Is it or is it not good practice to make an author the subject of sentences in a paper? When should their name be central, discreet, or absent in the sentence? Here are five variations showing different approaches, four of which include the author’s name. Are there particular identifiable use cases for these styles?

  1. Johnson argued that this will never work.
  2. Per Johnson, this will never work.
  3. The argument was advanced by Johnson that this will never work.
  4. This will never work (Johnson).
  5. This will never work [42].
  • 1
    What? Of course it's good practice to use names as subjects. Why is this even a question?
    – JeffE
    Jan 17, 2015 at 23:30
  • 1
    JeffE, FWIW, a researcher i trust told me he had a strong preference against it, so i wanted to ask the group. Jan 18, 2015 at 15:56

4 Answers 4


There is more here than style: the first and third options mean something quite different from the fourth and fifth. In the former cases, what you are asserting is that someone else argued for X. In the latter cases, you are asserting X and using the citation as evidence/support/proof. (That I can't quite tell where the second one fits into this dichotomy is a strike against it.) In an academic paper, that is a not so subtle difference.

I find the style question less critical. It is a matter of general good writing rather than anything specifically academic or it is specific to the journal at hand (so we need not discuss it here).

Of course you can use an author's name as a subject of a sentence: you can write what you want, you know! As a matter of style, to my ear the first option sounds good, the second option sounds weird, and the third option sounds weaker and wordier than the first, but maybe the surrounding text gives you a good reason to write it that way.

The difference between options 4 and 5 is just a difference in citation style. First that is very field dependent; in my field (mathematics), we would do 5 rather than 4; in much of the humanities it would be the other way around. Second, unless your choice is so so strange that it prevents your readers from finding the references in your bibliography, the whole issue can probably wait until your paper gets accepted, in which case they'll either do it for you, tell you exactly what to do, or tell you that you did it wrong (and ask you to fix it).

  • 1
    I hate references which are by number only. I much prefer alphanumeric references. And the first time you use an important reference you should use the author's name also, e.g., "...will never work (Johnson [42])."
    – Kimball
    Jan 17, 2015 at 9:34
  • 2
    @Kimball Unfortunately, you will typically get zero input into this decision for peer-reviewed papers, as it is almost always dictated by the venue in which you are publishing.
    – jakebeal
    Jan 17, 2015 at 13:22
  • 1
    @Kimball: There is actually a subtlety here that in math, whatever else we do, a formal citation is enclosed in square brackets [Ki15] (and I was referring to that as much as anything in my answer) but the grammatical status of [Ki15] is slightly fluid. I think it is a good practice to drop the authors' names the first time you make a citation (unless you're rattling off a long list), but whereas the [Ki15] can be tucked in at any convenient point in the sentence, the name should be part of the sentence. Thus: "a result of Kimball [Ki15]", not "a result (Kimball [15])". Jan 17, 2015 at 14:36
  • 1
    I prefer (superscripted) numerical references, but in general the numerical types are not part of the sentence. So you might say "This is a fact[12]" in most cases. You can also use "Einstein stated that this is a fact[12]" or some similar form where it's important to do so. Sometimes you have to use "The data was taken from reference 12". Note the lack of square brackets -- in this case "reference 12" is part of the text. Thus numerical styles can support both a full range of cited assertions without breaking the flow of the text.
    – Chris H
    Jan 17, 2015 at 19:33
  • 2
    @JeffE Why are they footnotes? The way they seem to generally be used in math are as short strings referring the paper, and can thus be used nicely as part of a sentence. This provides a lot more flexibility when one would like to specify more precisely how the cited paper is relevant, or for example wants to point out differences in notation. Jan 18, 2015 at 19:22

Version 2 seems stilted; 3 is an unnecessarily verbose use of the passive voice. The choice between 4 and 5 is really a matter of the journal's style guidelines rather than an active decision you will get to make as the author.

So the real choice here is between 1 or 4/5. The key thing to note is that you sentence draws attention to an actor and an action through your choice of subject and verb. (See Joseph Williams, Style: Lessons in Clarity and Grace for a superb discussion of this and many related principles)

In 1, the actor is "Johnson" and the action is "argued". In 4/5, the action is "This" and the action is "will not work". Thus if your intent is to draw attention to the fact that Johnson made this claim, as you would e.g. if discussing a history of ideas, version 1 might be preferred. If instead your intent is to draw attention to the claim itself, and the reference to Johnson is simply a matter of good scholarship, 4/5 will be preferred.


In addition to other good points made, there is the question of whether you wish to assert a thing, or only assert that someone else asserts it. That is, if you write "Johnson asserts X." then (from that sentence alone) it is not clear whether you agree, disagree, or are neutral. If, instead, you write something like "One might consider X. For example, see [Johnson]." then you are at least tentatively asserting X, with Johnson for corroboration. This distinction might matter more than style... although I'd agree that avoiding circumlocutions and verbosity is generally good (=more readable).


To me, formulation 1 would appear useable only in special contexts, such as

Johnson argued that this will never work ([Jo97]), but later Miller found a way to get rid of the obstacles in many important cases ([Mi04] and [Mi04a]).

In other words, you want to express Johnson's opinion without sharing it (and of course you back this up with citations). In general, 4 or 5 seems to be preferred (e.g., [Corvus])

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .