Guest Post: Chaturvedi and the Remission Standard the Supreme Court Declined to Write

(This is a guest post by Kshitij Saruparia)

On 15 May 2026, in Rohit Chaturvedi v. State of Uttarakhand, a bench of Justices B.V. Nagarathna and Ujjal Bhuyan quashed the Ministry of Home Affairs’ refusal to concur in the premature release of a convict in the 2003 Madhumita Shukla murder case. The refusal was a three-line non-speaking order, and on that ground alone it was indefensible. The Court did not stop there. It held in paragraph 7 that the heinousness of an offence “stand[s] exhausted at the stage of sentencing” and cannot be revived to deny remission, and that to refuse remission on that ground “is to reconvert remission into a retrospective reaffirmation of guilt”.

The order quashing the MHA letter is plainly correct. The "exhausted at sentencing" formulation is not, and the judgment itself supplies the proof: the same judgment rests on two authorities that contradict it. The result will not be litigated. That sentence will be.

A Statutory Veto, and the First Test of how it may be Used

The case is decided within a frame the judgment sets out with care. The crime was investigated by the CBI, so by the Court’s own order of 31 January 2025 the State of Uttarakhand could exercise the remission power under Section 473(1) of the Bharatiya Nagarik Suraksha Sanhita only after the concurrence of the Central Government under Section 477(1). The textual change from Section 435 of the old Code appears slight, “consultation” has become “concurrence”, but the operational change is slighter still. The Constitution Bench in Union of India v. V Sriharan Murugan [(2016) 7 SCC 1] had already held, in 2015, that “consultation” under Section 435(1) meant “concurrence”, with the result that the State could not effect remission in CBI-investigated cases without the Centre’s agreement. That reading had operated as settled law for nearly a decade before the Sanhita was enacted. Section 477 does not create the Union veto; it codifies it, entrenching the stricter construction beyond judicial dispute and insulating it from future legislative attempts to restore the provision's literal meaning.

The Standing Committee on Home Affairs, which examined the Bill across twelve sittings in 2023 and submitted the 247th Report, made no observation on what became Section 477 in its clause-by-clause recommendations. The Home Secretary’s presentation of “important changes” to the Committee listed a Central Government concurrence requirement only in the context of sentences awarded by military courts, which is the subject of Section 476, a separate provision. The change at Section 477 converting a Constitution Bench interpretation of a federal power into statutory text drew no recorded scrutiny. Parliament codified Sriharan without Parliament deliberating on whether codifying it was the right choice.

For every centrally-investigated case, the Centre therefore holds a statutory veto over release that is now unambiguous on the face of the legislation and immune to the challenge that a future bench might read “consultation” more narrowly. Chaturvedi is the first judgment to test how that veto may lawfully be exercised, which is why its reasoning, and not only its outcome, matters beyond the petitioner.

The Court’s organising distinction is between sentencing and remission. Sentencing looks back and fixes punishment by the gravity of the act. Remission looks forward to conduct in custody, reform, and reintegration. Allowing the gravity of the offence to defeat remission lets the executive re-impose, through an administrative refusal, a punishment the sentencing court did not pronounce. That objection is sound. It is also not new. The Court itself cites Satish v. State of UP (2021), which held that neither the length of the sentence nor the gravity of the offence can, by themselves, be the sole basis for refusing premature release, and that the risk of reoffending must be assessed from the prisoner’s antecedents and conduct in custody.

The Contradiction is Inside the Judgment

Satish is the careful statement of the rule. Gravity is not the sole basis; it is one input, and it feeds the forward-looking question of recurrence. Chaturvedi adopts something larger. To say gravity is “exhausted” at sentencing is to say it has no residue at the remission stage at all.

That cannot be squared with the test the same judgment adopts a few paragraphs earlier. At paragraph 5.3 the Court sets out the Laxman Naskar factors as the governing criteria for premature release noting that those factors were also reproduced in Rohit Chaturvedi in paragraph 5.3.. The first factor is whether the offence is an individual act of crime without affecting society at large. That is, unavoidably, an inquiry into the nature and gravity of the offence, addressed to the remission authority.

So, the judgment instructs the executive to weigh the character of the offence and, in the same document, declares that character spent. Read literally, the two directions cannot both be operative: an authority cannot apply a factor that turns on the gravity of the offence while simultaneously treating that gravity as having no residue. A remission authority handed both has been given a contradiction, not a standard.

The way to read Chaturvedi without breaking it is to treat “exhausted” as rhetoric and the Satish rule as the holding. Heinousness cannot be a freestanding and conclusive ground of refusal, because used that way it does convert remission into re-sentencing. But the nature of the offence does not disappear. It re-enters as one input into the prognostic question the Court itself frames, the risk of recurrence and the safety of the public, exactly as Satish and Laxman Naskar contemplate.

The Court’s instinct is defensible, and worth stating at its strongest. If offence gravity is freely available at the remission stage, executive review collapses into a standing reconsideration of desert, and the sentence imposed by the court is no longer treated as settled, but is repeatedly revisited by the Ministry through successive remission decisions. Holding the line between punishment and post-conviction reform is a serious constitutional purpose, and it is plainly what the bench was reaching for. The difficulty is that the bench already had the instrument for the task. Satish keeps gravity out as a sole or dominant ground while letting it inform the prognosis, and that protects the distinction without pretending the offence has vanished. On that reading the case decides nothing the law did not already hold. “Exhausted” does no more work than Satish; it does the same work less accurately, and leaves behind a sentence the next litigant will weaponise.

The Standard the Court did not Write

The judgment’s nearest approach to a positive standard is paragraph 7.1, where the decision on remission “must emerge from a holistic assessment of the prisoner and after balancing societal interests with the prisoner’s right to be considered for release on fair and reasonable criteria”. This is not a standard. It does not tell the Ministry what a lawful refusal must contain, what evidentiary threshold attaches to a finding of dangerousness, or, most importantly, through which doorway the seriousness of the offence may legitimately enter the “societal interests” side of the very balance the Court prescribes. Having narrowed the admissible grounds on one side, the Court left the content of a permissible refusal undefined on the other.

It then demonstrated the gap instead of closing it. Rather than remit to the Ministry with directions on how to decide, the Court declined to remand at all, called a remand “an empty formality,” and performed the remission assessment itself, on parity with the released co-accused, time served, conduct in custody, and the State’s recommendation. The executive’s function was not corrected. It was substituted. That may be defensible on these facts, where the petitioner had served twenty-two years and a co-accused convicted in the same case was already free. It is no guide at all to the Ministry in the next case, where the facts are worse and the offence graver.

What is settled after Chaturvedi is what was already settled after Satish: gravity alone will not sustain a refusal. What is open, and now sharper, is the question the Centre’s Section 477 veto makes unavoidable. What must a reasoned, non-arbitrary concurrence refusal actually say, on what material, and through which doorway, if any, may the seriousness of the offence return as part of “societal interests”? The “exhausted at sentencing” line will sit at the centre of that fight, pressed by convicts to shut the offence out entirely and distinguished by the Union as loose language unsupported by the judgment’s own authorities. The Court had the materials to write the standard. It offered a philosophy of punishment where a workable standard was needed, and left that standard for another bench and another prisoner. The door it shut needed shutting. The one it left open is the one that decides cases.

Guest Post: Renumbering as Repeal? The Allahabad High Court on the interplay of BNSS with the SC/ST Act

(This is a guest post by Ammar Shahid)

Two orders from the Allahabad High Court raise a peculiar question: does the bar on anticipatory bail under Section 18 of the Scheduled Castes and Scheduled Tribes (Prevention of Atrocities) Act, 1989 (“SC/ST Act“) continue to operate now that the Bharatiya Nagarik Suraksha Sanhita, 2023 (“BNSS“) has replaced the Code of Criminal Procedure, 1973 (“CrPC“)?

In Dinesh Kumar Srivastava v. State of UP (December, 2025), the applicant faced an FIR under Sections 376, 506, 406, 323, and 504 IPC along with Section 3(2)(v) of the SC/ST Act, arising from allegations of a relationship formed on the matrimonial website jeevansathi.com allegedly on a false promise of marriage. He sought anticipatory bail under Section 482 of the BNSS. The informant’s counsel opposed this on the ground that Section 18 of the SC/ST Act barred it. The Court disagreed, holding that firstly, Section 18 specifically refers to “Section 438 of the Code” and anticipatory bail applications are now filed under Section 482 BNSS, the bar had no application. Secondly, the Court held that prima facie the material on record did not establish an offence under Section 376 IPC, given that the relationship appeared consensual. The petition was allowed, and bail granted.

Next, Sushil Kumar v. State of UP (May, 2026) arose from rejection of an anticipatory bail application solely on the ground of Section 18. The alleged facts involved a commercial dispute over unpaid sale consideration for food grains, accompanied by an alleged assault and caste-based insult. The High Court set aside the trial Court’s order on two grounds. The first was the trial Court’s failure to consider the principle laid down in Prathvi Raj Chauhan v. Union of India, where the Supreme Court held that courts retain the power to grant pre-arrest bail in SC/ST Act cases where prima facie the ingredients of the Act are not made out. The second was the BNSS transition argument drawn from Dinesh Kumar Srivastava. The High Court held that these two factors had not been considered by the trial court, and sent the matter case for fresh consideration.

Section 8 of the General Clauses Act

Section 8 of the General Clauses Act, 1897 provides that where a Central Act is repealed and re-enacted with or without modification, references in any other Central Act to the repealed enactment shall, unless a different intention appears, be construed as references to the corresponding provisions of the re-enacted legislation.

Section 482 BNSS corresponds with Section 438 CrPC. Both carry the heading of anticipatory bail and provide the same relief to the same class of persons in the same procedural context. When Parliament enacted the BNSS, it was re-enacting an existing one under a different statutory number. By operation of Section 8 of the General Clauses Act, Section 18’s reference to Section 438 CrPC must now be read as a reference to Section 482 BNSS.

Neither Dinesh Kumar Srivastava nor Sushil Kumar consider Section 8, which is a serious flaw. The legal question, at its foundation, was one of statutory interpretation across a legislative transition. Section 8 of the General Clauses Act is the rule Parliament has itself put in place for resolving exactly this question. Its absence from the reasoning of both decisions is a significant gap.

The only caveat Section 8 carries is the phrase “unless a different intention appears.” One would need to find, within the SC/ST Act or Section 18 itself, some indication that Parliament intended the bar to be tied specifically and permanently to Section 438 CrPC as a provision, rather than to the institution of anticipatory bail as such. No such intention is apparent. Section 18 was aimed at the relief of anticipatory bail, not at a particular section number. The number was simply the address at which that relief resided at the time.

What the 2018 Amendment Tells Us

The legislative history of Section 18 adds important context. In Subhash Kashinath Mahajan v. State of Maharashtra, the Supreme Court permitted anticipatory bail in SC/ST Act cases subject to certain conditions. Parliament responded within months by enacting the SC/ST Amendment Act, 2018, inserting Section 18-A which, among other things, reinforced the bar on anticipatory bail. Parliament actively closed a space that the Supreme Court had opened (which it has, since, reopened - more on that below).

The suggestion that this same Parliament, in 2023, inadvertently allowed the Section 18 bar to lapse through the BNSS transition without any express amendment is difficult to accept. When a legislature has demonstrated such attentiveness to a particular provision, the more reasonable inference is that it relied on established principles of statutory interpretation, specifically Section 8 of the General Clauses Act, to ensure continuity. If Parliament had wanted to lift the bar, it could have done so expressly in the BNSS. It did not.

The Prathvi Raj Chauhan Exception

Both decisions also place reliance on the principle formulated by Prathvi Raj Chauhan, where the Supreme Court held that in cases where no prima facie material exists to make out an SC/ST Act offence, Courts may grant pre-arrest bail through inherent jurisdiction under Section 482 CrPC. This reading of Prathvi Raj Chauhan, however, is broader than what the judgment actually supports.

The inherent jurisdiction of the High Court in criminal matters vested in Section 482 CrPC, which corresponds to Section 528 of the BNSS, not Section 482 BNSS, which is the anticipatory bail provision. Section 18’s bar has always applied to the anticipatory bail route, not to the exercise of inherent jurisdiction in exceptional cases. These are two separate heads of power, and conflating them misrepresents both.

Beyond this structural issue, the Prathvi Raj Chauhan exception is not a straightforward gateway to routine merits review. The Supreme Court was clear that this inherent power is available “sparingly” in “very exceptional cases” to prevent a “miscarriage of justice or abuse of process of law,” and was explicit that “a liberal use of the power to grant pre-arrest bail would defeat the intention of Parliament.” The Court in Sushil Kumar, having identified Prathvi Raj Chauhan as a primary ground for remand, does not specify how the trial court is to apply this demanding threshold on reconsideration. Without any guidance, the remand functions as an implicit invitation to conduct the kind of routine merits review that Prathvi Raj Chauhan cautioned against.

This issue is also visible in Dinesh Kumar Srivastava. The Court's own prima facie analysis found no offence under Section 376 IPC on the basis that the relationship appeared consensual and no force was alleged. But the Prathvi Raj Chauhan threshold is directed at the absence of SC/ST Act ingredients from the face of the FIR, not at the tenability of a connected IPC charge. Where an FIR contains specific allegations of caste-based insult or harassment, as it does in both the decisions under discussion, it is not obvious that the threshold for invoking the inherent jurisdiction exception has been crossed. Hitesh Verma v. State of Uttarakhand, which Sushil Kumar itself cites, reaffirms that the bar under Section 18 is absolute where SC/ST Act offences are prima facie made out. Citing Hitesh Verma while simultaneously remitting for fresh consideration of anticipatory bail, without a definitive holding on whether SC/ST Act ingredients are absent from the FIR, creates a real tension that the judgment leaves unresolved.

Conclusion

The question raised by these two decisions is ultimately not a complicated one. Section 8 of the General Clauses Act provides a clear and direct answer: the bar under Section 18 travels with the legislative transition from CrPC to BNSS because Section 482 BNSS is the corresponding provision to Section 438 CrPC. The legislative history of Section 18, including the 2018 amendment, makes it implausible that Parliament intended this bar to lapse silently. The Prathvi Raj Chauhan exception operates through inherent jurisdiction in exceptional cases, not through anticipatory bail as a matter of routine.

The SC/ST Act is protective legislation enacted for a specific purpose. Its provisions are intended to give effect to that purpose, not to be read in ways that drain them of content through procedural technicality. Courts interpreting special protective statutes are expected to favour readings that advance the statute’s objects. A reading that allows the bail bar to lapse because a provision number changed does the opposite.

[Note: at the time of writing, it is not clear whether either decision has been challenged in the Supreme Court, or has been questioned by a bench of larger strength in the High Court itself]

Guest Post: Reconsidering the Expulsion Requirement in The Offence of Causing Miscarriage

 (This is a guest post by Ananya Kumar)

Under the Indian Penal Code 1860 (IPC), causing the death of a foetus was a criminal offence, even if done with the consent of the mother. The adoption of the Medical Termination of Pregnancy Act (MTPA) in 1971 created an exception to this. It allowed registered medical practitioners to terminate pregnancies with the mother’s consent, i.e., to carry out abortions. The IPC continued to penalise acts done with criminal intent to harm the foetus, or the mother, or both, i.e., the offence of causing miscarriage. Now, Sections 88 to 92 of the Bharatiya Nyaya Sanhita 2023 (BNS) prescribe punishment for the offence of causing a woman to miscarry without her consent in its varying forms in the same manner as the IPC.

Despite this statutory framework, a fundamental issue remains unresolved. Neither the MTPA, nor the BNS define ‘miscarriage’, and how it differs from ‘abortion’. Consequently, courts assume that both abortion and miscarriage entail the expulsion of a foetus from the womb before its term with the intent of causing its destruction, the only difference being that the former is done with the mother’s consent (see para 4).

However, miscarriage and abortion have significant differences and implications, beyond mere consent. For instance, miscarriages happen spontaneously and can sometimes lead to the death of a foetus inside the womb without it getting expelled. Yet, as things stand, courts have interpreted ‘miscarriage’ under the IPC to necessarily require expulsion. Consequently, courts have incorrectly dismissed charges of the offence of causing a foetus’ death on the grounds that it was not expelled from the womb. 

In this blog, I highlight how the law has wrongly assumed that expulsion of the foetus is a necessary element for the offence of causing miscarriage. I then point out cases where this has led to the acquittal of probable offenders for causing miscarriage. Accordingly, I suggest two interpretations through which courts can correct their approach towards interpreting the offences of causing miscarriage, in the absence of legislative action.

The Difference between Miscarriage and Abortion

In medical sciences, miscarriage is also referred to as ‘spontaneous abortion’. By contrast, what we ordinarily call abortion is called ‘induced abortion’. It is perhaps this terminology that has led courts to conclude that principally, the only difference between miscarriage and abortion is the non-consensual and spontaneous nature of miscarriage. 

For instance, Modi’s Medical Jurisprudence And Toxicology (10th edn, 1949) explains that when a woman consents to the expulsion of the foetus from her body, it is abortion. On the other hand, when a foetus is made to be expelled from a woman’s womb without her consent, it is termed as miscarriage (see page 319). In light of this, even when Parliament has defined miscarriage, in social security legislations, it has described it as the ‘expulsion of the contents of a … uterus’ (see, Maternity Benefit Act 1961 and the Code on Social Security 2020). This definition replicates the same error courts have been interpreting under criminal law.

The Supreme Court in Dr. Jacob George v State of Kerala (1994) relied on this explanation to state that the distinction between abortion and miscarriage is ‘not material’ for the purposes of the IPC so long as there is an intent to cause miscarriage (para 5). However, the difference between spontaneous and induced abortion cannot be reduced to the presence or absence of consent alone. 

There are many circumstances and complications that lead to different kinds of spontaneous abortions that do not have similar characteristics to abortions. Sometimes, complications in miscarriages, can lead to death of the foetus in-utero, without the expulsion of the products of conception. This type of miscarriage is called an ‘incomplete’ or a ‘missed’ miscarriage. This is why, equating miscarriages to ‘abortions with consent’ can miss the complicated reality of miscarriages. 

Another distinction between the two is the intent behind the act. Abortions are done with the deliberate intent of terminating and removing the foetus from the womb. However, it is not necessary that an offence is committed against a woman with the specific intent of bringing about the expulsion of the foetus. The intention may simply be to cause harm or death to the mother, or the foetus. In some cases, the child may die in-utero due to extraneous circumstances like shock, stage of foetal development, or death of the mother. Therefore, courts should recognise that miscarriages can also take place without the expulsion of the foetus.

Case Dismissals Due to the Expulsion Requirement

Courts have previously dealt with cases of incomplete or missed miscarriages under Sections 312 and 314 of the IPC. The interpretation by the courts in these cases shows flawed reasoning that has led to the acquittal of probable offenders, only due to non-expulsion of foetus from the mother’s body. 

For example, in Prasoon Kumar Jaiman (2017) before the High Court of Rajasthan, the petitioner was accused of causing miscarriage (Sec 313) where there was no expulsion of foetus from the mother’s womb. However, the Court held that no offence was made out due to a ‘lack of medical evidence’ proving miscarriage. The ‘lack’ here referred to the fact that the foetus had not been expelled from the mother’s body, which had been termed by the doctor as a ‘missed abortion’ (page 3).

Along similar lines, the judgment of the High Court of Chhattisgarh in Rajesh Kumar (2024) was another case where the expulsion requirement was misused by the accused. There the accused had allegedly administered a poisonous liquid to the mother with the intent to kill the foetus conceived outside marriage. The mother died soon after. In the autopsy, it was found that the foetus had died inside the womb before the mother, without getting expelled. The Court observed that causing miscarriage necessarily requires the expulsion of foetus (para 14). Since no such expulsion had occurred, the accused were acquitted on the charges under Section 312 and 314 of the IPC (Sections 88 and 90 of BNS). These cases demonstrate how the absence of a definition of miscarriage, as well as the judicial insertion of the expulsion requirement, allows perpetrators to evade culpability.

Avoiding Miscarriage of Justice

The most straightforward resolution to this issue would be for the legislature to introduce a statutory definition of ‘miscarriage’ in the BNS, one that does not make expulsion of the foetus a necessary requirement. A definition that instead focuses on the death of the foetus before term, whether or not followed by expulsion, would bring penal law in line with medical reality.

Alternatively, so long as there is legislative silence on the definition of miscarriage under the BNS, courts can prevent its abuse through two approaches. First, by interpreting the offence in a manner consistent with its medical meaning. Alternatively, by proceeding against perpetrators under other offences.

Firstly, as established above, miscarriages differ from induced abortions for more reasons than just consent. Keeping this in mind, interpretation of ‘miscarriage’ for the specific purpose of these crimes must be in line with the medical realities of their outcomes. Thus, the focus of the definition of miscarriage and abortion should not be on the expulsion of the foetus. Both aim to destroy the foetus inside the womb of the mother before the completion of the gestation period. Thus, regardless of whether there has been an expulsion of the foetus or not is not relevant for establishing the crime of causing miscarriage. The death of the foetus before term should be enough to meet that threshold.

Secondly, pending an authoritative reinterpretation of miscarriage by a larger bench, two alternative charges may be available to prosecutors in cases where there has been no expulsion. The foremost is Section 62 of BNS, which criminalises attempt to commit a crime. Thus, even if an offence of causing miscarriage is not made out on grounds of non-expulsion, offenders may still be tried for attempting to cause miscarriage with criminal intent if evidence is found.

The other alternative is Section 91, which punishes the act of ‘preventing a child from being born alive’ amongst others. Unlike the miscarriage provisions, Section 91 makes no reference to miscarriage or expulsion. While no court has applied Section 91 to situations of missed miscarriages yet, a textual interpretation of this Section would be wide enough to cover cases of intrauterine foetal death. 

Conclusion

By equating miscarriage with abortion for legal purposes, legislators and courts have overlooked the medical realities of miscarriage and in doing so, unreasonably narrowed the scope of criminal liability. This approach has serious consequences, as it allows individuals who have caused the death of a foetus with mala fide intent to evade culpability simply because the foetus was not expelled from the womb. Prior precedents demonstrate how absence of a statutory definition of miscarriage has become a loophole that undermines the objective of the law.

An interpretation of ‘miscarriage’ that focuses on the death of the foetus in-utero rather than its expulsion from the womb would resolve this inconsistency. Pending such a reinterpretation by a larger bench, the alternative routes available under Sections 62 and 91 of the BNS offer prosecutors a meaningful, even if not entirely ideal, path forward. Change in this area will require courts to look to the harm the provisions were designed to prevent.

Update - IBC Moratoriums, Personal Insolvency, and Cheque Bounce Cases

About a year ago, the blog covered a decision of the Supreme Court [Rakesh Bhanot v. Gurdas Agro, 2025 INSC 445] dealing with the issue of interim moratoriums under the Insolvency and Bankruptcy Code, 2016 [IBC] and prosecutions for offences under the Negotiable Instruments Act, 1881 [NI Act]. It took the view that the decision in Rakesh Bhanot did not sufficiently explain its conclusions, especially in light of existing precedent in the field. At the same time, the post also noted that the concerns animating the decision were, arguably, originating in poor legislative design of the IBC.

Over the past year, it appears that the correctness of the view taken in Rakesh Bhanot has come under a cloud. The most recent doubts were raised when in May 2025 a bench of Three Justices — headed by the Chief Justice — issued notice and granted interim reliefs in a petition which has, at least going by the news article, directly challenged the correctness of Rakesh Bhanot [Karan Bhatia, W.P. (Crl.) 131/2026]. 

There are tons of such fact scenarios playing out causing much grief to (some) litigants. A Three Justices Bench granting an interim relief in the face of the existing judgment, is a decent sign of which way the wind may blow. Additionally, it would be nice if, this time, the issue can also get the clarity that it requires. As the previous post discussed, there are some thorny law and policy issues involved here. It requires careful reading of the Indian Contract Act, the IBC, as well as precedent (such as P. Mohanraj). 

One can only hope that the issue thus gets finally resolved soon, one way or another. That is eminently more desirable than a Vijay Madanlal kind of scenario where the Court is kicking the proverbil can down the road which clogs the docket, as more and more persons move for interim reliefs, and also continues the confusion in courts below.

  

Course Correction on Delays and Bail under the UAPA?

[This post was updated at 6PM on 19.05.2026 to add reference to this news article and on 23.05.2026 to add reference to the SC Order dated 22.05.2026]

On 18.05.2026, a Division Bench of the Supreme Court handed down its verdict in Syed Iftikhar Andrabi [2026 INSC 503]. The bail application was allowed, which if we see the facts is not entirely surprising: the accused had been in jail for more than five years, there were many co-accused already on bail, some of them quite similarly placed, the incriminating material was largely confessional, and the trial was nowhere near completion. 

It is not because of the factual analysis that the judgment runs into 102 pages, which is quite unusual for a bail petition. The length was necessitated as the Bench felt it had to clear the legal position on a key issue affecting bail adjudication in the UAPA context, and the broader context of special crimes more generally. This, of course, was the issue of what impact does lengthy undertrial incarceration have in bail petitions under the UAPA with its additional restrictions for grant of bail under Section 43-D.

According to the Bench, a position had been set out by the Bench of Three Justices in K.A. Najeeb [(2021) 3 SCC 713], holding that the rigours of merits-based statutory restrictions for granting bail would, in the case of grave delays, be trumped by the concerns for an accused's liberty. The value of liberty and to a speedy trial was seen as grounded in Article 21 of the Constitution; lengthy undertrial detention would, in case of grave delay, be seen as causing detention in a manner contrary to these fundamental rights.

This position seemed to have been doubted and departed from in two subsequent judgments delivered by Division Benches; the first was Gurwinder Singh [(2024) 5 SCC 403], and the second Gulfisha Fatima [2026 SCC OnLine SC 10] (both were opinions of the same author). These subsequent decisions had, seemingly, held that bail was not the rule in UAPA cases owing to the restriction placed by Section 43-D UAPA which evidenced a legislative preference for jail as the default, and no relaxation or mollification of this statutory restriction was caused by delays or by the lengthy undertrial detention of an accused. 

The validity of this line of reasoning from both these opinions has been expressly doubted in Syed Iftikhar Andrabi [Paras 26 to 41]. The simplest line of critique is that the Division Benches could not have doubted or disagreed with the judgment of Three Justices in Najeeb. But the Court does not rest there, perhaps due to the manner in which both Gurwinder Singh and Gulfisha Fatima made departures. Neither said that the law laid down in Najeeb was bad; instead, the fight was about what law, if any, did Najeeb lay down in the first place. 

Gurwinder Singh and Gulfisha Fatima departed from the opinion in Najeeb by reading the latter as tightly wound to its facts. In that case there was no prospect of speedy trial whereas in Gurwinder Singh notable progress had already been made, and in Gulfisha Fatima the delays were 'natural' owing to the complexity of the case. In other words, both opinions tried to denude the opinion in Najeeb of having laid down any law at all. Having done so, both opinions offered their reading of the law, which saw the statutory bar of Section 43-D UAPA remaining unaffected by any length of delay. Even if there was delay, the detention of an accused may yet be justified in the overall facts of a case.

Syed Iftikhar Andrabi has specifically disapproved of this interpretation of Najeeb, as well as the holding of both Gurwinder Singh and Gulfisha Fatima that the statutory bar is adamantine and remains unaffected by delays. It reminds us of the very clear observation in Najeeb, that in appropriate cases the statutory bar would 'melt' in the face of the pressure created by the prolonged and unconstitutional deprivation of an accused's liberty. The statutory bar is, and remains, subservient to constitutional interests.

The verdict is a helpful course correction by the Supreme Court in the domain of bail on grounds of delay which has seen a renaissance of sorts in the past five years across multiple statutory contexts containing a clause restricting grant of bail. Whether the course correction will hold or not is something to be seen. All it takes is another Bench of Three Justices to come and doubt Najeeb, after all. And it took only a day for the State to raise the possibility. Think about this for a moment. It is not a right to release on bail, but the mere possibility for considering bail pleas slightly more favourably for those who have spent inordinately long periods in custody pending trial, is what the fight is about. And even this possibility is precarious, dependent upon how some words in a statute are read. For now though, the possibility shines brighter than it did six months ago.

At the same time, let us not overstate the usefulness of this jurisprudence of apology created by our courts. Where, after depriving individuals of their liberty for years, at some unspecified point in the future their custodial detention mysteriously burnishes their claims to liberty with greater light. In fact, Syed Iftikhar Andrabi has ironically reinforced the sheer unfairness of it all quite drastically. At paragraph 42, the Court cites data suggesting extremely low conviction rates for UAPA crimes, and rhetorically asks how can the prolonged detention be justified merely because of serious allegations if there are such high possibilities of a future acquittal. If this reasoning is good for prolonged detention and delay on grounds of bail, it is also good for demanding greater accountability for deciding bail applications on merits and seriously assailing the constitutionality of the restrictive bail clauses under the UAPA and other laws.

The allegations against the accused were serious in Syed Iftikhar Andrabi, but they were based on hardly any credible material. So why did the courts below continue to shy away from granting bail on merits for the five years that it took for this judgment to pass? That, really, is the question which deserves a lot more scrutiny going forward in bail jurisprudence under the UAPA. 

The vigour with which the Court has sought to resolve the creases within the legal fabric on delay-based bail for UAPA and other such crimes requires urgent reciprocation in the realm of merits-based bail. Only then will the proclamation of the Court in Syed Iftikhar Andrabi carry any real force: that bail remains the rule, and jail the exception, in all circumstances in our law.

UPDATE: It did not take long for the issue to get referred to a bench of larger strength. On 22.05.2026, while granting interim bail to the applicants, a bench of the Supreme Court took note of the observations in Syed Iftikhar Andrabi and requested the Chief Justice to constitute a bench of appropriate strength to "clarify or expound the position of law laid down in K.A. Najeeb's case, particularly in the backdrop of the rigour of 43D(5) which imposes restriction consciously and has received the assent of the Parliament, which obviously was brought in keeping in mind the valuable right enshrined in Article 21 of the Constitution of India." (Para 24). All eyes on the Court, then.     

[On reading of restrictive bail clauses, readers may find the six-part series 'Jail is Not the Rule' helpful]

Snippet - The Prison Statistics India 2024 Report

This past week, the Prison Statistics India Report for 2024 ['PSI Report'] was published by the National Crime Records Bureau ['NCRB]. I could not spot any massively interesting data points as such, especially if we read this Report together with the 2023 PSI Report. Indian prisons continue to be filled overcapacity, mostly with undertrial prisoners, who are usually young and poorly educated. While we do know that the prisons were collecting both religion and caste demographics, these indicators continue to be out of sight in these Reports. Economic profiles of prisoners are also not provided. If we did have that data provided in the Reports, one suspects it would further confirm that our prisoners are, disproportionately, coming from marginalised backgrounds.

The five-year charts in the 2024 Report show us that there has been a tapering of the prisoner population. The highs seen in the 2022 Report have, thankfully, not been witnessed again. Any positivity generated by this data is offset when we consider that that reasons for that tapering is not systemic improvement. One big reason are the efforts made in 2023 by the Supreme Court to push for release of prisoners through undertrial review committees. This is coupled by consistent decrease in prisoner population seen in Uttar Pradesh over the past couple of years. Anybody can see that these are not long-term solutions.

The 2024 Report is going to be interesting historically, as hereon the legal framework will no longer speak to the Criminal Procedure Code of 1973 but the Bharatiya Nagarik Suraksha Sanhita of 2023. 1973 saw the introduction of a new Code with an impetus on ensuring speedy justice and provisions meant to reduce unjust undertrial incarceration. It went out in 2024 with an unpleasant legacy on these fronts. The new law has built upon the same rotten foundations but with some tweaks on aspects of arrest and bail, so it would be interesting to see what changes it brings on the ground. 

A case in point will be the impact on default bail under what was Section 436-A of the 1973 Code, which has now become Section 479 of the 2023 Sanhita. This scheme of default bail under the old Code had enabled release of prisoners who had served half their prison term. The revised scheme allows for release on serving one-third of the sentence, provided it is a first-time offender (with some more caveats which we needn't elaborate here). 

Implementation of 436-A is seen by some as a chokepoint resulting in unjust incarceration. However, the PSI data has consistently shown that this may not be the case, as there was hardly a sizeable number of the prison population eligible for release in the first place. The 2024 Report reaffirms this trend. As we have seen, 479 BNSS will change the eligibility criteria, potentially expanding the reach of the clause. It will be fascinating to see the impact it has. If the half-year's worth of data is anything to go by, the future of this species of default bail may not be very different from its past.

 

Thinking about Section 27 (and 8) of the Evidence Act

Come July 2026, it would be two years since the three new criminal codes were brought into force. That is hardly long enough for the heavy, centuries-long, shadows of the old codes to disappear from the scene. At some points, you even feel sad for the new laws, which nobody can get around to either understanding or applying without referring to that familiar array of old clauses and provisions. Take Section 23 of the new Bharatiya Sakshya Adhiniyam 2023 ['BSA'] for example. Will Section 23 ever have an identity or an existence independent of its illustrious ancestors, Sections 25, 26, and 27 of the Indian Evidence Act 1872 ['IEA']? I highly doubt it.    

Ruminating about the identity crisis that will plague the new criminal codes is not the point of this post, as fun a topic as that may be. My focus is on the proviso to Section 23 of the BSA, or more specifically, its predecessor in Section 27 of the IEA. It is prompted by a sobering realisation, that there has been no post on this topic so far on this Blog.

The Statutory Scheme

Sections 24 to 26 IEA placed different kinds of barriers on evidence of confessions. Section 24 rendered any confession given due to inducement or threat from a person in authority 'irrelevant'. Section 25 stated that no confession made to a police officer by a person accused of an offence 'shall be proved' as against him. Section 26 stated that no confession made while in police custody shall be proved against a person unless it was made in the 'immediate presence' of a magistrate. Then came the proverbial backdoor in the form of Section 27, providing that where any fact is discovered pursuant to information given by a person in the custody of police, so much of the information which relates distinctly to the fact discovered. may be proved. Note that this scheme has been condensed into two provisions under the BSA; Sections 22 and 23. Section 22 is the successor to Section 24 IEA, and Section 23 of the BSA contains the erstwhile Sections 25 to 27 IEA.   

Logic and Theory

It can be argued that there are two dominant rationales guiding the legal scheme on confessions: first, respect for the 'best evidence' rule, and second, a distrust of acts done or statements made by a person who is under police control.

The best evidence rule is a hallmark of modern evidence law theory, which has two components. Most of us are familiar with its court-facing part: courts ought to decide cases based on the best evidence available to them. There is a legislative corollary, too, which commands the legislature to tear down technical rules which stand in the way of courts accepting the best evidence. For its time, the Indian Evidence Act made a lot of progress on the legislative side of things by removing older common-law hearsay objections. Today, though, the scheme of the Indian Evidence Act again appears antiquated when tested on this anvil, as can be seen with the approaches taken in other jurisdictions over the 20th Century. And, with blind repetition of this antiquated 1872 scheme in 2023, the process of law on the whole continues to suffer (yet another digression). 

In India's context, the pursuit of best evidence for courts under the adopted English system has coexisted with a general distrust of evidence purportedly given by persons under control of the police. A legal bar on using any statements, including of ordinary witnesses, recorded by police during investigation came in place since the start of codification in the 1860s. Such a step, no doubt, reduced the evidence potentially available before a court to decide a case; but the decision to remove such material from the adjudicative exercise was consciously made owing to fears about the police creating the evidence in question, which rendered it highly unreliable. Such fears were at their highest when it came to evidence of confessions taken by police from the mouth of an accused, prompting the total bar on evidence of confessions made to police.

I have offered a very generic sketch of a fascinating historical development in the mid 19th Century, but it serves our purpose for now, which is to highlight the two rationales in play guiding the legal scheme on the evidence of confessions. Understanding these two rationales, we can now proceed to appreciate the compromise arrived at by Section 27 IEA adopting the logic of confirmation by subsequent events. On its own, a statement by an accused may be worthless; but what if it leads to discovery of facts? Excluding the evidence of the statement even where it leads to discovery was seen as a step too far. As a result, Section 27 IEA allows proof of only so much of the information given by an accused which relates distinctly to the fact discovered.  

Interpretive Questions

One may be forgiven to think that the existence of strict rules barring admissibility of confessions, except in limited cases, would have been enough for the Indian criminal process to gradually move away from trying to use an accused as the primary source of gathering incriminatory material. The shift never really took place, apart from one of nomenclature where confessions are now increasingly called disclosures. Even today, a classic criminal investigation will almost always involve the accused making a disclosure, which may or may not directly lead to any recovery. The continued reliance on confessions for gathering evidence by police has naturally meant that courts are repeatedly faced with questions about admissibility of such statements. Which, in turn, means interpreting Section 27 IEA and figuring out of any part of the confession fits its contours.

I want to focus here on three interpretive questions which courts have faced and answered over time: (i) the meaning of 'fact discovered' for the clause, (ii) scope of the 'information' that is rendered admissible, and (iii) relevance of the conduct of an accused in such scenarios. There is a surprising level of coherence in judicial interpretation around these aspects, with of course existence of the odd incorrect view taken by courts from time to time.  

Fact Discovered

Since the logical premise underlying Section 27 IEA is that the discovery of a fact ensures reliability of the information given by an accused in custody of police, it is perhaps unsurprising to note that courts are strict about reading 'discovery'. If the police already knew about something, then there is no discovery of that fact made pursuant to the information given by an accused. Where this causal link is snapped, courts will not invoke Section 27 IEA at all.

This is all fairly straightforward, until you travel into the realm of determining what is a 'fact' that stands to be discovered pursuant to a statement given by an accused. A person's state of mind is also a fact, so can we argue that this is a fact discovered from the information given by an accused? The answer is no. It is not difficult to reason why this argument does not, and should not succeed. In this case, the information by an accused is the fact discovered — there is no independent confirmation of the reliability of information by any subsequent event. To apply Section 27 IEA this way would negate the very purpose of the clause and allow admission of confessions wholesale. Thankfully, this is one of the areas where the life of the law has remained closely tethered to logic, and courts have consistently insisted that the fact discovered must be a tangible object having an existence in the real world independent of the information by an accused. A great place to go and read about the development of law in this context remains the judgment in Navjot Sandhu [(2005) 11 SCC 600] 

'So Much' of the Information     

A common mistake is to think that discovery of a fact renders the evidence of a confession admissible in its entirety. The clause is much narrower in scope, admitting only evidence of so much of the information by an accused which distinctly relates to the discovery, and nothing more. Faithful application of Section 27 IEA thus requires to things in turn: first, clear and cogent material to show what was the information given by an accused, as the Supreme Court recently reminded police agencies in Boby v. State of Kerala [2023 INSC 13]. Second, courts to run a scalpel over such information to excise everything other than so much of the information which relates distinctly to the fact discovered. A good example of this approach of wielding a scalpel in action is Mohd. Inayatullah [(1976) 1 SCC 828]. 

What about Conduct?

It is quite often that an accused person does not merely give a statement, but either points out a spot from where an object is found or, even better, actually produces it himself. At the same time, an accused may straightaway point to a spot or produce an object without making any clear statement about the facts. Such scenarios throw up a different kind of evidence for a court to consider: the evidence of conduct, which is separately relevant under Section 8 of the IEA. How to sort such situations out?  

Conduct mixed with statements is a problem catered to by the statute. Section 8 itself clarifies that conduct does not include statements, except those which explain conduct. So it cannot be that by using conduct we can somehow broaden the scope of what is permitted through the narrow confines of Section 27 IEA, and say that the statements themselves are independently made relevant. There is another reason for this; any statements to police are unusable in evidence by virtue of Section 162 of the Criminal Procedure Code 1973. Section 27 is a statutory exception to this bar, but Section 8 is not.

When it comes to conduct which is not accompanied by any statements, we are faced with the same issue of the best evidence rule tussling with the distrust of evidence gathered from an accused by the police. Intuitively, an accused leading the police to material objects connected to the case is a relevant fact, which law should allow evidence for. But the risks of police creating evidence, which we saw associated with the evidence of statements made to police officers, do not magically disappear if we replace statements with conduct. 

Courts have held that these tests remain as important where police claim that the accused showed us a spot, or led to a recovery, although there is no comparable statutory prohibition in this scenario. As a rule, we see courts insist on some discovery of fact to render evidence of conduct relevant and admissible, and analyse whether it was truly the conduct which led to the discovery or was it a fact already known to the police. Additionally, courts also repeatedly insist that conduct evidence alone is quite a poor indicator for proof, and requires solid corroboration for it to further along the chain of circumstances being created in any case.                  

Constitutional Conundrums?

After the introduction of the Constitution in 1950, the evidentiary scheme of confessions has come within the constitutional crosshairs on a two notable occasions before the Supreme Court (there have been some High Court decisions as well). 

The first of these was Deoman Upadhyay [AIR 1960 SC 1125] where the validity of Section 27 was questioned on grounds of arbitrariness. The clause only applies to an accused person in police custody, prompting the unsuccessful argument that this invidiously discriminates between those in the custody of police and others. The second, was Kathi Kalu Oghad [AIR 1961 SC 1808]. Unlike Deoman Upadhyay, the constitutional question was only indirectly raised, as the Supreme Court was asked to decide whether there was a presumption of coercion accompanying statements made in police custody, thereby rendering Section 27 IEA contrary to the right against compelled self-incrimination. The Court unanimously rejected this argument and insisted that mere interactions with the police could not be deemed coercive [Note, though, that 15 years later a different albeit smaller bench in Nandini Satpathy did consider such issues more favourably].

The Future

It is quite likely that the current state of affairs will remain undisturbed in the short-term future; for no reason other than the legislative call to stick with the antiquated IEA approach even in the 21st Century. It will be interesting, though, to see what happens in the next 10-15 years as more interactions with the police have the benefit of audio-video recordings. Will a contemporaneous recording of the interaction prove to be the antidote for centuries of distrust regarding police affairs? With sufficient safeguards in place, there is no reason for it not to. Once this hurdle is crossed, the underlying basis of Section 27 will also be open to contestation. But not before that.