zompist bboard

Ahh, that's unfortunate (for my purposes). There must be a polysynthetic natlang out there with lots of preverbs and a ZBB poster or a linguistics paper that can give tons of examples.



I'd like to see morpheme-by-morpheme glosses, if you're able to provide such.

Also, how would Lakota handle a sentence like this?

c. Alex, a mutual friend, brought the package to me for you. He wanted to save you the trouble.

The main verbs in the glosses aren't especially exciting:

itȟáŋčhaŋ=kiŋ | tȟa-wo-wa-ší=kiŋ oyás'iŋ | wa-kȟatÁ-yA-pi waŋží~RED=la | opȟé<wičha-kiča>tȟuŋ='
chief=DEF | 3.POSS-NMLZ-ABST-command=DEF all | ABST-hot-CAU-NMLZ one~DIST=DIM | buy<3P.ANI-DAT>=DECL

wa-uŋspé-wičha-khiyA=kiŋ | tȟa-wa-yawá=kiŋ hená | itȟáŋčhaŋ=kiŋ etáŋ | wa-wičha-kiči-lá='
ABST-learn-3P.ANI-CAU=DEF | 3.POSS-ABST-read=DEF DIST.PL | chief=DEF from | ABST-3P.ANI-BEN-request=DECL

ABST = 'abstract' indefinite patient, devalentizer; lá X requests Y; walá X makes a request.


And c. uses an adverb too, you can't mark both dative 'to me' and benefactive 'on your behalf' both at once:

Alex okȟólauŋye kiŋ | iyéš čha | nitȟáŋtaŋhaŋ | wópȟaȟte kiŋ | makáhi. | Aníkpataŋ kta čha | íničiksape.
"Our friend Alex | he was the one who | on your behalf | the package | he brought it to me. | In order that you spare yourself | he took the trouble on your behalf."

o-kȟolá-uŋ-yA=kiŋ | iyé-š=čha | ni-tȟáŋtaŋhaŋ | wo-pȟaȟtÁ=kiŋ | ma-ki-a-hí=' | a-ni-k-pataŋ=ktA=čha | í<ni-kiči>ksapA='
NMLZ-friend-1DU-CAU=DEF | 3.COP-ADVS=SUB | 2.POSS-behalf | NMLZ-bundle=DEF | 1SG-DAT-bring-come.PRF=DECL | APP-2-REF-hold.back=IRR=SUB | take.trouble<2-BEN>=DECL

There's only a very fine line between dative and benefactive, cf.:

kahí (dative)
ki-a-hí
DAT-bringing-come.PRF
X has come bringing Y to Z

kíčahi (dative)
ki-[kahí]
DAT-[bring.PRF]
X has come bringing Z's Y to Z (without Z's knowledge/permission)

kíčičahi (benefactive)
kiči-[kahí]
BEN-[bring.PRF]
X has come bringing Z's Y to Z (as a favour to Z)

, but there's no such verb as X has come bringing Y to Z on A's behalf. The only verb I can think of that allows four separate arguments in one go is kíčič'u X gives A's Y to Z:

wičhámiyečič'u
wičha-mi-ya-kiči-k'ú='
3P.ANI-1SG.DAT-2.ACT-BEN-give=DECL
"you gave my Y to them"

, where the DAT argument works sort of like a malefactive (you gave Y to them 'to my detriment'), but it's quite a unique example.

I forgot this thread where I had some unaddressed questions.

Regarding Dyirbal and australian dependentmarking polysynthesis this view is advanced by Nicholas Evans and Hans-Jürgen Sasse in a 2002 edited volume called "Problems of Polysynthesis". Thety are argueing for a definition of polysynthesis that is closer to the classical definition where it describes languages that have high morpheme to word ratios and "sentence words". This would mean that languages with heavy case stacking could be considered dependent marking and polysynthetic at the same time. It would also mean that languages like Esimo-Aleut and Athabascan could be considered polysynthetic. I think the problem with this definition is that it becomes very broad almost so as to become meaningless, but on the other hand it also does show that word length itself is not simply related to headmarking syntax. It also shows that we probably ought not to talk about polysynthesis but about different kinds of polysynthesis that seem much more likely to be predictive of other grammatical features.

What do you think would be good minimally biassed sources of glossed texts in a large variety of different languages to check if that classification has any statistical merit to it? I'd be interested to see or do a study of the cross linguistic distributions of the mean morpheme to word ratio and the proportion of "single word sentences" (let's say sentences consisting of a single finite predicate word) to all sentences in a corpus.

My initial guess is that studying these two distributions won't reveal any distinct populations among the languages and we'd just see monotonously decreasing tails towards high morpheme to word and "sentence word" ratios. Still, this doesn't mean that a traditional definition for "polysynthesis" is useless despite not perhaps being corresponding to any natural distinction demonstrable from actual data. Even such vague definitions describing some part of a continuum can be useful for giving people a quick first impression of the type of language they are discussing. On the other hand, if the study would reveal any natural populations, we'd immediately know where to draw the border between "poly" and "non-poly" languages according to a classical definition.

Maybe an easier task to start with would be to study word frequencies in the sample languages. In more highly inflected languages word shapes vary a lot more and identical words appear in much lower frequency than in mildly inflected languages. This kind of study would need a strict orthography for the whole corpus of a single language, which might be too much to ask for some smaller languages without a written tradition, but wouldn't need any preanalysed and already glossed text material unlike the first two study methods.

The first question about the possible text corpora I dont know how to answer well. Since I mostly work with languages that are underdocumented I dont know which languages in the world really have good corpora. What I would do would be to use reference grammars of individual languages which tend to have a small corpus of texts presented in a glossed and a systematically represented and analyzed way. These are very small corpora however, and may not yield sufficient material. On the other hand if you are going for comparative breadth rather than depth then maybe that is OK.

Now, would it yield distinct populations? That is an interesting question that can ultimately only be answered empirically. The trouble would be making operational definitions that accurately represent the morpheme and meaning structure of each language. For example it is possible to have sentence words without having many morphemes. And often it is a question of analysis whether morphemes are clitics, relatively free function words or bound morphemes (the by now classic example of this problem is French). I for example consider Otomi to be polysynthetic but most experts don't (they consider time and subject to be expressed with proclitics which I consider prefixes). In so far as the example were to yield distinct populations I think that would be much more likely to be a result of historical process language spread and diffusion than to something inherent in linguistic typology - so the question is what we would learn from that fact.

Those aren't usually too long texts at all but might still be enough to get a stable estimate of the mean or median morpheme to word estimate within the language. Comparative breadth is after all more important to get a well populated sample distribution than the precision of individual data points taken to construct it.



Correlation between the typologies of languages within either closely related or areally connected groups is a problem as is sampling in general in typological studies. If you sample the languages sparsely enough such individual phenomena of spreading features should cancel out, but then you run into the problem of gathering enough data to actually make any conclusions at all. Another study philosophy would be to just gather a mass of data, merely making sure that it's sampled in a consistent way, do the analysis and only after that start to find out what findings are due to biases and what are maybe more underlying.

Can you recommend any papers on how polysynthetic languages deal with quantifiers?

Not really, the only one I know who has worked on that is Mark Baker who uses it as one of his traits that they dont have true quantifiers. I dont think anyone else has paid any attention to quantifiers (partly because if ones definition of polysynthesis is purely morphological as opposed to syntactic then there is no reason to think there would be any particular relation between quantification and polysynthesis)

OK. Well, I don't currently have access to university-accessible books/articles anyway. My previous googling hasn't really helped. Could you explain this phenomenon of "no true quantifiers" for everyone interested in polysynthetic tongues? (I've been working on such a conlang of this sort.)

Edit: Extra material added.

A "true quantifier" in the sense used by Baker is a quantifier that acts as the head of a noun phrase and "governs" the noun (e.g. forcing it to agree with the quantifier). Using generative jargon Baker defines it as quantifiers that C command a pronoun that it coindexes, shows weak crossover effects and triggers singular agreement on the verb. In English these are quantifiers like each and every and no, but not all, two, many.


Here are two papers about the phenomenon, the one about Navajo by Yazzie and Speas is particularly good (they argue against Baker):

http://link.springer.com/chapter/10.1007/978-94-017-2817-1_3#page-1
http://books.google.com/books?hl=en&lr=&id=G6J7Tki_WBcC&oi=fnd&pg=PA35&dq=Mark+Baker+true+quantifiers+polysynthetic&ots=Oi-5eFwxD3&sig=MaVQhhdKpFUqS1zsnc1TqTnXDdY#v=onepage&q=Mark%20Baker%20true%20quantifiers%20polysynthetic&f=false

Thanks for the info.

Unfortunately, I'm not eligible to read Springer material, and my screenreading program can't handle Google Books content as it appears in images. Might you have any other recommendations?

Hmm, if you pm me your email I can send you the springer article by baker. The other article is really better but it is probably not possible to get that as something your screen reader can handle, maybe through your local library you can get it if you have some way to read physical books. If at some point I get time maybe I could read it and send it to you as an mp3. But that would be a little while before Icould do that.

Radagast Revived--I sent you a PM.

Sorry if this has been answered somewhere and I missed it (so many pages!).

Let's take a transitive verb such as from dhok's question from almost a decade ago: "My mom gave me milk". If there's a verb milk-give and you can say a thing such as my-mom she-milk-me-gave, is the verb still transitive or does the incorporation turn it intransitive? Could you say something like my-mom she-milk-me-gave skim milk?

_________________
vec

Is the morpheme order on the English gloss intentional? I'm not aware of any polysynthetic lang that has verbobj incorporation but allows a person marker to split the two morphemes up again.

In my conlangs, "serve milk " would behave like any other verb, being transitive when the object is specified and intransitive otherwise. The incorporated object is indefinite, so it doesn't affect transitivity. E.g. like the English verb "hunt" , though one always hunts for something, when used alone is intransitive.

_________________
Sunàqʷa the Sea Lamprey says:

@Soap: Is the conlang in your signature polysynthetic? The word lamempambo appears to incorporate two nouns along with the subject into one free-standing word.

Perhaps another way to word waht I wrote:

The transitivity is determined by the person marking of the verb. The incorporated object is analogous to an English adverb, not an English (in-)direct object. With a content verb like "hunt" or "fish", one can say in English "I hunted for deer" where for deer is an adverbial phrase. This is our best analog to the incorporated nouns of polysynthetic languages. But in "I hunted for deer", the verb hunted is intransitive because there is no specified object in the sentence.

I used a different example because "give" is an open-ended verb and I wouldnt expect many languages to have incorporated forms like that since almost any noun could go there. Whereas with hunt, I could see a language having a specific word for hunting deer, hutning rabbits, and hunting birds, but nothing else.


Yes. But the gloss isn't quite what you'd expect ... lamempa by itself means "ice cream in a bowl", from lamem "ice cream" (itself a compound, of course) plus -pa "in a container". Therefore, lamempa behaves as a single noun and it is only one noun being incorporated. To be honest, it looks like a narrower translation of that verb would be "I want ice cream in a bowl", without "my", since that would require a separate morpheme, but I think I was translating loosely on purpose. (I havent actually worked with Poswa for a while so I had to look up the root.)

_________________
Sunàqʷa the Sea Lamprey says:

In many (most?) languages, give is ditransitive, so incorporating the theme would typically result in a verb that's still transitive. So in my-mom she-me-milk-gave, me would likely count as the object of the verb (even if it's only referenced by a pronominal marker on the verb and not by a free pronoun).

Also, several polysynthetic languages do in fact allow something like my-mom she-me-milk-gave skim milk, where the incorporated milk acts more or less like a classifier. In Marianne Mithun's typology of noun incorporation this is type IV; "classificatory incorporation". Compare this example from Onondaga (Barrie 2015, "Two kinds of structural noun incorporation", Studia Linguistica 69/3):

waˀgnasgwahní:nǫˀ neˀ gwíhsgwihs
waˀ-k-naskw-a-hninǫ-ˀ neˀ kwihskwihs
FACT-1.SG.AG-animal-EPEN-buy-PUNC NE pig
‘I bought a pig.’
(lit. ‘I animal-bought a pig.’)

_________________
Blog: audmanh.wordpress.com
Conlangs: Ronc Tyu | Buruya Nzaysa | Doayâu | Tmaśareʔ

So one way is treating the incorporated noun like an adverbial, one like a classifier.





Is there a functional difference between the two?

_________________
vec

viewtopic.php?p=955528#p955528

in the first page of this thread is an answer that i think answers the question there.

a language that has classifier based incorporation must also have arbitrary incorporation. that is, if you can say "hunt for animals", you can also make arbitrary incorporation like the above. Therefore, classifier-based incorporation is a subset of arbitrary based incorporation, not the other way around. there are exceptions butt i suspect they are not polysyntheticlanguages, but rather languages that have classifiers in general such as Swahili. where verbs are classified according to concord with the subject if the phrase.

_________________
Sunàqʷa the Sea Lamprey says: