The Terrorist and Disruptive Activities (Prevention) Act, the statute everyone calls TADA, was Parliament’s answer to a genuine emergency, and it was written to die young. The TADA Act 1987 carried its own expiry date in its first section: two years of life, and no more. A sunset clause of that kind is a promise made in advance. It tells the citizen that extraordinary powers are being taken for a stated season, that the season will end, and that anyone who wants the powers continued must return to the legislature and argue for them in the open. TADA kept that promise in form and broke it in substance. The two-year statute was renewed in 1989, renewed again in 1991, and renewed again in 1993, so that a law sold to Parliament as a brief emergency measure governed Indian criminal justice for ten years, from 1985 to 1995, and its prosecutions ran for decades beyond its death.

The history of TADA is therefore two histories at once. It is the history of a set of legal devices: a definition of terrorism wide enough to cover a speech, a confession rule that reversed a century of evidence law, Designated Courts sitting behind closed doors, a bail test turned on its head, presumptions that did the prosecution’s work for it. And it is the history of a renewal machine, a sequence of parliamentary decisions, each individually defensible, which together converted a temporary exception into a standing feature of the legal order. To understand how Indian anti-terror law developed afterwards, through POTA and into the UAPA amendments, the second history matters as much as the first. The sunset that never set is the claim that organises everything that follows, and the renewal record is the evidence for it.

One discipline should be declared before the record is opened, because TADA’s numbers have been ammunition in Indian politics for three decades. Wherever this history gives a figure for arrests, convictions or prisoners, the figure is attributed to the organisation that published it, and where reputable sources disagree, the disagreement is reported rather than resolved. The statute’s own governments never produced a definitive official account of its use, and that absence is itself part of the evidence: a law renewed three times on the strength of necessity was never, in ten years, subjected by its owners to a settled statistical audit. The reader should therefore treat every precise-sounding TADA total, from any quarter, including the quarters sympathetic to this history’s conclusions, with the suspicion that the absence of an audit was designed to licence.

The Punjab Emergency and the Road to a Special Law

The statute cannot be understood apart from the violence that produced it. Punjab in the early 1980s was descending into an armed separatist insurgency centred on the demand for Khalistan. The movement had a charismatic centre in Jarnail Singh Bhindranwale, fortified inside the Golden Temple complex at Amritsar, and a widening campaign of assassinations, bank robberies, and attacks on officials, journalists and ordinary travellers. In June 1984 the Union government sent the army into the Golden Temple in Operation Blue Star. Bhindranwale was killed; the temple complex was scarred; and the political consequences ran far beyond Punjab. On 31 October 1984, Prime Minister Indira Gandhi was assassinated by two of her own Sikh bodyguards, and the assassination was followed by organised anti-Sikh violence in Delhi and elsewhere in which thousands were killed. Rajiv Gandhi’s government, elected in December 1984 with an enormous majority, inherited a state whose monopoly on force in Punjab was contested and whose ordinary criminal process was visibly failing to answer an armed campaign.

The failure was not only one of firepower. The Punjab insurgency of those years operated through intimidation of witnesses, assassination of officials, and attacks timed to humiliate the state, and India’s ordinary procedural law had been designed for a different world. Investigation depended on witnesses who could be terrorised into silence. Trials depended on public proceedings in courts that could not protect the people inside them. Sentencing law treated an assassination, a bank robbery and a bombing as ordinary crimes differentiated only by result. The government’s case, stated at its strongest, was that a campaign of this character could not be fought with the ordinary tools of the Indian Penal Code and the Code of Criminal Procedure, and that a government which refused to legislate specially was choosing abdication over constitutional difficulty. That case had force in 1984 and 1985, and any honest account of TADA has to concede it before recording what the special law became.

Parliament had, in fact, already taken one step before TADA. The Terrorist Affected Areas (Special Courts) Act 1984 had created special courts for scheduled offences in areas declared terrorist affected, an admission that the trial forum itself needed redesign. What the 1984 Act did not do was define terrorism, create new substantive offences, or alter the law of evidence and bail. The killings of 1985 closed the remaining distance. When the Statement of Objects and Reasons of the first TADA Bill described the situation, it pointed to terrorist violence that had begun in Punjab and Chandigarh and had spread to Delhi, Haryana, Uttar Pradesh and Rajasthan, with explosions in transit points and crowded places from 10 May 1985 onwards, carried out by persons the statement called heartless and desperate. A statute that defined the terrorist act, armed police and prosecutors with special procedure, and confined itself, on paper, to two years of operation was the result.

The insurgency’s own trajectory across the decade forms the backdrop against which every renewal was argued, and it is worth tracing because the renewals did not track it in any simple way. The late 1980s saw the campaign at its most lethal, with massacres of bus passengers and targeted killings of officials, journalists and political workers, and a second army and police clearing of the Golden Temple complex in Operation Black Thunder in 1988. The early 1990s saw the counter-insurgency prevail: intensified police operations under the Punjab Police leadership, the return of an elected state government after the 1992 elections, and the progressive collapse of the militant organisations’ capacity. By 1993 and 1994, informed opinion across the political spectrum treated the Punjab insurgency as defeated. Yet the statute built for Punjab was renewed in 1991 without serious difficulty and again in 1993, after the insurgency’s back had been broken, on the strength of a bombing campaign in Bombay conducted by an organised crime and smuggling network with external sponsors. The emergency, in other words, had changed its address, and the statute followed the emergency rather than the other way round. That mobility, the capacity of a special law to migrate from the theatre that justified it to whatever theatre next frightens the republic, is the deepest structural fact in TADA’s history, and it explains why the argument about the Act was never really an argument about Punjab.

The 1985 Act: A Two-Year Statute

The first TADA was the Terrorist and Disruptive Activities (Prevention) Act 1985, Act 31 of 1985. It received the assent of the President on 23 May 1985 and came into force on 24 May 1985. Its duration clause did the work that defines the whole story: the Act was to remain in force for a period of two years from its commencement. The substance of the later, harsher 1987 statute was already present in outline. The Act defined the terrorist act and the disruptive activity, created Designated Courts, modified the Code of Criminal Procedure in its application to such cases, restricted bail, and gave the statute overriding effect over inconsistent laws. It applied to the whole of India, with one carve-out that did not last the year: as enacted, the provisions creating the terrorist act offence did not extend to the State of Jammu and Kashmir. The omission was removed within weeks by the amending Act 46 of 1985, and the Act applied to Jammu and Kashmir from 5 June 1985. Even the geography of exception, in other words, expanded almost immediately.

The 1985 Act’s internal architecture is worth a moment more, because it shows how much of the supposed escalation of 1987 was present from the first draft of the idea. Its opening sections defined the offences and their punishments. A middle band constituted the Designated Courts, regulated their procedure and powers, provided for public prosecutors to conduct the cases before them, and dealt with transfers between the special and ordinary courts. Its later sections modified the application of the Code of Criminal Procedure to TADA proceedings and provided for appeals. The scheme consciously overlapped with the Terrorist Affected Areas (Special Courts) Act 1984, which continued to exist beside it, and the two statutes were eventually challenged and upheld together. What the 1985 Parliament had built, in other words, was not an experiment in one device but the whole institutional world at once: offences, forum, procedure, evidence and appeal. The 1987 Act’s contribution was to sharpen the devices inside a world that already existed. This matters for the renewal argument, because it means the legislature’s real decision about TADA’s design was taken once, in the emergency spring of 1985, and every later Parliament, in 1987, 1989, 1991 and 1993, was deciding only whether to continue a world already built, which is a far easier decision to take lazily than the decision to build it would have been.

The two-year limit deserves attention because it was not decorative. A sunset clause in a security statute performs three functions. It forces the executive to justify continuation with evidence rather than momentum. It gives the legislature a scheduled occasion to examine the record of use, including misuse, before the powers are prolonged. And it signals to courts, which are asked to sustain extraordinary provisions against constitutional challenge, that the legislature itself treats the measures as exceptional and temporary. The Supreme Court’s eventual defence of TADA in Kartar Singh drew real support from the statute’s temporary character: a law Parliament kept under periodic review was easier to present as a proportionate response than a permanent code would have been. The irony that runs through the decade is that the sunset clause performed its judicial function, helping the Act survive challenge, while failing its legislative function, because each renewal debate became shorter, more routine and less evidentiary than the last.

Events kept supplying the government’s justification while the ink dried. The explosions the Statement of Objects described, in buses, markets and transit points across Delhi and the neighbouring states from 10 May 1985, were the domestic face of the emergency. Within a month of commencement, on 23 June 1985, the bombing of Air India Flight 182 over the Atlantic killed 329 people, an atrocity attributed to Sikh extremist networks and treated in India as proof that the Punjab campaign had acquired an international reach that ordinary policing could not answer. In July 1985 the government signed the Punjab Accord with the Akali Dal leader Harchand Singh Longowal, an attempt to settle the political dispute beneath the violence; in August 1985 Longowal was assassinated by militants who rejected the settlement. The statute and the settlement were the two arms of the same policy, and the assassination of the settlement’s author was read, fairly or not, as evidence that the coercive arm had to be kept strong precisely because the political arm was fragile. A Parliament legislating in that atmosphere was never going to let the two-year clock become an argument about first principles, and the renewals that followed inherited the atmosphere even after the Accord generation had passed.

The 1985 Act’s internal life was brief because its expiry was real. Two years from 24 May 1985 meant 23 May 1987. As that date approached, the Punjab insurgency had not ended; if anything the violence had deepened, and the government of the day faced exactly the choice the sunset clause was designed to force. It could let the Act die and fight on with ordinary law. It could seek a fresh enactment with a new justification. Or it could continue the regime with the least possible interruption. It chose continuation, and the manner of continuation set the pattern for the decade: first an ordinance to bridge the gap, then a new Act that was presented as a replacement and functioned as an intensification.

The TADA Act 1987: Expiry, an Ordinance and the Remaking

When the 1985 Act reached its expiry on 23 May 1987, Parliament was not in session, and the government was unwilling to accept even a short interval in which the special regime did not exist. The President accordingly promulgated the Terrorist and Disruptive Activities (Prevention) Ordinance 1987, Ordinance 2 of 1987, on 23 May 1987, and it came into force on 24 May 1987, the day after the old Act expired. The ordinance device mattered. Article 123 of the Constitution permits ordinances when Parliament is not in session and circumstances require immediate action, but an ordinance is a holding device that must be laid before Parliament and replaced by an Act. The 1987 sequence therefore had two steps: the ordinance preserved the regime without a day’s break, and Parliament then debated and passed a full statute during the following session. When the Terrorist and Disruptive Activities (Prevention) Act 1987, Act 28 of 1987, received presidential assent on 3 September 1987, its commencement clause reached back: most of its provisions were deemed to have come into force on 24 May 1987, while four of its most consequential sections, Sections 5, 15, 21 and 22, came into force at once on assent. There was no interregnum. The sunset of the first Act had set on schedule, and the regime had continued without missing a day.

The 1987 Act was not a mere reprint of its predecessor, and calling it a renewal understates what Parliament did that September. It was a fresh enactment, passed as a new statute with its own duration clause, and it strengthened the regime at several points that defined TADA for the rest of its life. Section 5, the unauthorised possession offence, made the mere possession of specified arms, ammunition or explosives in a notified area a TADA offence by itself, punishable with a minimum of five years and up to imprisonment for life, without any need to prove a terrorist act or an intention to commit one beyond the possession itself. Section 15 restated the confession rule in the form in which it became the most argued-about provision in Indian evidence law, making a confession recorded by a senior police officer admissible against its maker and, where they were charged and tried together in the same case, against a co-accused, abettor or conspirator. Section 21 armed the prosecution with presumptions. Section 22 gave identification by photograph the same legal value as a test identification parade, a provision the Supreme Court would later strike down as unintelligible and prone to grave injustice. The statute Parliament passed in 1987 was therefore the real TADA: the 1985 Act was its draft, and the 1987 text is the one the courts, the police and the victims of misuse lived under for the next eight years.

The ordinance bridge deserves a constitutional footnote, because it normalised a method. An ordinance under Article 123 must be laid before Parliament when it reassembles and ceases to operate six weeks from reassembly unless approved, so the 1987 ordinance was always a promissory note for a statute. The government honoured the note within the same year. But the sequence, expiry at midnight, ordinance the same week, statute within the session, taught every later government the choreography by which a lapsed security law could be continued without ever facing the clean question a sunset clause is meant to pose: should this exist at all? POTA itself began as an ordinance in October 2001 before its troubled passage through a joint sitting, and the habit of treating the ordinance as the natural first draft of emergency legislation is one of TADA’s quieter legacies. The 1987 Parliament did debate the replacement Bill, and the opposition criticism of that session anticipated every theme of the following decade: the breadth of the disruptive activity definition, the confession rule’s break with 1872, and the danger that a Punjab statute would become an all-India police instrument. The Bill passed substantially as drafted. The debate is worth remembering chiefly for how completely it predicted the record, and how little the prediction availed.

Its duration clause repeated the original promise. Section 1(4) provided that the Act would remain in force for two years from 24 May 1987, and attached to that clause the savings language that would outlive the statute: expiry would not affect the previous operation of the Act, anything duly done under it, any penalty incurred, or any investigation, legal proceeding or remedy in respect of rights and liabilities acquired while it lived, and all such proceedings could be instituted, continued and enforced as if the Act had not expired. Two features of that drafting deserve to be held together. The two-year clock created the renewal cycle that dominates the middle of this history. The savings language created the long tail that dominates its end, because it meant that the Act could die in 1995 and yet its trials, its convictions and its prisoners would continue, lawfully, for decades. Parliament in 1987 was building both the machine that had to be renewed and the mechanism that made renewal, in one sense, unnecessary: the cases already inside the system would finish under TADA whatever happened to the statute.

The Offences: Terrorist Act and Disruptive Activity

The scope of a special law is set by its definitions, and TADA’s definitions were drawn wide from the first line. Section 3 defined the terrorist act through intention and means together. Whoever, with intent to overawe the Government as by law established, or to strike terror in the people or any section of the people, or to alienate any section of the people, or to adversely affect the harmony amongst different sections of the people, used bombs, dynamite, other explosive or inflammable substances, lethal weapons, poisons, noxious gases, chemicals or other hazardous substances, in a manner that caused or was likely to cause death, injury, loss or damage to property, disruption of supplies or services essential to the life of the community, or the detention of a person coupled with threats to kill or injure that person to compel the Government or another person, committed a terrorist act. The punishment structure matched the rhetoric of emergency. Where the act resulted in the death of any person, the offence was punishable with death or imprisonment for life. In any other case the sentence ran from a minimum of five years to imprisonment for life, with fine in addition. Conspiracy, attempt, advocacy, abetment, incitement and facilitation were folded into the same punitive band, and harbouring or concealing a terrorist carried the same minimum of five years and the same ceiling of life.

Two elements of that definition did quiet but decisive work. The first was the intention list. Alongside overawing the Government and striking terror, Parliament included alienating a section of the people and adversely affecting harmony among sections, intentions that describe effects on social feeling as much as designs on the state, and that are correspondingly hard to prove or disprove by external evidence. The second was the means list, which reached beyond explosives to any hazardous substance and beyond killing to disruption of supplies and services essential to the community. Read together, the definition made the terrorist act turn on the prosecutor’s characterisation of motive as much as on the physical act itself, and because the procedural consequences of a TADA charge were so severe, the characterisation itself became the real punishment decision. A charge sheet that invoked Section 3 moved the accused into a different legal universe: a different court, a different bail test, a different law of confessions and a different investigation clock.

A further feature of the charging practice shaped everything the statistics later showed. A bombing or an assassination was already murder, conspiracy and waging war under the Indian Penal Code, and prosecutors habitually charged the ordinary offences and the TADA offences together, so that the Designated Court tried one act under two codes at once. The TADA label was thus, in many files, an overlay rather than a substitute: it changed the forum, the clock, the bail test and the evidence rules for a prosecution whose underlying facts the ordinary law already fully criminalised. Two consequences followed. First, the decision to invoke the overlay was where the system’s discretion actually lived, and it lived with the police at the moment of the FIR. Second, the conviction arithmetic needs care in both directions. Many accused who were acquitted of the TADA charge were convicted of the underlying Penal Code offence in the same trial, so the very low TADA conviction rate overstates the number who walked free entirely; and many others were acquitted of everything after years of custody, so the rate equally understates nothing about the human cost of the overlay for those it captured and then released. Honest use of the figures requires holding both corrections together, which neither the Act’s defenders nor its abolitionists consistently did.

The 1993 amendment rounded out the offence structure at its edges. Section 3(5) made membership of a terrorist gang or organisation involved in terrorist acts itself an offence within the same punitive band, and Section 3(6) reached property derived from or obtained through a terrorist act. Around the two headline offences stood a ring of supporting provisions that widened the Act’s daily usefulness to investigators. Section 6 raised the penalties for contraventions of the Arms Act 1959, the Explosives Act 1884, the Explosive Substances Act 1908 and the Inflammable Substances Act 1952 where the contravention was committed with intent to aid a terrorist or a disruptionist. Section 7 allowed the Central Government to confer the powers of police officers on its own officers for the Act’s purposes, and Section 7A, inserted later into the scheme, empowered an investigating officer, with the approval of the Superintendent of Police, to seize or attach property reasonably believed to be derived from a terrorist act, with a duty to inform the Designated Court within forty-eight hours. Section 8 completed the property regime by providing for forfeiture of the property of a person convicted under the Act. None of these provisions attracted the argument that Sections 3, 4, 5 and 15 attracted, but together they explain the statute’s texture as working law: it was not only a definition of terrorism with a trial attached, but a complete investigative code covering membership, money, weapons, seizure and forfeiture, and it could be entered at any of those doors.

Section 4 went further, and it is Section 4 that explains why TADA escaped the boundaries of counter-terrorism altogether. The disruptive activity was defined in Section 4(2) as any action taken, whether by act or by speech or through any other media, which questions, disrupts or is intended to disrupt the sovereignty and territorial integrity of India, or which is intended to bring about or support any claim for the cession of a part of India or the secession of a part of India from the Union. An Explanation widened it again, deeming advocacy or incitement of the killing of persons bound by oath to uphold the sovereignty and integrity of India to be a disruptive activity. The punishment, in the Act as amended, ran from a minimum of five years to imprisonment for life. The drafting contains a paradox that critics pressed for a decade. A law justified by bombs and assassinations criminalised, at the same punitive level, an action taken by speech. The verb list begins with the weakest imaginable form of conduct: to question. A speech that questioned the territorial settlement of the Union, without violence, without conspiracy, without an audience shown to be moved, could be charged as a disruptive activity and tried before a Designated Court under TADA procedure.

How did speech become a punishable disruptive activity under Section 4?

Section 4(2) defined a disruptive activity to include any action taken by act, by speech or through any other media that questions the sovereignty and territorial integrity of India. Because the definition named speech directly, no violent act was needed: the words themselves, charged as questioning or supporting secession, completed the offence.

The breadth was defended, in its strongest form, as a response to the way separatist campaigns actually worked. The Punjab and Kashmir movements of the period were sustained by propaganda, fundraising and recruitment as well as by arms, and the government’s argument was that a statute confined to completed violence would arrive too late in the chain. Critics answered that the provision inverted the proper order of a criminal law: the further an offence moves from physical harm toward expression, the more precisely it must be drafted, and Section 4 moved in the opposite direction, pairing the vaguest conduct with the harshest procedure. The history of the Act’s use gave the critics most of their material. The charge of disruptive activity, like the arms-possession charge under Section 5, could be laid by a police force against conduct far removed from any bomb, and once laid it carried the accused into custody under Section 20’s extended clock and reversed bail test. The definitions, in short, were the gateway through which the whole apparatus entered ordinary policing in places far from Punjab, and no account of the Act’s misuse is complete without them.

The Supreme Court, when the definitions were challenged in Kartar Singh, sustained them by locating them within the Constitution’s own vocabulary of permissible restriction. Speech directed at the cession or secession of Indian territory, the bench reasoned, fell outside the protection of the free speech guarantee because the interests of the sovereignty and integrity of India are themselves a ground of reasonable restriction written into Article 19(2), and a legislature confronting armed secession was entitled to strike at the advocacy that sustained it. The reasoning is coherent on its own terms, and it has been the template for the defence of speech-adjacent offences in Indian security law ever since. What it could not answer, and did not attempt to answer, was the administrative question: whether a definition that made questioning sovereignty an offence would be operated, by the police forces of a dozen states, with the discrimination the paradigm case assumed. The definition’s other working terms deserve note for the same reason. A terrorist, in the Act’s circular shorthand, was a person who indulged in terrorist or disruptive activities; a notified area was simply an area the State Government chose to notify for the purposes of Section 5, which meant that the boundary between an ordinary Arms Act recovery and a TADA offence carrying a five-year minimum was a government notification, issued without legislative debate and revocable at will. The Kartar Singh bench commended periodic review of those notifications precisely because it saw the risk: a power to create zones of exception by executive signature, inside a statute already excepted from ordinary law, is exception squared, and its exercise was never going to be self-limiting.

Section 15: The Confession Rule That Changed Evidence Law

Every special law has one provision that carries its philosophy, and in TADA that provision was Section 15. Indian evidence law had, for more than a century, treated police custody as a place where confessions go to be manufactured. Sections 25 and 26 of the Indian Evidence Act 1872 provide that no confession made to a police officer shall be proved against an accused person, and that a confession made in police custody is inadmissible unless made in the immediate presence of a Magistrate. The reasoning of 1872, recorded in the framing of those provisions, was institutional rather than sentimental: the police officer who investigates has an interest in the confession, the custody that produces it is invisible to any court, and the only safe rule is a rule of exclusion. Generations of Indian trials were conducted inside that rule. A confession could convict only if it was made to a Magistrate under the safeguards of the Code, with warnings, reflection time and a certificate of voluntariness.

Section 15 of TADA displaced the rule for TADA trials in a single sentence of remarkable economy. Notwithstanding anything in the Code of Criminal Procedure or in the Indian Evidence Act, it provided, a confession made by a person before a police officer not lower in rank than a Superintendent of Police, and recorded by that officer either in writing or on a mechanical device such as a cassette, tape or sound track, would be admissible in the trial of that person, or of a co-accused, abettor or conspirator charged and tried in the same case, for an offence under the Act or its rules. The provision layered its own safeguards over the displacement. Under Section 15(2), before recording a confession the officer had to explain to the person that he was not bound to make a confession and that, if he did, it could be used as evidence against him; and the officer was not to record any confession unless, on questioning the person, he had reason to believe it was made voluntarily. The TADA Rules added the working detail in Rule 15: the confession was to be recorded in the language in which it was made, read back or played back to its maker, signed by him, and certified by the officer in a prescribed memorandum recording his belief that it was voluntary; and every recorded confession was to be sent forthwith to the Chief Metropolitan Magistrate or the Chief Judicial Magistrate of the area, who would forward it to the Designated Court trying the case.

The case for the provision, put at its strongest by the governments that defended it, ran as follows. Terrorist offences are committed in secrecy by organisations that murder witnesses. The ordinary route to a confession, before a Magistrate, was said to be practically unavailable because an accused held in the middle of an ongoing conspiracy investigation would rarely speak in a Magistrate’s chamber, while the knowledge in his possession was exactly what investigators needed to prevent the next attack and to dismantle the network. The safeguard, on this argument, was relocated rather than abolished: the recording officer had to be a Superintendent of Police, a senior officer with rank to lose; the warning and the voluntariness inquiry were written into the section; the Rules required prompt transmission to a Magistrate; and the trial court retained the duty to test voluntariness before acting on the statement. A confession under Section 15 was not self-proving. It had to survive a judicial inquiry into the circumstances in which it was taken, and the Supreme Court in Kartar Singh added further conditions of its own.

The recording ritual itself repays a close look, because the whole controversy is contained in it. Rule 15 required the confession to be recorded in the language in which it was made or, if that was not practicable, in the language the person understood, with the memorandum of voluntariness appended in the officer’s own certification: his recorded belief that the confession was voluntary, that he had explained the person was not bound to make it and that it might be used against him, and that it was recorded in his presence and hearing and read over to its maker, who was found to have recorded it correctly. The mechanical recording the section permitted, by cassette, tape or sound track, was strikingly modern drafting for 1987, an attempt to substitute technological verifiability for the Magistrate’s presence: the tape could not prove voluntariness, but it could prove words, sequence and demeanour of speech, and it denied the later dispute about what had been said. The courts enforced the ritual’s formal elements with real strictness in one respect at least: the rank requirement was treated as jurisdictional, so that a confession recorded by an officer below the rank of Superintendent of Police was not a Section 15 confession at all, and delay or failure in transmitting the recorded statement to the Chief Judicial Magistrate as the Rules commanded exposed the confession to exclusion. What no formality could settle was the question the ritual was designed to avoid, whether months of custody under the Act’s own clock had made the recorded voluntariness true in any sense the law of 1872 would have recognised.

The case against the provision was the case of 1872 restated. Rank does not abolish interest: a Superintendent of Police is the officer responsible for the investigation and its results in his district, and the warning he administers is administered by the very authority the warning is meant to guard against. A mechanical recording proves that words were spoken, not that they were spoken freely after days or weeks in police custody under Section 20’s extended remand. And the extension of the confession to co-accused tried in the same case multiplied its power in the way that mattered most: one confession, extracted from one prisoner, could be read against every other person in the dock, each of whom had no opportunity to cross-examine its maker as a witness. The Nineteen Eighties debate was therefore not really about whether terrorists should be convicted. It was about whether the epistemology of the police station, distrusted since 1872 for documented reasons, could be rehabilitated by rank, tape and certificate. Section 15 answered yes. The controversy the answer generated followed the section for the rest of the Act’s life, into Kartar Singh, and out of TADA into every anti-terror statute drafted afterwards, because POTA copied the device and the UAPA debate has never escaped its shadow.

Admissibility, it should be added, was never the same as conclusiveness, and the distinction structured the trials. A Section 15 confession could be retracted at trial, and routinely was; the Designated Court then had to decide first whether the statement had been voluntary when made, on the safeguards record, and second whether it was true, a question on which courts looked for corroboration in material particulars before acting on a retracted confession, particularly against a co-accused. In principle this two-stage inquiry preserved a judicial core inside the police station’s product. In practice it concentrated the whole trial on a single contest, the circumstances of one recording session months or years earlier, attested by the officers who conducted it, against the word of an accused who had every incentive to disown the statement whether or not his account of its making was true. Judges of the Designated Courts resolved that contest case by case, and their resolutions are the real jurisprudence of Section 15: a body of decisions in which some confessions carried convictions to the gallows and others were discarded as unreliable on records that looked, from the appellate report, uncomfortably similar. The instability was inherent in the device. A rule that makes admissibility turn on institutional rank and truth turn on a reconstruction of invisible events will produce both kinds of outcome, and it did. The full working of the ordinary rule that Section 15 displaced belongs to the separate account of the Evidence Act linked at the confession sentence earlier in this section; what matters here is the displacement itself, because it was the single most consequential legal act the statute performed.

What did the courts do with Section 15 confessions after Kartar Singh?

The Supreme Court applied Section 15 as written. In State v Nalini in 1999, the Court treated a confession recorded under Section 15 as substantive evidence usable against a co-accused charged and tried in the same case, and in the Bhullar appeal of 2002 a TADA conviction resting substantially on such a confession was upheld by a majority.

The Nalini decision mattered beyond its facts because it settled the largest open question in the section’s working: whether a confession that the Evidence Act would have excluded altogether could not only be admitted but could carry the case against men and women who had never made it. The Court’s answer, faithful to the section’s text, was that it could, subject to the voluntariness inquiry and the safeguards Kartar Singh had prescribed. The Bhullar appeal showed the human edge of that answer. Devinder Pal Singh Bhullar’s prosecution for the 1993 Delhi bombing rested substantially on his confession recorded under Section 15; a three-judge bench upheld his conviction and death sentence by a majority of two to one, with the dissenting judge holding that the confession should not sustain the conviction. A death sentence carried by a police-recorded confession, over a judicial dissent about that confession’s worth, became the standing exhibit in the argument that Kartar Singh’s guidelines, faithfully recited, could not cure the structural defect of 1872: the statement’s birthplace was a police station, and no certificate changed where it was born. At the same time, appellate courts did quash or discount confessions where the Rule 15 and Kartar Singh conditions had not been observed, in transmission, warning or voluntariness, so the safeguards were not wholly ornamental. The record supports neither the claim that Section 15 was an automatic conviction machine nor the claim that its safeguards tamed it. It supports the narrower finding that the section moved the decisive contest in a TADA trial from the courtroom to the police station, and that everything else in the trial was arranged around what had happened there.

The Trial Machine: Courts, Custody, Bail and Presumptions

A definition and a confession rule do not make a system. TADA’s system was completed by its trial machinery, and the machinery deserves to be read as a whole, because its parts reinforced each other. Trial took place before Designated Courts constituted under Section 9, staffed by judges appointed for the purpose, sitting in proceedings that could be held in camera under Section 16. The same section armed the court with powers to protect witnesses: a Designated Court could take measures to keep the identity and address of a witness secret, choose a protected venue, and direct that a witness’s name not appear in its orders, judgments or the records of the case. Proceedings could, for reasons recorded, continue in the absence of the accused. Investigation ran on an altered clock. Section 20(4) rewrote Section 167 of the Code so that the periods within which a charge sheet had to follow arrest were read as sixty days and one hundred and eighty days in place of the ordinary shorter periods, and a proviso allowed the Designated Court, on a report of the Public Prosecutor showing the progress of the investigation and the specific reasons for further detention, to extend the custody period up to one year. The Constitution Bench in the Sanjay Dutt case settled the consequence of the clock running out: an accused who was not charge-sheeted within the period, as extended, acquired a right to be released on bail on filing his application, and that right was not extinguished by a charge sheet filed afterwards.

The extension power was itself litigated into a discipline. In Hitendra Vishnu Thakur v State of Maharashtra, decided in 1994, the Supreme Court insisted that an extension under the proviso was not a routine administrative favour: it had to rest on the Public Prosecutor’s report indicating the progress of the investigation and the specific reasons for detention beyond the ordinary period, the accused was entitled to notice of the request, and the Designated Court had to apply its mind to the report rather than extend custody mechanically. The pairing of the two decisions, Hitendra Vishnu Thakur on how an extension may be earned and Sanjay Dutt on what happens when the extended period expires unused, built the only genuinely accused-protective doctrine inside Section 20, and it was built by judges reading a prosecutor’s report requirement seriously. The episode is a useful corrective to two simplifications at once. The Act’s machinery was not operated without judicial friction; and the friction had to be manufactured case by case, at the cost of appeals, out of a drafting detail, because the statute’s main lines all ran the other way.

How long could a TADA accused be held before the charge sheet?

Section 20(4) of TADA rewrote the remand periods of the Code of Criminal Procedure as sixty days and one hundred and eighty days. On a report by the Public Prosecutor showing the progress of the investigation and specific reasons for more time, a Designated Court could extend detention up to one year before the charge sheet fell due.

Bail was where the machine bore down hardest on the person inside it. Section 20(7) disapplied Section 438 of the Code altogether, so the anticipatory bail that an ordinary citizen could seek on apprehension of arrest simply did not exist for a person accused under TADA. Section 20(8) then reversed the ordinary bail inquiry. No person accused under the Act was to be released on bail unless the Public Prosecutor had been given an opportunity to oppose the application and, where the Public Prosecutor opposed it, unless the court was satisfied that there were reasonable grounds for believing that the accused was not guilty of the offence and that he was not likely to commit any offence while on bail. The ordinary sequence, in which the prosecution must justify continued detention, was inverted: the accused had to persuade the court of reasonable grounds for believing in his innocence, at the start of the case, on the prosecution’s own papers, before the evidence had been tested. Section 20(9) added that these limitations sat on top of, not in place of, the restrictions in the Code. The combined effect, recorded in the undertrial figures examined later in this history, was that custody under TADA was measured in years for thousands of people, a large proportion of whom were never convicted of anything.

Could a person accused under TADA obtain anticipatory bail?

No. Section 20(7) provided that nothing in Section 438 of the Code of Criminal Procedure applied to a case involving the arrest of a person accused under TADA. The pre-arrest protection available to an ordinary accused was removed entirely, so arrest under the Act could not be anticipated by an application for bail in advance.

Proof, finally, was assisted. Section 21 instructed the Designated Court to presume, unless the contrary was proved, that the accused had committed the offence under Section 3 where arms or explosives of the kind used in the offence were recovered from his possession with reason to believe that similar items had been used in its commission, or where expert evidence showed his fingerprints at the site of the offence or on anything, including arms and vehicles, used in it. A parallel presumption attached to a person who had rendered financial assistance to someone accused or reasonably suspected of the financing offence under the Act. As enacted in 1987, Section 21 had contained two further presumptions, one flowing from a co-accused’s confession naming the accused and one from the accused’s own confession to a person other than a police officer; both were omitted by the 1993 amendment, in the single most concrete act of statutory self-correction in the Act’s life. Appeals completed the design: under Section 19, an appeal from a Designated Court lay directly to the Supreme Court, ordinarily within thirty days, cutting the High Court out of the ordinary appellate path for a class of trials in which the High Court’s distance from the local forum might have been thought a virtue.

Why did appeals from TADA courts go directly to the Supreme Court?

Section 19 of the Act made the judgment, sentence or final order of a Designated Court appealable directly to the Supreme Court, ordinarily within thirty days. The design traded the High Court’s intermediate scrutiny for speed and finality at the top, on the stated logic that terrorism trials should not spend years moving through two layers of appeal.

Read as a whole, the machine had a coherent theory. Terrorism was said to be an offence category defined by secrecy, intimidation and organisation, so the statute answered secrecy with protected witnesses, intimidation with a court empowered to sit in camera, and organisation with conspiracy-wide confessions, long investigation periods and presumptions drawn from possession and fingerprints. Its critics had a coherent counter-theory: that each departure from ordinary process was defended by reference to the paradigm case of the trained insurgent, while the machinery, once created, was operated by ordinary police forces against whoever the definitions could be stretched to cover, and the definitions had been drafted to stretch. Which theory better fits the record is the question the renewal debates of 1989 to 1993 should have answered, and to which this history turns next, after the renewals themselves.

Two closing provisions complete the architecture and should not be skipped, because they show the draftsman’s intent for the whole. Section 24 protected action taken in good faith under the Act, shielding officials from suit and prosecution for the exercise of its powers, a clause whose breadth mattered greatly to the officers asked to work the statute and correspondingly to the citizens who alleged abuse, since it raised the threshold for every collateral challenge to an arrest or a search. Section 25 gave the Act overriding effect over anything inconsistent in the Code, the Evidence Act or any other law. The override clause is the statute’s signature on its own work: where the ordinary law and TADA conflicted, TADA was to win, expressly and in advance. A reader who wants the measure of the Act in one line can take it there. Parliament had not merely added offences to the criminal law; it had legislated a hierarchy in which the exceptional text prevailed over the ordinary codes wherever the two met, and it renewed that hierarchy three times.

The Life of a TADA Case

The provisions read differently when they are followed in the order in which a human being met them. A TADA case typically began not with a bomb but with a police decision: the decision to record the information under this Act rather than under the ordinary codes. After 1993 that decision had a gatekeeper, because Section 20A(1) required the prior approval of the District Superintendent of Police before information about a TADA offence could be recorded, and the Hussein Ghadially judgment later showed what happened to prosecutions in which that approval had been taken from the wrong hands. Before 1993 there was no statutory gatekeeper at all, and the FIR itself performed the transformation: the same facts that would have founded an ordinary murder or arms case founded instead a case in a different legal universe.

Arrest followed, with the ordinary constitutional discipline of production before a Magistrate preserved in form, but the remand conversation that followed was governed by Section 20(4): the prosecution could look to sixty days, then one hundred and eighty days, and on the Public Prosecutor’s report up to a full year of custody before the charge sheet was due, while the accused’s bail application ran into Section 20(8)’s demand for reasonable grounds to believe in his innocence and Section 20(7)’s refusal of anticipatory bail altogether. It was inside that long custody that the case’s decisive event usually occurred, the visit of the Superintendent of Police with the warning, the recording apparatus and the certificate prescribed by Rule 15. Defence lawyers of the period described the sequence with bitterness and consistency: a client who would say nothing for months in the hope of release would speak in the end, because the statute had arranged time itself as an instrument of pressure, and the confession, once recorded and transmitted to the Chief Judicial Magistrate, reorganised every subsequent judgment a court would make about bail, charge and sentence.

Cognizance after 1993 required the previous sanction of the Inspector-General of Police or the Commissioner of Police under Section 20A(2), a second filter placed deliberately at the courtroom door. Trial then lay before the Designated Court, which could also try connected offences under other laws in the same proceeding, could sit in camera, could shield witnesses under Section 16, and could in recorded circumstances proceed in the accused’s absence. Judgment carried the Act’s heavy minima, five years at the floor and death where a terrorist act had killed. Appeal lay once, directly to the Supreme Court under Section 19. A case that began with a district police decision thus ended, if it ended in court at all, at the apex of the judicial system, having passed through a forum built for it, under evidence rules written for it, on a timetable written for it. That completeness is TADA’s real institutional achievement and its real danger in a single description: the statute did not bend the ordinary system at a few points, it replaced the system end to end for everyone its definitions captured, and the definitions captured far more people than the emergency that justified them.

One figure in that process deserves a separate note: the Public Prosecutor. The Act provided for prosecutors appointed for the Designated Courts, and its machinery made the prosecutor a gatekeeper at two decisive points, the report on which custody could be extended toward one year under Section 20(4) and the opposition that triggered the reversed bail test under Section 20(8). In ordinary criminal process the prosecutor is the state’s advocate at trial; under TADA he was additionally the certifying officer of the investigation’s incompleteness, whose report the Designated Court was asked to trust in place of a charge sheet. The Hitendra Vishnu Thakur discipline, requiring the report to show the progress of the investigation and specific reasons, and requiring notice to the accused, was the courts’ recognition that this certification power had become a liberty power in disguise. Where the discipline was observed, the prosecutor’s report was a genuine account of an unfinished investigation. Where it decayed into form, and the undertrial record suggests how often it did, the case’s timetable was set by the very office whose interest lay in prolonging it, and the accused’s year in custody was authorised by the signature of his adversary.

The Renewal Cycle: 1989, 1991 and 1993

The 1987 Act’s two years ran to 23 May 1989. The first renewal was accomplished by the Terrorist and Disruptive Activities (Prevention) Amendment Act 1989, Act 16 of 1989, which substituted four years for two in Section 1(4), carrying the statute’s life to May 1991. The political setting was the last year of Rajiv Gandhi’s government, with Punjab still violent and a general election due. No serious voice in that debate suggested that the emergency had ended; the argument was about how the Act was being used, and the government’s answer was that the instrument remained necessary and that its working did not justify surrendering it. The second renewal followed in 1991 through Act 35 of 1991, which substituted six years with effect from 2 May 1991, extending the life to May 1993. That renewal was carried through in a year of political discontinuity: the Chandra Shekhar government, the assassination of Rajiv Gandhi by the LTTE in May 1991 in the middle of a general election, and the arrival of P. V. Narasimha Rao’s Congress government in June. A statute that had begun as one government’s emergency device had, by its second renewal, been continued under governments of three different formations and a caretaker interlude, which is one reason the renewal story cannot be told as the story of a single party’s attachment to the law. Every formation that held power between 1987 and 1993 chose continuation.

The third renewal, Act 43 of 1993, substituted eight years for six, taking the final expiry to 23 May 1995. It followed the Bombay blasts of March 1993, the single deadliest terrorist attack in India to that date, in which coordinated explosions killed more than two hundred and fifty people; a Parliament asked to let the anti-terror statute lapse weeks after such an attack was being asked for a political act very few legislatures would perform. Yet the 1993 amendment is the most interesting of the three renewals, because for the first time the price of extension included safeguards written into the Act itself. The same amending Act that added two years inserted Section 20A, which provided that no information about an offence under the Act could be recorded by the police without the prior approval of the District Superintendent of Police, and that no court could take cognizance of a TADA offence without the previous sanction of the Inspector-General of Police or, in a commissionerate, the Commissioner of Police. It also removed the two confession-linked presumptions from Section 21. The legislative logic was explicit in the structure: renewal was being traded against filtration, an attempt to ensure that the Act’s door was guarded at the point of entry rather than reviewed only after years of custody.

Between the first and third renewals, the statute also performed a quieter feat: it survived the fall of its own authors. The government that enacted the 1987 Act left office in 1989. The coalition that replaced it included parties which had attacked the Act’s record in opposition, and the governments that followed, of a third formation and then of the Congress again, each had the opportunity at a renewal point to let the Act die and each declined. No cabinet between 1987 and 1995 ever put the question of abolition to Parliament as a serious proposition. That cross-party continuity is the single strongest evidence for the institutional reading of the renewal cycle offered later in this history: TADA persisted not because any one party loved it but because the office of governing, whoever held it, found the Act useful and its surrender dangerous. The opposition of each period, correspondingly, criticised the Act’s use while taking care not to promise its repeal in terms that a subsequent bombing could convert into a liability. The sunset clause was thus guarded from both sides of the House: by governments that would not surrender the powers and by oppositions that would not pledge to surrender them.

That trade defines the renewal cycle’s real character. In form, the cycle worked exactly as a sunset clause promises: three times, Parliament had to pass a fresh law to keep TADA alive, and each time the justification was publicly demanded. In substance, the cycle decayed in a predictable direction. The 1989 renewal was an extension. The 1991 renewal was an extension carried through a change of government, almost an administrative reflex. The 1993 renewal was an extension purchased with safeguards, which is a concession that the critics’ central charge, indiscriminate invocation, was true enough to require a statutory answer. What the cycle never produced was the one thing a sunset clause exists to make possible: a moment at which the evidence of use was weighed and found heavy enough to end the experiment. Each renewal debate began from the imminence of expiry rather than from an audit of the previous period, and the question put to Parliament was never whether the Act had worked, but whether the government of the day could be the government that let it die. That asymmetry, between the burden of continuing and the burden of ending, is the political engine inside the sunset that never set, and it recurs in every temporary security law that followed.

There is also an arithmetic of presentation that the renewal record exposes. Parliament never once voted to give India a decade of TADA. It voted, in 1985, for two years; in 1987, for two years; in 1989, for two more; in 1991, for two more; in 1993, for two more. Each individual vote was defended as short, exceptional and proportioned to a passing emergency, and the defence was available each time precisely because the increment was small. The decade was assembled the way such decades usually are, in slices too thin to alarm. A legislator who would have hesitated to enact a ten-year anti-terror code in 1985, and might have lost the vote, could vote five times for two years without ever confronting the sum. Sunset clauses create this slicing opportunity by design: their discipline depends on a legislature willing to aggregate, to judge the accumulated record rather than the marginal request. The TADA record shows a legislature that never aggregated until its own human rights commission did the arithmetic for it, in a letter addressed to every member individually some three months before the end.

The drafting form of the renewals tells its own story. Each extension was accomplished by a short amending Act whose operative act was the substitution of one figure for another in Section 1(4): two years became four, four became six, six became eight. No renewal re-enacted the statute, reopened its definitions, revisited Section 15 or re-argued the bail test. The legislative question was thus framed, by drafting choice, at its narrowest: not whether this body of law should exist, but whether the number in its first section should be increased. Opponents who wished to attack the confession rule or the definition of disruptive activity found themselves procedurally confined to a debate about duration, in which the government needed only to establish that terrorism had not vanished, a proposition no one in any of the three debates seriously contested. A sunset clause disciplines a legislature only if the renewal instrument reopens the substance, and the TADA renewal Acts were drafted to ensure that it did not. That is a design lesson independent of one’s view of the Act itself: the form of a renewal provision is as important as its existence, and a renewal that asks only about time will always be answered with time.

The Record of Use: Scale, Spread and the Conviction Question

The figures for TADA’s use are politically contested, and every number in this section is therefore attributed to its source rather than asserted as neutral fact. Human Rights Watch, in a backgrounder on India’s anti-terror laws published in 2001, recorded that over 76,000 people were arrested while TADA was in force from 1987 to 1995, and that the conviction rate was less than one per cent. Other published estimates of the conviction outcome range higher, from one per cent to four per cent, without displacing the central picture those figures agree on: the overwhelming majority of the people the Act processed were not convicted under it. Amnesty International’s reporting, as carried in the Indian press of the period, put the arrests at about 77,000, with about 72,000 of those arrested later released. The spread of estimates is itself information. A statute whose own state apparatus never produced a settled, authoritative account of how many people it had arrested and what became of them was being renewed, year after year, in an evidential fog that the renewal debates did little to disperse.

The geography of use told a sharper story than the totals. TADA had been justified by insurgencies in Punjab and, later, Kashmir, but its working life was national. A widely cited magazine account of the Act’s record, published in The Caravan, gives figures for 1990 in which the Gujarat police booked over 5,000 people under TADA in a single year, most of them Muslims, against about 1,600 booked in Punjab and about 2,000 in Jammu and Kashmir in the same year. Gujarat had no insurgency. What it had was a police force equipped with a statute that converted possession of specified arms in a notified area into a terrorism offence, removed anticipatory bail, reversed the bail test, and gave investigators a custody clock of up to a year. The same pattern was reported from Maharashtra and other states: the Act used in communal cases, in cases against habitual offenders, in labour and agrarian disputes, in the ordinary quarrels of local policing, wherever the procedural advantages of TADA were worth more to the local force than the discipline of the ordinary codes. The definitions examined earlier in this history supplied the stretch; the trial machinery supplied the incentive. A charge under TADA was a force multiplier for any investigation, and police forces across the country responded to the incentive exactly as the Act’s critics had predicted.

The defence of the record deserves to be stated as its makers stated it. Successive governments argued that low conviction numbers measured the wrong thing: that TADA’s purpose was preventive and disruptive, that arrests broke up networks and deterred attacks that consequently never appeared in any statistic, and that the relevant comparison was not convictions against arrests but the trajectory of the Punjab insurgency itself, which was defeated in the same decade. K. P. S. Gill, who led the Punjab Police through the final phase of that insurgency, defended extraordinary legal powers as operationally indispensable, and the argument that Punjab was pacified with TADA among the instruments available is part of the historical record, not a debating point. The critics’ answer was equally structured. They argued that the same police leadership credited firmness of ordinary counter-insurgency, not the Act’s procedural departures, when it described the Punjab victory; that a law defended as a scalpel had in practice been used as a net in states with no insurgency at all; and that the cost side of the ledger was not statistical but personal, counted in the years that tens of thousands of unconvicted people spent inside a process designed to make release difficult. Between those two accounts the statute’s own figures, so far as any were agreed, supported one proposition beyond dispute: very few of the people processed under TADA were convicted under TADA, and the Act was nevertheless defended and extended three times by the governments that possessed those figures. Explaining that combination is the real work of this history, and the explanation lies partly in the asymmetry of the renewal debate already described, and partly in the constitutional validation the Act received from the Supreme Court in 1994, which removed the last institutional pressure for repeal while leaving every power intact.

That documentation had an infrastructure, and the infrastructure mattered to the Act’s eventual fate. Civil liberties organisations, including the People’s Union for Civil Liberties and the People’s Union for Democratic Rights, maintained the unglamorous work of listing names, districts, charges and durations of custody, state by state and year by year, so that the misuse argument could be conducted with records rather than adjectives. Their reports recorded the pattern that the aggregate figures only suggest: TADA charges laid in communal disturbances against members of the minority community in numbers grossly disproportionate to any organisational involvement, charges laid in industrial and agrarian disputes, charges laid against habitual offenders and small arms carriers in notified areas whose connection to any terrorist design was never articulated even in the prosecution’s own papers. Defence lawyers in the Designated Courts added the trial-level record: cases in which the TADA charge was dropped or the accused discharged after years in custody, cases compounded or withdrawn when the political weather changed, and the steady conversion of the Act’s terror jurisdiction into a remand jurisdiction, in which the point of the charge was the custody it purchased rather than the conviction it promised.

Why the Act Was Hard to Kill

The combination recorded in the previous section, enormous use, minimal conviction, repeated renewal, demands an explanation, and the explanation is institutional rather than conspiratorial. Five interests aligned in favour of continuation, and no comparably organised interest stood on the other side until the very end. The first was the police interest. TADA was, from a district officer’s desk, a productivity instrument: it lengthened investigation time, weakened the accused’s exit routes, and made the senior officer’s recorded conversation with a prisoner evidentially priceless. No police leadership in any state asked Parliament to take the instrument away, and several defended it in public with the authority of operational experience. The second was the state government interest. Law and order is a state subject in its daily working, and state governments of every party discovered that the Act cost them nothing politically with their own electorates while giving their police a charge of rare coercive weight; the misuse that filled the reports occurred inside their jurisdictions, but the statute itself could always be blamed on, or credited to, Delhi.

The third interest was the Union government’s interest in never owning an attack. Incumbency under conditions of terrorism is an exercise in blame management, and the asymmetry recorded earlier in this history, that renewal is blamed by few and lapse is blamed by everyone after the next bombing, operated on every cabinet that considered the question between 1989 and 1995. The fourth was the judiciary’s inadvertent contribution: by upholding the Act in Kartar Singh on the assumption of good faith, the Supreme Court removed unconstitutionality from the opposition’s vocabulary, leaving critics to argue policy against a government arguing necessity, which is the weakest ground a reform movement can occupy. The fifth was the absence of a statutory reviewer. The Act created no inspector, commission or reporting duty of its own; the figures that eventually damaged it were assembled by journalists, civil liberties groups and finally the National Human Rights Commission, all working from outside, all deniable as partisan until the Commission’s letter of February 1995 made denial untenable. A sunset clause surrounded by these incentives is a clock in a room where every occupant profits from ignoring it, and the surprising fact in the record is not that TADA was renewed three times but that a route to ending it was found at all.

To these five a sixth may be added, quieter than the rest: the constituency of the already-charged. At every renewal point after 1989, thousands of prosecutions were pending under the Act, and lapse, for all the savings clause’s protection of pending cases, was represented in debate as a signal that would demoralise investigators and embolden the accused in live trials. The argument confused the statute’s future with its past, since the savings language answered it completely, but it was rhetorically serviceable precisely because it wore the dress of concern for victims and witnesses in real cases. A temporary law that generates a large pending caseload thereby acquires an argument for its own continuation drawn from the caseload itself, and the longer it runs the larger the argument grows. TADA’s renewal history is the working demonstration.

Kartar Singh: Validation With Safeguards

The constitutional challenge to TADA reached a five-judge Constitution Bench of the Supreme Court and was decided on 11 March 1994 in Kartar Singh v State of Punjab, reported at (1994) 3 SCC 569. The leading judgment, delivered by Justice S. Ratnavel Pandian, upheld the Terrorist Affected Areas (Special Courts) Act 1984 and both TADA statutes against the challenge that Parliament lacked the competence to enact them and that their provisions offended the guarantees of equality, free expression, protection against self-incrimination and life and personal liberty. The Court treated public order and terrorism as subjects on which Parliament could legislate for the whole country, accepted that the mischief of organised terrorism differed in kind from ordinary crime, and proceeded, as it said, on the assumption that those entrusted with the Act’s extraordinary powers would act in good faith and that lapses or excesses in individual cases would not condemn the statute. One provision alone did not survive: Section 22, which gave identification of a proclaimed offender by photograph the same value as an identification parade, was struck down as invalid, the Court holding it unintelligible and prone to grave injustice. Everything else, including Section 15, survived. Section 15 was sustained by the majority, but the judgment surrounded the confession power with the strictest procedural conditions in the Act’s history.

Those conditions are the judgment’s most durable legacy, and they repay attention because they show a court rewriting a statute’s safeguards while declaring its deference to the legislature. For confessions, the Court prescribed guidelines built on Rule 15 and going beyond it: the recording officer must warn the person plainly that he is not bound to confess and that a confession may be used against him; the confession must be recorded in an atmosphere free from threat or inducement, in the language of its maker, read back to him, and certified with a memorandum of voluntariness; the recorded statement must be sent forthwith to the Chief Metropolitan Magistrate or Chief Judicial Magistrate; and the maker, produced before the Magistrate, was to have the opportunity to complain of ill-treatment, with medical examination available where torture was alleged. Beyond the confession, the Court looked for systemic filtration of the Act’s use: it commended screening or review committees to examine whether TADA had been rightly invoked in individual cases and periodic review of the areas notified for the arms-possession offence, so that the geography of exception would not expand by inertia. The pattern of the judgment is the pattern later repeated, with variations, for POTA and the UAPA: uphold the statute, read safeguards into its operation, strike little or nothing, and leave the legislature’s powers exactly where they were. Critics of the judgment, including civil liberties organisations that had supplied much of the misuse record, argued that safeguards administered by the same institutions that produced the misuse were paper controls, and that the assumption of good faith was the one finding in the case contradicted by the largest volume of evidence. Defenders answered that a court which had struck the Act down in March 1994, a year after the Bombay blasts, would have been governing by judgment rather than judging, and that the guidelines route changed real outcomes in real confession cases, as later benches proved by quashing convictions where the safeguards had not been followed.

The screening committee recommendation had an administrative afterlife that partially vindicated both sides of that argument. Following the judgment, and again under the pressure of the undertrial litigation after the lapse, state governments constituted screening or review processes for pending TADA cases, and the reviews contributed, alongside the Shaheen bail directions, to the steep fall in the undertrial population the NHRC recorded between 1995 and 1997. The committees’ existence proved the critics’ premise, that very large numbers of cases should never have carried the TADA label at all, since a screening mechanism finds little to screen in a system that was invoked with discrimination. Their lateness proved the defenders’ difficulty in reverse: the filtration the Court commended in March 1994 was applied in earnest only when the statute was already dying, and to prisoners who had spent the intervening years inside it. A safeguard that arrives at the end of a statute’s life, however honestly administered, is a remedy for the record rather than for the prisoner, and the gap between the judgment’s design and its decade is the measure of how much of Kartar Singh was description of the Act India had and how much was a portrait of the Act the Court wished it were judging.

What Kartar Singh undeniably did was settle the constitutional argument while the political argument was still running. After March 1994, no government could say that TADA was legally fragile, and no opposition could say that its continuance was unconstitutional. The judgment converted the remaining question into a purely legislative one: whether Parliament, knowing the record, wished to keep the Act. Fourteen months later, Parliament answered by silence, and the way it answered is the hinge on which the whole story turns.

One Statute, Many Laws in Practice

A central statute in a federal republic is operated by state police forces, and TADA’s national text concealed sharply different local laws in practice. In Punjab and in Jammu and Kashmir, the Act was worked, whatever else is said of it, against armed organisations engaged in campaigns of assassination and bombing, and its Designated Courts dealt with the conspiracy trials of a genuine counter-insurgency. In Gujarat, on the figures already recorded, the Act became in 1990 an instrument used overwhelmingly in communal and ordinary policing contexts, with more bookings than Punjab and Kashmir combined in that year. In Maharashtra, the Act’s career ran from communal cases toward the organised crime and smuggling networks that culminated in the 1993 blasts investigation, in which TADA supplied the trial forum and the confession rule for the largest prosecution of the decade. The same Section 5 charge that in one state recorded the recovery of a militant cache recorded in another the recovery of a country-made pistol from a man with no politics at all.

The variation was structural, not accidental. The Act’s triggers were in state hands: notification of areas under Section 5, the district-level decision to invoke the Act (unchecked until Section 20A in 1993), the choice of forum and the pace of investigation. The Union government that renewed the statute controlled none of those triggers directly, which produced the renewal debates’ characteristic evasion, in which the Centre defended the Act by pointing to Punjab while its critics attacked it by pointing to Gujarat, and both were describing the same statute accurately. The lesson generalises beyond TADA: a special law’s real content is the sum of the incentives it hands to the most local official empowered to use it, and any assessment confined to the paradigm case in the capital will misdescribe the law that citizens actually met.

The notification device shows the mechanism at its purest. Section 5’s offence existed only inside an area the State Government had notified, and notification was an executive act requiring no legislative approval and attracting no publicity obligation that a villager or a slum resident could be expected to encounter. A state could, and states did, notify districts wholesale, converting every recovery of a scheduled weapon inside them into the raw material of a terrorism charge. The documented record of the Act’s use in the communal disturbances of the early 1990s, compiled by the civil liberties organisations and reflected in the aggregate figures, follows the notification map more faithfully than it follows any map of insurgency: where the area was notified and the local force wished to hold men whom the ordinary bail law would have released within days, TADA supplied years. The Kartar Singh bench’s commendation of periodic review of notified areas was aimed at exactly this quiet geography, and its neglect in the years when it might have mattered is one of the clearest instances of the judgment’s safeguards existing principally in the judgment.

The competence holding deserves a paragraph of its own, because it decided the ground on which all later anti-terror statutes stand. The challenge had argued that public order and police are State subjects, and that a Union statute creating offences, courts and procedure for what was in substance public order trespassed on the federal distribution of powers. The Court rejected the argument by characterising terrorism as a national phenomenon that no single State could confront, drawing on the Union’s defence and residuary powers and on the scale and cross-border character of the threat the statutes recited. Once that characterisation was accepted, the fundamental rights challenges were argued provision by provision, and the bench’s method was consistent: each departure from ordinary process was measured against the paradigm of the organised terrorist conspiracy, and each was found justified by the paradigm, with the safeguards the Court prescribed offered as the answer to the cases that fell outside it. The judgment’s treatment of Section 20(8)’s bail reversal and of Section 21’s presumptions followed the same pattern of acceptance disciplined by direction. Lawyers who later challenged POTA and the UAPA amendments inherited this structure entire: the competence question was closed, the paradigm method was entrenched, and the space left for argument was the space Kartar Singh itself had created, the adequacy and observance of safeguards. In that sense the judgment was not only TADA’s vindication but the constitutional template of Indian anti-terror law as such, and its most quoted sentence in the years since has been its assumption, hopeful in 1994 and harder to repeat with each year of the record, that those entrusted with the powers would exercise them in good faith.

The Lapse of 1995

The end, when it came, came from an accumulation of pressure rather than a single decision. The misuse record had become impossible to confine to opposition pamphlets: the figures of mass arrest and negligible conviction were in the reports of international organisations, the press of record, and, decisively, the publications of the state’s own human rights institution. The National Human Rights Commission, which had from its early years opposed the renewal of TADA, escalated its intervention in the Act’s final months. Its Chairperson, Justice Ranganath Misra, wrote on 20 February 1995 to every Member of Parliament, strongly urging that TADA be allowed to lapse, and the Commission’s annual report later recorded that this intervention contributed substantially to the final decision. The institutional route of that criticism matters for this series, because the Commission’s powers and limits are the subject of the linked account of the NHRC; within this history the point is simpler. By the winter of 1995 the case against renewal was no longer being made only by the Act’s traditional opponents. It was being made by a statutory commission chaired by a former Chief Justice of India, addressed individually to every legislator who would have to vote.

The government moved in the same direction. The Union Home Ministry acknowledged in the mid-1990s that the Act had been misused by sections of the police in several states, an admission that emptied the renewal case of its central premise, since the government’s defence of the Act had always rested on the claim that abuse was incidental rather than structural. On 16 May 1995 the government decided that TADA would not be renewed. The statute’s eight-year term under the 1993 amendment expired on 23 May 1995, and the Act lapsed. The choice of ending deserves precision, because it is the ending the whole design of the statute had pointed to. TADA was not repealed. No repeal bill was debated, no final parliamentary judgment on a decade of operation was delivered, and no findings on the misuse record were adopted. The government instead let the clock run out, which required no vote, no defence of the record and no repudiation of the powers themselves. A replacement proposal, the Criminal Law Amendment Bill 1995, was prepared as a possible successor and was not enacted. The sunset clause, renewed three times, finally operated in the one way it could not be argued with: nobody moved to stop it.

What lapsed on 23 May 1995 also deserves precision, because the word covers less than a newspaper reader of the day might have assumed. No new TADA case could be instituted for conduct after the lapse, no area could be newly notified, no confession could be recorded under Section 15 for a fresh investigation, and the Designated Courts lost their prospective jurisdiction. Everything already inside the system, by contrast, was preserved by the savings language: pending investigations could be completed, pending trials continued to judgment, sentences already imposed remained executable, and appeals lay and were decided under the dead Act’s provisions. The files that state governments had opened under the Act did not close; they merely stopped growing. Screening and review of pending cases, on the lines the Supreme Court had commended in Kartar Singh, became the operative question of the Act’s afterlife, and the undertrial litigation described in the next section was its first public test. Lapse, for TADA, was therefore not an event but a change of phase, from a living statute generating new cases to a closed corpus of cases working itself out through courts that outlived it by decades.

The lapse was tested almost immediately, in the cruellest available way. On 31 August 1995, three months after the Act expired, the Chief Minister of Punjab, Beant Singh, was assassinated by a suicide bomber in Chandigarh. The killing was precisely the category of event the renewal asymmetry predicted would revive a dead security law: a spectacular attack, in the very state whose insurgency had justified the statute, within weeks of its surrender. Demands for the Act’s return were made, and the Punjab government of the day faced the argument that it had been disarmed by Delhi’s conscience. The Act was not revived. The investigation and prosecution of the assassination proceeded under the ordinary law and the saved provisions, and the republic absorbed the lesson in both directions at once: the lapse had not disarmed the state, and the temptation the sunset clause exists to outlast had arrived on schedule and been outlasted. That August remains the strongest single piece of evidence that the ending of TADA was a settled national decision rather than an administrative accident, because a decision that survives the assassination of a chief minister is a decision.

The institution that had done most to engineer the ending recorded the outcome with a restraint that should not disguise its completeness. The National Human Rights Commission, whose chairperson’s letter the Commission itself credited with contributing substantially to the decision, treated the lapse in its annual reporting not as a victory to celebrate but as a file to keep open: its monitoring turned from the Act’s renewal to the Act’s residue, the thousands of undertrials whose cases the savings clause preserved, and it continued to place their numbers before the Supreme Court and the public until the population had fallen, on its own figures, from 6,060 in June 1995 to 1,502 at the start of 1997. The sequence captures the mature division of labour that ended TADA. Parliament supplied the silence, the Commission supplied the documentation, the Court supplied the bail directions, and the Act’s own savings clause supplied the problem all three were managing. No one of those institutions could have ended the Act’s working life alone, which is the final reason the ending took a decade: the statute had distributed the power to continue it among so many hands that ending it required, in the end, an alignment of all of them.

After Lapse: Savings, Undertrials and the Long Tail

Lapse is not erasure, and TADA’s afterlife is the proof. The savings language in Section 1(4) preserved the previous operation of the Act, penalties incurred under it, and every investigation and legal proceeding in respect of liabilities acquired while it lived, all of which could be instituted, continued and enforced as if the Act had not expired. The immediate consequence was a population of prisoners whose statute had died around them. The National Human Rights Commission recorded 6,060 TADA undertrials still in jail on 30 June 1995, weeks after the lapse, and placed its information before the Supreme Court, which on 27 February 1996 issued detailed directions in the Shaheen Welfare Association proceedings for the release on bail of categories of TADA undertrials, graded by the gravity of the charge and the time already served. On the Commission’s figures, the number of TADA undertrials fell to 1,502 by 1 January 1997. The sequence is the statute’s afterlife in miniature: a parliamentary silence ended the Act, and a combination of a rights commission’s documentation and the Court’s bail directions was needed to end its custody.

The trials ran far longer. The Rajiv Gandhi assassination case, investigated and tried under TADA among other laws, reached the Supreme Court in State v Nalini, decided in 1999, which applied Section 15 to confessions recorded from co-accused and sustained the Designated Court’s working in the most closely watched prosecution of the decade. The 1993 Bombay blasts case, tried in a TADA Designated Court in Mumbai, produced its verdicts in 2006 and 2007, more than a decade after the Act had expired, and Yakub Memon, convicted under TADA among other statutes, was executed on 30 July 2015, twenty years after the lapse. Devinder Pal Singh Bhullar, convicted under TADA for the 1993 Delhi bombing directed at M. S. Bitta, had his death sentence commuted by the Supreme Court in 2014. Section 20A, the entry-guard inserted in 1993, was still deciding appeals in that decade: in Hussein Ghadially v State of Gujarat, decided on 18 July 2014, the Supreme Court set aside TADA convictions from the Surat blasts cases because the approvals and sanctions on which the prosecutions rested had not been given by the authorities the section named, holding that strict compliance with the safeguard was mandatory and that a prosecution built on approvals from the wrong officers could not stand. A general account of a lapsed statute cannot serve as advice on any individual case, pending or concluded; the position of a particular accused depends on the record of that prosecution. What the record does establish, across two decades of appellate decisions, is that the savings clause did what it was drafted to do, and that a statute which Parliament allowed to lapse in 1995 was still being authoritatively interpreted, applied and, where its safeguards had been ignored, used to undo convictions, a full generation after its death.

The Commission’s submission to the Supreme Court in the undertrial proceedings supplies the connecting tissue between the misuse record and the afterlife, and it is worth stating in its own terms. As the Commission’s annual report summarised the position it had laid before the Court, a large number of the persons inside the TADA system had been arrested for offences which had nothing to do with terrorist and disruptive activities, and the system’s own filters had failed to separate them from the paradigm accused for whom the statute existed. The Shaheen directions were the Court’s administrative answer to that submission: undertrials were grouped by the seriousness of the charge and the custody already undergone, and bail followed for the categories whose continued detention the record could not defend. The most famous single beneficiary of the same corrective process illustrates the pattern from the other end of the celebrity scale. The actor Sanjay Dutt, arrested in the Bombay blasts investigation and charged under TADA, had the TADA charges against him dropped by the Designated Court in 1995 and was in the end convicted, in 2006, under the Arms Act alone. Thousands of unknown accused travelled the same arc, from a terrorism charge that organised their custody to an ordinary-law outcome or an acquittal that arrived years later, without the epilogue ever attracting a headline.

The Interregnum: India Without a Terror Statute, 1995 to 2002

Between the lapse of TADA in May 1995 and the enactment of POTA in 2002, India fought its insurgencies and its terrorist campaigns without a dedicated central anti-terror statute, and the interregnum is the control experiment the TADA debate had always lacked. The ordinary law did not fall silent: the Indian Penal Code’s offences against the state, the Arms Act and Explosives Acts, and the Unlawful Activities (Prevention) Act 1967 in its then form, with its machinery for declaring associations unlawful, all remained in force, and the insurgency in Kashmir and the militancy in the North-East were policed under them together with the armed forces legislation applicable in disturbed areas. The state-level response foreshadowed the future in one important respect: Maharashtra’s Control of Organised Crime Act 1999 carried the TADA device of the police-recorded confession, in its Section 18, into state law for organised crime, evidence that the confession rule had acquired a constituency independent of the terrorism justification that had created it.

At the centre, the question of a successor statute stayed open and unresolved. The Criminal Law Amendment Bill 1995, the Narasimha Rao government’s own draft successor, had failed with the lapse it was meant to soften. The idea returned in systematic form with the Law Commission’s 173rd Report of 2000, which recommended fresh prevention of terrorism legislation and offered a draft bill built with procedural safeguards, a lineage the Prevention of Terrorism Ordinance of October 2001 and the statute that followed it in 2002 visibly inherited. The interregnum thus cuts both ways in the argument. Defenders of TADA point to the return of the demand for a terror statute within six years, and to the attacks of the period, as proof that the lapse had left a genuine gap. Critics answer that the state functioned, prosecuted and convicted under ordinary law throughout the period, that the gap years produced no collapse, and that the eventual successor reproduced TADA’s devices so faithfully, and generated its own misuse record so quickly, that the interregnum’s real lesson is about drafting habits rather than about necessity. Both arguments are live in the POTA history to which this series turns next, and both begin in the silence of 23 May 1995.

The interregnum’s closing events fixed the terms on which the successor would be demanded. The hijacking of Indian Airlines Flight IC-814 in December 1999 ended with the release of imprisoned militants and was remembered as a humiliation of the state’s resolve. The attack on the Red Fort in Delhi in December 2000 and, decisively, the attack on Parliament itself on 13 December 2001 converted the Law Commission’s draft from a file in the ministry into an ordinance and then a statute within months. None of that alters the analytical value of the six gap years for judging TADA. For six years the Indian state confronted insurgency in Kashmir, militancy in the North-East, organised crime syndicates and externally sponsored attacks with the ordinary codes and the 1967 Act, and its courts convicted where the evidence supported conviction. Whatever the emergency case for a special statute in 1985, the interregnum demonstrated that the alternative to TADA was not impotence. It was the ordinary criminal law, working at ordinary speed, with ordinary safeguards, which is precisely the comparison the sunset clause had been designed, a decade earlier, to force Parliament to make.

Lineage and Lesson: The Sunset That Never Set

TADA did not begin India’s argument about special anti-terror law, and it did not end it. Its place in the sequence is set out in the companion history of India’s anti-terror laws as a lineage, from the Preventive Detention Act through MISA and the National Security Act to the UAPA. Within that lineage, TADA’s specific bequest was a template: define the terrorist act broadly, move the trial into a special court, admit the police confession, reverse the bail test, and let presumptions share the prosecution’s burden. When the next emergency statute came, its drafters worked from that template rather than from first principles, and the repeal of POTA in 2004 is best read as the second act of the drama TADA scripted, in which a successor law reproduced the confession rule and the special courts, generated a documented misuse record of its own, and was undone by an election rather than by its own short life. The provisions of the three statutes are set side by side in the head-to-head comparison of TADA, POTA and the UAPA, which shows how much of TADA’s machinery survives inside the law that replaced the replacements.

The migration of the template can be traced device by device, and the tracing is the quickest way to see which parts of TADA were genuinely of their emergency and which became permanent habits of Indian drafting. The police confession travelled furthest: POTA’s Section 32 reproduced Section 15 in substance, the Maharashtra Control of Organised Crime Act carried it into state law against organised crime, and its absence from the UAPA, whose Parliament chose extended custody and reversed bail but not police confessions, marks the one point at which TADA’s precedent was consciously refused. The special court travelled with it, reappearing in POTA’s Special Courts and later in the National Investigation Agency’s designated courts. The reversed bail test and the lengthened charge sheet period travelled into the UAPA’s own text, where they remain. The presumptions travelled least, trimmed at each copying. Read in this way, the lapse of 1995 ended a statute and barely interrupted a drafting tradition, which is why the Act rewards the attention this history has given it: TADA is the source code of Indian anti-terror law, and its bugs, as faithfully as its features, were inherited.

The lesson usually drawn from this history is about misuse, and the misuse record is real and documented. But the deeper lesson is about time. Every safeguard debate around TADA after 1987 assumed that the statute’s temporary character was itself a safeguard, that a law which must be renewed can be trusted with powers a permanent law could not carry. The decade refuted the assumption in its simple form. Renewal turned out to be a weaker discipline than its designers hoped, because the politics of renewal are asymmetric: continuation requires only that a government not wish to be blamed for the next attack, while termination requires a positive decision to surrender powers in public. Three renewals went through on that asymmetry, under governments of different parties, and the fourth decision, in 1995, was taken only when a rights commission’s letter, a Home Ministry admission and a decade of figures had made renewal more embarrassing than lapse. Any future sunset clause in Indian security legislation should be read with that record in mind, and readers preparing seminar presentations from this history can open and check a slide file in the browser through the ReportMedic presentation viewer before relying on the renewal figures quoted here, several of which remain contested between the sources that recorded them.

The TADA Renewal Timeline

The single findable artifact of this history compresses the argument into its instruments. Each row records one enactment or extension, the period it covered, the political setting in which it passed, and the criticism on record against it. The final row states the savings position, because the story’s last legal fact is that the Act’s death did not end its operation.

Instrument	Period covered	Political context	Criticism recorded
Terrorist and Disruptive Activities (Prevention) Act 1985, Act 31 of 1985	In force 24 May 1985 for two years	Rajiv Gandhi government; Punjab insurgency after Operation Blue Star and the assassination of Indira Gandhi	Breadth of the disruptive activity definition and the special court design criticised from enactment
Act 46 of 1985	Extension of the terrorist act provisions to Jammu and Kashmir from 5 June 1985	The original carve-out for Jammu and Kashmir removed within weeks	The geography of exception widened almost immediately
Terrorist and Disruptive Activities (Prevention) Ordinance 1987, Ordinance 2 of 1987	Promulgated 23 May 1987, in force 24 May 1987	Parliament not in session as the 1985 Act expired; no gap accepted in the regime	Use of the ordinance route to continue extraordinary powers without a prior parliamentary vote
Terrorist and Disruptive Activities (Prevention) Act 1987, Act 28 of 1987	Assent 3 September 1987; deemed in force from 24 May 1987 for two years	Fresh enactment replacing the 1985 Act in strengthened form	Sections 5, 15, 21 and 22 attacked as the regime’s harshest devices
Amendment Act 16 of 1989	Two years substituted by four; life to May 1991	Final year of the Rajiv Gandhi government; Punjab still violent	Renewal granted without an audit of the first period of use
Amendment Act 35 of 1991	Four years substituted by six, with effect from 2 May 1991; life to May 1993	Political transition year; Rajiv Gandhi assassinated in May 1991	Continuation treated as administrative reflex across a change of government
Amendment Act 43 of 1993	Six years substituted by eight; life to 23 May 1995	Passed after the Bombay blasts of March 1993 under the Narasimha Rao government	Extension traded against the Section 20A entry safeguards and the removal of two Section 21 presumptions
Lapse of the Act	Life expired 23 May 1995 without renewal	Government decision of 16 May 1995 not to renew; NHRC Chairperson Justice Ranganath Misra had written to every Member of Parliament on 20 February 1995 urging lapse	Ended by silence rather than repeal, with no parliamentary findings on the misuse record
Savings position after lapse	Section 1(4) preserved proceedings, penalties and investigations	Trials continued under the saved provisions; Supreme Court bail directions of 27 February 1996 addressed the undertrial population	Prosecutions and appeals ran for two decades after the statute’s death, including the Bombay blasts verdicts of 2006 and 2007

Verdict

TADA’s life answers the question its sunset clause posed. A legislature that writes a two-year limit into a security statute is making a claim about itself: that it will revisit the emergency with fresh eyes and end what it cannot justify. Between 1985 and 1995 that claim was tested four times, and it held only once, at the very end, when ending the Act had become easier than defending it. The statute’s devices were not marginal. Section 15 altered the law of confessions for every trial it touched and set the precedent every successor statute reached for. The bail reversal and the custody clock decided the lived experience of tens of thousands of accused persons, most of whom were never convicted. Kartar Singh preserved the whole structure at the price of safeguards the same structure was left to administer. The fairest verdict is the one the record supports without exaggeration in either direction: TADA was a real answer to a real emergency, drafted wider than its justification, renewed more easily than its record deserved, and ended by exhaustion rather than by judgment. The sunset that never set is a warning about drafting optimism. Temporary powers are temporary only if the institution that granted them is willing, on a named day, to say that the emergency has passed, and on the evidence of this decade that willingness, not the clause, is the safeguard that matters.

India had run this experiment once before, and the comparison instructs. The Preventive Detention Act 1950, the republic’s first temporary security statute, was likewise enacted for a short term, renewed year after year for two decades, and finally allowed to lapse in 1969 when the government of the day chose not to continue it. The pattern is older than TADA and survived it: temporary security legislation in India has repeatedly lived not by the strength of its justification at any single renewal but by the difficulty, at every renewal, of being the institution that ends it. What distinguishes TADA in the sequence is not the renewal habit but the completeness of the world it built while the habit ran: a definition, a court, an evidence code, a bail code and a clock, all of it validated by the Supreme Court, all of it outlived by its own case files. To read the Act whole, as enactment, renewal and lapse in a single arc, is to see how a republic can keep every promise it wrote into a sunset clause and still break the promise the clause was making.

One counterfactual haunts the whole subject and should be faced squarely before the verdict closes. Would Punjab have been pacified, and the republic protected, without TADA? The question is unanswerable in the form in which it is usually asked, because the Act operated alongside a massive counter-insurgency, a political accord, the exhaustion of the militant movement’s external support and the ordinary criminal law, and no historian can distribute the causation among them with confidence. What can be said is narrower and stronger. The necessity of some extraordinary response in 1985 is as well supported as any such claim can be. The necessity of this statute, in this breadth, renewed on this record, in states that had no insurgency, against accused persons the overwhelming majority of whom were never convicted under it, is not supported at all; it was simply never tested, because the renewal machinery was built, and then operated, so as never to test it. The gap between those two propositions is the whole history of TADA, and it is the gap into which every successor debate should be made to step.

Frequently Asked Questions

Q: What was TADA and when was it enacted?

TADA was the Terrorist and Disruptive Activities (Prevention) Act, India’s first statute to define terrorist and disruptive activities as distinct offences and to try them before special Designated Courts under altered rules of evidence, custody and bail. It was enacted twice. The first version, Act 31 of 1985, received presidential assent on 23 May 1985 and came into force on 24 May 1985 with a life of two years. When that Act expired, an ordinance continued the regime from 24 May 1987, and Parliament then passed the Terrorist and Disruptive Activities (Prevention) Act 1987, Act 28 of 1987, which received assent on 3 September 1987 and is the version under which the Act is generally discussed.

Q: Why did TADA carry a sunset clause?

The sunset clause limited the Act to two years so that extraordinary powers would exist only for the stated emergency and would lapse unless Parliament affirmatively renewed them. It served three purposes. It forced the government to return to the legislature and justify continuation with the record of the preceding period. It gave Parliament a scheduled occasion to examine misuse allegations before extending the powers. And it supported the Act’s constitutional defence, since a measure the legislature kept under periodic review was easier to present to the Supreme Court as a proportionate and temporary response rather than a permanent parallel code. In practice the clause produced three renewals, in 1989, 1991 and 1993, before lapse finally occurred in 1995.

Q: What did Section 15 of TADA allow that ordinary law did not?

Section 15 made a confession recorded by a police officer admissible in evidence, which the ordinary law forbids. Sections 25 and 26 of the Indian Evidence Act 1872 bar proof of confessions made to a police officer or in police custody, so under ordinary law a confession could be used only if recorded by a Magistrate under the Code’s safeguards. Section 15 displaced that bar for TADA trials: a confession made before an officer not lower in rank than a Superintendent of Police, recorded in writing or on a mechanical device, was admissible against its maker and, where they were charged and tried in the same case, against a co-accused, abettor or conspirator, provided the officer had warned the person that he was not bound to confess and had reason to believe the statement was voluntary.

Q: Did the Supreme Court uphold TADA in Kartar Singh?

Yes, in substance. In Kartar Singh v State of Punjab, decided on 11 March 1994 by a five-judge Constitution Bench and reported at (1994) 3 SCC 569, the Supreme Court upheld the TADA statutes and the Special Courts Act 1984 against challenges based on legislative competence and fundamental rights. Section 15’s confession rule was sustained by the majority, subject to strict procedural guidelines on warning, recording, transmission to the Chief Judicial Magistrate and production of the maker before that Magistrate. The one provision struck down was Section 22, which had given identification of a proclaimed offender by photograph the same value as a test identification parade; the Court held it invalid as unintelligible and prone to grave injustice. The judgment proceeded on the stated assumption that officials entrusted with the powers would act in good faith.

Q: Why was TADA allowed to lapse in 1995?

By 1995 the political cost of renewal had overtaken the case for it. The misuse record had been documented by human rights organisations and the press over years, with mass arrests and a very low rate of conviction as the recurring figures. The National Human Rights Commission opposed renewal, and its Chairperson, Justice Ranganath Misra, wrote to every Member of Parliament on 20 February 1995 urging that the Act be allowed to lapse. The Union Home Ministry had acknowledged misuse by sections of the police in several states. On 16 May 1995 the government decided against renewal, the eight-year term expired on 23 May 1995, and a proposed successor, the Criminal Law Amendment Bill 1995, was never enacted. The Act was ended by lapse rather than repeal, so no repeal debate or parliamentary findings on the record were required.

Q: How widely was TADA used before it lapsed?

The figures are contested and attributed differently by source. Human Rights Watch’s 2001 backgrounder on Indian anti-terror laws recorded over 76,000 arrests while TADA was in force from 1987 to 1995, with a conviction rate of less than one per cent; other published estimates put the conviction outcome between one and four per cent. Amnesty International’s figures, as reported in the Indian press, spoke of about 77,000 arrests and about 72,000 of those arrested later released. Use was national rather than confined to Punjab: a widely cited account in The Caravan gives figures for 1990 in which Gujarat booked over 5,000 people under TADA, most of them Muslims, against about 1,600 in Punjab and about 2,000 in Jammu and Kashmir in the same year.

Q: Did TADA trials continue after the Act lapsed?

Yes. Section 1(4) of the 1987 Act contained a savings clause providing that expiry would not affect the Act’s previous operation, penalties incurred, or investigations and legal proceedings in respect of liabilities acquired while it was in force, and that such proceedings could be instituted, continued and enforced as if the Act had not expired. Prosecutions therefore ran on for years. The Bombay blasts case produced verdicts in a TADA Designated Court in 2006 and 2007, the Rajiv Gandhi assassination case was decided by the Supreme Court in 1999, and appeals turning on TADA’s safeguards were still being decided two decades after the lapse, as in the Hussein Ghadially judgment of 2014.

Q: Which prosecutions are still associated with TADA?

Three prosecutions dominate the Act’s public memory. The Rajiv Gandhi assassination case was tried before a Designated Court and reached the Supreme Court in State v Nalini in 1999, a leading authority on Section 15 confessions. The 1993 Bombay blasts case, arising from the deadliest terrorist attack of the period, was tried in a TADA Designated Court in Mumbai, with verdicts delivered in 2006 and 2007 and the execution of Yakub Memon in 2015. The prosecution of Devinder Pal Singh Bhullar for the 1993 Delhi bombing aimed at M. S. Bitta also rested on TADA, ending with the Supreme Court commuting his death sentence in 2014. All three continued long after the Act lapsed, under the savings clause.

Q: How many times was TADA renewed between 1987 and 1995?

The 1987 Act was renewed three times. The Amendment Act 16 of 1989 substituted four years for the original two in Section 1(4), extending the Act’s life to May 1991. The Amendment Act 35 of 1991 substituted six years, with effect from 2 May 1991, extending it to May 1993. The Amendment Act 43 of 1993 substituted eight years, extending it to 23 May 1995, and in the same Act inserted the Section 20A safeguards and removed two presumptions from Section 21. No fourth renewal was moved, and the Act lapsed at the end of the eight-year term. Counting the 1985 predecessor, the regime Parliament repeatedly prolonged ran for ten years in total, from May 1985 to May 1995.

Q: Why was a new TADA Act passed in 1987 instead of simply extending the 1985 Act?

The 1985 Act had expired by its own two-year limit on 23 May 1987, and an expired statute cannot be prolonged by a simple amendment of its duration clause in the same easy way as a living one; the government also wished to strengthen the regime rather than merely continue it. It therefore bridged the gap with Ordinance 2 of 1987, in force from 24 May 1987, and then enacted a fresh statute, Act 28 of 1987, most of which was deemed in force from 24 May 1987 while Sections 5, 15, 21 and 22 operated from assent. The 1987 Act restated the scheme with a wider arms-possession offence, the confession rule in its mature form, detailed presumptions and photographic identification, so replacement, not extension, was the honest description of what Parliament did.

Q: What did the 1993 amendment to TADA change besides extending its life?

Act 43 of 1993 did three things besides substituting eight years for six in the duration clause. It inserted Section 20A, which barred the police from recording information about a TADA offence without the prior approval of the District Superintendent of Police, and barred courts from taking cognizance without the previous sanction of the Inspector-General of Police or the Commissioner of Police. It removed two of the presumptions in Section 21, the presumption drawn from a co-accused’s confession naming the accused and the presumption drawn from the accused’s confession to a person other than a police officer. And it left the rest of the machine intact. The amendment is remembered as the renewal that traded two more years of life for the Act’s first statutory entry safeguards.

Q: Did TADA apply only to Punjab or to the whole of India?

TADA applied to the whole of India from its commencement. The 1985 Act contained one early carve-out: as enacted, the provisions creating the terrorist act offence did not extend to Jammu and Kashmir, and that omission was removed by Act 46 of 1985, so the Act applied to that State from 5 June 1985. Within its field, the operative limitation was different: offences like the Section 5 arms-possession charge depended on areas notified by the State Government, and the choice of where to notify and where to invoke the Act rested with state police forces. That is why the recorded use of TADA was heavy in states with no insurgency, including Gujarat and Maharashtra, and why the Act’s history is a national history rather than a Punjab history.

Q: Was TADA repealed by Parliament, or did it expire on its own terms?

It expired on its own terms. The 1993 amendment had fixed the Act’s life at eight years from 24 May 1987, ending on 23 May 1995, and the government decided on 16 May 1995 not to seek a further renewal. No repeal bill was introduced or debated, and Parliament made no findings on the Act’s record. The distinction matters legally and historically. A repeal is a legislative judgment that a law should end; a lapse under a sunset clause is the working of the original bargain, and it allowed the Act to end without any government having to defend or repudiate the decade of its operation. The savings clause in Section 1(4) in any event preserved pending proceedings, so much of the Act’s practical operation survived the expiry it underwent.

Q: What did the National Human Rights Commission do to oppose the renewal of TADA?

The Commission opposed renewal institutionally and documentarily. It had recorded its opposition to the Act’s continuance in its reporting on the human rights situation, and on 20 February 1995 its Chairperson, Justice Ranganath Misra, a former Chief Justice of India, wrote to every Member of Parliament strongly urging that TADA be allowed to lapse; the Commission’s annual report later recorded that this intervention contributed substantially to the decision. After the lapse, the Commission documented the population the Act left behind, recording 6,060 TADA undertrials in jail on 30 June 1995, and placed its information before the Supreme Court, whose directions of 27 February 1996 on bail for categories of undertrials were followed, on the Commission’s figures, by a fall in the undertrial count to 1,502 by 1 January 1997.

Q: Which provision of TADA did the Supreme Court strike down in Kartar Singh?

Section 22. That provision gave the identification of an accused, in the case of a proclaimed offender, by photograph the same evidential value as identification in a test identification parade. The Constitution Bench in Kartar Singh held the provision invalid, describing it as unintelligible and prone to grave injustice, since a photograph-based identification carries none of the safeguards of a parade and the section’s own terms could not be given a coherent working meaning. It was the only provision of the Act struck down in that judgment. Everything else challenged, including Section 15 on confessions, the bail restrictions of Section 20 and the presumptions of Section 21, was upheld, in Section 15’s case subject to the strict procedural guidelines the Court prescribed.

Q: What happened to the TADA undertrials still in jail when the Act lapsed?

The lapse did not release them automatically, because the savings clause preserved pending proceedings and the charges against them survived the statute. The National Human Rights Commission recorded 6,060 TADA undertrials in jail on 30 June 1995 and took their position to the Supreme Court, which in the Shaheen Welfare Association proceedings issued detailed directions on 27 February 1996 for the release on bail of categories of undertrials, graded by the nature of the charge and the period already spent in custody. The Commission’s figures show the undertrial population falling to 1,502 by 1 January 1997. Cases not covered by the bail directions continued toward trial under the saved provisions, and some prosecutions ran for many more years.

Q: Did the government try to enact a replacement law after TADA lapsed?

A replacement was prepared but not enacted. The Criminal Law Amendment Bill 1995 was drafted as a possible successor that would have carried a version of the special regime into ordinary criminal legislation, and it was reported on as such by Amnesty International and the United States State Department’s country reporting of the period. Parliament did not pass it, and India had no dedicated central anti-terror statute between the lapse of TADA in May 1995 and the enactment of POTA in 2002, the gap being covered by the ordinary penal law and by the Unlawful Activities (Prevention) Act 1967 in its then form, including its provisions on unlawful associations. The failure of the 1995 Bill is part of why the lapse is read as a genuine ending rather than a rebranding.

Q: Why do demands to revive TADA surface after major terrorist attacks?

Because TADA occupies the place in political argument of a law remembered as strong, and because its strongest advocates present its repeal-era history as proof that firmness works: the Punjab insurgency was defeated during the Act’s currency, and police leadership of that campaign, including K. P. S. Gill, defended extraordinary powers as necessary. After a major attack, the devices associated with TADA, above all the admissibility of confessions recorded by senior police officers, are proposed again as the missing instrument, and the argument is made that procedural concessions to accused persons are a luxury a state under attack cannot afford. The counter-argument drawn from the same record is that TADA’s conviction outcomes were extremely low, that its use fell most heavily on people never convicted, and that its genuinely useful machinery had by then largely been carried into the UAPA. The demand revives the debate; it has not revived the Act.

Q: When did the trial in the 1993 Bombay blasts case, the longest-running TADA prosecution, conclude?

The trial before the TADA Designated Court in Mumbai concluded with verdicts delivered across 2006 and 2007, thirteen to fourteen years after the blasts of March 1993 and more than a decade after the Act itself had lapsed. The accused were tried under TADA together with other statutes, and the proceedings survived the lapse under the savings clause in Section 1(4). Sentencing and appellate proceedings extended the case further: the Supreme Court decided the principal appeals in 2013, and Yakub Memon, convicted under TADA among other laws and sentenced to death, was executed on 30 July 2015 after his final challenges failed. The case is the standard illustration of how a sunset clause ends a statute without ending its prosecutions.

Q: What was the Terrorist Affected Areas (Special Courts) Act 1984, and how did it relate to TADA?

The 1984 Act was Parliament’s first step toward special anti-terror procedure. It provided for special courts to try scheduled offences in areas declared terrorist affected, recognising that ordinary trial forums in the affected districts could not function safely or swiftly, but it created no new definition of terrorism and left evidence, bail and investigation law unchanged. TADA, enacted the following year and expanded in 1987, absorbed the special court idea into its own Designated Courts and added everything the 1984 Act had lacked: substantive offences, the confession rule, extended custody and the reversed bail test. The two statutes were challenged together and upheld together in Kartar Singh in 1994, and the 1984 Act is best understood as TADA’s institutional rehearsal.