ᴛʟᴅʀ: Yes, this construction is perfectly grammatical* in English, and perfectly common as well. There are subtle restrictions on it, however, so not all such transforms produce things that sound right, or at least good, to a native speaker.
ɴᴏᴛᴇ 1: It requires preposition stranding, which some decry — but they’re wrong. :)
ɴᴏᴛᴇ 2: Like Pullum, I won’t be using the confusing term “voice” here, though, because this is not about Latin.:)
This construction is called the prepositional passive construction. It is one of several constructions sometimes dubiously referred to as a “pseudo-passive” one.
Passives occur when a sentence (well, a clause) with an object is rearranged so that what had previously been an object is now that sentence’s grammatical subject governing its verb, and the old subject, if included, ends up being the object of a prepositional phrase starting with by. Another way to say this is that passives are clauses whose subjects and objects get swapped around, even though the agent and patient remain the same.
This reversal of subject and object is most commonly seen in clauses with direct objects, but it can also happen in clauses with indirect objects or even prepositional objects as well. Watch how in these pairs, the subject and object shown in the first of each pair exchange positions in the second, making it a passive:
Passive with Direct object
- The dog bit the boy.
- The boy was bitten by the dog.
Passive with Indirect object
- The school sent the winners prizes.
- The winners were sent prizes by the school.
Passive with Prepositional object
- A thousand passing schoolboys have peed off of this bridge.
- This bridge has been peed off of by a thousand passing schoolboys.
Just because the verb to pee is intransitive does not forbid it from being used in passives. It just can’t be a passive formed by rearranging a direct object into a subject, since there can be no direct object with an intransitive verb.
But intransitive verbs can certainly take prepositions, and these prepositional objects can be used to create passives, and indeed often are.
You’ll see in the examples I’ve provided that the passive formed by the direct object has nothing between the be + participle and the word by that precedes the original subject. But with passives formed by the indirect object or the prepositional object, there are extra bits after the particple. I’ve marked this by setting in italic the part after the participle up through the word by in my examples above.
Passives, per Pullum
In his most excellent essay on “The Passive in English”, Geoffrey Pullum writes that:
In English the prepositional passive is quite frequent, especially in relatively informal style. Most languages don't have anything like it (Norwegian is a rare example of a language that does).
Why does Pullum mention informal styles? Because prepositional passives require something we call preposition stranding, which is when the object of a preposition has been moved to some other position in a sentence than immediately following that preposition. Since with prepositional passives we're promoting that object to the sentence subject, this by necessity strands that preposition.
Of course, preposition stranding is perfectly normal in English, and always has been. But this is one of those constructions “up with which” certain pseudo-Latinate Victorian registers would so notoriously not put. That’s why it is sometimes considered less formal: because it’s something that informal registers are more likely to put up with outside of Victoria’s earshot.
Pullum goes on to observe that English prepositional passives do have restrictions about which ones are valid and which aren’t:
There are some peculiar restrictions on prepositional passives in English. One is that there can be a difference in acceptability according to whether the subject denotes an entity that is tangibly altered in state: This bottom bunk has been slept in is dramatically more acceptable than ??The bottom bunk has been slept above, apparently because sleeping in a bunk bed alters its state (the sheets are wrinkled and so on), while sleeping in the top bunk above it doesn't alter its state at all. Intuitively, you use a prepositional passive when the VP expresses a relevantly important property of the subject. That's a restriction on prepositional passives, because there is nothing peculiar about the active version Someone has slept above this bottom bunk. (Why would a language have a restriction like that? Who knows. I don't make or try to enforce any of the rules; I am merely trying to explain what the rules seem to be.)
A native speaker “knows” which prepositional passives are allowed and which ones are not, but non-natives attempting to learn the language by inferred rules should probably take some care with them.
This is the footnote for the asterisk in the opening summary:
- Remember that “grammatical” just means that something sounds “right” to a native speaker, and that “not grammatical” just means that something sounds “not right” to a native speaker.