Two weeks ago, I reported my run-in with a reviewer who wanted me to scrub common English contractions (like it’s, doesn’t, or we’re) from a manuscript. There’s a common belief that contractions mustn’t be used in scientific writing, although the genesis of this “rule” is unclear. So is the rationale. One that’s commonly suggested is that contractions make writing informal, and that that’s inappropriate – to which I say only “Harumph”. Another is much more important: the claim that they make writing less accessible to readers of English as an additional language (EAL).
I’ve been skeptical of that hard-for-EAL claim, but not being an EAL reader myself makes it hard for me to claim authority on the issue. So, I asked EAL readers of Scientist Sees Squirrel to weigh in – and they did. Today, poll results, and a couple of additional points raised by some folks who think about writing for EAL readers.
The most important result comes first. At least in my informal poll results, EAL readers report little difficulty with contractions:
There is some very modest evidence that contractions may be a larger issue for some EAL readers early in their learning of the language:
Reports that contractions were a bigger obstacle early on are no surprise – I would assume that everything is a bigger obstacle early on. But there were only 5 such votes (6.5% of n = 77 votes). Unfortunately, I didn’t allow (and probably couldn’t have allowed) respondents to quantify the magnitude of the difference.
Contractions are not, of course, the only thing an EAL reader might find difficult about reading in English. So another way to approach the question at hand is to ask respondents to prioritize the writing changes they’d like to see:
There was again a single vote for getting rid of contractions (1.2% of n = 87 votes), but it was swamped by votes for other changes, especially shortening and simplifying sentences. This is, I suspect, something on which EAL and English-native readers agree*!
Now, internet poll results are pretty terrible, as data go. So it’s important to think about ways that today’s data might be misleading. It’s especially important to consider effects of self-selection in the respondent population, and I can think of two ways that might matter here.
- First, the earliest and thus least practiced EAL readers are probably under-represented in the data, given that they’re less likely to visit a blog written in English. This could lead poll results to understate the difficulty of contractions – but only if the more seasoned EAL readers who do visit are misremembering or misreporting their own earlier experience. That was the point of the second question, really; and only a modest fraction of respondents said contractions were more difficult early in their learning.
- Second, opt-in polls tend to over-represent respondents who feel strongly about the issue. Since there probably aren’t a lot of readers – EAL or otherwise – who are in love with contractions, I’d suspect over-representation of respondents who hate them. But the actual results include so few of those that this bias doesn’t seem to have been a problem
So although I’ll cheerfully grant you that my poll data aren’t of publishable quality, my best guess is that they aren’t desperately misleading, either.
What about English vs. other languages? Are English contractions particularly difficult?
It doesn’t seem that way. If anything, contractions might be a little bit easier for EAL readers learning English than for English speakers learning other languages – but the sample size is small. Interestingly, 11 English-native respondents (28% of n=39 votes) report that their new language doesn’t even use contractions. This is quite implausible. All languages probably use contractions; it’s just that some of them don’t mark them with punctuation like apostrophes. So in a sense, in many languages contractions might be said to be too easy to even notice.
Finally: an additional perspective on contractions that I’d been thinking about, but hadn’t really managed to crystallize until some reader feedback helped me think it through. There are two ways to think about reading contractions. One is to believe that they’re an additional layer of complexity, a meta-construction that a reader must tackle in two steps: first decoding the contraction back to its constituent words (isn’t → is not), and then reading and understanding those constituent words. That does seem like it could be a burden (although perhaps a small one, since most of our common contractions are regular in construction). But the other way to think about contractions is as new words, shorter synonyms for their underlying phrases (isn’t = is not). Under this view, a contraction needn’t be decoded; it’s simply learned like any other word. (We learn synonyms all the time without worrying about it.)
If the shorter-synonym mechanism is how common contractions work for a new learner of a language, then at worst the use of common English contractions would require the addition of about 40 new words to an EAL reader’s vocabulary. But Iva Cheung, via Twitter, points out something I hadn’t thought of: actually, contractions are highly unlikely to be new words to these learners. That’s because the 35 most common contractions are already among the 1,000 most common English words** (using the contemporary-fiction word frequency list used in xkcd’s UpGoerFive cartoon and the text editor based on it). Even if we reduced our scientific-writing vocabulary to the 1,000 most frequent English words, we could still use contractions without placing a significant burden on readers. And while all kinds of people, including me, think the scientific literature could use a healthy does of writing more plainly, I’ve never seen any serious proposal to cut us down to the 1,000 most common words. (Actually, I’m not sold on the utility of that even for outreach writing, unless it’s nothing more than a fun mental exercise.)
So: I left the contractions in my manuscript, and if you want to do the same, you can point to this post as evidence that it’s probably not your worst sin against readability. Not even close.
© Stephen Heard November 6 2018
*^Which doesn’t stop it from being my own #1 writing problem. I love long sentences, with semicolons and parentheses and subordinate clauses and every other structure I can find to load them up with syntactical complexity. At least, I love writing them. Like most other readers, I dislike reading them, unless they’re meticulously crafted. When I revise, I target reductions in sentence length and complexity. I usually fail.
**^Beginning with I’m at #58. This is practically a first-day-of-lessons frequency of use, which I point out only because practically is #1,000 and I’m weird that way.