Photo: rows of theses in UNB’s library (S. Heard)
PhD theses are weird things. Now, it may seem weird that a person is expected to spend five years labouring to produce a single document that determines their career prospects* – but that’s not what I mean. Instead, I mean that they’re a weird writing form. They’re important, but nobody ever reads them. They’re singled-authored, except that they’re not**. They’re bulked up with trivia, but judged by their importance. They’re considered publications, except that republishing their content later doesn’t count as double-publication. What on earth is going on?
Let’s back up a moment, because it will help. Before you write anything, you should ask yourself three questions:
- Who is the audience you’re writing for?
- What is the function of the written piece for that audience (why are they going to read it)?
- What do you want your reader to take away from the piece?
The first two are the key to the thesis puzzle. Different bits of writing – a letter to the editor, a blog post, a journal paper – are intended for different audiences, and have different functions for those audiences. That shapes the way each should be written: the vocabulary (because what’s precise terminology for one audience is jargon for another), the style, the content – everything, really.
So what about a thesis? The reason theses are weird is that they don’t have a single audience or a single function. Instead, a thesis has to perform three different functions for three different audiences, and these can coexist a bit uncomfortably inside the single document. Here are those functions:
- Communicating science. A thesis makes a novel contribution to scientific knowledge, and must communicate that contribution. The audience for this function is other scientists in your field. Well, sort of. Nobody will actually read your thesis (sorry about that); instead, people will read the journal papers that your thesis becomes. Nonetheless, a thesis must be plausibly able to communicate science to hypothetical readers beyond the examining committee.
- Establishing credentials. A thesis is evaluated as evidence that its author merits the PhD. The audience for this function is the examining committee, and this explains much of the detail packed into the thesis that can make it painful to read for anybody else. The committee must be convinced that you understand the techniques you used and the results and implications of your work, and that you’re aware of the work’s context in the field. This may mean lengthy presentation of methods and analysis, and a comprehensive review of past literature – both well beyond the level of detail appropriate for journal publication. The conflict between this function and the first is obvious: including enough detail to convince a committee of your in-depth knowledge will inevitably get in the way of communication with a normal (peer scientist) reader.
- Archiving unpublished material. In many labs, theses serve as archives, recording things like highly detailed methodologies, records of failed experiments, annotated datasets, and data that were gathered but didn’t make it into the papers making up the core of the thesis. The audience for this function consists of the supervisor plus future researchers in the same lab; and the degree of detail needed for this function may be very large – far beyond even that needed for the credentialing function.
The need for one document to serve three functions for three audiences explains most of the weirdness I started this post with. It also explains the increasing dominance of the “papers format” over the “thesis format”.
A thesis-format thesis is a single document, with chapters that aren’t intended to stand on their own and would never be submitted to journals without major rewriting. This format prioritizes the credentialing and archiving functions over the communication function (the latter forced to be implicit, as the examining committee has to imagine the condensed and rewritten papers that will eventually result).
Thesis format was once the rule (hence the name), but it’s fading because the papers format has a huge advantage: its modular construction compartmentalizes the three functions. The core chapters are manuscripts prepared for journal publication, or even papers already published; they’re written to stand alone, and they fulfill the communication function. There’s usually an introductory chapter and a concluding one, which serve the credentialing function: they demonstrate knowledge of context in great detail, but nobody outside the examining committee will ever read them. The archiving function is nicely served by appendices, which can be conveniently skipped by nearly all readers (even the examining committee won’t normally read them). The compartmentalization in a papers-format thesis takes us back to where, as both writers and as readers, we want to be: with each module (each chapter or appendix) designed to play a single defined function for a single defined audience.
Does all this thought about audiences and functions remove the weirdness of the PhD thesis? Well, no, but it explains it, and lets us see past that weirdness to appreciate how each module fulfills its own function. That appreciation makes a thesis a lot less irritating to write, and a lot less irritating to read. I wish I’d understood it mumble-mumble years ago, when I wrote my own thesis.
© Stephen Heard (email@example.com) November 24, 2016
This post is based, in part, on material from The Scientist’s Guide to Writing, my guidebook for scientific writers. You can learn more about it here.
*^Although it’s really just the centuries-old tradition of apprenticeship, in which an apprentice becomes a master (and gains membership in a guild) by producing a substantive work – their masterpiece. We’ve largely forgotten this literal meaning of masterpiece.
**^I’ve seen members of examining committees protest when a thesis chapter (in a papers-format thesis) has a byline noting coauthors, apparently reasoning that as long as the inevitable collaboration isn’t actually acknowledged, everything will be OK. Take a deep breath, people; collaboration is a good feature of a thesis, not a bad one.