I said in the last post that I wanted to impart some tips on writing which have been very useful to me. These are all taken from a great article, The Science of Scientific Writing, by George Gopen and Judy Swan.
Gopen & Swan’s big theme is ‘reader expectations.’ Readers, it turns out, don’t just suck up all the words in a text indiscriminately, scramble them all in a mental blender, and absorb the resulting purée as pure information. Instead they look for a running story in any stretch of prose. And, semi-consciously, they – we – look for clues to that story in the structural choices made by the writer. I’ll explain what this really means below.
The article, as the title implies, is largely directed at scientists who have to write about their work. I think the advice it gives is fairly broad, but it’s still focused at just one facet of writing: trying to communicate information clearly. This is obviously a pretty worthy goal, but it’s not always a text’s primary goal. Sometimes one writes to persuade readers. Or to entertain them; or to impress them. Juggling such diverse aims always involves trade-offs.
With that caveat, here are some basic reader expectations to keep in mind when writing.
1. There must be a verb around here somewhere
Well-formed English sentences have a verb.* In the previous sentence, the verb is ‘have.’ The verb most commonly follows the subject of the sentence (the noun or noun phrase which the sentence is ‘about’). G&S note that readers tend to get antsy if the subject is separated from its corresponding verb by a lot of words. Consider:
The theoretical model, which had seven adjustable parameters, two of which were determined by best fit using a process which will be explained in further detail in the next section but does not significantly affect the conclusions, failed to explain the results.
The subject (‘The theoretical model’) and its verb (‘failed to explain’) are kept apart like forlorn lovers. A sentence like this is a headache to read. The reader has to keep the subject in mind until the verb shows up, making it very difficult to attend to intervening information. And until the verb appears, it’s hard to begin to interpret the sentence.
The other problem with constructions like this is to do with reader expectations. Material falling between the subject and the verb is regarded as less important by the reader. Sometimes this is fine. Some material is less important. There’s no harm in the occasional aside, but keep it brief.
Note that, although the example sentence is somewhat long, length itself is not the problem. The problem is its pathological grammatical structure.
2. The ‘topic position’
A sentence is like a story: the most important parts are the beginning and the end. These are where the reader is most attentive.
The beginning of a sentence, dubbed the ‘topic position,’ tells the reader what the sentence is going to be about. Consider the following sentences:
Latin America has made a religion out of football.
Football has been made into a religion in Latin America.
Both sentences contain the same information; but the first is a story about Latin America, and the second is a story about football.
Readers look for context and ‘old information’ at the beginning of the sentence. The topic position should link backwards to ideas you’ve already introduced. The first sentence might be appropriate in an article about Latin America, the second in an article about football.
3. The ‘stress position’
At the other end of the sentence, readers are ready to be dazzled. Gopen & Swan note that a reader will naturally pause a little as they reach the end of a sentence, mentally ‘taking a breath’ before the next one. They call it the ‘stress position.’ This is the best place to introduce new, important information. (Conversely, if you place minor information in the stress position, readers are likely to overemphasise its importance at the expense of your key points.)
Endings are inherently memorable. Use them wisely. Of course, ideas introduced in the stress position of one sentence, can be used as context in the topic position later on.
4. Fill in the gaps
Writing is hard because other people don’t know what we know ourselves. This is very evident when a passage of writing exhibits an apparently illogical structure. Often when this happens, there is an underlying logic that the writer is aware of, but has failed to articulate.
Consider the toy example:
Blue writing is hard to see on a blue background. We have a can of paint in the shed. We shouldn’t use the paint to write our name on the mailbox.
There’s obviously some missing information here! The final conclusion doesn’t follow at all from the information provided.
Let’s fill in the gaps:
Blue writing is hard to see on a blue background. We have a can of paint in the shed. The paint is blue. The mailbox is blue. We shouldn’t use the paint to write our name on the mailbox.
There’s now an identifiable chain of reasoning here. It still sounds kinda weird, though. Nothing draws attention to the logical structure interlinking the ideas in the passage.
Here’s the final version:
Blue writing is hard to see on a blue background. We have a can of paint in the shed, but the can is blue. Since the mailbox is also blue, we shouldn’t use the paint to write our name on the mailbox.
I’ve bolded the ‘logical’ words. These extra words don’t add to or alter the claims being made. (Blue writing hard to see, paint in shed, shouldn’t use, etc.) They just highlight the logical relationships between the claims. They turn a sequence of statements into an argument.
Of course, no one is likely to make these mistakes with paint cans and mailboxes. It’s quite common, however, with the sorts of complicated and subtle arguments that appear in scientific writing, and indeed in other academic domains.
In editing other peoples’ writing, I’ve noticed both types of omissions, both of whole units of information, and of logical linkages. The former tends to occur when the writer has elided a piece of information that is so ‘obvious’ to them, they forget that most of their readers won’t have it. (This is a seemingly universal professional hazard.) The latter suggests that the writer hasn’t yet figured out all the logical connections, or they’re not confident in their interpretation. It most often happens when they’re presenting a novel argument – their own argument, as opposed to restating conventional wisdom.
It’s difficult to repair these types of text problems, because it requires you to critically inspect your own thinking: to ask yourself what exactly you think, and why you think it. The flipside: it’s a great opportunity to pose exactly those questions. I find that writing is a uniquely valuable tool for uncovering logical and causal relationships. Human brains are adept at making connections between concepts; but a lot of the time, particularly for more abstract ideas, the default connection is, ‘this thing has something to do with that thing.’ Putting a logical argument to paper is an effective way of exposing those weak connections and articulating what the ‘something’ is.
5. Use a hard-working verb
As we noted, English sentences generally have a verb. Readers expect to learn what’s ‘going on’ in a sentence from the verb. A problem which is endemic in academic writing styles is to violate that expectation by using an uninformative verb.
Corrosion occurred of the metallic piece by the acidic solvent.
This sentence is harder to read than it needs to be. The verb, ‘occurred,’ tells us… well, that something occurred. You can’t get much more uninformative.
When the action of the ‘story’ is captured in the verb, the sentence is much easier to parse:
The acidic solvent corroded the metallic piece.
The second version looks much more like natural speech. So why is the ‘weak verb’ formation so popular? Maybe because it looks more ‘serious.’ I suspect it also makes it easier to elide the logical connections between semantic units. Uninformative verbs lend themselves readily to sentences like this:
When the metallic piece and acidic solvent were brought into contact, corrosion occurred.
Now it’s never spelled out explicitly that the acid is acting upon the metal. This sort of semantic ambiguity is bad for the reader, but makes life easier for the writer.
Whenever you see an uniformative, generic verb, ask if you can express the action in the sentence with a more specific verb. ‘Occurred’ and ‘took place’ are frequent offenders. ‘Is’ and ‘has’ are also often substituted where a better verb could be used. The other type of telltale to look for is nominalizations: nouns derived from a root verb. (Eg, corrosion, deforestation, inhibition, submission, and many many others.) Weak verbs and awkward nominalized nouns have a symbiotic relationship, and are often found together.** Often the solution is to unpack the nominalization into its underlying verb, as we did above.
I feel like I should soften this advice. Ambiguity is not inherently a bad thing. At the frontiers of knowledge, it comes with the territory. That’s all the more reason to eliminate unnecessary ambiguity. As Einstein probably wouldn’t have said: be as explicit as possible, but no expliciter.
That was longer than I expected! If you want to know more, I highly recommend Gopen & Swan’s paper. They take the time to walk through real examples of convoluted scientific prose, and subject them to a kind of ‘extreme makeover’ according to the principles above. The difference in clarity is staggering; if every scientist had a little Gopen and a little Swan sitting on their shoulder, reading journal papers would be much less of a chore.
One important thing: I don’t think you should worry too much about these rules, or any other rules, on the first draft. Get your thoughts down in a way that makes sense to you. Then, use these principles, and the overarching principle of respecting the reader’s natural expectations, to organise your written thoughts in a way that will make sense to other people.
*With the exception of the occasional short sentence fragment, whose meaning is obvious from the surrounding context. (Like this.)
**Because I don’t want to be held responsible for releasing a mischievous overgeneralization into the wild, let me point out that nouns formed from verbs are by no means all bad. Indeed, they’re staples of everyday English usage. But there is a definite tendency in academic writing to use them excessively and unnecessarily.