records the separation here because it is the single most expensive confusion in the practice of the field.

Why the Safeguards Are Procedural

A reader who has followed the architecture may notice what the safeguards never do. They do not require the anticipated harm to be proved. They do not give the detenu a hearing before the order is passed. They do not place the merits of the prognosis before a court. Every safeguard in Article 22 operates on the process: tell the detenu the grounds, hear the representation, interpose the board at three months, and let the courts audit each step. The design choice was deliberate and its logic is worth stating at its strongest, because it is the logic the retentionist case ultimately rests on. A prognosis about future conduct, formed on intelligence and pattern, cannot be subjected to ex ante judicial proof without ceasing to be the instrument it is; a state that must prove the future act before preventing it has been given the criminal law and nothing else. What can be subjected to discipline without destroying the instrument is the honesty and regularity of the process by which the prognosis is formed, communicated, reconsidered, and independently reviewed. Procedure is therefore not the safeguards’ weakness chosen by timid framers; it is the only dimension along which a preventive power can be disciplined at all, and the case law’s ferocious attention to intervals, files, and forms is the sound of that discipline working at the only points where it can bite.

The abolitionist reply is equally structural and should be left standing beside the logic rather than resolved into it. Procedure disciplines the formation of a prognosis; it does not test whether the prognosis was true, and a perfectly regular order may rest on material that is thin, partisan, or wrong in ways no procedural audit will ever expose, because the audit’s questions are about communication and timing and constitution of boards, and not about whether the man detained was in fact about to do anything at all. The system’s honest description is therefore double. It is a genuine discipline, which defeats real orders in real numbers and has improved the executive’s practice under its pressure. And it is a discipline with a designed blind spot at its centre, where the merits live. Every serious position in the field, abolition, retention, and each of the reforms, is at bottom a judgment about how much that blind spot matters, and the pillar’s architecture gives the reader the means to locate the judgment rather than a substitute for making it.

The Vocabulary of the Field

The field’s terms of art carry its doctrine in compressed form, and the pillar closes with them because a reader equipped with the vocabulary can parse any order, judgment, or statute in the area without a guide. The detenu is the person detained, the word itself a borrowing that marks the category’s civilian rather than criminal character. The grounds are the authority’s stated reasons, the operative document of the whole system, distinct from the material, which is the evidence and intelligence behind the grounds and may be partly withheld. The satisfaction is the authority’s recorded conclusion that detention is necessary, subjective in form and examinable for relevance, proximity, and honesty. Confirmation is the government’s act, after the board’s report, fixing the detention’s continuance within the statutory ceiling, and revocation is the government’s act of ending it at any stage, the power whose existence makes the representation a real remedy. A reference is the government’s submission of the case to the Advisory Board. A fresh order is a new order after an earlier one has expired or fallen, lawful on new material and constrained on old. Sufficient cause is the board’s constitutional question, the phrase clause (4) uses for what the board must find before custody crosses three months. Each term names a stage or an act in the life cycle this article has described, and the reader who meets an unfamiliar detention statute should translate it into these terms first, because the statutes differ in their labels far more than in their structure.

Who Is Detained: The Subjects of the System

The architecture’s subjects repay a direct look, because the distance between the power’s stated theatres and its lived population is where criticism and defence both draw their evidence. At the Union level the named theatres are espionage and sabotage against defence and foreign relations, organised smuggling and foreign exchange racketeering, and the narcotics trade’s organisers. At the state level the population is different in kind: local figures characterised by the state Acts’ labels, alleged bootleggers, land grabbers, drug peddlers, and men with long sheets of untried accusations, together with, in disturbed regions, persons detained on security grounds in circumstances far from the district pattern. Political detention, the use that dominates the field’s history and its public memory, cuts across the levels: it was the substance of the Emergency detentions, it attends insurgency and separatist movements, and it surfaces whenever a well-known figure is taken into custody without charge, which is the moment the wider public meets the institution and forms its view of the whole.

The pillar draws one conclusion from the population, and it is structural rather than statistical. The procedural discipline this article has described was designed in a constitutional text addressed to a security power, and it is applied overwhelmingly to a district population held under reputation-based state statutes. That mismatch does not invalidate the discipline, since the disciplines are general and the courts apply them without distinction, but it does mean that the system’s real test is not the rare espionage case for which the framers’ minds were formed. It is the ordinary goonda Act order, passed by a district officer on a police dossier, against a man whose alleged acts were never charged. A guide that prepared the reader only for the security paradigm would fail the population the system actually holds.

The Constitution’s Wider Custody Settlements

Preventive detention is the most prominent of several places where the Constitution contemplates custody and coercion outside the ordinary criminal process, and placing it among them sharpens what is distinctive about it. Article 33 empowers Parliament to restrict or abrogate the fundamental rights of members of the armed forces, paramilitary forces, police forces, and intelligence organisations, a settlement about discipline within the state’s own services rather than about custody of citizens. Article 34 contemplates indemnity and restriction in areas under martial law, a settlement for the breakdown of civil order itself. The Emergency provisions, already discussed, suspend enforcement of rights across the field. Against that company, preventive detention’s distinctiveness is its ordinariness. It requires no proclamation, no martial law, no special theatre; it is a standing peacetime power woven into the ordinary statute book and the ordinary district administration, available on a Tuesday in a peaceful state against a bootlegger. That ordinariness, more than its severity, is what separates the Indian settlement from the emergency detention powers of comparable constitutions, and it is why the field’s argument never settles: the power is simultaneously routine in its operation and exceptional in its character, and every participant in the debate is describing one of those truths and discounting the other.

The Eras of Use

Read across seven decades, the system has passed through recognisable eras, each of which shifted the population and the politics while leaving the architecture untouched. The first era, of the 1950 Act, was the founding state’s: detention aimed at insurgency, communal organisation, and the black market of a scarcity economy, defended by a government many of whose members had been detained under the colonial power and attacked by an opposition some of whose members the Act then detained. The second era was the Emergency, when the general statute of the day was turned to political custody at a scale that redefined the institution in public memory and generated the amendment settlement, half commenced and half withheld, that the field still works under. The third era was the specialist turn of the nineteen seventies and eighties, when the system’s growth came through theatre-specific statutes, smuggling and foreign exchange in 1974, blackmarketing in 1980, narcotics in 1988, reflecting a state that had identified the insulated organiser as its characteristic problem. The fourth era is the state era, the proliferation and heavy use of the reputation-based state Acts, in which the centre of gravity moved decisively to the districts and the doctrine moved with it, the circles test and the live link and the representation disciplines being forged in litigation over local orders rather than national security. The eras matter to the reader because the case law is era-stamped: a proposition about the institution drawn from Emergency litigation answers a political paradigm, and a proposition drawn from a goonda Act case answers a district one, and confusing the two produces the confident generalisations about the field that its specialists learn to distrust.

The Teaching Life of Article 22

No account of this field is complete without its examination life, because Article 22 is among the most heavily taught and tested provisions in Indian public law, in the civil services and judicial services syllabi and in every law school’s constitutional course. The teaching tradition has fixed the provision’s presentation into a shape the pillar has deliberately resisted and should now name: the clauses memorised as a list, the safeguards recited as a matching list, and the cases attached as labels. That shape produces the errors collected above, because the lists sever the connections that carry the law: clause (3) from the safeguards it displaces, clause (4) from the uncommenced amendment that would have rewritten it, clause (5) from the litigation that made it the system’s working remedy. The method this article has used, architecture first and vocabulary last, is offered as the corrective the cluster’s study article will develop for examination purposes in its own territory. For the pillar it is enough to record that the provision repays structural reading better than any other in the detention field, since its meaning is genuinely in its arrangement: an authorisation, an exclusion, a clock, a voice for the detenu, a withheld file, and a reserve power, in that order.

The Comparison With Other Democracies

The claim that India stands nearly alone in constitutionalising peacetime preventive detention survives scrutiny, provided it is stated as a claim about constitutional design rather than about practice. Other democracies have detained without trial, and their histories are not cleaner than India’s; the difference is where the power lives in their law. The United Kingdom interned on suspicion in both world wars under emergency regulations, and interned in Northern Ireland in 1971 under emergency powers whose consequences are still litigated in public memory, but its ordinary constitution contains no standing peacetime detention code, and each episode arrived dressed as an exception and departed as a scandal. The United States interned its own citizens of Japanese ancestry in the Second World War under executive and military authority ratified by a wartime Supreme Court, a precedent the same court has since disavowed in terms, and again the power claimed was emergency and racial, not a codified peacetime system with boards and clocks. The Indian settlement differs in kind: a permanent textual authorisation, enumerated grounds, constitutional safeguards, and a standing statute book administered in ordinary times by ordinary governments. Internationally, the design sits in visible tension with the liberty guarantee of the International Covenant on Civil and Political Rights, whose Article 9 states the general norm against arbitrary arrest and detention, and India’s own instrument of accession carried a declaration reading that Article together with the preventive detention clauses of Article 22, a formal acknowledgment by the Indian state itself that the two texts pull in different directions and that the Constitution’s settlement was the one it intended to keep.

The comparison disciplines the domestic argument in both directions. Abolitionists may fairly say that no comparable democracy has found it necessary to write the power into its fundamental law in peacetime, and that the Indian exception therefore requires an Indian justification rather than an appeal to universal practice. Retentionists may fairly answer that the comparison of texts is not a comparison of behaviour, that states without the text have detained worse and with less procedure when frightened, and that India’s distinction is candour: the power is named, bounded, and litigated in the open, under a constitution that at least forces the argument to happen in courtrooms with reported judgments. The pillar records the comparison because the One Test of this article is the reader’s ability to place any Indian detention law in its system, and the system itself sits in a world of alternatives that its defenders and critics both invoke, usually in caricature.

The File: The Paper Record of a Detention

Preventive detention is a paper system, and the pillar’s last piece of practical anatomy is the file itself, because the file is what every actor in the system actually handles. A complete detention record contains the proposal or report that moved the authority, ordinarily originating in a police or agency dossier of incidents and intelligence; the authority’s noting recording its satisfaction and the grounds drawn from the material; the order itself, addressed to the detenu and directing custody at a named place; the grounds as served, with the copies of the material the law requires to accompany them; the proof of service and its date, from which the clocks run; the detenu’s representation and the file’s record of its transmission and consideration; the government’s approval and confirmation orders; the reference to the Advisory Board with the material placed before it; and the board’s report. Every discipline described in this article attaches to an item in that list. Delay is measured between two dated papers. Vagueness is a property of one document. The live link is tested by comparing the incidents paper with the order’s date. The board’s constitution is a fact recorded in a notification. A challenge is, in substance, an audit of the file, and the reason procedural challenges succeed so often is that files assembled at district speed, across several hands, under political time pressure, so frequently fail the audit the Constitution prescribes.

The file also explains the system’s memory. Because each order generates a complete paper trail that the writ court can call for, the executive’s practice is continuously documented in a form later cases can cite, and the doctrines of the field are best understood as rules the courts distilled from reading thousands of such files and finding the same failures in them: the satisfaction recorded in another’s words, the grounds served late, the representation answered after the board had reported, the stale incident dressed as a prognosis. A student who wishes to understand why the case law reads as it does could do worse than to imagine the file room, since the law of preventive detention is, to an unusual degree, the law of what a government file must contain before the state may hold a person without trial.

Legislative Oversight and the Renewal Habit

The system’s accountability runs through courts and boards in the main, but it has a legislative strand that the founding era took seriously and later practice let thin. The Preventive Detention Act of 1950 was temporary by its own terms, and its repeated extensions forced Parliament to debate the power’s continuation at intervals, with the Home Minister required to justify the detentions of the preceding period to a chamber that included the power’s sharpest critics. Whatever the outcomes of those debates, the mechanism imposed a rhythm of public accounting that a permanent statute does not. The successor statutes are, in the main, permanent, and their oversight runs through the ordinary instruments of questions, debates on demands, and occasional amendment, none of which compels a government to lay its detention record before the legislature as a condition of the power’s survival. The sunset reform noticed in the argument section is, in substance, a proposal to restore the founding rhythm by design rather than by accident. The pillar records the strand because a complete map of the system’s disciplines should show where a discipline used to be and is now attenuated, as well as where disciplines operate. The courts’ audit grew as the legislature’s renewal habit faded, and some part of the field’s judicial character, its dense proceduralism and its suspicion of executive paper, is the judiciary doing work that the founding design distributed more evenly.

A Note on Numbers in This Field

The reader will have noticed that this article, in a field where numbers seem to matter enormously, has offered almost none beyond the constitutional figures themselves, and the omission is a position rather than an oversight. Detention counts in India are generated by two levels of government under more than a dozen statutes, recorded in categories that shift between sources and years, and published, where they are published, in fragments that resist honest aggregation; quash proportions depend entirely on the subset of orders that reach a court and are reported, which is a small and unrepresentative sample of the orders passed. An article that announced a precise national detention total or a confident percentage of orders struck down would be manufacturing the appearance of knowledge, and the series rule against invented statistics applies with special force to a subject where a wrong figure will be quoted by readers who trust the page. What can be said, and has been said, is directional and structural: that state orders vastly outnumber central ones, that procedural failure is the characteristic judicial finding, that the specialist statutes serve narrower populations than the general and state Acts, and that every one of those propositions is visible in the reported case law and the statute book without recourse to a single contested total. The discipline is worth naming because it is transferable: in legal writing, a pattern stated at the confidence the record supports is worth more than a figure stated beyond it.

The Family at the Gate

One constituency of the system appears in no clause and in every case: the detenu’s family, for whom the detention arrives as an absence and an order served on someone else. The family’s practical position follows from the architecture. The detenu’s remedies are the detenu’s, the representation and the petition, and the family’s role is to set them in motion: to obtain the order and the grounds, to place them before a practitioner while the representation window and the board timetable are still open, to keep the dates, and to pursue the information the system does owe, since the grounds served on the detenu are the family’s map of the case against him. Conditions of custody, interviews, and correspondence are governed by the applicable statute’s rules and the orders of the detaining government rather than by the criminal process’s undertrial regime, and grievances on those matters run to the government and, on the human rights side, to the commissions whose route this article has mapped. The pillar ends its anatomy here deliberately. A system description that stops at institutions misses the hands that actually work the remedies, and in this field those hands, at the start of almost every case that reaches a board or a court, belong to a family that received a copy of an order and refused to treat it as the end of the matter.

The Ordinance Door

The system’s statutes have repeatedly entered through the ordinance door, and the route is part of the architecture rather than an accident of its history. The National Security Act arrived first as an ordinance in September 1980, as the A. K. Roy record itself notes, and was then enacted by Parliament in that form; the pattern of executive initiation followed by legislative adoption recurs across the security statute book, where governments habitually judge that a gap in the detention code cannot wait for a session. The constitutional discipline on the route is the general one: an ordinance has the force of an Act, lapses unless laid before and approved by the legislature within the constitutional period after reassembly, and counts as law for the purposes of Article 21, a proposition A. K. Roy also settled against the argument that an ordinance, being executive in origin, could not establish the procedure Article 21 requires. For preventive detention the route carries a particular irony the reader should register. A power whose justification is urgency is itself legislated by the urgency procedure, so the first detentions under a new statute may rest on a text no legislature has yet debated, and the safeguards debate happens, if it happens, after the power is already in hands. The pillar does not treat the route as an abuse in itself, since the Constitution provides it and the courts have sustained its products, but a complete map must mark the door, because several of the field’s most consequential texts came through it.

What to Ask of Any New Detention Proposal

The architecture yields a standing test for the detention proposals that legislatures will go on producing, and the pillar states it as its final piece of equipment for the reader. Ask first for the entry: which of the two Seventh Schedule doors the proposal claims, and whether its grounds stay inside that door’s subjects. Ask next for the gap: what failure of the criminal process, stated with instances rather than adjectives, the proposal answers, and whether the existing statutes, central and state, already reach the conduct. Ask for the officer: who may order, at what level, and what approval discipline and lapse rule govern an officer’s order in its first days. Ask for the clocks: the interval for serving grounds, the timetable for the representation’s consideration, the board reference date against the three-month rule, and the ceiling on total duration, and compare each against the floors and ceilings this article has mapped. Ask for the board: its composition against clause (4), and whether the proposal attempts any clause (7) displacement of the board, which is the single most revealing clause in any new text. Ask, finally, for the sunset: whether the proposal expires and must be re-earned, or enters the permanent book by default. A proposal that answers all six questions cleanly may still be unwise, but it is at least legible to the discipline the Constitution built. A proposal that blurs the board, stretches the grounds beyond its entry, or hides its ceiling has announced, in its text, the litigation that will follow it and the orders that will fall.

Foreigners, Enemy Aliens, and the Word Person

Article 22 speaks throughout of persons, not citizens, and the choice of word distributes the system’s protection and its exclusions with precision. The safeguards of clauses (4) to (7) protect any person detained under a detention law, citizen or foreigner, and the writ courts have entertained foreign detenus’ challenges on the same procedural terms as citizens’, since Article 21 likewise protects persons. Clause (3) carries the system’s one status-based exclusion in its first limb: the arrest protections of clauses (1) and (2) do not apply to an enemy alien, a category of wartime law aimed at nationals of a state at war with India, distinct from the detention limb of the same clause and far narrower than a general foreigner exception. The distinction matters in practice because foreigners meet the Indian state at several different legal doors that public discussion habitually merges. Detention under a preventive detention statute is the system this article maps, with its grounds, representation, and board. Custody pending deportation rests on the foreigners legislation and its own logic of removal, in which the state’s claim is not that the person will act prejudicially but that the person has no right to remain. A foreign national may meet both doors in sequence, and the remedies differ by door, so the first task in any such case is the architectural one of identifying which power holds the person. The pillar records the point because the system’s own texts insist on it: the Constitution distributes its protections by the word person and its exclusions by named categories, and reading either loosely produces confident errors.

Article 19 and the Detention Laws

The relationship between Article 19 and preventive detention completes the doctrinal frame, and it is the part of the frame a reader most needs in order to understand what the Gopalan settlement excluded and the later cases restored. Article 19 guarantees to citizens a set of freedoms, movement and residence among them, that a detention order restricts as completely as any state act can: a detained citizen can neither move throughout the territory of India nor reside where he chooses. Under the Gopalan silo reading, that restriction was simply not the detention law’s constitutional problem, since Article 22 supplied a complete code for detention and Article 19 addressed other state action. Cooper’s rejection of the silos and Maneka Gandhi’s reconstruction of Article 21 changed the premise: state action is to be tested against the fundamental rights in combination, and a detention law’s validity may be argued through the fairness of its procedure under Article 21 and, where a freedom of Article 19 is engaged by a law’s operation, through the reasonableness discipline that Article 19 carries. In the daily work of the writ courts the Article 21 and Article 22 disciplines do nearly all the labour, for reasons of fit: the detenu’s complaint is characteristically about custody and its procedure, which are the texts of Articles 21 and 22, and Article 19 is confined to citizens while detention is not. But the restoration matters at the level of principle, because it ended the era in which a detention law could be defended as constitutionally self-sufficient, answerable to one Article alone. Every detention statute and every order under it now stands inside the whole fundamental rights chapter, and the pillar’s architecture, which began with a single Article, properly ends with the chapter.

A Worked Reading of One Order

The method is best fixed by working it once against a representative case, stated in general terms so that no real person’s file is described. Suppose a District Magistrate, specially empowered under a state preventive detention Act, orders the detention of a local figure on grounds that recite four incidents spread across six years: two prosecutions that ended in acquittal, a later altercation at a market in which a crowd gathered and trade shut for an afternoon, and intelligence of a plan to obstruct a public procession. The order is served, the grounds follow nine days later without any recorded explanation for the interval, the representation is rejected by the government after the Advisory Board has already reported in favour of detention, and the confirmed period runs to the statute’s ceiling. The architecture locates every question this order raises. Competence is not in issue: public order is Entry 3 ground and the Act is a state Act within it. The satisfaction is vulnerable at two points the doctrines name exactly. The six-year spread, with the two oldest incidents being acquittals, tests the live link: the authority must show why incidents of that age and outcome support a prognosis, and a file that treats them as self-evidently sufficient invites the court to break the chain between that altercation and the rest. The grounds’ interval of nine days tests clause (5): the representation right runs from communication, and an unexplained delay of that length in serving the grounds is the kind of lapse the representation cases treat as vitiating, since the detenu’s earliest opportunity was postponed by the state’s own hand. The sequence of the representation’s rejection, after the board’s favourable report, tests the consideration discipline from the other side: the board’s report does not exhaust the government’s duty on the representation, and a rejection that appears to have waited on the board suggests the government treated its own clause (5) obligation as discharged by a different body’s proceeding. The single strong incident, the market altercation with its community-facing shutdown, is the order’s best ground under the circles test, and a court persuaded that the satisfaction genuinely rested on it might still save the order; the accumulated procedural failures are what place that rescue in doubt. No figure in the example is real, and that is the point of the exercise: the analysis required no view about the man’s character or the district’s politics, only the map, the clocks, and the file, applied in order.

How Different Readers Should Use This Guide

The advocate’s route through the article runs from the architecture map to the file: identify the layer, audit the papers against the clocks, and locate the failure the doctrines will recognise, since in this field cases are prepared as audits and argued as audits. The law student’s route runs from the clause-by-clause reading to the doctrinal spine: hold clause (3) and clause (4) as the twin facts, then watch Gopalan build the silo, Cooper break it, and Maneka Gandhi rebuild the standard, and the rest of the syllabus arranges itself as commentary on that movement. The researcher’s route runs from the replacement pattern to the eras of use: the statutes are the data, their succession is the argument, and the questions worth pursuing are the ones the pattern poses, why no repeal has ever reached the constitutional layer, why the commencement power has been the graveyard of the field’s chief reform, and why a discipline designed for security custody came to govern a district population. The three routes share one habit, which the guide has tried to instil in every section: never accept a proposition about this field until it has been located in the architecture, because a proposition that cannot be placed, in a clause, an entry, a statute, a stage of the file, or a doctrine of the courts, is not yet a proposition about preventive detention in India at all, whatever its confidence or its source.

The Meaning of Sufficient Cause

The board’s question deserves a section of its own, because the phrase sufficient cause is the only merits standard the Constitution itself supplies anywhere in the detention clauses, and its content determines what the three-month gate actually examines. Sufficient cause is not proof of an offence, and the board does not ask whether the detenu committed the acts the grounds allege to the criminal standard; a board that tried the allegations would exceed its design and duplicate a court without a court’s protections. Neither is sufficient cause a bare echo of the authority’s satisfaction, since a board confined to asking whether the authority was satisfied would add nothing to the government’s own confirmation and the clause would be ornamental. The working content the phrase has acquired sits between those poles. The board examines whether the material placed before it, taken as a whole, supports the prognosis the authority recorded: whether the incidents are real in the sense of being documented rather than asserted, whether they bear on the statutory ground invoked, whether they are proximate enough to speak to future conduct, and whether the inference from the material to the necessity of continued custody is one the material can carry. It is, in short, a review of the reasonableness of a prediction on the papers, conducted by persons of judicial standing, in private, without counsel, and its strengths and its blind spots are exactly those of that description. The reader who grasps sufficient cause grasps why the same detention can survive the board and fall in the writ court, or survive the writ court on procedure having troubled the board on substance: the two bodies ask different questions of the same file, and the system’s protection lies in the conjunction, not in either examination alone.

Verdict: A Safeguard System Whose Protections Sit Beside the Power

The series thesis for this cluster holds that the text of Article 22 looks like a set of safeguards, and the record shows a set of safeguards whose most protective clauses are expressly disapplied to the people who need them most. This pillar’s tour of the architecture confirms the thesis in its precise form. The protection against arrest in clauses (1) and (2) is real, and it stops at the detenu’s door by the express words of clause (3). The safeguards that replace it, grounds, representation, and the board, are real as well, and the case law shows them defeating orders in substantial numbers, and they operate inside the executive process rather than in open court, on material the detenu may not fully see, under a threshold Parliament voted to halve and no government has commenced. Reading the record rather than the text is the whole discipline of this field, because the text alone will tell the reader that India has safeguards, and only the record shows where they bite, where they are withheld, and where they wait on a notification that has been outstanding since 1979.

For the reader continuing into the cluster, the routes are fixed by the map. The founding statute and its passage belong to the history of the 1950 Act. The general central statute belongs to the National Security Act guide. The judicial extremes of the system belong to the habeas corpus article that follows in the cluster’s order, and the comparison with ordinary arrest belongs to the comparison article. Each of those articles assumes the architecture stated here and spends its length on its own layer. What this pillar asks the reader to carry into them is the two-fact discipline of its namable claim: the twin exclusion that defines the detenu’s starting position, and the three-month clock that defines the system’s rhythm. Hold those, and the cluster, and the field beyond it, stay legible.

Frequently Asked Questions

Q: What is preventive detention and who decides it?

Preventive detention is custody ordered to prevent anticipated future conduct prejudicial to specified public interests, such as the security of India, public order, or essential supplies, rather than custody imposed as punishment for a proved past act. It requires no charge, no trial, and no conviction. The decision is the executive’s: a detention order is passed by the Central Government, a state government, or an officer specially empowered under the applicable statute, acting on that authority’s own satisfaction recorded from the material before it. A court does not pass the order and does not confirm it; an Advisory Board reviews custody that is to extend beyond three months, and the constitutional courts examine the order’s validity if it is challenged.

Q: Which article of the Constitution allows preventive detention?

Article 22 of the Constitution is the provision that governs preventive detention. Its clauses (1) and (2) state the protections of an ordinarily arrested person, and clause (3) then provides that those two clauses do not apply to a person arrested or detained under a law providing for preventive detention. Clauses (4) to (7) build the detention regime itself: the three-month Advisory Board threshold, the right to be told the grounds and to make a representation, the power to withhold facts against the public interest, and Parliament’s power to prescribe maximum periods, exceptional classes of case, and board procedure. Legislative competence to enact the statutes comes separately from Entry 9 of the Union List and Entry 3 of the Concurrent List in the Seventh Schedule.

Q: Which central and state preventive detention laws are in force?

The standing central statutes are the National Security Act, 1980, the general security law; the Conservation of Foreign Exchange and Prevention of Smuggling Activities Act, 1974, aimed at smuggling and foreign exchange offending; the Prevention of Illicit Traffic in Narcotic Drugs and Psychotropic Substances Act, 1988; and the Prevention of Blackmarketing and Maintenance of Supplies of Essential Commodities Act, 1980. Alongside them, state legislatures have enacted their own preventive detention Acts under Entry 3 of the Concurrent List, including the Tamil Nadu Act of 1982, the Gujarat Prevention of Anti-Social Activities Act, 1985, and the Jammu and Kashmir Public Safety Act, 1978. State governments also pass large numbers of orders under the National Security Act itself.

Q: Can a state government make its own preventive detention law?

Yes, within the subjects the Constitution assigns to it. Entry 3 of the Concurrent List in the Seventh Schedule empowers Parliament and the state legislatures alike to legislate for preventive detention for reasons connected with the security of a state, the maintenance of public order, or the maintenance of supplies and services essential to the community. A state Act must stay within those subjects, because the State List contains no detention entry of its own. Detention connected with defence, foreign affairs, or the security of India is reserved to Parliament alone by Entry 9 of the Union List. Every state statute, whatever its subject, must also satisfy the safeguards of Article 22, clauses (4) to (7), which bind all legislatures equally.

Q: Who sits on the Advisory Board and what does it decide?

An Advisory Board under Article 22, clause (4) consists of persons who are, or have been, or are qualified to be appointed as Judges of a High Court, so its members carry serving, past, or attainable High Court judicial standing. The Forty-fourth Amendment’s proposal to reconstitute boards on the Chief Justice’s recommendation, with a serving judge as chairman, was enacted but never brought into force. The board decides one question: whether, in its opinion, there is sufficient cause for the detention of the person whose case the government has referred to it. If the board reports sufficient cause before three months expire, the government may confirm detention beyond that point. If it reports no sufficient cause, the government must revoke the order and release the detenu.

Q: How long can a person be held in preventive detention in India?

There is no single period, because three different figures operate at three levels. Article 22, clause (4) sets a threshold of three months: beyond that point, detention may continue only if an Advisory Board has reported sufficient cause, or a parliamentary law under clause (7) covers the case. Three months is therefore a gate, not a maximum, and statements that it is the longest lawful detention are wrong. Above the gate, each statute fixes its own ceiling for a single detention under it. The National Security Act fixes a maximum period of twelve months, and the specialist central statutes fix one year in the ordinary case. When the confirmed period expires, the detention ends unless a fresh order is lawfully passed.

Q: Do the protections of Article 22(1) apply to a person in preventive detention?

No. Article 22, clause (3) expressly provides that clauses (1) and (2) do not apply to a person arrested or detained under a law providing for preventive detention. The detenu therefore has no constitutional entitlement under clause (1) to be informed of the grounds of arrest in the arrested person’s form, no right under that clause to consult and be defended by a legal practitioner of choice, and no right under clause (2) to be produced before a magistrate within twenty-four hours. In their place, clauses (4) to (7) supply a different set of protections: communication of the grounds of detention, the earliest opportunity to make a representation, and the Advisory Board gate at three months. Articles 21, 32, and 226 are not disapplied and remain fully available to the detenu.

Q: Why does India allow preventive detention in peacetime?

The Constituent Assembly wrote the power into the Constitution at the Republic’s founding, in the aftermath of Partition, with mass violence, a war in Kashmir, and insurgency confronting a state whose ordinary criminal machinery was built for proved past acts rather than anticipated harm. Members who defended the clauses treated the power as a necessity a responsible government could not surrender, while members who opposed them warned against authorising custody without trial in a fundamental rights chapter, and the Assembly retained the power with written safeguards rather than abolishing it. Successive legislatures have renewed the settlement at the statutory level ever since: the 1950 Act, the Maintenance of Internal Security Act, and the National Security Act succeeded one another, because no governing coalition has accepted governing without the instrument and no amendment has removed the constitutional authorisation beneath it.

Q: What is the difference between preventive detention and punitive detention?

The two differ in purpose, in timing, and in what must be established. Punitive detention is imprisonment imposed after a trial and conviction as punishment for an offence already committed, and its length is a sentence proportioned to the proved act. Preventive detention is custody imposed before any offence is proved, indeed before the anticipated act has occurred at all, on an executive authority’s satisfaction that a person is likely to act prejudicially to a statutory subject such as public order or the security of India. Because nothing is being punished, preventive detention creates no conviction and no criminal record, is not measured against the gravity of a past act, and is disciplined by the procedure of Article 22 rather than by the criminal courts’ sentencing law.

Q: On what grounds can a person be preventively detained in India?

The grounds are those the Seventh Schedule entries and the statutes built on them name. At the Union level they are conduct prejudicial to the defence of India, the relations of India with foreign powers, or the security of India, together with the specialist subjects of the central statutes: smuggling and foreign exchange conservation, illicit traffic in narcotic drugs and psychotropic substances, and blackmarketing that threatens essential supplies. At the state level they are conduct prejudicial to the security of a state, the maintenance of public order, or the maintenance of supplies and services essential to the community. A ground must fall within the statute invoked and the entry supporting it, and the courts insist that ordinary crime, without community-facing impact, is a law and order matter for the criminal process and not a detention ground.

Q: What can Parliament prescribe under Article 22(7) about preventive detention?

Clause (7) reserves three matters to Parliament. It may prescribe the circumstances and the classes of case in which a person may be detained for longer than three months without obtaining the opinion of an Advisory Board under clause (4). It may prescribe the maximum period for which any person may be detained in any class or classes of case under a detention law. And it may prescribe the procedure to be followed by an Advisory Board in its inquiries. The first power is the deepest reserve in Article 22, because the board gate itself operates subject to it. The second explains why maximum periods differ between statutes, and the third why board procedure is a statutory matter built on the constitutional frame.

Q: What happens if the Advisory Board reports that there is no sufficient cause for detention?

The detention must end. Where an Advisory Board reports that there is in its opinion no sufficient cause for the detention of a person, the appropriate government is bound to revoke the detention order and cause the detenu to be released forthwith, a consequence the statutes state in terms and the courts enforce without qualification. The board’s unfavourable report is the one point in the system at which a non-judicial body’s opinion operates as a command rather than as advice, and a government may not confirm a detention over it or substitute its own assessment for the board’s. This is also why the board’s composition and the timing of its report matter so much in litigation: a favourable report obtained late, or from a board not constituted as clause (4) requires, cannot do the work the clause assigns to it.

Q: Can a preventive detention order be passed against a person who is already in jail?

Yes, an order may lawfully be passed against a person already in custody, but it is examined with particular care. The detaining authority must be aware that the person is in custody and must record a satisfaction that detention is nevertheless necessary, ordinarily because release is imminent or likely, as where bail is expected or the custody is near its end, and because the person’s anticipated conduct on release engages the statutory ground. An order passed in ignorance of the existing custody, or against a person with no realistic prospect of release, discloses a satisfaction formed without the central fact of the case and is vulnerable on that ground. Where the authority knows of the custody and shows why preventive custody must follow it, the courts accept the order’s validity.

Q: Why can part of the grounds of detention be withheld from the detenu?

Article 22, clause (6) provides that the detaining authority is not required to disclose facts which it considers to be against the public interest to disclose. The design protects the sources behind the grounds: informers whose identity would expose them, intelligence whose revelation would defeat the purpose of the detention, and material whose circulation would harm the very interest the detention defends. The withholding is of facts, not of the grounds themselves, which clause (5) requires to be communicated so that a representation can be made. The tension between the two clauses is structural and permanent: the detenu must answer a case whose full contents may be partly invisible, while the Advisory Board, which sees the material, tests a case the detenu could not fully meet. Courts police the boundary by asking whether what was withheld was genuinely a fact of that character rather than the substance of a ground.

Q: Is India unusual among democracies in writing preventive detention into its Constitution?

Yes, in the specific sense that the power is a standing part of the constitutional text for peacetime use. Other democracies have detained without trial, the United Kingdom by internment under emergency powers in two world wars and in Northern Ireland, and the United States by the wartime internment of citizens of Japanese ancestry, an episode its Supreme Court has since disavowed. In each of those systems the power claimed was emergency power, presented as an exception and remembered as a scandal, and the ordinary constitutional text contains no permanent detention code with grounds, boards, and clocks. India’s settlement is different in kind: the authorisation, the subjects, and the safeguards are all written into the fundamental rights chapter itself, to be used by ordinary governments in ordinary times, and defended or attacked as a feature of the constitutional design rather than as a departure from it.

Q: Can a person be detained under a fresh order after an earlier order has been quashed?

It depends on why the earlier order fell. Where a court quashed the order because the material did not support the satisfaction, the same grounds cannot be recycled into a fresh order, because the adjudication that the material was insufficient would be nullified if the authority could simply sign again. Where the earlier order fell on a procedural defect, such as delay in serving the grounds or in considering the representation, the state’s position is stronger, and a fresh order on the same material, passed with the defect cured, has a firmer footing, though the courts examine such sequences closely for good faith. Fresh grounds arising from conduct after the earlier order stand apart altogether and may support a new order on their own terms. In every case the fresh order must record its own satisfaction.

Q: On what grounds do courts strike down preventive detention orders?

Courts strike down detention orders chiefly on procedure, because the Constitution commits the assessment of danger to the executive and the courts audit the process by which it was reached. The recurring grounds are delay or failure in communicating the grounds of detention; failure to supply the material the law requires with them, so that the representation right is emptied; delay by the government in considering the detenu’s representation, or its consideration by the wrong authority; a reference to the Advisory Board made late or to a board not constituted as clause (4) requires; grounds that are vague, irrelevant to the statute, or built on stale incidents with no live link to a prognosis; a satisfaction recorded on another’s dictation; and detention used for a collateral purpose, such as defeating a bail order. Orders also fall where the conduct alleged amounts only to a law and order disturbance and never reaches public order.

Q: Who can file a habeas corpus petition on behalf of a person in preventive detention?

A habeas corpus petition under Article 226 in a High Court, or Article 32 in the Supreme Court, may be filed by the detenu, and it may also be filed by another person acting on the detenu’s behalf, such as a spouse, a parent, a relative, or a friend, since a person in custody is often unable to approach the court directly and the writ exists precisely for that situation. The petitioner must show the relationship or the reason for acting for the detenu, and courts scrutinise petitions by strangers with no connection to ensure the remedy is not being misused, but a genuine next friend is not turned away for want of the detenu’s own signature. The petition challenges the detention’s legality, and the court calls for the order, the grounds, and the file, testing them against the statute and Article 22.

Q: What is COFEPOSA and why does it treat smuggling as a subject for detention?

The Conservation of Foreign Exchange and Prevention of Smuggling Activities Act, 1974, shortened everywhere to COFEPOSA, is the central specialist statute that authorises preventive detention of persons engaged in smuggling, in abetting it, in transporting, concealing, or harbouring smuggled goods, in dealing in such goods, and in activities prejudicial to the conservation and augmentation of foreign exchange. Its premise is structural: organised smuggling and foreign exchange rackets are directed by principals who insulate themselves from the physical acts the criminal law can prove, working through carriers and intermediaries, so prosecution habitually convicts the instruments and leaves the organisation intact. Detention is aimed at the interval in which the network can be disrupted by removing the organiser. COFEPOSA litigation has made the Act the standard illustration of the representation discipline, since its orders have fallen repeatedly on delay in considering detenus’ representations and in supplying material.

Q: Why have preventive detention statutes been replaced repeatedly instead of disappearing?

Because the power that sustains them has never been removed. The Preventive Detention Act, 1950 was extended repeatedly and allowed to lapse at the end of 1969; the Maintenance of Internal Security Act, 1971 succeeded it and disappeared in the repeal that followed the Emergency; the National Security Act, 1980 restored the standing power and remains the general central statute. Each lapse or repeal was a statutory event, reversible by the next legislature, while the authorisation in Article 22 stood untouched beneath the sequence, and no repeal coalition has ever carried the argument to the constitutional level where the power resides. Replacement is therefore the pattern a permanent constitutional power produces at the statutory level: governments of every complexion have concluded, on taking office, that they require the instrument, and have drafted it afresh rather than governing without it, so the statutes change their names and their details while the institution continues without interruption.