This post is part of our Research in Reflection Spotlight Series, which will run through August 2025. After a decade of active and intense research and data creation, the WPHP is coming to a close this summer. To celebrate, this series shares reflections on the WPHP by our Research Assistants, Project Manager, past Project Manager turned Lead Editor, and Project Director, featuring memories, research, lessons learned, and ultimately the impact of the WPHP on the team who produced it.
Authored by: Michelle Levy and Kandice Sharren
Edited by: Kate Moffatt
Submitted on: 09/02/2025
Citation: Levy, Michelle and Kandice Sharren. "Reflections of a Project Director and a Lead Editor: Last Words." Women’s Print History Project, 2 September 2025, https://womensprinthistoryproject.com/blog/post/152.
Kandice Sharren and Michelle Levy at Chawton House Library, 2015.
Michelle: Well, Kandice, it is hard to believe we have come to the end of the WPHP line. It’s been a long road, mostly smooth, even though we had the pandemic and postdocs to navigate. I thought a dialogue would be a good opportunity for some final thoughts. What have we learned? What do we wish we had done differently? What are we most proud of?
I can start with regrets (and a few sorry-not-sorrys). I do regret not having built into our data model a way of tracking who worked on each record, and for how long. It would have been great to have that information – and to have seen the collaborative work that has gone into every record. As you know, throughout the project I shut down anyone who described our work as “data entry.” Having this feature would have made this fact palpably clear.
One non-regret: I do not for a minute regret including reprints, even though it increased our workload, not sure by how much, probably at least five-fold. But from early days, such as finding the number of editions of Maria Rundell’s Domestic Cookery, to yours and Kate’s discovery about the number of Evelina editions, most of which were unaccounted for, one of the most important lessons for me has been the critical role reprints played in the history of women’s writing, and to the period’s publishing and print history more generally.
Over to you—regrets?
Kandice: I’ve had a few! Mostly because I didn’t really know what I was doing at the beginning of the project, and I’ve had to live with my questionable choices ever since.
Like you, I don’t regret including reprints at all! It’s fascinating to see which books had staying power, as well as the different formats books could take over time. That said, I do have some regrets about how we recorded them—it would be great if all of the different editions of a single work were linked in some way. One of our programmers, Michael Joyce, played with a way to link related titles quite a few years ago, but it isn’t comprehensive and doesn’t recognize changes in titles. Same thing with firms, actually. Early on, we decided to create new records every time a firm added or lost a partner, or changed addresses. While that is important data, it does mean it’s harder to get a sense of the longer trends of a business. I would love to be able to see all of the titles that the various iterations of Longmans, for instance, published in one place, rather than going through several different records.
Which I guess feeds naturally into your question about what we’ve learned. What haven’t I learned? I feel like my brain pre-WPHP belonged to a different person. This project taught me so much about digital humanities and data and book history, to the extent that I think in relational databases now. But I think the most important thing I learned was how generative and joyful collaboration can be, whether it was sitting in a weekly work session trying to trouble-shoot a weird title record with our team of research assistants, or co-writing an article, or trying to make Kate laugh as much as possible when I read her a nineteenth-century ghost story for a podcast episode. Having the opportunity to collaborate in these ways has taught me so much about being a good colleague and mentor and friend.
What about you? What did you learn from the WPHP?
Michelle: I agree, the WPHP has been transformative not only of my understanding of the period, but of the way I think about what research is in our discipline and even what our discipline is. I am glad you mentioned what he called the related titles and related firms conundrum. That was definitely a flaw in our data model, and one that has proved very difficult to rectify. But not connecting titles and firms was, in a way, a victim of our ignorance, in not understanding how many new editions (and translations) there were, and how many firms there were. And we cannot bear the blame for that, entirely, because nothing had prepared us for the numbers of editions and firms. These discoveries are one of the key takeaways from the project: sometimes you have to start collecting to realize the extent of what is out there. As you may recall, one of the concerns voiced to me at the outset is that there would be too few women writers (and too few of their books), and that women’s contributions to print would look paltry. Well 10 years later, we can say we have definitively repudiated these concerns. And about Michael’s attempt to connect our titles—he used an earlier version of AI to do this. It does a decent job, in that it identifies our title records with similar words and phrases in the titles. But it is far from perfect. One of the main problems with it is that because we include full titles, and many titles include the “by the author” attributions, it picks up titles that are identified in this way. This is why, currently, we do not have this function turned on and available to the public. But we may decide to do so when we officially stop working on the project. Still, this use of AI brings up a deeper issue I want to raise.
We began this project in what might be called the second wave or generation (in keeping with our Romantic roots) of digital projects in the humanities, with the first beginning in the 1990s. As you know, we began working on the data collection and model in 2015, and the public site launched in 2019. Back then, digital humanities was infused with optimism and a sense of boundless possibility. We embraced this, as well as some of the potentially more problematic elements: do you remember we originally promised that the database would be “comphrehensive”? Anyway, my point is that I think we began at a moment when the enthusiasm and promise of the digital humanities was high, and though we thought about endings (we were one of the first humanities projects with a full data management and succession plan), we focused on what was possible and what we could do. By we, I mean all the incredible student Research Assistants who have worked on the project over the last nine years. I fear that if we had started the project now, there would have been pressure to use AI, and I can say with total confidence that AI could not have completed a single title, person or firm record to the level of detail and accuracy that we have accomplished. There is simply too much searching, sorting and evaluation of evidence (see note above re: wishing we had tracked the human individuals and time that went into constructing each record). Having said that, AI is improving and using more reliable sources. I think there may be an end-of-project project involving the use of AI to check and supplement our records, but I think every instance of correction and augmentation would require human attention. But I still think it would be interesting to see what AI can do, even at this late stage. And perhaps AI could be trained to do a better job in linking our titles and firms. Kandice, any thoughts about how the project might be different if we started it now? Or what we could do with new tools with the data we have now?
Kandice: That’s an interesting question. I agree that we might be under pressure to use AI from the start, and I think that would be a loss—not only to our data quality, but to our own knowledge of the dataset. A few years ago, I was on a panel at SHARP with Kirstyn Leuner and Kate Ozment where we talked about the unique knowledge you gain by working closely with a dataset, which we called “data intimacy.” I think that kind of knowledge is lost, or at least limited, by using tools that might increase efficiency. Would I be able to find things in the WPHP as instinctively as I can if I hadn’t literally hand entered data from Raven and Garside’s English Novel in 2015? Probably not. I also think that in order to help develop a tool that can do what we would need it to do, we need to have that kind of intimate knowledge of the data and the difficulties it presents—which we only have because we’ve done the work. There’s a cyclical logic there, but I think one of the things that has become abundantly clear in the last couple of years is that AI tools so often reproduce the assumptions of the people who create them, while a project like ours is built around questioning those assumptions. I definitely don’t know enough about AI to know how to reconcile those two things, but I think that issue needs to be at the forefront of the conversation about whether and how to integrate AI into humanities scholarship.
I also think if we were starting the project now, we might be doing so with a greater awareness of the fragility and vulnerability of digital projects. In the time that we’ve been working on the WPHP, many of those first-generation digital projects that you referenced have become vulnerable or ceased to receive the institutional support that they need. I’m thinking, of course, of the cyberattack on the British Library which resulted in the temporary loss of the ESTC and the ongoing lack of access to many of their digitizations, as well as the Bodleian Library’s decision to take down the British Book Trade Index. (The data has since been made available separately by The Grub Street Project and the University of Victoria Humanities Computing and Media Centre.) So while we have been thinking about endings and succession for a long time, I think that means something different now than it did when we started—and I think there’s a lot more anxiety around what happens to digital projects in the long term now than there was when we started.
While we’re talking about the endings and afterlives of projects, and maybe to wrap up, why don’t we talk a bit about what we’re most proud of?
Michelle: I really like your term, data intimacy, and what it means about the care we have taken with creating our records. I am so proud that we nurtured so many graduate and undergraduate students, teaching them the value of taking the time to look for women writers and those in the book trades. There have been so many discoveries: the extent of women’s involvement in the book trades has been one of the biggest surprises, especially as we had been told that they were anomalous and impossible to find. We now have over 800 women-led publishers, printers, or booksellers—just a phenomenal number of women, many of whom published hundreds, even thousands of titles (not all of which are in our titles). Many of these women have been hard to find, but some have not been so hard, it just takes time and labour and patience in working with the resources we have. One thing we have done exceptionally well—and I credit the podcast primarily and the Spotlights secondarily—has been to narrate our process. To me that is one of the most important feminist legacies of the project. Yes, we have recovered hundreds of women writers, women in the book trades, and their books, and making all of this knowledge accessible is hugely important for our field and more broadly (we know that our Spotlights are already being used for large language models).We have also worked so hard to be transparent and humble about our successes but also our failures. I think these narratives are so important, as we know from the work of Lauren F. Klein and Catherine D’Ignazio on Data Feminism: we look and fall down rabbit holes and sometimes we find something important but very often we don’t. We recognize this effort in our “attempted verified” field, and I think more broadly (again, the podcast has been key), we recognize our limits and the compromises we have made. I think it is important to make this understanding of databases and data work public, but also to train up new scholars with that understanding, that research is not always or even often about huge new discoveries; it is accretive and, at its best, a collaborative project. We relied on so many sources, the work of so many other scholars and librarians; and we have depended on technical expertise of the members of our Digital Humanities Innovation Lab to make our findings visible. We have done our utmost to credit them and to draw out the work of not finding; of making the labour that is so often invisible, visible.
I also am proud of my decision to cede some control and trust you and Kate with the podcast. I will admit that, at first, I thought perhaps I should be vetting scripts, and wondered about letting you two run wild with a pair of microphones. But very soon—and I do mean, very soon, as in after the first episode—I realized not only that I could trust you to represent our work in the most clarifying (and at times, hilarious) light, but that giving you that kind of autonomy was important for your development and confidence as researchers and scholars. As you will be (are) learning as you transition into your own supervisory roles, teaching and mentorship need not be hierarchical and perhaps are most successful when students are allowed to voice their own knowledge and authority. So even though the podcast was almost entirely your’s and Kate’s own, when I reflect on my early decisions, I am pleased that I let you do it! Especially since the podcast has proven to be the most popular and possibly the most durable output of the project.
On a perhaps more basic level, the project (and support from SSHRC as well as various grants awarded by SFU’s Research Office and the English Department) have made it possible to fund so many students, including through rough times brought about by the pandemic and the very dismal job market that still persists. Beyond you, we have had students working on the project for many years, and this continuity has been so important, both for the project’s success but also, I like to think, for the students as well. We have students who have moved on to graduate school elsewhere who continue to want to work for us, and to me this is a testament to the mentorship and community we provide. We have worked really hard to sustain that community both online and through our many gatherings. Whether it was berry picking or watching Persuasion in my basement, we have spent a lot of time together, and these are some of my fondest memories as well as what I think ultimately what counts. We have lots of records, which as you say may or may not survive, but we have shaped and supported and cared for each other, and had fun together, and to me, this is what matters.
What about you? What are you most proud of?
Kandice: Argh! I was going to say exactly the same things as you! So, all of the above. Having a big database with thousands and thousands of records is a huge achievement, but I agree that the community around the project is what has made it worthwhile and sustainable. I’ve benefitted so much—personally, practically, intellectually—from your approach to mentorship, but it’s also given me a model for how to collaborate with undergraduate and graduate students, and how to keep the humans who do the research at the forefront of digital humanities. Having the opportunity to do that from such an early stage in my career, starting as a PhD student, has been incredibly fulfilling. So, that’s not so much what I’m proud of, but what you should be proud of. And speaking of you ceding control, I guess I’m also proud of myself for stepping outside of my comfort zone of very analogue scholarship ten years ago and rising to the challenge. So many elements of this project—from building a database, to training RAs, to starting a podcast—were completely new to me when I started them. I was pretty nervous about something at every stage of the project! But I did it!
Once again we’ve written way more than 1000 words for our Spotlight. I’m glad we’re ending as verbosely as we started (and as we mean to go on, probably). Any final words?
Michelle: I cannot end without saying what an amazing mentor you have been to so many of our Research Assistants: from helping and encouraging them to write that first conference proposal, to actually giving those papers and applying for graduate schools elsewhere, I am so grateful for all the work you put in as Project Manager and Lead Editor (such inadequate terms), a role that Kate has taken over so seamlessly! I love that we have had this multi-generational (not quite the right term, but you know what I mean) aspect to our project.
Fortunately, it is not goodbye as we will collaborate on our new project, The Women’s Manuscript History Project. Yes, you heard that right, it’s WMHP for short. I am excited to move forward, building on what we have learned and how we have worked together. Thanks to the Social Sciences and Humanities Council of Canada, here’s to at least five more years of collaboration to share what we learn about the incredible women of the period we study.
Kandice: To WMHP!
WPHP Monthly Mercury Episodes Referenced
Season 4, Episode 2: Ghosts of Print Culture Past
Works Cited
British Book Trade Index. The Bodleian Libraries, University of Oxford, https://bbti.bodleian.ox.ac.uk/. Currently offline.
British Book Trade Index. Grub Street Project, ed. Allison Muri, University of Saskatchewan, https://bbti.grubstreetproject.net/.
British Book Trade Index. Humanities Computing and Media Centre, eds. Janelle Jansted and Martin Holmes, University of Victoria, https://hcmc.uvic.ca/project/bbti/.
D’Ignazio, Catherine and Lauren F. Klein. Data Feminism. The MIT Press, 2020.
Garside, Peter, James Raven, and Rainer Schöwerling, eds. The English Novel, 1770–1829. 2 vols. Oxford University Press, 2000.
Levy, Michelle. Data Management Plan for Women’s Print History Project (1750–1850). McMaster University, 2020. https://doi.org/10.5281/zenodo.4064167