Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The article aims to investigate the textual outputs of Large Language Models as „synthetic storytellers” when they are prompted to tell a traditional folktale. Its main contribution is demonstrating a „method for studying the structures and cultural bias of AI generated texts” – by comparing the generated tales to traditional folklore variants. The methodology is clearly described, while the detailed description of the plot, structure, and key motifs of the chosen folktale („The Sweetheart in the Forest”), as well as the cultural background of the informant and the presumed audience, adds depth to the comparison. As the author states that this is a „preliminary study in a planned series of articles”, it will be interesting to see how this method is further developed and if general principles can be formulated based on the results of further analysis on a larger sample of folktales.

My only comment is that the following references in the main text and endnotes are missing from the list of references: Iser 1978; Foley 1991; Pilipets 2025; de Seta, Pohjonen, Knuutila 2024; Alver 1980.

Author Response

Thank you for your generous reading and for pointing out the missing references. They have now been added. The reference "Alver" is corrected, as it should be Gravem. 

Reviewer 2 Report

Comments and Suggestions for Authors

I was sceptical at first, but this is a sound and often insightful article; the abstract and some of the introduction do not do the argument justice.

 

I think the article would benefit too from amplification of the methodology; in the ‘LLMs telling tales’ section especially I’d like to see more justification of the approaches, and this will help with the overall justification: by the end the question “so what?” still arises, even though it should not be difficult to add further reflections on the wider significance of the findings.

 

The comparison with the ephemerality of performance is fascinating and might be extended in relation to the endless variation generated by LLMs. What are the implications for the identification of the “absurd” and the “uncanny” that LLMs here produce?

 

The interpretations are very good but there is also quite a lot of speculation. Often this is justified, working between the known and the unknown (e.g. on OpenAI’s educational strategy); but there might also be more specificity and referencing at times, which I indicate indicatively below. Might there not be more reflection on language difference, and on the use of Norwegian especially? Could links between “floating motifs” in traditional tales and those produced by LLMs not be reflected upon?

 

Specific issues:

 

Abstract: “which means they need to be analyzed using literary scholarship”: this does beg some questions: what evidence is there that most people likely read only the texts of LLMs? And more is needed on why these traditional analytical tools are needed.

23: introduce the classification system fully?

28-30: as above, evidence needed and fuller justification.

35: “studying the structures and cultural bias of AI generated text”: this is a wholly justifiable thesis. Makes sure the whole set-up is justified.

67: no capital for “firstly”

203: “at all times been an issue”: this sounds a little naively generalised. Make it stronger? And more specific as well, with data, could be powerful, depending on the historicity required by the argument (I would recommend more).

253: there is a missing full stop.

280: define what “corpus” signifies here.

359: “integrate” not “integrates”

389: this is compelling but resembles speculation; are there any sources, e.g. for the tradition of night-courting?

441: “are” not “is”

444: “focus” not “focuses”

Author Response

Thank you for your thorough reading and helpful comments. Please find my response below and the edited article attached:

23: Explanatory endnote added

202-203: I think the point still stands, but I changed the phrasing to make my point more precise.

389: A reference on the custom of night-courting was provided in line 194.

Spelling errors have been corrected and some phrasing has been changed for clarity. I have started the process of  making the generated texts available as a dataset, and as a result of this I changed the numbering of the generated texts. These have all been corrected in the text to correspond with the dataset, which will be available shortly.

Reviewer 3 Report

Comments and Suggestions for Authors

The article takes on a necessary subject - LLMs and the folktale tradition. It is a novel contribution that should spur further research (a project that the author notes is in the works). Some light copy-editing can be done to improve the already excellent text.

Comments for author File: Comments.pdf

Author Response

Thank you for your generous reading. Please find the revised manuscript below.