This is a great question – and one not easily answered. Every situation is different. I have written/co-written/edited/co-edited sixteen books and even with this small sample of the thousands of possibilities there is considerable diversity.

Firs of all, part of being a trained historian is taking a wide variety of courses and readings that give one the background that you’re describing. Trained historians approach topics with a lot of the context you’re asking about in their back pocket. But that is not always the case, and you don’t need formal training to arrive there!!!

Take, for example, my most recent book manuscript (completed submitted for review last January). This book takes its inspiration from a seminar paper, dealing with early western folklore, given to me by a fellow grad student in 1980. He was not a folklorist and I recognized immediately that the topic deserved to be treated fully, so I put the paper in my files, and I allowed the topic to gnaw at the back of my mind. The problem I faced is much like the one you described. It had been a decade since I found inspiration in North American history. I was a committed Europeanist, but that didn’t stop me from feeling the inspiration and recognizing the potential behind my colleague’s quick overview of the topic.

In 1981, I went for a year to study folklore in Ireland, pursuing a research topic dealing with medieval diffusion of a Northern European folktale. The topic failed utterly, yielding a jejune article and considerable disillusionment. But while in Ireland, I began picking away at the topic of that damn seminar paper, pursuing European roots of American folklore. This eventually led to a significant article on tommyknockers – spirits in the mines, but the need to address the broader topic presented by that damn seminar paper lingered.

With employment in the field of the history and archaeology of the West, I rekindled a flame, finding new inspiration American history, but I lacked much of the background, so I started retooling – reading lots of books, writing one – my first and eventually enrolling in a Ph.D. program.

My dissertation was to be “Revolution in Exile: Irish Fenians in the American West.” Unfortunately, I quickly discovered that the required background to address this topic hadn’t been written: I needed a comprehensive, modern history of Virginia City and the Comstock Lode so that I could tackle the topic of the Irish who lived there and organized their revolution.

This resulted in the writing of my second book, The Roar and the Silence: A History of Virginia City and the Comstock Lode (1998). To write this book, I read a everything I could find on the topic, and I used my Ph.D. program to fill in the context of western history that I needed – a massive, quick retooling to fill in the sort of background you are describing.

To augment this, I initiated and co-edited the book Comstock Women: The Making of a Mining Community (1998). This harnessed the research interests and sometimes forgotten, unrealized efforts of a dozen or so diverse experts. This helped me to fill in context that I needed addressed.

Building the context, then for my original inspiration, “Revolution in Exile” was consequently considerable. And I never wrote the book! But the pair of tomes, The Roar and the Silence and Comstock Women, remain on the shelf, regarded by many as the standard resources on the topic – until some punk-ass grad student writes their replacement: I hope I live to see the day and I stand ready to assist!

Context, then is something that can be built to suit the current need, or it is something that one brings to the table. Some books require a tremendous amount of building of context, and others are narrow in definition and can be knocked off quickly.

My wife and I co-authored a sweet little treatment of an obnoxious millionaire and his remarkable home at Lake Tahoe: Castle in the Sky. This was a quick commission to serve as a fundraiser for the non-profit managing the facility. We brought a lot of context to the table, but the topic was so narrow that little more was needed. We knocked off that one (and its second edition) quickly.

But what of that damn seminar paper and that original inspiration? Taking it to Ireland and eventually pursuing the emigrating Cornish knocker which transformed into the American tommyknocker, I found inspiration in Cornwall. That became another gnawing inspiration that lingered for decades of context building. Eventually – in 2018 – it yielded The Folklore of Cornwall: The Oral Tradition of a Celtic Nation, appearing nearly four decades after the original inspiration in the folklore library at the Department of Irish Folklore. A lot of context building!

And my most recent manuscript: it has passed peer review and will almost certainly be accepted by the editorial board for publication next month (appearing perhaps in 2023), forty-two years after the original inspiration. It is tentatively titled “Monumental Lies: Early Nevada Folklore of the Wild West.” Embedded within those pages is that damn seminar paper that the grad student handed to me, … together with four decades of gathering material and reading all sorts of things to provide context. My problem had been finding the means to put my arms around the topic. The question ceased to be context long ago. Approach was what I lacked. Then lightning struck and I hammered out one of my longest manuscripts in five months.

So you can see – sorry for the long journey – that it all depends on the topic and the specific situation!

It never stops. I'm still down here. Send help.

But seriously, I think it comes down to recognizing what's directly relevant for the topic you're working on and what's ancillary information. Of course, you the historian need to know that ancillary information, but you're not communicating as much of it to the reader, so you don't have to go into the same level of depth. Of course, figuring out where to draw that line can be difficult. It's probably a bit easier for me than most people, because my field concerns a short, finite period of history (if it's ca. 1933-1945, it's directly relevant; if it's not, it's probably ancillary), but there are still gray areas, especially if you're working on something like my current project, which deals with the postwar period rather than the war itself.

In my dissertation/first book, I fell into the typical grad student tendency to basically go really into depth on background information because when you're writing a dissertation, you're really trying to prove to your committee that you know the history and the historiography well, and going into inordinate amounts of detail is basically a defensive mechanism. I had like two whole chapters devoted to ~80 years of preceding history for a project that only covered events from 1940 to 1944. It was really excessive, and when it came time to actually publish it as a book, I ended up having to cut like, a full chapter of background information and historiography because it just wasn't important enough to have that level of depth on topics that have been covered elsewhere. The reviewer was basically like "give your readers some credit, this information already exists, if you footnote it they can go find it, streamline this and distill it down to the truly critical points" (this is why negative reviews are usually more helpful than positive ones). It was a good object lesson in concision and knowing when to stop, but it's still hard to say for sure where that line is in general. I think it's going to vary from person to person and project to project. In the end, it boils down to giving your reader all the information they need and none of the information they don't, which is hard to gauge as the author who has all the information they need and all the information they don't.

This is probably not a very helpful answer because it's really more of a feel thing than something you can answer with clear, objective criteria, but at the end of the day you do have to know when to draw the line because if you try to follow every little thread of research you'll never finish anything. I guess the practical way to look at it, as I alluded to above, is "how much of this is going to make it into the actual published work?" If you're devoting a chapter of a book to something, you probably need to go deeper into the historiography, but if it's just something you're writing a couple of paragraphs about, you probably don't need to spend an inordinate amount of time on it. There are only so many hours in the day and you have to be practical about it on some level.

There's been lot of smart things said already, but I should point out that not all historical fields are alike, when it comes to drawing lines on where to stop. The publishing industry known as David McCullough has done books on John Adams, the Panama Canal, the Wright Brothers and the Brooklyn Bridge: all things that have good records and plenty of secondary sources for his research assistants to work with. That means he can swoop in, quickly digest, sort his sources, publish, and go on to the next. But there are some fields that lack those advantages, where scholars have to spend a lot of time with their sources, as they're often meager and/or difficult to understand, and/or are written in difficult languages. David McCullough can breeze into an archive or two and quickly put together something good about Truman. But he can't do that with, say, Odo of Cluny. And it's not that these topics are unimportant. For example, there was a population of free Blacks in the 18th c. colonies living at the same time as John Adams that we now would love to know about. But they were often poor, and whatever they left in writing has not tended to survive. You find some of their names in old store ledgers, making purchases or sales, and in tax records, paying property taxes on something they own. But, perhaps predictably, whatever letters and diaries they wrote have pretty much vanished.

Working in a field like this requires a lot of detective work, and time to do it, and time to think about it. There is no choice but to go down the rabbit hole, and stay there.

I would say this is a much wider issue with research as a whole, I am in a completely different field (ML) and sometimes when I read into a relevant subfield it feels like I am swimming in very deep water (not sure how else to describe it).

For example sometimes I can spend a week reading up on maths I only need to briefly introduce to contextualise in whatever I am working on, but it feels important to understand in the moment, even if, in the end paper I will only talk about it for a sentence or two. Learning when to stop for what you need is both practise and recognising to draw the line at exactly what you need. It is something I am still getting the hang of.

Reading as you are is also very different from publishing, as in publishing the line is a bit clearer (include exactly what the reader needs to understand what you have done). Reading is up to you, you can go as deep or as shallow as you want to. I find it helps to set a timeline for certain topics: in my case I might want to understand X model, how it is derived and how its used by Y date. This will inevitably lead me down the road to new maths and ideas, which I can easily waste a lot of time on. At that point I can ask myself if I am on track to meet my goal, and if not I can make the decision to drop that line of reading if it is not helping me towards my original goal and timeline.

Edit: Sorry if this perspective is not what this sub is for! Just thought it might be useful to show it is a broader problem