Scholars are sometimes called upon to adjudicate charges of plagiarism, that is, of the unacknowledged “lifting” of one writer's words or ideas by another. When literal copying of whole expressions or sentences is involved, judgments are straightforward. But when the accuser claims that his or her ideas were plagiarized, as in the recent court case against Random House over Dan Brown's Da Vinci Code, the situation is much more complicated. Passing judgment in such cases requires a reflection on the meaning of “plagiarism” and the best ways of assessing it, which this brief article seeks to provide.

Plagiarism is the act of “stealing” or “abducting” (Latin: plagiare) somebody else's proposition and depicting it as one's own without properly acknowledging the source of inspiration. “Proposition” here includes the formulation of a specific phrase or the content of its ideas. The process of “stealing” somebody else's proposition is usually not directly observable and demonstrable, however. Therefore, assessments of plagiarism often must rely on inferences about the alleged process from the outcome, namely judgments about the degree of similarity between propositions advanced by B (the person accused of plagiarism—“the accused”) and propositions advanced by A (the originator of the idea that is said to have been plagiarized—“the source”).

Sometimes, two people independently come up with the same observation or idea. For instance, the periodic table of chemical elements was developed independently by Russian Dmitri Mendeleev and German Julius Lothar Meyer in the 1860s. By contrast, a charge of plagiarism alleges that B did not design the respective propositions on his or her own, but took them from the source. The alleged “kidnapping” of A's proposition thus had a decisive impact on B's propositions; it changed what B would otherwise have advanced as his or her own propositions.

A precondition for charging plagiarism of A's propositions is that A must have been the true source of the proposition; that is, the phrase or idea that B allegedly plagiarized must have originated with A. A must have clear intellectual “ownership” of the respective proposition. This indispensable precondition excludes cases of common knowledge or obvious observation. “Common knowledge” means that many people have already made an observation so that it is in the public domain (e.g., that the USSR dissolved in 1991 because of the failure of Communism); even if A repeats that observation and B takes the idea captured in that observation from A, this would not be a case of plagiarism—unless B copies (almost) the exact phrase with which A advanced the observation (and that phrase is not common knowledge). An “obvious observation” means that any able-minded person could easily have come up with this observation; e.g., although A may have been the first person to publish that it did not rain today in Austin, Texas, B—living in Austin and observing the weather—would not need to cite A.

For A to be the originator of a proposition—and the potential source in a plagiarism case—that proposition must have a significant degree of originality. That is the case if A coins a new phrase or composes a specific text (even if the ideas expressed therein are not novel); if A designs a specific new idea (e.g., that extraterrestrials caused the downfall of the USSR); or if A in a distinctive, novel, and original way combines several ideas that A or others have advanced previously—e.g., that Nigeria has suffered from political instability because it has oil wealth (“resource curse”), a tropical climate, pernicious legacies of colonialism, ethnic division, AND institutional problems.

Thus, as a necessary (but not sufficient) condition for a charge of plagiarism to stick, A's propositions need to be distinctive and novel and have a significant degree of originality. In fact, the more divergent from commonly accepted views (the “conventional wisdom”) and the more counterintuitive A's ideas are, the more clearly and convincingly can A claim intellectual ownership to those propositions.

Assuming that this precondition is met, a charge of plagiarism can be proven most clearly if there is evidence of the process of B taking away a proposition from A. For instance, if B or B's research assistant comes forward and admits to have written literal excerpts of A's work on note cards and if B admits that s/he has incorporated these note cards into her own work (cf. the Doris Kearns Goodwin case), “lifting” is proven clearly. But frequently, such evidence is not available, at least not to the accuser and those who judge the accusations.

Often, however, only one precondition for the act of “lifting” can be documented, namely that B knew A's work. Besides references to A's work in B's text, evidence of such knowledge may consist of credible statements from A that s/he shared her published or unpublished work with B before the date of B's publication. This precondition excludes cases in which two researchers independently develop the same idea; it also relieves scholars of the need to conduct a perfectly exhaustive search of the literature to make sure that a scholar in Outer Mongolia or Malawi has not published in some obscure publication on the topic that B plans to analyze. But evidence that B knew A's text is only a necessary yet by no means sufficient condition for inferring plagiarism.

Absent clear and conclusive evidence concerning the whole process of “stealing” or “abducting” somebody else's proposition, assessments of plagiarism must rely on inferences from outcomes, namely the texts produced by the source and the accused. In particular, the more similar B's propositions are to A's (assuming that A's propositions are novel and distinctive, i.e., different from common knowledge), the more likely plagiarism has occurred. Since scholars who work on the same topic may well independently come up with fairly similar ideas and since assessments of plagiarism should probably follow the common maxim in dubio pro reo (“when in doubt, judge in favor of the accused”), similarities in propositions must be clear and striking to sustain a charge of plagiarism.

Obviously, similarities are especially striking when propositions are identical or very similar in formulation. Coming up independently with (nearly) identical phrases on complex scholarly topics is so unlikely that close correspondences in words provide firm proof of plagiarism.

Judgments are more difficult with respect to the alleged “lifting” of ideas. With respect to ideas, assessments best proceed in a two-step fashion: The more similar B's ideas are to A's, and the more different A's (and B's) ideas are to common knowledge and the conventional wisdom, the more well-founded is a charge of plagiarism. That is, there is reason to infer plagiarism if there is a significant difference between A's ideas and common knowledge, on the one hand, and a clear and striking similarity of B's ideas to A's. Only if A advances a distinctive, reasonably novel, and original idea and if B's ideas are very similar can one infer plagiarism. Thus, the clarity of A's intellectual “ownership rights” (as discussed above) plays a significant role in the assessment of plagiarism when it comes to the alleged “stealing” of ideas.

Accordingly, plagiarism can be inferred in two situations (provided that the above-mentioned precondition is fulfilled, namely that B knew A's work). First, if A advances a distinctive single idea that other scholars had not proposed before (e.g., extraterrestrials destroyed the USSR) and B afterwards advances a similar idea (e.g., space aliens undermined the USSR), one can reasonably infer plagiarism. Second, if A advances a new multi-component argument that takes a number of ideas advanced by others yet combines them in a distinct and novel way (see Nigeria example above) and if B repeats that complex, configurational argument in a clearly similar fashion but depicts it as his or her own, one can reasonably infer plagiarism. Where the ideas that form part of such a complex argument are—on their own—common knowledge or the “intellectual property” of other authors, B's rendition of the complex argument must be very similar to A's for a charge of plagiarism to be well-founded. If B advances some ideas that form part of A's complex argument but criticizes or rejects others—especially if these component ideas have before been advanced by other scholars (such that A cannot claim intellectual ownership of these specific ideas, but only of their distinctive combination in a complex argument)—there is no reason to charge B with plagiarizing A's multi-component argument.

In conclusion, since there usually is no evidence of the act of plagiarism, inferences based on judgments of similarity between propositions advanced by the source and the accused are crucial. These judgments should apply a two-step “similarity of distance” test: Plagiarism can be inferred if A's propositions are clearly different from common knowledge and the conventional wisdom and if B's propositions (specific phrases, distinctive ideas, or multi-component arguments) are clearly similar to A's.