There are two ways of writing a dissertation, broadly speaking. The first is to do all of the research that you will need to do before you even think about putting pen to paper: outline your argument, cover your bases, and then begin. The second is to start writing before you have completed your research: get a sense of where you’re going, be ready to change your mind, and write what you can when you can. Each approach has significant benefits and drawbacks.
With the first, the benefit is that you have charted a more or less clear path through the argument you’re going to be making. You know where you’re headed, who has been there before, and the path you’re going to take to get there. At least in theory, this approach makes the writing itself easier. The drawback to this approach is that it allows you to postpone the writing process indefinitely in favor of more research and reconceptualizing (a fancy term for procrastination). That is, if you’ve decided to do all the research before you begin writing, it is difficult to tell when enough is enough.
With the second approach, you discover much about where you’re headed while you’re on your way. You know where you want to go, and you have a vague idea of what the path you’re going to take looks like (or at least what you want it to look like). But because you are granting that you will uncover many things in the course of your writing, the second approach requires a degree of flexibility. And therein lies the drawback: you have to be willing to change your mind or alter your course after you’ve already begun.
Regardless of which approach you choose (I myself prefer the second), you will at some point encounter the phenomenon that some refer to as the “rabbit hole.” The rabbit hole is comparable to discovering a loose thread on a sweater. You see the thread and you pull it. Sometimes it will be an inch long, and you can remove it without difficulty and forget about it. Other times, it is much longer, and continuing to pull at it will undoubtedly destroy a portion of your sweater (if not the entire garment).
The same is true for rabbit holes. When you see such a hole, you have no idea how long the tunnel behind it is or where it leads. The only way of determining the nature of a rabbit hole is to dive in and take a look. You may discover that it leads nowhere, that it is little more than a shallow crevice. Or, you might find yourself in a veritable maze of new, unexplored territory. The question, of course, is whether to include what you find in your dissertation, to save it for another venue (an article, presentation, etc.), or to simply ignore it (not a smart choice).
Rabbit holes can be frustrating for a number of reasons. If the material you discover is relevant to your topic, then you have to make time (and space) to incorporate it. If it contradicts something you’ve already said (or planned on saying) then you may have to defend or alter your position. If it is truly interesting but simultaneously irrelevant to what you’re trying to say, you might have to find another outlet for it. To be sure, resisting the temptation to “say it all” is a challenge that all dissertating students share.
But rabbit holes can also be invigorating. If you are researching a topic in order to confirm a hunch (or a “thesis,” if we are being fancy), then you may uncover material that confirms your hunch in a way that you didn’t expect. That is, you might uncover a path to your desired location that is different from and perhaps better than that which you intended. If this is the case, the challenge becomes determining what to do with the path you’ve already charted. Does it continue to exist as an alternate, albeit less-desirable route (i.e., in footnotes), or does it simply fall by the wayside? The latter of these two options is painful, to say the least, as it involves essentially trashing what may amount to weeks (or months!) of research. Yet even paths that end up being “less good” are valuable in the grand scheme of things, as we learn much from the roads we travel, regardless of whether they end up being the most efficient.