Avoiding plagiarism

Previously published as: Hart, G. 2020. Avoiding plagiarism. https://www.worldts.com/english-writing/eigo-ronbun75/index.html

Plagiarism is most often defined as using another author’s thoughts, and specifically their words and sentences, but presenting them as if they were your own thoughts. This is a complex subject for scientists, since it involves both ethical and legal considerations. From an ethical perspective, all modern science relies on the thoughts of previous researchers, which form our collective heritage; that is, these thoughts create the body of scientific knowledge that should be shared freely and openly among all researchers. To give credit for the hard work of previous researchers, we use literature citations to acknowledge their contribution to that heritage.

From a legal perspective, complications arise as a result of the constraints imposed by international copyright law. Simply citing someone’s work as the source does not give you the right to repeat their words exactly, since those exact words are protected by copyright law. When you write a paper, you therefore cannot create the entire text of your paper by copying the words of other authors, even if you cite the source of each group of copied words.

Note: “Fair use” guidelines exist to define how much of another author’s words it is legal to repeat in your own manuscripts. This is a complex issue, and details vary between academic fields and countries. It’s safer if you do not rely on fair use to justify copying another author’s words.

To protect themselves against the legal consequences from intentional or accidental copyright violations, journal publishers increasingly use plagiarism-detection software to prevent authors from copying published (thus, copyrighted) manuscripts. There is an obvious problem with this approach: There is a finite number of ways that you can communicate a given concept in any language, and the number of ways can be very small for very specific descriptions. At some point, it may become impossible to describe something without using the copyrighted words that someone has already published. This problem is particularly serious for descriptions of experimental methods, since there is a very small number of ways to describe procedures such as adding a specific concentration of a chemical to a specific laboratory solution or performing a specific measurement with a specific instrument. Although you can increase this number by using synonyms (words with similar meanings), this adds little value; stating that you magnified or enhanced the temperature of a reaction offers no benefit over stating that you increased the temperature; on the contrary, synonyms usually communicate less clearly and effectively than the one best word for a given concept. Yet if you use those best words often enough, you increase the risk that a journal’s plagiarism-detection software will describe your manuscript as plagiarism.

This problem occurs even when you are only quoting your own words from a previous paper. Although the words are yours, it is generally necessary to transfer the copyright for those words to the journal that published your paper. In theory, and sometimes in practice, that journal can accuse you of violating their copyright if you repeat your own words in a subsequent paper. This is problematic if you are performing a long-term study of a single phenomenon or a single study system. For example, it’s difficult to vary the wording if you use the same crop cultivar for several years of research and must describe its genetic characteristics, if you perform long-term research at a specific site and must describe its climate, or if you examine the electrical properties of a given ceramic capacitor with different dopants added to alter its behavior.

The ideal solution would be to fix the plagiarism-detection software so that it accounts for the unique characteristics of how scientific authors must write a manuscript. For example, repeating descriptions of standard methodologies should not be treated as plagiarism because scientists understand that these methods are shared among all researchers. Unfortunately, there are strong economic incentives for journal publishers to not fix their software.

There are several ways to solve this problem. Possibly the strongest solution is to create an online database of descriptions that can be reused. For example, researchers could develop a database of standard methods for their field of research, and copyright the information in this database using the Creative Commons license. The Creative Commons “Attribution 4.0 International” specifies that text protected under this license can be copied and redistributed in any medium or format, even for commercial purposes, with three constraints. You must:

• acknowledge the original source
• provide a link to the license description, and
• clearly indicate any changes that you made to the original description.

Under this approach, text from a database of standard methods could be freely reused in any journal article without any problems related to plagiarism. In practice, it will be difficult to accomplish this without modifying the plagiarism-detection software so that it accepts the repetition of text that is published under the Creative Commons license.

Note: The Public Library of Science (PLOS) uses this approach.

This is not just a theoretical suggestion. The journals published by the Nature Publishing Group have begun to investigate the possibility of a “protocol exchange". The database does not require peer-review of protocols, although they are screened to detect obvious problems such as health risks. Nonetheless, it’s in the interest of authors to rigorously review their protocols before publication, and update them if necessary after problems are detected. This both protects their professional reputation, and helps them to improve their methodology in response to feedback from the research community. For this approach to work, it will need to become an accepted part of the submission guidelines for all journals. As this is clearly in the interest of journal editors and readers, this should become an increasingly accepted part of the future journal publishing process.

Until this approach becomes universally accepted, and until all key methods have been included in such a repository, authors must look for compromise solutions. For example, when my authors are repeating their own published methods, I encourage them to add words such as the following in their letter to the journal editor:

“The Methods section of this paper repeats text that was previously published in [one or more literature citations]. We have done this to ensure that the peer reviewers can fully understand the present study without having to find and read [this previous paper / these previous papers]. We will be happy to work with you and the reviewers to determine how much of this repetition should be retained in the final published manuscript.”

I encourage you to copy those words freely if they meet your needs; I am releasing them under a Creative Commons license! By clearly indicating what text the journal’s software may identify as plagiarized, providing a strong justification for this repetition, and by emphasizing that you will work with the editor to solve any problems, you provide a strong argument for allowing the repetition.