Contested Meaning in Justin Trudeau’s Official Speeches, 2015–2024: A Hermeneutic Text Mining Approach

AHL3900 project description

Research Project and Objectives

The ten years during which Justin Trudeau was prime minister of Canada were tumultuous. At first, the Trudeau government worked to distinguish itself as progressive, in contrast to the United States during Donald Trump’s first presidency. Then it faced crises related to Covid-19, to the discovery of gravesites at residential schools, to the Freedom Convoy in Ottawa, and finally to the return of Trump as the U.S. president. Each of these events had an impact on Canadians’ sense of identity in frequently polarizing ways.  

In this project, I propose to describe how this impact changed over time by focusing on Trudeau’s o icial speeches as sites of negotiation over meaning. He gave 326 speeches between 2015 and 2024, as archived on the Prime Minister’s website (www.pm.gc.ca), reflecting the ways he was both responding to Canadians and trying to persuade them. However, the corpus of speeches is so large – nearly half a million words – that analyzing it using conventional approaches such as close reading or discourse analysis would be a herculean task. Thus, I am proposing to use text mining tools to overcome the problem of scale.  

My objectives are twofold. First, I will describe the forces shaping Canadians’ sense of themselves in relation to those they perceived as “other.” Words such as feminist, reconciliation, mandate, protest, and even freedom became focal points for controversy, metonyms standing in for larger debates about policy and identity. I will link changing patterns of word use to changing patterns of meaning and, by extension, Canadians’ contested collective self-perception. Second, I will contribute to theories of meaning, using the corpus of Trudeau’s speeches to address the challenge and promise of large corpora for observing change in meaning over time. I will build on alreadyexisting tools that use statistical analysis to identify relationships between words in a corpus. I will show how these tools complement hermeneutics, linking patterns of use to patterns of interpretation. The implications of my analysis of Trudeau’s speeches will reach beyond Canada’s recent history: my account of meaning will provide insight into situations in other contexts (such as journalism) where words acquire an evolving ideological charge.  

This project builds on a pending SSHRC Insight Development Grant, submitted in February 2026. If the grant is funded, the student(s) participating in the project proposed here will join me and a PhDlevel researcher. (I will go through with the project even if it is not funded, albeit on a smaller scale.) I have a strong track record in mentoring undergraduate students through the research process. Similar projects in Winter 2025, for instance, led to publications with undergraduate co-authors (DOI 10.1080/0907676X.2025.2590066, published in 2025, and DOI 10.20381/9082-d564, currently under review). This project will lead to a similar outcome: I will be editing a collective work on the evolution of keywords in Trudeau’s speeches, and participating student(s) will co-author a chapter tracing the evolution of one keyword over the course of Trudeau’s mandate.  

Research Approaches and Methods

This project combines text mining and hermeneutics. Text mining is a form of statistical description allowing for the extraction of “implicit knowledge from textual data,” such as speeches, produced without additional computer markup (T. Jo, Text Mining, DOI 10.1007/978-3-031-75976-5). We will be using techniques such as co-occurrence K. Conway, AHL3900 proposal, submitted April 2026, p. 2 analysis (mapping pairs of words observed together in a corpus), correlation analysis (mapping pairs of words based on the likelihood that they will appear together), and correspondence analysis (comparing the expected distribution of words to their observed distribution). We will be using the statistical programming language R for our analysis, in particular the tidytext package (J. Silge and D. Robinson, DOI 10.21105/joss.00037).  

Hermeneutics is a method for describing the relationship between the objective and subjective (or interpretive) dimensions of a text or corpus. As Paul Ricoeur writes (Hermeneutics and the Human Sciences, DOI 10.1017/CBO9781316534984), it consists in making and validating “guesses,” or hypotheses about meaning that can be tested against what a text says. We validate guesses by assessing their degree of congruence: to choose between two interpretations is to identify which “takes account of the greatest number of facts furnished by the text [... and] o ers a qualitatively better convergence between the features which it takes into account” (Ricoeur, p. 137). We also assess guesses’ degree of plenitude, understood as the degree to which “all the connotations which are suitable [are] attributed” (Ricoeur, p. 138), by situating speeches within the social and political context where they were delivered.  

Skills that students will acquire

Students will gain a familiarity with quantitative (text mining) and qualitative (hermeneutic) tools, with an emphasis on the ways that their synthesis can contribute to the solution of open-ended problems. They will gain experience programming in R and interpreting the results. They will also be able to describe, using concrete examples, how meaning and word use evolve over time. With respect to the content of our analysis, students will understand Canada’s recent political context in greater historical depth. Finally, with respect to knowledge creation, students will gain first-hand experience of the publication process, from writing to peer-review and revision. They will also have the chance to participate in the steps (such as copyediting and proofreading) that take place in order for a book to appear in print.  

Although this project is anchored in the field of communication, the skills students learn will also have applications in other fields, related to analytical synthesis (students will interpret disparate forms of evidence), critical interpretation of data (students will explore how methodological choices shape what data reveal or do not reveal), and creative problem-solving (students will learn strategies for answering questions for which there are no predetermined methodological frameworks). These skills will prepare students for the types of jobs that cannot be automated because they require creative, contextual thinking.  

Breakdown of the 90 hours of student activities

Background reading (theory and method of text mining, hermeneutics, political context of contested meaning), 20 hours; instruction in specific skills (text mining, hermeneutics), 15 hours; analysis, 15 hours; preparation of book chapter, 20 hours; revision of book chapter after peer review, 10 hours; weekly meetings, 10 hours  

Preferred semester: Winter 2027  

Note sur le bilinguisme : Bien que cette proposition soit rédigée en anglais, le projet peut se faire également en français. Les logiciels que nous emploierons fonctionnent dans les deux langues, et je suis en train de compiler des corpus parallèles (en anglais et français) des discours de Justin Trudeau. J’inviterai donc les étudiant·es à postuler dans la langue de leur choix.