Avoiding Plagiarism in Scientific Writing

17 Nov 2020 . Category . Comments #service #reviewing #teaching #advising

When you present someone else’s work as your own, you’re committing plagiarism. Most guidelines I’ve seen focus on plagiarism from the perspective of an individual. While this perspective is not invalid, it seems to lead to an approach to plagiarism that I disagree with overall, namely very elaborate rules for paraphrasing ^{1 2}.

I argue the more appropriate perspective in scientific writing is a collectivist perspective that focuses on the scientific enterprise. Science is fundamentally a process of constructing explanations through argumentation. Arguments rely on claims, evidence, and reasoning. Scientists are accountable for their arguments, which are published as part of the public record. Plagiarism distorts that record by obfuscating who is responsible for claims, evidence, and reasoning. This is extremely problematic regardless of whether the arguments are correct or incorrect. If the arguments are incorrect, the community does not know to whom to turn to correct the argument, assuming it can be corrected. If the arguments are correct, the community does not know who is the authority on the topic and should be trusted or rewarded with more resources.

Examples:

• An argument is believed to be incorrect, but is actually correct. Plagiarism puts the wrong person(s) in charge of establishing the validity of the argument. They fail, and a correct line of inquiry is abandoned.
• An argument is believed to be correct, but is actually incorrect. Plagiarism puts the wrong person(s) in charge of extending the line of inquiry. They continue and advance a flawed line of inquiry.
• A student wishes to be trained as a scientist in a particular area. They choose a mentor who has plagiarized work in that area. The student receives suboptimal training.
• A funding agency wishes to advance a line of inquiry. They fund a scientist who has plagiarized work in that area. The scientist receives funding, and since funding is a zero-sum game, the scientist who originated the idea does not.

The examples hopefully are sufficient to show that plagiarism damages the scientific enterprise in profound ways.

The key to avoiding plagiarism is to understand scientific argument. I believe the individualistic perspective on plagiarism is completely subsumed by this, with trivial formatting-level extensions, e.g. using double quotes and page numbers for direct quotations.

Understanding scientific argument through citations

Perhaps the most useful way to understand scientific argument generally is by looking at citations.

When are citations used?

Citations are typically used in the introduction section and the discussion section of a research paper. Citations in the introduction section support what relevant work has previously been done on the topic and why the current work is interesting/important. Citations in the discussion section compare findings to other work, for example elaborating why the findings are to be expected or not given previous work.

What is important to realize as you read other papers, and as you write them, is that every time an author says something, they are implicitly making decisions about whether a citation is needed, and if so, what to cite. Citation behavior (should) follow these conventions:

• A citation is needed if the writing is referencing other work, either as the source of an idea or as an example.
• A citation is not needed if the writing is advancing a novel idea or interpretation of the literature.
• A citation is not needed if the writing is referencing common knowledge in the area that is not directly attributable to specific work.

In other words, lack of citation implicitly claims either that the writing advances a novel idea/interpretation or that the writing is referencing common knowledge. How can one tell the difference between novel ideas and common knowledge? While the major novel ideas are sometimes marked in other ways, e.g. “We argue that...", this is not always the case. For this reason, it is extremely important to be well-read in the area you are writing about so that you know what is considered to be common knowledge and what is not. You should assume ideas/interpretations are original (and citable) unless you can prove that they are not.

What do citations mean?

Citations have a meaning beyond giving credit to other work. Because science is argumentation, the act of citing work implies you believe that work is sound (i.e. valid and reproducible) and you believe that work supports the argument you are making. So clearly you cannot cite something that you have not read, because you can only make these commitments for work that you have read. Of course you will sometimes want to cite flawed work or cite things in a particular way. When you do this, you will naturally disclaim aspects of the cited work, e.g. “Previous work has found that bosons are leptons that extend into the 5th dimension (Choux Pastry et al., 2051), but this work is flawed because..." However, if aspects of the work are not specifically disclaimed, then the default assumption is that you believe in the integrity of the work and that it supports the argument you are making.

Citations are extremely important to the scientific enterprise because of the meaning they convey. Consider that most publications are not cited that many times; even publications in “decent" journals like PLoS ONE are only cited about 3 times on average ³. Every time you cite a paper, you are voting for its importance in the mindshare of all scientists working in an area. Papers that are not continually voted for will eventually drop out of everyone’s consciousness, either becoming part of common knowledge or being forgotten entirely. So citations are arguably as important as the original papers themselves. A fantastic paper that is not cited will have less effect on the progress of science than a terrible paper that is widely cited. In other words, citations filter the literature.

Avoiding citation plagiarism

Hopefully the previous sections make clear why it is so important to avoid “citation plagiarism" or citing work that someone else has cited without reading it and vetting it. But how can you avoid citation plagiarism in practice? Since every paper starts with a lit review, you’ll always read someone else’s lit review before writing your own. So how can you possibly write your own lit review without plagiarizing them? The answer to these questions is actually very simple: you can avoid citation plagiarism by reading the original work and taking notes.

Your notes should be a commentary on what you are reading. While you read, you want to evaluate the argument that the paper is making in terms of evidence, claims, and reasoning. Once you have read the cited papers, you can go back to the lit review you started with and compare it to your own notes. Does that lit review have an idea or interpretation that wasn’t in your notes? Is it relating different pieces of work in a way that you didn’t that is important to your argument? If so, you should cite the lit review paper for its specific contributions. In contrast, you do not need to cite the lit review paper for citations that are factual, e.g. “Jaconde Sponge et al. (2032) used an autoregressive DNN with a denoising objective," assuming you have read Jaconde Sponge et al. (2032) and believe citing it is sound/appropriate for your argument. Finally, it is possible that as you read the cited papers, you came to an idea/interpretation that relates different pieces of work and is otherwise not factual in the sense of the previous example. In this case, it is important to find a way to cite the lit review for its specific contributions even though you feel that you might have independently come to say the same thing. Not only is it fair (after all, they said it first), but it also strengthens your argument to cite them because they are independent researchers who agree with you (or you agree with them). There are many ways to do this stylistically, so if you’re unsure of how to express this in writing, take the time as you read papers to understand when, how, and why other scientists are using citations. This will help you understand how to cite work appropriately and avoid plagiarism.

References