Aleut morphosyntax: Cool shit, or COOLEST shit?

[Xephyr]

There aren’t enough topics about Aleut morphosyntax floating around, and that just will not stand, goddammit! So I decided to make one, in order to spread the gospel of Aleut and its awesomeness.

(My sources are “The efficacy of anaphoricity in Aleut” by Jerrold M. Sadock in Variations on Polysynthesis: The Eskaleut Languages (which gave me the idea) and Aleut Grammar: Unangam Tunuganaan Achixaasix̂ by Knut Bergsland, with (hopefully accurate) Yup’ik examples by myself, going off of A Practical Grammar of the Central Alaskan Yup’ik Eskimo Language by Steven A. Jacobson.)

The thing with Aleut morphosyntax is that it uses a “non-anaphoric” vs. “anaphoric” distinction in its marking of both nouns and verbs. Anaphoric marking functions like the possessed suffixes in the other Eskimo-Aleut languages, only a bit more stripped down.. though more versatile! Lemme show you what I mean.

First, let’s look at Yup’ik. Yup’ik marks all its nouns for case and number (sg, du, pl) as well as for person and number (again: sg, du, & pl) of its possessor—whether overt (in which case it appears in the “relative” (ergative) case, or covert—in a bewilderingly abstruse matrix of endings that are enough to make any learner yearn for the more intuitive, tractable paradigms of oh, Sanskrit. But enough about my problems. Here’s some examples:

Code: Select all

Yup'ik
qetunraa-nka
son     -pl.1sgPOSS
“my sons”

qetunrar-put *
son     -sg.1plPOSS
“our son”

tan’gurra-m   aata  -a
boy      -REL father-sg.3sgPOSS
“the boy’s father”

tan’gurra-at  aata  -at
boy      -pl  father-sg.3plPOSS
“the boys’ father”

angute-m       qetunra-i
man   -REL.sg  son    -pl.3sgPOSS
“the man’s sons”

angute-t  qetunra-it
man   -pl son    -pl.3plPOSS
“the men’s sons”

* Archaic form used for clarity blah-ba-dee-blah

Not quite anything to write home about, other than maybe its thoroughness. As you can see, the possessed noun is marked for the number both of the possessum and the possessor. This works with both overt and covert possessors: remove the first noun in the last four examples and they all mean “his/their X[es]”. Inuktitut and Kalaallisut both, I believe, behave similarly.

It’s a bit different with Aleut, though! First of all, Aleut nouns can be either absolutive or “anaphoric” (possessed), but the latter series only has three forms per person per case… in the 3rd person these are sg. -V, du. -kix, pl. -s for the absolutive and sg. -gan, du. -kin, pl. -ngis in the relative case. So what marks the number of what and how? With overt possessors, these suffixes mark the number of the head noun, the possessum. Watch:

(As in Yup’ik, possessors are in the relative case.)
(The “A” in the breakdown means that the suffix is in the anaphoric series.)

Code: Select all

Aleut
hla-m      ada   -a
boy-REL.sg father-ABS.A/sg
“the boy’s father”

hla-s  ada   -a
boy-pl father-ABS.A/sg
“the boys’ father”

tayaĝu-m      hla-ngis
man   -REL.sg boy-ABS.A/pl
“the man’s sons”

tayaĝu-s  hla-ngis
man   -pl boy-ABS.A/pl
“the men’s sons”

“Ok, so the possessum isn’t marked for number of its possessor,” I hear you saying, “so what?” What’s what is what happens when the possessor is covert. Then, the anaphoric series is used to indicate the presence of a possessor, and its number marks the number of the possessum… OR the possessor… OR both! In short, the highest-number element determines the number marked on the head noun.

Code: Select all

hla-a
boy-ABS.A/sg
“his son”

ada   -ngis
father-ABS.A/pl
“his fathers” OR “their father” OR “their fathers”

“Ok, Cev, that’s a little interesting, but I don’t see why it was worth making a whole new thread about.” Just keep your britches on, boys and girls, cause I’m about to go into the REALLY interesting part about Aleut morphosyntax: the verbs.

Again, compare first with Yup'ik. I won't give another laundry list of Yup’ik examples, so suffice it to say that its nouns are marked S/O = absolutive case and A = relative case, and that its verbs are marked for transitivity and for person and number of both agent and patient in transitive verbs. Again, it’s deathly thorough. And again, Aleut has a morphologically simpler system, with a non-anaphoric conjugation (sg. -x̂, du. -x, pl. -s in the 3rd person) standing in opposition to an anaphoric one (sg. -V, du. -kix, pl. ngin), with only one argument agreed with for either. But this time, the anaphoric series doesn’t mark possession, but indicates and agrees with a “missing” complement in the sentence. What do I mean? Well, if a sentence is intransitive with either an overt or a covert subject, its verb has plain old regular non-anaphoric conjugation:

Code: Select all

tayaĝu-ẋ           hanga –ku -ẋ
man   -ABS.nonA/sg ascend-PRS-nonA/3sg
“the man is ascending”

hanga –ku -ẋ
ascend-PRS-nonA/3sg
“he is ascending”

… and if a transitive sentence has an overt patient, the verb still has non-anaphoric conjugation (and the agent, like in any well-behaved ergative language, is in the relative case):

Code: Select all

Ivaana-ẋ           kidu-ku -ẋ
Ivan  -ABS.nonA/sg help-PRS-nonA/3sg
“he is helping Ivan” †

But if a transitive sentence has a covert patient, the verb has anaphoric conjugation, which agrees in number with the patient.

Code: Select all

tayaĝu-m      hanga –ku -u
man   -REL.sg ascend-PRS-A/3sg
“the man is ascending it”

This applies whether the agent is overt or covert. So you can distinguish the transitivity of the following two sentences by their anaphoricity:

Code: Select all

qa –ku -ẋ
eat-PRS-nonA/3sg
“he is eating”

qa –ku –u
eat-PRS-A/3sg
“he is eating it”

Interestingly, though, and NOT like a well-behaved ergative language, if a transitive sentence has both agent and patient both represented with an overt noun, the nouns are both in the absolutive case. Essentially, Aleut marks for case only when it has to.

Code: Select all

hla-ẋ           suunaada-ẋ           agu –ku -ẋ
boy-ABS.nonA/sg toy.boat-ABS.nonA/sg make-PRS-nonA/3sg
“the boy is making a play boat”

Fine stuff, but that’s just frill really. All that crap was just to get to the real goods, the really trippy shit of Aleut morphosyntax. Consider the sentence:

Code: Select all

igachi-ngin  qitmigiku-ẋ
sinew –A/pl  cramp    -nonA/3sg
????????????????????????????

So there’s a noun, in the plural anaphoric, which as we’ve seen can mean either “his sinews”, “their sinew”, or “their sinews”. And there’s a verb with nonanaphoric conjugation, so there’s no covert patient. In fact, the verb is intransitive-only, so igachingin must be the subject, and since the verb is in the singular then the plural -ngin must refer to a plural possessor, so I bet you think this sentence means “Their sinew is cramped”. Don’t ya? DON’T YA?

WROOOOONG!!!!!

Apparently, Aleut has this nuts-as-hell rule that says if a clause’s “outer subject” (which here means the possessor of the subject) is not overtly indicated, that is if it is covert, then the verb in the clause agrees in number with it, and not with the actual subject of the sentence. So igachingin qitmigikux̂ in fact means “his sinews are cramped”, with the singular verb indicating the number of the possessor “his”. ISN’T THAT JUST NUTS?

The same thing happens with covertly-possessed patients of transitive verbs, only here since the “missing” element is in the object phrase I guess..., the verb now receives anaphoric conjugation.

Code: Select all

Piitra-m      ada   -ngis      kidu-ku -ngis
Peter –REL.sg father-ABS.A/3pl help-PRS-A/3pl
“Peter is helping their father”

So you see that even though Aleut’s person and number endings are more meager than those of other Eskimo-Aleut languages, they are also more versatile. Yup’ik may indicate the person and number of the agent and patient BOTH on the verb, but it doesn’t indicate the number of a covert possessor, does it?

Code: Select all

Yup'ik
Piitra-m      aata-at            ikayur-aa
Peter –REL.sg father-3sg.3plPOSS help  -3sg/3sg
“Peter is helping out their father”

Nope, didn’t think so.

And it isn’t just possessors that can be indicated by anaphoric conjugation of the verb, either. Objects of postpositions can, too. The first sentence here is a normal, non-anaphoric one, and the second drops the object of the postposition hadan and marks the dropped object with anaphoric conjugation:

Code: Select all

Aleut
Piitra-ẋ      tayaĝu-m      had   -an        huya-ku -ẋ
Peter –ABS.sg man   -REL.sg direc.-LOC.A/3sg go  -PRS-nonA/3sg
“Peter is going toward the man”

Piitra-m      had   -an         huya-ku -u
Peter –REL.sg direc.-LOC.A./3sg go  -PRS-A/3sg
“Peter is going toward him/her.”

You might say that that’s just more of the same, and that hadan isn’t a postposition, but a possessed noun, and you might be right, but none of the previous possessed nouns were in the locative. And besides, we’re not done yet. There's one more thing that anaphoric conjugation can agree with, and that is “an element of a nonfinal, that is to say, subordinate clause”:

Code: Select all

igaẋta-ẋ        kimiku   -ẋ        ukuẋta-l    angali  -q
airplane-ABS.sg come.down-nonA/3sg see   -CONJ do.today-nonA/1sg
“I saw the airplane coming down”

kimiku   -ẋ        ukuẋta-l    angali  -ng
come.down-nonA/3sg see   -CONJ do-today-A/1sg
“I saw it coming down”

Interestingly, as you can sorta see in the last example, anaphoric verbal conjugation does not agree in person with a covert argument, but with the “normal” subject of the verb. Nor does it agree in number with only one argument, if the latter is first or second person. So we still get a little of the Eskimo-Aleut bivalent agreement. Here’s a more clear example.

Code: Select all

hada  -ngin    huya-ku -ning
direc.-LOC.3pl go  -PRS-A/3pl/1sg
“I am going towards them”

hada  -ngin    huya-ku –mas ‡
direc.-LOC.3pl go  -PRS-A/3pl/1pl
“We are going towards them”

And that’s the Aleut morphosyntactic system. Pretty cool, huh? I think so. Run tell your friends.



† - You may be wondering why this isn’t “Ivan is helping”, since the sentence is built the same as tayaĝux̂ hangakux̂ above. The short answer is “I have no fucking idea”, and the long answer is “Kidu is a transitive-only verb so there has to be an object, but some verbs in Aleut are apparently labile so I don’t know how those do it, but possibly it uses a valency-changing suffix since Yup’ik does something similar.”
‡ - This sentence is of my own making, but it aligns with the previous sentence except for a simple morpheme-switch, so I highly doubt it’s incorrect.

[Soap]

What does covert and overt mean? Is covert just a way of saying that it's a pronoun?

[Yng]

p. fuckin' crazy

[Miekko]

awesome

[äreo]

That's fucking cool! Morphosyntastic.

[Salmoneus]

Iirc some Amazonian languages also do some pretty cool things when 'expected' information elements aren't overtly present. Not quite as cool as this though, I don't think, and I'm afraid I have no details to hand.

[Jipí]

Where's the fucking like button when you need it? This is the stuff I'd like to see on the Conlang Blog Aggregator more often rather than just random "Daily Words" in a handful of conlangs. At least we have William Annis on the Conlangery Podcast

[TomHChappell]

What does covert and overt mean? Is covert just a way of saying that it's a pronoun?

Apparently "covert" means it doesn't appear explicitly as an independent word:
if a covert participant or possessor shows up at all, it does so only in that some other word (e.g. the verb or the possessum) agrees with it.

"Overt" OTOH means it does show up explicitly as an independent word, either as a full noun-phrase or as an independent pronoun.

Of course the O.P., or one or more of the sources they quoted, should really be the one(s) to answer your question.

[merijn]

This is one of the coolest morphosyntactic things I have ever read.


++++

Covert means that something that is present in the semantics is not expressed in the pronounced form of a sentence.

[Jipí]

What does covert and overt mean? Is covert just a way of saying that it's a pronoun?

Stuff

From context it should be clear that "covert" means "only indicated by a pronoun/anaphoric ending" (i.e. the referent is implicit information), while "overt" means "is present as a proper NP" (i.e. the referent is explicit information).

[Cathbad]

So is this "just" that the verb ending disambiguates w/r/t the number of a non-expressed argument? So number agreement on the verb isn't "fixed" to refer to a specific slot/argument, but is just understood to cover whatever's missing?

(I'm sorry if this sounds a bit thick to all you MS-whizkids, but I always try to simplify things to myself as much as possible )

[Xephyr]

So is this "just" that the verb ending disambiguates w/r/t the number of a non-expressed argument? So number agreement on the verb isn't "fixed" to refer to a specific slot/argument, but is just understood to cover whatever's missing?

Provided that the "non-expressed argument" is one of the things that can trigger anaphoric conjugation-- objects, object possessors, etc., then yes that's the gist of it.

[Whimemsz]

Iirc some Amazonian languages also do some pretty cool things when 'expected' information elements aren't overtly present. Not quite as cool as this though, I don't think, and I'm afraid I have no details to hand.

They do, sort of, according to The Amazonian Languages, ed. R. M. W. Dixon and Alexandra Y. Aikhenvald. Or at least, some Macro-Jê languages are said to mark "the head of a nominal, verbal, or postpositional phrase for textual contiguity...or non-contiguity...of the determiner" (pg 181). Unfortunately, the author (Aryon D. Rodrigues) doesn't really go into any further detail, beyond providing a handful of examples. Here are some, though:

Code: Select all

Ofayé

pɨkɨtɨɛn ʃ-ɛnʃɨh
caiman CNT-heart
"the caiman's heart"

vs

h-ɛnʃɨh
NCNT-heart
"its heart"

Or

Code: Select all

Timbíra

i   tɛ       pĩco       j-ũʔkʰər
1sg ERG.PAST tree.fruit CNT-buy
"I bought fruit"

vs

i   tɛ       h-ũʔkʰər
1sg ERG.PAST NCTN-buy
"I bought it"

At least, I'm assuming this is the same thing you're talking about. But it doesn't seem to be nearly as cool as the Aleut system.

[Radius Solis]

That kicks mountains of ass, Xephyr. Thanks!