Figuring out How to Write Again...
Thursday, August 2, 2012 at 9:30PM There was a point just as I transitioned between my Ph.D. and first postdoc where I felt like I was being crazy productive. I had just written a thesis and was working on three separate manuscripts (all of which were actually published). All of this material came out of projects that I'd had a large part in designing, and so I didn't feel as though I had much difficulty charting out drafts for publication.
Then two years passed during which I wrote little of my own design, but rather focused on contributing analysis to larger efforts. I did draft at least one manuscript, but it fizzled because key co-authors had different opinions about where the emphasis should lay. It's not easy to write a good manuscript when you disagree on which results are the most significant and interesting1.
Unfortunately, I feel that I've become 'rusty'. Blogging helps with grammar and syntax (which is part of the reason that I've worked to keep it up), but it doesn't necessarily help with good 'science writing'. Actually, I've found that there are so many opinions regarding what 'good science writing' is, that I'm beginning to think that it's one of those 'you'll know it when you see it' things.
In the midst of working on flexing my typing fingers again, a colleague pointed me to an excellent journal article that both highlights some of the common mistakes that scientists make when trying to convey complex information to their peers, and suggests many alternatives. Some were even revelatory to me, at least in terms of focusing my own though process; while all the while being rather simple to implement. I wanted to share the article, which can be accessed by anyone here:
Gopen GD, Swan JA. 1990. The Science of Scientific Writing. Am Sci 78: 550–558.
I think that a few key insights from the paper are worth discussing. After pointing out some of the common pitfalls of scientific writing2, the authors suggest that there's a reason for why manuscripts aren't always written with a major focus on clarity: scientist are much more concerned with making sure that all of the information that must be conveyed is there rather than making sure that the 'flow' is adequate.
However, as the authors point out, improving flow requires only a few rules of thumb (you can go read them for yourself). The most significant suggestions that have changed the way I've been looking over my drafts are summarized in the following three points:
Place in the stress position the "new information" you want the reader to emphasize.
Place the person or thing whose "story" a sentence is telling at the beginning of the sentence, in the topic position.
Place appropriate "old information" (material already stated in the discourse) in the topic position for linkage backward and contextualization forward.
Once you figure out the 'stress' and 'topic' positions, it seems simple. Despite this, I think that it says a lot: Scientific writing is all about clarity and communication. While basic rules about sentence construction still apply (avoid redundancy, etc.), the overall construction of sentences, paragraphs, and sections works best when they follow a defined flow linking each section explicitly into the next. Again, this may seem intuitively obvious, but when you begin looking at the examples given in the text, I think you'll notice that many of the times you've had to read over a section multiple times, were because the expected flow of information was interrupted3.
Regardless, the article helped me reorganize some troublesome sections and lay out areas of the text more efficiently. I hope that others will find it useful as well.
1I've found that this is surprisingly easy, particularly when your data is being interpreted by people with widely different interests or backgrounds. An observation that is fascinating to you can be completely (and disappointingly) uninteresting to someone outside of your field.
2For instance, scientific papers often, in order to convey more information, separate subjects from verbs via overly long clauses, which leads to a difficulty in parsing due to playing against expectations. Mmmm... bittersweet.
3This occurs more often than not because information is referenced externally, or knowledge is assumed of the reader due to space constraints.
Reader Comments (2)
I've been working on my writing flow, too. Apparently with some success, one of my co-authors, a professor here with decades of experience at this kind of thing, told me my recently-submitted-to-journal second paper flowed very well, better than the first one from most of a year ago. It's certainly a skill within the broader skillset of "good scientific writing", and it can be improved with practice.
Thanks for the link to that paper, it looks useful.
Speaking of, I know what I should be getting back to right now...
Yeah, there's an aspect to my struggles that I didn't blog about in that the bar's been raised in terms of where we expect to get some of this stuff published. All of my Ph.D. work went to modest journals - it was pretty specific and therefore mid-range. You can be wordy, assume a certain familiarity from your readers, and, most of all, be honest about pros and cons assuming that folks will interpret them correctly. With my current stuff, we're shooting much higher, and so it's all about being very concise, positive, and obsessively self-hyping... something I'm not very good at, and honestly, I find a bit repulsive. I've always been somewhat disturbed by people who shamelessly self-promote instead of letting the work 'speak for itself'. But in realistic terms, grant/manuscript reviewers don't have time to 'suss out' the cool results from your papers, and big journals are so much more general that it's not a given that an academic editor will be qualified to do so. And so you need to point out what makes your work awesome 'without humility'. A good skill to learn.