INTRODUCTION
A simple question often creates confusion in arbitration law: when does arbitration actually begin? At first glance, Section 21 of the Arbitration and Conciliation Act 1996 seems to provide a clear answer. It states that arbitral proceedings commence when a request to refer the dispute is received by the respondent.
However, this does not fully resolve the issue. In many cases, the dispute itself arises much earlier. A breach of contract does not automatically lead to a dispute. A dispute comes into existence only when a claim is made and denied. This stage is commonly referred to as the crystallisation of the dispute.[1]
The difference between these two moments can have serious consequences. A party may formally commence arbitration within time but still fail because the claim had already become time-barred before the notice was sent. This makes it important to understand how Section 21 operates alongside the concept of crystallisation. This blog examines that distinction and explains why the real question is not just when arbitration begins, but when the dispute actually arises.
THE STATUTORY ANCHOR: SECTION 21 AND SECTION 43
Section 21 of the Act provides the primary rule for determining when an arbitration begins. It states that, unless the parties agree otherwise, arbitral proceedings regarding a specific dispute commence on the date the respondent receives a request to refer that dispute to arbitration. This provision serves as a procedural marker to eliminate uncertainty regarding dates.[2]
The significance of Section 21 is reinforced by Section 43 of the Act, which makes the Limitation Act 1963 applicable to arbitrations. Section 43(2) clarifies that for the purposes of the Limitation Act, an arbitration is deemed to have commenced on the date referred to in Section 21. [3]Consequently, the receipt of a notice under Section 21 stops the limitation clock for the substantive claims intended for arbitration. However, judicial interpretation has clarified that Section 21 is procedural rather than jurisdictional.[4] It does not create or validate the arbitration agreement but merely identifies the start of the formal process.
THE CONCEPT OF CRYSTALLISATION OF DISPUTES
While Section 21 identifies the formal start of the arbitration, it does not always coincide with the birth of the dispute. A primary distinction exists between a “claim” and a “dispute”. In Major (Retd) Inder Singh Rekhi v Delhi Development Authority, the Supreme Court observed that a dispute arises only when a claim is asserted by one party and denied by the other.[5]
A mere failure or delay in payment does not automatically create a dispute. There must be a positive element of denial. Therefore, a cause of action for the purpose of limitation arises when the right to claim accrues, which is typically when the “final bill” becomes due or when a clear denial of a right occurs. The period between the accrual of the cause of action and the issuance of a Section 21 notice is the critical window for determining whether a claim has become “stale” or “dead” under the Limitation Act.
THE “BREAKING POINT” DOCTRINE AND NEGOTIATIONS
In commercial relationships, parties often engage in prolonged negotiations before resorting to formal legal action. To address the impact of these negotiations on limitation periods, the Supreme Court introduced the “breaking point” doctrine in M/S Geo Miller & Co Pvt Ltd v Rajasthan Vidyut Utpadan Nigam Ltd.
The Court held that the period during which parties are engaged in bona fide negotiations toward an amicable settlement may be excluded when computing the limitation period for referring a dispute to arbitration. The “breaking point” is defined as the moment at which any reasonable party would have abandoned settlement efforts and contemplated a referral to arbitration. This moment marks the true crystallisation of the dispute.[6]
However, the threshold for this breaking point is lower in commercial disputes than in family settlements. In a commercial context, the primary interest is typically the recovery of payment, and the law expects parties to be vigilant. Parties cannot indefinitely postpone the accrual of a cause of action by merely sending reminders or interactions. Courts or tribunals must examine the negotiation history to determine if the “breaking point” occurred more than three years before the Section 21 notice was issued.
SECTION 11 AND THE THREE YEAR LIMITATION WINDOW
A common point of confusion in Indian arbitration practice is the difference between the limitation period for the substantive claim and the limitation period for filing an application under Section 11 to appoint an arbitrator. The Supreme Court in BSNL v Nortel Networks India (P) Ltd (2021) clarified that these are two distinct clocks.[7]
The limitation period for the substantive claim is governed by various articles of the Limitation Act depending on the nature of the breach. The right to file a Section 11 application, however, arises only after a Section 21 notice has been served and the respondent has failed to appoint an arbitrator within the statutory 30-day period. Because Section 11 does not prescribe a specific limitation period, courts apply the residual Article 137 of the Limitation Act, which allows for a three-year window from the date the “right to apply accrues”.
In BSNL v Nortel, the Court noted that this three-year period is “unduly long” and potentially defeats the Act’s objective of expeditious resolution. While the Act mandates that tribunals conclude proceedings within specific timeframes, the preliminary stage of seeking court intervention can currently drag on for years. The judiciary has suggested that Parliament amend Section 11 to prescribe a shorter, specific limitation period.[8]
THE EVOLUTION OF JUDICIAL SCRUTINY AT THE REFERRAL STAGE
The scope of judicial interference under Section 11 has been a subject of sustained debate. The prima facie standard was articulated in Mayavati Trading Private Limited v Pradyut Deb Burman, where the Court confined the enquiry largely to the existence of an arbitration agreement. This approach was subsequently developed in Vidya Drolia v Durga Trading Corporation, which recognised a limited power in courts to decline reference in rare cases involving manifest non-arbitrability or clearly time-barred claims. The objective was to prevent parties from being compelled into arbitration where the claims were, on their face, unsustainable.[9]
However, the Supreme Court has recently narrowed this scope of enquiry to favour “arbitral autonomy”. In SBI General Insurance Co. Ltd v Krish Spinning (2024), the Court clarified that the referral court under Section 11(6) should limit its enquiry to just two aspects:
- The prima facie existence of an arbitration agreement.
- Whether the Section 11 application itself was filed within the three-year limitation period from the date of refusal to appoint.
Intricate evidentiary enquiries into whether the substantive claims are time-barred, which involve determining the “breaking point” or “crystallisation”, are now the exclusive domain of the Arbitral Tribunal. The Court reasoned that since there is no right of appeal against an order passed under Section 11, a court rejecting a claim as time-barred at the threshold runs a serious risk of leaving a claimant remediless without a full hearing.[10]
ADMISSIBILITY VERSUS JURISDICTION: THE TVC TEST
To justify this shift, the judiciary has adopted the “Tribunal versus Claim” (TVC) test to distinguish between “jurisdiction” and “admissibility”. Jurisdiction refers to the power of the tribunal to hear a case, whereas admissibility refers to the nature of the claim, such as whether it is time-barred.
A challenge based on a statutory time bar is considered an admissibility issue because it attacks the claim itself rather than the tribunal’s authority. Under the doctrine of kompetenz-kompetenz, the tribunal is the preferred first authority to decide all questions of admissibility. This ensures that the arbitral process is not delayed at the threshold by preliminary objections that require a detailed examination of facts and evidence.
CONCLUSION
Rethinking Section 21 in light of the crystallisation of disputes reveals a complex dual system for commencing arbitration. Procedurally, arbitration commences under Section 21 upon the receipt of a notice for the purpose of stopping the limitation clock. Substantively, however, the validity of the claim depends on the crystallisation of the dispute, which occurs when a claim is denied or when negotiations reach a “breaking point”.
The current judicial trend, culminating in Krish Spinning, underscores a commitment to minimal judicial interference. By restricting the referral court’s role to the examination of the arbitration agreement’s existence and the timeliness of the Section 11 petition, the law empowers the Arbitral Tribunal to adjudicate the complexities of dispute crystallisation. For practitioners, this means that the burden of proving that a claim is “live” and not “deadwood” has shifted from the High Court to the Arbitral Tribunal, strengthening the principle of arbitral autonomy.
Author(s) Name: Shraddha Dongare (Maharashtra National Law University, Nagpur)
References:
[1] Major (Retd) Inder Singh Rekhi v Delhi Development Authority (1988) 2 SCC 338
[2] Arbitration and Conciliation Act 1996, s 21
[3] Ibid s 43
[4] Bharat Sanchar Nigam Ltd and Anr v Nortel Networks India (P) Ltd (2021) 5 SCC 738
[5] Major (Retd) Inder Singh Rekhi v Delhi Development Authority (1988) 2 SCC 338
[6] M/S Geo Miller & Co Pvt Ltd v Rajasthan Vidyut Utpadan Nigam Ltd (2020) 14 SCC 643
[7] Bharat Sanchar Nigam Ltd and Anr v Nortel Networks India (P) Ltd (2021) 5 SCC 738
[8] Anujay Shrivastava, ‘COURT’S REFUSAL TO REFER PARTIES TO ARBITRATION IN DISPUTES INVOLVING ‘TIME-BARRED’ CLAIMS: ANALYSING THE BSNL VS. NORTEL DECISION’ (2021) 10(2) International Journal of Legal Studies and Research <https://pure.jgu.edu.in/id/eprint/1233/1/IJSLR2021.pdf> accessed 05 May 2026
[9] Vidya Drolia and Ors v Durga Trading Corporation (2021) 2 SCC 1
[10] SBI General Insurance Co Ltd v Krish Spinning (2024) 12 SCC 1