Emotion and Legitimacy in Presidential Speech

• In Event: Open Refereed Roundtables
  In Refereed Roundtable (90 minute): Table 21: Political Behavior and Identity

Abstract

This study examines how historically shifting governance tasks shape emotional appeals and agenda structure in U.S. State of the Union (SOTU) addresses from 1934 to 2024. Using lexicon-based sentiment analysis and topic modeling, it analyzes institutional speech rhetoric, affective tone, and thematic organization across four historical regimes—World War II, the Cold War, the post–Cold War era, and the contemporary era of polarization. The purpose of this study is to identify whether the SOTU’s emotional register and narrative agenda are recalibrated when the presidency’s core task shifts from security mobilization to economic management and stabilization. To do so, the analysis examines how SOTU sentiment changes across years and presidents, compares topic patterns across historical periods, and assesses whether issue agendas become more concentrated or diversified over time. The results suggest that SOTU emotion is not merely an episodic reaction to events or a personal trait of presidents, but a patterned feature of institutional performance: affective tone remains broadly positive yet varies systematically across periods and administrations, while topic structure reorganizes by regime. In particular, the long-run trajectory shows a rebalancing from war/security-centered agendas toward economy/managerial governance agendas, with parallel shifts in the affective conditions under which legitimacy is made plausible. This matters because institutional political speeches are not simply policy reports, but performative vehicles that renew authority, organize national priorities, and manage collective feeling. By linking agenda transitions to changes in affective tone and agenda form, this study highlights emotional governance as a mechanism of political legitimacy in an institutional ritual setting.
Key words: State of the Union, affective tone, topic modeling, sentiment analysis, legitimacy

Author

• Xinshi Wang, Senshu University