The Ethics of Self-Citation: Good Reasons and Bad to Cite Oneself
Joseph S. Fulda · Journal of Information Ethics · 2008
Pioneering information scientist Eugene Garfield lists among the reasons one (should) cite(s) others to pioneers and to peers.1 We differ; articles are addressed to readers; homage is for a dedication or, perhaps, an acknowledgment. Garfield, of course, founded ISI, which derives its income in some positive correlation to the number of citations in the articles it indexes. This is not to call Garfield's motives into question: Most successful entrepreneurs believe in their enterprises, and Garfield is no exception. That's all.Even Garfield, however, would no doubt agree that paying homage to oneself is ignoble. In this brevity, we consider how rules 1, 2, and 3, in the essay following this column (see p. 54ff, below), are to be applied, with this caution in mind.Rule 1. If it is for the readers, it stays in, even if the source is oneself.Exception: If, however, it is merely an example used to illustrate a point and not itself related to the argument, it should be omitted, if not from the body of the article, then at least from the references that PsycInfo, Sociological Abstracts, ISI's databases, etc., will record.To illustrate this, let me give several examples-from my own work. I choose my own work for several reasons: (a) As I remark in the succeeding essay about editors and referees who routinely suggest that their own work be cited, they do so no doubt for the reason I do: I, like them, am most familiar with my own work. (b) Moreover, I have what philosophers call privileged access to my own thoughts. I know exactly why I chose to cite or not cite particular references (at least for my more recent work). (c) Com- ments, including critical comments, cannot be taken by the author as subjecting his work at random to a microscope, perhaps displaying small particles not visible at the macroscopic level. This last reason, then, has an ethical component.On to the examples: (1) In Multiple Publication Reconsidered (this journal, Fall 1998), I cited myself 13 times. The entire argument-that in some cases multiple publication is constructive, even ethically imperative- was made by example* in that piece: Without the examples, there was no paper at all! Moreover, many claims supporting the constructive and ethical purposes of multiple publication were advanced, none of which the scholar (i.e., the intended reader) could verify without full bibliographic citations. The one other self-citation, to prior work, was a reference to the idea that publication itself is ethically imperative, for one cannot argue that multiple publication can be ethically imperative, if publication itself is not, and it is better to cite oneself than to repeat an article under circumstances that themselves clearly do not meet the criteria for multiple publication!(2) In Peer Review: Three Suggestions in the Spirit of 'Mend It, Don't End It' (this journal, Fall 2007), I referred to both a large and a small contrarian idea that had difficulty getting published in peer-reviewed journals. The large idea was creationism; the small idea was the assertion that virtually all uses of the conditional in ordinary English have truth-functional semantics. For the former, I gave the name of the journal that finally broke the ice; for the latter-my own contrarian views-I gave no specifics. Since the purpose of that piece was to provide suggestions for mending the peer-review process, it would have been entirely out-of-place to list the arguments and the references giving my views on this other subject. And, if readers doubted that contrarian views do have difficulty getting past the hurdle of peer review, they were referred to a more general article on just that subject by another scholar. Here, it would clearly have been inappropriate to cite myself. (Although when a reader asked me why I did not do so, I promptly sent him a list of references.)(3) In What Happens When an Author Does Not Provide an Abstract (Journal of Scholarly Publishing, January 2006), I again made an argumentby-example, an inductive rather than analytic, approach. …