Here are some of my opinions about writing.1
Write plainly and straightforwardly, using concrete words whenever you can. From Robertson Davies’ novel The Manticore (1972):
Ramsay always insisted that there was nothing that could not be expressed in the Plain Style if you knew what you were talking about. Everything else was Baroque Style, which he said was not for most people, or Jargon, which was the Devil’s work.
I don’t have a favorite book on writing. I grew up with Strunk and White’s Elements of Style. Although the linguists at Language Log routinely (and justifiably) ridicule their grammar advice, their stylistic advice is a good starting point. Don’t adhere to their rules (e.g. “omit needless words”) rigidly, but keep them in mind, especially if you have fallen into an overly complex style. I also like Orwell’s Politics and the English Language — also denigrated by Language Log, also not to be taken as gospel, but a useful and enjoyable read.
Prefer the active to the passive voice, and language that attributes agency to vague language.2 (Some people recommend the passive voice in a methods section, because this emphasizes the procedures rather than who carried them out; I still prefer the active voice here.) For example, replace “It has been observed that ducks like water (Schmoo et al 2015)” with “Schmoo et al (2015) observed that ducks like water” (but see below).”
Using the first-person singular in scientific writing is controversial, but do it if you are writing a sole-authored piece and the audience will let you. If you’re writing with co-authors you can use “we”; you may also be able to use “we” in the sense of “I and you, the reader” (this approach is common in mathematics textbooks: “We next define … Therefore we can easily conclude that …”).
Scientific readers are generally more interested in the current state of the field than in its history. While you may sometimes want to illustrate a controversy or the outline the historical development of some research area, it’s generally better to write about what’s known now, i.e. where your study is starting from. Similarly, unless you want to refer to a particular study several times, “Ducks like water (Schmoo et al 2015)” is better than “Schmoo et al (2015) observed that ducks like water”.3 Don’t say “Research shows that …” — this is both vague and focuses on the research rather than the information that comes from it. Unless you are specifically writing a historical paper, focus on information and ideas rather than the people and studies that gave rise to the information and ideas.
When referring to observations or conclusions based on your own work described in the paper, you should say “Ducks like water” rather than “We observed (concluded that, found that, etc.) ducks like water”; it’s more succinct, and your readers will be able to infer that such statements are based on your observations and/or logic.4.
Don’t introduce your paper by saying that many people have long been interested in the topic: “your paper should introduce the biological topic and explain why it’s interesting and important, not say that other people think the topic is interesting and important” (Jeremy Fox).
Be deliberate about where you put the strong point or punch line of a sentence or paragraph. Depending on the flow of your argument it should go either at the beginning or at the end, rarely in the middle. Consider which ordering of clauses in a sentence will best engage the reader.
Verbs are usually punchier than nouns, including gerunds.
As Strunk and White say, “omit needless words”. After writing a draft, go through and see what qualifications and waffles you can delete without changing the meaning of your sentences.
Reading your own prose out loud is a good way to proofread and find awkward phrasing.
Phrase statements positively rather than negatively: “ducks are likely to be larger in the tropics” rather than “ducks are unlikely to be smaller in the tropics”, or “our attempt to catch ducks failed” rather than “our attempt to catch ducks did not succeed”. (Orwell specifically deprecates the “not un-” construction.)
I had it beaten into me in high school that “this” should never be used without an antecedent (“Ducks are bigger in the tropics. This means that …”) “This” what? Although I no longer believe it’s a hard-and-fast rule, include an antecedent whenever you can (“This enlargement means that …”); it will make your writing clearer.
Avoid redundancy, particularly when hedging. Instead of saying “in most cases, ducks tend to be larger” say either “in most cases, ducks are larger” or “ducks tend to be larger”.
The verb “to be” is vague; you can usually replace it with more concrete and specific phrasing. “There is/are” is a vague way to start a sentence. (Think about why you’re giving the reader this piece of information, and try to work that motivation into your transition.)
“May” or “may be” can replace “could suggest that” or “is probably” or “is likely” and other waffle-phrases
“and” splices: writers often use “and” to string together clauses that are only loosely connected: “Ducks are large and when they are approached by humans they make noise”. Putting these clauses in separate sentences (“Ducks are large. When they are approached by humans they make noise.”) or separating them with a semicolon (“Ducks are large; when they are approached by humans they make noise.”) can be better.
“effect” is overused and vague. “has a positive effect on”, “has a positive impact of”, “positively influences” → “increases”. “Affects” is better than “Has an effect on” (verbs stronger than nouns), but even here you should try to be more specific. In many cases you can make a directional statement about the effect (“increases”, “decreases”); “increases” is better than “affects positively”. I don’t like “impacts” as a verb (I recognize that this preference is old-fashioned/peevish). “Negatively impacts” should be “hurts” or “decreases” or …
For ecologists: “density-dependence” is vague. I try to avoid reifying density dependence (i.e., convert it from a label that describes a class of phenomena to a thing). “Density-dependent mortality” (or fecundity or whatever) is better than “density dependence” alone; the more specific you can be (e.g. attributing the effects to competition for limited resources or nest sites or whatever), the better.
A similar example from evolutionary biology: “species X has larger female-biased sexual size dimorphism” → “species X has larger females relative to males” or “females are larger in species X”
Avoid constructions of the type “A and B have effects C and D, respectively” — you save a few words this way but the reader has to work harder to associate A with C and B with D
Using possessives appropriately can make your prose punchier (e.g., use “hosts’ ability to …” instead of “the ability of hosts to …”)
Avoid misplaced modifiers5
Always spell-check. Spell-checking won’t catch everything, and you may have to skip over a lot of technical vocabulary that’s not in your spell-checker’s dictionary, but there’s no excuse for not doing it.
Everyone overuses their favorite punctuation marks, words, or turns of phrase. Learn what yours are, and use your revision process to thin them out. (For example, I deleted many instances of the word “usually” in this document, keeping only the ones I thought were really necessary. I sometimes get carried away with em-dashes or semicolons.)6
If you’re using BibTeX, check for capitalization! Most journal styles automatically set titles in all lower-case, so you have to protect any words that should be Capitalized or CAPITALIZED with curly brackets {} (some reference managers automatically {protect} {every} {word} {in} {the} {title}, but I prefer that only necessary words are protected). I also prefer journal names to be set in title case (Journal of Important Stuff) rather than sentence case (Journal of important stuff).
You should have read, or at least looked at, every reference that you cite. Thanks to the internet it is now plausible that you found a copy of Laplace’s “Recherches sur le calcul intégral aux différences infiniment petites, et aux différences finies” (1771) online and read it in the original French, but I will definitely challenge you about it during your committee meeting or thesis defence. If you read something in translation, you should cite the translation/translator; if you feel it’s important to cite work that’s cited in another work (because it is historically important or foundational), state that explicitly (e.g. “Laplace Recherches sur le calcul … [1771], cited in Schmoo et al [2015]”).
It’s fine to cite an earlier source without reading it if you attribute it (similarly to the example in the previous paragraph, e.g. “Smith (2017), cited in in Jones (2020)”), but: (1) if the information can be considered reasonably general, it may be OK to cite the later source without giving the earlier one, (2) if the earlier source feels important enough to cite, it’s probably best to go back and read the earlier source so you can cite it directly. (It is educational and sometimes shocking to follow a trail of references back to see how information from an original source gets distorted in subsequent citations.)
I prefer author-date style citation. It is less compact than using numbered references, but it immediately lets readers who are familiar with the literature in your field know what study you’re citing, without having to skip to the reference section.
Use a bibliographic citation manager. I prefer Zotero because it is free, powerful, and convenient, but you can use anything that works for you and your co-authors. Using a citation manager means you can keep track of everything you’ve read whether or not you use it in your paper; any modern citation manager + document preparation system will make the process of switching reference formats nearly painless.
When citing scholarly work that is available online (e.g. a
peer-reviewed journal article or conference proceedings, or an article
in a preprint archive), give the full citation information, not just the
URL or DOI (digital object identifier); it’s more formal, and
readers who are familiar with the literature in your area will be able
to know what you’re citing without having to click through a link. Do
include the DOI in the citation if one is available; don’t include both
the DOI and the URL. URLs that link to the publisher’s web site are
often unreliable in the long term; URLS that include the DOI, such as
http://www.aimspress.com/article/doi/10.3934/mbe.2024321,
are redundant with the DOI itself.
Information that is only available on a standalone web site should provide information about the date when you accessed it but you don’t need to do this with journal articles or other information that has a DOI; such sources are assumed to be stable.
As George Orwell says, “Break any of these rules sooner than say anything outright barbarous.”
I’ve titled the document “peeves” in recognition of the fact that opinions about language and writing are necessarily subjective; the linguists at Language Log, who are also referred to elsewhere in this document, call prescriptivist opinions about language “peeves”↩︎
the folks at Language Log, a blog by linguists, are snarky about people who criticize the use of the passive voice when what they mean is “language that obscures agency”↩︎
My mother, Joan Bolker, wrote a book on Writing Your Dissertation in Fifteen Minutes a Day, edited an anthology of essays about writing called The Writer’s Home Companion, and gave advice for medical researchers and writers in Writing Medicine. All are highly recommended, but they focus on the process of writing rather than on mechanics like word choice and structure.↩︎
This advice contradicts the previous advice about agency, attributing ideas to particular actors. “A foolish consistency is the hobgoblin of little minds” (Emerson: Wikipedia points out that this is a misuse of the original quotation, but I like it anyway.) Or: “Do I contradict myself? Very well then I contradict myself, (I am large, I contain multitudes.)” Whitman, Song of Myself↩︎
this example from a student paper: “When delving into the complex life cycles of the lamprey, female adult lampreys will lay thousands of eggs” is sort of a misplaced modifier, but an odd one. I think it’s actually a vague connector gone wrong, i.e. the author meant something like “Now let’s delve into the complex life cycle of the lamprey. Female adult lampreys will …”↩︎
I also like footnotes, but they’re rarely used, and rarely a good idea, in scientific writing. (They’re common in the humanities.)↩︎