Polysynthesis for Novices

The best topics from Languages & Linguistics, kept on a permanent basis.
User avatar
Legion
Avisaru
Avisaru
Posts: 522
Joined: Sat Mar 05, 2005 9:56 pm

Re: Polysynthesis for Novices

Post by Legion »

chris_notts wrote:
Legion wrote: Well this depends how we analyse those verbs and what polysynthetic languages can do, which is like I like Whimemsz to tell us: can polysynthetic languages have transitive verbs with null object, used as a generic statement like in Indo-European languages (eg "He's eating" vs "He's eating chicken")?
I personally don't see any reason that polysynthetic languages can't have ambitransitive verb roots. What reason is there to think that languages with a lot of verbal morphology must only have transitivity alternations via overt derivation?
Well the "morphological visibility constraint" could be interpreted to imply this, but I don't know if it actually implies this, which is why I'm asking.

Ambitransitive verbs may perform quitte different fonctions in different languages (and in iirc the fact that Indo-European ambitransitive verbs do not always perform the same function, depending on particular verbs and arguments, is a notable (though not uniquely) IE feature). I know some languages may have few or no ambitransitive verbs, requiring overt valency adjusting morphology in all cases. And I wonder if this correlates with polysynthesis (or not, which is fine too, but I'd like to know either way).

User avatar
Whimemsz
Avisaru
Avisaru
Posts: 690
Joined: Fri Jun 20, 2003 4:56 pm
Location: Gimaamaa onibaaganing

Re: Polysynthesis for Novices

Post by Whimemsz »

Radagast revived wrote:I just came back to see what was going on and I found this amazing thread by whimesz so I thought I'd do some shameless plugging.
Hello again, Radagast! Good to see you.


Legion: I'm not sure if there's any robust correlation; certainly there are some polysynthetic languages with many ambitransitive verbs, though. Apparently in some Arawak languages (like Tariana and Baniwa of Içana), all "transitive" verbs are actually ambitransitive. Unfortunately my source for this (Aikhenvald and Dixon's The Amazonian Languages) doesn't go into any further detail or provide any examples, so...it's not actually 100% clear what Aikhenvald intends by that statement...

chris_notts
Avisaru
Avisaru
Posts: 275
Joined: Wed Dec 15, 2004 9:05 am
Location: Nottingham, England
Contact:

Re: Polysynthesis for Novices

Post by chris_notts »

Legion wrote: Ambitransitive verbs may perform quitte different fonctions in different languages (and in iirc the fact that Indo-European ambitransitive verbs do not always perform the same function, depending on particular verbs and arguments, is a notable (though not uniquely) IE feature). I know some languages may have few or no ambitransitive verbs, requiring overt valency adjusting morphology in all cases. And I wonder if this correlates with polysynthesis (or not, which is fine too, but I'd like to know either way).
Well, a quick Google search reveals that Classical Nahuatl had some ambitransitive verbs (e.g. sow,plant,bury,grind,crush,steal,work,till (soil),dream). And doesn't Radagast say in his papers that at least some of the modern dialects count as polysynthetic? So that suggests (but doesn't prove) that having ambitransitives is not incompatible with being a "polysynthetic" language.

Having said that, I'm not sure that polysynthetic languages are even a separate type of thing, rather than just being a particularly interesting or notable cluster of mostly independent features that shade off gradually in all directions (with different people placing emphasis on slightly different sets of features). If that's the case, then:

(a) polysynthesis is not something a language is or is not, but a gradient phenomenon.

and

(b) I'm not sure that saying a language is polysynthetic has any additional non-tautological implications. E.g. if you define a polysynthetic language as one with X,Y,Z then obviously polysynthesis -> X,Y,Z, and polysynthesis -> the typological implications of X,Y,and Z, but this is really nothing more than a restatement of your initial definition. If that's the case then the question of whether French is polysynthetic or not is not a particularly interesting one.

Also, if anything I would think it would be easier for a language with verbal agreement with both transitive arguments to have ambitransitives. This is because the use of two agreement markers is also effectively a transitivity marker, so why also use an overt detransitiviser if the lack of an agreement marker already tells you that the verb is being used as an intransitive?
Last edited by chris_notts on Sun Apr 08, 2012 5:52 pm, edited 1 time in total.
Try the online version of the HaSC sound change applier: http://chrisdb.dyndns-at-home.com/HaSC

chris_notts
Avisaru
Avisaru
Posts: 275
Joined: Wed Dec 15, 2004 9:05 am
Location: Nottingham, England
Contact:

Re: Polysynthesis for Novices

Post by chris_notts »

Whimemsz wrote: Legion: I'm not sure if there's any robust correlation; certainly there are some polysynthetic languages with many ambitransitive verbs, though. Apparently in some Arawak languages (like Tariana and Baniwa of Içana), all "transitive" verbs are actually ambitransitive. Unfortunately my source for this (Aikhenvald and Dixon's The Amazonian Languages) doesn't go into any further detail or provide any examples, so...it's not actually 100% clear what Aikhenvald intends by that statement...
A grammar of Tariana wrote: Each transitive verb in Tariana can also be used in intransitive clauses, and thus is ambitransitive (however, ambitransitive verbs do not share all the properties with intransitive verbs, c.f. discussion under 13.3). Most verbs are S=A ambitransitives, e.g. -ñha `eat', -inu `kill, fight', -ña `hit'; -ka `see', -hima `hear'; -sata `greet', etc. This means that the object NP can always be optionally omitted, as in English `eat' (`he is eating dinner' or `he is eating'). Example 3.1 illustrates an ambitransitive verb.

3.1 (ãsi) nu-ñha-ka
(pepper) 1sg+A-eat-REC.P.VIS
`I ate/have eaten (pepper)'
The difference between ambitransitives and intransitives (section 13.3) seems to be related to morphological markers such as causatives. Some derivational morphemes such as the causative force a strictly transitive interpretation of otherwise ambitransitive verbs. However, the causative marker can only be used with a relatively small number of transitive verbs, so presumably for most ambitransitives this is not an issue because they simply don't take the morphological causative in the first place.

Also, by at least some definitions Tariana is not polysynthetic. It has crazy amounts of morphology, but it typically does not have object agreement, or even consistent agreement with intransitive subjects (it is split-S and only active verbs show agreement with their actor). So if you just go by the sheer amount of morphology Tariana would count as polysynthetic, but by other criteria (e.g. my understanding of Baker's definition) it would not count.
Try the online version of the HaSC sound change applier: http://chrisdb.dyndns-at-home.com/HaSC

Radagast revived
Lebom
Lebom
Posts: 87
Joined: Sat Apr 07, 2012 7:41 pm

Re: Polysynthesis for Novices

Post by Radagast revived »

Well, it boils down to definition of polysynthesis. For Baker Polysynthesis is something a language is or isn't - and because he takes this definition he can posit non-tautological consequences - such as specific predictions about quantifiers, word order and clause types. The problem with his definiton is that many prototypically polysynthetic languages fall outside of his definition - such as Greenlandic.

I think that the kinds of languages characterized by the kind of polysynthesis defined by Baker as TEH polysynthesis (I would prefer a term like obligatorily syntactically headmarking or indexing) by necessity do have a much stricter separation of intransitive and transitive roots - and consequently few ambitransitives. So few that you can always argue that they are basically just homophones (I think some formalists would prefer that kind of argument).

chris_notts
Avisaru
Avisaru
Posts: 275
Joined: Wed Dec 15, 2004 9:05 am
Location: Nottingham, England
Contact:

Re: Polysynthesis for Novices

Post by chris_notts »

Radagast revived wrote: I think that the kinds of languages characterized by the kind of polysynthesis defined by Baker as TEH polysynthesis (I would prefer a term like obligatorily syntactically headmarking or indexing) by necessity do have a much stricter separation of intransitive and transitive roots - and consequently few ambitransitives. So few that you can always argue that they are basically just homophones (I think some formalists would prefer that kind of argument).
Can you give a brief summary of the consequences that Baker claims follows from being a polysynthetic language by his definition?

I did Google it to try to remind myself, and the closest (free to access) thing I found was this paper here.

http://www.rci.rutgers.edu/~mabaker/Map ... a-txt2.pdf

It is his attempt to show the difference between a zero agreement marker and a lack of agreement to support his theory, because otherwise he potentially has problems. He points out that there are Bantu languages where objects show different behaviour with and without agreement (object agreement is optional in most Bantu languages), and then there are languages like Mapudugun where agreement can surface or not without any obvious difference in object behaviour. I don't really find his argument convincing since it seems to rely on a lot of additional assumptions, and also seems to assume part of his conclusion at various points (that the polysynthesis parameter and what he believes are its implications hold up).

That is, he concludes that Mapudugun alternates between overt agreement and zero agreement where Bantu languages alternate between agreement and lack of agreement, and that this explains why Mapudugun behaves like a polysynthetic language by his definition (always allowing dislocation of objects etc.). But many of his arguments seem to be from the claimed consequences of polysynthesis, so if his theory is not true then the argument here falls apart as well. The paper seems to say "this must be true for me to be right, therefore it is true".
Try the online version of the HaSC sound change applier: http://chrisdb.dyndns-at-home.com/HaSC

Radagast revived
Lebom
Lebom
Posts: 87
Joined: Sat Apr 07, 2012 7:41 pm

Re: Polysynthesis for Novices

Post by Radagast revived »

The traits he posits as following from polysynthesis (which is a requirement of marking of arguments of phrases on phrasal heads by agreement morphemes or compounding) are - non-configurationality (free word order, allowing disjunct phrases), valency reducing incorporation, no "true" quantifiers, no NP arguments (they're adjuncts). And a couple of more corollaries to those ones that I don't remember off the top of my head.

Except for the notion that this is a 'parameter', I think there seems to be something about his argument that is right. Languages like Nahuatl and Mohawk do seem to have a basic requirement that all direct arguments must be indexed on the phrasal head in some way. And I think that the general syntactic freedom that they also exhibit follows logically from that, as does syntactic incorporation. In Nahuatl and I think in Mohawk there is definitely a difference between a zero agreement morpheme and a lack of agreement. Zero morphemes are in paradigmatic relations with non-zero morphemes and they can be referred back to by anaphora for example.

Baker was exceptional in being a Chomskyan who started studying a language that was actually different - Mohawk. Then he realized that this was truly different from English syntax and posed challenges for the generative framework and he developed a way to analyze that within the generative framework - but then in true Chomskyan fashion he went on to overextend his definitions to a large group of languages that he had not studied in detail - and in doing so making the concept of polysynthesis near meaningless.

I think his description reasonably captures a very important aspect of syntax in Mohawk and Nahuatl. I don't think it is a good definition of polysynthesis.

chris_notts
Avisaru
Avisaru
Posts: 275
Joined: Wed Dec 15, 2004 9:05 am
Location: Nottingham, England
Contact:

Re: Polysynthesis for Novices

Post by chris_notts »

Radagast revived wrote:The traits he posits as following from polysynthesis (which is a requirement of marking of arguments of phrases on phrasal heads by agreement morphemes or compounding) are - non-configurationality (free word order, allowing disjunct phrases), valency reducing incorporation, no "true" quantifiers, no NP arguments (they're adjuncts). And a couple of more corollaries to those ones that I don't remember off the top of my head.
Does he claim a one-way implication (polysynthesis -> non-configurationality) or "if and only if"? Because there are certainly non-configurational languages without verbal agreement with all arguments, if by that we mean free word order and discontinuous noun-phrases. Could you also expand on what is meant by no-true quantifiers?
In Nahuatl and I think in Mohawk there is definitely a difference between a zero agreement morpheme and a lack of agreement. Zero morphemes are in paradigmatic relations with non-zero morphemes and they can be referred back to by anaphora for example.
Oh, clearly in some languages zero does form part of a paradigm that contrasts with other overt markers. But in the paper I referenced both languages displayed a choice between overt agreement and zero for the same (at least on the surface) controlling NP. Zero in these languages does not uniquely own any part of the agreement space, but instead represents a neutralisation of contrast.

Zero vs non-zero is determined in Bantu, and (from reading elsewhere) in Mapudugun, by pragmatic factors such as definiteness and topicality. The difference between them is that in the Bantu languages that Baker discusses, left dislocation is only possible with overt agreement, whereas in Mapudugun word-order is free whether there is an overt agreement morpheme or not.

Is this really due to a lack of agreement vs an alternative zero agreement morpheme?

Well, on the Bantu side, the simplest explanation to me seems to be that, if object agreement in Bantu only occurs if the object is topical, and left-dislocation is a way to mark topical NPs, then it follows that left-dislocated NPs must trigger agreement. The agreement morpheme does not "license" any dislocation; instead the pragmatic function of left-dislocation forces agreement. This is the simplest explanation because agreement is known to be affected by definiteness, topicality etc anyway, so we don't need to add any further complicated constraints to get the observed behaviour.

Conversely, on the Mapudugun side, Baker discusses the free word order regardless of whether an agreement marker is present or not, and the lack of any difference in the syntactic behaviour of objects regardless with agreement choice. By comparing this with Bantu, where the free word-order etc are associated with agreement, he concludes that in Mapudugun zero is also a 3rd person agreement morpheme even though there is also an overt 3rd person agreement morpheme.

However, this really doesn't make any sense because there are plenty of languages with free word order, non-configurationality etc and no object agreement. Comparing Mapudugun with exactly one other family of languages does not seem like a valid approach to proving anything, especially when the simplest and most intuitive explanation is that the lack of the overt 3rd person agreement marker represents a lack of agreement trigger by pragmatic factors. If he had compared it instead with a language with free word order and no object agreement, he'd presumably come to a very different conclusion.

He also, of course, did not do a detailed comparison of the pragmatic effects of word order in Mapudugun and the Bantu languages. If the Bantu phenomenon is explained by the pragmatic effect of left-dislocation and the exact pragmatic factors determining presence vs absence of agreement, then the question is - how does the pragmatic basis of word order and agreement in Mapudugun compare? If fronting the object is less marked in Mapudugun, then is it performing exactly the same function?

Did he do a decent typological survey in his other work on his polysynthesis parameter? Or did he just look at a very small number of languages?
I think his description reasonably captures a very important aspect of syntax in Mohawk and Nahuatl. I don't think it is a good definition of polysynthesis.
I can understand that having certain features like a lot of agreement may affect the structure of the rest of the grammar or create biases in how it's structured. I just didn't agree with the argument in the first paper of his about polysynthesis that I could find via Googling.
Try the online version of the HaSC sound change applier: http://chrisdb.dyndns-at-home.com/HaSC

Radagast revived
Lebom
Lebom
Posts: 87
Joined: Sat Apr 07, 2012 7:41 pm

Re: Polysynthesis for Novices

Post by Radagast revived »

chris_notts wrote: Does he claim a one-way implication (polysynthesis -> non-configurationality) or "if and only if"? Because there are certainly non-configurational languages without verbal agreement with all arguments, if by that we mean free word order and discontinuous noun-phrases. Could you also expand on what is meant by no-true quantifiers?
No, to Baker it is a parameter setting that causes the array of features to appear together. There is no reason why it wouldn't be possible to have either of the single features without having the parameter setting. So it is a one way implication.

Non-true quantifiers are quantifiers that are not heads of a noun phrase as quantifiers usually are supposed to be - that is they do not form an actual phrase with the noun they modify and the noun does not express any agreement with the quantifier. There is also some semantic tricks about the scopes of quantifiers such as "some", "every", "each" that become complicated by the fact that they don't form phrase with the nouns but are simply co-adjuncts to a main phrasal head.

So with a true quantifyier you'd get a tree structure like this

Code: Select all

[[i] broke [two [cups]]
But in Nahuatl you'd get something like this [(it was) two][(it was) cups]

chris_notts wrote: Zero vs non-zero is determined in Bantu, and (from reading elsewhere) in Mapudugun, by pragmatic factors such as definiteness and topicality. The difference between them is that in the Bantu languages that Baker discusses, left dislocation is only possible with overt agreement, whereas in Mapudugun word-order is free whether there is an overt agreement morpheme or not.

Is this really due to a lack of agreement vs an alternative zero agreement morpheme?

Conversely, on the Mapudugun side, Baker discusses the free word order regardless of whether an agreement marker is present or not, and the lack of any difference in the syntactic behaviour of objects regardless with agreement choice. By comparing this with Bantu, where the free word-order etc are associated with agreement, he concludes that in Mapudugun zero is also a 3rd person agreement morpheme even though there is also an overt 3rd person agreement morpheme.

However, this really doesn't make any sense because there are plenty of languages with free word order, non-configurationality etc and no object agreement. Comparing Mapudugun with exactly one other family of languages does not seem like a valid approach to proving anything, especially when the simplest and most intuitive explanation is that the lack of the overt 3rd person agreement marker represents a lack of agreement trigger by pragmatic factors. If he had compared it instead with a language with free word order and no object agreement, he'd presumably come to a very different conclusion.


I don't really remember the mapudungun article, its a long time since I read it. But in Nahuatl I make the argument myself that there is an alternate zero form for the overt 3rd person possessive marker. This is because the zero form only occurs with third person singular possessors that are recoverable from context.

chris_notts wrote: He also, of course, did not do a detailed comparison of the pragmatic effects of word order in Mapudugun and the Bantu languages. If the Bantu phenomenon is explained by the pragmatic effect of left-dislocation and the exact pragmatic factors determining presence vs absence of agreement, then the question is - how does the pragmatic basis of word order and agreement in Mapudugun compare? If fronting the object is less marked in Mapudugun, then is it performing exactly the same function?

Did he do a decent typological survey in his other work on his polysynthesis parameter? Or did he just look at a very small number of languages?

I can understand that having certain features like a lot of agreement may affect the structure of the rest of the grammar or create biases in how it's structured. I just didn't agree with the argument in the first paper of his about polysynthesis that I could find via Googling.


No, he doesn't do typology really - being a formalist he is interested in setting up a model and then see how things fit it. And being a generativist pragmatics doesn't matter much to him either - the questio of dislocation does because that has to do with the boundaries of the sentence - so if the element is dislocated outside of the sentence structure then it no longer counts as syntactically present inside the phrase structure and would have to be expressed somehow else. That may be the same with the Mapudungun example.

Again I am not trying to defend his theory in general - just to say that it accounts fairly well for my observations in one language.
Last edited by Radagast revived on Mon Apr 09, 2012 8:14 pm, edited 2 times in total.

User avatar
Ser
Smeric
Smeric
Posts: 1542
Joined: Sat Jul 19, 2008 1:55 am
Location: Vancouver, British Columbia / Colombie Britannique, Canada

Re: Polysynthesis for Novices

Post by Ser »

Legion wrote:
Serafín wrote:
Legion wrote:Though I realise there are also cases of direct object dropping outside of ditransitive verbs, very common sentences like "Mozart, j'aime" (for literary "Mozart, je l'aime") or "Il est venu hier, je sais" (literary "je le sais").
Out of curiosity why do you analyze these are direct object dropping? Is "if a COD, it should reappear with a pronoun in the verb" an actual rule?
Well, you have the basic sentence structure of French, so say: "Marie aime Mozart".

Then you have the so called "dislocations", where an argument of the verb is moved around, and ends up removed from the verb position. When this happens, the rule is that whatever element has been removed must be replaced by a personal clitic on the verb, so:
But my question is, is this a real rule? What if it's not always necessary to replace the dislocated argument with a pronoun, but rather, it depends on more factors? The rules for anaphoric pronouns the Spanish Academies recommend for dislocation are somewhat complex, sometimes they're added and sometimes not. I highly suspect this analysis of French is mistaken.
I'm not quite sure of that, I'm fairly certain sentences like "fais voir" "let (me) see (it)" or "mange!" "eat (it)" occure frequently in speech.
What if these imperatives are more about the actions though? The first one as in, "act in order to let the act of seeing possible!"; or "eat!" as in "start the act of eating", I mean... In Spanish (or at least in my dialect of it), you can say dejá ver and ¡comé! too, and that's more or less the sense I feel I'm giving. I don't feel there's a direct object missing, it's all about the action as it is.
Well this depends how we analyse those verbs and what polysynthetic languages can do, which is like I like Whimemsz to tell us: can polysynthetic languages have transitive verbs with null object, used as a generic statement like in Indo-European languages (eg "He's eating" vs "He's eating chicken")?
What I really meant to say was, what if these are just imperatives of intransitive verbs and not transitives with null objects?

chris_notts
Avisaru
Avisaru
Posts: 275
Joined: Wed Dec 15, 2004 9:05 am
Location: Nottingham, England
Contact:

Re: Polysynthesis for Novices

Post by chris_notts »

Radagast revived wrote: Again I am not trying to defend his theory in general - just to say that it accounts fairly well for my observations in one language.
Well, thanks for giving a summary of his theory.

Also, I didn't realise that Nahuatl had discontinuous / non-integrated NPs. I've got a book that's supposed to teach you Classical Nahuatl ("An Introduction to Classical Nahuatl" by Michel Launey), but I haven't really read that much of it.
Try the online version of the HaSC sound change applier: http://chrisdb.dyndns-at-home.com/HaSC

User avatar
Legion
Avisaru
Avisaru
Posts: 522
Joined: Sat Mar 05, 2005 9:56 pm

Re: Polysynthesis for Novices

Post by Legion »

Serafín wrote:
Legion wrote:Well, you have the basic sentence structure of French, so say: "Marie aime Mozart".

Then you have the so called "dislocations", where an argument of the verb is moved around, and ends up removed from the verb position. When this happens, the rule is that whatever element has been removed must be replaced by a personal clitic on the verb, so:
But my question is, is this a real rule? What if it's not always necessary to replace the dislocated argument with a pronoun, but rather, it depends on more factors? The rules for anaphoric pronouns the Spanish Academies recommend for dislocation are somewhat complex, sometimes they're added and sometimes not. I highly suspect this analysis of French is mistaken.
Well, as far as I can tell and observe, this rule is extremely solid in most context.

For subjects it's clearly there, null subject in French are rare outside of dropping impersonal "il" (and even there not all is permitted: colloquial French allows "faut partir" (we/you should go) or "faudrait" (that should be) but not "*faut" (that must be)).

In case of subject dislocation, a personal clitic is obligatory:

"Marie aime Mozart" > "Elle aime Mozart, Marie". I don't think any native speaker would see "*aime Mozart, Marie" (meaning "Marie likes Mozart") as grammatical, and could even parse it correctly; that sounds like an imperative here, indeed: "Aime Mozart, Marie!" ("Love Mozart, Marie!")

For 1st and 2nd person direct object, there is obligatory dislocation (because French true personal pronouns of 1st and 2nd person cannot function as verbal arguments: "*moi veux" or "*François aime toi" are gross barbarisms; 3rd persons pronouns are a bit more flexible: "Faites ceci!" (do this) is a perfectly fine sentence even in the most sophisticated registers), and the personal clitic is thus generally obligatory as well.

"François t'aime, toi" (François likes you) > "*Français aime, toi" sounds like the misformed sentence above.

For 1st/2nd indirect object there is some flexibility.

"Marie me l'a donné, à moi" (Marie gave it to me) > "Marie l'a donné, à moi" (less formal, but acceptable).

The most flexibility you get is with 3rd person direct and oblique objects, but even there you find case were dropping is difficult.

"Je vois le chien" (I see the dog) > "Le chien, je le vois" > "?Le chien, je vois" (the latter would rather be parsed as "Yes, I see, the dog, I understand what you mean")

And all this is colloquial French.

Literary French forbids those kind of clitic dropping in dislocation, and to that account is more polysynthetic than coloquial French. I don't know to which extense those rules are "artificial", but they have a pretty solid foundation even in coloquial language in practice.
What I really meant to say was, what if these are just imperatives of intransitive verbs and not transitives with null objects?
I don't know; at least in the case of "fais voir", the implication is generally "show me that thing", there's generally an implied object. Like "J'ai trouvé un lézard!" "Fais voir!" (I found a lizard!—Let me see it!)

Radagast revived
Lebom
Lebom
Posts: 87
Joined: Sat Apr 07, 2012 7:41 pm

Re: Polysynthesis for Novices

Post by Radagast revived »

chris_notts wrote:
Radagast revived wrote: Again I am not trying to defend his theory in general - just to say that it accounts fairly well for my observations in one language.
Well, thanks for giving a summary of his theory.

Also, I didn't realise that Nahuatl had discontinuous / non-integrated NPs. I've got a book that's supposed to teach you Classical Nahuatl ("An Introduction to Classical Nahuatl" by Michel Launey), but I haven't really read that much of it.
That's the best one. His doctoral dissertation that it is based on is a lot more comprehensive and linguistically sophisticated. Launey refers to the same phenomenon described by Baker as "omnipredicativity" - arguing that every phrases (NP and VP) is syntactically a predicate. His 1500 page doctoral dissertation is available here: http://celia.cnrs.fr/FichExt/Etudes/Launey/tm.htm

User avatar
Vuvuzela
Avisaru
Avisaru
Posts: 317
Joined: Fri Apr 13, 2012 1:34 pm

Re: Polysynthesis for Novices

Post by Vuvuzela »

"Eskimoan languages, for example, lack noun incorporation under most definitions, but they're highly polysynthetic"
Really? I'm somewhat familiar with Inuktitut, and I think it actually does incorporate nouns into it's verbs. For example:
ᐲᑕ ᐊᒻᒪᓗ ᒪᕆ ᐃᓪᓗᒨᖅᖃᐅᔪᓯᒃ
IPA: piː'ta amːa'lu ma'ri ilːumuːqːaudʒu'sit
(Piita and Mari House-OBJ/go.toPAST-SEMI-IMED/SUBJ-3p.du. [I think, I'm terrible at glossing])
Piita and Mari went to the house.
In this example, "House" gets shoved into "go to" so that one word can mean "the two of them go to the house." and stand on it's own as a sentence. Unless I'm not as familiar with Inuktitut as I think, or that doesn't count as noun incorporation.

User avatar
Xephyr
Avisaru
Avisaru
Posts: 821
Joined: Sat May 03, 2003 3:04 pm

Re: Polysynthesis for Novices

Post by Xephyr »

Vuvgangujunga wrote:"Eskimoan languages, for example, lack noun incorporation under most definitions, but they're highly polysynthetic"
Really? I'm somewhat familiar with Inuktitut, and I think it actually does incorporate nouns into it's verbs. For example:
ᐲᑕ ᐊᒻᒪᓗ ᒪᕆ ᐃᓪᓗᒨᖅᖃᐅᔪᓯᒃ
IPA: piː'ta amːa'lu ma'ri ilːumuːqːaudʒu'sit
(Piita and Mari House-OBJ/go.toPAST-SEMI-IMED/SUBJ-3p.du. [I think, I'm terrible at glossing])
Piita and Mari went to the house.
In this example, "House" gets shoved into "go to" so that one word can mean "the two of them go to the house." and stand on it's own as a sentence. Unless I'm not as familiar with Inuktitut as I think, or that doesn't count as noun incorporation.
"House" in usually considered the root there, and "go to" a verb-forming post-base*. The justification is that you can't inflect postbases as independent verb roots.

* - At least, I'm assuming, based on what I know about how Yup'ik works, and more-or-less consistently hearing that Inuktitut behaves the same way. Maybe what you have there is a totally different kind of construction though, I dunno.

EDIT: Oh shit. Apparently Iñupiaq (another Inuit language) allows noun-incorporation with full verbs, so I was totally talking out of my ass there.

By the way, you should probably use the standard transliteration instead of IPA; it's way easier to read. So that'd be "piita ammalu mari illumuuqqaujusit"

By the way too, what is "ammalu"? You gloss it as "and", but wouldn't "Peter and Mari" just be "Piita Marilu"?
"It will not come by waiting for it. It will not be said, 'Here it is,' or 'There it is.' Rather, the Kingdom of the Father is spread out upon the earth, and men do not see it."
The Gospel of Thomas

User avatar
Vuvuzela
Avisaru
Avisaru
Posts: 317
Joined: Fri Apr 13, 2012 1:34 pm

Re: Polysynthesis for Novices

Post by Vuvuzela »

Xephyr wrote:
Vuvgangujunga wrote:"Eskimoan languages, for example, lack noun incorporation under most definitions, but they're highly polysynthetic"
Really? I'm somewhat familiar with Inuktitut, and I think it actually does incorporate nouns into it's verbs. For example:
ᐲᑕ ᐊᒻᒪᓗ ᒪᕆ ᐃᓪᓗᒨᖅᖃᐅᔪᓯᒃ
IPA: piː'ta amːa'lu ma'ri ilːumuːqːaudʒu'sit
(Piita and Mari House-OBJ/go.toPAST-SEMI-IMED/SUBJ-3p.du. [I think, I'm terrible at glossing])
Piita and Mari went to the house.
In this example, "House" gets shoved into "go to" so that one word can mean "the two of them go to the house." and stand on it's own as a sentence. Unless I'm not as familiar with Inuktitut as I think, or that doesn't count as noun incorporation.
"House" in usually considered the root there, and "go to" a verb-forming post-base*. The justification is that you can't inflect postbases as independent verb roots.

* - At least, I'm assuming, based on what I know about how Yup'ik works, and more-or-less consistently hearing that Inuktitut behaves the same way. Maybe what you have there is a totally different kind of construction though, I dunno.

EDIT: Oh shit. Apparently Iñupiaq (another Inuit language) allows noun-incorporation with full verbs, so I was totally talking out of my ass there.

By the way, you should probably use the standard transliteration instead of IPA; it's way easier to read. So that'd be "piita ammalu mari illumuuqqaujusit"

By the way too, what is "ammalu"? You gloss it as "and", but wouldn't "Peter and Mari" just be "Piita Marilu"?
Oops. I was blanking on how to say "and", so I looked it up in a dictionary. Apparently, the dictionary I was using was using "ᐊᒻᒪ/amma" as a placeholder or something, unless their translation of "and" meant something awkward like "also". I have to start using my brain when trying to remember foreign grammars.
And I'll do that next time I'm transliterating Inuktitut.

User avatar
Whimemsz
Avisaru
Avisaru
Posts: 690
Joined: Fri Jun 20, 2003 4:56 pm
Location: Gimaamaa onibaaganing

Re: Polysynthesis for Novices

Post by Whimemsz »

Huh. I didn't know that. Thanks, Vuvgangujunga, I edited the first post accordingly.

User avatar
Aurora Rossa
Smeric
Smeric
Posts: 1138
Joined: Mon Aug 11, 2003 11:46 am
Location: The vendée of America
Contact:

Re: Polysynthesis for Novices

Post by Aurora Rossa »

What are some ways in which polysynthetic languages typically form dependent clauses as opposed to independent ones? Vohpenonomae said that some of the languages he studied drew no distinction between independent clauses and relative ones. He really commented on adverbial clauses or anything like that as far as I can recall. I have read that some polysynthetic languages, at least, have entirely separate verb paradigms for different kinds of clauses.
Image
"There was a particular car I soon came to think of as distinctly St. Louis-ish: a gigantic white S.U.V. with a W. bumper sticker on it for George W. Bush."

User avatar
Xephyr
Avisaru
Avisaru
Posts: 821
Joined: Sat May 03, 2003 3:04 pm

Re: Polysynthesis for Novices

Post by Xephyr »

Sure. Here are some of the dependent moods from Nunavik Inuktitut.

Imperfective a.k.a. Conditional:

aullaruma "if I leave; when I leave"
utaqqigukkit "if I wait for them; when [future] I wait for them"
aullaruvit aliasulaaqtunga "if you leave or when you leave, I'll be glad"

Dubitative:

aullalaarmangaarma "whether I shall leave"
takummangaaqpiuk "whether you see him/her"
nalujunga qanga aullalaurmangaaqpit "I didn't know when you left"

Perfective Appositional:

tusaaqsunga aliasuttunga "while hearing, I am glad"
ikajuqsugu aliasuttuq "while helping him/her, he/she is glad"

Imperfective Appositional:

tusarlunga aliasulaartunga "while hearing, I shall be glad"
ikajurlugu aliasulaaqtuq "while helping him/her, he/she will be glad"

Negative Appositional:

tusarnanga aliasunngitunga "while not hearing, I am not glad"
ikajurnagu aliasunngituq "while not helping him/her, he/she is not glad"
takutinnak aliasuttut "while you are not seeing, they are glad"
"It will not come by waiting for it. It will not be said, 'Here it is,' or 'There it is.' Rather, the Kingdom of the Father is spread out upon the earth, and men do not see it."
The Gospel of Thomas

User avatar
Salmoneus
Sanno
Sanno
Posts: 3197
Joined: Thu Jan 15, 2004 5:00 pm
Location: One of the dark places of the world

Re: Polysynthesis for Novices

Post by Salmoneus »

Radagast revived wrote:
chris_notts wrote:
Xephyr wrote:
chris_notts wrote:
Radagast is back! Did you write that article? I remember you doing work on similar things at one point.
The way you say that makes it sound like you have access to "Anthropological Linguistics", chris... *Ahem, ahem*...
I'm afraid not. I just read the abstract. I wish I did have access, but as I'm not affiliated with a university getting access is expensive.

Yes, I did. You can also find a preprint here. I just came back to see what was going on and I found this amazing thread by whimesz so I thought I'd do some shameless plugging.
Radagast! Stay! (How've you been?)
Blog: [url]http://vacuouswastrel.wordpress.com/[/url]

But the river tripped on her by and by, lapping
as though her heart was brook: Why, why, why! Weh, O weh
I'se so silly to be flowing but I no canna stay!

User avatar
Whimemsz
Avisaru
Avisaru
Posts: 690
Joined: Fri Jun 20, 2003 4:56 pm
Location: Gimaamaa onibaaganing

Re: Polysynthesis for Novices

Post by Whimemsz »

Jabechasqvi wrote:What are some ways in which polysynthetic languages typically form dependent clauses as opposed to independent ones? Vohpenonomae said that some of the languages he studied drew no distinction between independent clauses and relative ones. He really commented on adverbial clauses or anything like that as far as I can recall. I have read that some polysynthetic languages, at least, have entirely separate verb paradigms for different kinds of clauses.
Actually, I originally thought about including a section on this -- but then the whole project spun out of control and I was like, "screw it." Plus I haven't actually looked into this question in any depth at all. So all I can go by is how it's done in the polysynthetic languages I have some familiarity with.

In these cases, subordination seems to quite often be marked at least partly with verbal inflection, as one might imagine given how verb-centric polylangs often are. In Algonquian languages, there's a distinct inflectional paradigm for dependent verbs, called the "conjunct" verbal order (though subordination of various kinds is frequently marked in addition with verbal prefixes and with subordinating adverbial particles and conjunctions). For example, Ojibwe Giishpin gimiwang, gaawiin niwii-pi-izhaasii, "If it rains, I'm not going to come" (with the subordinating conjunction/particle giishpin, "if", and the verb gimiwan "to rain" inflected with the inanimate third person singular conjunct suffix -g).

Muskogean languages like Koasati (sect. 5.5 above) have various subordinating suffixes as well (as exemplified by the consequence suffixes and others given above). They also have switch-reference suffixes, which mark whether the subject of a following or dependent verb is coreferent with the subject of the preceding or matrix verb (these go in suffix slot 14 in Koasati). For example: athómmak yomáhlin chalakkí hokáhchok, "They called the wandering Indians Cherokees" (with the different subject marker -n on the verb yomáhli, "to go about").

There's some languages with much more intricate switch-reference/subordination systems, but the ones I know of aren't really polysynthetic. I don't know if that's a general pattern or not. A number of Amazonian languages use nominalizations (or nominalization-like processes) to form relative clauses and other subordinate clauses, but again many Amazonian languages are only moderately synthetic (and some are closer to English or German-level synthesis).

But, I can give some examples from Nambikuara, which has a crapload of subordinating suffixes (I've omitted the superscript numbers marking tones, because they're way too distracting):

ADVERSATIVE
-nxahate, "attention switch"
-ta, "contrast"
-tota, "contra-expectation"

SEQUENTIAL
-katṵ, "normal sequence"
-tekaka, "expected climactic sequence"
-tei, "contra-expected climactic sequence"
-tḭkxailu, "concomitant"

LOGICAL
-kxainãntṵ, "high probability"
-kelatekxai, "low probability"
-kxahata̰nxãntṵ, "concessional"
-kxaya̰ntxisu, "precautionary"

COMPLEMENT
-jutsu, "static"
-jausu, "thought"
-kxesu, "global"

NONPERSONAL RELATIVE
-sxã, "immediate sequence" (/present participle)
-nũla, "delayed sequence"

CLIMACTIC RELATIVE
-tena, "immediate response"
-nana, "delayed response"

DYNAMIC RELATIVE
-hakxai, "causal"
-kxayusu, "additional"

Some examples of these in use (where "TNS" stands for various combinations of tense, aspect, evidential, etc. suffixes):

Kaxyuhxa ĩnakatṵ ãuxinahẽla = "After I ate the meat, I slept" (= meat eat-1sg-katṵ sleep-1sg-TNS)
Ĩxainjũta ko̰nxainnxahẽla = "I didn't know about their coming home" (= come.home-3pl-jũta know-3pl-1sg.NEG-TNS)
Walinxa ũwhxetsxã sata̰ainnahẽla = "Grating the manioc, they set it aside" (= manioc grate-sxã set.aside-3pl-TNS)

Unfortunately, the grammar doesn't actually specify what all of these descriptions MEAN... which kind of reduces the usefulness. But whatever.


[NB: These Nambikuara examples are from Menno Kroeker (2001). "A Descriptive Grammar of Nambikuara." International Journal of American Linguistics 67(1): 1-87]

Gojera
Lebom
Lebom
Posts: 96
Joined: Sun May 22, 2011 8:46 pm

Re: Polysynthesis for Novices

Post by Gojera »

Ainu probably isn't the most polysynthetic language in the world. But it handles relative clauses in much the same way Japanese does. Japanese, of course, is a dependent-marking, strongly head-final language: the verb comes at the end of a clause, and a verb that precedes a noun forms a relative clause describing that noun. So yama ga noboritagatteiru otoko "the man that wants to climb the mountain". Ainu does things similarly. So compare:

a-maci
1SG-wife
"my wife"

e-kor mat
2SG-have wife
"the wife that you have"

Radagast revived
Lebom
Lebom
Posts: 87
Joined: Sat Apr 07, 2012 7:41 pm

Re: Polysynthesis for Novices

Post by Radagast revived »

Salmoneus wrote: Radagast! Stay! (How've you been?)
Good thanks, lots of work including fieldwork on Otomi - now living in the US and preparing to go to the field to work on a dissertation on Language policy in Mexico. I trust it you are also doing well.

User avatar
ol bofosh
Smeric
Smeric
Posts: 1169
Joined: Tue Dec 06, 2011 5:30 pm
Location: tʰæ.ɹʷˠə.ˈgɜʉ̯.nɜ kʰæ.tə.ˈlɜʉ̯.nʲɜ spɛ̝ɪ̯n ˈjʏː.ɹəʔp

Re: Polysynthesis for Novices

Post by ol bofosh »

I've now read through Whimemsz work; well done, that's given me something to read again and again. :)

I didn't understand all of it, but what I did understand fascinated me. Now I'll go through this thread with some (little) idea about what is being said. :-D
It was about time I changed this.

User avatar
Ser
Smeric
Smeric
Posts: 1542
Joined: Sat Jul 19, 2008 1:55 am
Location: Vancouver, British Columbia / Colombie Britannique, Canada

Re: Polysynthesis for Novices

Post by Ser »

I was reading your posts more carefully this time, and...
Whimemsz wrote:8 - How Does Polysynthesis Arise?

[...]

8.2 - Polypersonal Marking

The pathway by which polypersonal marking on verbs develops is quite straightforward: through grammaticalization, independent pronouns become cliticized with the verb root and eventually become inseperable affixes, often with some phonological reduction from their earlier form.

Various Romance languages actually offer good examples of this. Most Romance languages continue the Latin system of already marking the subject with an inflection on the verb (recall the Spanish examples above: hablo, "I speak"; hablas, "you speak", etc.). However, Romance languages also mark objects on the verb as well, using pronominal clitics, though the placement of these clitics varies from language to language and depending on the exact situation. Some examples (again from Spanish) can help demonstrate this:
  • no me lo digas, "don't tell me that" (more accurately: no me=lo=dig-as, don't to.me=it.OBJ-say:SUBJUNCTIVE-you.SUBJ). Here, the subject is marked with a verb suffix (-as), as usual in Romance, and both the direct object (me=, "me") and indirect object (lo=, "it, that") are indicated with clitics that are attached to the beginning of the verb.
  • ayer la vi, "I saw her yesterday" (more accurately: ayer la=vi, yesterday her=I:saw). Here again, the verb is inflected to mark the subject, and the direct object (la=, "her") is marked with a clitic preposed to the verb.
  • melo, "give it to me" (more accurately: dá=me=lo, give:2sg.SUBJ.IMPERATIVE=to.me=it.OBJ). Once again the verb inflects to mark its subject, and the object(s) are marked with clitics, though in this case they follow the verb (and by Spanish spelling convention, are written as one word).
Note that these object pronouns are reduced versions of the full pronouns of Latin: thus, lo for example is from Latin illum, "that", which has irregularly lost its initial vowel in the process of gramaticalization into an unstressed clitic. Note also that unlike in Latin, these Romance object clitic pronouns cannot freely occur in many different positions. Instead, there are a limited number of places within the clause where they can occur (generally either directly before or directly following the verb): this is another indication that they have become grammaticalized and are no longer completely independent pronouns, but rather are partly on the way to becoming verbal affixes marking person. Many Romance languages, thus, are a good demonstration of the beginning stages of the creation of polypersonal marking.

In fact, one Romance language is well known for having already developed true polypersonal marking, and is sometimes called "polysynthetic": French. In its evolution from Latin, French has undergone a number of phonological reductions, which have ultimately resulted in the French verb no longer effectively inflecting to mark the person of its subject, as other Romance languages are capable of (see here for an overview of some of these changes). As a result, French makes much more frequent use of personal pronouns than other Romance languages -- but in unmarked contexts these "pronouns", in fact, have become fused to the verb, both phonologically and morphosyntactically. Though they are still (sometimes) written as separate pronouns in the standard orthography, there are arguments for considering them true verbal affixes marking person.

Take the French sentence, je vais le lui donner, "I'm going to give it to him/her". Though written as several separate words, this is phonologically a single word, [ʒvɛləlɥidɔne]. There are also syntactic criteria for considering this a single word, though I'm not going to get into them here -- if you're curious, see this PDF for a number of arguments in favor of considering sentences like this as a single grammatical word in French).[...] The point here, though, is not to prove whether spoken French should be considered "polysynthetic"; the point is that this provides an example of how polypersonal marking could arise, whether or not French's pronominal verbal markers are clitics or true affixes.
I'd suggest that this section could be better improved with examples where not only these verbs in Spanish and French have pronoun clitics (the verb and the pronouns form really one word), but moreover, these clitics in both languages are developing into affixes where the verb agrees with its arguments in polypersonal agreement.

In Spanish making the verb mark that it has an indirect object is already the most common construction in both literary and colloquial Spanish, especially the latter. And it's even obligatory with certain verbs even in literary.
  • Le he dicho a él que venga mañana. (Madrid Spanish)
    le-he-dicho a=él que veng-a mañana
    3SG.IO-have.1SG.SBJ-told to=3SG that come-3SG.SBJ;SUBJ tomorrow
    'I've told him to come tomorrow.' (IO = indirect object, SBJ = subject, SUBJ = subjunctive)

    Si haces eso le contaré a mis papás. (Mexico City Spanish)
    si haces eso le-contar-é a=mis=papás
    if do.2SG.SBJ;PRES that 3.IO-tell-1SG.SBJ;FUT to=my.PL=parents[M.PL].
    'If you do that I'll tell my parents.'
Moreover, in colloquial Spanish, the polypersonal agreement is also expanding to direct objects there and there, at least within certain limitations...
  • Ya los leí los libros. (San Salvador Spanish)
    ya los-le-í los=libros
    already 3SG.M.DO-read-1SG.SBJ;PRF the=books (DO = direct object)
    'I already read those books.'
    (I should note there's not even a pause or a break in intonation there.)

    Ayer lo traje las llaves. (Quito Spanish)
    ya lo-traj-e las=llaves
    yesterday 3.DO-bring.PRF-1SG.SBJ the=keys
    'I brought the keys yesterday.'
Legion wrote:I've said it before, but the main obstacle to analyse French as polysynthetic is that it doesn't satisfy the "morphological visibility constraint". Not only all arguments are not marked on the verbs, but in fact, modern colloquial European French has even a tendency to drop clitics that would be considered obligatory in literary French: "je vais le lui donner" often just becomes "je vais lui donner", with the direct object entirely unmarked, simply implied by the verb (donner = to give). In some instance this can go further to "vais lui donner", with dropping of the subject clitic as well (though in this case the verb does have a separate, audible inflection for 1sg, but that would work with a less marked verb: "viens d'lui donner" = "(I) just gave (it) to him/her").
I was thinking about this again... And are there any chances that in these cases it's not that the direct object affixes le and la are dropped, but rather, it's the indirect object affix lui that has lost its initial [l]? As in je lui ai dit ça > j'ui ai dit ça. So we basically have je vais l'ui donner, and not je vais lui donner?
When you get into imperatives, even literary French is not shy of things like "Donne !" for "Donne-le-moi !" ("give it to me!").
Also... What's the register of donne-le-moi? Is it really marked so high that donne ! is the most usual way to say "give it to me!" in colloquial French?
I don't know; at least in the case of "fais voir", the implication is generally "show me that thing", there's generally an implied object. Like "J'ai trouvé un lézard!" "Fais voir!" (I found a lizard!—Let me see it!)
Well, naturally, I'm not a native speaker in French, but this dialogue is also possible in my dialect of Spanish with the very same surface form (—¡Encontré una lagartija! —¡Dejá ver!), and I don't feel there's any trace (in the syntactical sense) in the verbs, even though there's an implicature (in the pragmatics sense) that I'm telling you to let me see something (the lizard), just like there's an implicature that I'm interested in it.

Examples like Marie l'a donné, à moi, along with the many examples in section 3 of the paper Whimemsz linked to where the verb can't agree with its arguments (e.g. relative clauses: le chèvrefeuille que maman t'a donné, not *que maman t'l'a donné), show that verbal agreement with objects in colloquial French is by no means absolute. But I can't help to suspect that here with fais voir ! you're comparing the construction with English syntax too much, where let me see (it)! requires those pronouns and *let see! is ungrammatical.
Last edited by Ser on Sun May 13, 2012 5:11 am, edited 3 times in total.

Post Reply