Law & Policy
The Index Is Part of the Law: What the 2024 Renumbering Actually Cost
11 August 2026 · 8 min read
On 1 July 2024 the Bharatiya Nyaya Sanhita, the Bharatiya Nagarik Suraksha Sanhita and the Bharatiya Sakshya Adhiniyam replaced the Indian Penal Code, the Code of Criminal Procedure and the Evidence Act. The public argument about that replacement was, understandably, about content: which offences were new, which were dropped, what the timelines require, how much of the procedure had been moved to an electronic record.
Very little of the argument was about numbering. Yet for the people who operate the system daily — trial courts, listing sections, and counsel appearing in several matters at different stages on the same day — the renumbering was the change that arrived first and has proved the most persistent. It is worth asking why a purely formal change should have that effect, because the answer says something general about what codification is for.
What a section number is doing
In a working legal system, a section number is not a label attached to a rule. It is an address in an index that everybody uses at the same time.
That index is what allows a four-word phrase to carry a complete procedural posture. "A 138 complaint" — under the Negotiable Instruments Act — tells a listing clerk the forum, the nature of the proceeding, the stage at which it will be taken up and the range of orders available. "A 482 petition" told a High Court registry, without further explanation, that the matter was an invocation of inherent jurisdiction and not an appeal or a revision. "A 156(3) application" identified, in a few characters, a particular power exercised by a particular court at a particular point before cognizance.
None of that compression is decoration. It is how three people who have never met — the clerk who lists the matter, the Bench that reads it, the advocate who filed it — arrive at the same object without first negotiating what they mean by it. That convergence took decades to accumulate and appears in nobody's budget. It is unpriced infrastructure, which is precisely the sort of asset a reform exercise will overlook, because nothing in the process obliges anyone to account for it.
Displacement is bounded; collision is not
Two costs are involved here, and they behave very differently.
The first is displacement. Murder moved from Section 302 IPC to Section 103 BNS. Cheating, previously distributed across Sections 417, 418 and 420 IPC, was consolidated into Section 318 BNS, with cheating relating to property at sub-section (4). The Magistrate's power to direct an investigation moved from Section 156(3) CrPC to Section 175(3) BNSS. The High Court's inherent power moved from Section 482 CrPC to Section 528 BNSS.
Displacement is a lookup problem. It costs a mapping table, a couple of years of hesitation, and a generation of practitioners who will carry both numbers in their heads for the rest of their careers. That is a genuine cost, but it is bounded and it is payable once.
The second cost is collision, and it is a different thing. The old numbers were not merely vacated; they were reassigned. This was in part unavoidable: the BNS runs to 358 sections and the BNSS to 531, so most numbers that meant something under the old codes now mean something under the new ones. Reuse was not an oversight. It was the arithmetic consequence of renumbering two codes of comparable length.
Two examples show that the cost of a collision has little to do with how well known the number is.
Section 302 is the most recognisable number in Indian criminal practice. Under the BNS it denotes the offence of uttering a word, making a sound or a gesture, or placing an object, with the deliberate intention of wounding the religious feelings of another person — carrying forward what was Section 298 IPC, and punishable with imprisonment which may extend to one year. The number that meant a capital offence for over a century now means a one-year offence about religious sentiment.
That collision is jarring, but it is not especially dangerous. Murder and wounding religious feelings never occupy the same conversational slot. Nobody drafts a charge, files a bail application or reads a cause list in which the two are plausible alternatives. An error here announces itself immediately, and errors that announce themselves are cheap.
Section 482 is the opposite case. Under the CrPC it was the High Court's inherent power. Under the BNSS it is the provision for directions granting bail to a person apprehending arrest — anticipatory bail, previously Section 438 CrPC. Both referents are High Court criminal jurisdictions. Both are commonly invoked at the pre-trial stage. Both appear in the same cause list, filed by the same counsel, on the same morning.
Here an error does not announce itself. It produces a note, an instruction or a description that reads as entirely coherent until somebody checks it against the file. This is the general rule: the cost of a collision scales with the semantic proximity of the old referent and the new one, not with the prominence of the number. A well-known number reassigned to a distant subject is a curiosity. An ordinary number reassigned to a neighbouring subject is a trap.
And in the one case where the trap could have been removed for nothing, it was not. Indian drafting practice has long avoided renumbering for exactly this reason: new provisions are inserted with letter suffixes — 304A, 498A, 153AA — and repealed ones are marked "omitted" rather than closing the gap, so that a number, once given a meaning, keeps it or keeps nothing. Leaving 482 vacant in the new Sanhita would have cost a single line of untidiness and would have eliminated an entire class of error. The convention existed. It was simply not applied at the points where it mattered most.
The transition is not a date
The second structural point is that the changeover was never going to be an event. It is a tail, and the length of the tail is set by pendency.
Two rules produce this. Article 20(1) of the Constitution prevents conviction for an offence except for violation of a law in force at the time of the act, and prevents a penalty greater than that which could have been inflicted under the law in force at the time. An offence committed before 1 July 2024 therefore remains an offence under the IPC, and will be charged and punished under the IPC however long it takes to reach judgment.
On the procedural side, Section 531 BNSS repeals the CrPC but saves what was already running: where an appeal, application, trial, inquiry or investigation was pending immediately before the Sanhita came into force, it is to be disposed of, continued, held or made in accordance with the CrPC. Precisely where the line falls between a proceeding that was already pending and one initiated afterwards has itself become a litigated question, which is the ordinary fate of savings clauses.
The practical result is that both codes are live law simultaneously, and which one governs a given file is settled by the date of the offence and the date of institution rather than by today's date. A criminal practitioner in 2026 is not migrating between codes. They are running two of them concurrently, and keeping track, file by file, of which regime each matter belongs to. That state of affairs will not end when the profession finishes learning the new numbers. It ends with the disposal of the last matter instituted before July 2024 — and anyone who has watched how long a criminal appeal can remain pending will not expect that soon.
The narrow claim
None of this is an argument that the codes should not have been replaced. Whether the substantive changes are improvements is a separate question, and each provision has to be argued on its own merits.
The narrower claim is that renumbering was treated as a costless by-product of replacement, when it is an independent decision with independent consequences. Retaining the existing numbering for retained provisions was always available; so was renumbering while declining to reuse a retired number for a neighbouring subject. Neither option would have constrained a single policy choice about what the criminal law should contain.
The cost of not taking them falls unevenly. A large practice absorbs it with a mapping chart and a junior. It is absorbed less easily by district courts working through long daily lists, by clerks and typists reproducing section numbers from dictation, and by litigants attempting to read their own orders. The burden of an index change lands on whoever has the least capacity to verify.
The deeper point is about what codification is for. A code's value has never rested only on the quality of the rules inside it. It rests on the fact that everyone can find the same rule quickly and refer to it in the same words. Findability is not a convenience laid over the law; for anyone who has to use the law under time pressure, it is a substantial part of what makes the law usable at all. When the index is replaced, a stock of accumulated coordination is spent — and unlike a badly drafted rule, that expenditure cannot be repaired by an amendment. It can only be waited out.
Authored by Eshan Kumar Gupta, Advocate, Hon'ble Allahabad High Court, Lucknow Bench.