— continued from here —

Thy Paper Shall Have a Story

Papers for publication are different from proposals, and they are also different from reports or theses/dissertations. (This insight brought to you by Captain Obvious.)

Before you write a paper, you first have to ask yourself:

a) Do you know what you did?
b) Do you know why you did it?
c) Do you understand what you found?

Most students know what they did, which in a paper goes somewhere in the methods section and is often easiest to write.

However, most students, especially young ones, don’t actually know why they did what they did (other than that the advisor nudged them towards it). Why you did what you did is what makes your introduction and it is a very important part of the paper.  Often, writing a poor introduction means not really understanding where you work fits with the state of the art, which in turn means that you have to go back and read the literature more broadly, and you have to talk to your advisor more.

Understanding what you found goes into the results part of your paper. That’s “the meat.” Students have an easier time writing the results than the introduction, but often I find the results to be written trivially, just reading trends of graphs. That’s generally not enough for any reputable journal.

Your paper has to tell a story. It needn’t be the world’s most complete story, but it has to have a beginning, a middle, and an end. In other words, it has to have the Why (Intro), the How (Methods), the What (Results), and the So What, a.k.a. how it all fits with what we know or don’t know (Results/Conclusion). This is the paper skeleton, and it’s a good idea to not start writing the full paper until you are comfortable with it. (Again, you and your advisor should be talking about the skeleton  several times.)

Once you really truly understand the why, the other parts are easier to write. The intro needs to connect between setting the stage for the reader to recognize where your field is and what the important open problems are, what you do to address an open problem and how, and what you found and why it’s important.

Introduction (each of the paragraphs can be more than 1 if lots of material)

Paragraph 1: Open with an overview of the state of the art in a broader area and perhaps note applications relevant for your paper.

Paragraph 2: Recent developments in a closely related subarea; what is known (who measured/calculated it), what is still being debated (who said what, if there are competing experiments/theories)  or what is relatively open (any relevant work, perhaps on related systems), and why it is particularly important that we find an answer to some of those questions. (This paragraph also mentions briefly, some of the common experimental or theoretical methods, as you discuss the work of others).

Paragraph 3: “In this paper/letter, we…” State concisely what you did. It needs to connect to the open problems you just discussed as important in the previous paragaph. State what you did, how, and what you found, and how it answers the question(s) posed in Paragraph 2.

Paragraph 4: When writing comprehensive papers in the physical sciences, there is this “Table of Contents” paragraph, where you say things such as “This paper is organized in the following manner. In Sec. II, we present the methodology, …” Many people keep it entirely generic (Sec. II Methods, Sec. III Results, Sec. IV Conclusion), but I like to put in more detail and use this paragraph to show how the main thread connects my story (“In Sec. III A, we show the vibrational properties of vibranium obtained using neutron scattering experiments… In Sec. III D, based on numerical simulation, we reveal that these unique mechanical properties of vibranuium make it an ideal material for intergallactic warfare.”)

The Vomit Draft

The whole paper — the material you put in and the order you put it in — is in the service of presenting your story.  The story should be reasonably clear in your head before you start writing. However, sometimes, after you start writing, you actually go and look some stuff up and think of things another way and, all of a sudden, you may need to change parts of the story somewhat, or even dramatically. This is natural — it happens in technical writing as well as in fiction — and is perhaps the most fun part of writing: the fact that writing helps clarify your thinking. Which brings me to the common saying that you make figures first, then write around the figures; this is true enough, but is hardly gospel. I say have a skeleton first, a decision on what the paper will be about (informed by dozens of figures you and your advisor already went through before writing), then make the figures to best support your story, and then write the results section around the figures. My students and I redo figures many times during the writing-editing process, as we distill our message.

However, you have to start somewhere, and that somewhere is a reasonably clear concept of your paper’s story. (Sheesh, have I said it enough times already?)
Then you start writing a rough draft, also sometimes colloquially referred to as the vomit draft, because it hints at people vomiting the inside of their heads onto the page; like vomit, the product is generally misshapen and not pretty, but is usually not smelly. [If your vomit draft actually smells like vomit (eeww!), stop spilling food all over your keyboard. You are gross.]

Now, I know there are folks in the blogosphere who will tell me that some people don’t write the rough draft but their sentences come out their heads perfectly formed right away. If you are like that, more power to you, you are a freakin’ unicorn — I bet the horn focuses your thoughts! However, most people need a rough draft, which they then edit, and for them it’s good practice to separate writing as things come to mind (which helps with the flow) from editing (which ensures that horrible grammar, spelling, and punctuation are not unleashed upon the world, or worse — your advisor’s desk!).

So, with a clear outline of the main story in your head, just start typing things as they come to mind, as you would tell them to another person. Imagine yourself giving a talk… unless giving a talk is even more terrifying than writing, in which case imagine you are on a beach, in a hammock, talking to a really hot hammock neighbor about your research project over some margaritas. At this point, please don’t worry about grammar at all. Just write how you think about your story.

A trick that helps is to start writing amidst a block of already existing (unrelated) text, so the whiteness of the page isn’t daunting. I always cite Finding Forrester and how Forrester helped his young protege by having him type on top of an old story. This trick has helped many a student. In my group we use LaTeX, so everyone starts learning LaTeX by working from someone else’s paper anyway, which provides examples of LaTeX commands and of semi-related existing prose wherein you can nestle your draft.

The point is to start; once you start, just dump the contents of your mind onto the screen. Write as things come and, as long as words pour out of you, don’t stop… unless you are starving, have a bladder that’s about to burst, or can barely keep your eyes open; in any of those cases, please stop; you will be able to get good flow later, I promise. I hate it that in the popular culture the ability to write is conveyed as some sort of mystic, hard-to-replicate experience (and nobody ever shows editing); a recent offender is the movie The Words I just saw on Amazon Prime (the movie is so-so otherwise).

Don’t worry about “the muse” in technical writing. You will get good chunks of text out of you on multiple occasions. Just write. It needn’t be pretty, it needn’t be super organized, just write. Consider it a chance to reveal your thought process and your knowledge and your excitement to other scientists. The more you write, the lower the barrier to writing.

Edit later

But when is later? When the flow stops. When you are done with a few paragraphs. When you are having a hard time getting the flow started. When you are almost done. When a few days has passed since you finished a draft. When office mate asks you to look at their paper. When it’s either edit or check references. Basically, don’t edit at the level of every word or every sentence you produce, but you can do it paragraph by paragraph or page by page or section by section, depending on personal preferences, available time, and editing stage.

Good editing means careful reading, putting yourself in the reader’s shoes, and — more often than not — murdering your darlings, so your other darlings can go on and get reviewed well at a fancy journal.

Don’t be afraid to let some grammar rules slide if they ensure good flow: e.g., sometimes long sentences are justified and work better than shorter ones, which can be choppy. Don’t be afraid to use punctuation in the service of your point: I use commas, dashes, and parentheses, which can all help separate a minor clause, depending on how closely it’s tied to the main clause. I love the semicolon and use it a lot in technical writing; it provides closer coupling between successive sentences than a period. Things like that. (Sadly, ellipses and exclamation points are not welcome in journal papers.) Proper punctuation will help pace your reader.

What say you, blogosphrere? These were lengthy posts, so I am a little (okay, a lot) out of steam, and I am sure I forgot a whole bunch of things. Please let J know what your favorite tricks for ensuring good flow in technical writing are. 

Here are also links to related posts on Academic Jungle and here on xykademiqz. 

Happy vomit-drafting!