Years ago now, I recall a senior scientist who'd read a piece of my work saying doubtfully, 'It's very well-written. Very literary.' The implication was clear: 'But is it good science?'
With hindsight, I can agree: that particular effort wasn't ace. I can't even remember its title, and the obscurity's well-deserved. What did stick was my surprise that my colleague (undoubtedly a good scientist) saw good science and good writing as not just independent, but even perhaps opposed, since science is all about precision and language is irretrievably vague.
Years later, I still have a problem with this.
The problem for science is that the vast majority of scientific communication, even within the fraternity, uses language, not mathematics. The better they write, the clearer their message. The more equations, the fewer people read them and cite them, at least in my research field, biology. Science, as we keep being told, needs good communication.
If you think language is vague, you may (incorrectly) infer that how scientists write their abstracts and articles doesn't matter. It does. A poorly-written abstract can be ambiguous, confusing and at times incomprehensible. Yet in science training, writing skills can be downplayed, or even seen as suspicious. Media dons have good writing skills, you know.
Hmm. Sure, language isn't mathematics. Thank goodness, it's much richer and more flexible. But even though individual word meanings aren't rigidly defined, they can be combined with surprising subtlety. A simple example: Microsoft Word assures me that 'pleasant' and 'pleasurable' are synonyms -- i.e. they mean the same thing -- but they don't. Saying to your lover 'Sex with you's so pleasant' is less complimentary than 'Sex with you's so pleasurable'. For another example, think about the difference between 'Microsoft Word assures me' and my first choice, 'Microsoft Word tells me', when it comes to deciding how much faith to place in MS Word.
A science example
Here's an instance from the research literature: an abstract of a 2012 article by Dannlowski et al. in Biological Psychiatry. It's not a bad abstract; I've come across far worse. It does however show how language can affect understanding.
Childhood maltreatment represents a strong risk factor for the development of depression and posttraumatic stress disorder (PTSD) in later life. In the present study, we investigated the neurobiological underpinnings of this association.
Nice clear opening statement. But ... which association? Between maltreatment and depression, maltreatment and PTSD, or both? I'm already unclear what's being argued here; depression ≠ PTSD.
Hmm. So I'll guess it means both.
Since both depression and PTSD have been associated with increased amygdala responsiveness to negative stimuli as well as reduced hippocampal gray matter volume, we speculated that childhood maltreatment results in similar functional and structural alterations in previously maltreated but healthy adults.
So depression and PTSD are associated with a) more amygdala responsiveness (a functional difference) and b) less hippocampal gray matter (a structural difference), but it's ambiguous as to whether a) and b) are themselves associated.
One hundred forty-eight healthy subjects were enrolled via public notices and newspaper announcements and were carefully screened for psychiatric disorders.
The word 'carefully' here has the unintended effect of making me wonder why the authors felt they needed to put it there. Who writes, 'Our screening was pretty crap, we just ticked a few boxes'? I'm now a smidgen less inclined to trust this research.
Amygdala responsiveness was measured by means of functional magnetic resonance imaging and an emotional face-matching paradigm particularly designed to activate the amygdala in response to threat-related faces. Voxel-based morphometry was used to study morphological alterations.
What sort of threat? Anger, disgust, a knife gripped between the teeth? And -- morphological alterations? Why not say 'gray matter changes' if that's what you mean? Or 'structural changes', to make the link with earlier sentences transparent. It's also easier to read. I like the idea of an emotional paradigm, though, stuck in the corner sulkily matching faces.
Childhood maltreatment was assessed by the 25-item Childhood Trauma Questionnaire (CTQ). We observed a strong association of CTQ scores with amygdala responsiveness to threat-related facial expressions.
Right. This tells me the functional result, and now I want to find out more about the control conditions, and I'm curious as to whether the CTQ's any good, so I'm more likely to read the article. Now for the structural result:
The morphometric analysis yielded reduced gray matter volumes in the hippocampus, insula, orbitofrontal cortex, anterior cingulate gyrus, and caudate in subjects with high CTQ scores. Both of these associations were not influenced by trait anxiety, depression level, age, intelligence, education, or more recent stressful life events.
Both of what? Guys, you lost me. I'm reading about associations between high CTQ scores and five brain regions. So -- retracking here -- maybe 'both' means 'both the functional and structural associations'? And what does 'depression level' mean -- now, over the past year, or what? Too much neurotransmitter got used decoding these sentences.
Childhood maltreatment is associated with remarkable functional and structural changes even decades later in adulthood. These changes strongly resemble findings described in depression and PTSD. Therefore, the present results might suggest that limbic hyperresponsiveness and reduced hippocampal volumes could be mediators between the experiences of adversities during childhood and the development of emotional disorders.
I thought participants were screened to be free of such disorders? Or are they just not reaching the required clinical thresholds? And, eek, that stretch from 'amygdala' to 'limbic'!
All in all, I'm confused, and that makes me feel stupid, and as if I've already wasted too much time on this abstract. There's an interesting-looking piece on reconsolidation in the same table of contents. I think I'll read that instead ...
In the fierce competition for readers' attention, clarity is crucial -- and clarity comes from using language well. Ease of reading is also important, because I (and I suspect most readers) are less inclined to struggle through an abstract, let alone a whole paper, if the authors make it hard work. I like to read in a smooth, fast flow, sucking up meaning without having to stop and think about how to interpret every sentence; if I wanted that kind of intellectual workout I'd read French. Translating from science into English can be quite hard enough even if the writing's limpidly transparent, given the jargon and surfeit of acronyms.
Of course, language can also be used to create deliberate ambiguity, as in the apocryphal job reference, 'You would indeed be lucky to get this person to work for you'.
Words can also be used to trigger particular ideas without saying so explicitly, as in 'that particular effort wasn't ace' (implying that all the others were). These and other skills come with good writing, and allow the writer to be remarkably adept at putting across ideas. Indeed, one problem with mathematics is that it can convey a false sense of clarity and hence security, as all those failed economic models should have taught us.
Even foggy language, as in government or scientific jargon, can be skilfully and carefully unclear if the aim is to obfuscate -- as when a politician claims to have refuted an argument (destroyed its force) when all they've done is rebutted it (argued against it, unsuccessfully in the case I'm remembering).
Then again, it may just be that the person is inept, or isn't quite sure what exactly they're trying to say. Writing is a great way to sort out thoughts. And a skilful writer has more armour against the linguistic trickery of others, another reason for learning to write well.
On a different level from the slightly prickly abstract above is simple abuse. Language-lovers get extremely frustrated by common linguistic slopperies, like the following horror from the BBC, who should know better:
(of a dead soldier) 'if he'd had the equipment, he may have survived'
NO! 'may' ≠ 'might'! The whole point is he DIDN'T survive! I hope the grieving family didn't hear that insensitive report.
Language-lovers understand that putting words in order can be a counterproductive mess, a workaday chore, or a useful and beautiful craft. Like athletes, they aim to use their skills as best they can because, basically, why settle for mediocre when you can do better? That athletes are more admired is unfortunate for writers trying to make a career of it, but like science, writing's more a compulsion than a career choice. Seeing the gift of language wasted makes skilled wordsmiths wince, just as a pianist might wince on seeing a Steinway grand being bashed about. Why have such a powerful language-enabled brain if you're going to use its powers so poorly, just to get by, leaving yourself open to manipulation and blind to the fabulous riches of your linguistic heritage? Besides, as hobbies go wordplay's a good choice for these fearful and thrifty times: safer than swordplay, easier than foreplay, and cheaper than gameplay.
Setting feelings aside, however, clear and fluent language is hugely important to communication. For that reason alone it's worth trying to do it better. In science, as elsewhere, good writing matters.
Copyright Kathleen Taylor (neurotaylor) 2013. This post was first published on neurotaylor.com in January 2013.