Re: Polysynthesis for Novices
Posted: Mon Mar 12, 2012 10:56 pm
I'd imagine a common way is to not allow incorporated nouns to take relative clauses.
Incorporated nouns are usually no longer available to be modified, so sentence 2 would not be expressed using an incorporated noun. Or rather, it wouldn't be if the speaker were trying to convey all that information with one sentence, But it I'm not actually don't know of any North American languages that have relative clauses in the first place. You're more likely to find that sort of information split up: "Mom gave me the milk. She bought it at the supermarket."dhokarena56 wrote:Question: if a nominal participant is incorporated, how do polysynthetic languages tag a relative clause onto it? I mean, if "Mom gave me milk" is expressed something like "Mom milk-me-gave," what about "Mom gave me the milk that she bought at the supermarket"?
How can you tell that it is a full NP that has been dropped and not a pronoun? In other words, what is the evidence for this?Whimemsz wrote: But secondly, they frequently drop entire NPs, not just pronouns. Even in mildly-configurational polysynthetic languages, like Ojibwe, it's very rare for both the subject and object of a verb to be expressed with NPs.
L&L doesn't prune topics (the only ones that do now are C&C Quickies and Ephemera), so the only way this could disappear is if a mod actively deleted it, which I believe a lot of people would screech at.Archaeopteryx wrote:What a great thread! It reminds me times, when I wanted to create a polysynthetic language myself... I hope it's not going to dissapear after some time (I remember that this was often the case with many threads before, at least while ZBB was on its former hosting).
I read the following book a while ago:Radius Solis wrote: That can be tricky for languages you do not speak, and to some extent the notion of word boundaries may not even apply fully to all languages. But for the most part is is generally possible to look at some of the behaviors that words display and usually they will line up with each other. The more that they don't, the more arbitrary the idea of word boundaries becomes.
Good question!merijn wrote:How can you tell that it is a full NP that has been dropped and not a pronoun? In other words, what is the evidence for this?Whimemsz wrote: But secondly, they frequently drop entire NPs, not just pronouns. Even in mildly-configurational polysynthetic languages, like Ojibwe, it's very rare for both the subject and object of a verb to be expressed with NPs.
Presumably, in cases where "milk-give" is a set lexical compound, you'd either use two separate independent clauses ("Mom milkgave me, she bought the milk at the supermarket); or the language might permit constructions like "Mom milkgave me from the supermarket" (I didn't mention this sort of this in my initial post on NI; I'm going to add it in at some point).dhokarena56 wrote:Question: if a nominal participant is incorporated, how do polysynthetic languages tag a relative clause onto it? I mean, if "Mom gave me milk" is expressed something like "Mom milk-me-gave," what about "Mom gave me the milk that she bought at the supermarket"?
Vohp said that the polysynthetic languages he studied lacked relative clauses. He claimed that least some of them drew no distinction between independent clauses, relative clauses, participles, or even nouns. A verb meaning "he works metal" could also mean "he who works metal", "metal-working (one)", or even just "metal-worker" without any morphological change. Does that sound like something you've encountered in the languages you've studied?Whimemsz wrote:Assuming the language even has relative clauses, which it doesn't necessarily need to.
You don't need polysynthesis to do that. Hebrew present-tense verbs can act as all four of those with no morphological change (because they're derived from participles in the first place).Jabechasqvi wrote:Vohp said that the polysynthetic languages he studied lacked relative clauses. He claimed that least some of them drew no distinction between independent clauses, relative clauses, participles, or even nouns. A verb meaning "he works metal" could also mean "he who works metal", "metal-working (one)", or even just "metal-worker" without any morphological change. Does that sound like something you've encountered in the languages you've studied?
Kind of? I only have intimate knowledge of one polylang, though (Ojibwe), and Jeff only had of two (Mohawk and Cheyenne), so you shouldn't draw any conclusions from that data. As Astraios says, non-polylangs can do the same thing.Jabechasqvi wrote:Vohp said that the polysynthetic languages he studied lacked relative clauses. He claimed that least some of them drew no distinction between independent clauses, relative clauses, participles, or even nouns. A verb meaning "he works metal" could also mean "he who works metal", "metal-working (one)", or even just "metal-worker" without any morphological change. Does that sound like something you've encountered in the languages you've studied?Whimemsz wrote:Assuming the language even has relative clauses, which it doesn't necessarily need to.
[28.01.2009 23:27:02] <vlad> Xephyr]: Relative clauses in Nahuatl are marked by in at the start of them.
[28.01.2009 23:27:19] <vlad> which is also used for some other kinds of subordinate clauses
[28.01.2009 23:27:47] <vlad> like nicnequi in nicquaz "I want to eat it" (lit. I-want-it that I-will-eat-it)
[28.01.2009 23:29:20] <vlad> it's ambiguous as to what the, thing, is.
[28.01.2009 23:29:36] <vlad> like it could be the subject or the object, you just have to know from context
[28.01.2009 23:29:50] <vlad> like Japanese
[28.01.2009 23:30:08] <vlad> although I think Japanese is even more flexible
[28.01.2009 23:30:32] <Xephyr]> vlad: Those are the most boring relative clauses ever.
[28.01.2009 23:31:24] <vlad> everything's a relative clause
[28.01.2009 23:31:29] <Xephyr]> Instead of "bird" it's "that which be's a bird"
[28.01.2009 23:31:32] <vlad> also uh
[28.01.2009 23:32:11] <vlad> eh nevermind
[28.01.2009 23:32:20] <vlad> it's not important
[28.01.2009 23:32:39] <vlad> what I think is important though
[28.01.2009 23:33:04] <vlad> is to distinguish between modifier-type relative clauses, and wossname relative clauses
[28.01.2009 23:33:13] <vlad> appositional ones
[28.01.2009 23:33:16] <Xephyr]> Oh
[28.01.2009 23:33:24] <Xephyr]> The ones that are necessary information vs the ones that aren't?
[28.01.2009 23:33:28] <vlad> also adjectives.
[28.01.2009 23:33:37] <vlad> no
[28.01.2009 23:33:50] <vlad> that's a different thing
[28.01.2009 23:34:04] <Xephyr]> VLad: Oh
[28.01.2009 23:34:08] <vlad> like in English most relative clauses modify some noun
[28.01.2009 23:34:11] <Xephyr]> What are modifier, wossname, and adjective type relcaluses?
[28.01.2009 23:34:16] <vlad> and can't stand on their own
[28.01.2009 23:34:28] <vlad> while in e.g. Latin, relative clauses are also noun phrases
[28.01.2009 23:34:39] <vlad> you can put them next to a noun if you want but that's the same as putting a noun next to a noun
[28.01.2009 23:34:53] <Xephyr]> Example?
[28.01.2009 23:35:01] <vlad> Nahuatl is like Latin, Japanese is more like English
One thing that Mithun argues later on, after discussing the Koryak whale-hunting and Mohawk examples, is that this kind of noun incorporation remains a lexical, not syntactic process for the Mohawk speakers. That is, when a person uses an incorporated noun for the first time, everyone is aware that they have coined a novel word, since they are aware of the whole inventory of such words. (If Baker argues that noun incorporation is a syntactic not lexical process, they I really need to read through him more thoroughly. But the criticism I've read of Baker is that he attempts to reduce polysynthesis to a single parameter, when it is better described as a confluence of traits. For example, it seems that there is a significant group of polysynthetic languages such as the Inuit group that do not synthesize multiple lexical roots into a word. Rather than making use of noun incorporation or verb root serialization, they instead have a very large palette of non-root bound morphemes that express adjectival or adverbial meanings. And there are a lot of languages that mix these approaches in various ways).What are some techniques for analyzing word boundaries? How do you tell the difference between a polysynthetic language and a less-synthetic language written without spaces?
Code: Select all
kina a-e rusuy
herbs 1SG-eat want
"I want to eat herbs"
kina-e-rusuy-an
herbs-eat-want-1SG
Code: Select all
mukacaraha a-tuye
his chest 1SG-cut
"I cut his chest"
a-mukcar-tuye
1SG-chest-cut
Code: Select all
UNINCORPORATED COMPLEMENT:
kih ʔi:y-aš-ka
ceremonial_brother RECIP-make-PST
INCORPORATED COMPLEMENT:
ʔi:-kih ʔaš-ka
RECIP-ceremonial_brother make-PST
"they made each other ceremonial brothers
UNINCORPORATED SUBJECT:
ʔas ʔa:w-allu-kka
hand PL.S-move_about-PST
INCORPORATED SUBJECT:
ʔa:w-as ʔa:w-allu-kka
PL.S-hand PL.S-move_about-PST
"he felt about with his hands" (lit. "the hand[s] moved about")
You can find incorporation vs. non-incorporation by changing the stress of the phrase, as in Lakota, where stress can only be on the first or second syllable of a word:Trailsend wrote:What are some techniques for analyzing word boundaries? How do you tell the difference between a polysynthetic language and a less-synthetic language written without spaces?
Nice. Too bad it's behind a paywall. Most people don't have access to JSTOR.Gojera wrote:Aha, neat. I spent the last week on a big writeup on polysynthesis that I was going to post, but you've covered most of what I had and more.
Definitely the two most interesting papers on this were Marianne Mithun's 1984 paper "The evolution of noun incorporation", which is available in JSTOR, and Johanna Mattissen's 2004 paper "A structural typology of polysynthesis", is only on actual paper. Everybody should read that Mithun paper, it's great.
Another step would be "noun stripping". For example, in Basque there are lexicalised noun + verb combinations where the noun is still a separate word, but it cannot take any case suffixes or modifiers. The noun is clearly a separate word despite this because it need not be next to the main verb.Xephyr wrote:As an addendum to your post on noun incorporation, I'd like to add that sometimes, the noun in question may be grammatically incorporated into the verb complex, but phonologically remains an independent word. An example from Zuni:
The prefixes ʔi:- and ʔa:w- are verbal prefixes that don't normally appear before nouns, indicating that kih and ʔas are grammatically part of the verbal complex. However, each is stressed as a separate word (stress is initial in Zuni): ʔi:kih, ʔa:was.Code: Select all
UNINCORPORATED COMPLEMENT: kih ʔi:y-aš-ka ceremonial_brother RECIP-make-PST INCORPORATED COMPLEMENT: ʔi:-kih ʔaš-ka RECIP-ceremonial_brother make-PST "they made each other ceremonial brothers UNINCORPORATED SUBJECT: ʔas ʔa:w-allu-kka hand PL.S-move_about-PST INCORPORATED SUBJECT: ʔa:w-as ʔa:w-allu-kka PL.S-hand PL.S-move_about-PST "he felt about with his hands" (lit. "the hand[s] moved about")
Maybe this represents a transitional stage, between isolation and full incorporation?
Source: Newman, Stanley. Zuni Grammar. 1965.
FWIW English incorporation could be called stage-3 as well; there's no real phonological merging going on in "deer-hunting" or "doctor-recommended" - there is a full stress in each word - but in order to use that construction the noun must be to the immediate left of the verb and it cannot generally take any marking of its own. One the other hand we do have a handful of cases that are more like stage 4 at least insofar as the resulting word has only one full stress, such as "babysit".chris_notts wrote: Another step would be "noun stripping". For example, in Basque there are lexicalised noun + verb combinations where the noun is still a separate word, but it cannot take any case suffixes or modifiers. The noun is clearly a separate word despite this because it need not be next to the main verb.
So I would guess it might go something like this:
full argument > loss of ability to accept modifiers / role marking > loss of grammatical independence (must occur next to verb) > loss of phonological independence (part of the same phonological word as the verb stem)
Basque is in stage (2) of this process, Zuni is in stage (3), and typical incorporating languages are in stage (4).
Well. It's significant that most of our NI examples are not easily employed as finite verbs, I agree with that - but for the record, along with our substantivized NI verbs there are also plenty that serve mainly as modifiers: a peer-reviewed journal, a time-tested recipe, a cherry-picked example, a steam-driven engine, a cheese-eating liberal, a Swahili-speaking community. So Whim was quite right to point out that most of them are restricted to being used with derivational morphology, just not that they are marginal - at least, not marginal in the sense of being rare or completely unproductive. Some can appear with more than one sort of affix: deer-hunting / deer hunter, window-washing / window washer, cherry-picking / cherry-picked.ná'oolkiłí wrote:It's significant, though, that all but one of your examples are substantivized.