RECONSTRUCTING NATIONAL AND PAN-AFRICAN IDENTITY THROUGH TRANSITIVITY SHIFTS IN HUMAN AND AI TRANSLATIONS OF NIGER’S NATIONAL ANTHEM L’HONNEUR DE LA PATRIE

This paper investigates how transitivity shifts reconstruct national and Pan-African identity across L’Honneur de la Patrie and its four English translations: a widely circulated human translation and outputs generated by ChatGPT, Google Translate, and DeepL. Drawing on Systemic Functional Linguistics (henceforth, SFL), particularly the transitivity system (Eggins, 2004; Fontaine, 2013; Thompson, 2014; Halliday & Matthiessen, 2014), and adopting a comparative descriptive mixed-methods research design, the study compares the distribution of process-types, participant roles, and circumstantial elements to examine how experiential meaning is reconfigured across translation modes. The findings reveal systematic transitivity shifts in all the four translations. Compared with the French source text, the translations generally increase material processes while reducing relational processes, thereby foregrounding collective agency over symbolic identification. AI-generated translations further introduce behavioral processes absent from the source text and display distinctive redistributions of participant roles and circumstantial elements, resulting in different patterns of experiential meaning and ideological reconstruction. Although the anthem’s overarching vision of national unity, patriotism, and Pan-African solidarity is largely preserved, the four English translations differ in how they construe agency, identity, and contextual framing. The study demonstrates that transitivity shifts constitute a key linguistic mechanism through which both human and AI translation systems reconstruct experiential meaning and ideological representation. It contributes to SFL-informed Translation Studies by extending transitivity analysis to the underexplored genre of national anthems and offers new insights into AI-mediated translation of ideologically sensitive texts.