Writers may have their noms de plume; revolutionaries may have noms de guerre. Here, though, we will speak of (to coin a phrase) the noms de litige, and ask: When pseudonymous litigation is allowed, what sorts of pseudonyms should be used? In particular, how can we avoid dozens of Doe v. Doe precedents or Doe v. University of __, all different yet identically named? This piece discusses some approaches to achieving the twin goals of pseudonyms: protecting privacy and avoiding confusion.
Courts generally disfavor pseudonymous litigation, but sometimes allow it.1 Indeed, they sometimes themselves pseudonymize cases for publication, even when the party names remain in the court records.2 Both courts and parties also sometimes pseudonymize the names of nonlitigant witnesses and victims. But what kinds of pseudonyms should be preferred? There are many options, including:
(I focus here on pseudonyms chosen for the purpose of litigation; when parties already have well-established pseudonyms, for instance as authors, there may be reason to retain them, assuming that such pseudonymity in litigation is found to be allowed.18)
Each option, unsurprisingly, has its strengths and weaknesses. Traditional pseudonyms strongly signal that the party is pseudonymous (though there are of course many thousands of real people with those names19); so do neutral descriptive pseudonyms. But both of these approaches make it harder to uniquely and clearly identify cases, especially when the other party is a frequent defendant, such as the federal government, a university, or a fellow Doe.20
That’s of course already a risk with common real names, such as Johnson,21 but it’s especially serious with Doe cases. To give just one example, there are six Doe v. Trustees of Indiana University cases just from 2020 to 2022 that have yielded opinions available on Westlaw, all in the same field (higher education law).22 These seem likely to be joined by new cases each year, and they will remain potentially citable for decades to come.
Common names (such as Smith or Johnson) have some of the problems of Doe, without the advantage of quickly signaling that the party is pseudonymous.
Pure initials avoid these problems, but can “make for poor readability” and be “dehumanizing” (or, perhaps more precisely, “depersonalizing”23) — “human beings are the subject of . . . cases, not acronyms.”24 This is likely even more true of alphanumeric combinations that occasionally appear in such cases, such as P3, V7, or JA-836 Doe. The initials of different people in the same case are also fairly likely to coincide, especially when the parties share a last name.25
First-name-plus-initial can be less ambiguous and less depersonalizing, but might be too revealing of the parties, especially when “the names . . . are fairly unusual.”26 Indeed, even pure initials can be identifying, when coupled with other indications, such as the small school that the party is attending.27
Famous pseudonyms can be distracting, and can also bring political spin that might subtly influence the judge or jury, as with Hester Prynne (the name of the heroine in The Scarlet Letter,28 who was publicly shamed for adultery, used in a challenge to a sex offender registration law). Some descriptive names, such as “Jane Endangered,” can have the same effect. So can descriptive names that are professional designations, such as “Navy Seal 1” — while a party’s occupation and accomplishments will of course often be part of the record, indicating them as part of the party’s name might sometimes give the party an unfair, if slight, advantage.
Arbitrary names can risk inadvertently implicating someone else who has that name. To give one example from the Ninth Circuit:
The plaintiffs in this case previously were denominated “James Rowe, Jane Rowe and John Doe.” One of the many persons genuinely named “James Rowe” wrote to the court while the appeal was pending, and said that his reputation was harmed by a newspaper story about the appeal, because careless readers might think erroneously that he is a convicted sex offender. . . . It is preferable for lawyers and courts to avoid harm to the reputations of real persons by using . . . traditional references for pseudonyms.29
Perhaps the best solution for a solo pseudonymous party is the one used by the Equal Employment Opportunity Commission in its decisions dealing with discrimination by federal employers: It uses an arbitrary first name (generally matched to the gender of the party but not to any other characteristics, such as ethnicity) coupled with an arbitrary initial, such as “Christopher M.”30 Court opinions including such names would presumably need to note that they are pseudonyms, since at least at first the public would assume otherwise. But once that is done, using such pseudonyms ought to avoid the bulk of the problems noted above.
This approach may be less effective when there are several pseudonyms that need to be used (either for parties or for witnesses or victims), since that may end up too confusing for some participants. As one court described,
Evidence already submitted highlights the problems pseudonyms might pose in the present action, and the confusion it can produce. At least two women [among the litigants] use the pseudonym “Gertrude Stein,” and nine women are referred to simply as “Guerrilla Girl.” One woman cannot remember “her Guerrilla Girl pseudonym” so she has adopted the name “Chansonetta Stanley Emmons” for this litigation. Other women “changed their minds about their pseudonyms” and adopted new ones part of the way through their association with the Guerrilla Girls.
Plaintiffs and defendants have conflicting accounts of the involvement of “Romaine Brooks” (Susan Doe 4) and “Zora Neale Hurston” (Susan Doe 5) in this litigation. The woman that plaintiff Erika Rothenberg knows as “Romaine Brooks” informed her in March 2004 that she had not agreed to participate in the present litigation, but the defendants have submitted an authorization dated October 2003 by “Romaine Brooks” permitting them to include her in it. Similarly, in March 2004 Rothenberg spoke to the woman she knows as “Zora Neale Hurston” and was told that “Hurston” had not agreed to sue, but defendants have submitted an authorization dated October 2003 and a declaration dated June 2004 from “Zora Neale Hurston” in support of her involvement in the litigation. To conduct a trial in such an atmosphere, all the while using only pseudonyms, promises trouble and confusion.31
The plaintiffs were members of a writers’ collective, and chose as pseudonyms the names of famous women writers; but similar problems might have arisen if they had chosen more arbitrary pseudonyms. Real-first-name-plus-initial (e.g., Erika R. if Rothenberg had wanted a pseudonym) would be much clearer, though at the price of increasing the risk that the plaintiff could be identified. Real initials (e.g., E.R.) might likewise have been clearer. And it would have been clearer still if the parties had appeared under their real names, which is what the court ultimately insisted on. Nonetheless, if there is just one pseudonymous party, there should be much less risk of such confusion.
If I’m right, then it may make sense for courts to adopt something like the following rule — whether for an entire court system, as a local rule, or as a chambers practice:
a. Whenever any party is pseudonymous, that party’s name in the caption must be distinctly labeled in the caption as a pseudonym.
b. If a known individual litigant, witness, or victim is using a pseudonym selected for purposes of the litigation, the pseudonym should ordinarily be a first name followed by an initial (e.g., “Gunther K.”), with neither corresponding to the person’s actual names.
c. Unknown defendants, witnesses, or victims may be labeled using “Doe” or a similar name.
d. If there are multiple pseudonymous individuals involved in a case, such that the practice set forth in (b) is likely to prove confusing, the individuals’ actual first name and last initial, or actual first initial and last initial, or actual first, middle, and last initials, may be used instead (unless this creates an unreasonable risk of harm stemming from the possible identification of the individuals).
e. This rule does not apply to pseudonymous corporate or organizational litigants.
f. This rule does not affect the procedure for deciding whether a party may proceed under a pseudonym, or the substantive criteria for when such pseudonymity is appropriate.
Plaintiffs would usually know to follow such a rule. But if they don’t (for instance, if they are unaware of it), a judge can order them to refile the complaint with the original pseudonym (e.g., Doe) replaced with a revised pseudonym — just as judges who deny pseudonymity altogether order Doe plaintiffs to file complaints with the pseudonym replaced with the real name.32 And judges can also follow this rule as a guide when they themselves pseudonymize parties in issuing opinions, and when they pseudonymize witnesses and victims.
This approach should help ensure that the “rare dispensation” of pseudonymity is performed in a way that both protects privacy and minimizes confusion.
EUGENE VOLOKH is the Gary T. Schwartz Professor of Law, UCLA School of Law (firstname.lastname@example.org).