"Interesting questions are raised by present speech patterns"

« previous post | next post »

A recent PhD Comics strip:

It's good to see that the author (Jorge Cham) actually uses (or has Tajel use) the terms "third person" and "passive voice" correctly. To see how rare this is, you can browse Geoff Pullum's summary of "passive voice" discussion, or the more recent list in "The Passivator reborn" (1/10/2021), or the results of searching LLOG for passive voice.

And the comic also accurately illustrates the pressure to chose fancier rather than simpler words (e.g. "acceptable" rather than "fine" or "OK").

"Teaching Zombie Rules" (2/26/2009) noted the common style-guide constraint about using passives to avoid first-person pronouns in technical writing, citing stylistic advice from (now-defunct) style guide from the Acoustical Society of America:

Use passives instead of pronouns "I" and "we," e.g., "It was noted" instead of "We noted."

The ASA's current style guide puts it differently, managing to make the avoidance of first-person pronouns consistent with the avoidance of passive:

Many authorities on good writing emphasize that authors should use the active rather than the passive voice. Doing so in scholarly writing, especially when mathematical expressions are present, is often infeasible, but the advice has merit. In mathematical derivations, for example, some authors use the tutorial “we” to avoid using the passive voice, so that one writes: “We substitute the expression on the right side of Eq. (5) into Eq. (2) and obtain …,” rather than: “The right side of Eq. (5) is substituted into Eq. (2), with the result being … .” A preferable construction is to avoid the use of the tutorial “we” and to use transitive verbs such as “yields,” “generates,” “produces,” and “leads to.” Thus one would write the example above as: “Substitution of Eq. (5) into Eq. (2) yields … .” Good writers frequently go over an early draft of a manuscript, examine each sentence and phrase written using the passive voice, and consider whether they can improve the sentence by rewriting it.

In general, personal pronouns, including the “tutorial we,” are preferably avoided in scholarly writing, so that the tone is impersonal and dispassionate. In a few cases, it is appropriate that an opinion be given or that a unique personal experience be related, and personal pronouns are unavoidable. What should be assiduously avoided are any egotistical statements using personal pronouns. If a personal opinion needs to be expressed, a preferred construction is to refer to the author in the third person, such as: “the present writer believes that … .”

(Of course it's worth noting that "Substitution of Eq. (5) into Eq. (2)"  names the object and goal of the substituting, while leaving the agent vague. So if the goal of avoiding passive voice is to avoid being vague about agency, as opposed to appeasing the ghosts of Strunk and White…)

And I have the impression that the constraint about avoiding "we" (or even "I") is increasingly relaxed, though I don't have time today to do an empirical study of the historical trends.


  1. David Marjanović said,

    January 16, 2022 @ 11:39 am

    Definitely. I have never referred to myself as "the present writer" or "the senior author" or any such circumlocutions of the obvious; they just seem embarrassing to me. I've never been criticized for this either.

    I've often written "we" because only published alone once yet… but in that paper I didn't make any special effort to avoid "I".

    But I wouldn't write about general facts like "substitution of Eq. (5) into Eq. (2) yields" in the first person anyway…

  2. D.O. said,

    January 16, 2022 @ 12:02 pm

    There are some grant agencies that explicitly prohibit using personal pronouns in grant applications. This makes for extremely interesting linguistic contortions because the whole text is about what the applicant intends to do. The author of this comment am not sure if the third person self reference is allowed.

  3. Tim Finin said,

    January 16, 2022 @ 12:13 pm

    A benefit of using "We" rather than "I" is that it suggests it's not just one person presenting their idea or opinion, but a whole team sharing their consensus!

  4. Laura Morland said,

    January 16, 2022 @ 12:53 pm

    Maybe my brain is just tired, but could someone give an example of a "FIRST person passive voice"… if it exists?

    "An affirmative answer is given by me" wouldn't cut it, IMHO, but I would be pleased to stand corrected.

    [(myl) The "first person" part means that the subject is "I" or "we". The "passive" part means that the verb is in the passive voice, e.g. "I was puzzled by their answer" or "We were swept away by the current".

    The relevance of "third person passive voice" in the context under discussion is that it can offer ways to avoid first-person pronouns, as in the example of "it was noted" in place of "we noted", or "a random sample was selected" in place of "we selected a random sample".

    "First person passive" constructions seem unlikely to come up in technical writing, though maybe I'm being misled by habit :-)…]

  5. J.W. Brewer said,

    January 16, 2022 @ 1:04 pm

    There is I think an important distinction between an authorial "we" referring to the authors of a co-authored paper (and/or a research team, whether or not all of its members are credited as co-authors) and the "tutorial we" referred to by the ASA style guide. The former refers to identifiable individuals who may or may not be correct about whatever they assert; the latter has no particular fixed referent in the world; it is, if not literally a dummy subject like "weather it," the functional equivalent of an indefinite/impersonal pronoun like "one" in a clause like "herbarium material does not, indeed, allow one to extrapolate safely." The tutorial we thus performs the same pragmatic function as agentless passive constructions, viz. it allows the author to be "vague about agency" when that is thought to be affirmatively desirable because the author is supposedly making objective scientific claims whose value does not depend on the identity and subjectivity of the particular writer(s). So objecting to the passive-construction style of scientific writing but thinking the better alternative is active-voice constructions with "tutorial we" as the agency-and-accountability-free syntactic subject does not seem like a meaningful improvement to me.

    Somewhat relatedly, I was musing about the use of imperative voice in mathematical proofs (and perhaps some other scientific/technical context?), e.g. "Let ABC be any right triangle, with right angle C.* Draw the altitude CF from the right angle to the hypotenuse …" This seems like the same use of the imperative found in genres like recipes and directions, and shares the same "impersonal" vibe, because the speaker is not explicitly syntactically referred to in the discourse and thus the speaker's specific identity is easier to treat as irrelevant. (Obviously, there may be situations in which whether or not you will try a particular recipe depends on your opinion of the culinary talents of who wrote it, etc.) What I find interesting on reflection is that there are so many other contexts in English in which speakers routinely shy away from pure imperatives because they seem aggressive or impolite, leading us to paraphrase "Do X" as something softer like "would you mind doing X?" Yet in the proof/recipe/directions genre, the bare imperative seems perfectly neutral in tone rather than impolite or pushy.

    *The "Let (PRO)NOUN VERB (OBJECT)" construction in English is functionally equivalent to what I think is called the jussive subjunctive in Latin. As is common, I never focused on the oddity of the English construction until I took high school Latin and was taught that as the English analogue of a particular Latin inflected form. Perhaps English is more honest in making it an overtly imperative construction, although that's probably too moralistic a way of describing cross-linguistic variation in morphosyntax.

  6. J.W. Brewer said,

    January 16, 2022 @ 1:06 pm

    Due to the infirmities of increasing age, my prior comment said "imperative voice" when it obviously should have said "imperative mood." Unless there's some better terminology to be used for English that Huddleston/Pullum (or whoever) have devised but which has failed to stick in my mind.

  7. David Eddyshaw said,

    January 16, 2022 @ 1:14 pm

    "personal pronouns, including the “tutorial we,” are preferably avoided in scholarly writing"

    Except, apparently, "it", of course, which they wrongly call an "impersonal pronoun" (along with "these.") They've progressed beyond not knowing what a passive is, to not knowing what a personal pronoun is.

  8. Scott P. said,

    January 16, 2022 @ 1:36 pm

    Maybe my brain is just tired, but could someone give an example of a "FIRST person passive voice"… if it exists?

    Loquor linguam latinam.

    [(myl) That phrase is meant to mean "I speak Latin", where the verb loquor is deponent, so that it has active meaning though passive morphology. (And also doesn't use a direct object to indicate the language or mode of speaking…) A genuine first-person passive in Latin would be something like "Numquam deterrebor" (= "I will never be deterred").

    Other English examples of "first person passive voice" phrases would be things like "I was impressed by her eloquence" or "We will be greeted by the class president".]

  9. Wanda said,

    January 16, 2022 @ 1:56 pm

    In my fields of biology, we typically use "we" in papers and grant proposals. "We" is appropriate in our papers because the science is almost universally done in teams. I would also say that plainer, punchier language is appreciated, although not all authors have the skill to write that way when there's a lot of necessary jargon. When I write with students, I have to train them out of a lot of bad habits that are based on what they think academic writing is like.

  10. David Marjanović said,

    January 16, 2022 @ 2:16 pm

    A benefit of using "We" rather than "I" is that it suggests it's not just one person presenting their idea or opinion, but a whole team sharing their consensus!

    But that only works if the paper actually has several authors. Otherwise it's just distracting (although still usual in French – with singular adjectives & participles for lone authors!).

    Maybe my brain is just tired, but could someone give an example of a "FIRST person passive voice"… if it exists?

    "An affirmative answer is given by me" wouldn't cut it, IMHO, but I would be pleased to stand corrected.

    "I am (being) given an affirmative answer."

    The "Let (PRO)NOUN VERB (OBJECT)" construction in English is functionally equivalent to what I think is called the jussive subjunctive in Latin. As is common, I never focused on the oddity of the English construction until I took high school Latin and was taught that as the English analogue of a particular Latin inflected form. Perhaps English is more honest in making it an overtly imperative construction

    It's (theoretically) different in that the subjunctive expresses a wish here: "may (PRO)NOUN VERB (OBJECT)".

    A few vestiges of this, or perhaps calques from Latin, do occur in English: Rule Britannia – Britannia rule the waves is a wish. The lack of commas makes explicit that you're not giving a presumptuous order to Britannia. Likewise America, America, God shed his grace on thee.

  11. m said,

    January 16, 2022 @ 2:55 pm

    "Unto us a child is given" … ?

  12. J.W. Brewer said,

    January 16, 2022 @ 3:21 pm

    For Laura Morland's query, here's an excerpt from a 1956 decision of the Supreme Court of Wisconsin:

    'Against the contention of appellant, there was testimony by Walter Lohman, director of purchases for Nesco up to March, 1953, presently employed by the Heil Company, and a disinterested witness, that there was a day-to-day agreement and that "no agreement or other answer was given by me which would have led to Wolinsky's assuming that he had such a [five-year] contract."'

    Obviously it's possible that either the context of the preceding questions or the formal/artificial nature of giving sworn testimony may have caused Mr. Lohman to use a syntactic structure he would not have typically used in more spontaneous conversation.

  13. A. Sasportas said,

    January 16, 2022 @ 4:34 pm

    "Loquor linguam latinam" does not exist in native Latin.

    Loquor is a form of the Latin deponent verb loqui. A deponent verb has the form of a passive verb and the meaning of an active one.

    It does not take a direct object (linguam latinam).

    The Latin for 'Do you speak Latin?', for example, is LOQVISNE LATINE, with the adverb LATINE (literally, 'in Latin'), to which one may answer:

    LATINE LOQVOR 'Yes, I speak Latin' or LATINE NON LOQVOR 'No, I do not speak Latin'.

  14. Max Wheeler said,

    January 16, 2022 @ 6:10 pm

    RE: A Sasportas
    LOQVERISNE LATINE?, since, as you remark, LOQVOR is a deponent verb.

  15. Roy Sablosky said,

    January 16, 2022 @ 8:21 pm

    I am a copyeditor in the peer-reviewed literature and I disagree with the Acoustical Society of America's advice that authors avoid referring to themselves as "I" or "we." This might make you *sound* more "impersonal and dispassionate," but it won't actually make you more impersonal or dispassionate, or help your reader in any other way. In my view "the present writer believes that . . ." is wordy, awkward, and pretentious.

  16. A. Sasportas said,

    January 16, 2022 @ 9:31 pm

    @ Max Wheeler. Of course you are right.

  17. Rick Rubenstein said,

    January 16, 2022 @ 9:42 pm

    The important thing is to ensure scholarly writing is sapped of any sense of excitement; otherwise people might be inclined to read it, and then where would we be?

  18. Bob Ladd said,

    January 17, 2022 @ 4:46 am

    Vaguely relevant fun fact: In Hungarian, verbs in recipes are generally in the 1st person plural (unlike the imperative or infinitive used in recipes in most Western European languages). This would seem to count as an instance of the "tutorial we".

  19. David Marjanović said,

    January 17, 2022 @ 9:16 am

    "Unto us a child is given" … ?

    Third person.

    no agreement or other answer was given by me

    Third person.

    This would seem to count as an instance of the "tutorial we".


    The obsolete but formerly ubiquitous way of beginning a recipe in German was Man nehme, with the impersonal pronoun (3rd person singular) and "take" in the present subjunctive, expressing more or less a wish that "one take" the following ingredients and do stuff with them.

  20. Thomas said,

    January 17, 2022 @ 12:15 pm

    Thank you @J.W. Brewer for your enlightening comment about the usage of tutorial we. After writing scientific mathematical texts for a long time, I don't even bat an eye anymore at this style. It is interesting to think about it. Writing a mathematical text is mainly wrapping up logical deductions into a cascade of tutorial we, “soft“ imperative formulations and passive voice. They are interchangeable and carry no special connotations to me anymore. The passive voice is not very popular because it takes more words to convey the same content.

    I am also quite interested in how we can separate this “tutorial we” from what one might call the “nurse we”, named after the profession most inclined to use it. I have to say, the conflation of these two styles sometimes comes to my mind when I have to dumb down mathematical passages to the layman.

  21. Martin said,

    January 17, 2022 @ 4:17 pm

    J. W. Brewer above is correct to draw a distinction between the authorial and tutorial 'we'. In scientific papers at least it's necessary to have a way of marking work done by the current authors rather than work that might have been done by others in the past, and authorial 'we' serves that purpose: and, of course, single-author papers then logically use the first person singular. I've written a number of single-author papers in more than one discipline in this style and no proofreader or editor has ever raised any objection.

  22. Alexander Pruss said,

    January 18, 2022 @ 6:18 pm

    Halmos' "How to write mathematics" says to use "we" for the author and the reader. This forces a certain thoughtfulness. Thus: "I claim that …" should be used when the reader is not yet in a position to make the claim but "We will see that …" is perfectly fine since one can charitably presume the reader will go along with it once it's proved. I've tried to follow this approach in my mathematical writing pretty consistently.

  23. Matt Juge said,

    January 19, 2022 @ 12:12 am

    A little more on loquor. While linguam latinam may not be typical with this verb, it can take a direct object (e.g., Plautus Persa, act 2 scene 3: Rem loquitur meram "He's speaking the very fact" (per Leo's edition).

  24. Bloix said,

    January 19, 2022 @ 10:32 am

    The New Yorker used to frown on the first person. In the long profiles of noteworthy individuals, which are based on interviewing and just hanging out with the subject, the writers would resort to work-arounds, like [this is an invented example] – "he motioned to his visitor to have a seat, but all the chairs were stacked with books, prints and paintings."

  25. Andrew Usher said,

    January 21, 2022 @ 7:17 pm

    Matt Juge:
    That difference is not about 'loquor' or even Latin, but about different kinds of object. A language is not something you speak, as in say, but something with which you may speak. I think it's English that's anomalous for not distinguishing with the verb 'speak', but neither 'I speak Latin' nor 'I am speaking Latin' mean the same category as with 'the truth' or similar as object. If you're American you may have contemplated what 'Se habla espanol' literally means, and why it's not 'We speak Spanish'.

    k_over_hbarc at yahoo dot com

  26. SusanC said,

    January 23, 2022 @ 7:04 pm

    In a mathematical proof, particularly, "I":seems out of place.

    Although a proof was probably originally found by somebody or something (e.g. a computer program) there's a notion of it being a replicable list of instructions that can be carried out by anyone,

  27. SusanC said,

    January 23, 2022 @ 7:32 pm

    Although scientific writing does often resort to the passive voice, there is another way in which it is very explicit about agency. The author list at the start of the paper explicitly names the people who did the experiments that are going to be described, and who paid them.

    So "we":seems quite in order.

    In my department at least, the convention is the the author list includes the people who actually carried out the experiments, not just the people who did the write-up. (Minor contributions can be put in the acknowledgements section, where you would actually name who did what. ); so the "we" really is either some unspecified subset of the author list or someone who is named in the acknowledgements.as having done the thing.

  28. Philip Taylor said,

    January 24, 2022 @ 5:41 am

    Susan, may I ask what the colon signifies in « "I":seems out of place » and « "we":seems quite in order » ?

  29. Jonathan said,

    January 26, 2022 @ 2:39 pm

    Is a nurse saying "How are we feeling today?" a tutorial we? How about a waiter saying "How did we find the steak?" (My stock answer to those questions are "Depends on your feelings and the aggregation function employed" and "It was right here. The kitchen didn't really do much to hide it.")

RSS feed for comments on this post