Linguistic commentary from a guy who takes things too literally

Whomever Is Never Actually Right

Posted by Neal on October 21, 2007

My wife and I watched this week’s episode of The Office last night, which featured the following scene (20:55 into the online version, accessible here):

Ryan: What I really want, honestly Michael, is for you to know it, so that you can communicate it to the people here, to your clients, to whomever.
Michael: [chuckle] OK.
Ryan: What?
Michael: It’s whoever, not whomever.
Ryan: No, it’s whomever.
Michael: No, whomever is never actually right.
Jim: Sometimes it’s right.
Creed: Michael is right; it’s a made-up word used to trick students.
Andy: No. Actually, whomever is the formal version of the word.
Oscar: Obviously, it’s a real word, but I don’t know when to use it correctly.
Michael: [aside] Not a native speaker.
Kevin: I know what’s right, but I’m not gonna say, because you’re all jerks who didn’t come see my band last night.
Ryan: Do you really know which one is correct?
Kevin: I don’t know.
Pam: It’s whom when it’s the object of a sentence, and who when it’s the subject.
Phyllis: That sounds right.
Michael: Well it sounds right but is it?
Stanley: How did Ryan use it, as an object?
Ryan: As an object.
Kelly: Ryan used me as an object.
Stanley: Is he right about that?
Angela(?): How did he use it again?
Toby: It was, “I wanted Michael” — subject — “to explain the computer system” — the object —
Michael: Thank you!
Toby: — “to whomever,” meaning us, the indirect object, which is the correct usage of the word.
Michael: No one asked you anything, so why don’t whomever’s name is Toby, take a letter opener and stick it in your skull.

The fact that such a discussion can realistically take place is an indicator of the moribund status of whom (nicely summed up by Geoff Pullum here). In the dialogue we get just about every view of the word: rejection of it as a hypercorrection (Michael, Creed); acceptance but with admitted ignorance of how to use it (Oscar, Kevin); a statement of what’s becoming the new, sociolinguistically rather than syntactically based rule for usage (Andy); and the fairly accurate statement of the Standard English rule regarding whom (Pam, Toby).

Toby, however, overlooks that in I want Michael to explain…, it’s not obvious that Michael is a subject. I is the subject of want, and Michael is its direct object, as evidenced by the fact that it can be replaced be him, but not he. On the other hand, Michael seems to be the subject of explain in some sense, as he’s the one who will do the explaining. A lot of syntactic theorizing has taken as its starting point facts like this one, arguing whether Michael is a direct object, or a subject that has been “raised” to become a direct object, or perhaps something else. Even so, Toby’s and Pam’s statements are remarkably accurate, in light of irritatingly common errors like calling He died a sentence in the “passive tense”, or saying that science is a verb. (Props to Oscar, too, for distinguishing dislike of a word from nonexistence of a word.)

Despite its accuracy, Pam’s rule doesn’t really address the problem of whomever. It covers whom just fine, but the trouble is that when you’re dealing with who(m)ever, you’re usually dealing with more than one clause. Let’s look at a simple example first:

Whoever did this is in big trouble.

At the top level, we have the clause [SUBJ] is in big trouble. So the whoever is part of the subject. But the entire subject consists of another clause, whoever did this, which has the unusual property of being able to function as a noun phrase. This kind of clause is often referred to as a fused relative, to remind us that (in our case) whoever did this is like a noun and relative clause, “fused” together into one structure: the person who did this. In this fused relative, then, whoever acts as the subject of did. So no matter whether you’re looking at the top-level clause (aka matrix clause) or the embedded clause, whoever is (part of) the subject. Therefore, you use whoever instead of whomever; case closed.

Another simple example:

I’ll make friends with whomever Tom hates.

The matrix clause is I’ll make friends with [OBJ], so the whomever is part of the object of the preposition with. Looking at the fused relative that makes up this object, we find that whomever is functioning as the direct object of hates. Object of preposition in the matrix clause; direct object in the embedded clause. Either way, the objective case form whomever is called for. Good thing English doesn’t have different case forms for direct objects and objects of prepositions, or we might have a problem.

But with these next examples, we do run into just that kind of problem:

I’ll kill whoever did this.
Whomever Tom invites will probably be boring.

In the first sentence, whoever did this is the object of kill, but within this object, whoever is the subject of did. In the second sentence, whomever Tom invites acts as the subject of will be, but within this subject whomever is the object of invites. So which clause wins, matrix or the embedded?

As might have been guessed by my choice of whoever or whomever, the rule in Standard English is that the embedded clause wins. Thus, it doesn’t matter that whoever did this is the direct object of the matrix clause; all that matters is that whoever is the subject of did. It doesn’t matter that whomever Tom invites is the subject of the matrix clause; all that matters is that whomever is the direct object of invites. In some languages, though, sentences like these would be a problem. In German, for instance, you simply can’t use fused relatives unless the who(m)ever equivalent serves the same function in both matrix and embedded clauses, like in the earlier two examples.

Unfortunately, this still doesn’t give a definite answer on whether Ryan should have said whomever or whoever: In the relevant sentence from Ryan, there is only one clause: you can communicate it to [OBJ]. But wait, that makes things simple, right? It’s the object of to, so what we want is whomever, right? Well, I don’t know. It could depend on how Ryan would have completed the sentence if he hadn’t stopped at whomever. He might have said:

…communicate it to whomever you need to communicate it to.

In that case, fine: In the embedded clause, whomever is (part of) the object of to. (In the matrix clause, too, but as I said earlier, that doesn’t matter.) But what if Ryan had the following sentence in his mind?:

…communicate it to whoever needs to know it.

Now in the embedded clause, whoever is the subject of needs to know, so whomever wouldn’t be right.

On the other hand, maybe it’s ridiculous to imagine how Ryan would have completed the sentence. Though I may be succumbing to the Recency Illusion, it’s my impression that standalone whoever (and whatever, wherever, whenever) developed years after the purely syntactic rule for who(m)ever had faded from common knowledge, so why would I expect it to have arisen with any allegiance to this rule? Given that it emerged among a generation of speakers who mostly forgot, remembered imperfectly, or never knew the syntactic rule regarding whomever, I’d expect standalone who(m)ever to follow the more prominent, emerging sociolinguistic rule — use whomever in formal settings. This seems to be how Ryan is using it, as a boss addressing a subordinate and wanting to project as much authority as possible.

UPDATE: I’ve learned that Ed Cormany of Descriptively Adequate has commented on this dialogue. I learned this from Ben Zimmer’s post on Language Log, which conveniently congregates links to all the LL posts dealing with whom. Ed’s and Ben’s posts made me realize I’m behind the times: Why didn’t I think to just include the YouTube video, like they did? I’d do it now, but heck, just go see it on Ed’s or Ben’s post.

20 Responses to “Whomever Is Never Actually Right”

  1. Frogman said

    “As might have been guessed by my choice of whoever or whomever, the rule in Standard English is that the matrix clause wins.”

    Er… The rest of your text makes sense to me only if the embedded clause wins.

    I’m not a native speaker, and I’m very happy to learn about these subtleties. As far as I’m concerned, I stick to “who” and “whoever”, and I will continue to do so. It seems a lot safer.

  2. Neal said

    Thanks, Frogman. Yes, I did intend to write “the embedded clause wins”, and realized the error during my shower th is morning. That’s what I get for blogging at 1:00 in the morning. The post has been corrected.

  3. Stalina said

    Could you give exmples of the German cases you mentioned? Danke.

  4. Neal said


    The following are from:
    Groos, Anneke, and Henk van Riemsdijk. 1981. Matching effects in free relatives: A parameter of core grammar. Theory of markedness in generative grammar, ed. by A. Belleti, L. Brandi, and L. Rizzi, 177-216. Pisa: Scuola Normale Superiore.

    I found them in:
    Bayer, Samuel. 1996. The coordination of unlike categories. Language 72.579-616.

    In these sentences, where there is a mismatch between case demands of the matrix and embedded clauses, the sentence is ungrammatical no matter which case one uses for ‘who’.
    1. Wer[NOM] nicht stark is, muss klug sein.
    ‘who(ever) isn’t strong must be clever.’ (matrix nom, embedded nom)
    2. *Wen[ACC] Gott schwach geschaffen hat, muss klug sein.
    ‘whomever God created weak must be strong.’ (matrix acc, embedded nom)
    3. *Wer[NOM] Gott schwach geschaffen hat, muss klug sein.
    same intended meaning as 2, but still ungrammatical.
    4. Ich nehme, wen[ACC] du mir empfiehlst.
    ‘I take whomever you recommend.’ (matrix acc, embedded acc)
    5. *Ich nehme, wen[ACC] du vertraust. (matrix acc, embedded dat)
    ‘I take whomever you trust.’
    6. *Ich nehme, wem[DAT] du vertraust.
    same intended meaning as 5, but still ungrammatical.

    However, if the conflicting cases happen to have the same form, then the sentence can still work. For example, the neuter form is the same in the nominative and the accusative:

    7. Ich habe gegessen was[NOM/ACC] noch uebrig war.
    ‘I ate what was left.’ (matrix acc, embedded nom)

    Do Groos and van Riemsdijk’s grammaticality judgments agree with yours?

  5. Matthew said

    Thanks for the excellent post.

  6. Richard said

    I think you confirmed Creed’s observation!

  7. Their grammaticality judgments agree with those of this native speaker of German.

    Though I’ll admit that I’d have used ? (questionable) instead of * (wrong) for ?”Ich nehme, wem[DAT] du vertraust.” (and similarly for ?”Ich nehme, wer[NOM] kommt.”).

  8. Jette said

    This native German speaker’s grammaticality judgements agree with Philip’s.

    I’d even go a bit further and say that
    “Ich nehme, wem[DAT] Du vertaust.” sounds really pretty acceptable to me and for some reason
    ?”Ich nehme, wer[NOM] kommt.” gets better and better the further “kommt” moves away from”wer”, as for example in “Ich nehme, wer[NOM] auch immer kommt.”

  9. ruidh said

    Perhaps you were best not to include the video — it’s been taken down by You-Tube. I was at least able to get the point of the example because you had reproduced the text. Beware relying on other people to maintain the material on which you are commenting.

  10. Ben Zimmer said

    The YouTube video still works fine for me.

  11. […] to address this whole whoever/whomever debate, Literal Minded, picks it out and explains the usage for us, quite technically. As might have been guessed by my […]

  12. Ben Zimmer said

    Hmm, now it appears that the linked video has been pulled from YouTube, but this one is still online (for now).

  13. […] korrekte Verwendung von who/whom achtet. Den fragwürdigen Grammatikalitätsurteilen der Kommentatoren kann ich mich allerdings nicht […]

  14. […] Whomever Is Never Actually Right « Literal-Minded […]

  15. […] by Neal on April 12, 2008 If you found this post on whoever vs. whomever interesting, you can find further discussion of whomever in this post from […]

  16. Spoon said

    OK. I do not remember where I learned it, but what I learned was to use whom when anyone could eventually become the who. In other words, if you have possible, specific suspects that may be the who one would use who. If, on the other hand, anyone could eventually be the who the speaker would use whom.

    Is that at all correct?

  17. […] Zwicky says many speakers use both systems, depending on “formality, emphasis, prosody, and the like.” He also addresses this at the Visual Thesaurus. Incidentally, whomever doesn’t occur at all in some idiolects, but, for the record, it’s not “a made-up word used to trick students”, as Creed says in The Office; this is analysed at Literal-Minded. […]

  18. […] Literal-Minded: Literal-Minded has linguistic commentary from a guy who takes things too literally. (Recommended Post: Whomever is Never Actually Right) […]

  19. […] is a subject or an object. It’s the same problem that confuses English speakers about whoever and whomever. So actually, what we have here is a translation that is faithful even in preserving the ambiguity […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: