Monthly Archives: March 2015

What style of language do scientists really prefer?

“Our readers are intelligent, well-educated scientists. Why should we make our language dumbed-down, patronising and imprecise in the name of ‘readability’?”

It’s a fair question. Here’s the answer.

Never talk down to your readers. But never waste their time, either. And scientists, while intelligent and educated, are also busy. As well as their research, they may have teaching, management or clinical duties to perform, funding applications to write, presentations to plan, journals to keep up to date with… They don’t have time to wade through verbiage in search of facts.

If you’re writing about something complex, then of course you need to give all the necessary detail. If you’re writing for specialists, you can use their specialist terms. But you don’t need to add verbal complexity beyond that. Keep it clear and direct. This makes your writing more efficient and more likely to succeed in communicating your message. It’s also courteous to your readers.

Einstein may or may not have said, “Everything should be made as simple as possible, but no simpler.” Either way, it’s good writing advice.

But we don’t all follow it.

Many people, when writing in a work context, feel that they’re supposed to use language that is more abstract, impersonal and convoluted than they otherwise would. The idea is that this makes the writing sound more professional. But the result is often that it’s unclear and off-putting, even to highly intelligent readers.

Language like that can be made more concrete, more personal and more concise without dumbing down the content, without losing important information, and without making the tone inappropriately casual. In fact, directness and clarity normally sharpen the tone and can even help to add precision. Clearing up overgrown language can show you previously hidden patches of ambiguity.

Testing the hypothesis

Scientists like evidence, so let’s have some.

John Kirkman, as part of his book Good Style: Writing for Science and Technology, ran several surveys of scientists.

In one, he gave two versions of a passage from a research paper to 2,781 readers from medicine and the life sciences:

Brown’s version

In the first experiment of the series using mice it was discovered that total removal of the adrenal glands effects reduction of aggressiveness and that aggressiveness in adrenalectomized mice is restorable to the level of intact mice by treatment with corticosterone. These results point to the indispensability of the adrenals for the full expression of aggression. Nevertheless, since adrenalectomy is followed by an increase in the release of adrenocorticotrophic hormone (ACTH), and since ACTH has been reported (Brain, 1972) to decrease the aggressiveness of intact mice, it is possible that the effects of adrenalectomy on aggressiveness are a function of the concurrent increased levels of ACTH. However, high levels of ACTH, in addition to causing increases in glucocorticoids (which possibly accounts for the depression of aggression in intact mice by ACTH), also result in decreased androgen levels. In view of the fact that animals with low androgen levels are characterised by decreased aggressiveness the possibility exists that adrenalectomy, rather than affecting aggression directly, has the effect of reducing aggressiveness by producing an ACTH-mediated condition of decreased androgen levels.

Smith’s version

The first experiment in our series with mice showed that total removal of the adrenal glands reduces aggressiveness. Moreover, when treated with corticosterone, mice that had their adrenals taken out became as aggressive as intact animals again. These findings suggest that the adrenals are necessary for animals to show full aggressiveness.

But removal of the adrenals raises the levels of adrenocorticotrophic hormone (ACTH), and Brain2 found that ACTH lowers the aggressiveness of intact mice. Thus the reduction of aggressiveness after this operation might be due to the higher levels of ACTH which accompany it.

However, high levels of ACTH have two effects. First, the levels of glucocorticoids rise, which might account for Brain’s results. Second, the levels of androgen fall. Since animals with low levels of androgen are less aggressive, it is possible that removal of the adrenals reduces aggressiveness only indirectly: by raising the levels of ACTH it causes androgen levels to drop.

Kirkman sums up the main differences:

  • shorter, less complex sentence structures in Smith’s version
  • all necessary specialist terms are present in both versions, but Smith uses familiar words in place of unnecessary specialist terms
  • fewer passive structures and less roundabout phrasing in Smith’s version
  • paragraphing.

The scientists were then asked what they thought of the two versions.

“Which style do you prefer to read when you read scientific texts?” 74% picked Smith and 21% Brown.

“Which style do you think is more appropriate for scientific texts?” 57% picked Smith and 25% Brown.

Clearly the Smith style was much preferred, but the disparity between these two answers is interesting. A significant minority, even though they favoured Smith’s style, were still reluctant to endorse it for professional use. Presumably they were worried that their peers didn’t share their preferences. But, the survey shows, most did (they also rated Smith’s version easier to read and more precise, and rated Smith as being more objective and having the better-organised mind).

This is a case of what Steven Pinker calls pluralistic ignorance:

a false consensus, in which everyone is convinced that everyone believes something, and believes that everyone else believes that they believe it, but in fact no one actually believes it. One example is the cachet that college students place on drinking till they puke. In many surveys it turns out that every student, questioned privately, thinks that binge drinking is a terrible idea, but each is convinced that his peers think it’s cool.

Convoluted prose isn’t so literally sickening, but many academics and journals still maintain the illusion that it’s cool.

Kirkman also ran other surveys, giving different versions of specialist texts to chemical engineers, ecologists and biochemists. The results were similar. He lists the features of the preferred style:

Direct, verbs mainly active, minimum of special vocabulary, judicious use of personal and impersonal constructions, sentences of varied length but mainly short and not complex.

How to do it

For advice on how to improve convoluted writing, you could try Kirkman’s book, which focuses on scientific writing. Joseph William’s book Style: Toward Clarity and Grace is very good and has a broader remit.

If you want something a bit shorter, Rachael Cayley’s Explorations of Style blog has some useful posts on that draw on Williams, among others, with a focus on academic writing. Look in particular at the ‘five key strategies’ posts listed on the right.

Pinker’s The Sense of Style I also recommend, particularly chapters 2–5 (don’t get too bogged down in the sentence diagrams). He too draws on Williams, as well as bringing in some psychological research on how we process language. He’s an engaging writer, if sometimes a bit combative.

I might try to write my own short guide one day. For now, I’ve done a case study [PDF] of a passage of text that isn’t from a journal article but is written for scientists. I’ve tried to improve it – cutting it in half but keeping the information, expressed more clearly and directly – explaining the changes as I go.

What the internet desperately needs is another blog post about the Oxford comma

The Oxford comma – also known as the Harvard comma, the serial comma, and the what the hell is wrong with you people why can’t you just get a life – provokes strong opinions.

It’s the difference between these two sentences:

1a) I ordered bacon, eggs and beans.
1b) I ordered bacon, eggs, and beans.

The Oxford comma is the last one, before the “and” in version 1b.

Should it be there?

Some people say no (loudly): it looks fussy and slows the sentence down. The “and” is quite enough to separate the last two items in the list.

Other people say yes (even more loudly): it’s helpful for clarity. Well, maybe not in this case, but it’s more important when the individual items in a list are grammatically more complex, especially if they contain “and”s. Compare:

2a) They ordered bacon and beans, chips and eggs and toast.
2b) They ordered bacon and beans, chips, and eggs and toast.
2c) They ordered bacon and beans, chips and eggs, and toast.

Sentence 2a, without the Oxford comma, is ambiguous about which ingredients make up which meals: it could mean either 2b or 2c.


3a) The hallway contains a chest of drawers, a mirror, a sculpture made of copper wire and a hatstand.
3b) The hallway contains a chest of drawers, a mirror, a sculpture made of copper wire, and a hatstand.

While 3a may take a moment to decipher – is the hatstand part of the sculpture? – 3b makes it clearer.

The Oxford comma can also help in cases like these well-known examples:

4a) We invited the strippers, JFK and Stalin.
4b) We invited the strippers, JFK, and Stalin.

5a) This book is dedicated to my parents, Ayn Rand and God.
5b) This book is dedicated to my parents, Ayn Rand, and God.

Here the problem is confusion between different types of comma. In 4a, is the comma after “strippers” there to keep items separate in the list or to introduce the extra information (an appositive phrase) that JFK and Stalin are the strippers?

Likewise, 5a might suggest that Ayn Rand and God are my parents. But 4b and 5b make the separation clear.

Now, sure, there’s no real danger of misunderstanding in these two cases; rather, the risk is of a brief sense of absurdity. There are more sensible examples, though:

6a) I asked my neighbours, an architect and a builder.
6b) I asked my neighbours, an architect, and a builder.

So the Oxford comma can be useful. But given that it hardly seems necessary in “I ordered bacon, eggs, and beans”, should we use it all the time?

The AP Stylebook says no, recommending it only in more complex or potentially ambiguous cases: “do not put a comma before the conjunction in a simple series: The flag was red, white and blue. He would nominate Tom, Dick or Harry.”

But the Oxford Guide to Style says yes: “Given that the final comma is sometimes necessary to prevent ambiguity, it is logical to impose it uniformly”. The Chicago Manual of Style also favours using it consistently, as do Bryan Garner, Steven Pinker and others. (It’s more common in American writing than in British.)

I agree that consistency is good – other things being equal. But other things are not equal. There are disadvantages to using the Oxford comma.

It can slow a sentence down. This is obviously subjective and depends on what you’re used to, but I find “I ordered bacon, eggs, and beans” pretty ponderous and deliberate.

The New Yorker’s Mary Norris has taste that goes the other way, favouring the Oxford comma everywhere: “I’ve gotten used to the way it looks. It gives starch to the prose, and can be very effective.” Well, if starch is what you want…

She adds:

The New Yorker isn’t asking you to pause and gasp for breath at every comma. … The commas are marking a thoughtful subordination of information. … It looks choppy, but you don’t have to chop it up when you read it.

Sure, this is thoughtful (it’s in a great extract from what looks like being a great book by Norris), but it makes two huge assumptions.

It assumes that readers consciously choose how commas come across to them. And it assumes that readers will understand the publication’s policy on commas and the reasons behind it. Even for the New Yorker, that feels like a stretch.

As a copyeditor, I’m a big fan of the fine distinctions copyeditors fret over. But we have to have some perspective about whether our readers understand those distinctions the same way we do – or even at all. Sometimes we might be zealously and ingeniously splitting hairs that are invisible to the untrained eye.

There’s a second, more serious problem with the Oxford comma: sometimes it creates the very ambiguity or absurdity that it’s supposed to remove. I’m amazed that its partisans so rarely acknowledge this, because you only need to tweak their examples slightly to see it:

7a) We invited the stripper, JFK and Stalin.
7b) We invited the stripper, JFK, and Stalin.

8a) This book is dedicated to my mother, Ayn Rand and God.
8b) This book is dedicated to my mother, Ayn Rand, and God.

Here the tables are turned: 7b might suggest that JFK was the stripper, 8b that Rand was my mother. 7a and 8a are clear.

So we have to choose whether to use an Oxford comma or not in each case. A blanket policy, pro or anti, just won’t work.

And it gets worse. Try this pair:

9a) I spoke to my uncle, a magistrate and a priest.
9b) I spoke to my uncle, a magistrate, and a priest.

Neither is clear. Does 9a mention three people or one? Does 9b mention three people or two? We need to rephrase somehow:

9c) I spoke to my uncle, who is a magistrate, and a priest.
9d) I spoke to a priest and my uncle, a magistrate.
9e) I spoke to my uncle, who is a magistrate and a priest.
9f) I spoke to a magistrate, a priest and my uncle.
9g) I spoke to my uncle as well as a magistrate and a priest.

We have to face the awful truth: the Oxford comma is not a magical blade that can chop any sentence into slices of perfect meaning. It’s just one fallible tool among many.

Use it when you must, avoid it when you must, choose as you prefer (or as your readers will prefer) when you can, and rewrite whenever that would be better.

Oh, and try not to get too worked up about it.