“On Liberty” Now Officially Has Two Authors
An edition of On Liberty published this month is the first to officially name Harriet Taylor Mill as a co-author alongside John Stuart Mill.
The new volume is edited by
Many know that John Stuart Mill said that his wife, Harriet Taylor Mill, “was the inspirer, and in part the author, of all that is best in my writings […]. Like all that I have written for many years, it [On Liberty] belongs as much to her as to me.” In his Autobiography, he says
With regard to the thoughts [expressed in the book], it is difficult to identify any particular part or element as being more hers than all the rest. The whole mode of thinking of which the book was the expression, was emphatically hers. But I also was so thoroughly imbued with it that the same thoughts naturally occurred to us both.
Yet when Mill published the book in 1859, he did so under his name only, and every version published over the years, until this one, listed only him as the work’s official author.
So why the change now?
In a 2022 article in Utilitas (on which part of the introduction to the new On Liberty is based), Schmidt-Petri, Schefczyk, and Osburg lay out one reason: “stylometric analyses” provide some strong (though not on their own decisive) evidence to “say with some degree of confidence that JSM did not write On Liberty all by himself and that HTM played a part in putting parts of the text into words.” And what are stylometric analyses? “Stylometry extracts the writing style of a person from his or her texts and then compares this ‘stylome’ to the stylome of texts the author of which is yet to be identified.”
Additionally, the authors of the Utilitas article point out some considerations that may have given Mill reason to claim sole authorship of the work that might not have such force today. For example, “the provision of misleading authorship information [might be justified] in bigoted and sexist Victorian times,” but today’s norms “would give no reason to deny HTM the unequivocal status of being ‘the co-author of this classic.’”
At the end of their article, Schmidt-Petri, Schefczyk, and Osburg say, “in the interest of historical accuracy and of giving credit where it is due, we suggest modern editions should list Harriet Taylor Mill as well as John Stuart Mill as authors of On Liberty.” A few years later, they were among the editors to implement that suggestion with the new version, making true their prediction that “it would be the first time that a central text in the history of political philosophy would change authors after its publication.”
The authorship of On Liberty was also discussed by the new edition’s lead editor, Piers Norris Turner, along with Helen McCabe (Nottingham) and Mark Philp (Warwick) on a recent episode of the BBC’s In Our Time.
The publication date of the new edition is listed as March 31st, 2026, but it appears to be already on sale.
Wonderful! When I had to prepare excerpts of On Liberty for my students a few years ago, I listed Harriett Taylor Mill as a co-author for just this reason. This was about a year before the Utilitas article came out, but of course Mill’s remarks in his autobiography have been well-known for a long time. Glad that this is finally getting traction. It would be interesting to know what else stylometry could reveal about authorship practices.
But there still is not decisive evidence that she wrote parts? I mean, my husband edits everything I write and gives me as much extremely beneficial feed back as any advisor, I will always acknowledge and thank them for their help – as they will mine – but that does not entail that either of them are my co-writers. It seems revisionist to change something the author(s) themselves did not make public.
When you acknowledge and thank your husband for your help, do you say that he was “in part the author of all that is best” in it? That “it belongs as much to him as to me”? That “the whole mode of thinking it expressed is emphatically his”? John Stuart made all of those claims public.
And? If they had intended there to be two authors they would have made that clear. It is not up to us to change their work, to change history.
The editors specifically address the claim that he would have made their co-authorship public, specifying several strategic and pragmatic reasons to suppress her authorship at the time and which all fail to hold today.
To name just two: First, she was dead by the time it went to press and so she could not personally endorse the final manuscript, even though no substantive changes had been made after she died. Second, sales and impact would likely be negatively impacted if a female co-author was listed.
It’s possible that the issue wasn’t even the gender of the co-author – how many books at that time were ever published with multiple attributed authors?
When I compare modern academic disciplines, I see that there are some with norms of few authors (like philosophy) and others with norms of many authors (like lab sciences) and there are many kinds of involvement with a paper that could get a person credited as an author in some contexts but just thanked in the footnotes in others.
The Communist Manifesto had two authors but I don’t know how well it sold.
It was published anonymously by the Communist Workers’ Educational Association. It did not sell appreciably for its first decades.
I think it’s worth comparing what we do for publications that were first published without an author listed at all, or published under a pseudonym. I think it’s standard for printings of The Federalist Papers, or of the works of Kierkegaard, to say the name of the author (though Kierkegaard’s pseudonyms are also listed for his works). I think Beatles albums still list “Lennon/McCartney” for the songs written by either one of them (partly because they are still within copyright, and this is the official copyright designation) but critical discussions tend to make clear which one of them wrote which songs.
I guess in large part it depends on whether the listing of an author is treated as part of the work itself (in which case you stick with the original) or as metadata that is important for modern readers (in which case you go with the most accurate designation).
The composers of the Qur’an attributed the authorship of their work to God, not to themselves. If we now investigate which humans are responsible for the text’s context and attribute authorship of the text to them instead of God, that’s not changing history. It’s doing history. It’s how the academic discipline of history works.
It is not for us to decide whether those expressions suffice to justify listing Harriet as a co-author. If one takes them to do so, that would be a leap beyond what those expressions themselves warrant.
I’m curious about “officially”; what is the relevantly authoritative office for such things?
But if we accept this reasoning, then what about his other books? And the books of others?
It’s worth looking into the technical analysis they did for this. That indicates that Mill’s other works show much less of HTM’s influence.
A confusing aspect of all this is that the editors’ own analysis indicates Mill’s remarks in the preface cannot be literally true (he says that “all that I have written for many years … belongs as much to her as to me” and says HTM is “in part the author of all that is best in my writings”). So their argument is: yes the preface is misleading and not to be trusted, but, nonetheless, the stylometric analysis suggests co-authorship of certain passages. The change to the authorship then rests on the stylometry, not on believing the preface. I don’t know what to make of this. A lot of confidence is being placed in the stylometry.
In fact, we take Mill’s own testimony very seriously and argue that the stylometry corroborates his claims about On Liberty. I encourage people to read the portion of our introduction that is about co-authorship here, available free from Hackett: https://hackettpublishing.com/on-liberty-with-related-writings
This seems straightforward to me: assuming stylometry is reliable, which I at present have little reason to doubt, we both should accept this reasoning, and then, if we learn that other important works (whether by Mill or others) are not properly credited to all authors, we should follow the practice of Turner et al. with respect those works, too, and extend proper credit to all authors. Whether anyone should go out of their way to discover facts about authorship is a separate question, and I can see a range of reasonable positions on this. In this case, however, not much hard work of discovery was required, over and above stylometric analysis, given J.S. Mill’s own public comments.
I had long assumed that she was an implicit co-author of The Subjection of Women, and several internet sources say so.
So what you’re saying is, I can finally stop correcting my students when they write “Mills” instead of “Mill”?
I had the same thought 🙂
The introduction includes a summary of the arguments that repeatedly employs “the Mills”. Mill, Mills, the Mills. Not sure we will solve the students’ problem.
I foresee requests for partial credit 🙂
This will make it worse. They’ll need to say something like “according to Mill and Mill” or “according to the Mills” but then also either “the Mills’ argument is that” or “Mill and Mill argue that” (but still NOT “Mills and Mills argue that”). Sadly, I see no end to our troubles.
Yes, that’s one important problem solved. But what do I do with my students who credit A Theory of Justice to Rawl?
Just rawl over and accept it, I guess.
Are you sure your students aren’t just speaking Dennettian (Dennettese?)?
From The Philosophical Lexicon, edited by Dennett.
This is very cool! I’m curious if it could change our interpretations of the texts. For instance, is there any discussion about whether the (arguable) tension between the Liberty Principle in On Liberty and Utilitarianism could be attributed at least in part to the fact that the two works don’t have exactly the same author(s)? (I skimmed the Utilitas article and didn’t see anything.)
People are raising good questions, and we *try* to address many of them slowly in the introduction, to make an overall case. Hackett has made the portion of the introduction where we address co-authorship available for free here: https://hackettpublishing.com/on-liberty-with-related-writings
Can someone clue me in on why there are six editors? It seems like a lot?
I ask because this gave me an stress-dream in which the journal had six editors who all needed to sign-off on the paper.
I assume it’s because those six people contributed in some way to the work of putting the book together.
Thank you for this question. It gives me a chance to point out that this volume is the work of a team that included a range of relevant expertise, all needed to feel confident about the co-authorship claim: multiple John Stuart Mill scholars, multiple Harriet Taylor Mill scholars, and people with detailed understanding of the stylometric analysis (sometimes individuals fit more than one camp). We challenged and checked each other repeatedly along the way.
The volume also provides extensive supplementary writings from both authors to help make sense of the main text of On Liberty. This is something not mentioned in the original post, but again this scholarly contribution of the volume benefited from teamwork.
The analysis is based on the controversial assumption that Principles of Political Economy was co-authored by JSM and HTM.* I.e., the Utilitas paper treats a match between the style of On Liberty and the style of Principles as evidence that On Liberty is co-authored, where one might equally view it as evidence of exactly the opposite.
A footnote in the paper says, “[i]n order to get a better sense of how strongly controversial labels influence our results, we removed the Principles from the corpus and reran our analyses with the same settings. Although we cannot give a detailed account here, the changes to the results were less significant than what might have been expected. The analyses (available on our website) are of comparable overall accuracy and still ascribe parts of chapter 3 to HTM.”
Does anyone have the link for the above website? It doesn’t seem to be the Utilitas website (which has the article’s supplementary material).
Interestingly, the footnote only says they reran the results without Principles in the corpus, not that they reran it with Principles labelled as JSM authored, which is perhaps still the standard view.
*Except for one chapter which is labelled as HTM authored.
Thank you for this question. The introduction to the new volume explains that the stylometric analysis from the 2022 paper has been updated to address this concern about the Principles. The team has run many different versions of the analysis. I know my co-editor colleagues Christoph Schmidt-Petri and Michael Schefczyk, both at KIT, would be happy to field specific questions about the stylometric analysis, and they have presented these updated results at conferences after the 2022 paper.
This is a legitimate concern and we also worried about the plausibility. We are currently working on a follow-up paper starting from other assumptions, it’s not really finished, though we do have results (apologies, it seems the link is down, we’ll fix that asap, it would be on the KIT website). The stylometric evidence is not strictly speaking compelling anyway, given the number of assumptions and decisions required (described in detail in the paper, it’s open access on the Utilitas page). It is another piece in a puzzle, but it’s a new piece, and it’s testable empirical evidence.
Great!
But, peevishly, why is she referred to as Harriet Taylor Mill? Having just disappeared down a rabbit hole of looking this up, she wasn’t referred to thus during her lifetime, and on her gravestone of which I just found an image her name is given as Harriet Mill. And she was born Harriet Hardy. It makes her sound like Alexis Carrington Colby!
I’m curious why someone is only doing this now? It’s my understanding that analyses of this sort have been around for a long time. (I remember a prof of mine mentioning them in connection with Platonic and Aristotelian works where authorship was dubious or debated in the late 90’s.)Has stylometry gotten that much better? Is there then a potential to settle a lot more disputed authorship questions?
Also when is someone going to do this with the Freud book attributed to Rieff that pretty much everyone thinks is actually mostly Sontag’s work? I’d love to see how that comes back.
Actually, I tried to get a co-authored edition published in 2000, after publishing the Complete Works and biography of HTM, based on the historical evidence. No publisher was interested. Adding the stylometric evidence to the historical made the difference–26 years later.
How to generate interest in an edition that reprints a widely available text? Hmmmm…. let’s see … what could one do to grab some clicks ….
When I taught On Liberty I would bring slips of paper with “Harriet Taylor Mill” printed on them with some glue and have my students paste it on the cover while telling them why only John Stuart Mill’s name was on it. My female students would have the biggest smiles on their faces as we did this (and to be fair, a few of the men).
I have a review of this impressive new edition forthcoming in Ethics. A typescript is available here: https://www.drdaleemiller.net/papers.