They interviewed several candidates who he thought had the experience and qualifications he required.

My test prep book says this should be "who" because of the subordinate clause's predicate:

They interviewed several candidates who he thought had the experience and qualifications he required.

I feel like it should be "whom" as it's the object of the main clause.

The interviewed several candidates whom he thought had the experience and qualifications he required.

Who's right?

  • 1
    Since who is the subject of the verb had, it cannot be whom. You could, however, change the sentence to read whom he thought to have the experience, which uses a different, less usual structure.
    – Anonym
    Aug 12, 2014 at 21:42
  • @Anonym but why doesn't the pronoun agree with the larger clause's use of it as an object? Aug 12, 2014 at 21:53
  • 2
    He thought is a sort of parenthetical remark to the rest of the clause; it doesn't really 'fit in' grammatically. Some punctuation should provide some clarity: They interviewed several candidates who, he thought, had the experience and qualifications he required.
    – Anonym
    Aug 12, 2014 at 22:01
  • 2
    He thought is not necessarily parenthetical. Whom is sometimes found in this type of construction; there is no single, finite answer one way or the other. Note that who(m) is not really the object of any clause—it's more an extrapolated subject in a subordinate clause that is the object of he thought. There is more detailed discussion in this question, which this I've flagged this one as a duplicate of. Aug 12, 2014 at 22:16

1 Answer 1


It's complicated. The reason whom is wrong is that it's not an object; it's a subject.
It just looks like it oughta be the object, but only if you don't know the syntax.
That's what makes it a great test distractor. Here's the example sentence, stripped of irrelevant details.

  1. They interviewed several candidates
  2. who he thought
  3. had the right experience and qualifications

3 main verbs, therefore 3 clauses:

  • (1) is the main clause, and has the statement in it.
  • (2) is a relative clause, modifying candidates.
  • (3) is a that-less tensed complement clause, the direct object of thought.

Note that (3) is subjectless, though it's clear that its subject refers to whoever who refers to.

So the question is, why is who correct? Because it really is correct here, and whom is incorrect.
(one more example, by the way, of why I tell people never to use whom at all; who is alway safe)

Let's take a look at some other sentences. Let's start be deleting "he thought", producing

  1. They interviewed several candidates
  2. who had the right experience and qualifications.

Note that who is in fact the subject of had in this relative clause; whom would be wrong.
And another sentence:

  1. I think
  2. that they had the right experience and qualifications.
    2'. they had the right experience and qualifications.

Clause (2) is tensed (it's in past tense; it's not an infinitive or gerund), it's a complement clause (the clause is the direct object of think), and it starts with the complementizer that, which is optional here, as (2') shows.

So what? Well, one more fact about tensed complement clauses:

  • a tensed complement clause with that is an island, but without that it's not an island.

And two more facts about relative clauses:

  • relative clause formation can extract relative pronouns from far away in the clause
    (several candidatesᵢ whoᵢ had the qualifications)
    (several candidatesᵢ whoᵢ he thought had the qualifications)
    (several candidatesᵢ whoᵢ she said he thought had the qualifications)
  • but they can't extract them from islands; add that and it becomes ungrammatical.
    (*several candidatesᵢ whoᵢ he thought that had the qualifications)

What that adds up to is that the relative pronoun heading a relative clause may come from indefinitely far away, and does not necessarily have anything to do with the clause that happens to be at the beginning.

In this case, the I think part is essentially an adverb and adds no information, but instead qualifies the speaker's assertion. It plays no part in the rest of the relative clause, and its object is a whole clause, not a pronoun. Who was extracted from the subject position of that object clause and moved to the front by relative clause formation.

  • I don't disagree with anything you've said - but, isn't one of the tests we can do here to extract the relative pronoun and see if the sentence is still grammatical. If it is, so the story goes, it must be a potentially accusatively marked pronoun. The pronoun here seems to be removable. Any thoughts? (I have some, but not good enough to share with anyone). I note, in fact, it's droppable in the last two of your trio of examples. Aug 18, 2014 at 23:36
  • Interesting; I've never heard the story that goes that way. Who told it to you? The rule for dropping relative pronouns drops any relative pronoun that is not the subject of the immediately following (i.e, top) relative clause; it says nothing about them being objects, of any sort. In this case, it is indeed not the subject of that relative clause -- instead, it's the subject of a clause farther down. That's what's interesting about ungoverned rules like Relative Clause Formation -- they aren't limited to only one sentence, so many bets are off. Like I said at first, it's complicated. Aug 19, 2014 at 0:39
  • Hmm, it's my own theory, but I would say that that's the same rule for whether you can use whom or not. However, that's just my opinion and very unorthodox it is too - but seems backed up by usage, perhaps (ok, that would need substantiating a bit more). I don't think that there's any evidence from usage that it's not the case. Some inflections and strong (versus potentially weak) forms seem to be an indication not to expect a particular following item (personal pronouns such as mine ours etc, strong forms of auxiliary verbs, pronouns etc.) We have alternations of form for many things Aug 22, 2014 at 0:14
  • ... many things which aren't specifically 'case' that is. This alternation (and I continue with my unsubstantiated theory) could merely be about indicating that it's not the subject of the following verb (ie expect another noun in-between). OK, I'm aware that this is a batty theory, but I can't shake it off ... Aug 22, 2014 at 0:17
  • If you can find a theory that explains all the examples, it's superior to a theory that doesn't. Occam's Razor, basic methodology. Aug 22, 2014 at 0:21

Not the answer you're looking for? Browse other questions tagged or ask your own question.