S. H. Friedman and J. O. M. Karlsson recently reported what they believed to be a novel paradigm (Nature 385 480 480; 1997).

They presented a graph of the proportion of articles in the Medline database with the word ‘novel’ in the title or abstract, plotted as a function of the year of publication. They extrapolated, from the exponential rise in the use of the word, that by 2020 all scientific papers will claim novel findings. It is ironic that their observation is itself not novel.

Naturepublished correspondence in 1991 (350 9 9; 1991) in which we noted the accelerating use of the word ‘novel’ and documented it with a graph of the frequency of use of ‘novel’ and two control words (‘control’ and ‘unusual’) in Medline database records. In addition, we recommended that authors reserve ‘novel’ for strikingly new discoveries, lest the word lose its impact.

We are sorry to see, from the updated data provided by Friedman and Karlsson, that our recommendation has had no discernible impact. It appears that the editors of Nature recognize the dire consequences of the novelty explosion and will provide regular updates of the situation until the message takes hold.