Abstract
This thesis investigates how teachers in South Africa's low-quintile schools exercise agency as leadership to foster innovative learning environments (ILEs) that develop 21st-century skills despite systemic inequalities and resource constraints. Using a quantitative cross-sectional research design, the study employed a census sampling strategy with data collected from 74 foundation phase teachers (74 completed questionnaires, 83% response rate) across 8 Quintile 1–3 schools in the Waterberg district of Limpopo Province participating in the National Education Collaboration Trust's (NECT) Sandbox Project. Data were analysed using Partial Least Squares Structural Equation Modelling (PLS-SEM) to examine relationships between temporal dimensions of teacher agency (iterational, practical-evaluative, and projective), distributed leadership practices, and innovative learning environment characteristics. The study integrated three theoretical frameworks: Priestley et al.'s ecological model of teacher agency grounded in Emirbayer and Mische's temporal-relational theory, distributed leadership theory reconceptualised through Ubuntu philosophy and African communal praxis, and the OECD's ILE framework adapted for resource-scarce African contexts. Findings revealed that projective agency (future-oriented aspirations) significantly predicted ILE practices (β = 0.452, p < 0.001, f² = 0.206), while both practical-evaluative agency (β = 0.396, p < 0.001, f² = 0.171) and projective agency (β = 0.381, p = 0.001, f² = 0.148) significantly predicted distributed leadership practices. Teacher leadership demonstrated a direct positive effect on ILEs (β = 0.343, p = 0.009, f² = 0.130). Critically, leadership fully mediated the relationship between practical-evaluative agency and innovation (indirect effect β = 0.136, p = 0.017), indicating that teachers' ability to navigate daily contextual constraints translates into innovation exclusively through the leadership practices such negotiations engender. The iterational dimension (past experiences) showed no significant direct effects, repositioning past experiences as latent resources selectively activated by present challenges and future goals rather than determinants of innovation. The conceptual model explained 52.8% of variance in leadership and 57.6% of variance in ILEs, demonstrating substantial explanatory power. These results led to the development of the Temporal Agency as Leadership for Innovation (TALI) Model, which represents the study's primary...