Does thinking like a scientist lead to bad science writing?

In his book A Short Guide to Writing About Science [1], David Porush suggests that the mindset useful for doing science isn't always the best mindset for communicating science. (It's more than a suggestion, actually -- the second chapter of the book is titled "Why Good Scientific Thinking Can Lead to Bad Science Writing.") Since it's connected to our prior discussion of ambiguous scientific writing, let's have a look at Porush's diagnosis of bad science writing and the ways he thinks it could be better.

Porush notes that most people learn to write scientific papers by reading a whole mess of scientific papers and trying to imitate their style. Unfortunately, this process seems to entrench a lot of bad habits (as Porush sees them) that are treated almost as commandments, or as Porush describes them (on page 11):

The Idols of Science Writing

  • Complex sentences are more precise and intelligent than simple ones.
  • Passive voice is more objective than active voice because active voice involves stating the agent, who is often a human, and humans introduce subjectivity.
  • Descriptive prose is imprecise and unscientific because descriptions rely on adjectives which are often unquantifiable.
  • Claims about scientific fact must always be cautious; thus, using very cautious language -- squirrel language -- in scientific prose is always prudent.
  • Be as wary of innovation or taking chances in language as you are in the lab.
  • Revision is unnecessary because knowledge is transparent.

If you think these rules define good scientific writing, you will strive not to violate them. However, Porush argues that worshiping at their altar can lead to sins against clear communication.

Take the first idol, the preference for complex sentences rather than simple, direct ones. If you read enough scientific papers with convoluted sentences, you might get the impression that this is what an expert sounds like -- so, if you want to be taken seriously, you should write this way, too. Alas, this seems to be the same sort of phenomenon one sees sometimes at scientific talks. The speaker gives an impenetrable talk and the audience members murmur to each other, "I didn't understand that at all; he must be really smart." Why isn't the overwhelming response, "I didn't understand that at all; this guy can't communicate clearly to save his life"? If the goal is really the communication of science, "sounding smart" by being impenetrable is a bad strategy. Porush writes:

... the real basis of your authority as a writer is the work you've performed as a scientist; the research you've done in collecting data, analyzing it, interpreting it, and matching it to current theories or knowledge; and how well your writing explains this work to your audience. (p. 12)

In other words, let your content do the talking rather than using your complex prose as an "expert" disguise.

The second and third idols (favoring the passive voice and eschewing description) both stem from an effort to be objective. Striving for objectivity is a good thing in science. Writing as if you individually have achieved it -- as if you have somehow become a transparent eyeball delivering an inerrant view of How Things Are -- is not such a good thing. If I may quote myself in an earlier post:

While some thinkers have framed the problem of building objective knowledge as one that depends on each individual scientist being highly objective and switching off his or her own biases, others (including Frederick Grinnell and Helen E. Longino) have put the burden of objectivity primarily with the community -- bias is stripped out of what ends up being identified as scientific knowledge when the community "checks the work" of individual scientists within it.

If it really is the coordinated efforts of a community that makes scientific knowledge more objective, then trying to disappear from the account of what you did, what you measured, and how you interpreted it is unnecessary. Your fellow scientists know that the data wasn't left on the lab bench by elves; it was obtained through the activities of actual human beings. Better would be to use your language to communicate clearly what you did, how you did it, what you observed, and so forth to scientists who were not there so they can reason about your experiment and interpretations, too.

And what of the "squirrel" words? Porush recognizes that conditional language is a natural response when your scientific training cautions against overgeneralizing from your data, overlooking other interpretations of your results, or forgetting that your inferences rest on assumptions that may themselves turn out to be wrong. (Also, you don't want to put forward a bold claim that you may have to retract later.) He isn't advocating that scientific writing fall to the level of ad agency puffery, but he does claim that scientific writing can overdo the caution to the point of making claims unnecessarily weak and vague. Porush thinks scientists are better off explaining their conditional claims -- the specific assumptions they're making, the limits to the conclusions that should be drawn from the findings, etc. I'd throw in that seeing science as a cooperative project that requires communication (rather than a contest to see which scientist can score the most points) would encourage more clarity about what we think we know and what we think is still uncertain. Being vague about that may work to protect your reputation, but not to get the community to a point of fuller understanding efficiently.

Indeed, Porush seems to think that the community-level process of building more objective knowledge depends on the individual scientists acting more like individuals. What I mean by that is the scientists have to make reports of how the bit of the world they're investigating seems to them -- reports that are clear enough that other scientists can understand them -- and trust the engagement of the community as a whole with these individual reports to result in a more objective account of that bit of the world. Porush writes:

Science writing is not as procedural as the scientific method itself. There are other parts to doing science: drawing conclusions, analyzing the data, speculating about hypotheses, synthesizing all you know to arrive at a theory. In all of these, there is a role for persuasion; for advocating one view or interpretation over another. There is certainly a need for clarity, for an absolute commitment to being understood by your audience without ambiguity. The writer interested in communication -- and this is especially true for the scientist-writer -- must make sure the audience receives the message. (pp. 19-20)

Taking a stand -- saying here's what I did, here's what I saw, here's what I concluded from that, and here's why that conclusions seems more reasonable to me than this other one -- is a good thing in a scientific paper. It makes it easier for other scientists to try to replicate or extend your work to see whether they get similar results and draw similar conclusions. And it may even make it easier for other scientists to identify your assumptions (and blind spots) and thus to challenge them -- which, in the end, should make the scientific community's body of knowledge less biased and more objective.

If the scientific literature is really supposed to be part of an extended conversation among scientists, trying to make it clearer and less painful to read only makes sense.

-----

[1] David Porush, A Short Guide to Writing About Science. New York: HarperCollins College Publishers, 1995.

More like this

On the post in which I resorted to flowcharts to try to unpack people's claims about the process involved in building scientific knowledge, Torbjörn Larsson raised a number of concerns: The first problem I have was with "belief". I have seen, and forgotten, that it is used in two senses in…
I'm late to this round of the discussion about scientists and journalists (for which, as usually, Bora compiles a comprehensive list of links). The question that seems to have kicked off this round is why scientists are sometimes reluctant to agree to interviews, especially given how often they…
Today, Inside Higher Ed has an article about the recent decline of peer reviewed papers authored by professors in top five economics departments in high profile economics journals. A paper by MIT economics professor Glenn Ellison, "Is Peer Review in Decline?," considers possible explanations for…
One of the things that happens when I lay out a problem (say, the difficulties for scientists in communicating with non-scientists about scientific matters) is that my excellent commenters remind me not to stop there. They press me for a solution. I started, in my earlier post, to gesture toward…

Have you seen this? Is there a fine line somewhere there - if you hit it, you get published; if you go too complicated or too simple, the peer-review punishes you? And is that line set at "too complicated" setting already?

Uh, what? I don't know what papers Porush has been reading, but they weren't written in my world:

Complex sentences

... are frowned upon, and the simple declarative is much preferred. The goal is complete transparency: the information should flow into the reader so easily that he or she does not notice the prose. This is why research papers contain so many stock phrases: when you mean to say the same thing, say it the same way, so that everyone gets used to that formulation.

Passive voice

... is now actively discouraged; students are taught to write in first-person active, "I did this", "we did that". My instructors were drilling that into me more than a decade ago.

Descriptive prose

... is OK if you're Natalie Angier, but otherwise best avoided. And you probably aren't Natalie Angier.

squirrel language

... is also discouraged, and in its place students are taught to make assumptions explicit so there will be no need for squirreling.

Be as wary of innovation or taking chances in language as you are in the lab.

Good advice. Why take a chance on the language in the one forum where you do not want the language to intrude? You want to take chances with language, write poetry.

Revision is unnecessary

If this is a claim that researchers do not revise their manuscripts, it's ridiculous on its face.

There are lots of things that could be said about improving scientific writing, but it would be best if the person saying them gave evidence of having read an actual paper in the last ten years.

"but it would be best if the person saying them gave evidence of having read an actual paper in the last ten years."

To be fair Bill Porush was writing more than ten years ago, 12 to be precise (The book was published in 1995).

I'll agree with you that science communication has improved for the better, but as a non-scientist who has to read a lot of ethics committee applications which specifically state on each page "make sure this is understandable by a lay person" there is still some way to go yet...

I think this is a reasonably helpful list of some common flaws in scientific writing.

Cheers David

My biggest disagreement is with calling the list "Idols of Scientific Writing". I think what he means to say is "Common Characteristics of Scientific Writing". As Bill so aptly wrote earlier, some of the practices Porush described ARE frowned upon by the scientific community. Most scientists would probably tell you that most scientific papers are poorly written. Scientists don't _strive_ towards these characteristics; however, most of them do _get away with_ them. That being said, I have a few specific comments on his list:

1. For a man who hates complex sentences, Porush (at least in the quotes provided) sure makes use of parenthetical expressions and semi-colons. Bill was right in saying that complexity is discouraged. Scientists love a simple explanation. Unfortunately, they don't always explain it simply. There's room for improvement here.

2. Personal pronouns are more accepted these days, but they should still be used sparingly. While personal pronouns remind us that the work was ultimately done by people (with all the human errors that entails), it's bad writing in general to keep saying "I did this..." and "I did that...". It's a scientific paper, not an ego trip.

3. Descriptive prose is still frowned upon because it can too often be lost in translation. Figurative language is most successful when everybody comes from the seem frame of reference, and that rarely happens across disciplines and across international borders.

4. You do need to be cautious about your conclusions. This is even more important these days to help combat the sensationalism and politicalization of science. There's nothing wrong with a weak claim, if that's all the data support. What is often a problem is vague terminology. Many authors hedge their bets by using words like "seems", "slight", or "tends to". Quantify as much as you can.

Regarding passive voice: using it varies with the section of the paper. In general, you'll find more passive voice in the Methods and Results sections, and much less in the Introduction and Discussion. This is because Methods and Results focus on the work, while the Introduction and Discussion focus more on the research activity (and hence the researchers).

By David Kellogg (not verified) on 18 Jul 2007 #permalink