The comparison between preventive detention vs arrest is not an academic distinction, and it is usually learned at the worst possible moment: in a police station corridor, on the telephone to a lawyer, in the first confused hours after a person has been taken away and nobody in the family can say under what power. The two regimes look alike from the corridor. In both, uniformed officers take a person into custody. In both, the person is lodged in a jail or a lock-up. In both, the family is told to wait. Yet almost everything that follows from that moment, the rights the person holds, the documents that must exist, the forum that can be approached, the clock that is running and the way release is obtained, depends entirely on which of the two regimes is in operation, and the two regimes share almost nothing beyond the physical fact of custody.

An arrest is the opening move of a prosecution. It asserts that an offence has been committed, that this person is suspected of committing it, and that a court will in due course decide whether the assertion is true. Preventive detention is not the opening move of anything. It asserts no offence, begins no prosecution and promises no trial. It rests on a prediction: an executive authority is satisfied that a person is likely to act in a manner prejudicial to public order, to the security of the State or to some other protected interest, and that the person must be confined now to prevent the predicted act. The first regime looks backward to a deed and forward to a verdict. The second looks only forward, to a deed that has not happened, and it ends not with a verdict but with the expiry of a fixed period. Getting the regime wrong in the first forty-eight hours wastes the only period in which anything can be done quickly, because the remedies are not interchangeable. A bail application filed in a detention matter is a document addressed to a court that has no case before it. A family that waits for a charge sheet in a detention matter is waiting for a document the law never requires anyone to file.

This article resolves the comparison completely. It explains how to identify the regime from the face of the order, sets out the machinery of each side, states the decision table that separates them row by row, and then examines the doctrine that decides most contested cases, the distinction between law and order and public order drawn by the Supreme Court in Dr. Ram Manohar Lohia v. State of Bihar in 1966. It addresses the hardest point in the field, the detention order passed against a person who is already in custody in a criminal case, and it deals honestly with the respects in which the comparison does not run in the direction most readers expect.

Two Regimes Under One Article

The Constitution holds both regimes in a single article, and the architecture of Article 22 is the fastest way to understand why the comparison matters. Clauses (1) and (2) protect the arrested person. A person who is arrested must be informed, as soon as may be, of the grounds of the arrest, and holds the right to consult and to be defended by a legal practitioner of choice. The arrested person must be produced before a magistrate within twenty-four hours of the arrest, excluding the time necessary for the journey from the place of arrest to the magistrate’s court, and cannot be detained beyond that period without the authority of a magistrate. Clause (3) then performs the decisive switch. It provides that neither of those two clauses applies to a person arrested or detained under a law providing for preventive detention. The grounds on arrest, the lawyer, the magistrate within twenty-four hours: the entire protective apparatus of the first two clauses is switched off for the detenu by the constitutional text itself, and clauses (4) to (7) substitute a different and thinner set of protections designed for the detention regime alone.

Those substitute protections are the spine of the detention side of this comparison. Clause (4) provides that no preventive detention law shall authorise detention for longer than three months unless an Advisory Board, composed of persons who are, have been or are qualified to be judges of a High Court, reports before the expiry of those three months that there is sufficient cause for the detention. Clause (5) requires the detaining authority to communicate the grounds of detention to the detenu as soon as may be, and to afford the detenu the earliest opportunity of making a representation against the order. Clause (6) permits the authority to withhold facts whose disclosure it considers against the public interest. Clause (7) empowers Parliament to prescribe the circumstances and classes of case in which a person may be detained beyond three months without the Advisory Board safeguard, and to prescribe the maximum period of detention. The constitutional design is therefore not one regime with exceptions. It is two regimes drafted side by side, one built around a court and the other built around an executive satisfaction supervised, at a distance, by a board and by the writ jurisdiction of the constitutional courts.

The legislative competence to build the second regime is distributed with some care. Parliament may legislate on preventive detention for reasons connected with the defence of India, foreign affairs and the security of India under Entry 9 of the Union List. State legislatures may legislate for reasons connected with the security of a State, the maintenance of public order and the maintenance of supplies and services essential to the community under Entry 3 of the Concurrent List, which is why so much of the detention actually imposed in India flows from state statutes aimed at habitual offenders, bootleggers, drug offenders and similar locally defined categories, administered by district magistrates and commissioners of police. The arrest side, by contrast, runs on the criminal procedure code of general application, the Code of Criminal Procedure 1973 and its successor the Bharatiya Nagarik Suraksha Sanhita 2023, which govern every investigation in the country. A reader who wants the internal working of the detention statutes themselves should go to the pillar treatment of the preventive detention law of India, which owns that ground; this article owns the comparison, and it keeps returning to the single practical question of which regime a given custody belongs to and what follows from the answer.

One consequence of this architecture deserves its own emphasis, because it inverts what a lay reader expects of the judicial hierarchy. A challenge to an arrest begins at the bottom of the judicial system and climbs: the magistrate who receives the production, then the sessions court, then the High Court, are approached in sequence for bail and for the supervision of custody. A challenge to a detention begins at the top. The habeas corpus petition lies directly to the High Court under Article 226 or to the Supreme Court under Article 32, because the order it attacks is an executive act unsupported by any judicial proceeding beneath it, and there is no lower court in seisin of the matter to approach first. The detenu, the person with the fewest procedural rights inside his regime, is the one litigant in the criminal field who walks straight into a constitutional court. The arrested person, surrounded by safeguards, litigates his custody before the humblest court in the system. Neither position is simply better. The magistrate’s supervision is daily and granular, and the constitutional court’s review is episodic and confined to the order’s legality, but the inversion explains why detention law is made in the High Courts and the Supreme Court while bail law is made in the magistracy, and why the decided cases that fill this article are writ cases.

The Constitutional Bargain Behind the Two Regimes

The coexistence of the two regimes was a deliberate founding choice, and understanding it prevents a common misreading of the comparison: preventive detention in India is not a colonial survival that the Constitution forgot to remove, nor an emergency device that escaped its cage. The Constituent Assembly debated preventive detention directly, in September 1949, with the experience of colonial detention laws fresh in the room, and several members who had themselves been detained without trial under British rule spoke against the provision. The Assembly nevertheless wrote it into Article 22, accepting the argument, pressed by the Drafting Committee’s leadership, that a new state facing communal violence, disorder and fragile administration could not be denied the power altogether, and choosing instead to discipline it with the safeguards of clauses (4) to (7). The bargain was explicit: the power was admitted, and in exchange it was denied the ordinary trappings of the criminal process and fenced with substitutes, the grounds, the representation and the Advisory Board.

The first great test of the scheme, A. K. Gopalan v. State of Madras in 1950, treated Article 22 as a self-contained code for preventive detention: a detention law that complied with clauses (4) to (7) was not to be struck down by importing the fuller protections that clauses (1) and (2) give the arrested person, because clause (3) had deliberately withheld them. That reading explains the comparison’s asymmetry at its root. The detenu’s thinner rights are not an oversight to be corrected by interpretation; they are the text. Later constitutional development, above all the recognition after Maneka Gandhi v. Union of India in 1978 that a procedure depriving a person of liberty must be just, fair and reasonable under Article 21, has softened the edges of the bargain without rewriting it, and the modern grounds of challenge, expedition, disclosure, application of mind, are best understood as Article 21 discipline applied inside Article 22’s structure.

Parliament has once voted to thicken the detenu’s side of the bargain and then declined to bring its own vote into force. The Constitution (Forty-fourth Amendment) Act 1978, passed in the reckoning after the Emergency, amended Article 22 to reduce the period of detention permissible without Advisory Board scrutiny from three months to two, and to require that Advisory Boards be constituted in accordance with the recommendations of the Chief Justice of the appropriate High Court, with a serving High Court judge as chairman. Section 3 of that Amendment Act, the section carrying these changes, was never notified into force, and the unamended text continues to govern. The episode matters to the comparison because it marks the constitutional ceiling of reform by amendment: the safeguards a detenu holds are those of 1950, improved by judicial doctrine rather than by the constitutional text, while the arrested person’s safeguards have been repeatedly thickened by the courts, in the arrest-memo discipline of D. K. Basu and the arrest-necessity doctrine of later cases. The two regimes have not stood still since 1950, but they have moved by different means, and the detenu’s protections are the more judge-made of the two.

A Short History of the Two Tracks Running Together

The two regimes have run in parallel since the Republic’s first months, and the parallelism is itself a fact about the Indian state. The Preventive Detention Act 1950, Act 4 of 1950, was enacted within weeks of the Constitution as a temporary measure, its life fixed initially to run to 1 April 1951, and Parliament then renewed it again and again for two decades before allowing it to lapse at the end of 1969. The Maintenance of Internal Security Act 1971 followed, and after the Emergency the Janata government repealed it outright in 1978, leaving a brief interval in which no central preventive detention statute of general application stood on the book. The National Security Act 1980 closed the interval and, unlike its predecessors, was enacted without a sunset clause, which is why it remains the standing central statute of the detention side. Through every one of these changes the arrest side ran on undisturbed: the Indian Penal Code and the successive procedure codes prosecuted offences in 1950, in 1971 and in 1980 alike, and no government has ever proposed that detention replace them. The detention statutes rise and fall with the politics of order; the criminal process is the permanent background against which they are, each time, an addition.

The pattern of enactment teaches the same lesson as the doctrine. Each detention statute was defended at its birth as a response to a specific condition, smuggling and foreign exchange manipulation for the Conservation of Foreign Exchange and Prevention of Smuggling Activities Act 1974, the security climate of 1980 for the National Security Act, and each was then retained for the general run of cases its administrators found convenient. The courts’ narrowing doctrines developed in reaction to that retention, case by case, as orders passed for ordinary criminality came up for review. The comparison between the regimes is therefore not a snapshot but a seventy-year negotiation, in which Parliament repeatedly supplies a broad power, the executive repeatedly uses it at the low end, and the courts repeatedly push it back toward the exceptional cases for which it was justified. A reader who understands that rhythm can predict the fate of most detention litigation without reading the file: the question will be whether this order belongs to the exceptional class, and the authority’s grounds will either show that it does or reveal that it does not.

Reading the Order: How to Tell Which Regime Applies

The regime is identified by documents, not by the manner of the taking. A person may be picked up from home at dawn by either regime, and the uniforms, vehicles and hour tell the family nothing. The documents tell everything. An arrest generates a record of a suspected offence: ordinarily a first information report recording the offence alleged, an arrest memo recording the time and place of arrest and the officer who made it, and within twenty-four hours a production before a magistrate who will begin to supervise the custody. A detention generates a detention order: a written order signed by a district magistrate, a commissioner of police or a government, reciting that the authority is satisfied that with a view to preventing the person from acting in a manner prejudicial to a named interest, such as the maintenance of public order, it is necessary to detain him, followed within days by a separate document setting out the grounds of detention. The presence of a named offence and a case number marks the arrest side. The presence of the formula of satisfaction and prevention marks the detention side.

How soon must an arrested person be taken before a magistrate?

An arrested person must be produced before a magistrate within twenty-four hours of the arrest, not counting the time the journey from the place of arrest to the court reasonably takes, and police custody beyond that point exists only under the magistrate’s authority. No corresponding production is owed to a detenu at all.

The twenty-four hour rule is the cleanest single discriminator between the regimes, because it is a right the arrested person holds against the police and a right the detenu simply does not have. Clause (2) of Article 22 states it, the procedure codes enforce it, and a custody that crosses the line without production is unlawful custody in the arrest regime. In the detention regime, the same lapse of twenty-four hours, or of twenty-four days, without any magistrate ever seeing the detenu is not a defect at all: no magistrate has any role in authorising, extending or supervising a preventive detention. The detenu’s counterpart protections run on different tracks entirely, the communication of grounds, the representation to the government, the reference to the Advisory Board and the writ petition. Families who demand production before a magistrate in a detention case are demanding the arrest regime’s central safeguard from a system that has deliberately replaced it, and the demand, however understandable, is addressed to a forum that cannot grant it.

The second discriminator is the offence. Behind every arrest stands an allegation that a defined offence, created by a penal statute and carrying a defined punishment, has been committed by an identifiable act at an identifiable time. The first information report that records it can be read, obtained and challenged, and the whole subsequent process, investigation, charge sheet, charge, trial, is the state undertaking to prove that allegation or release the accused from it. Behind a detention order stands no allegation of that kind. The grounds served on the detenu typically narrate past incidents, and those incidents may describe conduct that is also criminal, but they are not tendered as charges to be proved. They are tendered as the material from which the detaining authority drew its prediction. The detenu is not asked to plead to them, no court will ever return a finding on them, and an acquittal in a prosecution founded on the same incidents does not, by itself, dissolve the detention, just as a conviction is not required to sustain it. This is why the statement that a detention is illegal because no offence is alleged, heard constantly in public discussion of detention cases, is wrong: the absence of an offence is not an irregularity in the detention regime but its defining feature, and a challenge built on that ground alone will fail.

A fourth discriminator, invisible on the first day but decisive later, is the language and intelligibility of what is served. The grounds of arrest are spoken and written at the moment of the taking, in the transaction itself. The grounds of detention arrive days later as a composed document, and the courts require that the composition actually communicate: grounds so vague that the detenu cannot answer them fail the purpose of Article 22(5), because the representation right depends on knowing what is alleged, and grounds must be conveyed in a language the detenu understands if the right is to be real rather than formal. From the earliest years of the 1950 Act the Supreme Court treated effective communication, not mere delivery of paper, as the obligation. A family examining the papers should therefore ask not only whether the grounds exist but whether the person inside can read them and answer them, because a defect on that point is a ground of challenge that appears in no other part of the comparison.

The third discriminator is the identity of the decision-maker. An arrest is made by a police officer, but custody is owned almost immediately by a court: the magistrate who receives the production decides whether the person goes to police custody for questioning, to judicial custody in jail, or out on bail. A detention is ordered by the executive and is never owned by a court at all. The detaining authority is a district magistrate or commissioner of police acting under a delegation, or the Central or a State Government acting directly, and the satisfaction that matters is that authority’s own. Courts enter the detention picture only from the outside, on a writ petition attacking the order, and even then they do not substitute their own satisfaction for the authority’s. The practical consequence is that the addressee of every early step differs by regime. In an arrest, the family’s lawyer speaks to a court. In a detention, the first representations are made to the very executive that ordered the custody, and the court is approached not for custody decisions but for the destruction of the order itself.

The Arrest Side: A Prosecution With Clocks Attached

The arrest regime is best understood as a sequence of supervised steps, each with its own clock. The arrest itself is governed by the procedure code: a police officer may arrest without a warrant for a cognizable offence, the arrested person must be told the grounds of arrest and, where the offence is not bailable, informed of the right to bail, and the safeguards the Supreme Court laid down in D. K. Basu v. State of West Bengal in 1997, the arrest memo, the intimation to a relative or friend, the medical examination, discipline the manner of the taking. Production before the magistrate follows within twenty-four hours. At production the magistrate chooses among three dispositions: police custody, where the officer persuades the court that the accused’s presence is needed for interrogation or recovery; judicial custody, where the accused is lodged in jail under the court’s authority; or release on bail. Remand is not a single grant. Custody is authorised in stretches, police custody is confined to short periods at the start of the investigation, and every extension is a fresh judicial decision the prosecution must justify.

The investigation runs against the most consequential clock in the arrest regime. The charge sheet, the police report that closes the investigation and names the offences the prosecution will pursue, must be filed within a period the code fixes by the gravity of the offence: ninety days where the offence is punishable with death, with imprisonment for life or with imprisonment for a term of not less than ten years, and sixty days for other offences, under Section 167 of the Code of Criminal Procedure and the corresponding provision, Section 187, of the Bharatiya Nagarik Suraksha Sanhita. If the period expires without a charge sheet and the accused is prepared to furnish bail, the accused is entitled to be released on bail as a matter of right. This is default bail, also called statutory bail, and it is not a favour dispensed by a sympathetic court. It is the code’s penalty on an investigation that outruns its timetable: the state’s failure to complete its case within the period converts continued custody into custody the court must end. The right is indefeasible once it accrues and is claimed before the charge sheet is filed, and generations of litigation have refined the precise moment at which it accrues, which is itself evidence of how much turns on it.

Between production and charge sheet, the texture of custody differs by regime in ways families feel before they understand. On the arrest side, the first remand period is where interrogation happens: the magistrate may authorise police custody only in short stretches and, under the Code of Criminal Procedure, the total of police custody could not exceed fifteen days and had to fall within the first fifteen days of custody, after which the accused passed to judicial custody in jail. The Bharatiya Nagarik Suraksha Sanhita altered that distribution, permitting the magistrate to authorise police custody in instalments spread across the initial forty or sixty days of the custody period, depending on the charge sheet deadline that applies to the offence. On the detention side there is no interrogation architecture at all: the detenu is lodged in jail under the order, the police have no custodial right to question him under the detention, and any questioning belongs to whatever criminal cases may separately exist. A detenu who is being interrogated is being interrogated in some other capacity, and identifying that capacity is part of reading his legal position correctly.

Bail beyond the default situation is graded by the character of the offence. In a bailable offence, bail is a right, obtainable at the police station or before the magistrate under Section 436 of the Code and Section 478 of the Sanhita. In a non-bailable offence, bail is discretionary: the magistrate weighs the gravity of the accusation, the strength of the material, the risk of flight and the risk of tampering under Section 437 of the Code and Section 480 of the Sanhita, and the High Court and the Court of Session hold wider powers under Section 439 and Section 483 respectively. Anticipatory bail under Section 438 of the Code and Section 482 of the Sanhita allows a person who apprehends arrest for a non-bailable offence to obtain, in advance, a direction that he be released on bail in the event of arrest. The whole apparatus presupposes a case: a file, a court, an accusation, a hearing. Its full working belongs to the complete guide to the Code of Criminal Procedure 1973 and to the treatment of the Bharatiya Nagarik Suraksha Sanhita 2023, which replaced the Code for offences and proceedings from 1 July 2024 while carrying the arrest, remand and bail machinery across with renumbered sections and the same governing ideas. What matters for the comparison is the shape: arrest custody is court-owned from the first day, contestable at every remand, and terminable by bail through several independent doors.

Before the arrest even happens, the code now interposes a filter that has no equivalent on the detention side. For offences punishable with imprisonment of up to seven years, the police officer must record reasons why arrest is necessary at all, and the default instrument is a notice directing the person to appear, under Section 41A of the Code of Criminal Procedure and Section 35(3) of the Sanhita, rather than a taking into custody. In Arnesh Kumar v. State of Bihar in 2014, the Supreme Court converted that filter into a checklist discipline, directing that officers record the specified reasons for arrest in the lesser offences and that magistrates scrutinise those reasons when authorising detention, because arrest brings ignominy, dents reputation and curtails liberty even where it leads nowhere. The comparison this creates is instructive. The arrest regime has developed a threshold doctrine of necessity at its entry point, administered by magistrates at the remand stage. The detention regime has no magistrate at its entry point, and its threshold doctrine is the detaining authority’s own satisfaction, policed later and from outside. When critics say the detention power lacks a gatekeeper, this is the precise institutional content of the complaint: the gate exists, but it is the authority itself, and the court’s review arrives after the gate has closed.

One more feature of the arrest side completes the picture, because it decides the honest half of the duration argument later in this article. An undertrial’s custody is not capped by any single outer limit tied to the accusation’s seriousness in the way a detention is capped. The undertrial provisions, Section 436A of the Code and Section 479 of the Sanhita, do supply a ceiling of a kind: a person who has undergone detention during investigation, inquiry or trial for up to one-half of the maximum imprisonment specified for the offence is ordinarily to be released on bail, and the Sanhita improves the position of a first-time offender by fixing the fraction at one-third. Offences punishable with death or life imprisonment stand outside that protection, and a court may extend custody beyond the halfway point for recorded reasons. The result is that an undertrial in a grave case, refused bail and awaiting a slow trial, can spend years in custody without any finding of guilt, a duration no preventive detention statute in the country could lawfully impose. That asymmetry is real, and any honest comparison must carry it.

The Detention Side: A Prediction With Safeguards Attached

The detention regime replaces every element of that machinery with a different part doing a different job. The trigger is not an act but an apprehension about acts. A detention statute empowers the government, and by delegation the district magistrate or commissioner of police, to order the detention of a person where the authority is satisfied that it is necessary to do so with a view to preventing that person from acting in a manner prejudicial to one of the statute’s protected interests. Under the National Security Act 1980, Act 65 of 1980, those interests include the defence of India, the relations of India with foreign powers, the security of India, the security of the State, the maintenance of public order and the maintenance of supplies and services essential to the community. The statute book treatment of that Act, its grounds and its record, sits in the guide to the National Security Act 1980; the comparison here needs only its procedural skeleton, which is representative of the family of detention statutes.

The order is the act of an executive satisfaction, and the courts describe that satisfaction as subjective: it is the authority’s own conclusion, drawn from material it has considered, and a reviewing court will not weigh the material as an appellate body would. But subjectivity is not a charter of unreviewability, and the limits on it are the law of this field. The satisfaction must rest on material that exists and is relevant. It must be the satisfaction of the authority named, reached by that authority’s own application of mind and not dictated by a superior or copied mechanically from a police proposal. It must be directed to the interest the statute names, so that material showing at most a problem of law and order cannot sustain an order framed for public order. And the prediction must be about the future: the past incidents relied on must bear a live and proximate link to the apprehended conduct, because a detention founded on stale history is a punishment wearing preventive dress, which the Constitution does not permit.

The documents follow a fixed sequence. The order is made and executed, and the detenu is taken into custody under it. The grounds of detention must then be communicated to the detenu as soon as may be, under the National Security Act ordinarily not later than five days from the date of detention and, in exceptional circumstances recorded in writing, not later than ten days, so that the detenu can exercise the right the Constitution gives him under Article 22(5): the earliest opportunity of making a representation against the order. The representation goes to the government, which must consider it with expedition, and an unexplained delay in considering it is one of the classic grounds on which detention orders are struck down. Separately, the government must place the grounds and any representation before the Advisory Board within three weeks of the detention under Section 10 of the National Security Act, and the Board, a body of three persons who are, have been or are qualified to be High Court judges under Section 9, must report within seven weeks of the date of detention under Section 11 whether in its opinion there is sufficient cause for the detention. If the Board finds no sufficient cause, the government must revoke the order and release the detenu. If it finds sufficient cause, the government may confirm the order and continue the detention for such period as it thinks fit, subject to the statutory ceiling: twelve months from the date of detention under Section 13 of the Act, a ceiling the Constitution’s Article 22(7) permits Parliament to prescribe.

Two features of this skeleton deserve emphasis because they are the ones public discussion most often gets wrong. The Advisory Board is not a court and its proceeding is not a trial. The detenu has no right to appear before it through a legal practitioner, a bar written into Section 11 of the National Security Act, and although the detenu may appear in person and may in some circumstances take the assistance of a friend who is not a legal practitioner, the Board’s task is confined to the sufficiency of the cause on the material placed before it, the grounds, the representation and the government’s file. It does not hear witnesses, return a verdict or pronounce on guilt, and its favourable opinion does not convict anyone of anything. Equally, the confirmation of an order by the government and the Board does not close the courts. The writ of habeas corpus remains available throughout, and it is to the writ, not to the Board, that the serious challenges go.

The custody that follows also differs in its legal character, and the jail administration is required to respect the difference. A convicted prisoner serves a sentence, with the incidents the prison rules attach to punishment. An undertrial is held in custody on a court’s authority pending a process that will decide him. A detenu is held on neither footing: he is a person confined by executive order for prevention, prison rules classify detenus separately from convicts, and the punitive incidents of imprisonment have no application to him, because there is no punishment for them to attach to. The distinction is not a courtesy. It follows from the same preventive character that conditions the power’s validity, and a detenu subjected to the regime of a convict is being punished in fact whatever the order says. In practice the separation is imperfect, detenus and undertrials share the same prisons and often the same wards, and the comparison at the level of lived experience is far closer than the law’s categories suggest. But when conditions are challenged, the category is the argument: the detenu’s complaint is not that imprisonment is harsh but that he is not, in law, a prisoner under sentence at all.

Where the order is made not by a government but by a delegated officer, the statute builds an executive approval chain that the arrest regime has no counterpart to. Under Section 3(4) of the National Security Act, an order made by a district magistrate or commissioner of police must be reported forthwith to the State Government with the grounds and particulars, and it lapses after twelve days unless the government approves it within that period. An order made or approved by a State Government must in turn be reported to the Central Government within seven days under Section 3(5). The chain does not involve any court, and its discipline is bureaucratic rather than judicial, but it is a real step in the sequence: a delegated officer’s order that the government does not approve dies of its own limit, and the reporting duty creates the paper trail on which later challenges rely.

The order also remains the executive’s own throughout its life, which produces two consequences the arrest regime cannot replicate. The government may revoke or modify its order at any time, on a representation, on the Board’s view, or of its own motion, and a custody imposed by a court cannot be ended by the executive in that direct way. And when an order expires or is revoked, Section 14(2) of the National Security Act expressly provides that the expiry or revocation is no bar to a fresh order against the same person. The courts have confined that power within the comparison’s logic: a fresh order must rest on fresh material and fresh satisfaction, not on the grounds of an order a court has quashed recycled under a new date, because to permit the recycling would make the habeas corpus remedy pointless. A fresh order genuinely founded on later conduct stands; an order that re-serves the quashed grounds falls again, and for the same reason.

The State Statutes Where Most Detention Happens

The National Security Act dominates public discussion of preventive detention, but the greater part of the detention actually imposed in India flows from state statutes enacted under Entry 3 of the Concurrent List and aimed at locally defined categories of offender. The pattern is consistent across states: a statute declares categories, the dangerous activities of bootleggers, dacoits, drug offenders, goondas, immoral traffic offenders and land grabbers in the Telangana Act of 1986, the dangerous activities addressed by the Tamil Nadu Act 14 of 1982 directed at bootleggers, drug offenders, goondas and similar classes, the Gujarat Prevention of Anti-Social Activities Act 1985, and the Karnataka statute of 1985 aimed at goondas and allied categories, and empowers the government and delegated district officers to detain persons falling in those categories to prevent them acting prejudicially to public order. These are the orders that fill the habeas corpus lists of the High Courts, and they are the orders on which the circles doctrine does its daily work, because the underlying conduct is typically a dossier of assaults, extortion, illicit liquor or drug peddling and intimidation, directed in the main at identifiable individuals and neighbourhoods.

For the comparison, the state statutes change nothing structural and everything practical. Structurally they replicate the National Security Act’s skeleton: satisfaction, grounds, representation, Advisory Board, a maximum period ordinarily of twelve months. Practically they lower the altitude of the whole regime. The central statute is aimed at the security of India and the State; the state statutes are aimed at the locality, and their detaining authorities are the district magistrate and the commissioner of police acting on police dossiers compiled by the local force. The distance between a law and order problem and a public order claim is shortest exactly here, and the High Courts’ insistence that a dossier of ordinary offences against individuals does not become a public order case by being bound in a file is the operative check. A reader comparing the regimes should therefore hold two pictures at once: the constitutional architecture, which is national and stable, and the working docket, which is local, repetitive and dominated by state goondas legislation.

What is police preventive arrest, and how is it different from preventive detention?

Police preventive arrest is a short power under the procedure code, Section 151 of the Code of Criminal Procedure and Section 170 of the Bharatiya Nagarik Suraksha Sanhita, letting an officer who knows of a design to commit a cognizable offence arrest to prevent it. It lasts twenty-four hours at most without a magistrate’s order and needs no detention statute.

The confusion between these two powers is the most common vocabulary error in the whole field, and it is costly because the remedies differ completely. A person picked up under the preventive arrest power is inside the arrest regime: the twenty-four hour production rule applies, a magistrate takes charge of the custody the next day, and the police must then either proceed with a case or let the person go. A person detained under a detention statute is outside that regime altogether, and the twenty-four hours pass without any production being due. Media reports that describe a detention under the National Security Act as an arrest, and reports that describe a preventive arrest as a detention, mislead families about which clock is running and which forum matters. The words are not interchangeable, and the first task in any real case is to fix which power was in fact used, from the documents.

The Two-Regime Decision Table

Question	Arrest	Preventive detention
What triggers it	A suspected offence already committed	An apprehension of prejudicial conduct in the future
Is an offence required	Yes, a defined offence under a penal law	No offence is alleged or required
Who decides	A police officer arrests, a magistrate owns the custody	The executive: government, district magistrate or commissioner of police
What document is issued	First information report and arrest memo	Detention order, followed by the grounds of detention
What must be supplied to the person	The grounds of arrest, at once	The grounds of detention, ordinarily within five days under the National Security Act
What are the time limits	Production before a magistrate in twenty-four hours; charge sheet in sixty or ninety days	Grounds within days; Advisory Board reference in three weeks; Board report in seven weeks
Who reviews the custody	The magistrate, at every remand	The Advisory Board, and the government on the representation
Where is it challenged	Bail application before the magistrate or the sessions court	Habeas corpus petition under Article 226 or Article 32
Is bail available	Yes, as a right in bailable offences, otherwise in the court’s discretion	No, because there is no case in which bail could be granted
What is the maximum duration	Custody pending trial, with the undertrial ceiling at half the maximum sentence in most cases	Twelve months under the National Security Act and most detention statutes
Effect on a criminal record	A charge and trial; conviction, if any, is recorded	No charge, no trial and no conviction

Law and Order Versus Public Order: The Test That Decides Most Cases

Most detention litigation is not, in the end, about procedure. Procedural failures, the late grounds, the undecided representation, the missing document, account for many quashed orders, and they are dealt with below. But the ground on which detention law is intellectually built, the ground that decides the contested cases on their merits, is a distinction between kinds of disorder drawn by the Supreme Court in 1966 and applied ever since: the distinction between law and order, public order and the security of the State. This article owns that doctrine for the series, because it is the point at which the comparison between the regimes becomes a working tool. The same conduct can be an arrest matter, a detention matter or neither, depending on which circle of disorder it falls in, and the detaining authority’s recurring mistake is to reach for the detention regime while describing facts that belong only in the outermost circle.

Dr. Ram Manohar Lohia v. State of Bihar, AIR 1966 SC 740, arose from the Defence of India Rules, under which the District Magistrate of Patna had ordered Dr. Lohia’s detention to prevent him from acting in a manner prejudicial to the maintenance of law and order. Justice Hidayatullah’s judgment for the Court dismantled the order on a point of language that turned out to be a point of constitutional substance. One has to imagine, he wrote, three concentric circles: law and order represents the largest circle, within which is the next circle representing public order, and the smallest circle represents the security of the State. It follows that an act may affect law and order without affecting public order, just as an act may affect public order without affecting the security of the State. The rules authorised detention for threats to public order, and the magistrate, by framing his order in terms of law and order, had claimed a wider power than the law gave him and had shown that he had not applied his mind to the narrower question the statute asked. The order could not stand. The illustration the judgment gave has become the standard teaching example: when two drunkards quarrel and fight, there is disorder, but not public disorder; the contravention of law always affects order, but before it can be said to affect public order it must affect the community or the public at large.

Does the Advisory Board decide whether a detenu is guilty?

The Advisory Board decides nothing about guilt, because there is no charge before it. It reads the grounds, the representation and the government’s material, and reports one opinion only: whether sufficient cause exists for the detention. A trial court asks the opposite kind of question, about a past act proved by evidence.

Arun Ghosh v. State of West Bengal, (1970) 3 SCR 288, converted the circles into an operational test. Arun Ghosh had been detained under the Preventive Detention Act 1950 on grounds describing a series of assaults and acts of molestation directed at particular individuals, members of one family and two other persons. Hidayatullah CJ, for the Court, held that the question whether a man has committed a breach of law and order or has acted in a manner likely to cause a disturbance of public order is a question of degree and of the extent of the reach of the act upon society. The test the Court stated is the one every detention court since has applied: does the act lead to a disturbance of the even tempo of the life of the community, so as to amount to a disturbance of public order, or does it affect merely an individual, leaving the tranquillity of society untouched? Applying it, the Court found conduct that was reprehensible but directed at individuals, not at women in general or at the locality, and therefore a problem of law and order for the criminal courts rather than of public order for a detaining authority. The detention was set aside and the detenu ordered to be released. The case line that runs with it, Pushkar Mukherjee v. State of West Bengal in 1969 among others, works the same calibration on different facts: rioting with weapons across a locality can disturb the even tempo of public life, while a sequence of assaults on named persons, however ugly, may remain in the outer circle.

The doctrine has never been displaced, and its modern applications show its teeth. In Ameena Begum v. State of Telangana, (2023) 9 SCC 587, the Supreme Court quashed a detention under the Telangana preventive detention statute of 1986, holding that the detaining authority had failed to distinguish law and order from public order, that the grounds were vague, and that the alleged acts were already being dealt with under the ordinary criminal law without any demonstrated need for the extraordinary preventive power. The judgment is a modern restatement of the 1966 discipline: the authority must show, from the material, how the conduct reaches the community, and a recital that the detenu is a habitual offender or a danger to public order, unaccompanied by that showing, will not survive. High Courts deciding habeas corpus petitions under state detention statutes apply the same test term after term, and a substantial share of the orders they strike down fall at exactly this point, not because the detenu’s conduct was innocent but because the authority answered the wrong question.

The circles also organise the detention statutes’ own ground lists, which repay a moment’s attention. The innermost circle, the security of the State and the defence of India, is the province of the National Security Act’s central grounds, directed at espionage, sabotage and threats to the country’s integrity, and detention orders at that altitude are comparatively rare and comparatively rarely disturbed on the characterisation ground. The middle circle, public order, is the contested territory this section has described. The statutes also protect an interest that sits awkwardly inside the circles altogether: the maintenance of supplies and services essential to the community, the ground under which hoarding and black-marketing detentions are passed, including under the Prevention of Blackmarketing and Maintenance of Supplies of Essential Commodities Act 1980. Courts test such orders by a cognate question, whether the conduct in fact threatened the community’s essential supplies or merely violated the control orders in the ordinary way, and the discipline is recognisably the same: the breadth of the power is justified by the community reach of the harm, and orders that cannot show the reach fail. A detaining authority drafting grounds under any of these heads is, whether it knows it or not, locating its case inside a circle, and the court’s first task is to check the location.

This is the namable claim of the comparison, and it deserves to be stated plainly. The concentric-circles test decides most detention cases: the question a court asks is not whether the person is dangerous but whether the alleged conduct disturbed the even tempo of community life rather than merely breaching law and order, and detaining authorities lose because they answer the first question when the statute asks the second. The claim also explains the otherwise puzzling pattern of outcomes to a lay reader. A detenu with a long record of individual offences walks free on habeas corpus while the prosecution of those very offences continues in the criminal court, because the two proceedings ask different questions and the detention failed its own. The arrest regime never faces this test at all. An arrest needs an offence, and the gravity or triviality of the offence’s effect on the community is irrelevant to the arrest’s validity; it matters later, to bail and to sentence. The circles are a detention doctrine because detention, lacking an offence to anchor it, needs a substitute anchor, and the reach of the conduct upon the community is the anchor the courts have supplied.

The Test at Work: Four Situations

Doctrine earns its place in a comparison when it can be run against facts, and the circles test is best learned on situations rather than definitions. Take first a single grave assault: a man attacks a shopkeeper with a knife over a debt, in the shop, before a handful of witnesses, and is arrested the same evening. The criminal process has everything it needs, an offence, a victim, evidence, and the reach of the act is exhausted by the victim and the immediate onlookers. If a detention order follows, framed for public order on that incident alone, it carries the Arun Ghosh defect on its face: the act was directed at an individual, the even tempo of the community’s life was not disturbed, and the order will ordinarily fall. The arrest regime was not merely available here; it was the right regime, and the detention was the wrong instrument chosen for a case the ordinary law completely covers.

Take next a pattern of armed extortion across a market: a group demands weekly payments from every trader in a bazaar, beats those who refuse in public view, and over two months the market begins closing early, traders stop opening on the group’s days, and customers stay away. No single incident exceeds the first situation by much, and individual prosecutions may be pending for several of them. But the reach of the conduct is now the community of the market itself: the even tempo of its life is visibly disturbed, which is the Arun Ghosh formulation operating as a description and not a slogan. A detention order that narrates the pattern, the public setting and the effect on the locality, and connects them to the statutory ground, is the kind of order the circles doctrine sustains. The difference between the two situations is not the wickedness of the men, which may be identical, but the demonstrated reach of the conduct, and a detaining authority that writes its grounds to show the reach wins cases that an authority relying on the record of offences loses.

Take third the professional bootlegger of the state statutes, supplying illicit liquor across a district through a network, with a dossier of excise and assault cases that have repeatedly failed or crawled. The state statutes were designed for this figure, and the authorities’ case for detention writes itself in operational terms: prosecution has been tried and has not interrupted the supply. The court’s inquiry, however, remains the same inquiry, and it is not whether the trade is harmful, which it plainly is, but whether the grounds show the activity disturbing public order rather than merely violating the excise and penal law serially, and whether the satisfaction rests on proximate material or on a stale dossier bound together for the occasion. Bootlegger detentions have been sustained where the material showed the trade’s public effects, adulteration deaths, violence around the supply points, localities held in fear, and quashed where the grounds amounted to a list of pending cases. The category in the statute does not decide the case; the grounds do.

Take fourth the political organiser in a charged season: speeches at meetings, a procession that turned disorderly, cases for unlawful assembly and mischief pending, and a detention order asserting that his activities threaten public order. Here the circles meet the freedoms of speech and assembly, and the courts scrutinise the grounds with the additional care that any restriction on political activity attracts, asking whether the material shows conduct reaching the community’s even tempo or instead shows the ordinary friction of political life that the criminal law and the magistracy’s regulatory powers exist to manage. Orders against political figures have repeatedly failed where the grounds recited speeches and participation in agitation without material of public disturbance, and the pattern teaches the same lesson from the other side: the nearer the alleged conduct sits to the exercise of democratic rights, the more concretely the reach must be shown. In all four situations the machinery is identical and the outcomes differ, which is the demonstration the comparison exists to give. The detention regime is decided less by its forms than by one question of characterisation, and the characterisation is reviewable.

Where the Regimes Meet: A Detention Order Against a Person Already in Custody

The two regimes run in parallel for most purposes, but they meet, and collide, in one recurring situation: the detention order passed against a person who is already inside, held in judicial custody as an undertrial in one or more criminal cases. The pattern is familiar to every criminal lawyer. An accused is arrested, produced, remanded, and applies for bail or is expected to apply. The prosecution apprehends that bail will be granted, or bail is in fact granted and the release is imminent. At that point a detention order arrives, founded on the same incidents or on the person’s history, and although the criminal court’s bail order stands, the jail gate does not open, because the detention supplies an independent legal basis for the custody. To the family this looks like the state simply refusing to obey the bail order. In law it is a distinct order under a distinct regime, and the bail order does not touch it. But the courts have surrounded this manoeuvre with conditions, precisely because it is the point at which preventive detention comes closest to being used as a device for defeating the arrest regime’s own safeguards.

The governing requirements were worked out across a line of decisions and are settled. A detention order against a person in custody is not invalid merely because he is in custody; preventive detention is aimed at future conduct after release, and an authority may legitimately apprehend that a person about to be released will resume prejudicial activity. But the order must show, on its face or in its grounds, three things. The detaining authority must have been aware that the person was already in custody. The authority must have had reason to believe, on cogent material, that the person was likely to be released, whether on bail or otherwise, in the near future. And the authority must have been satisfied, for compelling reasons recorded in the order or the grounds, that detention was nevertheless necessary because the person, once released, was likely to act prejudicially. The Supreme Court stated the discipline in Masood Alam v. Union of India in 1973, insisting that custody already undergone does not extinguish the power but that the authority’s application of mind to the custody must appear. Binod Singh v. District Magistrate, Dhanbad in 1986 applied it to release a detenu where the order betrayed no such awareness, and Kamarunnissa v. Union of India in 1991 confirmed both the validity of the power and the strictness of its conditions. The requirement was restated for the modern period in Union of India v. Paul Manickam in 2003: awareness of the existing custody and a satisfaction grounded in compelling reasons are not courtesies but conditions of validity, and an order silent on them falls.

Can preventive detention run alongside a jail sentence after conviction?

A detention order can be made against a person serving a sentence, because the power looks to what he may do on release, but the authority must show awareness of his position and compelling reasons tied to that expected release. A sentence answers a proved offence; the detention answers a prediction, and neither consumes the other.

The in-custody doctrine is the clearest window into what the comparison really involves, because it shows the two regimes operating as checks on each other. The arrest regime’s bail decision is a judicial assessment that custody pending trial is not necessary. The detention order is an executive override of the practical effect of that assessment, justified by a different question: not whether the person should be confined for the case, but whether his release, whenever it comes, will endanger public order. The courts permit the override but police its honesty, requiring the authority to confront the custody and the impending release squarely rather than pretend the person is at large. An order that recites incidents without mentioning that the person has been in jail throughout, or that asserts he is likely to act prejudicially while omitting any basis for expecting his release, discloses the one vice the doctrine exists to catch: the detention is being used not to prevent future harm but to continue a custody the criminal court has declined to continue. When that is the truth of the matter, the order is quashed, and the person goes out under the bail the criminal court granted.

The mirror-image situation also arises and is worth separating. A detenu whose order is quashed, or whose detention period expires, may be arrested at the gate in a criminal case founded on the same incidents; there is no rule that the detention immunises him from prosecution, just as there is no rule that the prosecution dissolves the detention. Each regime proceeds on its own footing, and a person can lawfully move from one to the other in either direction on the same day. What the law forbids is not the sequence but the substitution: detention used instead of a prosecution the state cannot sustain, or a prosecution kept artificially alive to supply material for detentions. The complaint that preventive detention has become a routine substitute for investigation, examined in the next section, is at bottom a complaint that this boundary is not being held.

The physical experience of the meeting point is worth recording, because it is where families encounter the comparison in its rawest form. A person whose detention is quashed, or whose detention period expires, and against whom a criminal case is pending is not released at the prison gate in any simple sense: the jail authorities hold him under whichever authority survives, and where a production warrant or a remand order in the criminal case exists, he passes from the detenu’s ward to the undertrial’s custody without leaving the prison. Conversely, a person released on bail may be served with a detention order at the gate itself, a practice the courts have not forbidden but which sharpens their scrutiny of the in-custody conditions, since an order executed at the moment of a court-ordered release invites the question whether its satisfaction was genuinely formed or merely timed. The lesson for practitioners is administrative as much as legal: the jail’s custody record, showing under which authority the person was held on each day, is a document of the first importance in both proceedings, because days are counted differently under each regime and a period of detention wrongly continued after an order’s expiry or quashing is itself unlawful custody sounding in compensation.

How Detention Orders Fall: The Grounds of Challenge

An arrest is challenged by contesting the accusation and seeking bail; the merits are the battlefield. A detention order, having no merits in that sense, is challenged by attacking the order’s compliance with the discipline of its own regime, and the grounds that recur in the decided cases form a stable pattern. Stating the pattern is not the same as asserting a rate: no defensible figure exists for the proportion of orders quashed, published counts vary by court, period and statute, and this article asserts none. What can be said with confidence is which defects the courts treat as fatal, because the judgments say so repeatedly.

The first is the wrong circle. Grounds that disclose, at their highest, a problem of law and order cannot support an order framed for public order, for the reasons Lohia and Arun Ghosh establish. The second is the broken link: the incidents relied on are too old, or too remote from the date of the order, to ground a prediction about the future, and the delay between the last incident and the order is left unexplained. Preventive detention is justified by imminence, and an authority that waits months after the last alleged act before deciding that detention is urgently necessary has evidence to give about why the apprehension remained alive; orders have repeatedly fallen where that explanation was absent. The third is the withheld document. The grounds must be accompanied by the documents and material relied on in reaching the satisfaction, because the detenu’s right to make an effective representation under Article 22(5) is empty without them. In M. Ahamedkutty v. Union of India, (1990) 2 SCC 1, the Supreme Court treated the supply of the relied-upon material as an integral part of the constitutional right, and non-supply, or supply so late that the representation is prejudiced, vitiates the detention. The fourth is the undecided representation. The detenu’s representation must be considered by the government with reasonable expedition, independently of the Advisory Board’s timetable, and in K. M. Abdulla Kunhi v. Union of India, (1991) 1 SCC 476, a Constitution Bench systematised the obligation: unexplained delay in disposing of the representation renders the continued detention unconstitutional. The fifth is mechanical satisfaction: grounds that reproduce the police proposal verbatim, orders that recite the statutory formula against material that does not answer to it, and the single stale incident dressed as a pattern all disclose the absence of the authority’s own application of mind.

The mechanical satisfaction defect deserves a final word of its own, because it is the defect most within the authority’s control and the one that best illustrates the comparison’s discipline. A detention order is not a rubber stamp for a police proposal; it is a distinct juridical act by a named authority, and the grounds are that authority’s own statement of why prevention is necessary. Where the grounds reproduce the sponsoring officer’s dossier in its own words, errors included, or recite satisfaction in the statute’s formula against material that visibly belongs to a narrower head, the file discloses that the two acts the law requires, the proposal and the independent satisfaction, collapsed into one. High Courts treat that collapse as fatal not from pedantry but because the satisfaction is the only judicially reviewable substitute the regime offers for the magistrate it withholds: if the satisfaction is not genuinely the authority’s, the detenu has been confined on the word of the police alone, which is the one arrangement Article 22 was written to prevent. The drafting lesson travels back into the comparison. On the arrest side, a careless first information report can be repaired by investigation and tested at trial. On the detention side, the grounds are close to the whole case, and carelessness in them is irreparable.

Two further features of the challenge landscape complete the pattern. The representation has more than one reader, and confusion between them loses cases. The detenu’s representation is addressed to the appropriate government, and it is the government that must consider and decide it with expedition; where the order was made by a delegated officer, the representation is also liable to be placed before the detaining authority for its comments, but the deciding obligation remains the government’s own, and a file that shows the representation travelling in circles without a decision is the Abdulla Kunhi defect in documentary form. Separately, the Advisory Board stage has its own small procedure: the detenu may appear before the Board in person, may seek the Board’s permission to be assisted by a friend who is not a legal practitioner, and may ask to be heard, and although the proceeding is nothing like a trial, a detenu who treats it as a formality wastes the one occasion on which the sufficiency of the cause is examined by a body the executive does not control. The writ petition remains the principal remedy, but the representation and the Board are not formalities on the way to it; they are parallel tracks, each capable of ending the detention, and the competent practice runs all three at once.

Rekha v. State of Tamil Nadu in 2011 gathered several of these threads into the standard modern warning against the routine use of the power. The Court described preventive detention as an exception to the ordinary course of criminal law, to be used sparingly and in genuinely exceptional situations, and held that an order founded on a solitary incident of the kind the ordinary law handles every day could not stand. Read with the in-custody line and the circles cases, the picture is coherent. The detention regime survives constitutional scrutiny because it is narrow: narrow in its triggers, which must reach the community; narrow in its timing, which must be proximate; narrow in its procedure, which must be punctiliously followed; and narrow in its duration, which is capped. Every ground of challenge is an allegation that the order, in the particular case, escaped that narrowness. The comparison with arrest cuts the other way here: custody after arrest is broad, routine and unexceptional, and its safeguards are correspondingly ordinary, remand hearings and bail applications in the daily work of the magistracy. Custody without trial is the anomaly, and the law prices it accordingly.

The Honest Counter-Reading: Where Detention Is Not Simply Worse

The instinct that preventive detention is simply a worse version of arrest is understandable and partly earned. A detenu has no charge to meet, no bail to seek, no trial at which the state must prove anything, and no lawyer before the body that reviews his custody. On every axis of procedural protection the comparison is lopsided, and any treatment that concealed that would be dishonest. But the comparison has other axes, and on two of them it runs the other way, which is precisely why the institution persists and why abolitionist argument that ignores them fails to persuade the audiences it needs.

The first is duration. Preventive detention carries a hard outer limit fixed by statute: twelve months under the National Security Act and under the generality of detention laws, after which the custody ends by operation of law whatever the authority wishes. Undertrial custody has no equivalent certainty. It is bounded by the undertrial ceiling described earlier, half the maximum sentence in most cases, but for grave offences that ceiling is itself measured in years, and where the offence carries life imprisonment the statutory ceiling falls away and custody lasts as long as the trial lasts, bail having been refused. A person accused of a serious offence can spend longer in prison awaiting a verdict that acquits him than a detenu spends under an order that was never tested at all. The second axis is consequence. A detention ends with release and nothing else: no conviction, no sentence, no criminal record of the adjudicated kind, no disqualification that flows from a conviction. An arrest that matures into a trial can end in a conviction that follows the person permanently. Detention is a grave interference with liberty, but it is legally sterile; it proves nothing, records nothing and punishes nothing, and the person who emerges from it is, in the eye of the law, exactly who he was before it.

The comparative angle abroad sharpens the point. A general executive power to imprison without charge in peacetime is rare among constitutional democracies, most of which confine anything resembling it to war or proclaimed emergency, and India’s distinctiveness lies in having normalised it as a standing instrument with constitutional text of its own. That distinctiveness is the product of the founding bargain described earlier and of the conditions of 1950, and it has been questioned in every generation since, including by the members of the Constituent Assembly who lost the argument. The comparison with the arrest regime is where the questioning bites hardest, because the arrest regime demonstrates daily that the State possesses a complete alternative apparatus for dealing with dangerous people, staffed, courted and clocked. The defender of detention must therefore always answer a question the defender of arrest never faces: not whether the person is dangerous, but why the ordinary apparatus, with its magistrates and its bail hearings and its undertrial ceilings, is inadequate to this person in this case. Where that answer exists and is documented, the detention stands. Where it is assumed, the order is an accident of drafting waiting for a habeas corpus petition to reveal it.

These two axes do not defend preventive detention. They explain a behavioural pattern that pure condemnation cannot: why a detaining authority sometimes prefers detention to prosecution, and sometimes prefers prosecution to detention. Detention is preferred where the state’s problem is prospective and its evidence is weak, untestable or dependent on witnesses who will not come forward; it delivers a defined period of incapacitation without the burden of proof. Prosecution is preferred where the state wants the permanent consequences of conviction, or where the conduct needs public condemnation, or where the material will bear a trial. The choice between the regimes is therefore a real institutional choice with real trade-offs, and the courts’ insistence that detention not be used as a substitute for a prosecution the state is merely too idle or too ill-prepared to mount is directed at exactly the abuse that the trade-off invites.

A third family of detention statutes complicates the picture usefully, because it shows the power operating where no question of street disorder arises at all. The Conservation of Foreign Exchange and Prevention of Smuggling Activities Act 1974 authorises detention to prevent smuggling, abetment of smuggling and dealing in smuggled goods, and the concealing or harbouring of smugglers, and its detenus are typically traders and carriers rather than public order offenders. The procedural skeleton is the same Article 22 scheme, grounds, representation, Advisory Board, a twelve month ceiling, and the litigation under it generated some of the leading doctrine on the representation and the relied-upon documents, including Ahamedkutty, a COFEPOSA case. The existence of the economic detention family matters to the comparison because it demonstrates that the detention regime is not a public order instrument with exceptions but a general preventive technology the Constitution makes available across fields, deployed wherever Parliament judges that the arrest regime’s dependence on proof, witnesses and completed acts leaves a gap it wants filled in advance. Whether that judgment is sound in each field is the standing argument; that the technology is one and the same is the structural fact.

That live argument deserves to be put at its strongest on both sides, as the advocates of each position put it. The criticism, pressed by criminal lawyers and civil liberties organisations, is that in several states preventive detention has become a routine instrument of ordinary policing: orders passed on thin and repetitive grounds against alleged habitual offenders, drafted from police dossiers in standard language, used to hold persons the criminal process would release on bail, with the quashing of individual orders by High Courts making little impression on the practice that generates the next hundred. The counter-argument, pressed by police leadership, is operational: in the cases that dominate detention dockets, extortion, organised intimidation, bootlegging and violent local gangs, the victims and witnesses who could sustain a prosecution are the very people the accused has terrorised into silence, trials collapse for want of testimony, and a short, reviewable preventive custody is the only instrument that interrupts the conduct. Both accounts describe something real. The witness problem is genuine and is not answered by repeating that detention is exceptional. The routinisation is also genuine, and it is not answered by pointing to the hardest cases, because the doctrine set out in this article supplies the test that separates them: conduct that disturbs the even tempo of community life, shown on cogent and proximate material, justifies the power, and a dossier of ordinary offences against individuals, recycled from case to case, does not. The verdict the case law has already reached is that the power is constitutional but its routine use is not, and each order stands or falls on whether it can show the narrower justification.

A word belongs here to the consequences that outlast the custody, because the regimes differ there too and the difference is actionable. A person who has been unlawfully detained, or arrested in breach of the safeguards, is not confined to the writ that ends the custody. After the fact, the complaint route to the human rights commissions exists: the Protection of Human Rights Act 1993 arms the National Human Rights Commission and the State Commissions with power to inquire into violations of human rights by public servants, including unlawful detention, and to recommend compensation and proceedings, and a complaint can be the practical vehicle for placing an episode on the official record. The commissions recommend rather than command, and the Act bars inquiry into an incident more than a year old, so the route rewards promptness. It belongs to neither regime exclusively, which is the point: it is the forum for the grievance after both regimes have finished, and a family that has spent its energy entirely on release sometimes misses that the record of what happened can still be pursued.

There is also a remedy that belongs to the arrest side’s history and reaches across to detention: monetary compensation for unlawful custody, awarded by the constitutional courts themselves. In Rudul Sah v. State of Bihar in 1983, the Supreme Court ordered compensation to a man who had remained in prison for years after his acquittal, holding that the writ court’s power to release would be inadequate if it could not also compensate a grave violation of Article 21, and in D. K. Basu the Court confirmed compensation as a recognised remedy for custodial illegality, additional to the ordinary civil claim. Compensation does not distinguish between the regimes: it follows the unlawfulness of the custody, whether the custody began as an arrest that outran its authority or a detention that should never have been ordered. For the comparison it marks the outer boundary of the State’s position in both regimes. Custody is the gravest ordinary interference with liberty the law permits, and when the custody is unlawful the Constitution answers not only with release but with money, because release alone returns the person to a life from which months have been taken.

Can a detenu consult a lawyer about the detention?

A detenu may consult a lawyer, take advice and have the representation and the petition drafted with professional help. The bar in the detention statutes is narrower than commonly believed: it excludes appearance by a legal practitioner before the Advisory Board itself, not consultation for the purpose of challenging the order in court.

The documents in a detention matter are the case. The order, the grounds, the material supplied with the grounds, the dates on which each was served, the representation and the date of its decision are the whole evidentiary basis on which a habeas corpus petition is built, and a claim of delay or non-supply stands or falls on whether those dates can be proved. Families should keep every paper served on the detenu and every acknowledgement of anything filed, in order, from the first day. Where the record survives in spreadsheet and document copies, a reader who needs to open and read them privately, on the reader’s own device and without uploading them to a service, can use the ReportMedic office file viewer for reading documents and spreadsheets in the browser. The habit matters more than the tool: in detention litigation, the side with the complete dated file usually has the argument.

The First Forty-Eight Hours: Doing the Right Things in the Right Regime

Everything in this article converges on a short practical sequence, because the comparison framework exists to resolve a decision and the decision is what to do first. One practical feature distinguishes the regimes at the level of daily life and belongs in the sequence. The undertrial’s custody is punctuated by court dates: productions, remand extensions, bail hearings, each an occasion on which the family sees the person, instructs counsel in the corridor and learns the state of the case. The detenu’s custody has no such rhythm. Visits and interviews are governed by the jail rules applicable to detenus rather than by the calendar of a case, correspondence and meetings with legal advisers are regulated as jail matters, and weeks can pass in which nothing is scheduled to happen unless the family makes it happen, by filing the representation, by pursuing the Board reference, by moving the writ. The passivity that the detention regime induces is one of its practical dangers, and the sequence below is designed against it.

Everything in this article converges on a short practical sequence, because the comparison framework exists to resolve a decision and the decision is what to do first. The sequence begins with identification. Establish from the documents which regime holds the person: a first information report and an arrest memo, or a detention order. That single determination fixes the forum, the remedy and the clock, and no useful step can be taken before it. If the papers are not forthcoming, the demand for them is itself the first step in either regime, because the arrest side owes the grounds of arrest at once and the detention side owes the grounds of detention within days, and unexplained withholding of the grounds is the first defect a court will notice.

If the regime is arrest, the sequence runs through the magistrate’s court. Verify that production happened within twenty-four hours. Attend the remand hearing, contest police custody, press for judicial custody or bail, and note the date from which the sixty or ninety day charge sheet period runs, because the default bail entitlement that arises if the period expires is among the most valuable rights in the entire comparison and is lost if not claimed in time. If the regime is detention, the sequence runs elsewhere. Calendar the grounds deadline and obtain the grounds and the relied-upon documents the moment they are served. Draft the representation at once and file it with proof of the date, since the government’s expedition in deciding it is a ground of challenge that exists only if the filing date is proved. Do not wait for the Advisory Board: the writ petition under Article 226 of the Constitution before the High Court, or Article 32 before the Supreme Court, can be filed as soon as the order and grounds exist, and the grounds on which orders actually fall, the wrong circle, the stale incident, the missing document, the undecided representation, are all available from the first week. The Board’s proceeding runs in parallel and its opinion matters, but treating it as a stage that must be exhausted before approaching a court is an error that costs the detenu the period in which habeas corpus is fastest.

After the first forty-eight hours the detention calendar continues, and the family should hold it as firmly as the undertrial family holds court dates. The grounds arrive, ordinarily within five days under the National Security Act. The representation should already be moving. The government’s reference to the Advisory Board must be made within three weeks of the detention, and the Board’s report is due within seven weeks of the date of detention; a reference made late is a defect of the same family as late grounds. If the Board reports sufficient cause, the government confirms the order and fixes the period of detention within the statutory ceiling, and the confirmed order then runs to its expiry unless revoked earlier or quashed. Revocation is not a hope but a standing possibility throughout: the government may revoke on a fresh representation showing changed circumstances, and competent practice treats the confirmation as the start of a second phase of representation rather than the end of the matter. The undertrial calendar, by contrast, is a sequence of hearings at which things are decided in the open. The detention calendar is a sequence of dates on which things are decided in files, and the family’s task across all seven weeks is to make each date generate a paper record, because the record is what the writ court will read.

Four pieces of circulating advice should be corrected plainly, because each is confidently given and each is wrong. The first is that the family should file for bail in a detention matter. There is no bail in preventive detention, and an application framed as bail invites the response that no case exists in which bail could be granted; the correct application is the writ petition seeking the quashing of the order. The second is that anticipatory bail can be obtained against an apprehended detention. Anticipatory bail is a direction operative on arrest for an offence; a detention order is not an arrest for an offence, and the provision has no purchase on it. A person who apprehends a detention order has remedies after the order is made, and in an exceptional case a challenge to a threatened exercise of power may be attempted, but the anticipatory bail application is not the vehicle and time spent drafting one is time lost. The third is that the family should wait for the charge sheet, which confuses the regimes entirely: detention has no charge sheet, and waiting for one means waiting through the very weeks in which the representation and the writ do their work. The fourth is that a detention is automatically illegal because no offence is alleged. The absence of an offence is the design of the regime, not a defect in it; the illegalities that release detenus are the specific ones catalogued above, and a petition that leads with the automatic-illegality argument signals to the court that the real grounds have not been found.

What follows is general information about how the two regimes work, not advice on any individual case: anyone facing an arrest or a detention should take advice on their own facts from a legal practitioner without delay, because the limitation periods, the remand dates and the representation timelines in this field are short and unforgiving, and several of the rights described here are lost by the simple passage of days. That sentence is not a formality in this subject. Default bail unclaimed before the charge sheet is filed is gone. A representation filed late starts the government’s clock late. A human rights complaint filed after a year is barred. The comparison’s practical content is a set of clocks, and clocks are the one thing that cannot be recovered by a better argument later.

The Verdict: One Question, Two Answers

The comparison resolves into a single diagnostic question: is the State answering something the person has done, or preventing something it fears he will do? If the first, the custody belongs to the criminal process, with its offence, its magistrate, its remands, its charge sheet deadline, its bail and its trial, and every safeguard of Article 22(1) and (2) applies in full. If the second, the custody belongs to preventive detention, with its executive satisfaction, its grounds and representation, its Advisory Board and its statutory ceiling, and the protections of the first two clauses are switched off by Article 22(3) and replaced by the substitute scheme of clauses (4) to (7). The regimes are not stronger and weaker versions of one power. They are different powers with different legitimating ideas: the arrest is justified by an accusation the state must prove, the detention by a prediction the state must keep within narrow constitutional channels.

The recommendation differs by reader, because each reader meets the comparison at a different point. The detaining authority should treat the arrest regime as the default and detention as the instrument of a demonstrated gap: where prosecution is genuinely unavailable or genuinely inadequate to interrupt conduct that reaches the community, and it should write its grounds for the reader it will actually face, a habeas corpus bench applying Lohia. The advocate should read every detention order against the five recurring defects before reading it for its allegations, because orders fall on their discipline more often than on their facts. The teacher has in this comparison the cleanest available illustration of how the Constitution prices liberty: the same physical custody, two entirely different justifications, and a clause that switches the safeguards between them. The aspirant should note that the distinction is examined from both directions, as doctrine, through the circles and the even tempo test, and as procedure, through the grounds, the Board and the representation, and that an answer which covers only one direction is half an answer. The family, finally, should hold the operational rule: identify the regime from the documents on the first day, because every later step is regime-specific and the clocks do not wait for the identification to catch up.

The general verdict has edges, and a comparison that concealed them would fail its own standard. The diagnostic question assumes the peacetime operation of the Constitution. Between 1975 and 1977, during the proclaimed Emergency, the enforcement of the detenu’s Article 22 protections was itself suspended by presidential order, and the Supreme Court’s majority in the ADM Jabalpur case upheld the consequence that a detenu could not move any court for release; the comparison in those years described the text but not the reality, and the Forty-fourth Amendment’s protection of Articles 20 and 21 against suspension was the system’s answer, leaving Article 22 itself outside the protected pair, a gap critics of the settlement continue to name. The verdict also assumes the standard statutory families. A foreign national detained with a view to regulating his continued presence or arranging his expulsion, a ground the National Security Act itself carries, is held in a detention whose object is removal rather than prevention of disorder, and the circles have little to say to it. And the verdict assumes a detaining authority acting within a functioning writ jurisdiction: every doctrine in this article is, in the end, a doctrine about what a High Court will do with an order, and its practical force rises and falls with the speed and independence of that forum.

The deciding factor in contested cases, the factor a reader should carry away, is the reach of the alleged conduct upon the community. The concentric circles of Lohia, outer law and order, middle public order, inner security of the State, and the even tempo test of Arun Ghosh supply the measure by which a detention is judged, and they explain why dangerousness alone has never been enough. An advocate reading a fresh detention order should go first to the grounds and ask which circle the material in fact occupies. An aspirant should hold the doctrine as the examiners hold it, as the single most tested distinction in the preventive detention field. A family should take from the comparison the operational lesson with which this article began: name the regime from the documents, then use that regime’s own remedies, in that regime’s own forum, on that regime’s own clock. Almost every expensive error in this field is a failure of that first identification.

Frequently Asked Questions

Q: What is the difference between preventive detention and an ordinary arrest?

An ordinary arrest begins a prosecution. It rests on a suspected offence already committed, is recorded in a first information report, and leads to production before a magistrate within twenty-four hours, remand, a charge sheet and a trial at which the State must prove the accusation. Preventive detention begins nothing. It rests on an executive authority’s satisfaction that a person is likely to act prejudicially in the future, alleges no offence, involves no magistrate and leads to no trial. Its safeguards are different in kind: communication of grounds, a representation to the government, review by an Advisory Board, a fixed maximum period, and challenge by habeas corpus rather than by bail.

Q: Is an FIR needed before a detention order?

No. A first information report is the document that sets the criminal process in motion by recording a cognizable offence, and preventive detention is not part of that process. A detention order needs a detaining authority, a statute that empowers it, and material from which the authority draws its satisfaction that the person must be confined to prevent prejudicial conduct. The grounds served on the detenu often narrate past incidents, and some of those incidents may also figure in first information reports, but the order neither depends on a report nor creates a case. The absence of a first information report is therefore not a defect in a detention order, and its presence is not what validates one.

Q: Is bail available against a preventive detention order?

No, and the reason is structural rather than harsh. Bail is a mechanism inside a criminal case: a court that is seized of an accusation releases the accused on conditions while the case proceeds. A preventive detention has no case, no accusation, no charge and no court seized of anything, so there is no proceeding within which a bail order could operate. Anticipatory bail is equally unavailable, because it is a direction that operates on an arrest for an offence, and a detention order is not an arrest for an offence. The detenu’s route to release is different: persuade the government through the representation, obtain a favourable Advisory Board opinion, or have the order quashed on a habeas corpus petition.

Q: What is the difference between law and order and public order?

The Supreme Court explained the distinction in Dr. Ram Manohar Lohia v. State of Bihar in 1966 through three concentric circles: law and order is the largest circle, public order sits inside it, and the security of the State is the smallest. Every breach of law affects law and order, but it affects public order only when its reach extends to the community at large. Arun Ghosh v. State of West Bengal in 1970 supplied the working test: an act disturbs public order when it disturbs the even tempo of the life of the community, while an act directed at individuals, however serious, remains a law and order problem for the criminal courts alone.

Q: Can you be preventively detained without committing any offence?

Yes. Preventive detention is aimed at conduct that has not yet occurred, so the committing of an offence is not a condition of the power. The detaining authority acts on material, usually past conduct and antecedents, from which it predicts that the person will act prejudicially to public order, the security of the State or another protected interest. That prediction is the whole basis of the custody. The power is nonetheless fenced: the material must exist and be relevant, the past conduct must bear a live and proximate link to the feared conduct, and the satisfaction must be the authority’s own. A detention built on fantasy, on irrelevant material or on incidents too old to predict anything is liable to be quashed.

Q: Which is easier to challenge, a detention order or an arrest?

They are challenged in different currencies, so the comparison is not exact. An arrest is usually not attacked head-on; the accused seeks bail, contests remand and ultimately contests the charge at trial, where the prosecution carries the burden of proof. A detention order is attacked head-on by habeas corpus, and the attacker has an unusually rich set of technical grounds: late or missing grounds, non-supply of relied-upon documents, an undecided representation, stale incidents, a mechanical satisfaction, and grounds that disclose only a law and order problem. Detention orders are quashed on such defects with a regularity that surprises lay readers, because the courts construe a power to imprison without trial strictly. The merits, however, are never open: no court will reweigh the authority’s prediction as an appeal.

Q: Does preventive detention show up as a criminal record?

Not in the sense that matters legally. A criminal record in the strict sense is the record of charges tried and convictions returned, and preventive detention involves neither: there is no charge, no trial, no finding and no conviction, and the detention period cannot be counted as a sentence. The detenu emerges with his legal status unchanged. In a practical sense the episode is not invisible. The detention is recorded in police and prison records, the incidents narrated in the grounds usually correspond to cases or dossier entries that exist independently, and a detaining authority may place the earlier detention itself before itself when considering a later order. But nothing in a detention amounts to an adjudication of guilt.

Q: Why do police use preventive detention instead of arrest?

The honest answer has two parts. The legitimate reason is that arrest and prosecution depend on evidence that can be produced in court, and in cases of organised intimidation, extortion or gang activity the witnesses are often the very people too frightened to testify, so cases collapse while the conduct continues. Detention interrupts the conduct for a fixed period without needing a witness to survive cross-examination. The illegitimate reason, which courts repeatedly identify, is convenience: detention avoids the disciplines of investigation, the charge sheet deadline and the bail hearing, and can be used to hold a person the criminal process would release. The first reason explains the power’s existence. The second explains why courts police its use so strictly.

Q: How can a family tell whether a person has been arrested or preventively detained?

Read the documents, because the taking itself looks identical under both regimes. An arrest produces a first information report naming an offence, an arrest memo, and production before a magistrate within twenty-four hours. A preventive detention produces a written detention order signed by a district magistrate, a commissioner of police or a government, reciting the authority’s satisfaction that detention is necessary to prevent prejudicial conduct, followed within days by the grounds of detention. If there is a case number and a named offence, it is an arrest. If there is an order speaking of satisfaction and prevention, it is a detention. The answer decides the forum, the remedy and the clock, so it must be settled first.

Q: Can an undertrial spend longer in custody than a person who is preventively detained?

Yes, and in serious cases this happens as a matter of routine. Preventive detention is capped by statute, ordinarily at twelve months, after which the custody ends by operation of law. Undertrial custody after an arrest has no equivalent fixed cap. The undertrial provisions release a person who has spent half the maximum sentence for the offence in custody, and the Sanhita improves the first-time offender’s fraction to one-third, but offences carrying life imprisonment fall outside that protection and custody then lasts as long as the trial lasts where bail is refused. An accused in a grave case can therefore spend years in prison awaiting a verdict, longer than any detenu may lawfully be held.

Q: Why is there no charge sheet in preventive detention?

A charge sheet is the closing document of an investigation into an offence: it names the accused, the offence and the evidence the prosecution will lead at trial. Preventive detention contains no investigation into an offence, no accused and no trial, so there is nothing for a charge sheet to close. The corresponding document is the grounds of detention, which performs a different function. Instead of framing a charge to be proved, the grounds disclose the material from which the detaining authority drew its prediction, so that the detenu can answer the prediction by representation. Families who wait for a charge sheet in a detention matter are waiting for a document the law never requires, and lose the weeks in which the representation and the writ petition do their work.

Q: What happens to a criminal case when the accused is also preventively detained?

The criminal case continues on its own track, unaffected in law by the detention. The two proceedings are independent: the detention is not a punishment for the case offences, the case is not evidence-proof of the detention’s grounds, and neither outcome controls the other. In practice the detention changes the accused’s position inside the case. He remains in custody even if the criminal court grants bail, because the detention supplies a separate basis for confinement, and his defence must be conducted from jail. The trial itself proceeds, and an acquittal does not automatically dissolve the detention, just as a conviction is not needed to sustain it. Each regime must be fought with its own remedies.

Q: Why does default bail exist after an arrest but not in preventive detention?

Default bail is the criminal process fining itself for lateness. The code gives the investigation sixty or ninety days, depending on the offence, to produce a charge sheet, and converts custody beyond that point into a right to release because the State has failed to complete the accusation it arrested the person for. Preventive detention has no accusation to complete and no investigation running against a deadline, so the trigger for default bail never arises. Its functional substitutes are different in kind: the statutory maximum period, which ends the detention outright, and the procedural deadlines for grounds, Board reference and decision of the representation, breach of which entitles the detenu to have the order quashed rather than to be bailed.

Q: Can a detention order be challenged in court before the Advisory Board gives its opinion?

Yes, and in most serious cases it should be. The Advisory Board reference is an internal safeguard that runs on the government’s timetable; it is not a stage the detenu must exhaust before approaching a court. A habeas corpus petition under Article 226 before the High Court, or Article 32 before the Supreme Court, can be filed as soon as the order has been made and the grounds served, and the grounds that actually release detenus, the wrong circle of disorder, the stale incident, the missing document, the undecided representation, are all examinable from the first week. Waiting for the Board wastes the period in which the writ is fastest, and nothing in the Board’s eventual opinion bars the court from deciding the petition.

Q: Can a detention order rest on a single incident?

Yes, in principle, but the incident has to carry the whole weight of the prediction, and most single incidents cannot. Number is not the test; reach and propensity are. A single act of exceptional gravity with an unmistakable public dimension, an act of sabotage, a killing designed to terrorise a locality, an armed attack that empties a market, can disclose both the man’s propensity and the community impact that public order requires, and courts have sustained orders founded on one such act. A single ordinary offence, a solitary assault or theft of the kind the criminal courts process daily, discloses neither, and in Rekha v. State of Tamil Nadu the Supreme Court quashed an order built on a lone incident of that character, holding that preventive detention is an exception to the ordinary course of criminal law and cannot be invoked where that law is adequate to the act.

Q: What is the difference between preventive detention and remand to police custody?

Remand to police custody is a stage inside an arrest. After production before a magistrate, the court may authorise the police to hold the accused for short periods at the start of the investigation so that he can be interrogated and recoveries effected, after which he passes to judicial custody in jail. Every remand is a judicial order, time-limited and contestable at the hearing. Preventive detention involves no magistrate at any stage: the custody is authorised from beginning to end by the executive order and the statute behind it. Police custody on remand looks forward to a charge sheet within the remand framework of the case. Detention looks forward only to its own expiry, confirmation and review dates.

Q: Can a person be preventively detained for an act for which he was already prosecuted?

In principle yes, because detention is not a second prosecution and the constitutional bar on double jeopardy, which protects a person from being prosecuted and punished twice for the same offence, is not engaged by a preventive measure. Courts have accepted that the same incident can found a prosecution and figure in the grounds of a detention. But the prosecution’s history matters to the detention’s validity. Where the person was acquitted on the merits, or discharged because the material was found worthless, a detaining authority that recycles the identical material without anything more invites the conclusion that its satisfaction is mechanical or mala fide. The authority must apply its mind to what the criminal process did with the incident, and orders that ignore an acquittal entirely are vulnerable.

Q: What is the subjective satisfaction of the detaining authority, and can a court examine it?

The satisfaction is the detaining authority’s own conclusion, recorded in the order, that detention is necessary to prevent the person acting prejudicially. It is called subjective because the court will not substitute its own view of the material or ask whether it would have been satisfied itself. It is not, however, beyond examination. A court can ask whether the authority applied its mind at all, whether the material existed and was relevant to the statutory purpose, whether the grounds are vague or stale, whether the authority confused law and order with public order, and whether the satisfaction was dictated by someone else or reached in bad faith. On any of those findings the order falls, which is why habeas corpus in detention cases is a real remedy and not a formality.

Q: What should a family do in the first forty-eight hours after a detention order?

First, obtain the order and fix the regime from its face. Second, note the date and time of detention, because every deadline in the regime runs from it. Third, press for the grounds of detention and the documents relied on, which are ordinarily due within five days under the National Security Act and sooner under the Constitution’s requirement of communication as soon as may be. Fourth, instruct a lawyer to draft the representation and the habeas corpus petition in parallel rather than in sequence. Fifth, file the representation with proof of its date, since the government’s speed in deciding it is itself a ground of challenge. Nothing in the first two days should be spent on bail applications or on waiting for a charge sheet, because neither exists in this regime.

Q: What papers should a family preserve when someone is preventively detained?

Preserve everything, in date order, from the first day. The detention order itself, the grounds of detention exactly as served, and every document supplied with the grounds, because non-supply and late supply are grounds of challenge that can only be proved from the file. The acknowledgement or proof of filing of every representation, with its date, because unreasonable delay in deciding a representation vitiates the detention and the delay is measured from that proof. The Advisory Board’s communications, the government’s confirmation order, and every order extending or revoking the detention. A habeas corpus petition in this field is built almost entirely out of dates and documents; the family that keeps a complete file gives its lawyer the argument, and the family that does not must reconstruct one under pressure.