Finality is indispensable to the administration of justice. Adjudication must, at some point, come to an end; without finality, a judgment would mark merely another stage in litigation rather than its conclusion.

Yet, finality cannot itself justify perpetuating an obvious judicial error. The power of review is, therefore, founded upon a fundamental paradox: the court must ordinarily have the last word, yet the law must retain a limited capacity to correct the court's own error. This principle was recognised by the Supreme Court in S Nagaraj v. State of Karnataka.

The Indian legal system accommodates this paradox through a narrowly defined power of review. Section 114 and Order XLVII of the Code of Civil Procedure (CPC) provide the statutory framework in civil proceedings, while Article 137 expressly recognises the Supreme Court's constitutional power to review its judgments and orders. The existence of this power is itself significant. Article 137 recognises that finality does not necessarily imply infallibility.

Section 114 CPC provides the statutory basis for review, subject to prescribed conditions. Order XLVII Rule 1 identifies the principal grounds: discovery of new and important matter or evidence despite due diligence; mistake or error apparent on the face of the record; and “any other sufficient reason”. The Explanation to Rule 1 excludes, as a ground for review, a subsequent reversal of law by a superior court or a subsequent decision of a co-ordinate or larger Bench. The Constitution Bench reaffirmed this position in Beghar Foundation v. Justice KS Puttaswamy.

Article 137 operates at the constitutional plane. It provides that subject to parliamentary law and the rules made under Article 145, the Supreme Court has power to review any judgment it has pronounced or order it has made.

The distinction between the two sources matters: Section 114 creates the statutory framework, while Article 137 constitutionally recognises the Supreme Court's power of self-correction. Neither provision, however, contemplates an unrestricted rehearing.

The defining characteristic of review jurisdiction is restraint. From Moran Mar Basselios Catholicos (1954) to S Madhusudhan Reddy (2022), the Supreme Court has consistently held that review is fundamentally different from appeal. An appeal challenges the correctness of a decision; review addresses a legally recognised defect warranting reconsideration.

The distinction is especially important in relation to an “error apparent on the face of the record”. An error is not apparent merely because elaborate argument reveals another possible view of the matter, or because the alleged error requires reasoning, re-appreciation of evidence, or a fresh weighing of competing interpretations. Such a grievance belongs to appellate, not review, jurisdiction. An erroneous decision is not necessarily an error apparent on the face of the record; the error must be self-evident, not one discoverable only through a long chain of reasoning, as explained in Parsion Devi v. Sumitri Devi and Meera Bhanja v. Nirmala Kumari Choudhury.

This distinction protects the finality of judgments from being undermined by re-argument dressed up as review.

Finality serves certainty, judicial economy and institutional authority. But finality is a means to justice, not its substitute. Review, therefore, performs a corrective function. Where a judgment rests on a patent error, overlooks material that could not reasonably have been produced earlier, or otherwise falls within a recognised ground of review, the law permits the court to revisit its decision.

The phrase “any other sufficient reason” cannot be read as an open-ended power to reopen a concluded adjudication; its application must remain consistent with the restrictive character of review jurisdiction. Otherwise, the exception would swallow the rule and review would become appeal by another name. The proper inquiry, therefore, is not whether the judgment is wrong, but whether the defect is of a kind the law permits to displace finality, as recognised in Chhajju Ram v. Neki.

Review must also be distinguished from recall. A review presupposes a valid adjudication sought to be reconsidered on a recognised ground. Recall, in contrast, operates in a narrower procedural field where the validity of the adjudicatory process itself is impeached, such as in cases involving patent lack of jurisdiction, fraud or violation of natural justice. It cannot be used merely to circumvent the limits governing substantive review.

Beyond review lies the exceptional doctrine of curative jurisdiction. In Rupa Ashok Hurra, the Supreme Court recognised the curative petition as an extraordinary mechanism to prevent gross miscarriage of justice even after dismissal of a review petition. The jurisdiction is deliberately narrow because an unrestricted curative power would effectively destroy finality.

The existence of curative jurisdiction demonstrates that in the rarest circumstances, finality remains subject to an exceptional constitutional power of correction where the judicial process has failed in a sufficiently grave manner.

Delhi Metro Rail Corporation Ltd. v. Delhi Airport Metro Express Pvt. Ltd. illustrates the continuing constitutional significance of this safety valve, while equally cautioning that curative jurisdiction cannot become a further stage of ordinary appellate review.

The remedial framework may, therefore, be understood as a progression from the finality of adjudication, to limited review and, in the rarest cases, to curative intervention. At each stage, the threshold rises and the permissible scope of intervention narrows.

The Supreme Court’s jurisprudence on review can be understood through three controlling principles.

First, finality remains the governing presumption: a judgment should not be disturbed merely because another view is possible.

Second, review is corrective, not appellate: the applicant must demonstrate a recognised and material defect rather than merely challenge the merits of the decision.

Third, finality cannot legitimise a patent and material judicial error: where such a defect falls within a recognised ground of review, correction may be justified without converting review into appellate re-examination.

Review jurisdiction may, therefore, be understood as a controlled corrective jurisdiction in which finality remains the governing presumption. A judgment should be reopened only where a recognised and material error can be corrected without turning the review into an appeal.

The critical distinction, therefore, is between an erroneous judgment and a reviewable judgment. The former concerns correctness; the latter concerns whether the law permits finality to be displaced.

This understanding reconciles the apparent conflict between finality and correctness. The court does not review because it has changed its mind, but because the law recognises circumstances in which the integrity of the judicial process demands reconsideration.

The justification for review lies neither in distrust of judicial decisions nor in an unrestricted pursuit of correctness, but in recognising the fallibility of adjudication while preserving the authority of final judgments. Finality is essential because justice requires an end to litigation. Review is essential because justice also requires that an obvious judicial error should not become irreversible merely for having been judicially pronounced.

The constitutional balance is thus better expressed not as “finality versus correctness” but as “finality subject to controlled correction.” A mature system of judicial review must resist two opposite dangers: correction without finality - which breeds interminable litigation - and finality without correction, which converts judicial error into judicial authority. Review jurisdiction stands between these extremes: its purpose is not to give the litigant another opportunity, but to give the law one final opportunity to correct itself. That is, ultimately, the constitutional justification for review.

Bimlesh Kumar Singh is an Advocate-on-Record, Supreme Court of India.