The Official Secrets Act 1923 is the statute under which India punishes espionage and the unauthorised movement of official information, and it is a statute with a hole at its centre. The Act, cited as Act 19 of 1923, creates offences that turn on whether information is secret, official, calculated to be useful to an enemy, or likely to affect the security of the State, and yet it contains no definition of a secret, no definition of an official secret, and no test by which a document becomes protected. A reader who works through its fifteen sections looking for the definition around which everything else is built will not find it, because it is not there. What the Act protects is identified outside the Act, in executive instructions that classify government documents, and what the Act punishes is identified inside it, in elements that a prosecution must prove in court. That division, between an administrative act of classification that the statute never mentions and a criminal liability that the statute defines in detail, is the single most important fact about the Official Secrets Act 1923, and it organises everything that follows in this article.

The statute is short enough to be read whole, and this article reads it whole. It sets out the principal offences, beginning with the spying offence in Section 3 and the wrongful communication offence in Section 5, and it explains the supporting offences, the definition of a prohibited place, and the procedural conditions that govern any prosecution, including the requirement of a government complaint before a court may take cognizance and the power to hold proceedings in camera. It then turns to the two relationships that define the Act in a constitutional order very different from the one in which it was enacted: its relationship with the Right to Information Act 2005, which overrides it for the purposes of access while leaving its offences intact, and its relationship with the freedom of the press, addressed by the Supreme Court in a 2019 judgment on documents already published by a newspaper. It records the reform history, in which an official commission recommended repeal and replacement and a later committee examined the Act against the transparency statute, and in which neither exercise changed the statute book. And it gives both sides of the retention debate in their strongest form before reaching a verdict, because the case for a general official secrets law is a real case and the case against this particular law is also a real case, and a statute pillar article that pretended otherwise would fail its readers.

One discipline governs the treatment throughout. The Act is a statute under which journalists and officials have been charged, and this article describes the law and the debate about the law. It describes no prosecution, names no person who has been charged, and characterises no charge. Where decided cases matter, they are the cases that state what the law means, not the cases in which individuals stood accused. The distinction matters to the statute itself, because much of the Act’s practical force lies not in convictions but in invocation, in the fact of its existence on the statute book and the caution its existence induces, and that force can be analysed without narrating any individual’s proceedings.

How does the Act divide its offences between purpose and custody?

The Act divides its offences by their organising idea. Section 3 is organised by purpose: its acts become criminal only when done for a purpose prejudicial to the safety or interests of the State. Section 5 is organised by custody: it criminalises wilful, unauthorised communication of protected material by a person holding it through office, confidence or contract.

The Act penalises two families of conduct. Section 3 punishes spying: acting for a purpose prejudicial to the safety or interests of the State by entering or approaching a prohibited place, making a sketch, plan, model or note useful to an enemy, or obtaining, collecting, recording, publishing or communicating protected codes, documents or information. Section 5 punishes the wrongful communication of official information by a person who holds it through office or confidence, and it also punishes a person who voluntarily receives such information knowing it was communicated in contravention of the Act. Supporting sections punish the unauthorised use of uniforms and official documents, interference with police and armed forces personnel near prohibited places, attempts, and the harbouring of offenders.

That short answer conceals the structure that the long answer must expose. Section 3 is built around a purpose. The prohibited acts, approaching a place, sketching it, collecting information about it, are in themselves ordinary acts that surveyors, journalists, researchers, photographers and the merely curious perform every day, and they become offences only when done for a purpose prejudicial to the safety or interests of the State. Section 5 is built around a relationship. The prohibited act, communicating information, is in itself the ordinary act of every official who briefs a colleague, a minister, a journalist or a family member, and it becomes an offence because of the route by which the information reached the person communicating it and the absence of authority for the communication. The two offences therefore police different things. Section 3 polices intention directed against the State. Section 5 polices the custody of official information regardless, in the ordinary case, of any hostile intention at all. A person may commit the Section 5 offence without wishing the State any harm, by passing on a document that came to hand through office to a person not authorised to receive it. A person cannot commit the Section 3 offence, on the better reading of its text and of the jurisprudence under it, without the prejudicial purpose that its opening words require. Keeping that distinction clear is the key to reading the Act accurately, and much confused commentary on the statute comes from collapsing the two offences into a single idea of leaking.

A colonial statute on a republican statute book

The Official Secrets Act 1923 was enacted by the colonial legislature, modelled on legislation of the same period in the United Kingdom, and it replaced an earlier Indian statute on the same subject, the Indian Official Secrets Act of 1889 as amended in 1904. Its lineage explains its shape. The British official secrets legislation of the early twentieth century was built for an imperial state that treated official information as the property of the administration and disclosure as a species of disloyalty, and the Indian statute inherited that architecture: a general prohibition on the movement of official information, offences graded by their proximity to defence, and no statutory machinery for deciding what counted as secret, because in the imperial design that decision belonged to the executive that owned the information. The statute was a consolidating and strengthening measure, gathering the espionage offences and the disclosure offences into one short code, and it has survived, in substance, every constitutional transformation since: the Government of India Act 1935, independence, the Constitution of 1950 with its guarantee of freedom of speech and expression in Article 19(1)(a), and the transparency settlement of 2005.

Survival is itself a datum. The statute book of the Republic is full of colonial legislation that was retained because it did ordinary work, companies law, contract law, evidence, procedure, that any state must do in some form. Official secrecy is different in kind, because it sits in direct tension with a commitment the Republic made twice, once in the Constitution’s speech guarantee and once in the Right to Information Act’s disclosure regime, and a statute can be retained by inertia in a way that a statute in the centre of a constitutional argument cannot. The retention of the 1923 Act is therefore not adequately explained by the general survival of colonial codes. It reflects a series of decisions, or failures to decide, by governments that considered the Act, received recommendations about it, and left it standing. The reform record examined later in this article shows that the Act has been considered at the highest official level more than once, and that the consideration produced recommendations which were not implemented. A colonial statute retained after examination is a different thing, politically, from a colonial statute retained by oversight, even though the legal position is identical.

The United Kingdom comparison sharpens the point. The British legislation on which the Indian Act was modelled was itself reformed in stages, and the catch-all disclosure offence of the early British statutes was replaced in 1989 by a narrower statute confined to defined categories of information and, for several categories, requiring proof of damage. India made no equivalent move. The Indian Act retains the structure of the earlier British model, a general offence of wrongful communication unconfined by category and unqualified by any damage requirement in its text. Comparative lawyers make this point often, and it is a fair point as far as it goes, but it should not be overstated. The question for India is not whether its statute matches a foreign template but whether its statute fits its own constitutional order, and that question is answered by Indian materials: the text of the Act, the text of the transparency statute that overrides it in part, and the judgments of Indian courts on what the offences require. The comparison matters chiefly as evidence that the design choices in the 1923 Act are choices, that other democracies protecting the same interests have drafted differently, and that difference in drafting is available to Parliament whenever Parliament chooses to take it up.

Why does the Act contain no definition of a secret?

The omission is structural rather than accidental. The statute was drafted on the assumption that the executive would identify its own secrets through administrative classification, so Parliament legislated the offences and left the subject matter to departmental practice. No later amendment has disturbed that allocation, and the defining function remains outside the enacted text.

The Act contains no definition because its design assigns that question to the executive. The statute criminalises conduct relating to secret official codes, protected documents and information useful to an enemy, and then leaves the identification of what is secret to administrative practice: government documents are classified under departmental security instructions, and a document’s classification marking is the practical signal that the Act’s offences are engaged. Parliament in 1923 legislated on the assumption that the administration would know its own secrets and mark them, and no Parliament since has inserted a statutory definition or a classification scheme into the Act.

The consequence deserves to be stated plainly, because it is the article’s structural point and the brief for this pillar requires it to be faced rather than softened. In most criminal statutes, the definition of the prohibited thing is the Legislature’s central act: a drugs statute defines the substances, a corruption statute defines the gratification, a terrorism statute defines the terrorist act, and argument in court proceeds from the definition outward. In the Official Secrets Act 1923 the order is reversed. The offences are defined with some care, and the subject matter of the offences is not defined at all. Whether a particular document is the kind of document the Act protects depends on how the executive has treated it: whether it was classified, at what level, whether it was entrusted in confidence, whether it relates to a prohibited place. The criminal law thus incorporates by reference a set of administrative decisions that are themselves governed by instructions the statute does not contain, does not authorise in terms, and does not reproduce. That is not a drafting oversight that a court can repair, because a court asked to supply a definition would be legislating the central element of a criminal offence. It is a feature of the enacted design, and every debate about the Act, retentionist and repeal-minded alike, is at bottom a debate about whether that design remains defensible.

The architecture of the Act

The Act is arranged in fifteen sections, and its internal economy repays attention because the distribution of weight tells the reader what the statute is for. The opening sections define the field. Section 2 supplies definitions, of which the definition of a prohibited place in clause (8) is the operative one, examined in detail below. Section 3 states the spying offences and Section 4 adds provisions on communication with foreign agents, including evidential rules about when a person is taken to have been in communication with a foreign agent and the meaning of that expression. Sections 5 to 10 state the remaining offences: wrongful communication, unauthorised use of uniforms, falsification and impersonation, interference with officers on duty near prohibited places, the duty of certain persons to give information to the police and the armed forces about offences, attempts and incitement, and the harbouring of spies. Sections 11 to 15 are machinery: powers of search warrant, the sanction requirement for prosecutions, the power to exclude the public from proceedings, and the liability of companies and firms for offences committed with the consent or through the neglect of their officers. Nothing in the Act creates a regulator, a classification authority, a review body, a reporting obligation or a sunset. There is no institutional architecture at all. The Act is a set of offences and a set of procedural conditions, and the institutions that operate it, the classifying departments, the sanctioning government, the prosecuting agencies and the courts, are all institutions created by other laws.

That spareness has one consequence that commentary often misses. Because the Act creates no machinery of its own, every safeguard in its operation comes either from its procedural conditions, the sanction requirement above all, or from the general law of criminal procedure and evidence, or from the constitutional jurisprudence of the courts. When defenders of the Act point to the rarity of abusive prosecution, they are pointing substantially to safeguards located outside the offences themselves. When critics attack the Act’s breadth, they are attacking the offences read alone. Both are describing the same statute from different ends, and the full picture requires holding both ends at once: offences of considerable breadth, filtered by a political sanction requirement and by courts that have insisted on proof of the elements the text contains.

Which offences carry the Act’s real weight?

Section 3 and Section 5 carry the weight between them. Section 3 supplies the gravity, with imprisonment extending to fourteen years for defence-related spying. Section 5 supplies the reach, governing the everyday custody of official information by officials and others, and extending to knowing recipients. The remaining offences protect the perimeter these two create.

Two offences carry the practical weight of the statute. Section 3, the spying offence, carries the gravity: imprisonment extending to fourteen years where the offence relates to defence works, establishments or secret official codes, and to three years otherwise. Section 5, wrongful communication, carries the reach: imprisonment extending to three years for communicating protected official information without authority, extending also to the voluntary recipient. The remaining offences, on uniforms, documents, interference, attempts and harbouring, support these two and are charged far less prominently in the statute’s public life.

The grading inside Section 3 is one of the few places where the Act speaks with precision about degrees of harm, and it deserves notice. The fourteen-year ceiling is reserved for offences committed in relation to a work of defence, an arsenal, a naval, military or air force establishment or station, a mine, minefield, factory, dockyard, camp, ship or aircraft, naval, military or air force affairs, or a secret official code. In every other case the ceiling is three years. The same three-year figure governs Section 5 and Section 6. The statute’s own sentencing structure therefore distinguishes sharply between espionage directed at the defence establishment, which it treats as among the gravest offences in the ordinary criminal calendar, and the wider penumbra of prohibited-place and information offences, which it grades at the level of a moderately serious offence. Commentary that speaks of the Act as though every breach exposed a citizen to fourteen years misstates the statute. So does commentary that speaks as though the Act were trivial. The truth is in the grading, and the grading is in the text.

Section 3: the spying offence read closely

Section 3(1) opens with the words on which the whole offence turns: if any person, for any purpose prejudicial to the safety or interests of the State, does one of the three specified classes of act, that person commits the offence. The three classes are worth separating, because they protect different things and they fail in different ways when a prosecution is tested.

The first class, clause (a), is place-based. It is an offence, with the prejudicial purpose, to approach, inspect, pass over, be in the vicinity of, or enter any prohibited place. The verbs are deliberately wide. Entry is only the last of them. A person who never crosses a boundary, who inspects from outside or passes over in the air, is within the clause if the purpose is proved. The width is intelligible in a statute aimed at the reconnaissance that precedes sabotage or attack, and it is equally plain that the width transfers the whole weight of the offence onto the purpose element: standing near a dockyard is a crime or a commute depending on why the person is there, and the statute knows this, which is why the purpose is placed at the head of the section rather than buried inside a clause.

The second class, clause (b), is artefact-based. It covers making any sketch, plan, model or note which is calculated to be, or might be, or is intended to be, directly or indirectly useful to an enemy. The triple formulation, calculated to be or might be or intended to be, is cumulative in effect: the prosecution need not show that the maker intended the artefact to reach an enemy, if the artefact is of a kind that might be useful to one. Combined with the purpose requirement at the head of the section, the clause reaches the preparation of material about protected subjects by a person acting with a prejudicial purpose. The words directly or indirectly do further work, extending the clause to material whose usefulness is mediated, a note about a supply route, a sketch of an approach road, rather than material useful in itself.

The third class, clause (c), is information-based and the widest of the three. It covers obtaining, collecting, recording, publishing or communicating to any other person any secret official code or password, or any sketch, plan, model, article, note, document or information which is calculated to be or might be or is intended to be directly or indirectly useful to an enemy, or which relates to a matter the disclosure of which is likely to affect the sovereignty and integrity of India, the security of the State or friendly relations with foreign States. The final limb, added by amendment in the years after independence, detaches the offence from the enemy-usefulness test altogether for a defined class of subject matter: material relating to sovereignty and integrity, state security, or foreign relations. The presence of publishing among the verbs is the textual foundation for the recurring public question whether publication can attract the Act, and the answer the text gives is that publication is named in Section 3, subject always to the prejudicial purpose that governs the whole subsection. Section 3 is therefore not confined to the classic spy who steals and transmits. Its verbs include the acts of the researcher and the publisher, and its discipline is supplied not by narrowing the verbs but by the purpose element and the sanction filter examined later in this article.

Section 3(2) addresses proof. It provides, in substance, that on a prosecution under the section it is not necessary to show that the accused was guilty of any particular act tending to show the prejudicial purpose, and that the purpose may be inferred from the circumstances of the case, the accused’s conduct, or the accused’s known character as proved. This is an evidential provision of real significance and it must be read exactly. It does not remove the purpose element. It permits the element to be proved by inference from conduct and circumstances rather than by direct evidence of a specific act, which is how purpose is proved in most criminal law, and it sits alongside the insistence of the courts, examined below, that the prejudicial purpose is an element to be established and not a presumption to be indulged. The tension between Section 3(2) and that insistence is a genuine feature of the statute’s operation, and honest treatment requires acknowledging both: the text facilitates inference, and the courts discipline its use.

Must the prosecution prove a prejudicial purpose under Section 3?

Yes, the prejudicial purpose is an element of the offence and the prosecution must establish it, although Section 3(2) permits it to be inferred from the circumstances, the accused’s conduct or known character as proved. The courts treat the purpose as a matter for proof on evidence, not as a presumption arising from the prohibited act alone.

Yes. The prejudicial purpose is an element of the Section 3 offence, placed by the text at the head of the subsection and qualifying every act the section prohibits, and the prosecution must establish it. The courts’ jurisprudence under the Act treats the purpose as a matter for proof on the evidence, including proof by inference from conduct and circumstances under Section 3(2), and not as something presumed from the fact that a prohibited place was approached or that protected information changed hands. A prosecution that proved the act without the purpose would not have proved the offence.

This insistence matters beyond doctrine because it is the principal judicial contribution to the Act’s shape. Parliament wrote a broad offence; the courts, applying the ordinary presumption that a penal statute means what its elements say, declined to read the head words as decoration. The result is a Section 3 that is narrower in court than it appears on the page to a reader who stops at the verbs. Critics of the Act are entitled to respond that an element provable by inference from conduct and character is a thinner protection than an element requiring direct proof, and that Section 3(2) was written precisely to widen the inferential route. Both points are sound, and together they state the true position: the purpose element is real, it is the prosecution’s to prove, and the statute itself supplies generous means of proving it inferentially. Any reform draft that took the point seriously would have to decide whether inference from character, as distinct from conduct, should remain available at all, a question the unimplemented reform reports did not reach in terms.

Prohibited places: the geography of the Act

The definition of a prohibited place in Section 2(8) is the one extended definition the Act does contain, and its structure mirrors the statute’s priorities. The definition covers, in the first place, works and establishments of defence: any work of defence, arsenal, naval, military or air force establishment or station, mine or minefield belonging to or occupied by or on behalf of the Government, and factories, dockyards, camps, ships and aircraft belonging to or occupied by the Government in which munitions or defence material are made, repaired or stored. It extends to places where such material is kept in temporary custody and to buildings and structures in process of construction for such purposes. A second limb reaches beyond defence installations to any place belonging to or used for the purposes of the Government which is for the time being declared by notification of the appropriate Government to be a prohibited place on the ground that information about it, or damage to it, would be useful to an enemy. A third element brings in railways, roads, canals and other means of communication, and works connected with them, when so declared. The definition is therefore partly self-executing and partly declaratory. Defence establishments are prohibited places by the statute’s own force. Other government places become prohibited places by executive notification, which is the same pattern found at the centre of the Act: the statute marks the core, and the perimeter is drawn by the administration.

Which places does Section 2(8) place within the Act’s geography?

Section 2(8) places defence works, arsenals, service establishments and stations, mines and minefields, and government factories, dockyards, camps, ships and aircraft used for defence material within the definition by the statute’s own force. Government places, railways, roads, canals and other means of communication join the definition when the appropriate Government declares them by notification for the statutory reason.

A prohibited place includes defence works, arsenals, naval, military and air force establishments and stations, mines and minefields, and government factories, dockyards, camps, ships and aircraft where defence material is made, repaired or stored, together with places where such material is temporarily kept. It also includes any place belonging to or used by the Government, and railways, roads, canals and other means of communication, that the appropriate Government declares by notification to be a prohibited place on the ground that information about the place or damage to it would be useful to an enemy.

Two observations follow. The first is that the declaratory limb makes the geography of the offence partly invisible in the statute itself: a citizen reading the Act learns the categories but not the notified list, and must rely on notifications published by governments to know which civil installations, a particular railway work, a particular government building, carry the status. The second is that the prohibited place concept does work in several offences, not Section 3 alone. It anchors the interference offence, it frames part of Section 5, where information relating to or used in a prohibited place is one of the classes of protected material, and it shapes police powers in the vicinity of such places. The definition is thus a load-bearing wall of the statute, and it is significant that it is also the place where the Act comes closest to defining its subject matter. The statute can say what a prohibited place is. It cannot, or does not, say what a secret is. The contrast is instructive: the drafters were capable of definition where they chose to define, which strengthens the reading that the absence of a definition of secrecy is design rather than incapacity.

Section 5: wrongful communication and the law of custody

Section 5 is the provision that gives the Act its everyday significance, because it is the provision that governs the handling of official information by the people who hold it in the ordinary course of government. Its structure is a chain of custody. Section 5(1) identifies the material and the routes by which it reaches a holder: a secret official code or password, or a sketch, plan, model, article, note, document or information which relates to or is used in a prohibited place, or which has been made or obtained in contravention of the Act, or which has been entrusted in confidence to the holder by a person holding office under the Government, or which the holder has obtained or had access to owing to his position as a person who holds or has held office under the Government, or who holds a contract with the Government, or who is employed by such a person. The offence is then committed by a holder who wilfully communicates the material to a person other than a person to whom he is authorised to communicate it, or who uses the information in his possession for the benefit of a foreign power or in another manner prejudicial to the safety of the State, or who retains it when he has no right to retain it or contrary to his duty, or who fails to take reasonable care of it or conducts himself so as to endanger its safety.

Several features of this drafting deserve emphasis. The first is the breadth of the custody routes. Office is only one of them. Confidence is another: material entrusted in confidence by an office-holder is protected in the hands of the recipient-turned-communicator. Contract and employment under a contractor extend the chain beyond the ranks of government servants. The second feature is duration. The words holds or has held office carry the obligation past the end of service, so that the former official who communicates material obtained through office is within the section, a point of direct practical significance for retired officials, for memoir-writing, and for the consultancy and commentary that former officials undertake. The third feature is the absence of a harm element in the principal limb. Wilfulness is required: the communication must be deliberate, not accidental. Authority is the defence built into the offence: communication to a person the holder is authorised to communicate it to is no offence. But the principal limb does not require the prosecution to prove that the communication harmed the State, affected its security, or benefited anyone hostile. The protected status of the material, its route to the holder, the wilfulness of the communication and the absence of authority complete the offence. It is this structure that the Shourie Committee, whose work the reform record below examines, proposed to change, by confining the penal provision to violations affecting national security. That the proposal was made at all, by a committee the Government itself appointed, is the clearest official acknowledgement that Section 5 as enacted reaches conduct well beyond security harm.

Section 5(2) completes the circuit by reaching the recipient. A person who voluntarily receives any secret official code or password or any of the protected classes of material, knowing or having reasonable ground to believe, at the time of receipt, that it is communicated in contravention of the Act, commits an offence. The recipient limb is the textual answer to the question whether the Act reaches journalists and researchers, and it does so with a knowledge condition. Receipt is not an offence where the recipient does not know and has no reasonable ground to believe the communication breached the Act. Receipt is an offence where the recipient takes material in the knowledge that it is being passed in breach. The limb is discussed further in the treatment of the press below; here it is enough to note its place in the architecture. Section 5 governs a chain with two directions of liability: outward from the holder who communicates, and inward to the person who knowingly receives. Section 5(3) grades the whole offence at imprisonment extending to three years, or fine, or both, with a parallel provision for the offence in relation to certain material. Section 5(4), inserted by amendment, applies the section to material communicated in specified circumstances connected with ships and aircraft of a foreign State in defined conditions, a technical extension that confirms the section’s custodial character.

What does the Act leave undefined at its centre?

It leaves secrecy itself undefined. The Act uses expressions such as secret official code and describes classes of protected documents and information, but it defines neither a secret nor an official secret, sets no damage threshold, and enacts no classification scheme. The content of its central concept is supplied by executive instruction and tested, where tested at all, in court.

No. The Official Secrets Act 1923 nowhere defines a secret or an official secret, and it contains no classification scheme and no statutory criterion by which a document becomes protected. The expressions it uses, secret official code or password, documents and information of the kinds described in Sections 3 and 5, are descriptions of categories, not definitions of secrecy. In practice, whether a document is treated as protected is determined administratively, through classification under the departmental security instructions by which government departments grade and mark their records.

The point cannot be made too carefully, because a great deal of public writing about the Act gets it wrong in one direction or the other. Some writing asserts that the Act defines official secrets expansively, as though a definition existed and were broad. It does not exist. Other writing asserts that because there is no definition, the Act protects nothing until a court intervenes, as though protected status had no practical existence outside litigation. That is also wrong. The administrative classification system is real, it grades documents at levels descending from the highest grade for material whose unauthorised disclosure could cause exceptionally grave damage to national security, through secret and confidential, to restricted, and it governs how documents are marked, stored, carried and destroyed inside government. The Manual of Departmental Security Instructions issued by the Ministry of Home Affairs is the instrument through which this is done, and the Second Administrative Reforms Commission recorded the classification practice without treating the manual’s own confidentiality as an anomaly in itself, while insisting that disclosure should be the norm and secrecy the exception. The accurate statement is therefore twofold: the Act contains offences but no definition, and the definition function is performed, outside the Act, by executive instruction. A statute that protects secrets and does not say what a secret is is not a paradox invented by critics. It is a description of the enacted text.

The supporting offences and the procedural frame

The offences that surround Sections 3 and 5 are best understood as the statute protecting its own perimeter. Section 6 punishes the unauthorised use of naval, military, air force, police or other official uniforms, the falsification of reports, forgery, personation and the use of false documents in connection with prohibited places and protected material, the impersonation that is the classic instrument of the infiltrator. Section 7 punishes obstructing, knowingly misleading or otherwise interfering with a member of the armed forces or a police officer engaged on guard, sentry, patrol or similar duty in the vicinity of a prohibited place. Section 8 imposes on certain persons a duty to give information to the police and the armed forces about offences and suspected offences, and treats failure as an offence. Section 9 punishes attempts and incitement as though the offence attempted had been committed, and Section 10 punishes harbouring a person known or reasonably believed to have committed, or to be about to commit, an offence under the Act, where the harbouring is undertaken to prevent apprehension. Section 15 attributes liability to companies and firms where an offence is committed with the consent or connivance of, or is attributable to neglect by, a director, manager, secretary or other officer. The grading of these offences is modest beside Section 3’s defence ceiling, three years in the principal cases, which fits their character: they are offences of facilitation and perimeter, serious because of what they facilitate rather than because of what they are.

The procedural frame is where the statute keeps its safeguards, and two provisions matter most. Section 13 provides that no court shall take cognizance of an offence under the Act except upon a complaint made by order of, or under authority from, the appropriate Government. The requirement is a jurisdictional filter of genuine weight. It means that a prosecution cannot be set in motion by a private complainant, and cannot proceed at all unless the government of the day, Central or State as the case may be, has directed or authorised the complaint. Whatever else is said about the Act, this provision is a real procedural protection, and the reform debate has generally treated it as such: it converts the decision to prosecute into a decision for which a government is politically answerable, and it interposes that decision between an investigating agency’s suspicions and a court’s process. Its limits should also be noted. The filter is executive, not judicial. The sanctioning authority is the same state whose information the Act protects, and the sanction decision is not a public merits assessment. The protection is therefore against casual or unauthorised prosecution, not against a determined government, and critics who call the safeguard insufficient and defenders who call it substantial are both describing the same provision accurately from different premises about how much trust an executive filter deserves.

Who controls the gateway to a prosecution under the Act?

The appropriate Government controls it. Section 13 forbids a court to take cognizance of an offence except upon a complaint made by order of or under authority from that Government, so no investigation becomes a prosecution without a governmental decision at the threshold. The filter is executive in character, and it is the statute’s principal internal safeguard.

A court may take cognizance of an offence under the Act only upon a complaint made by order of, or under authority from, the appropriate Government, as Section 13 requires. No private complaint can set a prosecution in motion, and no prosecution can proceed without that governmental direction or authority. The requirement is a jurisdictional condition, placed at the threshold of the case, and it is the statute’s principal internal safeguard against casual or unauthorised prosecutions.

Section 14 supplies the second procedural feature: where proceedings under the Act are pending, the court may order that the public generally, or any particular person, be excluded from the court during the proceedings, notwithstanding that publicity is the ordinary rule of criminal trials, if the publication of evidence or statements made during the proceedings would be prejudicial to the safety of the State. The judgment remains publicly pronounced, but the route to it may be closed. The in camera power is the statute’s answer to a genuine dilemma, that a trial about secret material may itself disclose the material, and its existence is a reminder that the Act’s drafters thought about the litigation stage as part of the secrecy system. Section 12 applies the search warrant provisions of the criminal procedure code with modifications, authorising magistrates to issue warrants for places and material connected with offences under the Act. The frame, taken together, is coherent: an executive threshold before prosecution, closed proceedings where safety requires, and search powers adapted from the general code. A reader assessing the statute fairly will weigh this frame alongside the offences, because the Act as it operates is offences plus frame, and reform proposals that addressed only the offences, as several did, left the frame untouched.

Section 4: communication with foreign agents

Section 4 is the statute’s bridge between the spying offence and the intelligence world in which spying moves, and it is the provision least read by commentators who stop at Section 3. It provides that a person who has been in communication with, or has attempted to communicate with, a foreign agent shall be deemed to have committed an offence under Section 3 if the surrounding conditions the section specifies are present, and it supplies an extended definition of communication with a foreign agent for this purpose. A foreign agent, for the section, includes any person employed by a foreign power directly or indirectly for the purpose of committing an act prejudicial to the safety or interests of the State, or who has been or is reasonably suspected of having been so employed, or who purports to act for such a power. Communication is correspondingly extended: visiting the address of a foreign agent, consorting or associating with one, having a foreign agent’s name or address in one’s possession, and kindred contacts all fall within the evidential scheme the section builds. The deeming structure matters. Section 4 does not create a free-standing offence of contact with a foreign agent in the abstract. It ties communication of the specified kinds to the Section 3 offence, so that the gravamen remains spying, and the contact operates within the prosecution’s proof of it. Read alongside Section 3(2), which permits the prejudicial purpose to be inferred from conduct and circumstances, Section 4 completes the evidential architecture of the spying offence: the substantive verbs in Section 3, the inferential licence in Section 3(2), and the contact provisions in Section 4 together describe how a prosecution for espionage is expected to be built out of movements, materials and associations rather than out of confession or interception alone.

The breadth of the foreign agent concept deserves the same unflinching statement as the breadth of Section 5 received. A person reasonably suspected of employment by a foreign power for prejudicial purposes is within the definition, and suspicion is a state of the investigator’s information, not a finding of any court. Contact provisions built on such a definition are capable of reaching the ordinary traffic of diplomacy-adjacent life, journalism, academic exchange and commerce, and they are confined in practice by the same two disciplines that confine the rest of the Act: the requirement that the Section 3 offence itself, with its prejudicial purpose, be made out, and the Section 13 sanction that stands between investigation and prosecution. The section is a reminder that the 1923 drafters legislated for a world of physical rendezvous, addresses and letters, and that its categories have been carried, unamended in this respect, into a world of electronic contact whose every message is a communication and whose every contact list is an address book. A successor statute would have to rebuild Section 4 for that world, deciding which contacts with which foreign persons should bear evidential significance and which should be beneath the criminal law’s notice. The enacted text makes no such discrimination, and the discrimination is supplied, again, outside the statute, by prosecutorial judgement under governmental sanction.

Extent, application and the person of the offender

The Act extends to the whole of India, and its personal reach is wider than its territorial phrase suggests. It applies to offences committed by citizens of India outside the country, a standard provision of Indian penal statutes of its era, and it applies within the territory to every person regardless of citizenship, including the foreign national whose activities on Indian soil are the classic subject of the spying offence. Its offences are not confined to government servants, a point worth labouring because the statute’s popular image is of a discipline code for officials. Sections 3 and 5(2), and the supporting offences, are addressed to any person. Section 5(1) is addressed to persons who hold protected material through the custody routes it lists, a class that includes officials but extends, through confidence, contract and employment, well beyond the public service. The one provision with an official’s flavour, the duty to give information under Section 8, is itself addressed to persons generally in the situations it specifies. The Act is therefore a general criminal statute with an official centre of gravity, and its classification alongside service conduct rules or departmental instructions, a confusion encountered in public discussion, understates its character. Breach of it is crime, tried in the criminal courts under the ordinary procedure as modified by its own provisions, punishable with imprisonment and fine, and carrying the consequences of conviction that any criminal record carries.

The Act’s relationship to the general penal law is one of layering rather than displacement. Conduct within its offences will frequently also constitute offences under the general code, theft of documents, forgery, criminal conspiracy, and under other special statutes concerned with sovereignty and security. The existence of overlapping liability is unremarkable in Indian criminal law, but it has a specific consequence for this statute’s statistics and reputation. Prosecutions for espionage-adjacent conduct may proceed under other provisions for evidential or strategic reasons, leaving the Official Secrets Act uninvoked in cases that are, in substance, secrecy cases, while the Act is invoked by name in official discourse in cases that never become prosecutions at all. Both movements depress the ratio of prosecutions to invocations that the brief for this article notes, and both should caution readers against treating the Act’s charge sheet history as a measure either of its deterrent success or of its practical importance. The statute’s importance lies in the liability it holds in reserve across the whole field, available to a sanctioning government, structured by its elements, and shadowed by its undefined centre.

The amendments that did happen

A statute described as unamended in substance across a century invites the question what amendment it has in fact received, and the honest answer sharpens rather than blunts the description. The Act was amended in the decades after independence at points that mattered to its severity and its reach. The amendment of 1967 raised the ceiling of the Section 3 punishment for the defence-related class from seven years to fourteen, doubling the statute’s maximum at its gravest point, and the same era of amendment wrote the sovereignty and integrity and foreign relations limb into Section 3(1)(c), detaching part of the information offence from the enemy-usefulness test. Amendments adjusted the application of Section 5 to material connected with foreign ships and aircraft in defined circumstances. References within the Act were conformed, across the years, to constitutional change: the affairs of the Government replaced the affairs of the Crown in the grading clause, the appropriate Government succeeded the authorities of the colonial text, and the procedural references were aligned with the successive codes of criminal procedure. What no amendment did is equally enumerable. No amendment inserted a definition of a secret. No amendment introduced a classification scheme, a damage threshold for Section 5, a public interest defence, a declassification rule, a time limit on the custody obligation, or a reporting obligation on the sanctioning government. The amendments adjusted the statute’s voltage at its gravest point and widened one limb of its central offence. They left the architecture, and the missing centre, exactly as enacted. That pattern, legislative attention to severity without legislative attention to definition, is itself evidence in the reform debate: Parliament has been willing to make the Act heavier and has not been willing, or has not been asked with sufficient force, to make it more precise.

Classification outside the statute: the manual that does the defining

Because the Act defines nothing, the operative law of what is secret in India is a body of executive instruction. The Manual of Departmental Security Instructions, issued by the Ministry of Home Affairs, prescribes the classification of official records, the gradings of top secret, secret, confidential and restricted, the criteria attaching to each grade by reference to the damage unauthorised disclosure would cause, and the handling rules, marking, custody, transmission, review and destruction, that follow the grade. Classification decisions are taken inside departments by officers authorised for the purpose. Declassification and downgrading follow administrative schedules and decisions, not statutory events. The manual is an administrative instrument. It is not laid before Parliament as legislation is, it is not enacted, and its text has historically been treated within government as itself a protected document, although its existence, its grading structure and its general content are described in official reports, including the First Report of the Second Administrative Reforms Commission, which discussed the classification regime in the course of recommending the repeal of the Act.

The constitutional oddity of this arrangement is worth dwelling on, because it is the mature form of the article’s central point. In a legal order that insists, as a matter of basic criminal law principle, that offences be defined by law with sufficient certainty that a citizen can know in advance what is prohibited, the subject matter of the Official Secrets Act offences is fixed by an instrument that is not law in that sense at all. The citizen’s practical position is mediated by markings: a document marked secret announces its own protected character, and a person who handles marked material knows the executive’s position even if the statute’s boundaries remain untested. But markings are neither universal nor conclusive. Material may be protected under Sections 3 and 5 without bearing any marking, because the sections turn on the nature of the material and its route to the holder, not on a stamp. Material may bear a marking whose grade reflects departmental caution rather than any security judgement that a court would share. And the classification decision itself, taken inside the executive, is the very decision that determines the practical scope of a criminal offence the executive also sanctions prosecutions under. None of this makes the classification system improper as an administrative tool. Every administration must grade its records. The point is narrower and stronger: in this statute, the administrative tool does double duty as the effective definition of a crime, and it does so because Parliament placed no definition in the Act and has placed none there in a century.

Successive official examinations noticed this. The Second Administrative Reforms Commission treated the classification regime as part of the secrecy culture it was asked to reform, recommended that disclosure be the norm and secrecy the exception, and addressed the grading of information in the course of its recommendations on the transparency statute’s exemptions. The committee of secretaries that examined the Act against the Right to Information Act took up classification and declassification among its subjects. The reform record is examined as a whole below; here, classification belongs to the anatomy of the Act in operation. A map of the statute that omitted the manual would omit the place where the statute’s central term acquires its content, and a defence of the statute that omitted it would be defending an abstraction.

The classification regime in detail

Because the classification regime does the defining work of the statute, its internal structure deserves the detailed treatment a statute pillar owes to any provision that carries operative weight. The gradings descend by expected damage. Top secret is the grade for information whose unauthorised disclosure could be expected to cause exceptionally grave damage to national security or the national interest, and it is reserved for the closest secrets of the State. Secret covers information whose disclosure could be expected to cause serious damage. Confidential covers information whose disclosure could be expected to cause damage, and restricted covers information whose disclosure could be expected to cause undesirable effects short of damage in the graded sense, a category whose existence signals that classification reaches material whose sensitivity is administrative as much as strategic. Each grade carries handling consequences: who may originate a document at that grade, how it is marked, registered, copied, carried, stored and destroyed, who may see it, and how its custody is recorded. The regime is, in administrative terms, complete. It answers every question a records officer can ask. What it cannot answer is the criminal law’s question, because its grades measure expected damage in the executive’s estimation and its criteria are applied by the executive to its own documents, so that a court trying a Section 5 charge receives the classification as a fact about the material’s treatment rather than as a legal test the statute has supplied.

Two features of the regime compound the structural point. The first is over-classification, recorded by examiners of the system and familiar to every user of the transparency statute: classification is costless to the classifier and costly to omit, so material drifts upward in grade and outward in volume, and the protected universe grows beyond the material whose disclosure would damage anything. The Second Administrative Reforms Commission’s insistence that disclosure be the norm and secrecy the exception was aimed at exactly this drift, and its observation that information officers refused requests by pointing to classification markings states the drift’s transparency cost. The second feature is the weak discipline of the downward direction. Declassification exists in the regime as schedule and decision, but no statutory event forces it, no presumption of expiry attaches to any grade, and the decision rests in the department that classified. The result is an asymmetry with legal consequences: entry into the protected universe is an administrative act taken at the document’s birth, exit from it is an administrative act that may never be taken, and between the two the criminal statute waits, its offences indifferent to the age of the material. A reform draft serious about the definition problem would attack the asymmetry rather than the grades, by writing into the successor statute a presumption that classification expires unless renewed, a device the administrative regime contains in embryo and the criminal statute ignores entirely.

Invocation and prosecution: the gap that defines the Act in practice

The brief for this pillar directs attention to a fact that commentators across the debate record: prosecutions under the Act are rare relative to the frequency with which the Act is invoked. The invocation takes many forms short of prosecution. The Act is named in official statements when documents surface in public. It is named in correspondence declining to share records, sometimes accurately as a description of the material’s protection and sometimes loosely as a synonym for confidentiality, although the transparency statute’s exemptions, not the secrecy Act, are the operative law of access refusal. It is named in warnings to officials and former officials about their obligations, in departmental instructions, and in the advice tendered to editors and researchers who approach protected subjects. In each register the Act does work without entering a courtroom: it raises the perceived cost of handling official information, it supplies a vocabulary of prohibition, and it marks the boundary between the administration’s information and the citizen’s enquiry with the authority of a criminal statute. The work is real, and it is not measured by conviction statistics, which record only the small residue of cases that pass the sanction filter and the evidential tests.

The gap has rival explanations, and the stance charter requires them to be separated. The retentionist explanation is that the gap is the safeguard system working: a broad reserve power, a political sanction threshold, and courts that insist on elements will predictably produce many invocations and few prosecutions, because most conduct warned against is deterred or is found on examination not to meet the elements, and the residue is prosecuted. The critical explanation is that the gap is the design’s pathology: a statute whose offences are wider than its enforcers’ willingness to prosecute transfers the effective law from the statute book to enforcement discretion, and invocation without prosecution is discretion in its purest form, the threat of a law whose boundaries no one is required to test. Both explanations fit the observable gap, and the choice between them turns on a prior judgement about how far executive restraint can substitute for legislative precision in criminal law. This pillar has already given its answer in the verdict above, and consistency requires it to be repeated here in application: restraint is a virtue in a prosecutor and a defect in a statute, because a citizen ordering his affairs is entitled to the boundary the law states, and a boundary stated too widely and narrowed only by practice is a boundary the citizen cannot find. The gap is therefore counted, in this article’s assessment, on the critical side of the ledger, while the fact that Indian practice has produced no scandal of mass prosecution under the Act is counted, with equal honesty, on the retentionist side.

The transparency statute: override without repeal

The Right to Information Act 2005 changed the legal environment of the Official Secrets Act without amending a word of it, and the precise mechanics of that change are the most frequently misstated point in the whole field. The transparency statute must therefore be stated exactly. Section 22 of the Right to Information Act provides that its provisions shall have effect notwithstanding anything inconsistent therewith contained in the Official Secrets Act 1923, and in any other law in force, or in any instrument having effect by virtue of any law other than the transparency statute itself. The official secrets statute is named in the override. Section 8(1)(a) of the transparency statute exempts from disclosure information whose disclosure would prejudicially affect the sovereignty and integrity of India, the security, strategic, scientific or economic interests of the State, relations with a foreign State, or lead to incitement of an offence. Section 8(2) then provides that a public authority may allow access to information if the public interest in disclosure outweighs the harm to the protected interests, notwithstanding anything in the Official Secrets Act or any of the exemptions in Section 8(1). The reading of these provisions together, which this series’ treatment of the Right to Information Act 2005 states and which this article follows, is that the transparency statute governs access to information held by public authorities and prevails over the secrecy statute within that field, including, through Section 8(2), a public interest route to disclosure of material that would otherwise be exempt, while the offences in the Official Secrets Act remain on the statute book and continue to govern conduct that falls outside the transparency framework.

For a question of access, which of the two statutes prevails?

For access, the transparency statute prevails. Section 22 of the Right to Information Act gives it effect notwithstanding anything inconsistent in the Official Secrets Act, and Section 8(2) adds a public interest route to disclosure. Outside the access framework the position reverses: the 1923 offences stand unrepealed and govern unauthorised taking, communication and receipt.

The transparency statute overrides the secrecy statute for access, and does not repeal its offences. Section 22 of the Right to Information Act gives that Act effect notwithstanding anything inconsistent in the Official Secrets Act 1923, and Section 8(2) permits a public authority to allow access where the public interest in disclosure outweighs the harm to the protected interests, notwithstanding the Official Secrets Act and the Section 8(1) exemptions. Conduct outside the transparency framework, unauthorised taking, communication and receipt of protected material, remains governed by the 1923 Act’s offences, which stand unrepealed.

The boundary this creates is workable but not comfortable, and its discomforts are structural. A citizen who obtains a document through the transparency statute, after a public authority has applied Section 8(2), holds it lawfully, whatever its classification markings. A citizen who obtains the same document through an unauthorised channel holds it under the shadow of Section 5(2). An official who discloses under the transparency framework is protected by it; an official who discloses the same material outside it, even from a public interest motive identical in substance, faces Section 5(1) without a statutory public interest defence, because the 1923 Act contains none. The two statutes thus divide the world by channel rather than by content, and the division is deliberate on the transparency side and inherited on the secrecy side. It is also worth recording what the override did not do. It did not amend the secrecy Act’s offences, insert a definition, require declassification, or enact the replacement chapter that the Administrative Reforms Commission had recommended the year after the transparency statute commenced. The 1923 Act was left standing next to a statute that repudiates its premise, disclosure as exception, and substituted the opposite premise within the access field, disclosure as the rule. Statutes in that posture coexist until Parliament resolves the tension, and the resolution has been awaited through the whole period the reform record covers. The contrast with other secrecy-bearing fields in this series is sharp: statutes governing the wider security and disclosure field, from the citizenship code whose architecture the guide to citizenship law in India maps, to the secrecy provisions scattered through fiscal and regulatory law, each locate their secrecy rules inside a scheme Parliament has revisited. The 1923 Act alone carries a pre-constitutional scheme into the transparency era substantially untouched.

The exemptions deserve a final precise word, because Section 8(1)(a) is where the two statutes interlock most tightly. The exemption’s language, prejudicial effect on sovereignty and integrity, security, strategic, scientific or economic interests, relations with a foreign State, tracks the subject matter of the secrecy offences without copying their elements. An information commission deciding an access appeal under the exemption is performing a different exercise from a criminal court trying a Section 3 charge: the commission weighs disclosure of identified information against identified heads of harm, frequently with the Section 8(2) balance in view, while the criminal court asks whether an accused acted with a prejudicial purpose in relation to material of the statutory kinds. Confusion between the two exercises produces bad law reporting and worse public debate, and the discipline of keeping them apart is one of the practical services a statute pillar can perform.

The courts and published documents: the 2019 judgment

The leading modern authority on the interaction between the Act and publication arose in proceedings for the review of an earlier judgment of the Supreme Court, and it reached the Court in the form of a preliminary objection by the Government. The review petitioners relied on documents that had been published by a newspaper. The Government objected that the documents were privileged and had been improperly obtained, that reliance on them was barred by the provision of the evidence statute concerning unpublished official records relating to affairs of State, and that the documents were protected under the Official Secrets Act. A three-judge Bench, in two separate but concurring opinions delivered on 10 April 2019, the lead opinion of the Chief Justice speaking also for a second judge and a concurring opinion by the third, dismissed the preliminary objection and held that the documents could be considered. In accordance with the discipline this article observes, the underlying subject matter of the review is not described here. The holding is what matters, and the holding has three strands.

The first strand concerns the evidence statute. The privilege provision applies to unpublished official records, and documents that have been published in a newspaper are, by the time they reach the court, in the public domain. The Court reasoned that the fact of publication transformed the material’s position: the three documents had been published in the newspaper and were available in the public domain, a position the Government did not seriously dispute, and neither the evidence statute’s provision nor the Official Secrets Act furnished a ground to exclude material already public from the Court’s consideration in adjudicating the dispute before it. The second strand concerns prior restraint. The Court observed that no provision of the Official Secrets Act, and no other law brought to its notice, vested in the executive a power to restrain the publication of documents marked as secret, or to bar the placing of such documents before a court of law called upon to adjudicate an issue. The Act creates offences. It does not create a censorship power, and the executive cannot borrow from the existence of offences a power of suppression that Parliament never conferred. The third strand concerns the press. Recalling the Court’s jurisprudence on the freedom of the press within the guarantee of freedom of speech and expression, the Court treated the publication of the documents by the newspaper as consonant with the constitutional guarantee, no law enacted by Parliament specifically barring such publication on the grounds enumerated in Article 19(2) having been shown. The concurring opinion added a distinct emphasis: after the enactment of the Right to Information Act, the Government’s claim of privilege stood diluted, because the transparency statute had recast the premises on which secrecy claims are received, and the balance between openness and security had to be struck in that altered setting.

What did the 2019 Bench decide about documents already in print?

The Bench dismissed the Government’s preliminary objection and held that documents published by a newspaper and standing in the public domain could be considered by the Court, that neither the unpublished-records privilege nor the Official Secrets Act excluded them, and that the Act confers no power on the executive to restrain publication of documents marked secret.

Deciding a preliminary objection in review proceedings in April 2019, the Supreme Court held that neither the evidence statute’s privilege provision for unpublished official records nor the Official Secrets Act barred the Court from considering documents already published by a newspaper and standing in the public domain. It found no provision in the Act empowering the executive to restrain publication of documents marked secret, treated the newspaper’s publication as consonant with the constitutional guarantee of freedom of speech, and dismissed the objection.

The judgment’s limits are as important as its holdings and should be stated with the same care. The decision is a decision about the consideration of documents by a court in proceedings before it, and about the absence of a suppression power. It is not a decision that publication of protected material can never constitute an offence under Section 3 or Section 5, a question that was not before the Court in that form and on which the judgment decides nothing. It is not a charter for the taking of documents, and nothing in it qualifies the offences that govern unauthorised acquisition and communication. What it settles is narrower and still considerable: material that has entered the public domain through publication does not carry into court a privilege that excludes it, the secrecy statute confers no power of prior restraint, and the freedom of the press is a live constitutional presence in this field and not a courtesy. For journalists and publishers, the decision is the leading modern authority precisely because it separates questions that public debate habitually merges, the legality of obtaining, the legality of publishing, and the usability of published material in adjudication, and answers the third without pretending to answer the first two. The separation also disciplines the expressive freedom analysis that runs through this series, visible in the treatment of the Shreya Singhal judgment on free speech, where the Court likewise insisted that a statute be judged by the power it actually confers and the elements it actually contains rather than by the fears or hopes projected onto it.

The recipient, the journalist and the researcher

The recurring public question, whether a journalist can be prosecuted under the Act, has a textual answer that is more precise than the debate usually allows. No provision of the Act names the press. The Act reaches a journalist, as it reaches any citizen, through provisions of general application: Section 3, where publication is among the verbs and the prejudicial purpose must be proved; Section 5(2), where voluntary receipt of protected material, with knowledge or reasonable ground to believe it was communicated in contravention of the Act, is an offence; and the supporting provisions on attempts and harbouring. Against that reach stand three qualifications, each substantial. The knowledge condition in Section 5(2) confines the receipt offence to knowing receipt of material communicated in breach, and excludes the recipient who takes material without that knowledge or ground for belief. The purpose element in Section 3 excludes the publisher who publishes without a purpose prejudicial to the safety or interests of the State, and the prosecution bears the burden of establishing that purpose. And Section 13 interposes the appropriate Government’s complaint between any investigation and any prosecution, so that a decision to proceed against a journalist for receipt or publication is a decision a government must own at the threshold. The honest summary is that the Act does reach the press through its general provisions, that its elements and its sanction filter confine that reach in law, and that the distance between the reach in law and the apprehension in practice is itself one of the statute’s effects, since a provision invoked in argument, in correspondence and in official warnings may shape editorial behaviour without any prosecution being instituted or sanctioned.

Through which provisions does the Act reach the press?

The Act contains no provision addressed to the press as such. It reaches journalists through provisions of general application: Section 3, whose verbs include publishing and whose safeguard is the prejudicial purpose, and Section 5(2), which reaches voluntary receipt with knowledge of a breach. Section 13’s governmental complaint stands between any such case and a court.

A journalist stands in the same position as any other person under the Act’s general provisions. Section 3 names publishing among its verbs but requires proof of a purpose prejudicial to the safety or interests of the State. Section 5(2) reaches a person who voluntarily receives protected material knowing, or having reasonable ground to believe, that it was communicated in contravention of the Act. No prosecution can begin at all unless the appropriate Government complains under Section 13, which is the filter any case against a journalist must first pass.

The researcher’s position follows the same lines with one addition. Academic and journalistic work with archival material raises the question of age. The Act contains no provision under which material loses protection by passage of time. Declassification is an administrative event governed by departmental schedules, and disclosure through the transparency statute is a channel governed by that statute’s exemptions and balance. Material that remains classified is, in the Act’s terms, as protected in its fiftieth year as in its first, unless and until the executive declassifies it or a public authority releases it. This is one of the points at which the Act’s fit with a working historical discipline is most strained, and it featured in the considerations that led official examiners to take up declassification when they reviewed the Act against the transparency statute. It is also a point where the practical advice embedded in this article’s legal-information character should be plain: this article states the law of a statute in general terms for information, and a person handling specific protected material requires legal advice on the specific facts, because the elements turn on classification, custody and authority in ways that general exposition cannot resolve.

The disclosure side of the researcher’s position connects to a neighbouring statute in this series. An official who encounters wrongdoing and discloses it outside the transparency channels faces Section 5 without a public interest defence in the Act, and the protective scheme enacted for such disclosures stands in the Whistle Blowers Protection Act, whose machinery and limits this series has examined separately. The junction between the two statutes is one of the least satisfactory in the field: the whistle blower statute contemplates disclosure of corruption and abuse of power through designated channels, the secrecy statute criminalises unauthorised communication of protected material without exception for motive, and an official whose information falls in both territories must navigate the junction with the channel discipline of the one statute as the only safe route through the other. A reform settlement that took the junction seriously would say, in the secrecy statute itself, what protection a disclosure through the designated channel enjoys. The enacted texts do not say it with the clarity the subject deserves.

The artifact: protected by what

The table that follows states, for five recurring categories of information, whether the Act applies, what must be established, whether the transparency statute overrides, and what a recipient’s position is. It is the findable artifact of this pillar, and its last two rows are the point of the exercise: the row on published information is settled by the leading decision, and the row on classified documents rests on a classification the Act itself never addresses.

Category of information	Does the Act apply?	What must be established	Does the transparency statute override?	A recipient’s position
A document marked under the departmental classification manual	Yes, through Sections 3 and 5 where its elements are met; the marking itself is not an element of any offence	For Section 3, a prejudicial purpose; for Section 5, protected status, custody through office, confidence or contract, wilful communication and absence of authority	The marking does not bind an access decision; Sections 22 and 8(2) of the transparency statute govern access by public authorities	Receipt is an offence under Section 5(2) only where the recipient knows or has reasonable ground to believe the communication breached the Act
A document relating to a notified prohibited place	Yes; the prohibited place connection is an express class in Sections 3 and 5	The place’s status under Section 2(8), by the statute’s force for defence establishments or by notification for declared places	Access decisions remain governed by the transparency statute, subject to the Section 8(1)(a) exemption and the Section 8(2) balance	The same knowledge condition governs receipt; the place’s notified status is a fact the recipient can verify from the notification
Information entrusted in confidence to an official	Yes; confidence is an express custody route in Section 5(1)	That the material was entrusted in confidence by a person holding office under the Government, and wilful communication to an unauthorised person	Entrustment does not displace the transparency statute for material held by a public authority	A recipient who knows of the confidence and of the breach stands within Section 5(2)
Information already published	The Act’s offences are not a bar to a court’s consideration of it, and the Act confers no power to restrain its publication	For any prosecution, the ordinary elements of the offence charged; publication alone does not supply the prejudicial purpose under Section 3	The transparency question does not arise in the same form, the material being already public	The 2019 judgment treats newspaper publication as consonant with the freedom of speech guarantee; receipt of published material is not the receipt Section 5(2) addresses
Information obtained under the transparency statute	No offence is committed by holding or using material lawfully released through that channel	That the release was made by a public authority under the transparency statute, including under the Section 8(2) public interest route	The transparency statute is the governing law for the release, by force of Section 22	The recipient holds the material lawfully, whatever classification marking it bears

The table’s disciplined modesty is deliberate. It does not purport to decide cases, and its rows state positions that the text of the two statutes and the one leading judgment support. Where a category depends on facts, the classification actually borne by a document, the notification actually issued for a place, the authority actually held by a communicator, the table says what must be established rather than pretending the fact away. That is the most a reference table on this statute can honestly do, and the discipline generalises: on a statute whose subject matter is fixed administratively, every sound statement of the law is conditional in form, and unconditional statements about what the Act protects should be distrusted in proportion to their confidence.

The reform record: recommended for repeal, examined, and left standing

The Act has been officially examined, in whole or in part, at least four times since independence, and the record of those examinations is the strongest single body of evidence in the retention debate, because it is evidence generated by the State about its own statute. The Law Commission of India, in its Forty-third Report of 1971 on offences against national security, considered the field of security legislation and recommended consolidation of the laws dealing with national security, a recommendation the Administrative Reforms Commission later recalled with approval. The Shourie Committee of 1997, a working group on the right to information and transparency chaired by H. D. Shourie, recommended a recast of Section 5(1) that would have confined the penal provision to violations affecting national security, supplying a draft of the narrowed clause. The pattern of the early record is consistent: examiners who looked at the Act found its disclosure offence too wide and said so in terms, while stopping short, in the Law Commission’s case, of recommending the Act’s disappearance rather than its consolidation.

The central document in the record is the First Report of the Second Administrative Reforms Commission, Right to Information: Master Key to Good Governance, submitted in June 2006 under the chairmanship of Veerappa Moily, a year after the transparency statute commenced. The Commission’s treatment of the Act is unqualified in direction. It recorded that the Act had been regarded in many quarters as primarily responsible for the excessive secrecy in government, observed that in its form on the statute book the Act was an anachronism beside the transparency regime, and held that disclosure had to be the norm and secrecy the exception. Its recommendation, at paragraph 2.1.12 of the summary of recommendations, was that the Official Secrets Act 1923 should be repealed, and substituted by a chapter in the National Security Act containing provisions relating to official secrets, and that the equivalent of the existing Section 5 in the new law should be drawn on the lines the Shourie Committee had recommended, confining the offence by reference to national security. The Commission also agreed with the Law Commission’s older case for consolidating national security law. The Government’s response, recorded in the action-taken material on the report, did not accept the repeal recommendation: objections recorded at the time included that the offences of a substantive penal statute could not suitably be inserted into the National Security Act, an enactment directed to preventive detention, and that a fresh classification exercise of the kind contemplated was not practicable. The recommendation was therefore made by the Government’s own commission, considered by the Government, and not implemented. The Act continued.

Which official body recommended repeal, and with what replacement?

The Second Administrative Reforms Commission recommended repeal in its First Report of June 2006, proposing substitution by a chapter in the National Security Act containing provisions on official secrets, with the disclosure offence redrawn on the narrower Shourie Committee lines. The Government did not accept the recommendation, and no replacement chapter was enacted.

Yes, by an official body of the highest standing. The Second Administrative Reforms Commission, in its First Report on the Right to Information submitted in June 2006, recommended that the Act be repealed and substituted by a chapter in the National Security Act containing provisions on official secrets, with the successor to Section 5 confined on the lines the Shourie Committee had proposed. The Government did not accept the recommendation, and the Act remains in force unrepealed.

The later examination took a different route to a related destination. In 2015, following a decision taken at a meeting chaired by the Cabinet Secretary, the Government constituted a committee of the Secretaries of Home, Law and Personnel to study the Official Secrets Act in the light of the requirements of the Right to Information Act, including the provisions governing the declassification of older records. The committee reported in 2017 with recommendations directed to making the Act more transparent and more consistent with the transparency statute. Those recommendations, like the Commission’s before them, were not translated into amendment. The statute book therefore carries a peculiar and precisely describable position: a colonial secrecy Act, recommended for repeal by an administrative reforms commission in 2006, examined for harmonisation by a committee of secretaries reporting in 2017, and amended in neither respect, surviving alongside the transparency statute that overrides it for access and the constitutional jurisprudence that confines its use against published material. It also survived, untouched, the great codification of the criminal law enacted in 2023, in which Parliament replaced the Penal Code, the procedure code and the evidence statute with new codes. That codification did not absorb the Official Secrets Act, did not repeal it, and did not reproduce its offences in the new general code. The Act stands as a special statute outside the codes, exactly where it stood before them, and any account of its position that describes it as swept into the 2023 settlement misdescribes the record.

Longevity of this kind has two readings, and the stance charter requires both. On one reading, repeated examination followed by retention is a form of democratic verdict: governments of different compositions looked at the Act, received advice, and kept it, from which it may be inferred that the Act does work that its examiners’ alternatives did not credibly promise to do. On the other reading, the record shows advice converging on reform and decisions declining to act on it, from which it may be inferred that inertia and the convenience of an available secrecy statute, rather than a reasoned preference for this statute over its alternatives, explain the retention. The evidence in the record supports the second reading more strongly than the first, for a specific reason: the recorded governmental responses engaged with the machinery of the Commission’s proposal, the destination statute for the replacement chapter, the practicability of reclassification, rather than with the Commission’s central charge, that Section 5 as enacted is wider than national security requires. A rejection that answers the packaging and not the substance is evidence of a question deferred, not of a question decided. This article’s verdict section returns to the point.

The case for retention, in its strongest form

The retention case begins from a premise no serious participant denies: every State protects defence and intelligence information, and a State that enacted no secrecy offences would not thereby become more virtuous, only less able to protect material whose disclosure can kill. The Indian State maintains armed forces, intelligence agencies, nuclear and space programmes, and negotiation positions, and the information those activities generate requires legal protection against espionage and unauthorised disclosure. The question is never secrecy or no secrecy. It is the design of the secrecy law, and the retentionist argues that a general statute has design advantages that its critics underrate. A general statute avoids the gaps of an enumerative one. A scheme of specific provisions, each tied to a category of information or a class of holder, invites argument at every boundary, and hostile intelligence work is precisely the activity best placed to exploit boundary argument. The 1923 Act’s generality, its coverage of material by custody and character rather than by list, is in this view not a defect but the point.

The retentionist continues with the statute’s operating record. The Act contains, in Section 13, a sanction requirement that makes every prosecution a governmental decision, and the courts have insisted that Section 3’s prejudicial purpose be proved. The combined effect, on this account, is a statute that is broad in text and narrow in operation: prosecutions are rare relative to the frequency with which the Act is invoked in official discourse, a disproportion the brief for this article directs attention to and which commentators across the debate record, and rarity of prosecution under a sanction filter is what a well-designed reserve power looks like. The retentionist adds a comparative caution. Foreign models are not obviously happier: the British replacement statute of 1989 generated its own controversies about category coverage and damage tests, and the United States, which has no general official secrets statute of the Indian type, protects the same material through a dense thicket of espionage, classification and computer misuse law whose aggregate breadth rivals the Indian Act’s. On this view, repeal and replacement would exchange a known instrument, glossed by a century of practice and confined by known safeguards, for an unknown one whose boundaries would take a generation of litigation to settle, during which the protection of defence information would be at its weakest. Finally, the retentionist makes a constitutional point: Article 19(2) permits reasonable restrictions on the speech guarantee in the interests of the sovereignty and integrity of India and the security of the State, and a secrecy statute confined to those interests is within the restriction power, as the courts’ treatment of the Act, including the 2019 judgment’s careful refusal to strike at its offences, tends to confirm.

The case for repeal or replacement, in its strongest form

The repeal case begins from the text. A statute enacted in 1923, modelled on the legislation of an imperial power, framed for an administration that answered to no electorate and recognised no citizen’s right to information, is asked to operate in a republic whose Constitution guarantees free expression and whose Parliament has enacted a transparency statute premised on disclosure. The repeal advocate argues that fit matters: legal instruments carry the assumptions of their design, and the 1923 Act’s assumptions, official information as administrative property, disclosure as disloyalty, secrecy as the default, are the assumptions the Republic has elsewhere repudiated. The absence of a definition is then not a technical blemish but the signature of the design. A statute in which the scope of a criminal offence is set administratively, by classification under an executive manual, fails the elementary requirement that criminal liability rest on law stated with certainty, and fails it at the centre, on the element that decides what the offence is about.

The repeal advocate continues with Section 5’s structure. An offence of wrongful communication that requires no proof of harm, no proof of hostile intent, no proof even that the material’s disclosure could damage anything, and that extends to a knowing recipient, places the ordinary traffic of a democratic society, the briefing, the leak to expose wrongdoing, the handing of a document to a journalist or a researcher, inside the formal boundaries of crime, subject to executive discretion about enforcement. A law whose real content is supplied by enforcement discretion is, on this view, objectionable in itself, whatever the discretion’s record, because citizens are entitled to have the boundary of crime set by the law and not by the prosecutor’s restraint. The advocate then deploys the official record: the State’s own commission recommended repeal and a narrowed replacement, the State’s own committee recommended harmonisation, and nothing was done, so that the retention of the Act in its enacted breadth cannot be presented as the considered preference of the Republic’s institutions over the alternatives their examiners designed. The advocate closes with the chilling effect, the gap between prosecutions brought and behaviour altered. A statute invoked constantly and prosecuted rarely does its work in the second register: in the official who declines to share the harmless document, the editor who declines the story, the archivist who declines the file, none of which appear in any statistic and all of which are costs a transparency republic should count. The retentionist’s rarity of prosecution, on this reading, is not evidence of a safeguard working. It is evidence of a threat working, which is a different thing and a worse one.

Verdict

Both cases are honestly stateable, and this pillar is required to choose. The choice it makes is for replacement: the repeal advocate has the better of the argument on the statute as enacted, and the retentionist has the better of the argument only on the narrower question whether a general official secrets law of some kind should exist, which the repeal recommendation itself never denied. The Administrative Reforms Commission’s proposal was not abolition of secrecy law. It was repeal of this Act and substitution of a chapter in national security legislation, with the disclosure offence confined on the Shourie lines to violations affecting national security. That proposal threads the two cases together: it keeps the general protection the retentionist rightly insists every State needs, and it cures the two defects the repeal advocate rightly identifies, the undefined subject matter and the harm-free communication offence, by writing the successor offence around national security effect. The Government’s recorded objections answered the proposal’s destination, not its substance, and a decade and a half of further coexistence with the transparency statute has strengthened rather than weakened the underlying case, because every year of coexistence generates fresh instances of the channel-based boundary, lawful by one route, criminal by another, that a successor statute could rationally redraw. Replacement drafted on those lines, a definition or a damage threshold in the statute, a public interest consideration for the disclosure offence, the sanction filter retained, would be an improvement on the enacted position from every standpoint except that of an administration that prefers its secrecy law undefined. The statute’s survival is therefore best explained not as a verdict on its merits but as the residue of decisions taken about other things, and the series thesis finds here its most complete instance: the operative content of the law of official secrecy in India sits outside the statute that bears its name, in a classification manual, in an override located in another Act, in a safeguard located in a procedural section, and in recommendations located in reports that were never enacted.

A colonial statute retained: the Act among its contemporaries

The Official Secrets Act is not the only colonial statute the Republic retained after argument, and placing it among its contemporaries clarifies what kind of survival its survival is. The law of contempt, examined in this series in the treatment of the Contempt of Courts Act 1971, descends from a colonial contempt jurisdiction that the Republic re-enacted in fresh statutes, debated in Parliament, and amended, most significantly by inserting a truth defence in 2006. Whatever view is taken of contempt law, its Indian form is the product of post-independence legislative choice, repeatedly renewed. The sedition provision of the old Penal Code travelled the opposite road to a judicially supervised suspension and then to the codification exercise. The Official Secrets Act’s road is neither of these. It was not re-enacted in Indian form, as contempt was. It was not carried into the new codes, as the offences against the State were in recast form. It was left, in its 1923 text, amended at the margins across the decades, the fourteen-year ceiling raised by the amendment of 1967, the sovereignty and integrity limb added to Section 3 in the same era, and the substance, the offences, the undefined subject matter, the sanction frame, otherwise continuous with the enactment the colonial legislature passed. A colonial statute retained in that posture makes a distinctive claim on a reader’s attention. Its authority is the authority of enactment, as with any statute. Its claim to fit the constitutional order is only as strong as the examinations it has survived, and the examinations, as the reform record shows, recommended against it in its enacted form and were not acted on. Retention by survival is legally sufficient and intellectually thin, and the gap between those two judgements is the space in which the Act continues to operate.

The documentary life of the Act deserves a section of its own, because the statute is encountered by most citizens not as law but as paper: the classification marking on a file cover, the secrecy grading on a noting sheet, the marked annexure to a published report. Government records reach the public, when they reach it, as documents to be opened, read and compared, and the practical handling of released official records, files disclosed through the transparency statute, committee reports, action-taken memoranda and published judgments, is document work. Readers assembling a documentary record of a secrecy question, the Act’s text beside the Commission’s report beside the transparency statute, will often work with scanned and converted files, for which a browser-based viewer for office documents and record files is the kind of tool the work calls for. The observation is mundane and worth making in a pillar on this subject: the law of official secrecy is a law about documents, its offences are committed with documents, its classification is marked on documents, its leading modern judgment concerned documents published in a newspaper, and its reform history is a history of documents that governments commissioned, received and shelved. To understand the Act is to follow paper through institutions, and the institutions leave the paper trail by which they can be judged.

How does the Act treat a former official?

It treats the former official as it treats the serving one, because Section 5 fastens on the route by which material was obtained. Information obtained owing to a position as a person holding or having held office is within the section, and its wilful communication to an unauthorised person is an offence without any time limit in the text.

The Act reaches a former official through the custody language of Section 5. Material obtained, or accessed, owing to a position as a person holding or having held office under the Government is within the section, and a wilful communication of such material to an unauthorised person is an offence without any time limit in the section itself. The obligation the statute imposes is therefore not a term of service that expires with service. It attaches to the material by the route the holder obtained it, and it travels with the material into retirement.

That single feature reorganises a good deal of practical life around government. The retired official’s memoir, the former adviser’s newspaper column, the ex-officer’s evidence to a researcher, all move in the neighbourhood of Section 5 whenever they draw on material obtained through office, and the neighbourhood has no fence marked lapse of time. What separates the lawful memoir from the offence is the character of the material and the authority for its communication, not the seniority of the author or the public value of the recollection. Successive generations of official writing in India have been shaped by that discipline, visibly in the caution of the published record and invisibly in the record never written. A successor statute drafted for the Republic would have to confront the question the 1923 text answers by silence: whether the custody obligation should persist indefinitely for all grades of material, or diminish with the grade and the years, as declassification schedules already assume for the records themselves. The administrative system recognises that secrecy decays. The criminal statute does not. That asymmetry, records declassified by schedule while the personal obligation runs without limit, is one more instance of the article’s central pattern, the operative judgements about secrecy being taken outside the Act that punishes its breach.

For the serving official, the practical teaching of the statute is channel discipline, and it is worth stating as plainly as the pillar format allows. Material within the protected classes moves lawfully in three ways: within the chain of authority the official’s role supplies, through the transparency statute where a public authority applies its provisions, and under the direction of the sanction and disclosure machinery of government itself. It moves unlawfully by wilful communication outside those channels, whatever the motive, and the Act’s text gives no quarter to motive in the principal limb of Section 5. Officials who believe material evidences wrongdoing are directed by the structure of the law to the designated disclosure channels that other statutes maintain, and not to unauthorised communication, because the secrecy statute will judge the act and not the intention behind it. That is severe, and the severity is a deliberate part of the design the reformers proposed to soften by confining the offence to national security violations, a proposal whose non-enactment leaves the severity in force.

What a successor statute would have to decide

The reform debate is often conducted at the level of slogan, repeal or retention, and the pillar format permits a more useful exercise: stating the decisions any successor statute must take, on which the existing Act takes no position at all or a position by silence. The first decision is the definition. A successor must either define the protected subject matter in the statute, by category, by damage threshold, or by a classification scheme enacted in the statute itself, or must state expressly that classification under executive instruction determines protection, thereby owning in law the arrangement the 1923 Act produces by silence. The second decision is the harm element in the disclosure offence. The Shourie draft and the Commission’s adoption of it chose a national security confinement. A successor could instead adopt a damage test of the kind the British statute of 1989 applies to several categories, or a public interest defence, or some combination, but it must choose in text. The third decision is the recipient. A successor must decide whether knowing receipt should remain an offence, whether receipt by a person who publishes in the public interest should be treated differently from receipt by a person who transmits to a foreign power, and what knowledge standard should govern each case. The fourth decision is time: whether protection should decay by statute, through classification expiry or a long-stop for historical material, or persist until executive declassification. The fifth decision is machinery: whether the sanction requirement should remain an executive filter or be supplemented by reporting to Parliament, by a requirement of reasons recorded in writing, or by review. The sixth decision is the junction with neighbouring statutes, the transparency override and the whistle blower channels, which a successor should state within its own text rather than leaving to implication across three enactments.

Listing the decisions exposes why replacement has not happened, and the exposure is more instructive than another rehearsal of the arguments. Each decision is genuinely contestable. Security establishments will resist a statutory definition as a map for hostile enquiry. Transparency advocates will resist any definition wide enough to satisfy the establishments. The harm element divides lawyers who trust elements from lawyers who trust defences. The recipient question divides the press’s defenders from its critics in ways that align imperfectly with party or ideology. A government that takes up replacement buys six difficult fights at once, for a statute whose defects, however real, produce no daily administrative inconvenience to the government itself. Inertia in such a case is not mysterious, and it is not evidence that the existing settlement is optimal. It is evidence that the costs of the settlement fall on persons, officials, journalists, researchers and citizens, who do not control the legislative timetable, while the costs of change would fall on the institution that does. That distribution explains the century. It does not justify it, and the Commissions that examined the Act, staffed by persons who understood the distribution from inside, said as much by recommending change in full knowledge of its difficulty.

Reading the two statutes together: a worked discipline

For the practitioner, the official and the researcher alike, the usable product of this pillar is a discipline for reading the two statutes together, and it can be stated as a short sequence of questions that the texts answer in order. The first question is channel: is the material being sought through the Right to Information Act from a public authority, or is it moving outside that channel? If through the channel, the transparency statute governs, the Official Secrets Act is overridden by Section 22 for the purposes of the request, and the analysis proceeds under the exemptions in Section 8(1), the public interest balance in Section 8(2), and the severability and third-party provisions of that statute, not under the offences of 1923. If outside the channel, the second question is custody: how did the material reach the person proposing to communicate or receive it, through office, through confidence, through contract, or through none of these? The third question is element: for Section 5, is the communication wilful and unauthorised; for Section 3, is there evidence from which a prejudicial purpose could be proved; for Section 5(2), does the recipient know or have reasonable ground to believe the communication breaches the Act? The fourth question is forum: has the appropriate Government complained under Section 13, without which no court may proceed? The sequence does not decide cases, and it is offered as general legal information rather than advice on any facts. What it does is prevent the characteristic errors of the field: treating classification as an answer to an access request, treating the transparency override as a repeal of the offences, treating the absence of a prosecution as proof that conduct is lawful, and treating the presence of an invocation as proof that it is not. Each error is an error of reading one statute without the other, and the pair, secrecy and transparency, enacted eighty-two years apart for opposite premises, can in practice only be read as a pair.

Proof, search and trial: the Act in the courtroom

The Act’s courtroom life is governed by the interaction of its own procedural provisions with the general law, and the interaction repays a section of its own because it determines what the offences mean for an accused person in practice. The burden throughout is the ordinary criminal burden: the prosecution must prove the elements beyond reasonable doubt, and the special features of the Act adjust the route by which elements may be proved without displacing the burden itself. Section 3(2) is the chief adjustment for the spying offence, permitting the prejudicial purpose to be inferred from the circumstances, the conduct and the known character of the accused as proved, and dispensing with proof of a particular act tending to show the purpose. The provision is sometimes misdescribed as reversing the burden. It does not transfer the burden to the accused. It authorises the court to find the element proved by inference from proved facts, which is the ordinary mechanism for proving intention in criminal law, stated expressly because the drafters knew that direct evidence of a spy’s purpose would rarely exist. The discipline the courts have maintained, that the purpose remains an element to be established and not a presumption to be indulged, is the counterweight that keeps the provision within constitutional tolerances, and a prosecution that invited a court to convict on the act alone, purpose unexamined, would invite reversal.

Section 4’s communication provisions work the same evidential seam from the other side, supplying the contact material from which purpose and association may be inferred. The search provisions of Section 12, applying the warrant machinery of the criminal procedure code with modifications, govern the seizure of the documents and materials on which secrecy prosecutions characteristically depend, since the offences are documentary in their nature and the exhibits are the case. The trial itself is shaped by Section 14’s exclusion power, which permits the public to be excluded where publication of the evidence would be prejudicial to the safety of the State. The in camera device creates a characteristic tension with the open justice principle, which Indian courts treat as a norm of constitutional rank in criminal proceedings, and the section’s own terms supply the reconciliation attempted by the drafters: exclusion is discretionary, it is tied to a finding about prejudice to the safety of the State, and the judgment is pronounced in public even where the evidence was heard in private. A closed evidentiary stage with a public judgment is a compromise that satisfies neither the secrecy interest in full, since reasoning may reveal what evidence concealed, nor the publicity interest in full, since the public cannot test reasoning against evidence it never heard. The compromise is the statute’s own, and it has the merit, rare in this Act, of being visible on the face of the text rather than supplied from outside it.

Sentencing under the Act follows the grading already examined, and the grading itself performs a communicative function that sentencing theory recognises: the fourteen-year ceiling for defence-related espionage announces that Parliament ranks the betrayal of defence information with the gravest non-homicide offences, while the three-year ceilings for the communication, receipt and perimeter offences place them in the middle band of the criminal calendar, serious and imprisonable, short of the exceptional. A successor statute reconsidering the scheme would have to re-perform this communication deliberately, because the ceilings are the part of the Act that public debate most often misquotes, attributing the maximum of the gravest offence to the generality of breaches. Accuracy about grading is therefore not pedantry. It is the difference between a debate about a draconian statute and a debate about a graded one, and the graded description is the true one.

The appropriate Government and the federal distribution of the filter

Section 13’s expression, the appropriate Government, imports the federal structure into the sanction filter, and the importation has consequences that commentary centred on the Union overlooks. For offences connected with the Union’s subjects and establishments, the appropriate Government is the Central Government. For matters within the State field, it is the State Government. The filter is therefore not a single national gate but a distributed one, and the decision to complain about conduct touching, for example, a State-administered installation or a State official’s communication may be taken in a State capital by a government with its own political complexion and its own relations with the press and the opposition of the day. The design has a federal logic, since the government responsible for the information is the natural judge of the prosecution’s wisdom. It also multiplies the points at which the filter’s quality matters, and it means that the safeguard’s real-world strength varies with the strength of sanction practice across governments, a variation no central guideline in the Act disciplines. The point generalises a theme of this pillar: the Act’s safeguards are real and are located in discretion, and discretion distributed is discretion of variable quality.

The same distribution explains a feature of the Act’s enforcement history that puzzles readers who expect a national security statute to be enforced nationally. The offences protect the State in the constitutional sense, the Union and the States as one legal order, and their subject matter crosses the federal line constantly, since defence is a Union subject, police and public order are State subjects, and the information economy of government ignores the Seventh Schedule entirely. A document may be classified by a Union ministry, held by a State official, communicated to a journalist in a third jurisdiction, and published from a fourth. The Act’s elements are indifferent to these crossings, as a general criminal statute should be, but its sanction filter is not, and the identification of the appropriate Government in a cross-cutting case is a real preliminary question on which prosecutions can stall or founder. The drafting choice to distribute the filter was made in a unitary colonial state and adapted by amendment to a federal republic, and like so much in this Act it functions adequately in practice while resting on a design its makers did not choose for the polity that uses it.

Declassification, archives and the life of a secret after its usefulness

Every secrecy system contains an implicit theory of time, and the Indian system’s theory is administrative from end to end. A document is classified at creation by the department that originates it. Its grade is reviewed within the department’s own practices. Records of permanent value pass, under the public records framework administered through the National Archives, toward archival custody as they age, and declassification decisions are taken by the originating department in that passage. The criminal statute supervises none of this. The Official Secrets Act contains no period after which protected material ceases to be protected, no duty to review a classification, no presumption in favour of release for material of a given age, and no offence of over-classification to balance its offences of disclosure. The silence is symmetrical with the Act’s silence on definition, and it has the same explanation: in the design of 1923, the life cycle of a secret belonged wholly to the administration that owned it, and no subsequent amendment disturbed that allocation.

The consequences are visible in the recurring public controversies of the field, which are characteristically controversies about old paper: files of historical importance whose release is sought by scholars, families and legislatures, and whose retention is defended by reference to sensitivity that the passage of decades has plainly diminished. The transparency statute supplies the citizen’s only statutory lever in these contests, and its exemptions and public interest balance are doctrines built for present governance more than for history. The committee of secretaries that reported in 2017 took declassification within its remit precisely because the mismatch was recognised inside government: a records regime that moves paper toward the archive on one timetable and a secrecy statute that recognises no timetable at all will generate friction at every point where history is politically alive. Other democracies have answered the friction with statutory clocks of varying design, fixed periods after which classification lapses or must be affirmatively renewed, with exceptions for the narrowest categories. The merits of any particular clock are debatable, and the security case against automatic lapse for genuinely enduring secrets, agent identities, cryptographic material, is a strong one. The Indian settlement’s peculiarity is not that it rejected a clock after debate. It is that the criminal statute has never contained the question, and the debate, when it happens, happens in committee rooms and archival policy rather than in Parliament on the text of the Act. A successor statute that decided nothing else but wrote a review duty and a renewal presumption into the law of classification would change the practical operation of official secrecy more than many a louder reform, because it would move the system’s theory of time from administrative habit into law, which is where this pillar has argued, from its first section, the operative content of this statute ought to reside.

The Official Secrets Act in this series: the definition that sits outside

This series has repeatedly found the operative content of Indian statutes located outside their enacted text: in rules that carry the real scheme, in schedules that hold the figures that matter, in commencement notifications that decide which of two enacted regimes is alive. The Official Secrets Act is the pattern’s most complete instance, because what sits outside this statute is not an operative detail but the definition of the statute’s own subject matter. The Act punishes the mishandling of secrets and nowhere says what a secret is. The classification manual says it, administratively. The transparency statute qualifies it, by override. The courts confine it, by insisting on elements. The commissions indicted it, by recommending a successor that would say it in law. Each of those locations is a real location of the law in operation, and none of them is the statute.

A reader who has followed the pillar to this point holds the One Test answer the brief required. The statute does not define the thing it protects. Its interaction with the transparency statute is override for access, survival for offences. Its repeal has been recommended by an official body and not effected. Around those three propositions the rest is commentary, grading, purpose, custody, sanction and debate, and this article has supplied the commentary at the length the statute’s importance warrants. The Official Secrets Act 1923 remains what it was enacted to be, a general law against espionage and unauthorised disclosure, and it remains what its examiners found it to be, a law whose centre is missing and whose replacement has been designed on paper more than once. Both descriptions are true, and the statute book holds both, undisturbed, into the transparency age.

One further discipline belongs to the reader who uses this pillar as a working reference, and it concerns the difference between the Act as invoked and the Act as written. Official statements reach for the statute’s name as a general emblem of secrecy, in contexts ranging from the refusal of a record to the warning of a witness, and the emblem is broader than the law. The law is narrower in its elements, more specific in its grading, and more dependent on its sanction filter than the emblem suggests, and it is wider in one respect the emblem understates, since its communication offence requires no proof of harm at all. A citizen, an official or a publisher who needs to know where the boundary runs should therefore go to the text, to Sections 3, 5 and 13 in the first instance, and to the transparency statute’s Sections 8 and 22 for the access side of the line, and should treat every invocation of the Act’s name, from whatever quarter, as a claim to be checked against those provisions rather than as a statement of them. That habit of checking is the practical form of the rule of law in a field where the operative definition sits outside the statute, and it is the habit this article has attempted to equip. Where a specific document, a specific communication or a specific prosecution is in issue, the checking must be done with legal advice on the specific facts, since classification, custody, authority and purpose are all fact-sensitive, and general exposition, however careful, states the law and not the outcome of any case.

Frequently Asked Questions

Q: What does the Official Secrets Act 1923 penalise?

The Act penalises two families of conduct. Section 3 punishes spying: acting for a purpose prejudicial to the safety or interests of the State by approaching or entering a prohibited place, making a sketch, plan, model or note useful to an enemy, or obtaining, collecting, recording, publishing or communicating secret official codes or protected documents and information. Section 5 punishes the wrongful communication of official information held through office, confidence or contract, and the knowing receipt of such information. Supporting sections punish unauthorised use of uniforms and documents, interference with officers near prohibited places, attempts, and harbouring offenders.

Q: What counts as a prohibited place?

Section 2(8) defines a prohibited place to include works of defence, arsenals, naval, military and air force establishments and stations, mines and minefields, and factories, dockyards, camps, ships and aircraft belonging to or occupied by the Government where defence material is made, repaired or stored, including places where such material is temporarily kept. The definition also covers any place belonging to or used for the purposes of the Government, and railways, roads, canals and other means of communication, that the appropriate Government declares by notification to be a prohibited place on the ground that information about it or damage to it would be useful to an enemy.

Q: Does the Act define an official secret?

No. The Act nowhere defines a secret or an official secret, and it contains no classification scheme and no statutory criterion by which a document becomes protected. Its offences describe categories of material, secret official codes and passwords, documents relating to prohibited places, material entrusted in confidence or obtained through office, and leave the identification of what is secret to administrative practice. In operation, government documents are graded and marked under departmental security instructions, and that executive classification performs the defining function the statute omits. The absence of a definition is the statute’s most remarked structural feature and stands at the centre of the case for its replacement.

Q: How does the Act interact with the Right to Information Act?

The transparency statute overrides the secrecy statute within the field of access and leaves its offences intact. Section 22 of the Right to Information Act gives that Act effect notwithstanding anything inconsistent in the Official Secrets Act 1923. Section 8(1)(a) exempts information whose disclosure would prejudicially affect sovereignty and integrity, security and related interests, and Section 8(2) permits a public authority to allow access where the public interest in disclosure outweighs the harm to the protected interests, notwithstanding the Official Secrets Act. Unauthorised taking, communication and receipt of protected material outside that framework remain governed by the offences of 1923, which stand unrepealed.

Q: Can a journalist be prosecuted under the Official Secrets Act?

A journalist is in the same position as any other person under the Act’s general provisions, since no provision names the press. Section 3 includes publishing among its verbs, but requires the prosecution to prove a purpose prejudicial to the safety or interests of the State. Section 5(2) reaches a person who voluntarily receives protected material knowing, or having reasonable ground to believe, that it was communicated in contravention of the Act. Any prosecution also requires a complaint by order of or under authority from the appropriate Government under Section 13. Reach, elements and the sanction filter together define the true position more exactly than either alarm or reassurance.

Q: What did the Supreme Court say about published official documents?

Hearing a preliminary objection in review proceedings, a three-judge Bench held in April 2019 that neither the evidence statute’s privilege provision for unpublished official records nor the Official Secrets Act barred the Court from considering documents already published by a newspaper and standing in the public domain. The Court found no provision in the Act empowering the executive to restrain publication of documents marked secret, treated the newspaper’s publication as consonant with the constitutional guarantee of freedom of speech, and dismissed the objection. A concurring opinion added that the transparency statute had diluted governmental claims of privilege. The decision governs consideration of published material and decides no question of liability for obtaining it.

Q: Has repeal of the Official Secrets Act been recommended?

Yes. The Second Administrative Reforms Commission, in its First Report on the Right to Information submitted in June 2006, recommended that the Act be repealed and substituted by a chapter in the National Security Act containing provisions relating to official secrets, with the successor to Section 5 drawn on the narrower lines the Shourie Committee had proposed in 1997, confining the offence by reference to national security. The Government did not accept the recommendation, recording objections to the proposed destination and machinery. A committee of secretaries later examined the Act against the transparency statute and reported in 2017 with harmonisation recommendations, which were likewise not enacted.

Q: What sanction is required to prosecute?

Section 13 provides that no court shall take cognizance of an offence under the Act except upon a complaint made by order of, or under authority from, the appropriate Government. The condition is jurisdictional and stands at the threshold of every prosecution. A private complainant cannot set a case in motion, and an investigating agency’s conclusions have no legal route to a court unless the Central or State Government, as appropriate, directs or authorises the complaint. The requirement is the statute’s principal internal safeguard, converting each prosecution into a decision for which a government is politically answerable, though the filter is executive rather than judicial in character.

Q: What is the maximum punishment under Section 3 of the Act?

Section 3 grades its punishment by subject matter. Where the spying offence is committed in relation to a work of defence, an arsenal, a naval, military or air force establishment or station, a mine, minefield, factory, dockyard, camp, ship or aircraft, the naval, military or air force affairs of the Government, or a secret official code, imprisonment may extend to fourteen years. In all other cases the maximum is three years. The fourteen-year ceiling, raised by amendment in 1967 from the lower figure of the original enactment, is therefore reserved for espionage directed at the defence establishment and official codes, and the wider penumbra of Section 3 conduct carries the three-year grade shared with Sections 5 and 6.

Q: What must the prosecution prove about purpose under Section 3?

The prosecution must prove that the accused acted for a purpose prejudicial to the safety or interests of the State. The purpose words stand at the head of Section 3(1) and qualify every act the section prohibits, approaching a prohibited place, making a sketch or note, obtaining or communicating protected material alike. Section 3(2) permits the purpose to be inferred from the circumstances of the case, the conduct of the accused, or the accused’s known character as proved, and relieves the prosecution of the need to show a particular act tending to demonstrate the purpose. The courts’ jurisprudence treats the purpose as an element to be established on evidence, not presumed from the act alone.

Q: What does Section 5 punish a recipient for doing?

Section 5(2) punishes a person who voluntarily receives a secret official code or password, or a sketch, plan, model, article, note, document or information of the classes Section 5 protects, knowing or having reasonable ground to believe, at the time of receipt, that the material is communicated in contravention of the Act. The offence lies in knowing receipt of material passed in breach, not in receipt alone. A recipient who takes material without that knowledge, and without reasonable ground for believing a breach lies behind it, commits no offence under the provision. Imprisonment may extend to three years, or fine, or both, the grade the section sets for wrongful communication offences generally.

Q: How are government documents classified if the Act does not classify them?

Classification is performed administratively under the Manual of Departmental Security Instructions issued by the Ministry of Home Affairs. The manual prescribes gradings of top secret, secret, confidential and restricted, assigned by reference to the damage that unauthorised disclosure would be expected to cause, with the highest grade reserved for material whose disclosure could cause exceptionally grave damage to national security. It governs marking, custody, transmission, review and destruction of classified records inside departments. The manual is an executive instrument, not legislation, and its operation means that the practical scope of the Act’s offences is fixed by administrative decision, a structural feature the reform commissions addressed when they examined the secrecy regime.

Q: Can proceedings under the Act be held in camera?

Yes, where the court so orders under Section 14. If it appears to the court that the publication of evidence given or statements made during proceedings under the Act would be prejudicial to the safety of the State, the court may order that the public generally, or any particular person, be excluded from the court for the duration of the proceedings. The order displaces the ordinary rule of public trial for the evidentiary stage, while the court’s judgment is delivered in public. The power answers a practical dilemma of secrecy litigation, that an open trial about protected material could itself disclose the material the proceedings exist to protect.

Q: Does the Act punish an attempt to commit an offence under it?

Yes. Section 9 provides that a person who attempts to commit an offence under the Act, or who incites, conspires with or abets another to commit such an offence, is punishable as though the offence attempted had been committed. The provision imports the general logic of inchoate liability into the statute’s own text rather than leaving it to the general penal law. Its practical significance is greatest for Section 3, where the spying offence already reaches preparatory conduct such as approaching a prohibited place or making a sketch, and Section 9 extends liability one step further back to the attempt to do those acts, provided in each case that the elements, including the prejudicial purpose, are established.

Q: What does Section 6 of the Act cover?

Section 6 covers the unauthorised use of uniforms, falsification of reports, forgery, personation and false documents. It punishes, among other conduct, wearing or using a naval, military, air force, police or other official uniform, or a uniform resembling one, without authority, making or using forged passes, permits, certificates or documents connected with prohibited places and protected material, and personating a person holding office under the Government. The offence supports the perimeter the principal offences draw: impersonation and forged credentials are the classic instruments by which protected places are entered and protected material obtained, and the section grades the conduct at imprisonment extending to three years, or fine, or both.

Q: What happens to a person who shelters an offender under the Act?

Section 10 punishes harbouring. A person who knowingly harbours a person known or reasonably believed to have committed, or to be about to commit, an offence under the Act, where the harbouring is undertaken in order to prevent that person’s apprehension, commits an offence punishable with imprisonment extending to three years, or fine, or both. The provision mirrors the harbouring offences of the general penal law and is directed at the support network around espionage rather than at espionage itself. Its elements are deliberately specific: knowledge or reasonable belief about the other person’s position, and a purpose of preventing apprehension, must both be established before the shelter a person gives becomes criminal under the section.

Q: What did the Shourie Committee propose for Section 5?

The Shourie Committee of 1997, a working group on the right to information and transparency chaired by H. D. Shourie, proposed a recast of Section 5(1) that would have confined the penal provision to violations affecting national security. Under its draft, the wrongful communication offence would have attached to official secrets obtained by virtue of holding or having held office with or under government, and kindred custody routes, where the violation bore the national security character the existing section omits. The Second Administrative Reforms Commission adopted the proposal in 2006, recommending that the successor provision in a national security chapter follow the Shourie lines. The proposal was not enacted, and Section 5 retains its enacted breadth, requiring wilfulness and absence of authority but no effect on national security.

Q: Does the Act contain a public interest defence?

No. The Act contains no provision under which a communication otherwise within Section 5 is excused because it was made in the public interest, and no provision under which disclosure of wrongdoing is a defence to a charge under Section 3 or Section 5. The public interest balance in Indian law on this subject is located in the transparency statute, whose Section 8(2) permits a public authority to allow access where the public interest in disclosure outweighs harm to protected interests, and in the separate whistle blower protection scheme. Within the four corners of the 1923 Act, motive is relevant to sentence and to the sanction decision, but it is not an element-negating defence to the offences as enacted.

Q: Why is the Act described as a colonial survival?

The Act was enacted in 1923 by the colonial legislature, modelled on British official secrets legislation of the same period, and replaced the Indian Official Secrets Act of 1889 as amended in 1904. Its architecture, a general prohibition on movement of official information, graded espionage offences, and no statutory definition of secrecy, reflects the assumptions of an imperial administration for which official information was administrative property. It has survived independence, the Constitution’s speech guarantee and the transparency statute of 2005 without re-enactment in republican form, its substance amended only at the margins, and its repeal has been recommended and not effected. The description records that lineage and that legislative history, and is used by critics and by official commissions alike.

Q: What is the position of material that has entered the public domain?

Publication does not amend the Act, and the offences are not expressed to lapse for material once published. What the courts have settled is the litigation position: in its April 2019 judgment the Supreme Court held that documents published by a newspaper and standing in the public domain could be considered by the Court notwithstanding the privilege provision for unpublished records and notwithstanding the Act, and that the Act confers no power to restrain publication. The holding governs the use of published material in adjudication and the absence of a suppression power. It does not decide that obtaining or communicating protected material is lawful because publication later occurred, and the elements of Sections 3 and 5 continue to govern that conduct.