Once pushed beyond endurance by the insistent attentions of a junior copy editor obsessed with prosecuting the ultimate crime of the split infinitive, George Bernard Shaw famously told the hapless copy editor's boss: "I care not whether he goes quickly or quickly goes, but go he must."
I suspect that Shaw was victorious and his text retained its full complement of split infinitives. However, consecutive research assessment exercises have taught British scientists to bypass book publication as they travel the one true road to career success: distilling ever larger and more overhead-laden grants into extended abstracts for publication in top-ranked journals. Unfortunately, anyone hoping that this intense literary selection pressure would generate profound and enjoyable reads will be disappointed by the present content of most science journals. Increasing rejection rates and ever more intrusive editing are combining to render them superficial and tedious.
Tim Birkhead ("... and be quick about it", 14 May) has noted the damage caused by the pressure for near-instant peer review of manuscripts, but he did not fully explore the psychological consequences of formatting a review for a highly prescriptive website. Unsurprisingly, shoehorning hurried comments into electronic boxes has brought a formulaic box-ticking mentality to the entire review process. Allies say "yea" and detractors say "nay", typically with minimal justification. Nay-sayers are more likely to triumph in the present environment, where an editor's primary role is finding reasons to reject perfectly valid scientific outputs in order to maximise the journal impact factor. Gentle rejection is effected by recommending that further research be done, while firmer condemnation is easily achieved by identifying any element of the text as heterodox.
A scientist fortunate enough to survive the review phase - essentially a lottery largely dictated by the fortuitous absence of nay-saying reviewers - then faces two further stages in the publication process that barely existed a decade ago. Negotiation is first required with editors, who are often more critical than reviewers, and then with the rapidly expanding posse of copy editors employed to impose order on each scientist's chaotic prose. Such rigour might be justified if it brought greater clarity, but it is actually employed primarily to shorten manuscripts wherever possible so that more text can be crammed in, and to bring greater uniformity and false objectivity to science writing.
The grossly simplistic Introduction-Methods-Results-Discussion format familiar to any science-oriented schoolchild is ever more rigorously imposed on empirical studies. Literary quotes, linguistic flourishes and parenthetic asides are religiously expunged, along with any indication that the authors had prior expectations of what their studies might reveal. Routine suppression of the first person prevents researchers from taking overt responsibility for their work. The greatly expanded editing process is time-consuming, eroding the productivity and job satisfaction of researchers and editors.
This literary selection pressure has rendered most scientific journals homogeneous in both ideas and style - a trend that electronic journals have done little to redress. But the increasingly electronic nature of publishing and the subsequent filtering of scientific outputs through search engines does raise a more fundamental question: is there an ongoing need for peer review, or indeed for the present level of editorial intervention? If citation is as important as the RAE and research excellence framework suggest, why can't market forces alone determine the best science through the citation process?
The case for anarchy is strengthened by the fact that personalities, ideas and controversies have already been driven to the margins of "legitimate" science, where exploratory, innovative thinking has been kept alive primarily by the invaluable ministrations of a handful of maverick popularisers. Exciting and provocative science is most likely to be found in outlets that require little peer review or editorial intervention - either good old-fashioned books or ultra-hip blogs - where it often resides alongside, and can be difficult to differentiate from, pseudoscience.
We scientists can continue to sacrifice increasingly large proportions of sound research on the altar of orthodoxy, indulging in ever more time-consuming self-censorship. But we might make better use of our time by allowing readers, rather than editors, to be the primary arbiters of good science. If we maintain our present authoritarian course, science will continue to lose students and public support to the arts, where self-expression is still grudgingly permitted.