Polysynthetic Conlang

DF · Post by DF » Sat Mar 06, 2004 10:59 pm

I might as well post this here...this is my quick attempt at a polylang:

hkoamem?hka:htheahko
hk-o-a-me-m?hka-aht-he-a-hk-o
Masc.-Anim.-Sing.-3S-to go-(direction)away-1S-Sing.-Masc.-Anim.
He is going away from me. (He is leaving.)

Is that plausible for a polylang?

Soap · Post by **Soap** » Sat Mar 06, 2004 11:02 pm

Two questions:

1) why are the person markers (head markers?) at the end of the word aren't in the same order as the ones at the beginning?

2) I don't know about "hk" at the beginning of a word, unless it's something other than /hk/. Also I would find "hth" hard to pronounce, if the h's both mark aspiration.

vohpenonomae · Post by **vohpenonomae** » Sat Mar 06, 2004 11:08 pm

DarkFantasy wrote:I might as well post this here...this is my quick attempt at a polylang:

hkoamem?hka:htheahko
hk-o-a-me-m?hka-aht-he-a-hk-o
Masc.-Anim.-Sing.-3S-to go-(direction)away-1S-Sing.-Masc.-Anim.
He is going away from me. (He is leaving.)

Is that plausible for a polylang?

I suppose, but you wouldn't need to overtly mark first person as animate; that much is assumed of anything capable of producing (or understanding) speech.

BTW, what are the h+stops? Preaspirates, or {h} clusters?

DF · Post by DF » Sat Mar 06, 2004 11:10 pm

Mercator wrote:1) why are the person markers (head markers?) at the end of the word aren't in the same order as the ones at the beginning?

It has different orders for the suffixes and prefixes. The prefixes mark the subject; the suffixes mark the object. Here are the orders:

Prefix order: gender-animacy-number-pronominal-OTHER
Suffix order: OTHER-pronominal-number-gender-animacy

2) I don't know about "hk" at the beginning of a word, unless it's something other than /hk/. Also I would find "hth" hard to pronounce, if the h's both mark aspiration.

The <h> before plosives shows that the consanant is preaspirated. Anywhere else it is a voiceless glottal fricative. Also, the <hth> is a preaspirated /t/ + /h/ consanant cluster.

Aurora Rossa · Post by **Aurora Rossa** » Tue May 25, 2004 5:12 pm

Gkx'o?t?k!o?s? bv?duk???tsa?.
Gkx'o?t-?k!-o?s-? bv?[s]d[/s]-uk??-?ts-a?.
Read-past-sing.imm.ob-1sing write-passive-hab.asp-sing.imm.sub
I was reading a book.

How do you like this sample? I am having trouble finding a way to represent breathy vowels.

Whimemsz · Post by **Whimemsz** » Tue May 25, 2004 5:55 pm

Now that the thread has been reopened, there's something I've wanted to ask for a while. There's a problem I keep running into, which I can't find the solution to.

Say, for example, that the word for "tree" is he-grows-toward-(the)-sky. And let's say that's Ma-legam-et-kasis. Now, say "sky" itself is an incorporated noun root for ait-katsi-lis, or he-hugs-(the)-earth. BUT, now, "earth" is ALSO an incorporated noun root for elitoris, and so on ad infinitum. How do you avoid this problem? How do you make all 'nouns' in the language actually be verbs, without having this never-ending descent through more and more incorporated roots?

And, while I'm thinking about it, what sorts of methods do various languages use to make a 'noun' into a root for incorporation?

Thanks.

Aurora Rossa · Post by **Aurora Rossa** » Tue May 25, 2004 6:03 pm

Say, for example, that the word for "tree" is he-grows-toward-(the)-sky. And let's say that's Ma-legam-et-kasis. Now, say "sky" itself is an incorporated noun root for ait-katsi-lis, or he-hugs-(the)-earth. BUT, now, "earth" is ALSO an incorporated noun root for elitoris, and so on ad infinitum. How do you avoid this problem? How do you make all 'nouns' in the language actually be verbs, without having this never-ending descent through more and more incorporated roots?

You could always eliminate it on a case-by-case basis and change the words for tree, sky, etc. Bp@x????kx? doesn't have a whole lot of incorporation compared to many polylangs. Miami doesn't incorporate nouns at all, I heard.

Salmoneus · Post by **Salmoneus** » Tue May 25, 2004 6:09 pm

I can think of two ways to avoid this problem:

1.) Have some "axiomatic" noun roots on which to build - noun roots that just are, and don't mean anything other than what they mean.

2.) Build from nouns defined intransitively, or transitively with respect to other verbs, not nouns. "It-squawks-loudly" or "he-likes-running", for instance. In your example, you could make "earth" be simply "she-is".

Soap · Post by **Soap** » Tue May 25, 2004 6:26 pm

I think Sal has pretty much hit on the answer, for how verb-based polsynthetic langs such as Noyahtowa do it.
It'd be nice to see a language based on recursive incorporation though. Such that "tree" could be built up of "grows-from-the-earth" and "earth" could be built up of "holds-up-trees".
Or you could say that your language always uses incorporated roots, but in reality just use nonsense words for when you get to a word for which you can't think of a noun to incorporate.

vohpenonomae · Post by **vohpenonomae** » Tue May 25, 2004 7:36 pm

Eddy the Great wrote:Miami doesn't incorporate nouns at all, I heard.

Not sure who told you that, but Miami, like PA and all its daughters, uses the Algonquian medial, which is usually incorporated noun information. Noun incorporation is very much a part of Miami morphology.

Aurora Rossa · Post by **Aurora Rossa** » Wed May 26, 2004 3:02 pm

Not sure who told you that, but Miami, like PA and all its daughters, uses the Algonquian medial, which is usually incorporated noun information. Noun incorporation is very much a part of Miami morphology.

"This is a mininal requirement to be considered polysynthetic (some people think object incorporation is key, but it's not; Miami does not allow object incorporation, but is poly just the same)."

Whoops, misremembered.

doctrellor · Post by **doctrellor** » Tue Sep 21, 2004 10:25 pm

As far as Noun incorporation in a Poly lang

Lemme see if I get it

1. I-k'uru-k'euwe-m.
agr-dipper-old-pres
The dipper is old.

2. We-fan-lur-mi.
neg-snow-fall-pres
The snow isn't falling.

3. *0-khwien-teurawe-we.
agr-dog-fun-pres
Intended reading: The dog is running.
Onondaga

4. *H-ate-tsi?kti-?se:-?.
agr-refl-louse-drag-asp
Intended reading: The louse crawls.

a. Ta-'u'u-wia-ban hliawra-de
1s:A/A-baby-give-PAST woman-SUF
I gave the baby to the woman.

b. *Ta-hliawra-wia-ban
1s:A/A-woman-him-give-PAST
I gave him to the woman.

c. *Ta-hliawra-'u'u-wia-ban
1s:A/A-woman-baby-give-PAST
I gave the woman the baby.

Like this? where the noun-verb complex is seen?

doctrellor · Post by **doctrellor** » Tue Sep 21, 2004 10:34 pm

But then I am confused by this...

a. Yede seuan-ide a-mū-ban
that man-SUF 2sS-see-PAST
You saw that man.

b. Yede a-seuean-mū-ban
that 2sS:A-man-see-PAST
You saw that man.

with noun incorporation, I can see b. but what about a? but
nouns can't incorporate when...

a) subject of a transitive clause
b) subject of verbs like "run", "jump", etc
c) from PP
d) from an adjunct NP

I thought Poly langs ALWAYS incorporated?

Glenn · Post by **Glenn** » Tue Sep 21, 2004 11:42 pm

doctrellor wrote:I thought Poly langs ALWAYS incorporated?

Not at all--or at least, not necessarily; according to what Jeff has posted and what I've read, most (maybe all) polysynthetic natlangs also permit free-standing subjects and/or objects--for example, in the form of topical phrases.

Jeff once gave the following example of topic-comment structure in a polylang:

"A-banana, I-ate-it." (English: "I ate a banana.")

Here, the noun "banana" is free-standing; however, it is also reflected in the subject and object markers on the verb. (I suspect that sentence (a) above operates similarly: "That man, you-saw-him.") If the noun were then incorporated, a single construction (I-ate-a-banana) would result--but it's not necessarily mandatory.

I've seen at least one polylang grammar (can't remember the language, unfortunately

) that gave pairs of examples like the one above, showing both incorporated and non-incorporated nouns.

[Side note: I'm pretty sure that I've also seen examples (conlang or natlang?) of sentences where the object noun was incorporated, but the subject noun was not; Jeff told me once that while this construction can be found in some languages, the opposite--subject noun incorporated, object noun separate--generally cannot, since the object here becomes part of the verb stem. That's just going by vague memory, however...]

Then there are languages that I think of as "semi-polylangs"

, like Klingon, which is often called a polylang. In Klingon, the subject and object are always both marked on the verb (poly-style headmarking), and subject and object pronouns are generally incorporated, but subject and object nouns are not.

p@,
Glenn

Glenn · Post by **Glenn** » Tue Sep 21, 2004 11:49 pm

Addition: while writing that last post, I found the example sentence above; it's part of Jeff's capsule definition of polysynthism, which was as follows:

Jeff Burke wrote:A while back, when we were planning a newbie FAQ for the board here, I wrote a description of polysynthism that you may find helpful; it mentions that one of the key feature of poly languages is the marking of both subject and object within the verb. This is a mininal requirement to be considered polysynthetic (some people think object incorporation is key, but it's not; Miami does not allow object incorporation, but is poly just the same).

Here is that description:

Polysynthesis is a grammatical phenomenon most famously present in many North American Indian languages; the Algonquian and Iroquoian tongues are the most well-known. These languages, instead of using case endings on nouns and conjugatational endings on verbs, mark both subject and object inside the verb; further, subject and object inflections are considered the actual arguments of the verb, while any free-standing nouns associated with a verb are like dislocated topic phrases. Here is an example from Mohawk, an Iroquoian language spoken in upstate New York and southern Ontario:

Teiotahi?:kton w?:keke'

Nominalized and rendered in normal English, this is 'I-ate a-banana'; but a more faithful translation would be 'A-banana, I-ate-it'. The pronominals 'I' and 'it' (the arguments of the verb root -k- 'eat') are expressed via the prefix ke-, which indicates a singular first-person subject and a singular third-person object. ke- is, then, akin to noun case and verb conjugation fused into a single marker and attached to the verb. Mohawk has some fifty-odd of these pronminal prefixes that express subject-object combinations; and a prefix is mandatory on every verb. By contrast, the noun teiotahi?:kton 'banana' stands in relation to the verb w?:keke' only indirectly, as a topical phrase stands in relation to a central clause in English.

Cheyenne, an Algonquian language spoken in southern Montana and Oklahoma, also marks subject and object on the verb, but does so differently; instead of a fusional marker, it uses a prefix for subject and a suffix for object. Witness:

Hetane n?v??mo

Faithfully translated, this means 'A-man, I-saw-him'. Hetane is the noun 'man' standing in relation to the verb n?v??mo 'I-saw-him' like a topical phrase. N?- is the prefix for a singular first-person subject; -o is the marker for a singular third-person object.

In addition to marking both subject and object inside the verb, polysynthetic languages often make use of noun incorporation. This is a phenomenon where a noun becomes, in essence, an inflection on a verb. Incorporated nouns can be, depending on the language and circumstances, direct objects, indirect objects, subjects, locatives, and
of other kinds. Mohawk makes heavy use of incorporation, as in:

Watia?tawi?tsher?:io

This translates as 'It-is-a-good-shirt', where the noun root atia?tawi (a word that can be used to refer to basically any upper-body garment) is present inside the verb. Cheyenne also uses noun incorporation on a regular basis:

N?tahpe'emaheona

This verb, meaning 'I-have-a-big-house', contains the noun morpheme maheo 'house'.

Noun incorporation can look bewildering, but is best understood in terms of English compounding. Babysit is an example of a word with a incorporated direct object; and words like deer-hunter and manslaughter make use of the same kind of incorporation.

As you can see, noun incorporation is often a feature of poly languages, but it is not a feature of all of them (see Jeff's comment on Miami above). I believe that different polylangs differ regarding the extent of what can be incorporated into the verb, and how frequently incorporation is used (for instance, Mohawk can incorporate virtually anything, and often does (although not always--see Jeff's topic-comment example above), while, i.e., Miami is evidently more restrictive. Of course, in the case of a "verby" polylang, even these "independent" nouns might be verb phrases in and of themselves.)

p@,
Glenn

doctrellor · Post by **doctrellor** » Wed Sep 22, 2004 8:47 am

ok, cool. So it's a standard feature but not always, and even a language can have both...cause I heard that Poly lang syntax is minimal.. but with this, it makes syntax more complicated that I thought...

Thanks

chris_notts · Post by **chris_notts** » Fri Dec 17, 2004 10:11 am

jburke wrote:
Eddy the Great wrote:3rd person is default. I'll make suffixs for 1st, 2nd, and 4th person.
Diachronically speaking, first person would be more likely to be default; but no language I know of has any person as default.

In certain circumstances (ie ergative agreement) Basque has null agreement for 3rd person singular only. For example:

ikus-ten du
he sees it
ikus-ten du-t
I see it
ikus-ten du-gu
we see it
etc

As you can see, the marking for 3rd person ergative is null, but other person/number combinations require a non-null marker to be added.

vohpenonomae · Post by **vohpenonomae** » Fri Dec 17, 2004 3:35 pm

chris_notts wrote:
jburke wrote:
Eddy the Great wrote:3rd person is default. I'll make suffixs for 1st, 2nd, and 4th person.
Diachronically speaking, first person would be more likely to be default; but no language I know of has any person as default.
In certain circumstances (ie ergative agreement) Basque has null agreement for 3rd person singular only. For example:

ikus-ten du
he sees it
ikus-ten du-t
I see it
ikus-ten du-gu
we see it
etc

As you can see, the marking for 3rd person ergative is null, but other person/number combinations require a non-null marker to be added.

I should qualify what I said there; no modern language that I know has a default person. Proto Algonquian had third person as default.

valinta · Post by **valinta** » Thu Dec 23, 2004 11:19 pm

I just decided to start playing with poly languages, and I have a sentence for everyone to sample. I'm thinking about how to handle genitives/posession, and one idea is included here.

Tacu ironisuiałuîandiŋata, nimaciusuiata.

When I began to live in the city, my life improved.

Tacu iro-ni-suia-łuî-andi-ŋa-ta, ni-maciu-suia-ta.

Time inceptive-1st.sing.animate.subj.-live-3rd.sing.inanimate.obj.-city-static.locative.in*-past, 1st.sing.animate.subj.-improve-life-past.

First of all, I'm handling clauses like Japanese: place a modifying phrase directly next to a noun as if it were an adjective.

Are there any features that I'm implementing wrong or missing?

And the main question: The method of indicating posession that I'm using in the intransitive phrase after the comma is to match the posessor's subject marker with the thing posessed, which is the incorporated subject after the verb. Something feels wrong about that, though, and I don't think I could use it universally. Any other ideas?

vohpenonomae · Post by **vohpenonomae** » Fri Dec 24, 2004 12:20 am

First of all, I'm handling clauses like Japanese: place a modifying phrase directly next to a noun as if it were an adjective.

Don't think in terms of phrases or clauses; it'll mess you up. Think in terms of a basic idea (a "root") being expounded upon by attached afterthoughts; e.g., I don't really see a basic "root" morpheme in the first word. What morpheme is that word being built around?

And the main question: The method of indicating posession that I'm using in the intransitive phrase after the comma is to match the posessor's subject marker with the thing posessed, which is the incorporated subject after the verb. Something feels wrong about that, though, and I don't think I could use it universally. Any other ideas?

Attach possessives to the thing possessed; and have the possessives give the pronominal details of the possessor.

Post by **zompist** » Fri Feb 04, 2005 11:20 am

Thanks to Ahribar and Trebor for working hard and quickly to reduce this topic from a 75-page monster that made my heart sink every time I looked at it, to a sleek beast which should be an excellent permanent resource. (That's why it's in L&L, so it won't get pruned.) They shall be known henceforth, polysynthetically, as They-have-sifted-what-those-others-accumulated.

Try not to spam the thread. If you have simple questions or want to show off your own conlang, start a new topic. I'd like anything added here to be of similar archival quality.

Trebor · Post by **Trebor** » Fri Feb 04, 2005 7:13 pm

ok, i understand what a person hierarchy is, but how could one develop in a language?

vohpenonomae · Post by **vohpenonomae** » Fri Feb 04, 2005 8:18 pm

Trebor wrote:ok, i understand what a person hierarchy is, but how could one develop in a language?

In Algic/Algonquian, Alford theorized that it was due to the cultural practice of deferring to others, i.e., to a second person. The Algonquian hierarchy always places second person at the top.

doctrellor · Post by **doctrellor** » Wed Apr 27, 2005 10:03 am

jsburke wrote:
Trebor wrote:ok, i understand what a person hierarchy is, but how could one develop in a language?
In Algic/Algonquian, Alford theorized that it was due to the cultural practice of deferring to others, i.e., to a second person. The Algonquian hierarchy always places second person at the top.

In Dakota and Cheyenne, the 'place of honor' for a tepee or sweat lodge is opposite the entrance. I am not sure if this is common practice in Woodland cultures (Objibwe), but it seems widespread in Plains cultures.

blank stare II · Post by **blank stare II** » Sat Jun 25, 2005 5:24 pm

jburke wrote:
Kak'ula'ikemaa.
Ka-k'ula-'ike-ma-a
Imperf-building-make-I
I have been making a building.

Let's say you add -sa to the end of this word (making it Kak'ula'ikemaasa), indicating that the object (building) is singular third person. Or, let's say that 'building' is masculaine, -sa could indicate that the object is singular, third person and masuline all three.

Even though the building is the object of the sentence, does it have to be marked as third person singular? Isn't that already established?