Citations are Rebels

In summer 2025 I finished revising the manuscript and bibliography of my second book and started to sort out the image rights and commission artwork. I will have more to say about that project in 2026. In the meantime, I want to talk about some of the troubles which come up when trying to clearly indicate where your information comes from or where readers can learn more.

Historians work with multiple versions of the same text side by side. Often any one edition of the original text is not perfect, or they are not fluent in the original language and use translations to find specific passages which they want to look at closely. In classics this does not create major problems because there are standard methods of citing ancient Greek and Latin texts which let you find the same place in whatever edition or translation you are using.¹ But nineteenth-century philologists did not always bother to divide medieval texts into numbered chapters and sections, and even today they don’t always bother to note how their edition of a text differs from earlier editions as a classicist would note the differences in the apparatus criticus. So if you want to talk about the battle of Agincourt, its often best to cite a page in an edition of the French or Latin text, and a page in Anne Curry’s sourcebook. Curry did not include the original languages, and the editions generally did not include translations. There is no easy way to find the right place in the original language given the translation, or vice versa. Citing both is easiest for the reader, but it clutters up footnotes, especially when citing parallel passages in different sources (several of the earliest surviving writers on the battle seem to have talked to one another or read one another, and when Curry wrote nobody had figured out who came first).

Sometimes I check photos of a manuscript and a printed edition. The Internet has made original manuscripts much more accessible than the old days when you ordered a microfilm and waited for it to arrive in the mail so you could take it to a library and see if the black-and-white images were readable (I have one microfilm of a manuscript in my office which I bought in roughly 2011). Including the page in the manuscript and the location in the printed edition is helpful but makes the footnotes complicated to parse.

It seems straightforward that a work has an author or authors and sometimes editors, translators, and so on. Older editors often included the name of the author in the title, as in Chronica Johannis de Reading et Anonymi Cantuariensis, 1346-1367. There are also books which list people who contributed chapters or technical services, or simply acknowledge a number of named “assistants.” Archaeologists often need lengthy technical help from specialists in dendrochronology or small finds or textiles, and sometimes they contribute enough writing to get credited on the title page of the book. Another common division is for one person to edit a text, and another to translate it or provide a commentary. Which of the two is the author? If you sort publications alphabetically, which should you list first?

Some works have a publisher and a printer. Until well into the 20th century, it was common for learned societies to take on many aspects of publishing themselves, but contact for services with a local printer. Conference proceedings and editions of sources were often published in this way. Its not always clear whether the printer or the organizer belongs in the bibliography. For a few years there was a Medieval Electronic Scholarly Alliance which offered to provide peer review independent from publishers, so a self-published work could be reviewed before publication by someone who was not a friend of the authors. The peer-review service seems like it should be mentioned, but most style guides assume that the publisher lets you guess whether the book was peer-reviewed before publication.

Sometimes all parts of a publication appear at the same time, and are just bound separately for practical reasons. Other times they appear years apart, and are sometimes even printed by different printers. It even sometimes happens that the original author dies and his work is continued by someone else, so the last part has a different author. The Early English Text Society edition of Lydgate’s Troy book is a good example: shortly before the first world war, the editors had the budget to print and distribute one volume of a certain size per year, but this poem was too big, so they divided the poem into several volumes which each had an appropriate page count. The subscribers got their annual volume, and the whole poem was edited eventually.

Nobody worries about exactly what printing of a book they used. But I found one reprint of a collection of sources which removed a few pages from the end because the reprinters thought the original editors had included things which did not belong (The Household Books of John Howard, Duke of Norfolk, 1462-1471, 1481-1483 by Anne Crawford).

There are works which specialists cite in specific ways, such as legal cases or acts of parliament. Those special ways are clear and concise if you check legal citations, Egyptian papyri, or Late Babylonian texts every day, but not every reader will know how to decipher 5 Henry VIII c. 8, Canada (Attorney General) v. Bedford, or Dar. 253. On the other hand, using the long form can become repetitive. Older works like this do not always have a single canonical version, so if a reader has a different version than the writer, they might see different things. Ancient historians take for granted that readers will know where to find say Ammianus Marcellinus in original or translation, but medievalists often want to be directed to a specific edition or translation. That also makes footnotes longer and more cluttered.

Scholars of the ancient and medieval world often use art as sources. We can’t always print a good photo or drawing of everything we used. Pointing to the inventory number and the institution which holds them is helpful, but is not necessarily enough to find an image. Large architectural sculptures do not always have individual inventory numbers at all. Citing a book which has a photo or drawing of an object is useful if the reader has that book handy.

Many scholarly publications now have a DOI or a stable URL which is supposed to point you towards a website where you can download an up-to-date version. Some are even published online first, and in print later or never. When you delve into them, you discover public-domain works which have been assigned a paywalled DOI by their publisher, and stable URLs which have broken. Two years after the website of the British Library was taken down by crackers, their EThOS collection of links to theses and dissertations is still down, and the best they can promise is that by the end of 2025 they will point to theses archived by individual universities (archive link). Recreating their collection of files, so that if universities were hacked in the same way the British Library copy would still exist, will begin in 2026. It is very much an open question whether to bother listing URLs for pre-Internet publications in a print bibliography, because they may not work in ten years let alone a hundred years.

The Web Gallery of Art was briefly down in 2025 as well. Providing a link is not necessarily a way for someone in 10 or 20 years to reliably find a photo of a painting or sculpture, because those links break over time. Even government projects such as Bildindex are often transformed.

Citing informal web publications is even messier. Some of the ones which I cite have already gone offline and are only preserved by the Internet Archive. Sometimes once-useful domains go dormant then are taken over by sleazy businesses who take advantage of their Google rank. There are new archiving services such as archive.ph which may or may not last. They give tidy short URLs, but those URLs are only meaningful as long as the original service works. I point to websites when they were my first or best source of information on something, but I can’t promise that readers will be able to find them.

All of these came up in one book with about 250 items in the bibliography focused on one part of the world. Imagine what it would have been like in a longer or broader book!

If you went to university a long time ago, or you only publish in one field, you can forget many of these difficulties. An advantage of not working in academia is that I can publish on my different interests in different places, so I often see that the way I want to cite things is not what the venue expects, or what is clear to me is not clear to the reviewers.

There are style guides which make a heroic effort to standardize all of this. I am experimenting with Zotero which is designed to scrape metadata out of a PDF or a webpage, store it in a database, and spit it out in one of those rigid formats. I hope you can see why I think this will save time, but not solve the central problem. There is no way to reduce this wild jungle of diversity into a monoculture pine plantation. Even if academics agreed to simplify how they publish things tomorrow, the past 200 years of academic publications would remain. The actually existing Internet is fundamentally unsuitable for the demands of scholarship, which require that if I cite something, you can find the same thing a hundred years from now, with no more intervention than keeping it somewhere dry and updating the catalogue every few decades.

I know I usually ask my gentle readers to support this site, but donating to the Internet Archive would not be a bad idea either!

(scheduled 14 September 2025)

1. Except for a few works like Pliny’s Natural History! ↩︎