On the bare skeleton of the Prevention of Terrorism Act, 2002, the two sides of India’s longest running argument about security law agree completely. Both accept that POTA was passed at a joint sitting of Parliament, a device the Constitution reserves for deadlock between the two Houses and one that had been used only twice before in the history of the Republic. Both accept that the Act governed India’s response to terrorism for less than three years, that allegations of misuse gathered around it from its first months, that a change of government in 2004 brought its repeal, and that on the very day the repeal took effect the core of the statute was written into another Act, where its substance remains. The disagreement begins only when these agreed facts are characterised. To one side, POTA was a necessary law destroyed by politics, enacted in the shadow of an attack on Parliament itself and repealed as an act of appeasement that weakened the state’s defences. To the other side, it was a draconian law whose repeal was overdue, a statute that armed the police against the citizen and was predictably turned against minorities, political opponents and the poor. Each characterisation cannot survive contact with the agreed skeleton, and the reason is the same in both cases. A repeal that re-enacts the repealed law’s substance on the same day is not the weakening of a state’s defences, and a law whose substance the repealing government chose to preserve was not, in that government’s own judgment, intolerable in substance either. What the record shows is narrower and stranger than either slogan. Call it the joint sitting that changed nothing: the most extraordinary legislative procedure the Constitution offers was deployed to enact a statute whose penal core outlived its own repeal without interruption, so that the fiercest parliamentary fight in the history of Indian security law settled almost nothing about the power the state actually holds.

This account follows the order the controversy requires. The passage record is set out first, because the joint sitting is the fact about POTA that every other account assumes and few explain. The statute is then described as a text, in the terms in which Parliament enacted it, so that the arguments can be tested against provisions rather than reputations. The case for the Act is stated at its strongest, in the voice of the government that piloted it, and the case against it at its strongest, in the voices of the parliamentarians, litigants and official bodies that opposed it. The misuse record is then separated into what the Review Committees and the courts established and what remains allegation, before the repeal and the same day transplant are examined together, which is the only honest way to examine either. A verdict follows on what the record settles, what it leaves factually contested, and what it cannot settle because the underlying dispute is about values rather than facts.

The Winter in Which the Law Was Made

POTA was a product of a single season. On 11 September 2001, hijacked aircraft were flown into buildings in the United States, and on 28 September 2001 the United Nations Security Council adopted Resolution 1373, which called on member states to legislate against terrorism, against its financing and against the networks that sustain it. The resolution was adopted under the Council’s enforcement powers and was not hortatory: it decided that states should criminalise the financing of terrorism, freeze the assets of those who commit or facilitate terrorist acts, deny safe haven and support to terrorist groups, and bring their domestic law into conformity with those duties. A government minded to enact a special statute in the autumn of 2001 could therefore present the project not as a domestic preference but as the discharge of an international obligation binding on India, and the Vajpayee government presented it in exactly those terms, in the Statement of Objects and Reasons and on the floor of the House. The presentation mattered because it relocated the argument. Opposition to the bill’s provisions could be characterised, and was characterised, as indifference to an obligation the whole organised world had assumed, while the opposition’s answer, that Resolution 1373 required legislation against terrorism but nowhere prescribed confessions to police officers or six month charge sheet periods, identified the gap between the obligation and the use the government made of it. That gap, between what the moment required and what the statute contained, is where the POTA controversy lived from its first day. The government of Prime Minister Atal Bihari Vajpayee, a National Democratic Alliance coalition in which the Bharatiya Janata Party was the largest party, had been considering an anti-terrorism statute for longer than that autumn. A Prevention of Terrorism Bill had been drafted in 2000 and had drawn a negative opinion from the National Human Rights Commission. The demand within government for a successor to TADA had been building for some years before that autumn. The hijacking of an Indian Airlines aircraft to Kandahar in December 1999, ended by the release of imprisoned militants, had exposed to public view the state’s dependence on ordinary process against networks that operated across borders, and the Prevention of Terrorism Bill of 2000 was the first fruit of that exposure. The Terrorist and Disruptive Activities (Prevention) Act, known as TADA, which had carried India’s special anti-terror procedure through the Punjab insurgency and the Bombay bombings, had been allowed to lapse in 1995 amid the controversy its own working had generated, and for six years the Union had operated without a central special statute, relying on the Indian Penal Code, the ordinary criminal procedure, state legislation such as the Maharashtra Control of Organised Crime Act of 1999, and the Unlawful Activities (Prevention) Act of 1967, a permanent statute concerned principally with unlawful associations. The longer sequence of these instruments, and the way each generation of Indian anti-terror law inherited the drafting of the one before, is traced in the account of the lineage of India’s anti-terror laws. What matters for the passage of POTA is that the government entered the autumn of 2001 believing the gap was real, and that events then gave that belief a force no drafting committee could have supplied.

On 13 December 2001, five gunmen attacked Parliament House in New Delhi while Parliament was in session, and nine people were killed before the attackers themselves were shot dead. The investigation attributed the attack to the Jaish-e-Mohammed, and the two countries on either side of the Kashmir dispute mobilised their armies along the border in the standoff that followed. The Statement of Objects and Reasons attached to the eventual Act recorded the government’s framing of the moment in its own words: an upsurge of terrorist activities, an intensification of cross-border terrorism, insurgent groups in different parts of the country, and an existing criminal justice system that was, in the government’s submission, not designed to deal with the types of heinous crimes the proposed law addressed. That framing was contested at every point by the opposition parties, who did not dispute the gravity of the Parliament attack but disputed the conclusion drawn from it, arguing that the ordinary law had proved sufficient to investigate the attack itself and that a special statute on the TADA model would repeat a documented failure. The contest between those two readings of the same winter runs through everything that follows, and it is worth noticing at the outset that the disagreement was never about whether terrorism was a grave threat. It was about what kind of law, if any, a grave threat justifies, and that is a disagreement about institutional design rather than about the facts of December 2001.

Two Ordinances Before a Single Vote

The statute began as executive legislation, twice. Parliament was not in session in October 2001, and on 24 October 2001 the President promulgated the Prevention of Terrorism Ordinance, 2001 under Article 123 of the Constitution, bringing the special regime into force without a vote in either House. The government’s justification, recorded in the Statement of Objects and Reasons, was that circumstances existed which rendered immediate legislation necessary. The opposition’s answer, made at the time and repeated in the debates of the following spring, was that Parliament was due to reconvene within weeks for the winter session and that a measure of this consequence had been deliberately routed around the scrutiny a bill would have faced, including scrutiny by the National Human Rights Commission, which had not been consulted on the ordinance. Both characterisations of the same act were available then and remain available: an urgent response to an emergency the calendar could not accommodate, or a decision to govern a contentious subject by decree for as long as the Constitution’s six week and six month limits on ordinances allowed.

The ordinance route then failed on its own terms. The winter session of Parliament met, but it adjourned sine die on 19 December 2001 with the replacing bill not introduced, the session having been consumed in large part by the aftermath of the Parliament attack six days earlier. An ordinance lapses six weeks after Parliament reassembles unless it is replaced by an Act, and the life of the October ordinance was running out. On 30 December 2001 the President promulgated the Prevention of Terrorism (Second) Ordinance, 2001, giving the regime continuity. The second ordinance was not a mere reissue. The government used the interval to modify its own scheme in response to criticism: the proposed life of the legislation was shortened from five years to three, the Schedule of terrorist organisations grew from twenty-three entries to twenty-five as two further organisations were declared, and the power to forfeit the proceeds of terrorism was moved from a designated executive authority to the Special Court. The modifications matter beyond their detail. They show a government bargaining with its critics through the text of a decree, conceding duration and judicialising forfeiture while conceding nothing on the provisions that carried the real controversy, the confessions, the custody periods and the bail standard, and they show the opposition’s pressure registering in the statute before Parliament had voted on it at all.

How long was POTA meant to last?

Three years. Section 1(6) deemed the Act in force from 24 October 2001, the date of the first ordinance, and limited its life to three years from that commencement, so the statute would have expired in October 2004 by its own terms without any repeal. The first ordinance proposed five years; the second reduced the period to three.

The ordinance route also carried a discipline of its own, and the government’s handling of it shows how completely the emergency method governed the statute’s birth. An ordinance under Article 123 must be laid before Parliament when it reassembles and ceases to operate six weeks after reassembly unless a replacing Act is passed, and it cannot survive beyond a maximum of six months and six weeks in any event. The October ordinance therefore confronted the government with a fixed parliamentary deadline from the day it was issued, and the failure of the winter session to take up the bill was not a postponement but a defeat that only a second ordinance could disguise. Re-promulgation, the device by which an executive renews its own decree when the legislature has not endorsed it, had been held by the Supreme Court in earlier decades to be a constitutionally suspect practice when used to bypass the legislature as a matter of routine, and the opposition in 2001 charged the government with exactly that design. The government’s answer rested on the calendar and the crisis: the session had been consumed by the Parliament attack and its aftermath, the ordinance would have lapsed in the middle of a military standoff, and continuity of the security regime was a necessity the Constitution’s own time limits had created. The episode matters to the passage story because it meant that by the time Parliament voted in March 2002, the regime the bill proposed had already been the law for five months. Members were not deciding whether India should have a special anti-terror code. They were deciding whether the code India already had, under which arrests had been made and investigations opened, should continue, lapse, or be re-founded on a parliamentary vote, and that framing, in which the status quo wore the uniform of the emergency, was the government’s first and quietest victory in the passage of POTA.

The three year limit deserves emphasis because it shaped the politics of everything after. A permanent statute invites a fight about the permanent distribution of power between citizen and state. A statute with a three year life invites a different calculation on both sides: the government needed only to hold its majority together once, and the opposition needed only to wait, or to win one general election, for the law to die. POTA’s repeal in September 2004, a month before the statute would have expired by its own sunset, is unintelligible without this clause. The repeal was, among other things, a race to claim credit for a death that the calendar had already scheduled, and the sunset clause is the first piece of evidence for this article’s central claim, that the drama of POTA’s passage and the drama of its repeal both exceed, by a wide margin, the legal change either event produced.

The Passage: A Rejection, and the Third Joint Sitting

The Prevention of Terrorism Bill, 2002 reached the Lok Sabha in the Budget Session, and on 18 March 2002 the House passed it after a combined discussion in which a statutory resolution disapproving the second ordinance was negatived and the bill, moved by the Home Minister, L. K. Advani, was taken into consideration and passed. The debate of 18 March had itself been a combined proceeding of a kind that signalled the government’s difficulty: the statutory resolution disapproving the second ordinance, moved by opposition members, was discussed together with the bill, so that the House voted in one sequence on whether the executive had been entitled to legislate at all and on whether the legislation should be enacted. The resolution was negatived, the bill was passed, and the government had its first House. The government’s parliamentary position was comfortable in the lower House and precarious in the upper. The National Democratic Alliance did not command the Rajya Sabha, where the Congress party and its allies held the larger strength, and on 21 March 2002 the Rajya Sabha negatived the motion that the bill be taken into consideration, defeating it by 113 votes to 98. The message the Rajya Sabha sent to the Lok Sabha the following day recorded the bare constitutional fact: the bill, passed by one House, had not been agreed to by the other. In the ordinary course, a bill so rejected is dead. The sequence through which an ordinary bill travels, and the points at which disagreement between the Houses can end it, are set out in the general account of how a bill becomes law in India. For this bill, the government chose not to accept the ordinary course, and the Constitution supplied the exception it needed.

Article 108 of the Constitution provides that where a bill passed by one House is rejected by the other, or where the Houses finally disagree on amendments, or where six months pass without the other House acting, the President may summon both Houses to meet in a joint sitting to deliberate and vote on the bill together. At a joint sitting the bill is decided by a majority of the members of both Houses present and voting, which means that the much larger Lok Sabha, with its 543 elected members against the Rajya Sabha’s 245, normally determines the result. Two categories of bill are excluded from the device altogether. A Money Bill cannot go to a joint sitting, because the Rajya Sabha’s powers over Money Bills are already limited to recommendation. A Constitution Amendment Bill cannot go to a joint sitting either, because Article 368 requires each House to pass it separately by a special majority, and the amending power therefore cannot be exercised by the combined arithmetic of a joint sitting, a point developed in the account of the Constitution amendment process in India. The exclusion of amendment bills is what makes Article 108 a mechanism for ordinary legislation only, and it is why the most consequential law of that parliamentary session, a security statute, could be rescued by a device unavailable for constitutional change.

President K. R. Narayanan summoned the joint sitting, and it met on 26 March 2002. It was only the third joint sitting in the history of the Republic. The first, in 1961, had concerned the Dowry Prohibition Bill; the second, in 1978, the Banking Service Commission (Repeal) Bill. Neither predecessor had carried anything like the public charge of the third, and no joint sitting has been summoned since, so the POTA sitting stands in the parliamentary record as both the most consequential use of Article 108 and the measure of how rarely the device is used at all. The sitting also produced the detail on which a striking number of secondary accounts are wrong. The Speaker of the Lok Sabha, G. M. C. Balayogi, had died on 3 March 2002 when his helicopter crashed in Andhra Pradesh, and the office of Speaker stood vacant on the day of the sitting; Manohar Joshi was elected Speaker only in May 2002. Under the order of precedence that Article 118 and the rules of procedure prescribe, the joint sitting was therefore presided over not by a Speaker but by the Deputy Speaker of the Lok Sabha, P. M. Sayeed, and it was Sayeed who declared the result. Accounts that name a Speaker as presiding officer of the POTA sitting are naming an office that was empty on the day, and the correction matters to the aspirants and journalists for whom the sitting is a standard examination and reference point.

The vote, when it came, was 425 in favour and 296 against, a margin of 129 in a combined House. The arithmetic deserves a moment’s attention, because it explains what the joint sitting did and did not decide. The Rajya Sabha had rejected the bill by fifteen votes. The joint sitting reversed that rejection not by persuading the upper House but by submerging it: the Lok Sabha’s surplus over the Rajya Sabha’s strength was large enough that the government’s lower House majority, even with defections, absorbed the upper House’s entire opposition and produced a comfortable combined majority. Article 108 is often described as a deadlock breaking mechanism, and so it is, but the POTA sitting shows the precise sense in which it breaks deadlocks. It does not reconcile the two Houses. It counts them together, in proportions the Constitution fixed, and the proportions decide. Whether that is a legitimate design is a genuine constitutional question, and it was argued in those terms at the time. The government’s answer was that Article 108 is the Constitution’s own provision, that the directly elected House is meant to prevail on ordinary legislation when the two Houses irreconcilably differ, and that a security statute delayed is a security statute denied. The opposition’s answer was that a law restricting liberty ought to require the consent of both Houses in a federal legislature, that the Rajya Sabha represents the states whose police forces would operate the Act, and that carrying such a law on combined arithmetic against the stated refusal of one House converted a safeguard into a formality. Both answers are arguments about constitutional design rather than about terrorism, and the record does not refute either. What the record does establish is the political fact the joint sitting revealed: in March 2002 the government could pass its anti-terror law, but only by the one route that advertised, to every observer, that it could not pass the law in the ordinary way. The political cost of the device deserves recording alongside its mechanics, because the cost was part of what the device was for. A government that summons a joint sitting announces that it cannot carry the Council of States and that it regards the bill as important enough to spend the Constitution’s rarest procedure on it. The announcement raises the stakes of the vote itself: defeat at a joint sitting, after rejection in one House, would have destroyed the bill and damaged the government’s authority in a way that quiet withdrawal never could. The Vajpayee government accepted that exposure deliberately, and its managers treated the sitting as a demonstration, to the country and to the adversary across the border during a military standoff, that the Indian state could unite its legislature behind a security law when the occasion required. The opposition, for its part, treated the same sitting as a demonstration of the opposite proposition, that the government had been driven to a procedural extremity because the merits of its bill could not persuade a majority of the states’ House. Both demonstrations were performed before the same audience on the same day, and both were believed by the audiences disposed to believe them. What neither side could know on 26 March 2002 was that the demonstration would be the most durable part of the achievement. The statute the sitting enacted was within thirty months of its repeal. The sitting itself entered the permanent record of the Constitution’s working, where it remains the standard illustration of Article 108 in every textbook, examination and commentary, long after the Act it passed had been repealed and its provisions rehoused.

The bill received the assent of the President on 28 March 2002 and became the Prevention of Terrorism Act, 2002, Act 15 of 2002, deemed by Section 1(6) to have been in force from 24 October 2001, the date of the first ordinance it replaced.

What the Prevention of Terrorism Act Actually Said

Begin with the definition, because everything else in the Act is machinery built around it. Section 3(1) defined a terrorist act as an act done with intent to threaten the unity, integrity, security or sovereignty of India, or to strike terror in the people or in any section of the people, by using bombs, dynamite or other explosive or inflammable substances, or lethal weapons, poisons, noxious gases or other chemicals, or any other hazardous substances, in a manner causing or likely to cause death or injury, loss of or damage to property, disruption of supplies or services essential to the life of the community, or damage to the monetary stability of India by the production, smuggling or circulation of high quality counterfeit currency. The same clause swept in the detention of a person, coupled with threats to kill or injure, intended to compel a government or an international organisation to act or abstain from acting. Section 3(2) fixed the punishment: where the act resulted in death, death or imprisonment for life; in other cases, imprisonment of not less than five years extending to life, with fine. Two features of the definition carried the later controversy. The first is that its operative filter is intent: the prosecution had to show that the act was done with the intention the clause specifies, and it was on that word that the Supreme Court’s eventual reading of the whole Act would turn. The second is breadth. A definition that runs from bombing to counterfeit currency, and from assassination to the disruption of essential supplies, places a great deal of conduct at the boundary of the special law and the ordinary law, and leaves the choice of which law to invoke substantially to the police officer and the government that sanction the prosecution. The strongest critics of the Act located its danger in exactly that choice. The strongest defenders answered that terrorism is defined by purpose rather than by instrument, and that any definition organised by instrument would invite evasion.

Around the definition the Act built a set of organisation offences. Section 18 provided for the declaration of terrorist organisations, which were listed in the Schedule to the Act; the Schedule contained twenty-five organisations when the Act was passed, including the Lashkar-e-Taiba, the Jaish-e-Mohammed and the Liberation Tigers of Tamil Eelam. Section 19 gave an organisation a route to apply to the Central Government for removal from the Schedule, with the Review Committee machinery available in the process. Section 20 made membership of a terrorist organisation or a terrorist gang an offence punishable with imprisonment up to life. Section 21 made it an offence, punishable with imprisonment up to ten years, to invite support for a terrorist organisation, to arrange or address a meeting in furtherance of its activities or to address a meeting knowing it would encourage support for it. Section 22 punished raising funds for a terrorist organisation with imprisonment up to fourteen years. These provisions moved the frontier of the criminal law away from acts and toward association, expression and finance, and they did so on the reasoning the Supreme Court later articulated for the government: that modern terrorist organisations are sustained by members, speakers and fundraisers who never touch a weapon, and that a law confined to the person who plants the bomb addresses the smallest part of the phenomenon. The counter argument, pressed by the petitioners in the constitutional challenge, was that offences of association and expression are the provisions under which the merely sympathetic, the careless and the politically inconvenient are prosecuted, and the Vaiko prosecution examined in a later section was brought under precisely these clauses.

The provision that reached furthest into ordinary civic life was Section 14. It imposed on specified categories of persons, including those connected with an offence under the Act and those in possession of information about specified matters, an obligation to furnish information on the written requisition of an officer not below the rank of Superintendent of Police. Failure to furnish the information, or furnishing information knowing it to be false, was an offence punishable with imprisonment up to three years, or fine, or both, triable summarily. No equivalent duty, enforceable as a free standing offence against the citizen at large, had stood in the ordinary criminal law, and the provision drew a distinctive objection: that it converted silence into a crime and placed journalists, lawyers, doctors and family members under a pressure to inform that the ordinary law reserves for the witness box. The government’s defence of the clause was practical. Terrorist conspiracies are closed worlds, the argument ran, and the state cannot prosecute what it cannot learn; a duty to answer a senior officer’s written requisition, confined by rank and by form, was presented as a measured substitute for the far wider powers of compelled interrogation that the Act might have contained.

The procedural departures were the heart of the controversy, and there were four. The first concerned confessions. Under the Indian Evidence Act, a confession made to a police officer is inadmissible, a rule resting on the documented unreliability of confessions obtained in police custody. Section 32 of POTA departed from it. A confession made to a police officer not below the rank of Superintendent of Police, recorded in writing or on an electronic device, was made admissible, subject to conditions: the officer had to explain to the person that he was not bound to confess and that a confession could be used against him; the confession had to be recorded in an atmosphere free from threat, inducement or violence; and the person had to be produced before a Chief Judicial Magistrate or Chief Metropolitan Magistrate within forty-eight hours, who would record his statement and direct a medical examination where a complaint of torture was made. TADA had contained a confession provision of the same family, and its defence in the Supreme Court in Kartar Singh v State of Punjab had established the template on which POTA’s clause was drafted and on which it would later be defended. The second departure concerned custody. Section 49(2) rewrote the ordinary timetable of Section 167 of the Code of Criminal Procedure: the period within which police custody could be taken was enlarged, the investigating agency had ninety days to complete its investigation, and the Special Court could extend that period to one hundred and eighty days on the report of the Public Prosecutor showing the progress of the investigation and specific reasons for continued detention. Only after that period, if no charge sheet had been filed, did the accused acquire the right to release on bail for default. A further safeguard in the same clause required an officer seeking police custody of a person already in judicial custody to file an affidavit stating his reasons and explaining any delay. The third departure was bail. Sections 49(6) and 49(7) required the court to hear the Public Prosecutor before granting bail and provided that, where the Public Prosecutor opposed bail, no person accused under the Act was to be released unless the court was satisfied that there were grounds for believing that he was not guilty of the offence. The condition relaxed after one year of detention, when those sub-sections ceased to apply. The fourth departure was a set of institutional provisions: trials were to be held before Special Courts with the status of courts of session; a chapter of the Act authorised the interception of communications on the orders of senior officers, with review machinery for interception orders, and made intercepted material admissible in evidence; and a court could take cognizance of an offence under the Act only with the prior sanction of the Central Government or the State Government, a filter intended to screen prosecutions at the threshold.

The Act’s final group of provisions concerned its own discipline. Section 60 required the constitution of Review Committees, one by the Central Government and others by State Governments, each chaired by a person who was or had been a judge of a High Court. As enacted, the committees’ role in relation to individual prosecutions was limited and in practice advisory, a weakness the 2003 amendment was later passed to repair. The Act required annual reports on its working to be laid before Parliament and the state legislatures. It made the corrupt or malicious exercise of its powers by an officer itself punishable, and it carried the sunset already described: a life of three years from 24 October 2001, with savings that preserved proceedings begun under it even after expiry. Read whole, the statute was neither the blank cheque of its opponents’ rhetoric nor the carefully hedged instrument of its defenders’. It was a severe procedural code attached to a broad substantive definition, hedged by filters of rank, sanction and seniority whose effectiveness depended entirely on how the officers and governments operating them chose to behave, which is why the argument about POTA was always, in the end, an argument about its working as much as its text.

Two further parts of the scheme complete the textual picture, and both were directed at the infrastructure of terrorism rather than at individual acts. The first was forfeiture. The Act provided for the attachment and forfeiture of the proceeds of terrorism, the funds and property derived from or intended for terrorist activity, on the reasoning that an organisation deprived of its money is weakened more durably than an organisation deprived of a single operative. The history of this power inside the statute’s short gestation is a small lesson in institutional design: the first ordinance had vested forfeiture in a designated executive authority, and the second ordinance moved it to the Special Court, so that the taking of property required a judicial order rather than an administrative one. The second was interception. A dedicated chapter of the Act authorised the interception of wire, electronic and oral communications on the order of a competent authority at senior level, for the investigation of offences under the Act, subject to review of interception orders by a committee, and it made the intercepted material admissible in evidence. Interception as an investigative technique existed under the Telegraph Act and the safeguards the Supreme Court had laid down for telephone tapping in the nineteen nineties; what POTA added was an admissibility regime of its own, under which material gathered by interception could be placed before the Special Court as evidence rather than used only as intelligence. Together with the requirement of annual reports on the Act’s working, laid before Parliament by the Central Government and before the state legislatures by the State Governments, these provisions show the draftsman’s complete design: a statute that reached money and communications as well as persons, and that purported to account for its own use to the legislatures once a year.

The design’s weakness, visible in the text itself to a reader who compares the promise with the machinery, was that almost every safeguard in it was a safeguard of rank or of report rather than of decision. A requisition for information had to come from an officer of a stated rank, but no one outside the police could refuse it. Cognizance required a government’s sanction, but the sanctioning government was ordinarily the same government whose police had investigated. Interception required a senior officer’s order, but the review of that order was internal to the executive. The annual reports were to be laid before the legislatures, but nothing in the Act made any consequence follow from what a report disclosed. The one safeguard in the original scheme that involved a decision by a body independent of the prosecuting executive was the Review Committee, and as enacted its powers in relation to prosecutions were the weakest in the Act. That distribution, generous with seniority and reporting, parsimonious with independent decision, is the textual fact on which the whole later history turns: the amendment of 2003 and the repeal settlement of 2004 both consisted, in essence, of converting the Review Committee from the scheme’s weakest element into its strongest, after the prosecutions it might have prevented had already been launched.

Did POTA reverse the burden of proof?

No clause of POTA presumed an accused person guilty or required him to prove his innocence at trial; the prosecution’s burden remained the ordinary one. The Act’s severity operated before trial: a charge sheet could take one hundred and eighty days, and bail, where the Public Prosecutor opposed it, required grounds for believing the accused not guilty.

The Case for POTA, Stated at Its Strongest

The government’s case began from the character of the threat. Piloting the bill in the Lok Sabha on 18 March 2002, the Home Minister, L. K. Advani, told the House that there was a universal appreciation in almost all democracies that terrorism was a new kind of challenge to the security of a nation, a challenge for which the existing legal regimes were inadequate, and he pointed to Security Council Resolution 1373 as evidence that the obligation to legislate was not India’s invention but an instruction the international community had issued to every member state. The argument had a domestic limb as well. India in 2002 faced a proxy war in Jammu and Kashmir sustained from across the Line of Control, insurgencies in the North East, an armed Maoist movement in the central states, and, as December 2001 had shown, the capacity of a five man squad to strike at the seat of the legislature itself. The ordinary criminal law, the government argued, is built for the investigation of completed acts by identifiable offenders within the jurisdiction. It presumes witnesses who can be protected by the ordinary machinery, evidence that survives delay, and accused persons whose networks end where the evidence ends. Terrorist offending inverts each presumption. Conspiracies are proved from the inside or not at all, which is why confessions recorded under safeguards were said to be indispensable; investigations cross state and national boundaries, which is why ninety days was said to be too short; and organisations renew themselves through members, recruiters and financiers who commit no violent act, which is why the association offences were said to be the core of the Act rather than its excess.

The second limb of the government’s case was comparative and historical. Democracies facing terrorism, the argument ran, had all legislated specially: the United Kingdom had maintained prevention of terrorism legislation for decades and had consolidated it in the Terrorism Act 2000, and the United States had answered September 2001 with the USA PATRIOT Act within weeks. India, on this view, was not departing from democratic practice but converging with it, and doing so with greater restraint than its own past. POTA, its defenders pointed out, was not TADA reissued. It contained no general presumption of guilt of the kind TADA had carried. Its confession clause was hedged with the rank requirement, the warning, the forty-eight hour production and the medical examination that the Supreme Court had prescribed in Kartar Singh. It had a three year sunset where TADA had been extended again and again. It created Review Committees. It demanded sanction before cognizance and seniority before requisition. Arun Jaitley, speaking for the government in the parliamentary debates of that period, answered the charge of inadequate consultation by enumerating the process the bill had undergone: consultation of state governments, of the consultative committee of Parliament, meetings with chief ministers and with opposition leaders, a process he described as wide scale. The defence, in its complete form, was therefore not that POTA was gentle. It was that a state under attack is entitled to procedural weapons proportioned to the attack, that the weapons chosen were those every comparable democracy had chosen, and that the safeguards written into the text, the sunset, the committees, the sanction filter, the seniority requirements, were the answer to the misuse objection in advance: if the law was abused, the remedy lay against the abuser, under a statute that itself punished malicious exercise of its powers.

The third limb was federal and operational, and it was made most often by the governments and police leaderships of the affected states. Terrorism and organised crime, they argued, are investigated by state police forces whose officers work under the ordinary code and who face accused persons financed and directed from outside the state and sometimes outside the country. A central special law gave those forces a uniform instrument across state boundaries, a Schedule of proscribed organisations that applied in every state at once, and forensic and interception powers that a patchwork of state laws could not reproduce. The government also drew support from the Act’s own early operation in the prosecution that arose from the Parliament attack, examined in a later section, in which the confession provision and the Special Court delivered a completed trial within little more than a year of the offence. For the Act’s defenders, that prosecution was the design working as intended: a grave case, tried fast, on evidence the ordinary law might have excluded, in a court built for the purpose.

The Case Against POTA, Stated at Its Strongest

The opposition’s case began where the government’s case ended, with TADA. Between 1985 and 1995 India had operated a statute of the same family, and the record of that decade was the opposition’s central exhibit: a law presented as an emergency instrument for Punjab had spread to states and subjects far from its justification, had produced mass detention as a routine instrument of policing, and had ended in a political consensus so complete that no government defended renewing it. The review committees appointed in the mid 1990s to examine TADA cases state by state had recorded the pattern in official form. Against that record, the opposition argued, the burden lay on any government proposing to re-enact the model to show what had changed, and POTA’s text showed, in their submission, that nothing structural had changed. The definition remained broad enough to cover a riot, a dacoity or a protest that disrupted an essential service if a prosecutor chose to allege the specified intent. The confession clause restored the single provision of TADA that civil liberties opinion most condemned, for reasons the Evidence Act’s contrary rule embodies: a confession taken in police custody is the evidence most easily manufactured and least safely tested, and the seniority of the officer recording it does not change the conditions under which it is obtained. The bail clause inverted the premise of criminal procedure by requiring the court, before trial, to be satisfied of grounds for believing the accused not guilty, which in practice, the opposition argued, meant that arrest under POTA was itself the punishment, served in advance of any adjudication and for a period the custody provisions stretched to six months before a charge sheet was even required.

The second limb of the case against was about process and motive. The National Human Rights Commission had recorded its opposition to the near identical Prevention of Terrorism Bill of 2000, and the ordinance of October 2001 was promulgated without consulting the Commission, a sequence Amnesty International documented in its briefing on the ordinance, in which it also questioned the stated urgency of legislating by decree when Parliament was to sit within weeks. The People’s Union for Civil Liberties carried the objection into the Supreme Court as a constitutional challenge. In Parliament, the Congress party and the parties of the Left opposed the bill through every stage, voting against it in the Rajya Sabha in sufficient numbers to defeat it there, and their parliamentary argument had a federal edge the joint sitting could not answer: policing is a state function, the states’ forces would operate the Act, the states’ governments would answer for its misuse, and yet the Act was being imposed over the refusal of the House in which the states are represented. The opposition also named the political temptation the statute created with a precision that events then illustrated. A law that lets a state government arrest an opponent for a speech, hold him for months without a charge sheet, and oppose his bail on the strength of the accusation itself, is a law whose abuse does not require corruption of its text. It requires only an ordinary political motive and an ordinary police force. The Vaiko arrest in Tamil Nadu, in which the government of one party prosecuted a former ally under the speech offence within four months of the Act’s passage, was the opposition’s proof of concept, delivered by the government’s own side of the political spectrum.

The confession clause drew a further objection that went to the sociology of the police station rather than to the text. The safeguards of Section 32, the rank of the recording officer, the warning, the production before a Magistrate, the medical examination, all operate, the critics observed, inside a system in which the accused has spent days in the custody of the same force that records his statement, in which the Magistrate sees him for minutes, and in which a complaint of torture must be made to the very authorities holding him, against the officers who will continue to hold him. A safeguard that requires the victim of coercion to initiate it, in the place of coercion, within forty-eight hours, is a safeguard whose activation the coercion itself discourages. The government’s answer, that the Supreme Court had prescribed precisely these conditions in Kartar Singh and that Parliament had enacted the Court’s own design, was formally complete and practically unresponsive, because the objection was never that the safeguards were absent from the paper. It was that paper safeguards of this kind had been present under TADA as well, in the interpretation the Court had imposed, and had not prevented the custodial practices that made TADA’s confessions notorious. On this point the two sides were arguing about different things: the government about the law’s design, the critics about its environment, and the PUCL judgment, which sustained the clause by re-stating the safeguards, decided the design question while leaving the environmental one exactly where it found it.

The third limb was constitutional, and it is the limb that outlasted the statute. The petitioners in the Supreme Court argued that Sections 20, 21 and 22 punished association and expression without requiring any connection to a violent act, offending the freedoms of speech and association in Article 19; that Section 32 offended the protection against self incrimination and the fair trial guarantee; and that the disclosure offence in Section 14 compelled speech of a kind the Constitution protects the citizen against. Behind the particular articles stood a structural claim, made by civil liberties advocates throughout the Act’s life: that special anti-terror procedure in India had never remained special, that each statute’s exceptional powers had migrated into ordinary policing, and that the true precedent POTA set was not for the trial of terrorists but for the treatment of any accused person a government found inconvenient. That claim could not be tested against POTA alone, since it is a claim about sequences, and it is examined as such in the verdict section of this account, against the transplant that followed the repeal.

The Misuse Controversy: What the Record Establishes

The misuse allegations against POTA were specific, and they repay being stated specifically rather than as an atmosphere. They fell into three groups, corresponding to three kinds of case. The first group concerned communal application in Gujarat. After the train burning at Godhra on 27 February 2002, in which fifty-nine passengers died, and the violence across Gujarat that followed it, the state government prosecuted the Godhra accused under POTA, and POTA was invoked in a wider set of cases arising from that period. Critics in Parliament and outside it alleged that the Act was applied in Gujarat overwhelmingly against members of one community, and Amnesty International’s reporting on detentions in Ahmedabad in 2003 documented Muslims held under security legislation in circumstances the organisation characterised as illegal detention. The Gujarat government rejected the charge of communal application and defended the Godhra prosecution as a response to a terrorist conspiracy against a train full of pilgrims. The second group concerned political application, and its exhibit was the prosecution of Vaiko, the leader of the Marumalarchi Dravida Munnetra Kazhagam, examined in detail shortly. The third group concerned the Act’s use in the states fighting the Maoist insurgency, where human rights organisations documented arrests of alleged sympathisers in Jharkhand that included, on their reporting, elderly persons and minors, and where the distance between the statutory definition of terrorism and the conduct alleged was at its widest. Contemporary legal commentary surveying the Act’s first years recorded hundreds of questionable and prolonged detentions and noted that the most visible cases were those in which political figures were arrested by rivals controlling the state police machinery, while the least visible fell on members of disempowered minorities with no forum in which to complain. The reporting on which that record rests depended throughout on sources inside police stations, prisons and affected families who spoke at personal risk, which is why the craft question of how journalists use VaultBook to protect their sources belongs to this part of the story: a public account of a security law’s working exists only where the people who experience that working can describe it without being identified.

Two case histories carry more weight than any aggregate, because both were eventually decided by the Act’s own machinery and by the courts, and both therefore belong to the record rather than to the controversy. The first is Vaiko’s. A case was registered against him on 4 July 2002 under Sections 21(2) and 21(3) of POTA, among other provisions, over speeches at a public meeting at Thirumangalam in Madurai district in which he had expressed support for the Liberation Tigers of Tamil Eelam, a proscribed organisation in the Act’s Schedule. He was arrested on 11 July 2002 at Chennai airport on his return from Chicago and produced before a magistrate at Madurai the following day, and the case went to the Special Court at Poonamallee. The prosecuting government was the All India Anna Dravida Munnetra Kazhagam government of Tamil Nadu, and Vaiko’s party had been its ally; the prosecution of a sitting Member of Parliament and former ally for the content of a political speech, under a statute his prosecutors’ allies at the Centre had enacted over opposition warnings about exactly this use, made the case the defining political controversy of the Act’s life. Vaiko remained in custody for roughly nineteen months before his release on bail in February 2004. He was also a petitioner before the Supreme Court, his writ petition being heard with the People’s Union for Civil Liberties challenge discussed in the next section. On 8 April 2004 the Central Review Committee, exercising the binding power the 2003 amendment had just conferred, held that there was no prima facie case against him: the speeches, it reasoned, could not be said to encourage support for the LTTE or to further its activities, sympathy or oral support for the cause of the Eelam Tamils not being support for the organisation’s ideology or methodology, and the conduct alleged therefore fell outside Section 21(3). The committee directed the state government to have the prosecution withdrawn. The legal end of the matter nevertheless took a further decade, for reasons the repeal section explains, and the case was closed in law only in October 2014, when the Madras High Court held that the committee’s direction had withdrawn it years earlier.

The review machinery was also tested in the prosecutions arising from the attack on the Akshardham temple in Gandhinagar in September 2002, the other great atrocity of the Act’s first year, and the litigation over the committees’ authority in those cases travelled with the Godhra litigation to the Supreme Court’s judgment of 2008. The pattern repeated there in form: a state prosecution under the special law, a committee’s scrutiny of whether the special law applied, and a constitutional contest over who had the last word.

The second case history is Godhra, and it is the harder one, because the underlying crime was not in dispute. The committee that the repeal legislation charged with reviewing every POTA case examined the Godhra prosecutions and, by its opinion of 17 May 2005, held that the accused could not be tried under POTA, though they could be tried under the other laws their alleged acts attracted, and it directed the Public Prosecutor to move without delay for withdrawal of the POTA charges. The opinion, delivered by the Central POTA Review Committee headed by Justice S. C. Jain, a retired judge, was the single most consequential finding the review machinery produced: the flagship prosecution of the Act’s flagship case did not, in the committee’s assessment, disclose a prima facie terrorist act within the statute. The Gujarat government supported the victims’ challenge to that conclusion, and the question travelled to the Supreme Court, which in October 2008 upheld the review provisions of the repeal Act and the binding character of the committee’s opinion, and to the Gujarat High Court, which on a victim’s father’s petition upheld the committee’s recommendation, holding that the incident, though a shocking and serious criminal act, fell short of the requirement of an act of terror on the material available. The trial then proceeded under the Indian Penal Code, the Railways Act and the damage to public property legislation, and in February 2011 the special court convicted thirty-one of the accused and acquitted sixty-three, sentencing eleven to death. The Godhra sequence is the complete misuse argument and the complete answer to it, contained in one record. A state government had characterised a massacre as terrorism and prosecuted it under the special law; the special law’s own review machinery, upheld by two courts, held that the characterisation failed the statute’s own test; and the ordinary law, to which the case was returned, proved capable of trying it to conviction. What the sequence does not show is that the accused spent the intervening years, in many instances the better part of a decade, in custody under a legal label two courts eventually held inapplicable.

Did the Review Committees find that POTA had been misused?

The committees made no general finding about the Act. They decided case by case whether a prima facie case existed, and in the two most prominent instances, the Vaiko prosecution and the Godhra prosecutions, they found that it did not and directed withdrawal. Their pattern of outcomes, not any verdict on the statute, is the official record of its misuse.

Set against these decided instances, the larger aggregate claims made on both sides remain contested, and this account does not pretend otherwise. Critics of the Act asserted that misuse was systematic and that the decided cases were the visible portion of a practice that reached thousands of unremarkable detentions. Governments that operated the Act asserted that the charges were exaggerated for political effect, that the statute’s filters of sanction and seniority were observed, and that the Review Committees’ willingness to direct withdrawals proved the safeguards worked. The record cannot fully adjudicate between those aggregate claims, because no authoritative count of POTA detentions, withdrawals and convictions was ever laid before Parliament in a form that settles the question, and the annual reports the Act required did not produce one. What the record does settle is the instances. In the cases that were tested by the machinery Parliament itself built, the machinery repeatedly found that the Act had been applied where its own definition was not met, and that finding, repeated across a speech case in Tamil Nadu and a massacre case in Gujarat, is the hardest fact in the whole controversy for the Act’s defenders to answer, and the one its critics are entitled to rest on without exaggeration.

The Courts and POTA: The PUCL Judgment

The constitutional challenge reached the Supreme Court while the Act was still young, and the judgment in People’s Union for Civil Liberties v Union of India, delivered on 16 December 2003, governed the statute for the remainder of its life. The petitioners attacked the Act on the grounds set out in the case against it: legislative competence, the association and expression offences, the confession clause, the disclosure duty and the bail provisions. The Court dismissed the petitions and upheld the Act, but it upheld it in the manner the Supreme Court had established a decade earlier for TADA in Kartar Singh, by sustaining the text while narrowing its meaning, a template whose origins and operation are traced in the account of TADA, the predecessor statute POTA replaced. Two features of the judgment define it. The first is doctrinal. The Court held that the mere possibility of abuse of a statute is not a ground for declaring it unconstitutional, a proposition that answered the opposition’s strongest rhetorical case by classifying it: misuse, on this reasoning, is a fact about administration, to be corrected case by case, and not a fact about the law, to be corrected by invalidation. The second feature is interpretive, and it qualifies the first in practice, because the Court sustained the contested provisions only after reading into them the conditions that made them, in its judgment, constitutional.

The most consequential reading concerned the organisation offences. The Court held that Sections 20, 21 and 22 could be sustained only if they were understood to require a positive inference that the person acted with intent to further, encourage, promote or facilitate terrorist activity: a person who professes sympathy, invites support or addresses a meeting without that intention does not commit the offence. Understood in that way, the Court reasoned, the sections could not be misused, and with that clarification it upheld them. The confession clause was sustained on the Kartar Singh reasoning, with the safeguards of Section 32 treated as real protections rather than formalities: the production of the maker before a Magistrate within forty-eight hours was an additional safeguard, the Magistrate was required to enquire into any complaint of torture and to provide for medical examination and treatment, a confession shown to have been extracted by threat or inducement was a nullity in law, and the trial court bore the responsibility of deciding, with reasons, whether a confession was voluntary and admissible. The bail provisions in Sections 49(6) and 49(7) were upheld as not unreasonable, the Court accepting that offences under the Act were of a complexity that justified the extended investigation period and the conditions attached to release, while noting the relaxation the statute itself provided after one year of detention. The result was a statute that survived entire and meant less than its bare text suggested. Every safeguard the Court identified was a safeguard the police and the prosecuting governments were from that point bound in law to observe, and every later failure of those safeguards, in the cases the Review Committees examined, was therefore not merely a political embarrassment but a departure from the Act as the Supreme Court had authoritatively construed it.

The Kartar Singh inheritance shaped the judgment at every point, and it is worth stating plainly what that inheritance was. When TADA’s confession provision had been challenged a decade earlier, the Supreme Court had sustained it by the same method: the clause survived because the Court surrounded it with safeguards, the rank of the recording officer, the warning to the maker, the prompt production before a magistrate, and a judicial duty to test voluntariness, and the Court treated those safeguards as the conditions of the provision’s validity rather than as courtesies. Parliament drafted Section 32 of POTA with that judgment open on the table, writing the Kartar Singh safeguards into the statute’s own text, and the PUCL Court, faced with a provision that embodied its predecessor’s conditions, sustained it in its turn. The PUCL Court also disposed of the challenge to Parliament’s competence to enact such a law at all, holding that legislation against terrorism fell within Parliament’s powers under the Union List, so that the Act could not be defeated at the threshold as a trespass on the states’ field of public order. The pattern across the two judgments is the real doctrine of Indian anti-terror jurisprudence in this period: the courts would not deny the legislature its special statutes, and they would not leave those statutes with the meanings their broadest words suggested. Every generation of the legislation was therefore, in practice, co-authored by the bench that reviewed it, and a reader who wants to know what POTA meant in law must read the PUCL clarifications as part of the Act, which is how the Review Committees and the Madras High Court later treated them when the Vaiko direction applied the intent clarification to a set of speeches and found the offence unmade out.

The judgment’s limits are as much a part of the record as its holdings. PUCL was an interpretation, not a rescue, and interpretations bind only where institutions comply with them. The Vaiko prosecution, which the committees later found disclosed no prima facie case, had been launched eighteen months before the judgment, on a reading of Section 21 that the Court’s clarification excluded; the judgment did not itself free him, and the instrument that did was the Review Committee power Parliament created in the amendment of the same year. Nor did the judgment touch the political question, which was never before the Court: whether a statute that requires this volume of judicial narrowing, and this machinery of ex post review, in order to operate constitutionally is a statute a legislature should enact at all. The courts’ role in the POTA story is therefore double, and both halves matter to the verdict. The judiciary kept the Act within the Constitution by construing it, and the judiciary’s own later decisions, in the Godhra litigation of 2008 and the Vaiko litigation of 2014, gave binding force to the review machinery that undid the Act’s most prominent applications. The statute was upheld, and the cases that made it infamous were, one by one, held not to be cases under it.

A Central Law in State Hands: The Federal Friction

One structural feature of the Act explains more of its history than any other, and it receives less attention than the confession clause or the bail standard. POTA was a central statute that only state machinery could operate. Policing in India is a state function. The officers who arrested under POTA, recorded confessions under it, opposed bail under it and laid charge sheets under it were state police officers, answerable to state governments, and the sanction without which no court could take cognizance of a POTA offence was, in the ordinary case, the sanction of the State Government concerned. The Central Government had enacted a law whose daily use it did not control, and the states that controlled its use included states governed by parties opposed to the coalition at the Centre, states governed by the Centre’s own allies pursuing their own quarrels, and states, Gujarat foremost among them in the controversy, governed by the Centre’s own party and applying the Act with an enthusiasm the Centre then had to defend or disown. The federal design of the Act was therefore not an incidental feature. It was the mechanism by which a statute debated in New Delhi as an instrument against cross-border terrorism became, in Thirumangalam, an instrument in a Tamil Nadu political rivalry, and in Gujarat, an instrument in the legal aftermath of communal violence, and no reading of its text in Parliament could have predicted or prevented either use, because the text delegated the operative decisions, whom to arrest, what to characterise as terrorism, whether to sanction prosecution, to governments the enacting Parliament did not command.

The review machinery of the Act has to be understood against that design, because the machinery was, from 2003 onwards, a federal instrument as much as a liberties safeguard. The amendment of that year created a hierarchy inside the review system: a State Committee’s direction bound the State Government and its officers, the Central Committee’s direction bound everyone, and where the two committees differed on the same offence the Central Committee prevailed. Parliament was thereby giving a body appointed by the Central Government power to terminate prosecutions launched by state governments, a striking transfer of authority in a field the Constitution assigns to the states, and it did so with the support of the coalition that had enacted the statute, because the cases that embarrassed the Act were state cases for which the Centre was taking the blame. The repeal Act of 2004 completed the transfer by routing the universal audit of POTA cases through the Central Committee alone. The two great confrontations that followed were both, in form, federal confrontations. In Gujarat, a state government aligned with the victims resisted a central committee’s direction about a state prosecution, and the Supreme Court’s judgment of 2008 settled the matter in the committee’s favour as a question of statutory construction. In Tamil Nadu, a state prosecution launched in 2002 was found by the central committee in 2004 to have no prima facie basis, and the state machinery’s resistance to that conclusion kept the case alive in the courts for a further decade. The lesson the federal history teaches is distinct from the lesson about civil liberties, and it survives the repeal. A central anti-terror statute in India is a licence issued to twenty eight separate political executives, and its real content in any given year is the sum of the uses those executives choose to make of it, checked only by whatever review machinery the statute builds with power to overrule them.

The POTA Passage and Repeal Record

The whole controversy can be reduced to a single sequence of legislative acts, and the sequence, set out in full, is the argument of this article in its most compressed form. The final column records, for each step, what that step changed in the actual quantum of anti-terror power the Indian state held: not the power asserted in speeches, but the offences, procedures and organisations regime in force in law.

Step	Date	What happened	Change in the quantum of anti-terror power
First ordinance	24 October 2001	The President promulgates the Prevention of Terrorism Ordinance under Article 123, Parliament not being in session	The full special regime is created by executive order: new offences, police confessions, extended custody, restricted bail
Second ordinance	30 December 2001	The ordinance is re-promulgated with modifications after the winter session ends without a bill	No change in the powers; the regime’s proposed life is cut from five years to three, and the Schedule grows from twenty-three organisations to twenty-five
Lok Sabha passage	18 March 2002	The Lok Sabha passes the Prevention of Terrorism Bill, the disapproval resolution against the ordinance being negatived	No change; the regime continues as an ordinance and gains the first House’s approval
Rajya Sabha rejection	21 March 2002	The Rajya Sabha defeats the bill by 113 votes to 98	No change in law; the bill is blocked in the ordinary course and survives only because Article 108 exists
Joint sitting	26 March 2002	Both Houses, sitting together under Article 108 with Deputy Speaker P. M. Sayeed presiding, pass the bill by 425 votes to 296	The ordinance regime becomes statute law, deemed in force from 24 October 2001; the quantum is the full POTA code, carrying a three year life
Presidential assent	28 March 2002	The bill becomes the Prevention of Terrorism Act, 2002, Act 15 of 2002	Formal completion of the same step; no further change
Amendment	27 October 2003, by ordinance, later the Amendment Act of 2003, Act 4 of 2004	Section 60 is amended to give the Review Committees power, on the application of an aggrieved person, to direct that a case disclosing no prima facie basis be dropped, the direction binding the government and the investigating officer	The first subtraction: a working mechanism for ending prosecutions, confined to cases under the Act
Repeal	21 September 2004	The Prevention of Terrorism (Repeal) Act, 2004, Act 26 of 2004, first enacted by ordinance, repeals POTA, with savings preserving pending proceedings and subjecting every registered case to Review Committee scrutiny within one year	On paper, the whole special code is subtracted; in practice, pending cases continue and the subtraction awaits the committees’ findings
Simultaneous UAPA amendment	21 September 2004, by the Amendment Act of 2004, Act 29 of 2004, deemed in force the same day	The Unlawful Activities (Prevention) Act is amended to insert a chapter of terrorist act offences, a forfeiture chapter, a terrorist organisations chapter and a Schedule of proscribed organisations	The substantive offences, punishments, organisation bans and forfeiture powers are restored inside a permanent statute; the special procedure, confessions, extended custody, bail conditions and sunset are not restored
Net effect of the full sequence	2001 to 2004	A statute enacted by the rarest procedure in the Constitution is repealed, and its penal substance continues without a day’s interruption in another Act	Approximately zero on offences and organisations; negative on procedure; positive for the state on permanence, since the transplanted provisions carry no expiry date

Two features of the record in this form deserve notice before the narrative resumes. The first is that every step in the passage sequence before the joint sitting changed nothing in the law at all: the power the state held on 22 March 2002, the day the Rajya Sabha’s message reached the Lok Sabha, was identical to the power it had held since October, because the ordinance carried the whole regime and the parliamentary votes, until the last one, were votes about the ordinance’s future rather than about the existence of the powers. The second is the symmetry at the end of the table. The repeal and the transplant were executed by a pair of ordinances promulgated on a single day, and their combined effect is the mirror image of the passage: an enormous political event, the destruction of a named and hated statute, accompanied by a legal change that its own authors confined to procedure, duration and address. The sections that follow examine the three steps in the sequence that did change the law: the amendment of 2003, the repeal of 2004, and the amendment of the Unlawful Activities (Prevention) Act that accompanied it.

Parliament Repairs Its Own Statute: The Amendment of 2003

The amendment of 2003 is the least remembered step in the sequence and, on the analysis this article defends, the most instructive. By the autumn of 2003 the Act had been in force for two years, the misuse controversy was at its height, the Vaiko prosecution was a standing embarrassment to the coalition that had enacted the statute, and the Review Committees that Section 60 had created were understood on all sides to lack the power that mattered: they could consider the Schedule and entertain grievances, but they could not end a prosecution. The government’s own assessment, embodied in its remedy, was that the defect was real. On 27 October 2003 an ordinance amended Section 60, and the amendment was later enacted as the Prevention of Terrorism (Amendment) Act, 2003, Act 4 of 2004, deemed in force from the ordinance’s date. The amendment inserted sub-sections (4) to (7) into Section 60. Under them, any aggrieved person could apply to the Review Committee for a review of whether a prima facie case existed for proceeding against him under the Act. The committee’s directions were made binding on the Central Government, the State Government and the investigating police officer alike. Where the Central Committee and a State Committee had both reviewed the same offence and differed, the Central Committee’s direction prevailed. And where a committee found no prima facie case, the proceedings pending against the person were deemed to have been withdrawn from the date of the direction.

The amendment is worth dwelling on because it was a concession of a precise and limited kind, and its limits define the government’s position better than any speech. Parliament conceded that prosecutions had been, or might have been, launched without a prima facie basis sufficient to survive independent scrutiny, and it built a mechanism with teeth to end them: binding directions, a deemed withdrawal that required no further order, and a central override to prevent a state committee and a state government from protecting a state prosecution. Parliament did not concede that the Act itself was misconceived. No offence was narrowed, no custody period shortened, no bail condition softened, and the amendment was carried by the same political formation that had enacted the statute the year before. The first major exercise of the new power was the Central Committee’s direction of 8 April 2004 in Vaiko’s case, and the direction’s reasoning applied the Supreme Court’s PUCL clarification to the facts: speeches expressing sympathy with the Eelam Tamil cause did not, without intent to further the LTTE’s activities, amount to the support offence. The amendment therefore stands in the record as the moment the enacting coalition accepted the critics’ narrowest and strongest claim, that the Act as operated was capturing cases its own definition did not cover, while continuing to reject the critics’ larger claim about the Act itself. When the government changed in 2004, the machinery the amendment created was already in place, and the repeal legislation simply universalised it.

The amendment also altered what could honestly be said in the Act’s defence, and the alteration cut in both directions. Before October 2003, the government’s answer to a documented bad case was that the Act contained safeguards and that courts could correct errors on appeal, a process that ran in years. After the amendment, the answer was that any person prosecuted without a prima facie basis had a remedy that ran in months and ended the case. Critics answered, with equal justice, that the amendment was a confession: a statute that needs an emergency exit built into it a year after enactment, because its own filters of sanction and seniority have failed to keep out cases that cannot survive a prima facie test, is a statute whose front end was defectively designed. Both characterisations are supported by the amendment’s text, and the choice between them is the whole controversy in miniature. What is not contestable is the direction of travel within the statute’s own life. POTA began, in 2001, as a grant of power moderated by seniority and reporting. It ended, in 2004, with a binding external audit of every case it had generated. The distance between those two designs is the measure of what the enacting coalition learned from operating its own law, and it is a distance no speech in the joint sitting debates of March 2002 anticipated travelling.

Repeal in 2004: The Promise, the Twin Ordinances and the Savings

The general election of 2004 returned a Congress led coalition, the United Progressive Alliance, to power under Prime Minister Manmohan Singh, and the repeal of POTA was among the commitments its Common Minimum Programme carried. The commitment was redeemed with a symmetry that the Act’s critics were quick to note and that this article treats as the centre of the story. In September 2004 the Union Cabinet approved the course, and on 21 September 2004 two ordinances were promulgated on the same day. The first repealed the Prevention of Terrorism Act. The second amended the Unlawful Activities (Prevention) Act to receive its substance. Parliament then enacted both measures in statutory form: the Prevention of Terrorism (Repeal) Act, 2004, Act 26 of 2004, in force from 21 September 2004, and the Unlawful Activities (Prevention) Amendment Act, 2004, Act 29 of 2004, which received its form later in December but was deemed to have come into force on 21 September 2004, so that at no point in the sequence did a day pass on which the terrorist act offences existed in neither Act. The opposition that had enacted POTA charged that the repeal weakened national security. The government answered, accurately as far as it went, that the offences, the punishments, the organisation bans and the forfeiture powers continued in the amended UAPA, and that what had ended was a discredited procedural code and a name. Both the charge and the answer were self serving in the way the brief for this controversy predicts: the charge ignored the transplant, and the answer declined to dwell on it, preferring to present the repeal as the destruction of a draconian law rather than as its rehousing.

The repeal Act itself is a short statute with an unusually careful savings scheme, and its terms governed POTA cases for years after the repeal. Section 2(2) preserved the ordinary consequences of repeal in the citizen’s disfavour where they had already accrued: the repeal did not affect the previous operation of the Act, rights and liabilities acquired under it, penalties incurred, or investigations and legal proceedings in respect of them, all of which could be instituted, continued and enforced as if the Act had not been repealed. A proviso set the outer limit: no court was to take cognizance of an offence under POTA after one year from the repeal Act’s commencement, which closed the door to fresh POTA prosecutions after 21 September 2005. Section 2(3) then did the work the savings alone would not have done. Notwithstanding the repeal of Section 60, the Central Government’s Review Committee was required to review every case registered under POTA within one year of the repeal Act’s commencement, whether or not the person affected applied for review. Where the committee found no prima facie case, cases in which cognizance had been taken were deemed withdrawn, and investigations still pending stood closed, in each case from the date of the committee’s direction. Section 2(4) armed the committee with the powers of a civil court to secure documents and records, and Section 2(5) allowed the Centre to constitute as many committees as the caseload required. The design is the 2003 amendment universalised and made compulsory: review ceased to depend on an aggrieved person’s application and became an audit of the entire POTA caseload, conducted against the calendar of a single year.

The audit’s consequences took years to finish working through the courts, and their chronology is the true epilogue of the Act. The Godhra opinion of May 2005 was delivered under this machinery. The Gujarat government’s resistance to it produced the Supreme Court’s judgment of 21 October 2008 in the Godhra and Akshardham litigation, which upheld Sections 2(3) and 2(5) of the repeal Act as constitutional and settled the mechanism’s central doubt: a committee’s finding of no prima facie case withdrew the case by operation of the statute itself, without any need for the Public Prosecutor to apply, or the court to consent, under the ordinary withdrawal provision of the Code of Criminal Procedure. The Vaiko prosecution required the same principle to be stated again by the Madras High Court in October 2014, which set aside the Special Court’s order of September 2004 refusing effect to the withdrawal and held that Section 321 of the Code had never had any application once the committee had spoken. A prosecution begun in July 2002 was therefore closed in law in October 2014, twelve years after the arrest and a decade after the committee direction that the statute said had already withdrawn it. Persons convicted under POTA before the repeal stood outside this machinery altogether: the savings preserved their convictions and sentences, which continued to be executed and appealed as if the Act remained in force. The repeal governed the future of the Act’s procedure, not the fate of those the Act had already condemned, and the most serious demonstration of that limit is the prosecution that arose from the attack on Parliament itself, to which this account turns.

The Parliament Attack Prosecution: The Act’s Exhibit and Its Warning

The prosecution of the Parliament attack case was the use for which POTA had, in the government’s presentation, been designed, and its progress through the courts is the most complete single test of the statute in operation. The investigation was conducted while the regime stood on the authority of the second ordinance, and the charge sheet was filed on 14 May 2002 before the Special Court designated under the Act, within weeks of the Act’s passage at the joint sitting. The trial was conducted at the speed the statute promised. On 18 December 2002, a year and five days after the attack, the Special Court convicted all four accused and sentenced Mohammad Afzal, Shaukat Hussain Guru and S. A. R. Geelani to death, and Afsan Guru, tried under her married name and known to the record as Navjot Sandhu, to five years’ imprisonment. For the Act’s defenders, that completed first instance trial, in a case of this gravity and complexity, was the design vindicated: the Special Court, the confession provision and the custody rules had together produced what the ordinary system, in their argument, could not have produced at that pace. The appeal history complicates the vindication in a way that matters more than the trial. The Delhi High Court, in October 2003, acquitted Geelani, a lecturer at a Delhi college, and Afsan Guru outright, holding the evidence against them insufficient, while confirming the death sentences of Afzal and Shaukat. Half of the prosecution’s capital case had therefore failed at the first appeal, on the insufficiency of the evidence the special machinery had gathered against two of the four accused, and the trial court’s judgment against them stands in the record as a warning about what the accelerated procedure could do to the assessment of evidence in its lower courts.

The Supreme Court’s judgment of 4 August 2005 in State (NCT of Delhi) v Navjot Sandhu completed the appellate record. The Court confirmed Afzal’s death sentence, resting its conclusion on the confession he had recorded under Section 32 of POTA together with the circumstantial evidence, including the telephone records and the recoveries, that connected him to the slain attackers. It allowed Shaukat Hussain Guru’s appeal in part, setting aside his convictions on the principal charges and convicting him only of concealing the conspiracy, under Section 123 of the Indian Penal Code, with a sentence of ten years’ rigorous imprisonment. It dismissed the State’s appeals against the acquittals of Geelani and Afsan Guru, observing of Geelani that although a needle of suspicion pointed toward him, suspicion could not substitute for proof. Afzal’s sentence was carried out in February 2013 after the rejection of his mercy petition. The case thus ended with one execution, one conviction for concealment, and two acquittals, out of a charge sheet that had sought four convictions and three death sentences. Each side of the POTA argument can quote that outcome, and does. The government of the day quoted the completion of the trial and the confirmation of its central conviction in the gravest case the Republic had prosecuted in a generation. The critics quoted the acquittals and the partial collapse of the case on appeal as evidence that the special procedure had overcharged first and discriminated later. The sober conclusion the record supports is narrower than either. The ordinary appellate structure, applying ordinary standards of proof, corrected the special court’s excesses without needing the special law to be repealed for the purpose, and the conviction that survived did so on evidence whose centrepiece, the Section 32 confession, belonged to the very category of evidence Indian law has excluded for a century and has excluded again in every anti-terror statute enacted since.

The Transplant: What Moved into the UAPA, and What Did Not

The second ordinance of 21 September 2004, and the Amendment Act that followed it, rebuilt the Unlawful Activities (Prevention) Act, 1967 from a statute about unlawful associations into the general anti-terror code of the Indian state. The amendment of the Unlawful Activities (Prevention) Act carried out on the same day as the repeal is analysed in its own right in the account of the UAPA amendments and what each of them changed; what matters here is the inventory, because the inventory is the verdict on the repeal. The amendment substituted new chapters in the 1967 Act. A chapter on punishment for terrorist activities defined the terrorist act in Section 15 in terms that track Section 3 of POTA closely, the same intent to threaten the unity, integrity, security or sovereignty of India or to strike terror in the people, the same catalogue of means, the same reach to property damage, essential supplies and counterfeit currency, and Section 16 attached the same punishments, death or life imprisonment where death resulted, and imprisonment of five years to life in other cases. Companion offences punished the raising of funds for a terrorist act, conspiracy, organising camps, recruitment and harbouring. A chapter on forfeiture reproduced the machinery for seizing the proceeds of terrorism. A chapter on terrorist organisations reproduced the listing scheme: a Schedule of proscribed organisations, an offence of membership, offences of support and fund raising, and a route for an organisation to seek removal. The organisations proscribed under POTA continued proscribed under the UAPA without any fresh examination of their cases. For the organisations themselves, the repeal was a pure formality of address: an organisation listed in POTA’s Schedule in 2002, with the membership, support and fund raising offences that listing attracted, found itself listed in the UAPA’s Schedule in 2004 with the same offences re-enacted around it, its members exposed to the same prosecution under a new section number. The continuity was deliberate and was defended as such. A repeal that had struck at the listings would have required the state to rebuild every proscription from the evidence outward, in the middle of the conflicts that had produced the listings, and no participant in the 2004 settlement proposed that course. The point is worth isolating because it shows where the repeal’s drafters located the boundary of their reform. The organisation regime, the definitional core and the punishment scale were treated as the settled inheritance of the Indian state, to be carried from statute to statute as administrations changed. The procedure was treated as the variable. Almost everything that is distinctive in the POTA story, its confessions, its custody periods, its committees and its sunset, belongs to the variable side of that line, and almost everything that survived belongs to the settled side. On the substantive side of the ledger, therefore, the repeal subtracted almost nothing. A person who committed, financed, supported or belonged to a terrorist organisation on 20 September 2004 faced the same offences and the same punishments on 22 September 2004, under a different Act’s name.

The subtraction was real, but it lay entirely in procedure, and its inventory is equally precise. The confession provision did not move: no clause of the amended UAPA made a confession to a police officer admissible, and Section 32 of POTA remains the last occasion on which Indian anti-terror law contained the device in the form TADA had pioneered. The interception chapter did not move: the amended UAPA contained no scheme making intercepted communications admissible under its own authority. The Special Courts did not move: UAPA prosecutions proceeded before the ordinary courts under the ordinary Code. The custody timetable did not move: investigation of a UAPA offence ran on the Code’s ordinary periods, not on the ninety and one hundred and eighty days of Section 49, and the bail condition requiring grounds for believing the accused not guilty did not move either. The disclosure offence did not move: the amended UAPA imposed no free standing duty on citizens to furnish information on pain of imprisonment. And the Review Committees did not move: the amended statute created no body able to direct that a prosecution be dropped, the repeal Act’s committees being a winding up device for POTA cases alone. Parliament in 2008, after the Mumbai attacks, returned to the subject and inserted Section 43D into the UAPA, restoring an extended investigation period of up to one hundred and eighty days and a stringent bail condition for UAPA offences. The sequence is instructive. The procedural code that the repeal discarded in 2004 as the disreputable part of POTA was reassembled inside the UAPA four years later, under the pressure of a new atrocity and with far less parliamentary resistance than its original enactment had required, while the confession provision, the one device no government has reintroduced, marks the true outer limit of what the repeal permanently removed.

What happened to POTA’s three year limit when its provisions moved into the UAPA?

It disappeared. The Unlawful Activities (Prevention) Act, 1967 is permanent legislation with no sunset clause, and the chapters inserted into it in 2004 carried no expiry date. Provisions that Parliament in 2002 would enact only for three years therefore became, by the transplant of 2004, part of the permanent law of India without any further vote on their duration.

The disappearance of the sunset is the single largest structural change in the whole sequence, and it passed almost without comment at the time, because it arrived disguised as continuity. In 2002 the three year limit had been the government’s principal textual answer to the charge of authoritarian design, and the opposition had treated it as a small concession extracted under pressure. In 2004 the repealing government, elected on the promise to destroy the Act, transferred its offences into a statute with no limit at all, and the enacting opposition, by then defending the repeal’s security credentials, did not press the point. Both coalitions thereby agreed, in conduct if not in speech, on the proposition that the terrorist act offences were fit to be permanent law. The argument that remained genuinely open between them concerned procedure: how confessions, custody, bail and review should be arranged around a substantive code neither side any longer proposed to repeal. That is the argument the following decade of amendments, and the litigation under the UAPA, has actually been about.

The Afterlife: Demands to Restore POTA

POTA has never left Indian public argument, and its returns follow a rhythm the transplant analysis predicts. After every major terrorist attack, the demand is heard, in Parliament and outside it, that POTA be brought back, as though the repeal of 2004 had disarmed the state and only the restoration of the named statute could arm it again. The demand reached its greatest intensity after the attacks on Mumbai in November 2008, and the legislative answer given then is the most instructive episode in the Act’s afterlife. The government of the day did not restore POTA. It amended the Unlawful Activities (Prevention) Act again, inserting Section 43D with its extended investigation period of up to one hundred and eighty days and its stringent bail condition, and it enacted a separate statute creating the National Investigation Agency for the federal investigation of terror offences. That is to say, confronted with the strongest political pressure the repeal’s critics could generate, the state rebuilt inside the UAPA substantially the procedural code whose removal in 2004 had been the repeal’s real content, while leaving unrestored the confession provision and the disclosure offence. The state’s own conduct across 2004 and 2008 is therefore the best evidence of what the repeal meant. The name POTA was expendable in 2004 and remained expendable in 2008. The offences were indispensable throughout. The procedure was negotiable, surrendered in a season of liberties argument and recovered in a season of grief, within the same permanent statute.

The recurring demand for restoration also rests on a pair of factual confusions that this article’s record corrects, and correcting them is a service the passage and repeal history can perform for every later debate. The first confusion is that the repeal removed the terrorist act offences from Indian law, leaving a gap that attacks then exploited. No such gap existed for a single day, as the twin ordinances and the deemed commencement of the UAPA amendment establish. The second confusion is the romantic one, that POTA in its life was an effective instrument frustrated only by its repeal. The Act’s own review machinery found its most prominent prosecutions unsustainable, its flagship massacre prosecution was returned to the ordinary law by two courts, and its one completed capital prosecution ended with half its accused acquitted. A demand to restore POTA is, when analysed, a demand for three specific things and no more: the admissibility of confessions to senior police officers, a dedicated court structure, and a free standing disclosure offence, since everything else the Act contained either never left the law or returned to it in 2008. Debates about restoration would be shorter and more honest if they were conducted as debates about those three provisions, on their merits, rather than about a name that functions in public argument as a synonym for resolve.

Facts, Values and the Verdict

The disputes that survive in this controversy divide into three kinds, and separating them is the condition of judging any of them. The first kind is settled fact. The passage record, the votes of 113 to 98 and 425 to 296, the identity of the presiding officer, the dates of assent, amendment and repeal, the text of the savings and of Section 2(3), the holdings in PUCL, in the Godhra review litigation and in the Vaiko litigation, and the contents of the 2004 UAPA amendment are all matters of public record on which this account has relied and on which no serious dispute exists. The second kind is contested fact. Whether misuse of POTA was systematic or aberrational is a question of aggregate fact that the surviving record, for the reasons given in the misuse section, cannot finally answer: the decided instances establish misuse in prominent cases across two states and two kinds of prosecution, and no authoritative aggregate was ever produced against which to measure them. Whether POTA deterred attacks that would otherwise have occurred is likewise a factual question in form, asserted confidently by the Act’s defenders on the strength of the state’s survival of a grave period and denied by its critics on the strength of the attacks that occurred during its life, including the attack on the Akshardham temple in September 2002; neither side’s evidence can carry the weight placed on it, since deterrence is measured in events that did not happen. The third kind is value dispute, and no accumulation of record will settle it. How much procedural protection a democracy may suspend in the face of organised political violence, whether the directly elected House should be able to override the states’ House on a law the states’ police must operate, and whether association and expression may be criminalised in advance of any violent act, are questions about the kind of state India should be, and the POTA record informs them without deciding them.

One of the value disputes deserves a fuller statement before the verdict, because it is the dispute the passage record most directly engages. The joint sitting device decides ordinary legislation by combined arithmetic whenever the two Houses irreconcilably differ, and the POTA sitting is its strongest modern instance. The design’s defence is constitutional and coherent: Article 108 is not an oversight but a settlement, under which the House that represents the people directly, renewed at every general election, is intended to prevail over the House that represents the states on the ordinary business of the Union, the states’ check operating through debate, delay and the publicity of rejection rather than through a permanent veto. The design’s critique is equally coherent: a veto that can always be overridden by arithmetic is a check that exists only until a government decides to pay the political price of overriding it, and for laws that restrict liberty, and that state police forces must operate, the framers’ choice means that the one House in which the states are represented can be constitutionally silenced by the size of the other. The POTA record contributes one fact to that old argument. The device has been used three times in the Republic’s history, and on the third occasion the law it enacted was repealed within thirty months and its substance re-enacted by the ordinary route it had bypassed. Whatever else that sequence shows, it shows that the override purchased nothing that ordinary legislation could not have delivered two years later, which weakens the pragmatic defence of the device, that it exists for measures too necessary to wait, without touching the constitutional defence, that it exists because the settlement says so.

On the matters the record does decide, this account takes the verdict the evidence supports, and it is a verdict against both of the slogans with which the article opened. The repeal of POTA did not weaken India’s anti-terror law in substance, because the repealing government moved the substance into the UAPA on the same day and made it permanent; the charge of appeasement by repeal fails against the transplant record, and it fails on the words of the two Acts of 2004 rather than on any interpretation of motive. The enactment of POTA was not the indispensable rescue its defenders describe either, because the provisions that distinguished it from the law that replaced it were procedural, and the flagship prosecutions of its life ended, in Godhra, with the courts holding the special law inapplicable and the ordinary law sufficient, and, in the Parliament attack case, with a conviction that survived on a confession the law has never admitted since. What the joint sitting decided, in the end, was the location of a set of offences for two and a half years and the terms on which the prosecutions under it would be run. The Constitution’s rarest legislative device was used with perfect regularity and considerable political courage or recklessness, depending on the value judgment the reader brings, to produce an outcome the next Parliament substantially reproduced by ordinary means while claiming to reverse it. The intensity of a legislative fight in India, POTA therefore demonstrates, is a poor predictor of the durability of its outcome, because the fight is about names, symbols and the custody of the state’s reputation for resolve, while the outcome is decided in the drafting of definitions and savings clauses that the fight barely touches.

What the Episode Teaches About Security Legislation

Four lessons stand out of the record, and each is a lesson about drafting rather than about politics. The first concerns sunsets. The three year limit in Section 1(6) was the most consequential sentence in POTA’s design, because it converted every argument about the Act into an argument with a deadline, and it made the statute’s survival depend on a future Parliament’s willingness to renew it in the light of its working. The transplant of 2004 abolished that discipline for the transplanted provisions without debate, and later Indian security legislation has generally been permanent from enactment. A legislature that wants emergency powers to remain emergency powers must write the expiry into the text and must treat any proposal to rehouse the powers in a permanent statute as the substantive decision it is. The second lesson concerns review. The Review Committees were written into POTA in 2002 as an advisory ornament, given binding power in 2003 after the ornament proved useless, and universalised in 2004 as an audit of the whole caseload. The machinery worked when it was given teeth, and its findings in the Vaiko and Godhra cases are the most credible official assessments of the Act’s operation that exist. Review that cannot direct an outcome is not a safeguard but a decoration, and the sequence from 2002 to 2004 is a controlled experiment proving the point within a single statute. This article is legal information about a repealed statute and the legislation that succeeded it; it is not legal advice, and it is no substitute for advice on the text of the anti-terror law in force in any live matter.

The third lesson concerns definitions. The committees and courts that dismantled POTA’s prominent prosecutions did not hold the Act unconstitutional or its procedures unlawful. They held, case by case, that conduct the police had characterised as terrorism did not meet the statute’s own definition of a terrorist act or its own requirement of intent. The definition was the battlefield all along, and the prosecution’s habit of reaching for the special law in cases of communal violence, political speech and insurgent sympathy was a habit of characterisation that a narrower definition, or a requirement of independent approval before the special law was invoked, might have disciplined at the source. The fourth lesson concerns the ratchet. Powers enacted as temporary and exceptional do not retire when the emergency passes; they migrate. TADA’s devices passed into POTA after TADA’s lapse had been celebrated as their burial, and POTA’s offences passed into the permanent UAPA after POTA’s repeal had been celebrated in the same terms, with parts of its procedure following in 2008. Any honest accounting of a security statute must therefore be an accounting of the sequence, not of the single Act, because the Act is the one element of the sequence guaranteed to be temporary.

Settled, Contested and Unresolved

What is settled can be stated briefly. POTA was Act 15 of 2002, passed at the third joint sitting in the Republic’s history on 26 March 2002 by 425 votes to 296, with the Deputy Speaker, P. M. Sayeed, presiding because the Speaker’s office was vacant, and it was repealed with effect from 21 September 2004 by Act 26 of 2004 on the same day that Act 29 of 2004 wrote its substantive offences into the Unlawful Activities (Prevention) Act. The Supreme Court upheld the Act in PUCL while narrowing its offences to conduct intended to further terrorist activity, and the review machinery Parliament built in 2003 and universalised in 2004 found no prima facie case in the Act’s two most prominent prosecutions. What is contested is the scale of the Act’s misuse beyond those decided instances, the deterrent value the Act may or may not have had, and the weight to be given to the political motives on every side of its enactment and repeal. What is unresolved is the value question the episode poses in its purest available form: whether a democracy is better defended by special laws whose severity is proclaimed and temporary, or by permanent laws whose substance is the same and whose enactment attracts no fight at all. The joint sitting of 26 March 2002 was summoned to answer a question about a bill. The record it left answers a different and more useful question, about how little of a security state’s real power is ever decided in the chamber where the fighting happens.

There is a last, quieter sense in which the sitting changed nothing, and it concerns the people the Act prosecuted. The men and women whose cases fill this account, the parliamentarian held for nineteen months for a speech, the Godhra accused held for years under a label two courts withdrew, the Parliament attack accused of whom two were acquitted and one was executed, experienced the statute not as a controversy but as a sequence of custody orders, bail hearings and charge sheets. For them the difference between POTA and the law that replaced it was never the difference between the joint sitting and the repeal. It was the difference made by the machinery that decided their cases: a committee with power to direct a withdrawal, a court willing to read an intent requirement into a broad offence, an appellate bench insisting that suspicion is not proof. The legislative history of POTA is a history of two dramatic parliamentary events and one undramatic administrative invention, and it was the invention, the binding Review Committee, that touched the most lives. Legislatures that debate the next security statute would do well to argue about the invention first.

Frequently Asked Questions

Q: What was POTA and why was it enacted?

The Prevention of Terrorism Act, 2002, Act 15 of 2002, was India’s special anti-terror statute, enacted to give the state offences and procedures beyond the ordinary criminal law for dealing with terrorist activities. It defined the terrorist act, punished membership of, support for and fund raising for proscribed organisations listed in its Schedule, admitted confessions made to senior police officers, allowed up to one hundred and eighty days before a charge sheet, restricted bail, and created Special Courts and Review Committees. It was enacted in the aftermath of the September 2001 attacks in the United States and the attack on Parliament in December 2001, after the earlier anti-terror statute, TADA, had lapsed in 1995 leaving no central special law. It was deemed in force from 24 October 2001 and carried a three year life.

Q: Why was POTA passed at a joint sitting of Parliament?

The bill passed the Lok Sabha, where the governing National Democratic Alliance held a majority, on 18 March 2002, but the Rajya Sabha, where the opposition held the larger strength, rejected it on 21 March 2002 by 113 votes to 98. A bill rejected by one House ordinarily fails. Article 108 of the Constitution, however, allows the President to summon a joint sitting of both Houses where one House rejects a bill passed by the other, and the combined House decides by a majority of the members present and voting. President K. R. Narayanan summoned the sitting, and because the Lok Sabha is more than twice the size of the Rajya Sabha, the government’s lower House majority produced a combined majority of 425 to 296. It was only the third joint sitting in the history of the Republic.

Q: Who presided over the POTA joint sitting?

The Deputy Speaker of the Lok Sabha, P. M. Sayeed, presided. A joint sitting is normally presided over by the Speaker of the Lok Sabha, but that office was vacant on 26 March 2002. The Speaker, G. M. C. Balayogi, had died on 3 March 2002 in a helicopter crash in Andhra Pradesh, and his successor, Manohar Joshi, was elected only in May 2002. Under the order of precedence prescribed for a joint sitting, the chair therefore passed to the Deputy Speaker, and it was Sayeed who conducted the sitting and declared the bill passed by 425 votes to 296. Accounts that name a Speaker as the presiding officer of the POTA sitting are in error, because no Speaker held office on the day.

Q: Did the Supreme Court uphold POTA?

Yes. In People’s Union for Civil Liberties v Union of India, decided on 16 December 2003, the Supreme Court dismissed the constitutional challenges and upheld the Act, holding among other things that the mere possibility of a statute being abused is not a ground for striking it down. The upholding came with important clarifications. The offences of membership, support and fund raising in Sections 20, 21 and 22 were read as requiring an intention to further or facilitate terrorist activity, so that sympathy or speech without that intent is no offence. The confession provision in Section 32 was sustained subject to its safeguards, including production before a Magistrate within forty-eight hours and medical examination on a complaint of torture. The restrictive bail provisions in Section 49 were also upheld.

Q: What were the main allegations of POTA misuse?

The allegations fell into three groups. In Gujarat, critics alleged that the Act was applied overwhelmingly against Muslims, including in the prosecution of the Godhra train burning accused and in detentions arising from the violence of 2002, a charge the state government denied. In Tamil Nadu, the arrest of the MDMK leader Vaiko for speeches supporting the LTTE was alleged to be the use of an anti-terror law against political speech, by a state government against the leader of a former ally. In Jharkhand and other states facing the Maoist insurgency, human rights organisations documented arrests of alleged sympathisers, including, on their reporting, elderly persons and minors. Underlying all three groups was the structural allegation that the Act’s broad definition, long custody periods and restrictive bail standard made such use predictable rather than accidental.

Q: Why was POTA repealed in 2004?

The United Progressive Alliance, which won the general election of 2004 under Prime Minister Manmohan Singh, had committed itself in its Common Minimum Programme to repeal the Act, presenting repeal as the remedy for a law it held responsible for widespread misuse against minorities and political opponents. The repeal was carried out by ordinance on 21 September 2004 and then enacted as the Prevention of Terrorism (Repeal) Act, 2004. Two qualifications belong to the record. The Act was in any event within a month of expiring under its own three year sunset clause, so the repeal anticipated a scheduled death. And the repealing government simultaneously amended the Unlawful Activities (Prevention) Act to carry POTA’s substantive offences forward, so the repeal was directed at the Act’s name, its procedure and its political history rather than at the offences themselves.

Q: What happened to pending POTA prosecutions after the repeal?

They did not collapse. Section 2(2) of the repeal Act saved the previous operation of POTA, so investigations and proceedings already begun could continue as if the Act had not been repealed, although no court could take cognizance of a POTA offence more than one year after the repeal commenced. Section 2(3) then subjected every case registered under POTA to review by the Central Review Committee within one year, whether or not the accused applied for review. Where the committee found no prima facie case, cases in which cognizance had been taken were deemed withdrawn, and pending investigations stood closed, in each case from the date of the committee’s direction. The Supreme Court upheld that scheme in 2008, holding that the deemed withdrawal operated by force of the statute itself. Cases the committee allowed to stand continued under the saved provisions.

Q: Did repealing POTA actually reduce anti-terror powers?

In substance, very little. On the same day the repeal took effect, 21 September 2004, the Unlawful Activities (Prevention) Amendment Act inserted POTA’s core into the UAPA: the definition of a terrorist act, the punishments, the offences of membership, support and fund raising, the forfeiture machinery and the Schedule of proscribed organisations all continued, and because the UAPA is permanent legislation they continued without the three year limit POTA had carried. What the repeal genuinely removed was procedural: the admissibility of confessions made to police officers, the one hundred and eighty day investigation period, the stringent bail condition, the disclosure offence, the Special Courts and the Review Committees. Some of that procedure was restored to the UAPA by amendment in 2008. The net reduction in the state’s substantive anti-terror power was close to zero.

Q: When was POTA repealed, and by which Act?

POTA was repealed with effect from 21 September 2004 by the Prevention of Terrorism (Repeal) Act, 2004, Act 26 of 2004. The repeal was first effected by an ordinance promulgated on that date, as one of a pair of ordinances issued together, the second of which amended the Unlawful Activities (Prevention) Act to receive POTA’s provisions, and Parliament subsequently enacted the repeal in statutory form with the same effective date. The repeal anticipated the Act’s own timetable by about a month. Section 1(6) of POTA had limited its life to three years from 24 October 2001, so the statute would have expired in October 2004 even if the government of the day had done nothing at all.

Q: Why did POTA begin as an ordinance in 2001?

Parliament was not in session in October 2001, and the President may legislate by ordinance under Article 123 when Parliament is not sitting and circumstances require immediate action. The government stated in the Statement of Objects and Reasons that the security situation, an upsurge of terrorist activity and the intensification of cross-border terrorism, rendered immediate legislation necessary, and the Prevention of Terrorism Ordinance was promulgated on 24 October 2001. The choice was contested at the time: Parliament was due to reconvene within weeks for the winter session, and critics argued that a statute of this consequence had been routed around parliamentary scrutiny and around consultation with the National Human Rights Commission. When the winter session ended without a replacing bill, a second ordinance followed on 30 December 2001.

Q: What did the 2003 amendment to POTA change?

The Prevention of Terrorism (Amendment) Act, 2003, Act 4 of 2004, first promulgated as an ordinance on 27 October 2003, strengthened the Review Committees created by Section 60 of the Act. As originally enacted, the committees could not end a prosecution. The amendment inserted sub-sections (4) to (7) into Section 60, allowing any aggrieved person to apply for a review of whether a prima facie case existed for proceeding against him. The committee’s directions were made binding on the Central Government, the State Government and the investigating police officer. Where the Central Committee and a State Committee had both reviewed the same offence and differed, the Central Committee’s direction prevailed. Where no prima facie case was found, pending proceedings were deemed withdrawn from the date of the direction. The power was first prominently used in the direction ending the prosecution of Vaiko in April 2004.

Q: How did POTA define a terrorist act?

Section 3(1) defined a terrorist act by the intent with which specified means were used. The act had to be done with intent to threaten the unity, integrity, security or sovereignty of India, or to strike terror in the people or in any section of the people, using bombs, dynamite or other explosive or inflammable substances, lethal weapons, poisons, noxious gases, chemicals or other hazardous substances, in a manner causing or likely to cause death or injury, loss of or damage to property, disruption of supplies or services essential to the life of the community, or damage to the monetary stability of India through the production, smuggling or circulation of high quality counterfeit currency. Detaining a person, coupled with threats to kill or injure in order to compel a government or an international organisation to act or abstain from acting, was also brought within the definition.

Q: How long could a person be held under POTA before a charge sheet was filed?

Up to one hundred and eighty days. Section 49(2) of the Act modified Section 167 of the Code of Criminal Procedure for POTA cases: the investigating agency had ninety days to complete its investigation, and the Special Court could extend that period to one hundred and eighty days on the report of the Public Prosecutor indicating the progress of the investigation and the specific reasons for the detention of the accused beyond ninety days. The same provision enlarged the period within which police custody could be taken and required an affidavit stating reasons where police custody was sought over a person already in judicial custody. If no charge sheet was filed within the period as extended, the accused became entitled to be released on bail for default of the prosecution.

Q: What safeguards did the Supreme Court read into POTA in the PUCL case?

In People’s Union for Civil Liberties v Union of India, decided in December 2003, the Court upheld the Act only with clarifications that narrowed it. Sections 20, 21 and 22 were read as requiring a positive inference that the person acted with intent to further, encourage, promote or facilitate terrorist activity, so that association or speech without that intent is no offence. For confessions under Section 32, the Court treated production before a Magistrate within forty-eight hours as an operative safeguard: the Magistrate is to enquire into any complaint of torture and provide for medical examination and treatment, a confession extracted by threat or inducement is a nullity in law, and the trial court must decide the admissibility of a confession for reasons. The bail restrictions in Section 49 were upheld as not unreasonable for offences of the complexity the Act addressed.

Q: Could a person be prosecuted under POTA for failing to disclose information about a terrorist act?

Yes. Section 14 of the Act imposed an obligation to furnish information on specified categories of persons, on the written requisition of an officer not below the rank of Superintendent of Police. Failure to furnish the information, or furnishing information known to be false, was itself an offence punishable with imprisonment for a term extending to three years, or with fine, or with both, and the offence was triable summarily. The provision was among the most criticised in the Act, because it made silence punishable outside the witness box and bore directly on journalists, lawyers, doctors and family members who might hold information about suspects. It was not carried into the Unlawful Activities (Prevention) Act when POTA’s substantive provisions were transferred there in 2004.

Q: What did the Review Committee decide in the Godhra case?

The Central POTA Review Committee, headed by Justice S. C. Jain, reviewed the cases of the accused who had been charge sheeted under POTA for the Godhra train burning of 27 February 2002 and, by its opinion of 17 May 2005, held that in its opinion the accused might not be tried under POTA, though they might be tried under other laws, and it directed the Public Prosecutor to apply without delay for withdrawal of the POTA charges. The Gujarat government supported the victims’ challenge to that outcome. The Supreme Court upheld the binding review machinery of the repeal Act in October 2008, and the Gujarat High Court afterwards upheld the committee’s recommendation, holding that the incident, though a shocking and serious criminal act, fell short of the requirement of an act of terror on the material available.

Q: Why was Vaiko’s arrest under POTA controversial?

Vaiko, the leader of the Marumalarchi Dravida Munnetra Kazhagam and a Member of Parliament, was arrested in July 2002 over speeches at a public meeting at Thirumangalam in which he had expressed support for the Liberation Tigers of Tamil Eelam, a proscribed organisation, and he was charged under Sections 21(2) and 21(3) of POTA, the support offences. The prosecution was brought by the All India Anna Dravida Munnetra Kazhagam government of Tamil Nadu, whose party had been an ally of Vaiko’s, under a statute the opposition had warned could be turned against political speech, and the case became the standing illustration of that warning. He remained in custody for roughly nineteen months before his release on bail in February 2004. The Central Review Committee held in April 2004 that the speeches disclosed no prima facie case, sympathy for the cause of the Eelam Tamils not amounting to support for the LTTE within the meaning of the section.

Q: Was POTA used in the Parliament attack prosecutions?

Yes. The attack of 13 December 2001, which had supplied the immediate case for the statute, was prosecuted under it. The charge sheet was filed in May 2002 before a Special Court designated under POTA, and in December 2002 that court sentenced Mohammad Afzal, Shaukat Hussain Guru and S. A. R. Geelani to death and Afsan Guru to five years’ imprisonment. The Delhi High Court acquitted Geelani and Afsan Guru in October 2003 and confirmed the other two death sentences. The Supreme Court, in August 2005, confirmed Afzal’s death sentence, relying heavily on his confession recorded under Section 32 of POTA together with the circumstantial evidence, reduced Shaukat Hussain Guru’s conviction to one of concealing the conspiracy, with ten years’ imprisonment, and upheld the acquittals. Because the repeal Act saved pending proceedings, the repeal did not disturb the case.

Q: What happened to people who had already been convicted under POTA when the Act was repealed?

Their convictions stood. The savings in Section 2(2) of the repeal Act preserved the previous operation of POTA, the penalties incurred under it and the proceedings in respect of them, so sentences continued to be executed and appeals continued to be heard and decided under the repealed Act as if it had not been repealed. The Review Committee machinery in Section 2(3) addressed cases still at the stage of investigation or trial, not concluded convictions. The most serious instance is the Parliament attack case, in which the Supreme Court decided the appeals in August 2005, after the repeal, confirming a death sentence imposed under POTA, and that sentence was carried out in February 2013 following the rejection of a mercy petition.

Q: Did POTA apply to acts committed before the Act was passed?

In part. POTA received the President’s assent on 28 March 2002, but Section 1(6) deemed it to have come into force on 24 October 2001, the date of the first Prevention of Terrorism Ordinance, which it replaced and which had itself operated from that date. Conduct on or after 24 October 2001 could therefore be prosecuted under the regime continuously, first under the ordinance and then under the Act. Conduct before that date could not be prosecuted under POTA at all: Article 20(1) of the Constitution bars conviction for an act that was not an offence under a law in force when it was committed, and neither the ordinance nor the Act purported to reach back beyond the ordinance’s commencement. The attack on Parliament of 13 December 2001 fell within the regime because it postdated the ordinance.