8/ In other words — this isn’t just analysis of language; it’s analysis of the run-up to transition decisions. Understanding discourse is understanding transition timing.

Read the paper #openaccess here:

doi.org/10.1016/j.er...

\end

7/ Why this matters beyond South Africa: we see discourse as part of policy sequencing. Before policies shift, narratives shift. Discourses shape what becomes politically thinkable, when coalitions emerge, and how “windows” for phase-out open (or close).

6/ Media space is dominated by Eskom, political elites, and major coal firms. Civil society, labour, and international actors appear far less. Sentiment toward mining and “transition politics” is generally positive — reinforcing continuity over disruption.

5/ Renewables are framed largely as grid relief, not replacement for coal. Carbon-management technologies (CCS, hydrogen, offsets, “clean coal”) often appear as future-fix narratives that sustain incumbency. Explicit coal-exit discourse is rare — except briefly around the #JETP announcement.

4/ The core insight: coal remains discursively entrenched. Coverage centres on mining, politics, and the energy crisis — not coal phase-out. “Transition” language rises, but is mostly tied to crisis response and technology narratives, rather than structural change.

3/ We combine dynamic topic modelling, named-entity recognition, and sentence-level sentiment analysis, supported by careful manual reading and validation. This lets us track not just what is being said over time, but who says it and in what tone - without losing interpretive depth.

2/ We analyse ~8,000 national news articles (2010–2024) to understand how South Africa’s #coal transition is narrated — and what that reveals about #power, #politics, and pathways for change.