1. The duplicated text on Page 1 & 2 and the path on Page 8: This one is probably just a copy-paste artifact, but on Page 1 at the bottom and Page 2 at the top, the same long string of evolutionary path numbers is repeated. More importantly, on Page 8, you describe the path as “1-1, 1-2, 1-3, 1-5, 1-6, 1-9, 1-10- 2-1, 2-4… 5-8, 5-9”. But then in Figure 4 (the Sankey diagram) , I’m trying to trace that exact sequence, and it doesn’t seem to match the visual flow. The lines in the diagram look much more interconnected than that single string suggests. How did you derive that specific textual path from the network of lines in the figure? It feels like there’s a disconnect between the quantitative data and the narrative you’re building from it.
2. You mention choosing a Rao-Stirling value > 0.7 to select the “high interdisciplinary literature,” citing a distribution graph (Figure 2) and a reference (49). Looking at the histogram, the distribution seems pretty continuous, with no obvious natural break at 0.7. Was this cut-off arbitrary? The choice of this threshold directly determines the entire dataset for your topic modeling (18,573 articles). A different threshold would have given you a completely different set of “interdisciplinary” papers. Without a strong justification, this feels like a major methodological weak point that undermines the foundation of the analysis.