First some terminology, so that we can agree what we are talking about. I will distinguish semantic/thematic roles from grammatical functions:
Semantic roles
- Agent=doer of the action
- Theme=directly affected/acted upon by the action
- Goal=end state/location of the Theme
Grammatical functions
- Subject=noun phrase which agrees with the verb/receives nominative case (I/he/they, etc.)
- Direct Object = noun phrase adjacent to the verb which gets a semantic role from that verb; receives accusative case (me/him/them, etc.)
- Indirect Object = noun phrase introduced by the preposition to.
A large part of the confusion in describing verbs like give in English comes from not distinguishing Goal from Indirect Object, and not distinguishing Direct Object from Theme.
Verbs like give in English have two alternate forms:
- John gave a book to me.
- John gave me a book.
Now lets apply the definitions above to these two sentences.
In both (1) and (2), John is the Agent (doer of the action), and also the Subject (agrees with the verb/would be he as a pronoun).
In (1) a book is the Theme (directly affected by the action) and also the Direct Object (adjacent to the verb and getting a semantic role from it. Me is the Goal (end state/location of the Theme) and also the Indirect Object (introduced by the preposition to).
In (2), however, things change. English has a rule commonly called "Dative Shift" which turns an Indirect Object (introduced by to) into a Direct Object (adjacent to the verb). This doesn't affect the semantic roles of the verb, but it does affect the grammatical functions. So in (2) although me is still the Goal (end state/location of the Theme) it is now the Direct Object (adjacent to the verb). A book is still the Theme (directly affected by the action) but is no longer the Direct Object. We can call it a second object if you like.
Now we can look at the Passive rule in English. Very roughly (because strictly speaking this is not quite correct), the passive in English makes the Direct Object of the active sentence the Subject of the passive sentence, and makes the subject of the active sentence an optional argument introduced by the preposition by. Since there are two active forms in (1) and (2) there should be two corresponding passive forms, and in fact, there are:
- A book was given to me (by John).
- I was given a book (by John).
In (3) the Direct Object of (1) became the subject, and in (4) the Direct Object of (2) became the subject. Since Passive doesn't care about the semantic roles, only the grammatical functions, we retain the same semantic roles as before: a book is the Theme in both sentences, and me/I is the Goal. Crucially, a book in (4) is not the direct object, but still the "second object" it was in the active form.
Now remember in (2) I said that a book is no longer the direct object of give because the Dative Shift rule makes the Indirect Object into a Direct Object. This predicts that the second object in (2) should not be able to undergo passive, because it it no longer a direct object, and this is in fact the case, as we can see in (5):
- *A book was given me.
This contrasts minimally with (3) which contains the preposition to. In that sentence, a book was the direct object in the active form, and therefore is able to be the target of the passive rule.
John Lawler's comment sums it up:
Technically, the teacher is wrong. The teacher is confusing question
formation, pied-piping, case marking, preposition stranding, and
formality levels. The sentence produced was the passive of the
sentence given. There are other rules beside passive involved,
however. All of the following are grammatical:
- The vase was broken by who?
- The vase was broken by whom?
- Who was the vase broken by?
- Whom was the vase broken by?
- By whom was the vase broken?
They are not always appropriate for the same circumstances, like any different sentence.
Here's a paper that describes some relevant phenomena: "On the Grammatical Status of PP-Pied-Piping in English: Results from Sentence-Rating Experiments," by Seth Cable and Jesse A. Harris.
In English, the "neutral" position for question words is generally at the start of the sentence. Many grammatical analyses treat this as the result of a process, "wh-movement," that moves these words from another position.
Cable and Harris give the following example:
(1) Simple Wh-Movement in English
a. She left John
b. Who1 did she leave t1 ?
It's also grammatical to leave a question word in place ("wh-in-situ"), as in "She left who?", but in standard English this kind of structure is not the default: it's generally less appropriate than wh-movement except in certain circumstances (e.g. echoing a declarative sentence to express surprise). This is a valid stylistic reason why someone might object to your daughter's sentence. The teacher was wrong to call it "ungrammatical."
Sometimes, other words can also "move" with the question word. The term used for this is "pied-piping," since the idea is that the question word metaphorically abducts other words a bit like the Pied Piper of Hamelin abducted children.
This phenomenon is quite interesting to linguists so there has been a lot of study of it. When the question word is the object of a prepositional phrase (PP), it's generally optional to move the preposition to the front of the sentence along with the question word. Cable and Harris's explanation:
(2) Pied-Piping in English
a. She left that guy.
b. [ Which guy ]1 did she leave t1 ?
Interestingly, when a wh-element is complement to a preposition, there
appears to be some optionality in whether the preposition is
‘pied-piped’ when the wh-element undergoes movement. That is, it would
appear that English freely permits both the structures in (3). In
sentence (3a), the wh-word has not pied-piped the higher PP. Such
structures are commonly referred to as ‘preposition stranding’ or
‘P-stranding’. In sentence (3b), the wh-word has pied-piped the PP, a
structure referred to as ‘PP-pied-piping’.
(3) The Optionality of Pied-Piping PP in English
a. Who1 did she leave [PP with t1 ]?
b. [PP With whom ]1 did she leave t1 ?
Although both the structures in (3) are commonly reported in the syntactic literature as ‘acceptable’, the grammatical status of PP-pied-piping structures is somewhat unclear, especially when compared to their preposition stranding counterparts. While speakers recognize structures like (3b) as English, such structures are not particularly colloquial. It is sometimes said that such structures are limited to particular registers, but often what is meant by ‘register’ in this context is unclear. After all, structures like (3b) are no longer a regular occurrence in formal written English either.
So the teacher's sentence "Whom was the vase broken by?" is not ungrammatical either.
However, it may sound funny because in general, the contexts where people use "whom" are also contexts where most people try to avoid preposition-stranding. (This point is also made by the answer to the following question: "Prepositions at the end of sentence and whom".) This would be a valid stylistic reason to object to the teacher's sentence.
As other people have mentioned, this consideration means that the following sentence might be the best, stylistically speaking:
- By whom was the vase broken?
However, a valid stylistic objection to this sentence is that to most people it sounds highly formal, old-fashioned, or pretentious. Most people would say, and many people would write
- Who was the vase broken by?
People could object that this sounds too informal. I think that's silly though, so I won't call that a valid stylistic objection. Using "who" in this position (not "whom") and stranding prepositions are both generally considered acceptable by educated people. But if the teacher is a pedant or "stickler," that might not matter.
Best Answer
First of all, there's nothing inherently bad or wrong with using passive voice--often times, it is the better way to construct a clause.
Now, with that aside.
Does your example use passive voice? . . . Superficially, it might appear so, to an automated grammar checker.
A traditional grammar's version of a passive (voice) sentence is a sentence that involves a construction whose lexical verb is a past-participle verb form and whose auxiliary verb is a form of the verb "BE" (or of the verb "GET", in some uses).
Your example has the expression "am finished". The word "am" is a form of the verb "BE"; and the word "finished" has the shape of a past-participle verb form. But that is the crux of the matter: Is the word "finished" in your example a past-participle verb or a past-participle adjective?
There are some grammatical diagnostic tools/tests that you can try out to come to your own decision on this. One diagnostic test that can often be useful is the below--Is the candidate version similar in meaning to an active voice version:
"I am finished with the website for now". -- [original version]
"Something finishes me, for now". -- [active version?]
Hmm, that doesn't seem to support the passive argument. Compare to: "Today, the neighbors finished the job"; "Today, the job was finished (by the neighbors)".
(Aside: There's also the prepositional passive: "George Washington slept in that bed"; "That bed was slept in by George Washington". But that doesn't seem to be relevant here.)
Another diagnostic test, which can sometimes be useful, to see if the candidate word is an adjective, is to try to see if the candidate word can be modified by "too" or "very":
Nah, that isn't convincing either. Compare to: "Tom is very tired".
There are other diagnostic tests that you can try (which might be in your favorite grammar books), but at the end of the day, you'll have to make the decision for yourself. (I have my own opinion.) Someone else may very well produce a convincing argument for you for one position or the other. Of course, your decision might also be influenced as to the grammar that you are comfortable with, and with how that grammar defines passive voice.