Would you believe this man? Dr Fox

In my previous post I tried to allay some possible concerns about "writing to recipes", and one of my main points was that sticking to a schema can lead to clearer writing. In this post I want to tackle another "fear of the formula", one that seems the most paradoxical of all: an aversion to the very clarity that we were previously (and naively?) assuming to be a desirable feature of science reporting.

Nearly every time I give a course at least one student expresses this concern directly, e.g. "If my paper is easy to understand, people will think the science is too simple" or implicitly, e.g. "Most of the important papers in my field are difficult to read". Now you might think a writing teacher should give short shrift to such ideas and tell the students: "No more talk about that, it's just plain wrong". But is it?

Various lines of evidence suggest that clarity and intelligibility are, in fact, not always rated highly by academics, including scientists. In the much-discussed Dr Fox experiment a trained actor played the role of an authority on the application of mathematics to human behavior. He delivered a deliberately nonsensical talk to audiences consisting of psychiatrists, psychologists and other highly educated professionals. The talk, "Mathematical Game Theory as Applied to Physician Education", was very well received, and was probably enjoyed by no-one more than the attendee who said he had previously read some of Dr Fox's publications!

J. Scott Armstrong explored a similar theme, looking at how readers evaluate journal articles.  His study, which focused on management journals, suggests that readers rate content more highly if it is expressed in an obtuse writing style. More recently, in the well-known "Sokal affair", the physicist Alan D. Sokal, eager to demonstrate sloppy thinking in some branches of the humanities, managed to get a largely nonsensical hoax article published in the journal Social Text ("Transgressing the Boundaries:Towards a Transformative Hermeneutics of Quantum Gravity"). I find it a rather amusing piece, with many gems, such as:

"Thus, a liberatory science cannot be complete without a profound revision of the canon of mathematics (105). As yet no such emancipatory mathematics exists, and we can only speculate upon its eventual content. We can see hints of it in the multidimensional and nonlinear logic of fuzzy systems theory (106); but this approach is still heavily marked by its origins in the crisis of late-capitalist production relations."

Befitting a millennium where expectations of instant stardom are the norm, anyone can now perpetrate their own Sokal hoax, thanks to the efforts of several MIT computer science students. These students created SCIgen, an online program that automatically generates a fake, nonsensical paper complete with figures and bibliography. SCIgen papers composed entirely of gibberish have been accepted by at least one conference and one journal (Bruce Lewenstein, personal correspondence). Here is the first paragraph of the Introduction to a paper I "authored" in a one minute break from writing up this post:

"In recent years, much research has been devoted to the synthesis of I/O automata; contrarily, few have enabled the synthesis of voice-over-IP. In the opinion of steganographers, the usual methods for the improvement of Moore's Law do not apply in this area. A typical riddle in steganography is the exploration of the refinement of scatter/gather I/O [1]. The study of virtual machines would greatly degrade the exploration of superpages."

In all these examples, the implication is that the style of a paper or talk has a major effect on the evaluation of content, and in most cases, content very much takes the back seat. One would expect that at least some stylistic features of academic reporting have evolved to facilitate communication. It now seems obvious however that these features almost have a life of their own, life enough to sometimes persuade experts in the field that a nonsensical report must contain useful content even if they cannot personally detect it.

Now, it might be countered that the cases described are extreme, atypical examples that do not bear upon our initial question: should or should not a working scientist aim for clarity? It is true that a scientist never needs to write up a nonsensical report, at least not that he or she is aware of. But often we must report a study that seems so lean in terms of either the quality or quantity of its findings that it is "too simple" for the desired category of journal. In this situation lucid writing may well make the inadequacies clear to a reviewer and thus seem something to avoid. So just as in the hoaxes above, obtuse writing lavished with the stylistic hallmarks of academic discourse could keep some reviewers wondering where the problems of unintelligibility lie: with themselves or the paper. For an author, this "approach" could be particularly tempting because it does not take any particular skill or effort to write non-lucidly: for many it is the default style of an early draft, and for some their default style, period. Armstrong recounts an anecdote about an author who having had a paper rejected several times, gained success by sending in the first draft, which she considered incomprehensible.

So do I take the side of those students who fear obscurity if their papers are mean and lean? Never. No matter how small the kernel of "truth" in a study, I feel it may well have as much long-tern impact if presented clearly in a minor journal as when tarted up confusingly in a major one. But even more importantly, during the process of writing itself, the more precisely one can isolate the gist of the study, the better one is able to explore it intellectually and to see it in the broadest context. And in doing thus one may well end up with findings that are not so lean after all.