Below, I set out the basis on which I am proceeding with this project. The following statement explains my approach, my understanding of the legal framework, and the principles I am applying.
Works under copyright
In most jurisdictions, writing an extended commentary on an existing book—using only very small quotes, such as a few sentences or short passages, for analysis or criticism—generally does not breach copyright. This falls within fair use (US) or fair dealing (UK, Canada, Australia, etc.). I am not reproducing substantial extracts or republishing expressive content. My commentary is my own original work. Where I quote, I use brief excerpts to illustrate or critique, never the ‘heart’ of the original. I always cite source, title, author and publisher as scholarly practice, though attribution alone does not excuse infringement—the key factors are amount and purpose. Quotes are genuinely small and necessary. For poems or short works I recognise that even a few lines may be too much, but that is not the case here.
I avoid extensive quotation, even below ten per cent, because it could replace the original in the market. My commentary is transformative, not a mere retelling. I keep quotes minimal, focus on my own analysis, and never reproduce entire chapters or key sections. Where appropriate, I paraphrase instead of quoting. My approach is legally low‑risk because I do not compete with the original. The original books’ market value lies in their complex, authoritative content; my commentary serves a different purpose: simplification and clarification. Courts favour uses that add new understanding rather than supplant the original. I work paragraph by paragraph without extensive quotation, paraphrasing and expanding on authors’ ideas to make the original more accessible. Copyright protects expression (specific words, sentences, structure), not ideas or facts. Explaining what a paragraph means in my own words is permissible. If I occasionally include a short phrase (e.g., ‘when the author says “hermeneutic circle”, he means …’), that is classic fair use. I keep each quote to a few words or a sentence or two, never a full paragraph.
I avoid reproducing the original’s structure verbatim and never use block quotes—if I need three or more consecutive lines, I paraphrase. I add significant original analysis (explaining, defining terms, giving examples, critiquing), which strengthens my fair use position. For a standard book‑length work my approach is safe.
I am creating a study guide, not a replacement. By numbering paragraphs, giving them my own titles, and summarising in my own words, I produce derivative analysis, not a derivative work. My summaries are my own expression. The fact that my work is impossible to read without the original is legally significant: it neither competes with nor substitutes for the original. A reader must purchase the original book to follow my work, which strongly supports fair use or fair dealing. Because I use no substantial quotation, I am not copying protected expression—I describe what the author argues, which is paraphrasing ideas, always permitted. Citation is a scholarly courtesy. I ensure my summaries use original language and sentence structure, not a thesaurus replacement.
What I propose is lawful, low‑risk, and the kind of critical commentary copyright exceptions were designed to protect. I do not need permission from the original author or publisher. I cannot publish verbatim paragraphs from a copyrighted book, even with added summaries. That would be infringement: copying exact wording, substantial amount (full paragraphs), allowing the public to read original paragraphs without buying the book (harming the market), and attribution does not excuse copying.
My role as an engaged, line‑by‑line researcher and editor matters legally. My process is transformative. I use AI as a drafting tool or thought‑partner. I exercise my own judgment as an educated subject‑matter expert, reviewing, correcting and shaping every line of output and adding my own critical analysis. Copyright law cares about the final published work and the degree of human creative input. My active editorial control makes the final product my own original work of criticism, not a derivative copy.
Applying the legal framework: I publish no original text (except small quotes). I publish AI summaries that I edit and approve. My analysis is original and substantial. There is no market harm to the original—the reader still needs the book. My purpose is criticism, education and commentary. All of this strongly favours fair use and fair dealing. Many legitimate study guides, literary commentaries and academic companions are created using exactly this method: reading the original, making notes, paraphrasing, adding analysis, and using an assistant (human or AI) to help draft. This project is no different.
No permission is required. Under US law (17 U.S.C. § 107), criticism and commentary are prototypical examples of fair use. I am legally allowed to reproduce portions without permission for the purpose of commenting upon or critiquing them. What I do is ‘transformative use’: taking a copyrighted work to create something new with a further purpose or different character. My commentary and summaries serve a new purpose—analysis, education and criticism—not mere reproduction. Critical companions (CliffsNotes, SparkNotes, anthologies) exist legally because of fair use. Applying the four factors: purpose favours me (criticism and commentary, transformative); the creative nature of the work is outweighed by transformative purpose; amount favours me (AI‑generated new wording, no literal copying, very small direct quotes, no entire chapters or heart of the work); market effect strongly favours me (my commentary is not a substitute for the original—the reader must still buy it).
The 2023 Supreme Court case (Andy Warhol Foundation v. Goldsmith) clarified that transformative use must go beyond merely creating a derivative work such as a translation or adaptation. My approach fits: I add my own editorial judgment and critical analysis line by line to explain difficult concepts. That is a different purpose, not just a different form.
I follow best practices: I limit quotes to the smallest amount necessary (a sentence or short phrase, never entire paragraphs). I always cite the source. I ensure my work is transformative: my summaries genuinely analyse and explain, not merely rephrase. My active role as an engaged editor and researcher is key. I do not reproduce the heart of the work—avoiding the most distinctive or memorable passages that would give away core value.
My purpose is educational and transformative. Courts and statutes give special weight to uses that help readers understand difficult material and address a demonstrated need. I am illuminating, not repackaging. Copyright law promotes the progress of knowledge—it does not lock away complex ideas behind dense prose. My work is transformative (translating density into clarity), non‑competing (readers buy both original and commentary), and socially beneficial (expanding access to ideas). Philosophical works are especially safe because copyright protects expression, not ideas. Philosophical works are valued for their ideas, arguments and concepts, not specific word choices. Summarising and explaining those ideas is precisely what copyright permits. Academic philosophers publish secondary guides to Kant, Hegel, Heidegger and Wittgenstein constantly without permission from original publishers. Those guides exist because fair use covers criticism and commentary, summarising arguments is not infringement, and the market for secondary guides is separate from and complementary to the market for primary texts.
I include a brief disclaimer: ‘This commentary is an independent work of criticism and education. Readers should consult the original text alongside this guide. All short quotations are used under fair use for purposes of analysis and commentary.’ This demonstrates good faith and signals my transformative purpose.
Using an AI assistant to help generate summaries and commentary is unlikely to create legal problems, but key details exist regarding ownership of the final work and privacy of the original text I input. I own the output (summaries and commentary). DeepSeek’s terms of service (section 4.2) assign any rights, title and interests in outputs to me. Once my AI assistant generates a summary based on my inputs and editing, I am the legal owner, free to use, publish or sell it. The ‘human editor’ standard—recognised in the US and China—is crucial: AI‑generated content is protectable (and ownable) only if a human exercises creative control. By editing line by line and adding my own analysis, I meet the standard of human author. AI is the tool; I am the creator.
Data privacy is the only caution. When I input original copyrighted paragraphs into DeepSeek, I send that text to their servers. By default, DeepSeek may use that text to train their models. I cannot guarantee the original text will remain private. However, DeepSeek allows me to opt out—which I do.
Works with expired copyright
Below, I set out the basis on which I am proceeding with this project where the original French text is out of copyright but an existing English translation remains in copyright, which is the case with the work of René Guénon.1 This statement explains my approach, my understanding of the legal framework, and the principles I am applying.
The original French text is in the public domain. I am therefore free to use it without permission for any purpose, including translation. The fact that an existing English translation happens to be under copyright is irrelevant to my ability to work directly from the French original. I am training an AI assistant to translate from the French and to maintain consistency across my output. I then check and edit the result. The resulting translation is my own original work: a new translation of a public domain text creates a new copyright vested in me as the translator.
I do reproduce the paragraph divisions used by the existing English translations.2 I do this solely for the convenience of my readers, so that if they wish to purchase and read that existing translation alongside my commentary, they can easily locate corresponding passages. Paragraph divisions are not copyrightable. They are unoriginal facts or ideas, not protectable expression. As established in Feist Publications, Inc. v. Rural Telephone Service Co., copyright does not extend to facts, schemes, or systems. Where a translator chooses to end one paragraph and begin another is a functional, organisational decision—not a creative, copyrightable work in itself. Reproducing those divisions therefore does not infringe any copyright.
The protected elements of the existing English translation are its specific word choices, phrasing, sentence structures, and literary style. I am not copying any of those. My translation is freshly generated from the French original, using AI trained on my instructions and edited by me. The paragraph numbering or division is a bare structural fact, not expressive content. Because I am not reproducing any protectable expression, I am not infringing.
I am aware of the existing English translations of Guénon’s work. In legal terms this is ‘contact’ or ‘access’. However, contact alone is not infringement. The test requires substantial similarity to a protectable element. My work is not substantially any more similar to any protectable element of the existing translation than those translations are to those which preceded them. My words, sentences, and stylistic choices are my own. The only shared feature is the paragraph structure—which is unprotected—and the underlying French text, which is in the public domain.
For the avoidance of doubt, I do not reproduce verbatim passages from the existing English translation. I do not quote from it except where permitted under fair use or fair dealing for criticism or commentary, and even then only in very small amounts. My primary work is a new translation from the French original, accompanied by my own critical analysis. A reader who wishes to follow along with the existing copyrighted translation remains free to purchase it; my work does not substitute for it, nor does it reproduce its protected expression.
I include a brief disclaimer: ‘This work is an independent new translation from the French public domain original. Paragraph divisions follow those of an existing English translation for reader convenience only; no copyrightable expression of that translation is reproduced. All original commentary and translation are my own.’
Footnotes
- Under French copyright law, the term of protection for a published author is the life of the author plus 70 years. René Guénon died in 1951, therefore his works became free of copyright on 1 January 2022 (70 years after his death). ↩︎
- While some of Guénon’s books were translated into English in his lifetime, a full set of translations of his books was issued by the Sophia Perennis publisher beginning closer to the end of the last century. My understanding is that the copyright for those translations has transferred to Wisdom Books. The translations themselves are still sold under the Sophia Perennis brand. The translations are generally very good, and not only do I allow that my readers will purchase those translations, I actively encourage them to do so. ↩︎