D11 - Judicial Decision-Making
Date: Jun 13 | Time: 10:15am to 11:45am | Location: 680 Sherbrooke St. West 491
Chair/Président/Présidente : Emmett Macfarlane (University of Waterloo)
Discussant/Commentateur/Commentatrice : Emmett Macfarlane (University of Waterloo)
Governing through Intervention? Government Mobilization in the Supreme Court of Canada: Brendan Dell (Queen's University), Danielle McNabb (Brock University)
Abstract: Studies of the American and Canadian Supreme Courts illustrate that third-party interveners, also called “friends of the court,” have a statistically significant impact on judicial decision-making (Hazleton and Hinkle, 2022; Alarie and Green, 2010). Specifically, there is a rich body of empirical evidence to suggest that interveners influence the content of judicial decisions, with respect to legal doctrine, case outcomes, and ideology. At the same time, research shows that in the Canadian context, governments are overwhelmingly the biggest “repeat player” interveners (McNabb, 2023). This context matters because, as Marc Galanter’s instructive “haves and have nots” theory (1974) suggests, the more frequently an actor participates in legal cases, the more likely they are to influence judicial decision-making. Put simply, government interveners are likely to have an institutionalized advantage, whereby the Court is more receptive to their interventions compared to other types of actors. There is a scarcity of literature that explores the nature and impact of government interventions (especially provincial governments), and even more, the scholarship that does exist only provides coverage on Charter cases up to 2007 (Radmilovic, 2013). To address this important gap, this paper develops an updated empirical account of government interventions in all constitutional cases (including federalism and Aboriginal and treaty rights cases) at the Supreme Court of Canada during the past decade. In doing so, this paper investigates the success rates of government interventions and the extent to which governments intervene as a unified front in the three areas of constitutional law. Ultimately, this work makes important contributions to the legal mobilization and government strategic behaviour literatures.
Data-driven Decisions? Empirical Evidence, Interveners, and the Supreme Court of Canada: Danielle McNabb (Brock University), Kate Puddister (University of Guelph)
Abstract: The Supreme Court of Canada (SCC) is frequently called upon to make determinations on contentious issues, with important consequences for politics, law and public policy. Considering the high stakes and possible widespread consequences that can result from its decisions, the SCC allows third-parties or interveners, who are not direct parties to a case, to participate and inform the Court’s decision. Interveners, which include governments, business and civil society groups, have become a permanent fixture in the Court, with hundreds of these actors participating annually, and in more than half of all SCC cases. Given this, it is worth investigating the nature of intervener participation, and more specifically the types and sources of information interveners provide to the Court. Do interveners try to persuade the Court to stray from strictly legal sources in reaching its decision? What types of information do interveners provide to the SCC in their attempt to influence and inform the Court? In this paper, we systematically examine the content of written arguments submitted by interveners in the 103 Charter cases heard by the SCC during the past decade. We assess the types of materials, both legal and extralegal, that interveners provide to the Court and in particular, we analyze the use of social science and empirical data by interveners in their arguments. We consider the implications of using empirical evidence by judges, including how empirical data is presented to judges. We situate our findings into wider debates regarding the role of interveners in democratizing judicial decision-making.
THE INSTITUTIONAL IMPACT OF LANGUAGE ON CASE OUTCOMES AT THE SUPREME COURT OF CANADA: Jean-Christophe Bédard-Rubin (University of Toronto), Tiago Rubin (Collège Bois-de-Boulogne), David Côté (University of Toronto)
Abstract: The Supreme Court of Canada (SCC) is an officially bilingual institution. However, litigants only obtained the right to be heard by a panel of judges who could understand them directly in English and/or French without simultaneous interpretation in 2022 with the reform of the Official Languages Act. For most of the Court’s history, only a handful of judges, generally from Quebec, could hear cases argued in French. The linguistic makeup of the judiciary started to change in the 1970s when new procedural language rights were added to the Criminal Code. Over time, more judges fluent in both English and French were appointed to lower courts and judicial language training programs provided new opportunities for judges to perfect their language skills. Consequently, judges from lower courts elevated to the Supreme Court of Canada have been increasingly fluent in both languages, though very few have been perfectly bilingual. There remain linguistic differences between the judges and these differences continue to have an impact on judicial behavior. Though few qualitative studies have explored the actual impact of language on the Court and on judges individually, anecdotal evidence suggests that language plays an important role in how the Court organizes its work as an institution. Previous quantitative analysis has shown that there are several statistically significant correlations between language and institutional dynamics at the SCC; cases argued in French are heard by smaller panels, unilingual judges are less likely to hear cases argued in French, and, even when they do, they are less assertive than in English cases. Thus, while the Court as whole has become an increasingly bilingual institution, the persistence of observable behavioral differences suggests that even small variations in language proficiency might perpetuate the linguistic division of labour at an institutional level. Taking these empirical findings as its starting point, this article assesses the institutional impact of language on case outcomes. It does so in two ways. First, it considers how the effect of language on panel size and composition affects the ideological composition of panels in cases argued in French and English. It does so by observing how the ideologically median judge on the panel varies depending on the language of the case. Second, it considers how the effect of language on assertiveness affects the degree of collegiality of judges in cases argued in both languages. It does so by looking at how the language of the case affects how far judges deviate from their preferred ideological position as expressed in their written opinions when they join their colleagues in cases argued in either language. While the story needs to be further clarified, the paper shows that this behavioral model better captures the institutional rather than the individual impact of language on case outcomes.
Applying LLMs at the SCC: A Comparison of Three Approaches: Andrew McDougall (University of Toronto Scarborough), Andreea Musulan (University of Toronto), Robert Schertzer (University of Toronto Scarborough)
Abstract: Using a corpus of 3,400 Supreme Court of Canada (SCC) case summaries from 1980 to 2021, this study compares three approaches for determining the outcome of decisions: human judgement, traditional machine learning, and large language models (LLMs), specifically ChatGPT. Through this work, we will produce a publicly available dataset of the 3,400 SCC cases with all three sets of annotations for the outcome of appeals. This dataset will also contain metadata on: judge dissent from majority decisions for each case; biographic characteristics of the judges themselves, such as where they studied law and their hometown; and links to the pages on the SCC website that contain other data previously parsed by Lexum, such as subjects pertaining to each case. The goal of this project is to set benchmarks for confidence in applications of LLMs to legal documents, inspiring other scholars to explore their use further. By increasing confidence in the application of LLMs to legal research, we hope to reduce technical barriers for qualitative scholars interested in asking questions that require big data.