i was amazed by serial verbs when i first discovered them in chinese but i find them unusable in my conlang.
there are sentences that would end up ambiguous especially that i'd like to have a head initial language.
so a sentence like "i throw pen leave room" may be understood as "i threw the pen and left the room" or "i threw the pen out of the room" as i know it often happens in chinese that the latter of two verbs refers to the action or state of the object. a sentence like "i see man sad die" might be "i see a sad man die" or "i saw a sad man and died". hmmm.. also, a wikipedia example, "Amma take money give Kofi" looks like a monotransitive 'give', huh? i've always wondered if every idea can be expressed without using prepositions/case endings. the Ithkuil site lists all possible cases but can all of them be expressed by (serial) verbs alone? do you guys think chinese is primitive and handles a lot metaphorically and thus ambiguously? also, does chinese actually have main verbs and valency or can it just string together as many verbs as it wants?
the same thing happens in subordinate clauses: does "he see girl enter room" mean "he saw a girl that was entering the room" or "he saw a girl and entered the room"? my dream of a grammarless (only word-order-based) only-content-words language will always stay just that, a dream, i guess.. well, maybe i'm too much used to the precision of expression of english and don't realise languages like chinese can keep functioning without it..
on the other hand prepositions are apparently more complicated that they seem. i always dig into the underlying basic meaning of every lexical item i come across and prepositions could be and are rendered as verbs in chinese but they do more than any verb can do. i remember a sentence from a book about chinese verbs that went roughly as: "i cut apple become two" meaning "i cut the apple in two" and it seems clear the "in" means actually "make-(it-)become" than just "become" but maybe chinese doesn't need to be that precise..
the mechanics of serial verbs vs prepositions?
the mechanics of serial verbs vs prepositions?
Last edited by awer on Thu May 14, 2015 12:54 pm, edited 1 time in total.
Re: the mechanics of serial verbs vs prepositions?
1. "Ambiguity" is the bugbear of conlangers, but natural languages deal with it just fine. It wouldn't take much context to clear up who ultimately left the room, you or the pen.awer wrote:i was amazed by serial verbs when i first discovered them in chinese but i find them unusable in my conlang. there are sentences that would end up ambiguous especially that i'd like to have a head initial language.
so a sentence like "i throw pen leave room" may be understood as "i threw the pen and left the room" or "i threw the pen out of the room"
2. All it would take to disambiguate such a sentence is a simple conjunction. So "i throw pen leave room" for "I threw the pen out of the room" vs "i throw pen then leave room" for "I threw the pen and then left the room". (IIRC, this is exactly what Chinese does, but my Chinese is damn rusty at this point.)
Re: the mechanics of serial verbs vs prepositions?
Try to avoid the unfortunate wording about Chinese being "primitive" or "less precise".
If you have specific questions about Chinese, there's a thread for that where you're likely to get answers from fluent Mandarin speakers. Or check out a textbook on Chinese.
My general understanding, however, is that serial verbs don't really create ambiguity in Mandarin, for various reasons. For instance, coverbs (the preposition-like verbs) generally occur before the main verb ("I want with him talk"); they're also a small easily recognized class.
I think that where you can interpret a sentence as a pivot construction, that takes precedence. E.g.
Wǒ quàn tā niàn yī.
I advise 3s study medicine
I advised him/her to study medicine.
I guess you could force a reading "I gave him/her advice, then I studied medicine", but it would be a weird thing to say. What follows "I advised him/her" is normally going to be the advice, not what you did next. Probably you would need a different construction to get that reading (as I did in the English: "I gave him/her advice").
LB's point about conlangers and ambiguity is brief but deep. Part of the problem is we often don't have a widely accepted analysis for these pragmatic factors, so we fall back on appealing to "context". This sentence is ambiguous, but perhaps not at first glance:
I left him a pen.
Why is it we don't interpret it like "I left him a cripple"? Some combination of pragmatic and real-world knowledge. It's an ambiguity but it is not actually a problem.
If you have specific questions about Chinese, there's a thread for that where you're likely to get answers from fluent Mandarin speakers. Or check out a textbook on Chinese.
My general understanding, however, is that serial verbs don't really create ambiguity in Mandarin, for various reasons. For instance, coverbs (the preposition-like verbs) generally occur before the main verb ("I want with him talk"); they're also a small easily recognized class.
I think that where you can interpret a sentence as a pivot construction, that takes precedence. E.g.
Wǒ quàn tā niàn yī.
I advise 3s study medicine
I advised him/her to study medicine.
I guess you could force a reading "I gave him/her advice, then I studied medicine", but it would be a weird thing to say. What follows "I advised him/her" is normally going to be the advice, not what you did next. Probably you would need a different construction to get that reading (as I did in the English: "I gave him/her advice").
LB's point about conlangers and ambiguity is brief but deep. Part of the problem is we often don't have a widely accepted analysis for these pragmatic factors, so we fall back on appealing to "context". This sentence is ambiguous, but perhaps not at first glance:
I left him a pen.
Why is it we don't interpret it like "I left him a cripple"? Some combination of pragmatic and real-world knowledge. It's an ambiguity but it is not actually a problem.
- Nortaneous
- Sumerul
- Posts: 4544
- Joined: Mon Apr 13, 2009 1:52 am
- Location: the Imperial Corridor
Re: the mechanics of serial verbs vs prepositions?
The conlang I have that uses serial verbs for most everything has a 'subjectizer' particle -- the default is that the object of the first verb in the sequence is the subject of the next, but the subjectizer marks that the *subject* of the first verb is the subject of the next.
It seemed like the sensible way to handle a setup like this:
ñɛħ kɛʔ Caŋ ŋuɣ
cat bite Zhang die
A cat bit Zhang to death.
ñɛħ kɛʔ Caŋ lə ŋuɣ
cat bite Zhang SBJZ die
A cat bit Zhang and died.
But I don't know if it happens in any natlangs.
It seemed like the sensible way to handle a setup like this:
So:In Chinese, as in Southeast Asian languages, when a transitive verb is followed by an intransitive verb, the object of the combined verb may be understood as the object of the first verb and the subject of the second: lǎohǔ yǎo-sǐ le zhāng (lit. tiger bite-die PERF Zhang) "the tiger bit Zhang to death", where Zhang is understood as the direct object of yǎo ("bite") but as the subject of sǐ ("die"). In the equivalent construction in Hindi, the one who dies would be the tiger, not Zhang.
ñɛħ kɛʔ Caŋ ŋuɣ
cat bite Zhang die
A cat bit Zhang to death.
ñɛħ kɛʔ Caŋ lə ŋuɣ
cat bite Zhang SBJZ die
A cat bit Zhang and died.
But I don't know if it happens in any natlangs.
Siöö jandeng raiglin zåbei tandiüłåd;
nää džunnfin kukuch vklaivei sivei tåd.
Chei. Chei. Chei. Chei. Chei. Chei. Chei.
nää džunnfin kukuch vklaivei sivei tåd.
Chei. Chei. Chei. Chei. Chei. Chei. Chei.
Re: the mechanics of serial verbs vs prepositions?
For most verbs, serializing passes the object on as the subject of the next verb. You need a conjunction (and then, at the same time, as soon as, etc.) if the second verb shares the subject with the first, in which case the first verb and object looks like an adverbial clause of time.
For the rest, many of them follows their semantic meaning. Verbs meaning "help", "represent", "use" passes their subjects to the next verb, "accompany", "follow", "bring" means both together are the subject of the next verb. Coverbs also fall in this category, 把 (bǎ, the kind-of causative) passes both arguments to the main verb that they're both dropped there, 被 (bèi, passive) passes the subject as the object, and the object as the subject.
Some are just complicated. 找 "find", is the "accompany" type of verb regarding having a meal and recreation, but is the "most verbs" kind when the second verb is what the object do but the subject doesn't (e.g. Fix the car when the O is a car mechanic but the S isn't). But in practice, additional information is often inserted (e.g. 一起 "together", 幫X "help X, for X") to eliminate ambiguity.
For the rest, many of them follows their semantic meaning. Verbs meaning "help", "represent", "use" passes their subjects to the next verb, "accompany", "follow", "bring" means both together are the subject of the next verb. Coverbs also fall in this category, 把 (bǎ, the kind-of causative) passes both arguments to the main verb that they're both dropped there, 被 (bèi, passive) passes the subject as the object, and the object as the subject.
Some are just complicated. 找 "find", is the "accompany" type of verb regarding having a meal and recreation, but is the "most verbs" kind when the second verb is what the object do but the subject doesn't (e.g. Fix the car when the O is a car mechanic but the S isn't). But in practice, additional information is often inserted (e.g. 一起 "together", 幫X "help X, for X") to eliminate ambiguity.
Re: the mechanics of serial verbs vs prepositions?
There are infinite numbers of possible schemes for serial verb constructions (SVC). First of all it's important to understand not all SVC are born equal.
The most basic type is called nuclear SVC (nSVC) which involves two equally meaningful verbs concatenated. English has this for a few very restricted uses, especially in the imperative, such as: come look, go check it out. Nuclear SVC can either indicate simultaneity or cause-effect relationships. Typically any arguments in a nSVC will apply equally to each verb in the construction, but not necessarily. Some SVC languages will allow mixing together intransitives and transitives, for example, while others will not allow that.
Asymmetric SVC (aSVC) looks very similar to nuclear SVC, but they differ in that one of the verbs (the “minor verb”, mV) comes from a restricted set and may be semantically bleached to some extent, serving a grammatical function at least partly. Depending on the exact type of construction and the language's grammar, the minor verb may be placed either before or after the major verb.
Auxiliary SVC (auxSVC) is what we call it when a language with no deranked verb forms serializes an auxilary verb directly to an unmodified major verb. Instead of a indicative + infinitive type deal such as want to see, you simply get want see. The language may or may not treat these constructions the same as aSVC.
Exceptional case marking SVC (ECM SVC) or switch-function SVC (sfSVC) is when a argument is used as the object of one verb but the subject of another. This is analoguous to accusativus cum infinitivo in Indo-European languages such as English, where you can say: I want him to leave the house or she made me break the window, or it's close cousin the accusativus cum gerundivo: I saw him leaving the house. Unlike English and Latin, SVC language will often not require any kind of deranked verb form (like the infinitive or the gerund) for the secondary verb, using a fully finite verb form instead: she tricked me eat poison.
Not all SVC languages have ECM type SVC. Many, like Chinese, will interpret the subject as referring to both verbs: I saw girl enter room where I is the subject of saw and enter, but other languages will not. Some languages allow both, with only context to clear out the ambiguity, or they will have grammatical marking that indicates either switch-function or same-function use on the verb.
Coverbs, finally, are something that looks very similar to a preposition, but are in fact verbs. In many SVC languages, the coverbs can be used as semantically full verbs in addition to their more grammaticalizing thing. So you could have a benefactive coverb that, when used as a major verb, simply means help or something similar. You could have an adessive coverb meaning 'to, toward' which means 'come to' when used as a major verb. Or the coverbs may be more restricted in their usage. Coverb constructions are very varied from language to language and vastly differ in the extent of grammaticalization they have undergone.
Finally, an important concept to understand when designing an SVC language is the iconic order of events. It seems to be a near-universal for SVC languages that most of the time, the iconic order of events controls the order in which the various components of SVC appear in. So a verb for stopping or quitting will typically come last in an SVC phrase, whereas in a language with deranked verb forms such as English, it may appear before. So we say: I quit smoking, but a speaker of an SVC language is more likely to say I smoke quit.
I hope some of this helps.
The most basic type is called nuclear SVC (nSVC) which involves two equally meaningful verbs concatenated. English has this for a few very restricted uses, especially in the imperative, such as: come look, go check it out. Nuclear SVC can either indicate simultaneity or cause-effect relationships. Typically any arguments in a nSVC will apply equally to each verb in the construction, but not necessarily. Some SVC languages will allow mixing together intransitives and transitives, for example, while others will not allow that.
Asymmetric SVC (aSVC) looks very similar to nuclear SVC, but they differ in that one of the verbs (the “minor verb”, mV) comes from a restricted set and may be semantically bleached to some extent, serving a grammatical function at least partly. Depending on the exact type of construction and the language's grammar, the minor verb may be placed either before or after the major verb.
Auxiliary SVC (auxSVC) is what we call it when a language with no deranked verb forms serializes an auxilary verb directly to an unmodified major verb. Instead of a indicative + infinitive type deal such as want to see, you simply get want see. The language may or may not treat these constructions the same as aSVC.
Exceptional case marking SVC (ECM SVC) or switch-function SVC (sfSVC) is when a argument is used as the object of one verb but the subject of another. This is analoguous to accusativus cum infinitivo in Indo-European languages such as English, where you can say: I want him to leave the house or she made me break the window, or it's close cousin the accusativus cum gerundivo: I saw him leaving the house. Unlike English and Latin, SVC language will often not require any kind of deranked verb form (like the infinitive or the gerund) for the secondary verb, using a fully finite verb form instead: she tricked me eat poison.
Not all SVC languages have ECM type SVC. Many, like Chinese, will interpret the subject as referring to both verbs: I saw girl enter room where I is the subject of saw and enter, but other languages will not. Some languages allow both, with only context to clear out the ambiguity, or they will have grammatical marking that indicates either switch-function or same-function use on the verb.
Coverbs, finally, are something that looks very similar to a preposition, but are in fact verbs. In many SVC languages, the coverbs can be used as semantically full verbs in addition to their more grammaticalizing thing. So you could have a benefactive coverb that, when used as a major verb, simply means help or something similar. You could have an adessive coverb meaning 'to, toward' which means 'come to' when used as a major verb. Or the coverbs may be more restricted in their usage. Coverb constructions are very varied from language to language and vastly differ in the extent of grammaticalization they have undergone.
Finally, an important concept to understand when designing an SVC language is the iconic order of events. It seems to be a near-universal for SVC languages that most of the time, the iconic order of events controls the order in which the various components of SVC appear in. So a verb for stopping or quitting will typically come last in an SVC phrase, whereas in a language with deranked verb forms such as English, it may appear before. So we say: I quit smoking, but a speaker of an SVC language is more likely to say I smoke quit.
I hope some of this helps.
vec
Re: the mechanics of serial verbs vs prepositions?
Good info!
Is there anywhere to read up on this? I assume different languages have different orders, so it would be helpful to see what some of those look like.vec wrote:Finally, an important concept to understand when designing an SVC language is the iconic order of events.
Tibetan Dwarvish - My own ergative "dwarf-lang"
Quasi-Khuzdul - An expansion of J.R.R. Tolkien's Dwarvish language from The Lord of the Rings
Quasi-Khuzdul - An expansion of J.R.R. Tolkien's Dwarvish language from The Lord of the Rings
Re: the mechanics of serial verbs vs prepositions?
That isn't the case. Most verbs, including "see", are ECM in Chinese. May I see where you read that?vec wrote:Not all SVC languages have ECM type SVC. Many, like Chinese, will interpret the subject as referring to both verbs: I saw girl enter room where I is the subject of saw and enter,
Chinese is an exception, both the sponaneous "stop" and the causative "stop" are put before the verb of the action that's stopped.vec wrote:Finally, an important concept to understand when designing an SVC language is the iconic order of events. It seems to be a near-universal for SVC languages that most of the time, the iconic order of events controls the order in which the various components of SVC appear in. So a verb for stopping or quitting will typically come last in an SVC phrase, whereas in a language with deranked verb forms such as English, it may appear before. So we say: I quit smoking, but a speaker of an SVC language is more likely to say I smoke quit.
Re: the mechanics of serial verbs vs prepositions?
Sorry, I stand corrected. I was just inferring from previous parts of the conversation. Point was, languages differ in whether they have ECM at all, whether ECM and non-ECM coexist, or whether V NP V constructions are exclusively ECM.M Mira wrote:That isn't the case. Most verbs, including "see", are ECM in Chinese. May I see where you read that?vec wrote:Not all SVC languages have ECM type SVC. Many, like Chinese, will interpret the subject as referring to both verbs: I saw girl enter room where I is the subject of saw and enter,Chinese is an exception, both the sponaneous "stop" and the causative "stop" are put before the verb of the action that's stopped.vec wrote:Finally, an important concept to understand when designing an SVC language is the iconic order of events. It seems to be a near-universal for SVC languages that most of the time, the iconic order of events controls the order in which the various components of SVC appear in. So a verb for stopping or quitting will typically come last in an SVC phrase, whereas in a language with deranked verb forms such as English, it may appear before. So we say: I quit smoking, but a speaker of an SVC language is more likely to say I smoke quit.
vec
- 2+3 clusivity
- Avisaru
- Posts: 454
- Joined: Fri Mar 16, 2012 5:34 pm
Re: the mechanics of serial verbs vs prepositions?
Mark (or one of the other admins), please add this to the L&L after it calms down a bit.
---
Chinese has been mentioned as a well-known example of SVC. Within IE languages, many New Indo-Aryan languages, such as Hindi-Urdu (as part of the south Asian areal group), have system(s) somewhere between Vec's aSVC and auxSVC -- a largely closed class of verbs, semantic bleaching, etc. See Wikipedia's Hindustani grammar page, which is more complete than those for related languages.
Hindi~Urdu has three constructions of note (Hindi is SOV by default with morphological split-ergativity with quirky case marking and, as I understand it, syntactic nom-acc alignment):
The “conjunctive participle” construction. While in most cases not SVC-like, it provides a starting point for comparison --
Basic form: [ . . . Non-finite verb . . . Finite Verb]
features: typically share the same A/S, non-finite verb is subordinate, finite verb acts as the main verb. Examples from C. P. Masica 1976 at 108:
<usne uskii ããkhõ mẽ dekhaa. usne darwaazaa band kar diyaa.>
He looked into her eyes. He closed the door.
<usne uskii ããkhõ mẽ dekhkar darwaazaa band kar diyaa.>
He looked into her eyes and closed the door. OR Looking (having looked) into her eyes, he closed the door.
The “explicator verb” construction
basic form: [ . . . Non-finite verb (. . .) Finite Verb] – where only emphatic elements intervene between the verbs.
Features: non-finite verb acts as the main verb, finite verb acts is subordinate (auxiliary?), finite verb is lexically bleached to varying degrees. Examples from C. P. Masica 1976 at 141-2:
<. . . le jaanaa>
take away OR lit. “take (and) go”
<likh deenaa>
write down (for somebody) OR lit. “write (and) give”
Notes: Unlike simple “compound verbs” which co-occur with lexically selected helpers – eg. <saaf kaarna> to clean OR lit. clean do -- this construction occurs with pragmatically selected finite verbs. The source notes that the limited (not a closed class but close to it – language by language in the area about 20 or less verbs occur in this construction) set of verbs occurring in these constructions often fall into a directional pattern: away from the speaker: go, run away, give, send, release; towards the speaker: come, take arrive; up: rise, emerge; down: fall, throw, descend; motionless: sit, stand, put – though also sudden: fall and violent: throw. C. P. Masica 1976 at 143. The author goes on to add that the finite verb can add a sense of: completion, suddenness, directionality, benefaction, intensity, violence, stubbornness, reluctance, regret, forethought, thoroughness, etc. to the infinite verb. The author also notes that use of these verbs adds a sense of completeness (elaboration?), which may nevertheless co-occur with imperfective forms.
The continuous aspect marker
As discussed above, the explicator forms build upon increasingly bleached left-fringe finite verbs. In Hindi, this has led to <rah-> the frozen perfective of <rahnaa> “to remain” acting as a continuous aspectual marker for the Continuous.
The source also compares these constructions -- and others in S. Asia to Chinese-esque serial verb constructions, SE direction equivalents, and central and eastern European verb prefixes. Cool stuff,
---
Chinese has been mentioned as a well-known example of SVC. Within IE languages, many New Indo-Aryan languages, such as Hindi-Urdu (as part of the south Asian areal group), have system(s) somewhere between Vec's aSVC and auxSVC -- a largely closed class of verbs, semantic bleaching, etc. See Wikipedia's Hindustani grammar page, which is more complete than those for related languages.
Hindi~Urdu has three constructions of note (Hindi is SOV by default with morphological split-ergativity with quirky case marking and, as I understand it, syntactic nom-acc alignment):
The “conjunctive participle” construction. While in most cases not SVC-like, it provides a starting point for comparison --
Basic form: [ . . . Non-finite verb . . . Finite Verb]
features: typically share the same A/S, non-finite verb is subordinate, finite verb acts as the main verb. Examples from C. P. Masica 1976 at 108:
<usne uskii ããkhõ mẽ dekhaa. usne darwaazaa band kar diyaa.>
He looked into her eyes. He closed the door.
<usne uskii ããkhõ mẽ dekhkar darwaazaa band kar diyaa.>
He looked into her eyes and closed the door. OR Looking (having looked) into her eyes, he closed the door.
The “explicator verb” construction
basic form: [ . . . Non-finite verb (. . .) Finite Verb] – where only emphatic elements intervene between the verbs.
Features: non-finite verb acts as the main verb, finite verb acts is subordinate (auxiliary?), finite verb is lexically bleached to varying degrees. Examples from C. P. Masica 1976 at 141-2:
<. . . le jaanaa>
take away OR lit. “take (and) go”
<likh deenaa>
write down (for somebody) OR lit. “write (and) give”
Notes: Unlike simple “compound verbs” which co-occur with lexically selected helpers – eg. <saaf kaarna> to clean OR lit. clean do -- this construction occurs with pragmatically selected finite verbs. The source notes that the limited (not a closed class but close to it – language by language in the area about 20 or less verbs occur in this construction) set of verbs occurring in these constructions often fall into a directional pattern: away from the speaker: go, run away, give, send, release; towards the speaker: come, take arrive; up: rise, emerge; down: fall, throw, descend; motionless: sit, stand, put – though also sudden: fall and violent: throw. C. P. Masica 1976 at 143. The author goes on to add that the finite verb can add a sense of: completion, suddenness, directionality, benefaction, intensity, violence, stubbornness, reluctance, regret, forethought, thoroughness, etc. to the infinite verb. The author also notes that use of these verbs adds a sense of completeness (elaboration?), which may nevertheless co-occur with imperfective forms.
The continuous aspect marker
As discussed above, the explicator forms build upon increasingly bleached left-fringe finite verbs. In Hindi, this has led to <rah-> the frozen perfective of <rahnaa> “to remain” acting as a continuous aspectual marker for the Continuous.
The source also compares these constructions -- and others in S. Asia to Chinese-esque serial verb constructions, SE direction equivalents, and central and eastern European verb prefixes. Cool stuff,
linguoboy wrote:So that's what it looks like when the master satirist is moistened by his own moutarde.