The ease with which large chunks of text can be digitally scanned and compared with what has previously been published has produced a new breed of academic watchdog. Plagiarism-detection software has opened up scrutiny of scientific publications to non-experts and text that has been copied and pasted without proper attribution is now a common reason for papers being retracted. Hobby­ists and political opponents have made a cottage industry out of searching the back catalogues of high-profile individuals for evidence of such misdeeds.

Such plagiarism is unethical and it is a form of misconduct, but scientists are not writers. We value the originality of ideas more than of language. There are worse offences than text plagiarism — such as taking credit for someone else’s research ideas and lifting their results. These are harder to detect than copy-and-pasted text, so receive less attention. This should change. To help, academic journals could, for instance, change the ways in which they police and deal with such cases.

To scientists, plagiarism of an idea strikes at the heart of research as a creative enterprise. An idea could be a hypothesis to explain observations, or an experiment designed to test a hypothesis. Such plagiarism is difficult to uncover unless errant authors make fortuitous slip-ups, or it is accompanied by text plagiarism. I have been a victim of both.

In the first, scientists published a paper in a physics journal that failed to give me and my colleagues credit for a method that we devised to test our idea of a magnetic glass state. We called it cooling and heating in unequal field (CHUF), and the plagiarism might have gone unnoticed had the errant authors not also copied our protocol’s acronym.

After we complained, the journal published a swift correction from the authors, who apologized and acknowledged that we had published the idea first.

In the second instance, another set of authors duplicated one of our ideas, along with large chunks of text. In this case, the correction addressed the text plagiarism only.

There is a third form of copying: results plagiarism. This is different from fraud, in which the claimed experiments are often not carried out. In results plagiarism, scientists can repeat an experiment and obtain valid data. Such reproduction is, of course, a useful and common feature of science. The deception comes when they fail to mention the original work.

It is important to appreciate why scientists may indulge in the three forms of plagiarism.

Giving credit to those who did the original work is essential. Credit is a driving force for humans and quality of original thought is one of the most fundamental ways to judge scientists’ work and determine who progresses in their career.

In my opinion, scientists who plagiarize ideas and results are usually established researchers who can assess the validity of published work and probably to have publications of their own accepted and cited.

How should this be dealt with? It is important to appreciate why and how scientists may indulge in the three forms of plagiarism. Copied text in a paper’s introduction or concluding paragraph may happen simply because the authors lacked sufficient command over the language (usually English) to express the concept in a different way, or had read previous works that left an indelible mark on their subconscious. (This is a charitable interpretation and does not apply if the copying is extensive). Such examples of plagiarism can be enough for some journals to retract a paper. But who benefits from such a penalty? If the experiments are performed as described and due credit is given, then the results in the errant paper are robust and benefit the relevant field.

Second-hand text in the methods and results sections could indicate a more severe case of plagiarism — of ideas and data. Computer software cannot determine whether this is the case; that requires the eyes of independent experts.

Authors who have plagiarized ideas or results have crossed a serious ethical line and should be sanctioned by their institutes, as should other perpetrators of misconduct. But need the errant papers be withdrawn, as many journals currently insist? I argue that this withdraws new and useful results from the scientific record.

Speaking as someone who has been plagiarized, I believe that a correction to the errant paper is sufficient. The wording of the correction must make clear that the offence was plagiarism, not fraud, and include reference to the original work. Most importantly, the correction should be attached to the PDF file of the paper, so that every download also carries the correction. This does happen, but not enough: many journals provide only an online link to the correction. If viewers, including future referees, do not notice that link, then the corruption of the scientific flow of ideas and credit continues. If corrections were to be made prominent, it could dissuade plagiarism in future.