Plagiarism is the most sincere form of flattery, they say (or rather, this is said of imitation). In arts - literature, music, painting - it can at times be tolerated, as an artist might want to take inspiration from others, elaborate on an idea, or give it a different twist. In art it is the realization of the idea which matters.

As opposed to that, in scientific publications plagiarism is almost univocally bad, as what it refers to is not the copying of an idea: we always fruitfully and guiltlessly copy ideas in our scientific work. Plagiarism in scientific publications concerns the direct cutting and pasting of pieces of other articles. It is bad because a scientific publication should be something that advances human knowledge, while articles constructed by assembling pieces of previously published material only add worthless noise.

Furthermore, plagiarism threatens the system of judging the quality of a scientist by the number of publications he or she has authored. I do not particularly like such a yardstick (and I say this regardless of having over 900 publications with my name on them), but I like even less the pollution from non-original articles.

As an editor of an Elsevier publication, "Reviews in Physics", I thus keep the guard high against unsolicited submissions. I love to receive those -any editor of scientific journals does, I guess- but I am suspicious at first. Who is the author ? How established is he or she in the field ? And above all, is the text original ? This last question comes up even before I try to assess whether the article describes interesting physics and/or whether it is good or bad.

There exist refined tools on the market to spot pieces of a text that are identical to others already in print. iThenticate is the tool used by Elsevier editors for this task. But then there is Google. Google alone works wonders, if you know how to search. But it is not just Google: you also need to develop an eye to spotting the parts of a draft article which are most likely to have been copied. Plus, of course, you need to have the stimulus to search with Google in the first place.

Yesterday I was happy to prove to myself that I do have such an eye for plagiarism. I was evaluating a possible article for "Reviews in Physics" and something led me to sense that the work could in fact be copied from others. 

First of all, the paper did not make it to our page limit, which is a strict 15 pages. 15 pages is little space for a review of even a specialized physics topic, and usually authors have to make an effort to squeeze within that limit all they want to put in their article. The one I was evaluating, however, was 12 pages long. Huhm, you really don't have more to say on the topic ?

As a second indicium, the paper was a unsolicited submission by three authors, all reasonably well established in the field. As our journal has been launched only less than a month ago, I found it strange that they would submit their text at once: we ask for title and abstract only. Was this an article rejected by some other journal, which they were trying to submit to us now ?

So I set out to google parts of the text. And I soon found a perfect match: a sentence most likely copied from a 2005 review on the same topic. However, the draft did cite that review in the references: a case of imitation rather than plagiarism ? I gave them the benefit of doubt, and continued to dig.

It took me quite a while to find a proof of my suspicions, but I did get there as I googled a part of a sentence which looked like it was in poor match with the rest of the text. Google brought up the same sentence from a "chapter 7" document with no title or author. The geographical origin of the document, however, was the same as the prospective authors! Could it be a case of self-plagiarism ? Self-plagiarism is totally acceptable in scientific papers in my opinion. I would have been happy if it had been so, but first I had to determine who was the author of that chapter. Was it a book or something else ? 

Finally I got around to the site which contained the full text. It was a repository of Ph.D. theses. Unfortunately there were hundreds of theses on the topic of the review, and the site did not allow text searches. I thus wasted some 20 minutes trying to check a dozen "chapter 7" sections of Ph.D. theses. And finally, I found it. Several pieces of text matched perfectly the manuscript I was evaluating. And crucially, the author of the Ph.D. thesis was not one of the authors of the manuscript! Nor was the thesis cited in the references. Case closed. Those guys shamefully tried to exploit a Ph.D. thesis to bump up their publication rate. 

I have described the above incident here because I wish to stress that it is not so easy to detect plagiarism - in this case, iThenticate failed to find a match, as it did not check PhD theses as Google can do. But it took also some ingenuity to understand that the manuscript was worth checking in detail.

Of this I am quite proud of!