• the manuscript is a descriptive endeavor, leveraging the forma mentis approach to then elaborate about the intepretation of the observed semantic and emotional reconstructions. It is to note however, that the sample size is relatively small, at least in breadth (only 2 individuals analysed), and potentially in width (the total number of sentences analysed for each individual is not reported, but should be). This will surely affect the probability to observe certain (co-)occurences of words, and results in some statistical uncertainty that is for now hard to assess. An example of this uncertainty would be the emotional variation from speech to speech, with some of them potentially being situated in a context or platform promoting some emotions above others. I would find the analysis strengthened if the authors could evaluate the statistical significance of their observed outcomes, or at least comment upon this.

• More generally, the limit of the low sample size should be discussed in perspective, in particular the potential limit of representativeness with regards to the general discourse from each "side"

• Finally, I feel that more precise details should be given in the methods section as to the network reconstruction method, in addition to the reference provided, to help guide the reader. While Eq1 is useful to describe closeness centrality, it does not really give information on network computation itself. How are links defined and weighted? Are these number of co-occurences, or over-representations? Are weights used in the centrality computation? How is emotional richness computed exactly? When referring to reconstructions around a given word, what the authors mean are ego-centered networks, with ties between neighbors?

• abstract: "emotional patterns linked to a quick and pervasive spread of falsehoods" -> maybe "emotional patterns intended to trigger a quick and pervasive..." would be more precise

• abstract could end with a take home / summary sentence opening perspectives, such as the ones in the conclusion, e.g. "mindset reconstruction could be an important tool to deal with infodemic communication materials facilitating the climate divide "

• the networks are a bit hard to interpret because of their relatively density, and because of the fact that they might have a same word that can be laid out in different places. I wonder whether it would be possible to either use the same layout using all possible words from both cases, or to guide the eye with some arrows or boxes to highlight important results that are discussed in main.

• p5 l110  "from to key public speeches"  -> "from key..."

• p7 l152 Closeness centrality is defined as the inverse average distance between a word and all its neighbours (Metcalf & Casey, 2016). -> you mean and all other words in the full network, or in the ego-centered subnetwork?

• p19 l358 in vivo flagging *of* online...

General:

I really love the idea of reconstructing forma mentis/mindset networks from text. I think it’s super innovative and a really interesting and important deep dive into the communication strategies of key influencers. Great work. Nonetheless, I do think a few things could be better emphasised and/or made more clear. I describe these points below.

More specific comments:

• The abstract could offer more clarity as to what you do in the study and what you find. It seems some of the content you currently have there might be better suited to the introduction as it pertains more to the background of the study instead of what it does.
• Perhaps there could be more discussion of how and why you chose your key influencers, including alternative choices? You do this really well in your discussion of how and why you chose to you closeness centrality later in the presentation of the network analysis.
• You say you have 11 Thunberg speeches and 3 Monckton: is the total volume per influencer roughly equivalent? It seems you might have a lot more content to work with for Thunberg relative to Monckton. It might be good to add a bit more detail on how long the speeches are in total, how many words, if you undertake any pre-processing steps to prepare the texts for analysis (e.g. removing stopwords, stemming/lemmatizing,...), etc.
• Perhaps a personal preference but I would also like to see details on what tools you used to process the texts, build the networks, and then visualise them.
• In the method section, it might be good to clarify what you mean by ‘concepts’ in the networks. I assume you mean the emotions from the NRC Emotion Lexicon?  
• Throughout the paper, I would pick one of “climate infodemics” and “climate disinformation” and stick with it instead of switching between them. I might also consider referring specifically to Thunberg and Monckton instead of generalising to climate activism and climate infodemics/disinformation, respectively, as you are specifically analysing those two.  
• Related to text box 1, the information contained there could be easily translated to a table with three columns: one for the term, one for what it means for climate activism, and one for climate disinformation/infodemics. You could also consider a fourth column to contain the references for the explanation of the content you provide for each term. This format may make the content more digestible and less repetitive.
• The discussion and conclusion section currently seems to reiterate content from the introduction. I would consider emphasising your results more and their implications given the literature and contributions to future work.

Editorial:

• At the end of sentences, you commonly put quotations only around the word and not including the punctuation, e.g. “like this”. instead of “like this.” — something to consider revising.
• Line 82-83: I think there is a preposition missing where the X is:  “climate communication linked X the climate divide”
• Line 144: Should “network neighbourhood” be plural?
• Line 179: It seems “infodemics” should not be plural?
• Line 213: Same thing as in line 179
• Line 214: “activism” has an apostrophe; perhaps you meant climate activists’?
• Line 227: Should the “on” at the end of the line be “in” instead?
• Line 240: Should the “to” at the end of the line be “within” instead?
• Line 247: “resorts” should not be singular if you are referring to “climate infodemics” in the plural.
• Line 251: The “in” at the end of the line should be “on”
• Line 256: “concentrates” should not be singular if you are referring to “climate infodemics” in the plural.
• Line 257: The “to” at the end of the line be “within.”
• Line 259: “associates” should not be singular if you are referring to “climate infodemics” in the plural. 
• Line 263: “displays” should not be singular if you are referring to “climate infodemics” in the plural. 
• Line 268: Same thing for “projects”
• Line 272: Same thing for “appears”
• Line 292: “entwined to” —> “entwined with”
• You may want to provide a source for the claim beginning on line 298 and ending line 301
• Line 336: “alike” —> “like”