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.
First, to get it out of the way, it is not the case that 'Indian English is rhotic'. There is widespread variation - some speakers are and some are not (more details and references later).
But in explanation, there are multiple things going on here. There's the 'why' question, there's the 'r' question, and then there's India.
There are three kinds of answers to 'why' questions in linguistic phenomena.
it just happened (random drift). Why did the upper class accent in Southwest England in the early 19thc turn non-rhotic? Dropping r's ain't hard, they're pretty breathy consonants already. It may have just gone away all on its own. Sociologically, any kind of change can be an in-group marker, continued by the group to exclude outsiders. Of course the changes aren't totally arbitrary: a phoneme will drift into another more easily if it is nearby, t > d > th > 0, but not very often t > ng
Exposure to a different community. Why did the upper class accent in Boston become non-rhotic in the late 19th c? Probably to mimic the British accent (a sociologically desirable accent).
There's a documented case of one person saying something one weird way once and then everybody starting to follow. This is rare. Everybody started saying the solecistic 'My bad' (for 'My mistake') after the movie 'Clueless' popularized it (it existed before then but the widespread use came from the movie). This is rare to confirm, but may well be a non-trivial source of changes in languages.
But these why answers are somewhat conjectural and because of lack of ancient source material, phonological ones are hard to pin down unless there's datable poetry with rhyming for scholars to match up.
Asking about 'r' is extra interesting because there are so many varieties of it: a tap, a retroflex, a lateral, a trill, a velar fricative, a coloring on a vowel (all of which can drift in different directions according to tongue position). And becoming non-rhotic has many flavors, sometimes having intrusive or linking r's, sometimes one, and sometimes neither.
Asking about India and Indian English is complicated because India is kinda a big country, with many highly populated local standard languages (Hindi, Bengali, Telegu, Tamil, etc, etc, etc) each with different phonological systems, though it does have some Sprachbund characteristics. See The effects of native language on Indian English sounds and timing patterns
for details.
And then there's the English colonization, which in no way replaced any languages in India but, to oversimplify, the Indian variety of English became a lingua franca, a language of higher education/business, but such crossover languages aren't necessarily homogeneous, with prestige and colloquial versions. And then also it is who one is listening to, people in India, immigrants to the UK or to the US or media from different places.
Oh, and also English itself is complicated enough. It has rhotic and non-rhotic varieties. American, Scots, Irish, Wessex are rhotic, and English colonists may have taught English to Indians rhotically 150 years ago, or maybe the overwhelming number of American Youtube videos are teaching them now.
So to combine these all together, it's all one big caveat. There's no single Indian English, r's are different among the local standard languages plus non-rhoticity, there may have been some imitation of the RP r-dropping as well as a tendency to be rhotic or non-rhotic from the local standard. In the New Delhi area, one study (Postvocalic (r) in urban Indian English ) shows that there's a good mix of rhoticity versions. The study breaks it all down by many factors, in each group there's some people that are rhotic, sone not, tendencies in one direction or the other, but no 100% percent either way.
If there is a particular IE accent, say of Punjabis in SW England or bank clerks in Hyderabad, then one might be more specific. In the first example, Punjabi is rhotic (with a postvocalic r being a tap). So if they are r-dropping in English then it is probably an imitation of RP (or Estuary which is also non-rhotic).
At this point, you're probably wondering why I don't just answer what you know I think you're asking about, the stereotypical Indian speaking English, like Maz Jobrani or Apu on the Simpsons. Actors are pretty good at imitating
others, but they're not perfect, they're just making up stuff. They're just not particularly trustworthy.
On the English Dialects Archive there are 17 instances of Indian English, and my (rhotic-ear biased) survey showed only 2 of those were rhotic. But it could be that those were not reflective of the generic IE accent that you're thinking of. And as far as it goes, in Indian media (movies or news produced in India with Indian actors, speaking non-dubbed English) I personally hear strong non-rhoticity ('here' - /hjaw/, 'nurse' - /nʊs/,/nɜːs/)
Best Answer
This is a very interesting example. I'll give a short answer here, and leave a full explanation to someone more familiar with the syntax of this particular verb.
Verbs often take preposition phrases as a Complement. In the active voice version of the Original Poster's sentence, we see the verb add take a Direct Object and a prepositional phrase functioning as Locative Complement:
In the sentence above we see water as a Direct Object, and the phrase to the mixture functioning as the Locative Complement.
Usually when we see verb plus preposition phrases like this, the preposition phrase cannot become the Subject of a passive construction. So for example if we passivise Someone slept in the bed, we get:
Here the Complement of the Preposition in has become the Subject of the passivised sentence. In contrast, the preposition phrase in the bed itself cannot become the Subject:
This is the pattern that we generally see with such sentences. We cannot move the prepositional phrase into the Subject position. We can only move the noun phrase from within the preposition phrase. So the first examples in the following pairs of sentences are grammatical, and the second ones are not:
The active version of the Original Poster's example is:
We can of course make water the Subject of a passive version of the sentence:
But we should expect to be able to make the noun phrase the mixture the Subject of a passive version of the sentence as well:
But this gives us a poor result. Very strangely, we do seem to be able to move the whole preposition phrase to the Subject position here. Many grammarians say that this is not possible. However, there is no doubt that the only viable solution to the Original Poster's malformed sentence, without moving mixture out of the Complement phrase, is to move the entire prepositional phrase into Subject position:
Now, exactly why this should be possible, I have no idea. However, very strangely indeed, it actually is. Maybe someone else can explain why ...
Edit note
Rathony has suggested that maybe there is some phrase movement here. He is correct, I believe. I think my version of the sentence is actually one where the prepositional phrase has been fronted and the Subject post-posed. In other words my version of the sentence is a convoluted version of:
Here the Subject of the sentence is actually, of course, water. So, this made me think that perhaps this post should be deleted. However, there still remains the puzzling question, as detailed above, of why the following isn't grammatical:
So I've turned it into a wiki instead.