Zompist Bboard Again

...again. >__< Sorry to bring this up once more. I'm super confused about this, so I tried to write down all possible combinations of transitive and intransitive sentences, and how relativization interacts with that. Like, we have a main clause and a relative clause, and one of the arguments in the main clause is at the same time an argument in the relative clause. Hopefully you'll understand what I mean when you see the following.

Intransitive main clause's subject is relativized and appears as the subject in the intransitive relative clause
The man who runs stumbles < the man runs | the man stumbles

Intransitive main clause's subject is relativized and appears as the subject in the transitive relative clause
The bird that wakes up first catches the worm < the bird wakes up | the bird cathes the worm

Intransitive main clause's subject is relativized and appears as the object in the transitive relative clause
Is this even possible?

Transitive main clause's subject is relativized and appears as the subject in the intransitive relative clause
The girl who sprained her ankle cried < the girl sprained her ankle | the girl cried

Transitive main clause's object is relativized and appears as the subject in the intransitive relative clause
The apple that John brought ripens < John brought the apple | the apple ripens

Transitive main clause's subject is relativized and appears as the subject in the transitive relative clause
The boy who cried wolf told a lie < the boy cried wolf | the boy told a lie

Transitive main clause's subject is relativized and appears as the object in the transitive relative clause
Is this possible?

Transitive main clause's object is relativized and appears as the subject in the transitive relative clause
A dog chased the cat that caught a mouse < a dog chased the cat | the cat caught a mouse

Transitive main clause's object is relativized and appears as the object in the transitive relative clause
I'll eat what you cook < I'll eat what | you cook what

I'm not sure if all the clauses that are supposed to be main clauses in these examples actually are main clauses. Can someone more knowledgeable please check that for me? There are two cases here where I couldn't come up with an example. Are these possible in English? If not, are they possible in other languages? And if I've misanalyzed anything here, please let me know.

You've got 'main' and 'relative' the wrong way around in some cases.

So, for instance, if we take "The apple that John brought ripens"...

The main clause is "the apple [...] ripens". This is what can stand alone as a clause, and it's also the 'new' information.

The relative clause is "that John bought". This cannot stand alone as a clause, and it serves to provide further information to help identify which apple is being discussed, rather than providing new information.

Let's go through these possibilities... for simplicity, we'll call the subject of an intransitive 'S', the subject of a transitive 'A' and the object of a transitive 'O'. This is obviously dangerous and misleading in general, but with this stated clearly and our discussion only being on this issue (and not roles vs cases, alignment, etc), it shouldn't be a problem for now, and saves a lot of time. I'll underline main clauses, and italicise relatives.

[actually, not true. The whole thing is the main clause, in a sense, because the relative becomes embedded in the matrix clause as part of the noun phrase. A relative is just a clause that's been subordinated in such a way as to embed it into a noun phrase. So I'm actually underlining what the main clause WOULD be if the relative weren't there]

So:
S(matrix)-S(relative)
The apple that fell was ripe

S-A
The bird that ate the poisoned wolverine died

S-O
The apple that the wolverine ate had been poisoned

A-S
The couple who eloped poisoned the apple

A-A
The wolverine that ate the apple also tore out Bob's entrails

A-O
The wolverine that Mike saw ate all of the childrens' lunch

O-S
The couple saw the apple that fell

O-A
The custard poisoned the evil couple who hated all the children

O-O
The dead bird had fatally polluted the custard that the couple had stolen from the starving children

I'm not quite sure what the question is. I discuss this quite a bit in my syntax book; you can also look at various strategies in WALS:

https://wals.info/chapter/122

https://wals.info/chapter/123

For a conlang, you probably want to worry about a few things:

* what exactly do you do to mark relativization?
* did your method leave subclauses ambiguous?
* is the NP in the subclause corresponding to the head noun simply omitted, or is a pronoun left behind?
* what case is assigned to the head noun? Almost always this is what's appropriate for the main clause, but there's a phenomenon called case attraction where the subordinate clauses's case is used instead.

More complicated things can be relativized: indirect objects, locatives, NPs in sub-sub-clauses. What can be relativized varies by language.

zompist wrote: ↑Tue Feb 26, 2019 6:55 pm I'm not quite sure what the question is. I discuss this quite a bit in my syntax book; you can also look at various strategies in WALS:

https://wals.info/chapter/122

https://wals.info/chapter/123

For a conlang, you probably want to worry about a few things:

* what exactly do you do to mark relativization?
* did your method leave subclauses ambiguous?
* is the NP in the subclause corresponding to the head noun simply omitted, or is a pronoun left behind?
* what case is assigned to the head noun? Almost always this is what's appropriate for the main clause, but there's a phenomenon called case attraction where the subordinate clauses's case is used instead.

More complicated things can be relativized: indirect objects, locatives, NPs in sub-sub-clauses. What can be relativized varies by language.

Actually, everything must be able to be relativized, at least indirectly. Those who can't are actually able to do it by raising. For example:

In Indonesian, we cannot relativize on direct object directly. But it doesn't prevent us to translate an English sentence like these:

Andrew eats a meat that Sarah bought.

We just use passive construction:

Andrew memakan daging yang dibeli Sarah. (literally Andrew eats a meat that was bought by Sarah)

zompist wrote: ↑Tue Feb 26, 2019 6:55 pm I'm not quite sure what the question is. I discuss this quite a bit in my syntax book; you can also look at various strategies in WALS:

https://wals.info/chapter/122

https://wals.info/chapter/123

For a conlang, you probably want to worry about a few things:

* what exactly do you do to mark relativization?
* did your method leave subclauses ambiguous?
* is the NP in the subclause corresponding to the head noun simply omitted, or is a pronoun left behind?
* what case is assigned to the head noun? Almost always this is what's appropriate for the main clause, but there's a phenomenon called case attraction where the subordinate clauses's case is used instead.

More complicated things can be relativized: indirect objects, locatives, NPs in sub-sub-clauses. What can be relativized varies by language.

The Wikipedia article on relative clauses is also a very good resource for this kind of thing - it breaks down each part of relative clause formation in great detail, giving lots of examples. I used it a lot before I found the WALs entry.

I need to do some reading before I can reply better. I'll just adress a couple of things though...

Salmoneus wrote: ↑Tue Feb 26, 2019 3:51 pmS-O
The apple that the wolverine ate had been poisoned

Thanks for your new examples! But can you remake this one without using passive voice please? I was intentionally avoiding passive voice when I tried to make examples, because that complicates things (see my next post).

zompist wrote: ↑Tue Feb 26, 2019 6:55 pmwhat exactly do you do to mark relativization?

My conlang has a brain-melting system that involves word order. In certain cases it's also necessary to use passive voice (so I need examples that are all active in English). I wrote this part of the grammar a long time ago, but I didn't have a good methodology, so that's why I now need all these combinations of S, A and P. I'm trying to see how many of these sentence types can be translated with the current rules.

Qwynegold wrote: ↑Wed Feb 27, 2019 9:37 am I need to do some reading before I can reply better. I'll just adress a couple of things though...

Salmoneus wrote: ↑Tue Feb 26, 2019 3:51 pmS-O
The apple that the wolverine ate had been poisoned

Thanks for your new examples! But can you remake this one without using passive voice please? I was intentionally avoiding passive voice when I tried to make examples, because that complicates things (see my next post).

The apple that the wolverine ate was blue.
The apple that the wolverine ate stank.
The apple that the wolverine ate imploded unexpectedly.

lol, thanks!

Salmoneus wrote: ↑Tue Feb 26, 2019 3:51 pm[actually, not true. The whole thing is the main clause, in a sense, because the relative becomes embedded in the matrix clause as part of the noun phrase. A relative is just a clause that's been subordinated in such a way as to embed it into a noun phrase. So I'm actually underlining what the main clause WOULD be if the relative weren't there]

Yeah, is there no word for this? I'll just call it main clause too when I actually main clause sans the relative clause.

Anyway, I've now learned that sentences of the type "I'll eat what you cook" are called free relative clauses; and they are a problem of their own.

I've worked out how those bound relative clauses should be translated now. I thought the rules I had written in my grammar were insufficient for translating all of these sentences, but it turns out they're enough. But it took me a looong time and several tries before I succeeded. So that section in my grammar needs to be completely rewritten to make it clearer. The rules are basically that the relative clause comes first, then the NP with the head noun, and then the main clause. The main clause must have SVO order.

S-S
The apple that fell was ripe
PST fall, the apple-NOM PST be ripe

S-A
The bird that ate the poisoned wolverine died
the poisoned* wolverine-ACC PST eat, the bird-NOM PST die
*Using poisoned as an adjective for now, because otherwise I might have to use a relative clause inside a relative clause, and I don't even dear to think of such things.

S-O
The apple that the wolverine ate stank.
by wolverine PASS-eat, the apple-NOM PST stink

A-S
The couple who eloped poisoned the apple
PST elope, the couple-NOM PST poison the apple-ACC

A-A
The wolverine that ate the apple also tore out Bob's entrails
apple-ACC PST eat, the wolverine-NOM PST tear PREP entrails of Bob

A-O
The wolverine that Mike saw ate all of the childrens' lunch
by Mike PST PASS-see, the wolverine-NOM PST eat lunch-ACC of all children

O-S
The couple saw the apple that fell
PST fall, the apple-NOM PST PASS-see by the couple

O-A
The custard poisoned the evil couple who hated all the children
the all children-ACC PST hate, the evil couple-NOM PST PASS-poison by the custard

O-O
The dead bird had fatally polluted the custard that the couple had stolen from the starving children
by the couple from the starving children PST PASS-steal, the custard-NOM PST PRF.3 PASS-pollute fatally by the dead bird

So you use the passive voice when the head noun is a subject in one clause and an object in the other. I'm not sure yet how to formulate a rule that defines which of the two clauses, main or relative, or both, that the passive will appear in.

I also need to figure out:

1. What happens if you want to relativize an indirect object or some other thing.
2. How the adversative case fits into this. This case marks someone who is negatively affected by the action, and can be used on the patient (which is usually the object, but not in passive sentences).
3. Some good way to handle free relative clauses. The above rules would produce "by 2SG PASS-cook, food-NOM PASS-eat by 1SG", which feels way too cumbersome for such a simple sentence.

Oh btw, a question that arose during all of this. Nominatives are zero-marked in this language, and nouns preceded by a preposition are usually not (never?) case-marked. Would it be more correct to say that these nouns have nominative (or direct) case rather than that they are unmarked for case?

So only subjects can be relativised? That's a bit extreme...


Yes, bound vs free relatives are an issue. So are 'definite' or 'restrictive' vs 'indefinite' or 'descriptive' clauses. Eg in English "I like what I see!" vs "I like whatever I see!", and "I like soldiers who didn't get captured" vs "I like (some/these/the) soldiers who didn't get captured". [English generally doesn't distinguish the two, but does often have disambiguating devices].

There's often a distinction between direct and indirect relatives - where the former modify subject or object, and the latter modify obliques. It may be that you're describing a direct relative structure here, and you have a different structure for obliques?

And you might want to consider sentential relatives: "My name's Bob - which you already knew", "I discovered that Mongolia was actually very large, which came as an unpleasant shock". These can be regarded as relatives that modify an entire clause (although I think some languages deal with similar meanings without relative structures).

Looking at the WALS info, it seems there are several languages that drop the relativized element when it is a core argument, but require a pronoun for arguments that play a more peripheral/adverbial role in the relative clause. WALS lists Ewe, Irish, and Arabic, though I'm not familiar enough with those languages to confirm this characterization. Additionally, some languages, especially Austronesian ones, do not allow relativization on such non-core arguments at all, requiring a bit of creative periphrasis instead. The take away for conlangers is that it's pretty common for languages to have no really elegant way to deal with oblique relatives. This appears to be happening colloquially in some dialects of English ("That's the house that we were partying inside of it earlier.") and many languages might have the potential to shift to such a system very quickly and easily, depending on their syntax. A really simple way to add a bit of complexity would be to take your 100% regular relativising strategy and just rip out all the wires when it comes to obliques.

Moose-tache wrote: ↑Sun Mar 03, 2019 1:56 am Looking at the WALS info, it seems there are several languages that drop the relativized element when it is a core argument, but require a pronoun for arguments that play a more peripheral/adverbial role in the relative clause.

Ancient Egyptian did this.

Moose-tache wrote: ↑Sun Mar 03, 2019 1:56 am Looking at the WALS info, it seems there are several languages that drop the relativized element when it is a core argument, but require a pronoun for arguments that play a more peripheral/adverbial role in the relative clause. WALS lists Ewe, Irish, and Arabic, though I'm not familiar enough with those languages to confirm this characterization.

I can confirm this for Modern Irish.

In Irish, relative clauses are constructed with the help of a relative particle whose basic form is a /ə/. It occurs immediately before the main verb and causes lenition:

an tigh a thóg an maighre mná "the house which the fine woman built"
an maighre mná a thóg an tigh "the fine woman who built the house"

In Munster, the form used with non-core arguments is go/gur /gə(r)/[*]:

an mairnéalach gur thógadar a thigh "the sailor whose house they built" (lit. "...that they built his house")

Where the object of a preposition is involved, the older method is to combine it with the relative particle at the start of the clause:

an tigh ina rabhadar ag obair "the house in which they were working"

Nowadays it's more common to see a resumptive prepositional pronoun:

an tigh go rabhadar ag obair ann lit. "the house that they were working in it"


[*] This looks identical to the generic subordinating conjunction but may actually be derived from a generalised form of a combination of the relative particle with a preposition such as ag or do, e.g. an mairnéalach aga bhfuil an tigh (= an mairnéalach go bhfuil an tigh aige) "the sailor whose house it is"/"the sailor who has the house".