Dissent is a Symbol: D.Y. Chandrachud's Dissent in Romila Thapar v. Union of India on Bhima Koregaon and the Chilling Effect

On 28 September 2018 a three-judge bench of the Supreme Court refused by 2:1 to constitute an SIT into the Bhima Koregaon arrests. Justice D.Y. Chandrachud dissented. A reading of his 41-paragraph opinion on fair investigation, Article 32, and the boundary between dissent and incitement.

On 28 September 2018, a three-judge bench of the Supreme Court declined by 2:1 to constitute a Special Investigation Team to inquire into the arrests of five rights activists in connection with the Bhima Koregaon violence. The majority — Dipak Misra CJI and A.M. Khanwilkar J. — left the accused to the ordinary remedies available under the Code of Criminal Procedure: bail, transit remand challenge, quashing under section 482. 

Justice D.Y. Chandrachud dissented. He would have constituted an SIT independent of the Maharashtra Police, with the investigation monitored by the Supreme Court; pending such inquiry, he would have ordered the five accused be released on house arrest.

The line at which the dissent’s position crystallises is at paragraph 29 of the opinion. Dissent is a symbol of a vibrant democracy. Voices in opposition cannot be muzzled by persecuting those who take up unpopular causes.¹ The sentence travels well beyond the procedural prayer the petition had sought. It articulates a particular reading of the constitutional architecture under which the criminal process operates — a reading on which the integrity of investigation is not merely a matter of administrative law but a precondition for the political space that civil and democratic dissent occupies. 

The dissent’s engagement with the facts, with the precedents on Article 32 jurisdiction, and with the boundary between protected dissent and unprotected incitement, is organised around this proposition.

The case the three-judge bench was asked to decide

The petition was filed under Article 32 by five distinguished academics — the historian Romila Thapar, the feminist economist Devaki Jain, the Rhodes scholar economist Prabhat Pattnaik, the sociologist Satish Deshpande, and the human rights advocate Maja Daruwala. The prayer was for an “independent and comprehensive inquiry” into the arrests by Pune Police, on 28 August 2018, of five rights activists: the advocate Sudha Bharadwaj, the journalist Gautam Navlakha, the writers Vernon Gonsalves and Arun Ferreira, and the poet Varavara Rao. 

The arrests, carried out simultaneously in Delhi, Faridabad, Mumbai, Thane, and Hyderabad, followed an earlier round of arrests on 6 June 2018 of five other individuals (Surendra Gadling, Rona Wilson, Sudhir Dhawale, Shoma Sen, and Mahesh Raut). The provisions invoked were sections 13, 16, 17, 18, 18B, 20, 38, 39, and 40 of the Unlawful Activities (Prevention) Act, 1967, along with sections of the Indian Penal Code.

The factual background lay in the violence that broke out at Bhima Koregaon on 1 January 2018. The site has its own history: a victory pillar raised by the colonial government to commemorate the 1818 defeat of the Peshwa army at the hands of an East India Company force that included Dalits, tribals, Muslims, and Christians. 

The annual commemoration on 1 January has long been a Dalit rallying ground. On 31 December 2017, the Elgar Parishad event was held at Shaniwar wada in Pune, addressed by retired Supreme Court Justice P.B. Sawant, retired Bombay High Court Justice B.G. Kolse Patil, and Prakash Ambedkar of the Bharatiya Republican Party. 

The next day, violence broke out at Bhima Koregaon. Two FIRs followed. FIR 2/2018, lodged by Anita Sawale, named Sambhaji Bhide and Milind Ekbote, leaders of two Hindutva organisations, as perpetrators; Bhide was never arrested, Ekbote secured bail within a month. FIR 4/2018, lodged by Tushar Damgude, a self-professed Bhide follower, named six speakers at the Elgar Parishad event and led, in subsequent months, to the chain of arrests under the UAPA.

The petitioners’ case was that the August 2018 arrests were not a step in a bona fide investigation but a campaign of selective prosecution. The first preliminary objection by the Additional Solicitor General was that the petitioners had no locus to seek relief on behalf of others. The dissent rejected the objection at paragraph 5. 

The jurisdiction under Article 32, the dissent observed, is “wide enough to reach out to injustice in any form and originating in any source.” And the question receded into the background once the five arrested persons had themselves subscribed to the averments in the petition and sought transposition as petitioners. The locus objection lacked substance.²

The fairness of the investigation, and what the record showed

Paragraph 1 of the dissent sets out the framing principle. The Court’s role, the dissent observes, “involves particularly sensitive balances when the state seeks to curb freedom to investigate perceived breaches involving offences against the state.” 

Custodial interrogation balances diverse and often conflicting values: effective criminal justice administration, an impartial process of investigation, and the liberty and reputation of the individual. The key to the balance between the two lies in a fair, independent and impartial investigation of crime.

The principle does substantial work in the dissent because the petitioners’ case turned not on whether the activists were guilty of the offences alleged — a question the Court would not adjudicate in a writ proceeding — but on whether the investigative process had been conducted with the integrity that Articles 14 and 21 demand. The dissent’s engagement with the record proceeds along three lines.

The first concerns the conduct of the Pune Police in the days surrounding the arrests. On 29 August 2018, within hours of the Court issuing notice on the petition and ordering the five arrested persons placed under house arrest, the Joint Commissioner of Police Shivajirao Bodhke held a press conference declaring that the police had “more than sufficient evidence” against the five. The dissent describes this conduct, at paragraph 22, as “disconcerting.” 

The Joint Commissioner had “sought in this oblique manner to respond to the interim order of this Court by recourse to the electronic media.” On 31 August 2018, a press conference was addressed by Additional Director General Parambir Singh and senior officers; during the course of the press conference, letters “(many of which should form part of the case diary) were selectively flashed and read out.” On 1 September, ADG Singh appeared on the NDTV programme Truth v Hype and conceded that the letters he had read out were still undergoing forensic analysis. The dissent treats this conduct as fortifying “the need for an investigation which is fair.”³

The second concerns the documentary record. Counsel for the petitioners had advanced specific submissions: that thirteen alleged letters — said to have been recovered from electronic devices of accused arrested earlier — were unsigned, undated, lacked headers, and were never produced before the courts hearing transit remand applications; that seven of the thirteen were addressed to or signed by “Comrade Prakash,” a pseudonym which the Gadchiroli Sessions Court had already determined was used by Professor G.N. Saibaba (then in Nagpur Central Jail), with the consequence that letters dated after his March 2017 incarceration could not authentically be attributed to him; that one letter purportedly authored by Sudha Bharadwaj — a Faridabad-based advocate who is not a Marathi speaker — contained Devanagari script in forms peculiar to Marathi, suggesting fabrication. 

The dissent expressly declined to delve into the authenticity question on the merits at the writ stage; those, the dissent observed, are “matters for a fair investigation.” But the pattern of selective public disclosure of unverified material was, for the dissent, indicative.

The third concerns the procedural irregularities at the point of arrest. Section 41B of the Code of Criminal Procedure requires that the memo of arrest be attested by an independent witness — either a member of the family of the person arrested or a “respectable member of the locality” where the arrest is made. 

In the present case, the two panch witnesses were employees of the Pune Municipal Corporation; they travelled as part of the police team that carried out the arrests. The independent-witness requirement — itself reinforced by the Constitution Bench in D.K. Basu v. State of West Bengal as emanating from Articles 21 and 22(1) — was, on the dissent’s reading, not complied with.⁴

The Article 32 jurisdiction, and the cases that mark its limits

The doctrinal question on which the bench divided was whether Article 32 was the appropriate forum for the relief sought. The majority’s position rested on a line of authority — Rajiv Ranjan Singh ‘Lalan’ (VIII) v. Union of India, Gulzar Ahmed Azmi v. Union of India, Simranjit Singh Mann v. Union of India, Ashok Kumar Pandey v. State of West Bengal — holding that PIL is not maintainable at the behest of third parties for reliefs that are available under the Code of Criminal Procedure to those directly affected. 

The general rule, as the majority emphasised, is that bail, remand, quashing, and ancillary criminal-process reliefs belong to the competent court under the CrPC. The dissent did not contest the general rule. “As a court which is governed by precedent, we are bound by them and by a consistent line of authority,” the dissent observed at paragraph 14.

What the dissent located in the precedent chain was a separate line of authority on court-monitored investigations and SIT constitutions. Narmada Bai v. State of Gujarat had directed the CBI to take over an investigation conducted by the Gujarat police where the state’s charge-sheet, in this Court’s reading, could not be “said to have run in a proper direction.” Sanjiv Rajendra Bhatt v. Union of India had constituted a court-monitored investigation where the integrity of the state investigative agency was in question. Babubhai v. State of Gujarat had held that fair investigation, and not merely fair trial, is part of the constitutional guarantee under Articles 20 and 21 — a holding the dissent treats as foundational. 

State of West Bengal v. Committee for Protection of Democratic Rights had affirmed the state’s constitutional duty of fair and impartial investigation. Centre for PIL v. Union of India had ordered the 2G court-monitored investigation. Bharati Tamang v. Union of India had affirmed that, in exceptional circumstances, the Article 32 jurisdiction extends to constituting an SIT, transferring an investigation to a central agency, or even ordering a de novo investigation.⁵

The dissent’s point is that these two lines of authority are not in conflict. The first — the bar on third-party PIL for CrPC reliefs — governs the general case in which an accused has access to the ordinary criminal-process remedies and there is no question about the integrity of the investigative process those remedies presuppose. 

The second — the SIT and court-monitored investigation line — governs the exceptional case where the integrity of the investigation itself is in question, and where the ordinary remedies would therefore be inadequate because they operate within and depend on the investigative process. The dissent reads the petition before the Court as falling within the second category.

Dissent, incitement, and the boundary between them

Paragraph 29 is the rhetorical and doctrinal centre of the opinion. It opens with the assessment of the record: there is, the dissent observes, “a serious allegation that the arrests have been motivated by an attempt to quell dissent and to persecute five individuals who have pursued the cause of persons who have suffered discrimination and human rights violations.” 

The dissent acknowledges the countervailing imperative: “the Court must be mindful of the need not to thwart a criminal investigation leading to the detection of unlawful acts.” But the balance does not, on the dissent’s reading, fall in favour of accepting investigative conduct that displays the features the record displays. “Individuals who assert causes which may be unpopular to the echelons of power are yet entitled to the freedoms which are guaranteed by the Constitution.”

From here the dissent moves to its most-quoted formulation: that dissent is a symbol of a vibrant democracy, and that voices in opposition cannot be muzzled by persecuting those who take up unpopular causes. The line carries the dissent’s normative commitment. But the dissent does not leave the proposition at the level of assertion. It immediately marks the boundary at which the protection ends.

Where, however, the expression of dissent enters upon the prohibited field of an incitement to violence or the subversion of a democratically elected government by recourse to unlawful means, the dissent ceases to be a mere expression of opinion. Unlawful activities which violate the law have to be dealt with in accordance with it.D.Y. Chandrachud J., para 29

The distinction does substantial doctrinal work. It carves out a category — expression of dissent on causes that may be unpopular with those in power — within which the constitutional protections of Articles 19 and 21 operate at full strength, and beyond which they yield. The boundary is drawn at two specific places: incitement to violence, and subversion of a democratically elected government by recourse to unlawful means. 

The dissent does not extend the protection of dissent to either. What it preserves is the space between political opposition and unlawful action, and what it refuses is the conflation of the two.

The framing matters because the UAPA, under which the activists were charged, criminalises a broad range of activity associated with unlawful organisations. Sections 17 and 18 of the Act, dealing with raising funds for and conspiracy to commit a terrorist act, are themselves serious provisions. The dissent does not suggest that they cannot be invoked, or that their invocation against persons engaged in genuine unlawful activity is improper. 

What the dissent insists on is the integrity of the process by which the line between expression and unlawful action is drawn. Where that process is itself in question — where selective leaks, fabricated materials, and procedural irregularities accumulate in the record — the constitutional courts have a duty to ensure that the line is drawn by a credible investigator, not by media trial.

The investigative process and the wrongful arrest

Paragraph 37 of the dissent draws on a striking near-contemporary judgment. Just two weeks earlier, on 14 September 2018, the same Chief Justice who would form part of the majority in Romila Thapar had spoken for a three-judge bench awarding Rs 50 lakh in compensation to the space scientist S. Nambi Narayanan. Narayanan had been wrongfully implicated by the Kerala police in 1994 on charges of leaking ISRO secrets, subjected to custodial interrogation, and exonerated some years later by a CBI investigation. 

The Chief Justice’s judgment recorded that the dignity of a person is shocked when such treatment is meted out to him, and that compensation under the public law remedy was warranted for the wrongful imprisonment, malicious prosecution, humiliation, and defamation the appellant had suffered.⁶

The dissent observes that compensation in Narayanan’s case was ordered “nearly 24 years after the wrongful arrest — a grim reminder about how tenuous liberty can be and of the difficulty in correcting wrongs occasioned by unlawful arrest.” The point is not directed at the Chief Justice individually; it is directed at the constitutional implication of the temporal gap. The remedy of post-hoc compensation is, however generous, a remedy that operates after liberty has been lost for a span of years that no monetary award can return. 

The dissent’s argument is that the constitutional duty of the courts at the front end — to ensure investigative fairness when the apprehension of unfairness is credible — is precisely the duty that prevents the situation in which compensation later becomes the only available remedy.

From this premise the dissent reaches the disposition at paragraph 39. “This Court has a constitutional obligation, where its attention has been drawn, in a case such as the present, to a real likelihood of the derailment of a fair investigative process to issue appropriate directions under Article 142 of the Constitution.” The disposition at paragraphs 40 and 41 follows: an SIT to be appointed, the investigation monitored by the Court, monthly status reports, and the interim house-arrest order to continue for three weeks to allow the accused to apply for bail before the courts of competent jurisdiction.

What the dissent did not become, and what it preserves

The majority’s judgment was the order of the Court. The SIT was not constituted. The investigation continued under the Maharashtra Police until the National Investigation Agency assumed jurisdiction in 2020. The five activists named in the petition spent extended periods in custody. Varavara Rao was 79 at the time of his arrest; the dissent had noted his age and the absence of any conviction in twenty-five prior prosecutions. 

Stan Swamy, an 84-year-old Jesuit priest who had been working with Adivasi communities in Jharkhand, was added to the same chain of UAPA arrests in October 2020; he died in custody on 5 July 2021 while awaiting a ruling on his bail application. The case has, in the years since, become a touchstone in Indian writing on civil liberties and the operation of anti-terror law.

What Chandrachud J.’s position in Romila Thapar preserves — against the backdrop of what followed — is twofold. The first is an articulated constitutional theory of when the Article 32 jurisdiction reaches into ongoing criminal investigations. The theory holds that the general rule of CrPC remedies operates within a presupposition: that the investigation those remedies regulate is itself conducted with the integrity Articles 14 and 21 demand. 

Where the integrity is in question, the ordinary remedies are inadequate by their own terms, and the constitutional courts have a duty under Article 142 to issue corrective directions. The framework is offered for future application across the spectrum of cases in which selective prosecution, media trial, or fabrication is credibly alleged.

The second is the dissent-vs-incitement distinction at paragraph 29. The distinction operates as a structuring principle for free-speech and political-dissent jurisprudence in cases that arise under anti-terror statutes. It preserves the space for political opposition on unpopular causes — the dissent’s phrase is “unpopular to the echelons of power” — while making clear that the protection does not extend to incitement to violence or subversion by unlawful means. 

The framing has been cited in subsequent free-speech and UAPA litigation, in writing about the chilling effect of selective prosecution, and in arguments about bail and remand in cases involving rights activists, journalists, and academics.

The opinion closes with a sentence that has, in the years since, been quoted more than perhaps any other line in the dissent. If this Court were not to stand by the principles which we have formulated, we may witness a soulful requiem to liberty. The line is a warning addressed to the institutional future of the Court itself. The principles the dissent refers to — fair investigation as part of Article 21, the SIT and court-monitored investigation jurisdiction under Article 32, the dignity-of-the-individual line that runs from Joginder Kumar through D.K. Basu — were principles the Court had itself formulated over decades. 

The dissent’s argument is that the same Court must stand by them in cases where standing by them is institutionally difficult, not only in cases where the application is uncontroversial. Whether subsequent benches read the dissent as forecasting what they ought to do, or whether they read it as recording a position from which the Court turned away, will depend on how the doctrines it draws on are developed in the decade following 2018. The position the dissent records, in either reading, is available.