Sunday, June 06, 2021

When do you use et al. as opposed to listing out the authors? First names? Middle initials? Jr?

 If I was refering to the paper with bibtex entry: 


  author    = {Jeffrey Bosboom and

               Spencer Congero and

               Erik D. Demaine and

               Martin L. Demaine and

               Jayson Lynch},

  title     = {Losing at Checkers is Hard},

  year      = {2018},

  note      = {\newline\url{}},


(The paper is here.)

I would write 

Bosboom et al.~\cite{BCDDL-2018} proved that trying to lose at checkers (perhaps you are playing a child and want to keep up their self-esteem, or a Wookie and don't want your arms to be torn off your shoulders   (see here),  or a Wookie child) is hard. 

Why did I use et al. ? Because it would be a pain to write out all of those names. 

How many names does it take to make you write et al. ? Are there exceptions? 

I have not seen any discussion of this point on the web. So here are my rules of thumb and some questions.(CORRECTION- a commenter points out that this IS discussed on the web. Even so, you can comment on my thoughts or give your thoughts or whatever you like.) 

1) If  there are 3 or more people than use et al. Exception: If the three are VERY WELL KNOWN as a triple. E.g., the double play was Tinker to Evers to Chance. Well, that would not really come up since in baseball nobody ever says the double play was Tinker et al.  More relevant examples:

Aho, Hopcroft, and Ullman

Cormen, Leiserson, and Rivest, also known as CLR

The more recent edition is

Cormen, Leiserson, Rivest, and Stein. I have heard CLRS but I don't know if people WRITE all four names. 

Lenstra-Lenstra-Lovasz also usually mentions all three. 

2) If there is one name do not use et al.  unless that one person has a multiple personality disorder.

3) If there are 2 people it can be tricky and perhaps unfair. If the second one has a long name then I am tempted to use et al. For example

Lewis and Papadimitriou (If I mispelled Christos's name-- well- that's  the point!- to avoid spelling errors I want to use et al. )

Lampropoulos and Paraskevopoulou (the first one is UMCP new faculty!). After typing in the first name I would not be in the mood to type in the second. 

Similar if there are lots of accents in the name making it hard to type in LaTeX (though I have macros for some people like Erdos who come up a lot) then I might use et al. 

(ADDED LATER- some of the commenters object to my `rule' of leaving out the last name if its complicated. That is not really my rule- the point of this post was to get a discussion going about the issue, which I am happy to say has happened.) 


There are other issues along these lines: when to include the first name (when there is more than one person with that last name, e.g. Ryan Williams and Virginia  Vassilevska Williams), when to use middle initials (in the rare case where there is someone with the same first and last name- Michael  J. Fox added the J and uses it since there was an actor named Michael Fox.)

I will soon give a quote from a math paper that amused me, but first some context.  The problem of determining if a poly in n variables over Z has an integer solution is called E(n). By the solution to Hilbert's 10th problem we know that there exists n such that E(n) is undecidable. E(9) is undecidable, but the status of E(8) is unknown (as of May 2021) and has been since the early 1980's. 

Here is the quote (from here).

Can we replace 9 by a smaller number? It is believed so. In fact, A. Baker, Matiyasevich and J.Robinson  even conjectured that E(3) is undecidable over N.

Note that Baker and Robinson get their first initial but Matiyasevich does not.

I suspect that they use J. Robinson since there is another mathematician with last name Robinson: Rafael Robinson who was Julia's Robinson's husband (to my younger readers--- there was a time when a women who got married took her husband's last name). There is at least one other Math-Robinson: Robert W Robinson. I do not think he is closely related. 

Baker: I do not know of another mathematician named Baker. I tried Google, but the Bakers  I found were   not in the right time frame. I also kept finding hits to an anecdote about Poincare and a man whose profession was a baker (see here though its later in that blog post). However, I suspect there was another mathematician named Baker which is why the author uses the first initial.  Its possible the author did not want to confuse Alan  Baker with Theodore Baker, one of the authors of Baker-Gill-Solovay that showed there were oracles that made P = NP and others that made P NE NP.  But somehow, that just doesn't seem right to me. I suspect there is only one mathematician with last name Matijsavic. 

Thomas Alva Edison named his son Thomas Alva Edison Jr.  This was a bad idea but not for reasons of authorship, see here.


  1. "I have not seen any discussion of this point on the web."

    I am very confused by this statement. There are tons and tons of discussions of this online, on style manual websites, blogs, and even twitter. Googling "how to use et al." brings up a page of websites discussing this issue. For example, the APA recommends using it starting at three authors, while the Chicago Manual of Style recommends listing three authors and using et al. starting at four.

  2. You should not be confused- I am simply WRONG which is not confusing. I have modified the post.

    Which of the sources you site do you agree with, if any?

    1. I use \citep, \citet, and \citeauthor and allow them venue’s style guide to determine how they are rendered in-line. In the bibliography proper, I list all authors with their full names regardless of length, language of origin, or how lazy I’m feeling.

      In the rare instances where I am not typesetting in BibTex I work on a case-by-case basis. For example, I recently co-authored a machine learning paper with 52 authors. The paper has two joint first-authors, while the rest of the authors are listed in reverse alphabetical order. I thought about how to list this on my website and CV, and I chose to list it as “Caswell*, Kreutze*, and 50 others (incl. Biderman).” This decision is largely predicated on two factors: a desire to acknowledge and explicitly credit the two joint first authors of the paper, and an understanding of the fact that when you’re reading my website or my resume you are most likely uninterested in the author list for its own sake but rather are interested in reading about my personal contributions.

      The paper I have coauthored with the second largest author list has ten authors, myself listed second. On my CV and website I list all ten authors. It doesn’t take up that much space, and in particular doesn’t push the citation on to an additional line so I see no reason to trim it. The 52 author paper would have had an authorship list several lines long, and longer than the rest of the citation. I looked at how that appeared written out in full on the page and decided that the full authorship list was overly distracting, and unlikely to server the reader’s interests.

      When lecturing or giving a talk, I follow a convention that I picked up from being a student of Babai (though I have no idea if it is original to him). Often time space is at a premium on the white board and on lecture slides, so I compress the authorship list by listing the first initial of the last name of each author. For example, I would refer to the authors of “Algebraic methods for interactive proof systems” as LFKN. However I make a point to include ending matter that has the full citation, and the first time I come across a slide with such an abbreviation I state this fact explicitly and say the names of each of the authors aloud the first time each paper is referenced. As an industry algorithms researcher I don’t spend much time lecturing a room in front of a white board, but I feel that any lecture wherein I feel obligated to cite the authors of a particular paper is a lecture wherein I should make a full bibliography available to the attendees. Depending on the context, this likely means via a follow-up email, printed copies provided at the door, and/or a full listing online.

  3. Who the hell is "Vanessa Williams"? The singer? Really, Bill?

    She's been a tenured professor for several years now, and she's been publishing in TCS for over 15 years. Maybe it's time to learn her name? (You aren't the only one who continues doing this, just the latest one.)

    Students reading: please do the opposite of all this horribly lazy and disrespectful advice. Be scholars.

    1. 1) I have made the correction-thanks for letting me know.
      2) My post and the comments have lots of advice, so when you say to do the opposite, I am not sure what you want are advocating. Please clarify.

    2. To clarify, I said "Be scholars." You were advocating the opposite of that.

  4. Why no just use \citeauthor and let the style take care of it?

  5. You can't look at it in isolation. When the TCS convention is to use alphabetical order for authors and to refer to their work as a collective its ill advice to propose using the first author alone.

  6. > Similar if there are lots of accents in the name making it hard to type in LaTeX (...) then I might use et al.

    This is de facto quite discriminatory. Please re-think your citing ethos.

    1. It is also discriminatory to mention only the last name.

      In the western world, it may seem sensible to say "Papadimitriou proved BLA about Nash equilibria". But in other parts of the world last names are less varied, and "Kim proved BLA about Kim equilibria" can quickly become confusing.

      I think Knuth's insistence on listing full names, including with the original spelling, has some merit: (But, yes, it requires extra effort.)

  7. No no ... I agree w/ u Bill.
    It's a good topic and I haven't seen it widely discussed
    from a TCS's point of dealing.
    @Stella, APA? Hardly any literate student of TCS would follow
    those guidelines.

  8. From STOC 2021 Call for Papers:
    "Recommended best practices for citations: Authors are asked to avoid "et al." in citations in favor of an equal mention of all authors' surnames (unless the number of authors is very large, and if it is large, consider just using \cite{} with no "et al."). When not listing authors' names, citations should preferably include the first letters of the authors' surnames (or at least the first three followed by a +, and possibly the year of publication). If using BibTeX, this can be accomplished by using \bibliographystyle{alpha}."

  9. I recently wrote out a list of six authors on the first occurrence of a citation because it was a very important related work. In the subsequent citations of the work I just wrote First author et al.. I think in fields where authors are ordered alphabetically it is important to try to avoid et al. because it discriminates against authors with last name late in the alphabet.

  10. > After typing in the first name I would not be in the mood to type in the second.

    How very embarrassing for you to admit to not knowing how to operate copy/paste.

  11. Embarrassing? Never! Whenever I make a mistake or admit a lack of knowledge or expertise, and someone points this out and corrects me, I feel enlightened, not embarrassed. Thanks for enlightening me (I am not being sarcastic).

  12. (1) Please correct Vanessa Williams to Virginia Vassilevska Williams. (2) Please correct Matijasevich's lastname. (3) Please reconsider at least your two-authors policy. Is this really what you're willing to continue doing? Is this really what you want younger readers to see here?

    1. (1) Done- including the middle name. Non-rheotorical question- why is putting the middle name in important? (I DID it so this is not an argument about it)
      (2) Done- I have seen his name spelled several different ways, though I suspect the version I had was none of them.
      (3) I don't really have a policy- this post was more of an attempt to get a conversation going, which it has. I have clarified that in the post.
      THANKS for your comments which have made me correct and clarify my post.

  13. This post is satire, right? Almost fell for it!

    1. Yes and No.
      The notion of not including the second author if the name is to long is, of course, a joke.

      The question of what IS a good convention that balances fairness and length-of-exposition, is being asked seriously.

  14. It should be "Tinker to Evers to Chance", not Tinkers.

  15. The convention in TCS is quite standard at this point, and is captured, e.g., by latex's \citet command: "first two authors" for two-author papers, and "first author et al." for >=3 authors. If you want to credit all authors (increasingly encouraged, as mentioned above), then use \citet*{} to get all author names. No danger of typos in people's names, and no need to worry about "complicated" (non-standard/not-sufficiently-Western?) last names. As someone who has been "et al.-ed" in a survey discussing his two-author paper (and is likely to be last author on most of his papers, due to alphabetical ordering), your point about just using et al. for two authors, despite the common convention of listing both, hit close to home for me. I join Sasha G's call for you to reconsider your two-author citation policy, and even removing this unfortunate line from your post. This is not the kind of message I would expect from Fortnow et al.'s blog.

  16. My thoughts: Remember who you are writing the paper for, it's not you and it's not the people you cite. Use "et al." when it makes it simpler for the reader not to have to peruse a long list of names.

  17. For Baker: How about Brenda Baker?

  18. Can we get a synoptic guideline conclusion on this?

    Lance says depending on complexity and number/length of names,
    "et al." is a viable alternative.

    Don Knuth doesn't always cite the full names in the body of
    the paper, what he does do, is cite the full names (including any middle names) in the reference section. That's a separate issue altogether. It turns out that some authors don't want their middle name to be spelled out in full, though. But Don believes that this is giving the author full credit for his or her work. (Does one nowadays also use 'its' work? Gender neutrality? Is this a common enough thing ... ?)
    Someone pointed out that Don also insists on the original spelling of the name ... yes, that's only done for TAOCP -- so hardly entertain this option for ordinary papers.

    Then there are LaTeX primitives that handle the assignment of names accordingly. Should we really use inbuilt LaTeX primitives to deal with this? Does the AMS-Latex style have anything on this? And how many people are using these to handle this issue?

    1. You're looking for "their," not, "its." It's a common enough singular pronoun, people. Let's not dehumanize.

  19. The following comments are based on

    The policy adopted this time by STOC regarding the use of the "first author et al." format is a wonderful step towards addressing problems due to alphabetical ordering norms (thank you to Virginia and other organizers). There are still primacy issues though, which coupled with the convention of alphabetical ordering, creates undesirable effects.

    We frequently use randomization in our algorithms -- why not use it here, and randomize the author ordering? :)  Yes, any individual paper will still have primacy effects, but for any author, effects would cancel out across multiple papers. Would love to know all your thoughts on this.

    Btw a similar issue holds in websites, and the theory group's website at CMU dynamically randomizes the ordering of people