A referee report


Recently, I got this as part of a referee report on one of my papers:

If I am allowed to be a bit more picky, I wish to suggest to avoid some of the wordings that are not quite rigorous in a scientific article. For example, “unexpectedly large correlations of peak positions …”, the word “unexpectedly” is quite subjective.
Similar problem with “surprisingly important …”

So, how do you answer that?    It’s not as if he/she is really wrong, but it’s also hard to agree.  One feels intuitively that a scientific paper should be a little more than just a report of a fact:

Measurement of the Off-diagonal Dicromatic Effusion Coefficient of CH4

G. W. Objective and O. T. Purist, University of Southern North Dakota, Hoople, ND.

Introduction:  It has long been known that the microwave Bon-Bon effect (Bonhamption and Bonning, 2003) can be explained terms of a dicromatic effusion matrix operating on the … these coefficients have never been experimentally determined.

Methods:  Spectroscopically pure [1] Methane Hydrate was chosen as especially suitable to show the effect.  The sample was accidentally strained during mounting, then exposed it to pulsed laser irradiation at a fluence of 1 TW/square-cm in sub-picosecond pulses …

Discussion:  … correcting for the Rath-Molson effect which is generally believed to … Three of the samples were chosen for detailed study and typical results are shown …

Conclusion:  The off-diagonal effusion coefficient is found to be 4.4±0.67, in natural units.

Certainly, one should be objective in reporting results, and they are the most important part of a scientific paper.  If your results cannot be reproduced by any competent experimenter, they aren’t really worth anything.   (Except perhaps as a teaser to get other people to think about the microwave Bon-Bon effect.)   Certainly, no one can build on your work if there are too many loose ends that aren’t explained.

Now, in Linguistics, subjectivity is always a problem.   Most papers in linguistics journals are subjective in one way or another.  For example, many experimental papers depend on a subjective evaluation of what something sounds like, or where to place the borders between one word and the next.   As another example, papers in linguistic theory typically select words/phrases to examine in an unexplained, presumably subjective way.   And, since the conclusions that are reached can depend on the examples used, that means the conclusions that they reach are subjective.   (But, that’s a rant for another day.)

Since this referee and I operate on the fringes of Linguistics, I approve of his/her effort to hold the line and to keep papers as objective as possible.  But there is another part of a paper which you miss if you try to force the entire paper to be objective.  You miss the chance to explain where the results fit into science and why they are interesting.   You also miss the chance to tell people what the research means and give them some ideas what it might be good for.

That part of the paper invariably contains subjective opinions; it is forward-looking into the future, and one certainly cannot objectively measure what a paper is good for.  However, if one gets rid of it in an attempt at scientific purity, the paper becomes a little less valuable and a little less interesting.  Often, it’s small.   But sometimes it is important.   Consider Watson and Crick’s paper describing a structure for DNA (F. H. C. Crick and J. D. Watson, “The complementary structure of deoxyribonucleic acid”, Proceedings of the Royal Society A, volume 223, pp. 80-96, 1954).   Who would eliminate the beautifully understated subjective opinion from the paper that set out the core idea for the next 50 years of biology:

In conclusion, we may mention that the complementary relationship between the two chains is very likely related to the biological role of DNA.  …  It seems to us that the presence of a complementary structure strongly suggests that the self-duplicating process will be found to involve the alternative formation of complementary chains, and that each chain will be found capable of serving as a template for the formation of a complement.

I’m unlikely to do that well, but I insist on the right to try.