Guest Post: Decoding the Doctrinal Shift in Section 223 BNSS (?)

(This is a guest post by Akshitha P. Hariharan)

The Bharatiya Nagarik Suraksha Sanhita, 2023 (“BNSS”) has replaced the Code of Criminal Procedure, 1973 (“CrPC”) with considerable fanfare about reform and expedition. Within a year of the BNSS coming into force in July 2024, Section 223 of the BNSS (“Section 223”) had produced a body of conflicting decisions and a procedural puzzle that courts across the country are navigating differently on any given day. The cause of all this is, on its face, the insertion of a new word and a proviso to the erstwhile statutory text.

Section 200 of the old CrPC required that “a Magistrate taking cognizance of an offence on complaint shall examine upon oath the complainant.” 

Section 223(1) of the BNSS now reads “a Magistrate having jurisdiction while taking cognizance of an offence on complaint shall examine upon oath the complainant.” And then, it carries for the first time: “Provided that no cognizance of an offence shall be taken by the Magistrate without giving the accused an opportunity of being heard.”

Two questions have consequently divided the High Courts. First, does the phrase “while taking cognizance” reposition the complainant’s examination as a step preceding formal cognizance, contrary to the settled position under the CrPC that such examination is post-cognizance? Second, at what point must the accused be heard under the new proviso: immediately upon perusal of the complaint or only after the sworn statements of the complainant and witnesses have been recorded?

This piece attempts a considered analysis of both questions. The competing readings of “while” are then examined, followed by an assessment of why the sequential reading, despite the weight of Supreme Court authority it must contend with, is the more defensible one. The scope of the hearing under the proviso is addressed thereafter, along with considering just what a pending reference in one of the High Courts may resolve. 

The Old Regime and Inherent Ambiguity

Under the CrPC, criminal procedure in private complaint cases followed a sequential model. At the stage of cognizance which was understood as the Magistrate judicially applying mind to the complaint, no hearing was required of either party. A Magistrate simply took notice of the complaint. Examination of the complainant under Section 200 followed, and this examination was understood, on the dominant view, to have been done after cognizance had already been taken i.e., post-cognizance. 

There was, however, an inherent textual ambiguity in the old Section 200 that courts largely papered over (as discussed on the blog), in arriving at this dominant post-cognizance view. 

The opening words, “a Magistrate taking cognizance of an offence on complaint shall examine upon oath the complainant”, are grammatically indeterminate. The lack of clarity over the statute’s use of active voice and present participle (“taking” rather than “having taken”) create a lacuna. It was unclear whether it described an examination that is part of the very act of taking cognizance itself, or if it prescribed a mandatory procedural step consequent to having already taken cognizance.

Left to resolve this friction through construction, the Supreme Court generated two parallel streams of judgments that vacillated on the exact moment cognizance was triggered. One view, articulated in decisions like CREF Finance Ltd. v. Shree Shanthi Homes (P) Ltd., postulated that cognizance is taken at the initial stage when a Magistrate peruses the complaint to ascertain if an offence is disclosed, treating it as a step strictly preceding the examination of witnesses. This was further mapped in S.K. Sinha, Chief Enforcement Officer v. Videocon International Ltd., where the Court held that taking cognizance under Chapter XIV operates as a “sine qua non or condition precedent” before the machinery of Chapter XV, which houses the examination of the complainant, can even commence.

A second, divergent view fused the two events. A four-Judge Bench in Jamuna Singh v. Bhadai Shah laid down the bedrock for this interpretation and noted that the very act of a Magistrate proceeding to examine a complainant under Section 200 of the CrPC “puts it beyond doubt” that cognizance has already been taken. This logic was reaffirmed by a three-Judge Bench in Devarapalli Lakshminarayana Reddy v. V. Narayana Reddy, which clarified that applying one’s mind for purposes of proceeding under Section 200 constitutes taking cognizance. The sequence, as per Nirmaljit Singh Hoon v. State of West Bengal, was that a Magistrate first takes cognizance under Section 200 and then examines the complainant to ascertain if a prima facie case exists to issue process.

Despite this oscillation between treating the examination as the absolute proof of cognizance (per Jamuna Singh) or as a distinct subsequent step (per S.K. Sinha), the dominant consensus, as ultimately reaffirmed by a Constitution Bench in Sarah Mathew v. Institute of Cardio Vascular Diseases, was that Chapter XV of the CrPC functioned unequivocally as a post-cognizance chapter. Therefore, under the old regime, the examination of the complainant followed from, rather than preceded, the taking of cognizance.

Relocating the Examination through the Addition of ‘While’ and the Scope of Hearing

The case for treating the examination as part of the cognizance process rather than as post-cognizance in the new regime runs as follows. Parliament offered no explanatory note on the insertion of “while”, but the word most naturally denotes simultaneity or an ongoing process. “A Magistrate while taking cognizance... shall examine” is most readily read as positioning the examination within the process of taking cognizance and not as something that follows it. 

When read alongside the proviso, which prohibits taking cognizance without first hearing the accused, the implication is that cognizance cannot formally culminate until both the examination and the hearing have occurred. The Madras High Court in Seeman v. Varun Kumar captured this shift and held that “cognizance is no longer a single instantaneous act but a process within which an intermediate point exists at which notice to the proposed accused is envisaged” (¶33). This suggests that the Magistrate is still in the process of arriving at the cognizance decision when examining the complainant.

The Karnataka High Court in Sri Basanagouda R. Patil v. Sri Shivananda S. Patil mapped the sequence to be comprising a complaint filed, complainant examined on oath, notice issued to the accused appending the complaint and sworn statements, hearing, and only then formal cognizance and issuance of process. The Court held that the “question of taking of cognizance would not arise at this juncture”. Instead, this examination generates the substantive material, which refers to the sworn statements, that must be appended to the notice issued to the accused to ensure the hearing is not an “empty formality” (¶10, ¶11). Therefore, notice is issued at that stage, and only after hearing the accused does a Magistrate take formal cognizance and regulate further procedure. So far, the High Courts in Delhi (1, 2), Allahabad, Calcutta (1, 2, 3), Punjab & Haryana, Chhattisgarh, Kerala (1, 2), and Gauhati have adopted the same sequence.

There is also an argument from the proviso’s workability if one attempts to strictly apply the historical Supreme Court standard of cognizance. Per S.K. Sinha, the Court ruled that taking cognizance is the “sine qua non or condition precedent” that must occur prior to the commencement of criminal proceedings. However, the first proviso to Section 223(1) of the BNSS strictly dictates that “no cognizance of an offence shall be taken” without first giving the accused an opportunity of being heard. On this reading, the proviso would be violated the moment the complaint is read. Indian courts have long resisted constructions that produce this kind of structural impossibility or cause procedural paralysis, under the maxim lex non cogit ad impossibilia, the law does not compel what cannot be performed. A provision should not, where another reading is available, be construed so as to make its own compliance impossible from the first moment it applies.

In defining the outer limit, in Seeman, the Madras High Court held that the pre-cognizance hearing permits the accused to make submissions and place defensive material to inform the Magistrate’s evaluation, but absolutely prohibits the cross-examination of witnesses or any full-blown adjudication. However, once again, there is ambiguity regarding the exact utility of this defensive material. Under Section 226 of the BNSS, the dismissal of a complaint must rest exclusively on the sworn statements of the complainant and witnesses (and the result of any Magistrate’s inquiry). Thus, while an accused might successfully highlight a jurisdictional bar or a lack of statutory sanction during the hearing, the Magistrate’s ultimate order of dismissal must be anchored in the inherent weakness of the complainant’s own case.

Furthermore, this threshold hearing also invites unresolved friction regarding Article 20(3) of the Constitution of India. If the accused’s preliminary submissions form part of the record, there is a severe risk that these early defensive assertions could be weaponised against their maker. Until courts clarify whether such pre-cognizance submissions enjoy privilege, this statutory protection risks becoming a self-incrimination trap before the practice of filing detailed written replies becomes entrenched.

The Rita Bakshi Reference and the Doctrinal Dilemma

Perhaps the most detailed exposition between the conflicting positions is found in the Delhi High Court’s March 2026 decision, Dr. Rita Bakshi v. Seema Bajaj & Anr. Dr. Swarana Kanta Sharma, J., while finding herself bound by coordinate-bench decisions (Brand Protectors and Neeti Sharma), noted the structural flaws in the High Courts’ new “sequential reading” and referred the questions of law to a Larger Bench.

While it is to be said that the sequential reading adopted by the High Courts offers a practical exit from the bottleneck created by Section 223(1), it has nonetheless invited severe criticism. This rests on three primary doctrinal pillars: first, the irrelevance of the statutory modifications, second, the impermissibility of retracing judicial steps, and third, the destruction of the speedy trial mandate. Each will be addressed in turn. 

First, it has been advanced that the sequential reading rests on a weak grammatical base and the phrase “having jurisdiction” in Section 223(1) BNSS is dismissed as mere statutory surplusage, since a Magistrate inherently cannot conduct an inquiry without possessing jurisdiction under the statute. Furthermore, the insertion of the word “while” does not alter the fundamental curial act of taking cognizance. Further, as per Supreme Court judgments from R.R. Chari to Jamuna Singh, cognizance is triggered the moment a Magistrate applies their judicial mind for the purpose of proceeding under the chapter. Therefore, on this view, the examination of witnesses cannot be shifted to a pre-cognizance stage.

However, while the addition of “while” is indeed a clumsy legislative change, the High Courts had to improvise to prevent Section 223(1) from becoming self-defeating. If cognizance is treated as occurring the instant a Magistrate reads the complaint, the new proviso would be breached at that very moment. To avoid this impossibility, courts in Brand Protectors and Basanagouda have relied on the word “while” to reconceptualise cognizance as a processual window.

Second, on the view that as per Jamuna Singh, examining the complainant on oath itself shows that cognizance has already been taken. The High Courts’ new “sequential” model (examining witnesses, then issuing notice, then deciding on cognizance) seemingly forces Magistrates to move backwards from post‑cognizance to pre‑cognizance, which they have no power to do since they cannot review or undo cognizance orders. Nonetheless, the counter is that High Courts (like in Basanagouda) avoid this by redefining the timeline and that the examination is treated as pre‑cognizance material-gathering, and formal cognizance occurs only at the end, after hearing the accused. So, when a complaint is dismissed at that stage, the Magistrate is not retracing steps or reviewing an earlier order, but rather simply refusing to take cognizance in the first place.

Third, the sequential reading is criticised for undermining the legislature’s goal of expeditious trials by drawing the accused into a pre‑cognizance “mini‑trial.” An accused could reasonably seek to cross‑examine the complainant, lead defence evidence, or move Section 94 BNSS petitions for document production, spawning revisional and appellate challenges if refused and stalling the case at the threshold. It is undeniable Parliament itself introduced delay through the first proviso to filter out frivolous complaints. The real question is therefore not whether delay exists, but how to make the delayed hearing meaningful. If the strict doctrinal view is followed and the accused is summoned on the bare complaint, the hearing degenerates into an “empty formality” on unsworn allegations. As Basanagouda recognises, a genuine opportunity of being heard requires that notice to the accused enclose both the complaint and sworn statements. The delay is thus an unavoidable consequence of poor drafting. The sequential reading merely ensures that this cost yields real procedural fairness to the accused rather than a hollow ritual.

The delay is the inevitable price extracted by the poor drafting of the BNSS. Regardless, the sequential reading would just ensure that this price buys actual procedural fairness for the accused rather than a hollow ritual.

Conclusion - What the Reference Must (hopefully) Resolve

Section 223 BNSS has plunged cognizance, already the subject of decades of contested jurisprudence, into more complications. This self-inflicted confusion arose owing to the Parliament’s introduction of a substantive safeguard without attending to its interaction with inherited doctrine, and without explaining why. 

A year in, the provision has plainly not made private complaint cases proceed faster. Section 223 BNSS has generated conflicting judicial sequences, inconsistent approaches to the timing of cognizance, and unresolved questions about the scope of the accused’s right to be heard.

The provision also raises broader questions that remain unexplored. As Sekhri observes, it is difficult to discern the principled basis for extending a pre-cognizance hearing to accused persons in private complaints, and in Enforcement Directorate prosecutions such as under the Prevention of Money Laundering Act, 2002 (as per Kushal Kumar Agarwal v. Directorate of Enforcement), while not giving an equivalent safeguard to persons proceeded against through the ordinary FIR route. Whether this differential treatment withstands scrutiny under Article 14 is a constitutional question that may eventually require independent consideration.

If the Rita Bakshi referral reaches the Supreme Court and it endorses the sequential reading, it will accept that some delay is the necessary cost of ensuring an accused is heard on a meaningful evidentiary foundation. If it rejects that reading, it must articulate an alternative construction that gives genuine effect to the proviso without rendering it practically unworkable. Until then, Section 223 will remain less a provision that streamlines criminal procedure and more a tale of caution of how a single poorly drafted legislative amendment can unsettle an entire procedural framework.

Guest Post: The Boundary between Trafficking and Possession in Counterfeit Cases

(This is a guest post by Jishnudeep Kolay) 

Imagine a person walking down a city street, with a bundle of 500 counterfeit notes of INR 500 in a bag. He has neither sold them nor attempted to tender them as legal. He is simply walking. Under the IPC, he could have been charged with an offence under Section 489C, for possession of Fake Indian Currency Notes (‘FICN’), with an intention to use as genuine — an offence that is punishable with a maximum sentence of seven years. 

However, the Calcutta High Court’s judgment in the reference case of Anikul v State (2025), somehow turns this into an offence under Section 489B, for trafficking in FICN, which carries a maximum sentence of life imprisonment. For doing so, it converts static possession into “active transportation.” 

The core issue in the reference was this: can recovery of a substantial amount of FICN raise a presumption of active transportation of FICN, thus make it trafficking, which is punishable under Section 489B. The High Court answered in the affirmative, holding that the presence of a substantial amount of FICN shall be deemed to be active transportation, punishable under Section 489B.

This post dissects the judicial doctrine and argues that the answer to the reference misconstrued the textual framework of Sections 489B and 489C. The two relevant Supreme Court judgements in Mammutti (AIR 1979 SC 1705) and Umashankar (AIR 2001 SC 3074) do not tackle the specific issue in raised in Anikul, but simply affirm that the requisite mens rea must be proved. Existing High Court judgements, mostly by the Calcutta High Court, conflict with each other, as this post shall show. Therefore, Anikul presented a great opportunity for a full bench to clarify the issue — which it missed.

The Textual Framework

Section 489B reads as follows:

“489B. Using as genuine, forged or counterfeit currency-notes or bank-notes.—Whoever sells to, or buys or receives from, any other person, or otherwise traffics in or uses as genuine, any forged or counterfeit currency-note or bank-note, knowing or having reason to believe the same to be forged or counterfeit, shall be punished with 4 [imprisonment for life], or with imprisonment of either description for a term which may extend to ten years, and shall also be liable to fine.”

While the maximum punishment for Section 489B is life imprisonment, it is 7 years for Section 489C, which reads as follows:

“Possession of forged or counterfeit currency-notes or bank-notes.—Whoever has in his possession any forged or counterfeit currency-note or bank-note, knowing or having reason to believe the same to be forged or counterfeit and intending to use the same as genuine or that it may be used as genuine, shall be punished with imprisonment of either description for a term which may extend to seven years, or with fine, or with both.”

Clearly, mere possession is intended to be punished under Section 489C and not 489B. While the mens rea element is same for both, the actus reus separates the two. The words preceding phrase “or otherwise traffics in” in 489B evince the need to prove a transaction. One might argue that the phrase “or uses as genuine” does not require the presence of another person. However, the same phrase is also used under 489C, with the same mens rea requirement. Thus, if a person is proved to knowingly use FICN as genuine then he can be convicted for an offence under Section 489C. However, the words preceding the phrase in Section 489B shows that to elevate the offence to Section 489B it, some kind of trafficking or transaction in FICN must be shown.

Conflicting Doctrine

The Supreme Court judgements on these issues are limited and have not addressed the issues that High Courts have struggled to answer. In Mammutti, the Court confirmed the requirement of knowledge that was already present in the provision. Umashankar also followed suit. 

The Calcutta High Court has had several chances to examine the provision. In Jiban Sasmal, the accused was found with 5 FICN of ten-rupee denomination. The High Court held that mere possession does not capture the actus requirement of section 489B, which requires some kind of transaction. 

While Jiban Sasmal had to deal with possession of a small quantity of FICN, in Jubeda Chitrakar, the accused was found with more than 500 pieces of FICN. There was no evidence of any kind of transaction. On the conduct aspect, the appellant argued that “otherwise traffics in” must be read ejusdem generis with the preceding words, thus requiring proof of a transaction with the presence of another person. However, the High Court held that “traffics” refers to dealing in illegal substances. Therefore, possession of a large quantity of FICN is not dormant, but active possession, which amounts to trafficking. Any other interpretation, as per the Court, would “dilute the rigour of law”. In reaching its conclusion, it relied on Section 106 of the Evidence Act, presuming that possession of such quantity of FICN is transportation with knowledge, unless rebutted. In essence, any person found with a large quantity of FICN can be made liable under both Sections 489B and 489C, even without a direct proof of any of the acts specifically mentioned in Section 489B. While the holding is in stark contrast with Jiban Sasmal, the only differentiating factor is the quantity of FICN recovered. 

Facts very similar to Chitrakar appeared before a single judge bench of the same High Court in Sarvesh Pathak. There, the accused was found in possession of over 8 lakhs worth of FICN. The state argued that such a large amount by itself proves an offence under Section 489B. The judge disagreed, holding that all that it proves is the conduct under Section 489C — possession or that they may use it as genuine — but not 489B..

The Kerala High Court has largely agreed with Chitrakar. However, its doctrine shows the importance of conduct of the accused rather than merely the quantity of FICN. In Md Islam v. CBI, the accused were found with more than 200 FICN pieces. Citing Chitrakar, the Court held the guilty as no explanation was offered. Vijayan acquitted the accused, as less than 20 notes were recovered. In Masidur Rahman (2024 SCC OnLine Ker 4650), a conviction was sustained despite recovering less than 50 notes from the accused because the accused had attempted to transact with the notes. Facts that the accused was tensed when the fake notes were discovered and his hurried attempt to leave the place, were used to prove knowledge.

The Delhi High Court has read the provisions strictly, without presuming active transportation barring proof. In Akil v. State, a substantial amount of more than 200 fake notes were recovered. However, the Court, citing Umashankar, held that mere possession or even transacting with the same without proper proof of mens rea cannot sustain a conviction under either Section 489B or 489C. Further, for Section 489B, the High Court held that there must be a transaction and mere possession is not enough. While the Akil Court did not cited Chitrakar (in any case, it was not bound by it), the logic employed was exactly the opposite. The recovery of a substantial amount of FICN was not taken as active transportation, without any evidence of transaction.

The Anikul Reference

The reference in Anikul was the perfect opportunity to clarify the position. The question asked was whether a recovery of a large quantity of FICN can be deemed to be active transportation, thus falling under the ambit of Section 489B. While it cited several judgements, the High Court failed to offer any independent reason of its own to back up the conclusions. It held that while the amount of FICN recovered is immaterial for Section 489B, a substantial amount shall be deemed to be active transportation. Before triggering the presumption, the prosecution must prove the foundational facts of Section 489B. 

Basically, the Court repeated the ratio of Chitrakar, without any independent reasoning. This is striking, since several cases cited by the Court itself go against this holding. For example, in Abu Sajeed, the Calcutta High Court itself had held that lack of evidence of a transaction shall lead to acquittal on the charges of Section 489B. 

Anikul’s view that possession of substantial quantity of FICN is “active transportation,” and therefore trafficking, conflates the offence of trafficking with simple possession which has lesser punishment under section 489C. The flaw originated in Chitrakar, which Anikul repeated without assigning any independent reasons.

The first few words under Section 489B: “sells to, or buys or receives from, any other person” indicates the requirement of some transaction. On the other hand, Section 489C simply requires possession “and intending to use the same as genuine or that it may be used as genuine.” The mens requirement under both is the same. Therefore, if an accused is found to be merely in possession of FICN, he can be charged only with Section 489C regardless of the quantity. Chitrakar held that “or otherwise traffics in” can capture possession of substantial amount of FICN as well. However, the same shall also be punishable under Section 489C, which limits the sentence to a maximum of 7 years as opposed to life imprisonment under Section 489B. If such a reading is adopted, then the difference between the two sections and the legislative intention of treated transacting in FICN as a more severe offence shall be vitiated. The defence in Chitrakar did argue for reading “or otherwise traffics in” ejusdem generis with the preceding words. However, the Court held that such a reading shall dilute the rigour of the provisions.

High Courts have also marshalled section 106 IEA. The High Courts of Calcutta and Kerala have used section 106 in the following manner: Prosecution proves substantial possession of FICN — section 106 is invoked and the onus shifts to the accused to explain the source — accused fails provide a satisfactory explanation — court presumes knowledge and the act of trafficking as well. However, section 106 can never be used to compensate for the prosecution’s failure to prove the ingredients of an offence (here and here). Thus, the Court cannot make a jump from mere possession to trafficking, by using 106 IEA, without proving any transaction of FICN. While section 106 can be deployed, the manner is different. If an accused is caught possessing a substantial amount of FICN, then 106 can be marshalled to presume that he was “intending to use the same as genuine or that it may be used as genuine.” However, presumption cannot change the conduct altogether. Anikul, by uncritically following Chitrakar, continued the flawed reading and missed a crucial opportunity to change course.

'Paper' Tigers: Why the Public Examinations Act 2024 (and other such laws) Don't Solve the Problem

A 2024 statute has come under the spotlight amidst protests in Delhi: the Public Examinations (Prevention of Unfair Means) Act, 2024. In a bid to allay protestors, it seems that this law is in line for amendments to give it more 'bite' if news reports are to be believed. Some steps have already been taken to boost the law's working by directing that prosecutions launched under it be handled by 'Fast Track Courts' — Delhi passed the necessary notification two days ago (23.07.2026).

It is fair to say that the news about amendments making this law stricter did not have the desired effect. Perhaps, it is a realisation that more laws do not matter. Perhaps, the realisation that the 2024 Act has been in place for a while now without results and it is not lacking in 'bite'. Perhaps, it is the knowledge that countless 'Fast Track Courts' have not made a difference in the past either to secure accountability through the system. 

These are all good realisations to have. The 2024 Public Examinations Act, even with its proposed amendments, is yet another example of legislators adopting the cheapest option available to resolve a socio-political problem. The formula is simple. Show that you are dealing with an issue by releasing paper tigers in the form of a new statute, which sounds impressive and costs nothing but the cost of the paper it is printed on, but do not invest in resources which would be required to operationalise the law or solve the problem on ground. We have been here many times before. 

In the present scenario, the point could not be starker. Robust public examinations would rely a lot on robust institutions conducting them. Investing in this would presumably prevent leaks, in turn reducing the need for hyper-punitive laws to punish prospective offenders. Surrounded by protests, the government has sacked 47 officers of the National Testing Agency (with very little clarity or cause), indicating that all may not have been not right under that roof. Which then leads to the following questions: What accountability and oversight existed for the institution? Lets see. Is there a governing law for it? No. Is there any delegated legislation / rule / regulation controlling its functioning? No. The NTA It is an 'autonomous body' under pure government control. In other words, no accountability for the institution itself, just pure opacity.

There is a lot more to criticise about the 2024 Act. It is a poorly drafted law with key terms going undefined and overlaps with other laws, which will create obstacles for any eventual prosecution whenever it is launched. It is a central law with limited coverage. Solely driven to satiate the lust for a punitive response to a bad event, it does not concern itself with the victims — the countless students — by (for instance) delineating measures for distributing the properties attached and confiscated upon conviction to them. It is telling that the Act itself did not create special courts to deal with prosecutions and the history of 'Fast Track Courts' demonstrates the measure may prove counterproductive and hurt ordinary citizens more by clogging up other parts of the criminal justice system. The list goes on.

The limited point here is to remind ourselves that whenever we see a statute being touted as the answer by a government in the aftermath of a crisis of confidence in public institutions, chances are that it is a distraction by calling upon your baser, more retributive, instincts.

Guest Post: The Gendered Blind Spot in India's Pre-Arrest Bail Architecture

(This is a guest post by Akhil Yadav)

The Allahabad High Court granted Swami Avimukteshwaranand Saraswati anticipatory bail on 25 March 2026 for a case which involved two boys according to the Protection of Children from Sexual Offences Act 2012 (POCSO) which the authorities had filed. The first informant has since challenged the order before the Supreme Court by way of a Special Leave Petition. The High Court only provides a short explanation of the important question which the author needs to explain. The core issue remains unidentified by the High Court and the involved parties and the following media coverage. The question investigates why Section 482(4) of the Bharatiya Nagarik Suraksha Sanhita, 2023 (BNSS) never applied to this case because of the facts and what that reveals about an existing gap in India's pre-arrest bail system which has persisted since 2018 and which the 2023 legislative changes did not resolve.

The Anticipatory Bail Bar and Its Pedigree

The High Court and Court of Session receive authority to decide pre-arrest release matters according to Section 482 BNSS which replaces Section 438 CrPC. The section establishes an exception which states that “Nothing in this section shall apply to any case involving the arrest of any person on accusation of having committed an offence under section 65 and sub-section (2) of section 70 of the Bharatiya Nyaya Sanhita, 2023.” The current provision derives directly from Section 438(4) CrPC which the Criminal Law (Amendment) Act 2018 introduced following the Unnao and Kathua incidents. The provision defined Sections 376(3) and 376AB and 376DA and 376DB of the IPC as offences which included minor titles that described ‘Punishment for rape on woman under sixteen years of age’ and ‘Punishment for gang rape on woman under sixteen years of age.’

The legislative purpose was visible and urgent: to deny perpetrators of the most heinous sexual offences against girl children the protection of pre-arrest bail. The IPC provisions from 2023 BNS and BNSS enactment consolidated into BNS Sections 65 and 70(2) which resulted in an amendment to the bail restriction. Section 65 BNS punishes rape of a woman under sixteen years with a minimum of twenty years' rigorous imprisonment. Section 70(2) prescribes life imprisonment or death for gang rape of a woman under eighteen years. The most important aspect of the bar refers to the separate provision which exists within POCSO itself. The comparison between the two laws shows how the Scheduled Castes and Scheduled Tribes (Prevention of Atrocities) Act 1989 works. The Act includes Section 18 which establishes that Section 438 CrPC which became Section 482 BNSS through the current law shows no applicability to cases that involve arrests under the Act. The system uses a complete prohibition which functions through its built-in mechanisms.

The design of the substantive offences creates a gap which exists between two parties. The BNS 2023 defines rape in gender-specific terms which determine that a man must commit the crime against a woman. The law only permits Sections 65 and 70(2) to be used for aggravated rape cases which involve female victims.

The POCSO uses a gender-neutral framework to make all sexual offences against children of any sex into criminal activities. The existing system faces a structural problem because POCSO recognizes identical offences against male and female children while BNSS 2023 anticipatory bail bar only applies to offences which fall under specific gender-based rules of the BNS. The law blocks male child offences which match serious status from the statutory prohibition.

The Gauhati High Court's Interpretation and Its Limits

The initial investigation into this disjunctive statement requires a resolution of Section 482(4) BNSS which the Gauhati High Court examined as a different textual issue. The literal interpretation of "and" which connects Section 65 to Section 70(2) creates a ridiculous situation because an accused person requires simultaneous charges under both laws yet this situation remains impossible to achieve since the laws define separate crimes which have different age requirements for victims and different penalties. The ruling in June 2025 establishes “and” should function as “or” because interpreters must follow purposive reading when literal interpretations lead to nonsensical results. Both Section 65 and Section 70(2) declare “woman” whether these sections connect through conjunction or disjunction. The court must interpret ‘woman’ according to its established meaning because no evidence exists to support a “child” definition which would expand a legal provision beyond its established boundaries through purposivism. The Gauhati HC's reading establishes wider restrictions for cases involving female victims while it maintains complete control over cases involving male child victims. The gender gap continues to exist after fixing the ‘and/or’ problem.

The Consequence in Practice

The FIR in the Prayagraj case was registered under POCSO Sections 3, 4(2), 5(1), 6, 16, and 17, alongside Section 351(3) of the BNS (criminal intimidation). The BNS Sections 65 and 70(2) charges could not exist because both sections require a female victim whereas the alleged victims are male. The anticipatory bail bar in Section 482(4) BNSS therefore never entered the picture. The Allahabad High Court used Section 482(1) framework to analyze the case according to Sushila Aggarwal v. State NCT of Delhi, which listed three factors to determine bail eligibility based on the offence's nature and severity and evidence tampering risk and flight risk.

The High Court's ruling about Section 29 shows that guilt presumption cannot start until prosecutors present charges, which matches the Delhi High Court's complete explanation in Dharmander Singh v. State (Govt. of NCT, Delhi) and the Allahabad HC's parallel decision in Monish v. State of U.P.. The court accepted the argument because pre-arrest application of Section 29 would demand a small trial and violate the right of silence while forcing an accused person to disprove an uncharged accusation. The blog has explained the Section 29 presumptive clause through its analysis of the court's reasoning, which builds a shared understanding among High Courts despite remaining areas of disagreement.

The Structural Asymmetry and What It Means

Step back from the facts of the Prayagraj case and consider what the structural gap produces in the abstract. A person accused of aggravated penetrative sexual assault of a child below twelve years, an offence under POCSO Section 5(m) carrying a minimum sentence of twenty years faces very different pre-arrest bail positions depending solely on the victim's sex. If the victim is a girl, the Explanation to Section 65 BNS brings the case within the ambit of BNS Section 65 (rape of a woman under twelve years), and Section 482(4) BNSS bars anticipatory bail entirely. The charge against him only covers POCSO when the victim is a boy since Section 482(4) does not apply to him; Section 482(1) gives the court authority to grant anticipatory bail. The gravity of the offence is identical, the minimum sentence may be identical, the harm to the child is identical but the accused's access to pre-arrest bail protection is radically different, turning entirely on whether the child is male or female.

This asymmetry is not an argument about whether anticipatory bail should or should not have been granted in any particular case. Courts retain ample discretion to refuse anticipatory bail even without a statutory bar and in cases of serious sexual offences against children that discretion should be exercised with great care. The statutory framework eliminates discretion for judges when girls become victims while boys maintain their right to judicial discretion.

Towards a Fix

Two distinct paths exist which lead to a solution. The narrower and more immediately achievable solution requires Section 482(4) BNSS to include POCSO Sections 5 and 6 which define aggravated penetrative sexual assault and its punishment. The proposed change would expand the anticipatory bail prohibition to include serious sexual offences against male children which POCSO recognizes. The BNS rape provisions should receive gender-neutral treatment of victims according to the recommendations made by the Law Commission and the Justice Verma Committee and the 2013 Ordinance since this change would extend BNS Sections 65 and 70(2) to male child rape victims which leads to automatic application of Section 482(4). The proposed reform has faced ongoing resistance which makes it improbable to become the speedier solution to the problem. Section 482(4) BNSS will maintain its protection yet which depends on the child's sex until one of these methods gets implemented. This result contradicts the gender-neutral child protection system which POCSO established through its implementation.

Codification of Criminal Law (Part 3): The Evidence Act and BSA

The third and final instalment of this series continues the discussion on India's criminal codes and turns our attention to the Indian Evidence Act 1872 [IEA], replaced in 2023 by the Bharatiya Sakshya Adhiniyam [BSA]. The IEA (and BSA) is not strictly a 'criminal' code, though, as the rules of evidence it lays down apply to civil trials too. That does not affect the points made in this post, I think.

Same pattern. Four themes, on aspects go beyond the substantive content — is the definition of hearsay good or not, etc. — but touch upon the impact the code has on the legal system.

Theme 1 — Evidence, for what kind of Trial? 

At the time when the IEA was passed for British India, the legal system was a peculiar creature. Different defendants would be entitled to different trials, on the basis of race, and the trials themselves were a mix of jury trials and bench trials. Why does that matter? A law on the subject of evidence is, obviously and undeniably, linked to how that legal system is processing cases. A set of rules governing reception of evidence in jury trials is not ideally suited to a trial conducted by a judge alone. In the former a jury decides questions of fact, requiring some care to ensure a jury is not influenced by bad quality evidence. In the latter, considering it is a judge helming the case, this is not seen to be a problem.

Part of the genius behind the IEA was that it successfully presented a system of evidence which did not trouble courts regardless of the trial — and also, civil trials. But as India has abandoned jury trials outright, there is no need to continue with a system designed to set out a kind of lowest common denominator for our courts. This point is developed more thoroughly in the next two, interlinked themes.

Added here is the problem of staggered trials. The trial envisioned by the IEA (and now BSA) is one continuing over day-to-day, focused on oral evidence delivered before a judge who is taking careful notes over the demeanour of witnesses, which remains fresh in the jury / judge's mind by the time judgment day comes. They also imagine a witness to remember things, assuming not a lot of time has passed between the incident and trial. Present day trials are anything but — they take years to commence, and once they do they are stop-start affairs, relying very little on the orality of proceedings but instead on the written copy of what the witness has deposed. Older rules on examining witnesses focused on orality end up creating more problems rather than solutions.      

Theme 2 — Admissibility and 'Un-ringing' the Bell

The jury trial / judge trial distinction becomes relevant especially when we think of how the IEA / BSA deal with the issue of objections to the 'admissibility' of evidence. An easy example will do: the IEA and BSA both render evidence of confessions made to police officers inadmissible (with a limited exception, which we needn't bother with for now). 

In a jury trial, if such a confession was being led, counsel could pre-emptively object and ensure the jury isn't using this material in its decision. But what about a trial held by the judge? This material would go before a judge, who will look at it, and if an objection is raised the judge will have to decide it. Even if the judge holds that the material was inadmissible, can we really, truly, say that the confession did not have any impact on their decision? Can we, in other words, really expect that the text of the law can 'un-ring' the bell which has rung when the material is first seen by the judge?

If we stick to a theory of judges being supreme beings always bound to go by the law, then sure, their ultimate decision can never be affected by them being shown inadmissible evidence at the trial. But, lets be honest, we all know that is not true: seen in how appellate courts across India routinely overturn decisions on grounds of improper consideration of evidence by trial courts. To think that this is a problem with judges is, I would argue, a shortsighted view. The scheme of the IEA / BSA contributes to this failing, by creating rules and then applying them unreasonably by asking the courts to artificially splice their mind over the life of a case.      

Theme 3 — Reconsidering the '6-55' Relevance Scheme

Flowing from this is a critique of the '6-55' scheme — this refers to a detailed set of provisions in the IEA from Section 6 till 55, explaining to us the concept of 'relevance'. Bear with me as I spell it out.

The IEA, as part of a logical turn in evidence theory in the 19th Century, follows a simple idea - anything relevant should be admissible. The shift was seen as welcome, doing away with many artificial rules barring all kinds of useful material from consideration of a court. The endorses this logic in section 5, and goes on to spell out what is meant by 'relevant' for purposes of trials in Indian courts through Sections 6 to 55 of the Act. So we find that evidence of cause and effect, of motive, of alibi — all of these are, naturally, relevant, and so naturally kinds of evidence that can be led at trial.

Dumbing down the concept of relevance was intentional. Stephen felt that it would allow for greater consistency across a jurisdiction not fully familiar with English law concepts. The scheme has its admirers till today — no doubt also in the Government which chose to retain it lock, stock and barrel (minus some pre-partition references in the illustrations). But having such a detailed scheme of relevance comes with its problems. There are limits to language, and breaking up the ordinary goings-on of life into Victorian prose creates confusions in a world which has moved far from that context. It also creates a problem of exclusion - what is not in Sections 6-55 is beyond the pale. Why should it be so? Can't we trust the judge — the same judge who we trust to exclude inadmissible evidence from consideration — to decide what is relevant and what is not?

Linked to this is another problem with the IEA approach to admissibility, adopted in the BSA. By stating that all that is relevant is admissible, without considering that some of this evidence many be tainted due to constitutional guarantees which came in after 1872 - either directly or by implication. Surely the constitutional barrier cannot be undermined by statutory rules, but we see time and again courts are stuck in this dilemma, most commonly in context of the right against compelled self-incrimination and most recently in context of the fundamental right to privacy.

The 'relevant = admissible' coda also comes unstuck because of the asterisk that an IEA placed at the time, which has become an increasing point of debate as the world has gone digital. Yes, all that is relevant is admissible, but for some kinds of material there are additional hoops that have to be crossed. Documents must be original; Wills must have two attesting witnesses; And, most relevant for us, Printouts must carry certificates of experts. Having such rules might have made sense at the time as a compromise to the logic of relevance, but does it make sense to retain them even after the IEA embedded its roots in the system over 150 years?                  

Theme 4 — Codifying Appreciation of Evidence?

The IEA and BSA extensively deal with admissibility and relevance of evidence and facts. But they conspicuously say very little about assessing appreciation of evidence by a judge. Does this mean that there are no rules governing this field, and they are left to the whims and caprices of a judge? Not at all. There are rules which have been cemented over time, operating at par with the codes while not being codified. A good example is the treatment of approver's testimony — not enough without corroboration — and consideration of a case based on circumstantial evidence.

Why did the IEA and BSA not spend so much time on weighing evidence? Because, traditionally, it was a task left for the jury which was guided by a judge in each case by way of the judicial instructions given at the close of evidence. It was these instructions which birthed many of these rules in the first place. Now, even as there are no juries, the value of these concepts has meant that judges continue to adhere to them as if they are edicts. If we are sticking to codes, then why not add these rules as well, to leave no matter of doubt and perhaps improve upon them?

Concluding the Series

The Indian Penal Code, Criminal Procedure Code, and Evidence Act, were living monuments by the time of their repeal in 2023. Like any other living monument, they stood still (mostly) while the world around them drastically changed. This created a paradox where legal codes hardwired on 19th Century problems and solutions for a former colony kept being used to deal with 21st Century problems faced by the most populous country in the world. Recognising this, in 2023 the Indian government finally decided to pull the plug on the codes. Thus, on July 1, 2024, they were replaced with three new laws which operate ... almost exactly the same way that the colonial codes did.

The three posts in this short series sought to highlight some themes for discussion surrounding the three codes governing the Indian criminal process (now, from beyond the grave). With a focus on systemic issues and concepts, rather than specific nitty-gritty type issues, the idea is not to suggest, even for a moment, that these are the only themes meriting discussion or that the thoughts here are the only take on those themes. The point, as always, is simply to sustain the ongoing conversation around the architecture of the Indian criminal process.