Balog: a strictly monovalent conlang

[Imralu]

So do you think du is expanding its range at the expense of sie?

Yeah, well, it's definitely not going in the other direction. Scandinavian languages are far ahead. Danish de (3P) is still used for the second person, but I think it's restricted to much more formal contexts than Sie in German. Swedish ni (2P) is all but gone as a polite form for 2S. It's pretty much been phased out in the last century.

I find it really jarring in subtitles when I'm watching something in Danish and Swedish, which I understand OK, but not well enough to just go without subtitles and then there's only German subtitles and the characters are all calling each other by their first names and saying "du" and the German subtitles are using "Herr/Frau _____" and they're calling each other Sie, even after they've known each other for a year and are clearly friends and have been through some big shit together.

(More generally, it looks like I need to read up on politeness marking — I think I may have totally misunderstood the concept.)

Yeah, I think it's generally misunderstood by speakers of different languages. Like, people wondering why God is spoken to informally in the Bible. At least in most European languages, formal doesn't actually mean more respect. It means distance. You can actually jokingly disrespect your friends by addressing them formally. It kind of has more of a feeling of "I don't know you" or "I don't consider you a friend" than "I respect you". I think that's probably a bit of a Sprachbund feature though. Apparently in Thai, you use respectful pronouns with your parents to show your respect and that would be positively disrespectful in German, like you don't recognise them as your parents. The closest thing I can think of in English would be calling your parents by their names instead of "Mum" and "Dad" - it just makes them some person rather than your parent (and I think it's wonderful ).

I personally really hate formality distinctions in languages, but I like to put a lot of features in my languages that I would actually hate to deal with in real life. This rank thing is horrible, like, fifty times worse than the most stuffy formality distinction, but to me that makes it a Fun Conlang thumbs_up.bmp

I’m not sure I understand your problem. Why would you need to embed a noun phrase within a verb phrase to apply a definite nonspecific article?

Because the article is a nominaliser and signals the subject and therefore as soon as you add it to a verb, it is not a predicate, so, currently, in Balog, I have no way of distinguishing between "The house is (a) big (one)" and "The house is the big one." The predicate is generally taken to be indefinite by default, so there should be some way of marking definiteness within the predicate.

Bad ibben.

bad

be.big

i=

DEF(e)=

ben

be.house

The house is big.

Ben ibbad.

ben

be.house

i=

DEF(e)=

bad

be.big

The big one is a house.

*Ibbad ibben.

*

*

i=

DEF(e)=

bad

be.big

i=

DEF(e)=

ben

be.house

*the big one ... the house ... (Two subjects with no predicate. This is not a valid way to say "The house is the big one")

The way I get around this in Iliaqu is with a verbaliser, which I gloss as "COP". Essentially, it gets added to the small, closed class of nominals (which function as articles and pronouns) and it prevents them from beginning an argument/adjunct phrase and transforms them into a verb (phrase). I'll show verbs in blue and nominals in red, with relativised verbal phrases enclosed within red brackets with their head.

Tta [xu mala].

tta

be.big

xu

NOM.3S.DEF.INAN

mala

be.house

The house is big.

Mala [xu tta].

mala

be.house

xu

NOM.3S.DEF.INAN

tta

be.big

The big one is a house.

... [xu tta] [xu mala].

...

...

xu

NOM.3S.DEF.INAN

tta

be.big

xu

NOM.3S.DEF.INAN

mala

be.house

... the big one and the house ... (Two nominal arguments, understood with an "and" linking them and assumed to be two different entities, but, nevertheless, not a valid sentence without a predicate or surrounding context such as a question.)

To solve this, there is the verbaliser/copula which is prefixed to the nominal which is to be verbalised and barred from beginning an argument.

[Nxu tta] [xu mala].

N-

COP-

xu

NOM.3S.DEF.INAN

tta

be.big

xu

NOM.3S.DEF.INAN

mala

be.house

The house is the big one. (Definiteness marked inside a predicate.)

This can also be used with indefinite nominals (either specific or non-specific), but it's pretty unnecessary in most cases as the predicate is generally interpreted to be indefinite (e.g. zu mala "a (specific) house" in the nominative can be verbalised to nzu mala "be a (specific) house", but it's seldom necessary).

However, it can also be used with other cases as well. I call it a copula because it is used to put a nominal phrase into a predicate (or in fact, any verb phrase, including the imbedded ones inside nominal phrases), but it seems less copula-like when it appears with certain cases, such as the dative (also functions as allative) or the ablative, in which case it is generally translated with "go" or "come".

[Nxua mala] [ju gole]

N-

COP-

xua

LOC.3S.DEF.INAN

mala

be.house

ju

NOM.3S.DEF.ICS

gole

be.woman

The woman is at the house.

[Nzua mala] [ju gole]

N-

COP-

zua

LOC.3S.SPEC.INAN

mala

be.house

ju

NOM.3S.DEF.ICS

gole

be.woman

The woman is at a (specific) house.

[Gexi mala] [ju gole]

N-

COP-

exi

DAT.3S.DEF.INAN

mala

be.house

ju

NOM.3S.DEF.ICS

gole

be.woman

The woman goes to the house.

[Nju gole] [ju [gexi mala]]

N-

COP-

ju

NOM.3S.DEF.ICS

gole

be.woman

ju

NOM.3S.DEF.ICS

N-

COP-

exi

DAT.3S.DEF.INAN

mala

be.house

The one who goes to the house is the woman.
(Most literally: The one who goes to the thing which is a house is the one who is a woman.)

Without N-, there would be no way to equate pronouns to each other either, because, for example, there is no verb meaning "to be me" other than the nominal "I" prefixed with this verbaliser.

Nna na.

N-

COP-

na

NOM.1S.ACS

na

NOM.1S.ACS

I am me.

This is not so in Balog because the person and number markers are themselves verbs. The nominalisers' only roles (so far), in addition to indicating subjects, are to mark definiteness, specificity and rank, not person or number as the nominals in Iliaqu do, so ...

Ž'oož

ž

1S

o=

DEF(c)=

ž

1S

I am me.

That means it's not such a necessary structure in Balog, as well as the fact that there are no cases that could also be useful to sneak into a verbal phrase.

Perhaps I will only allow it in Balog with a two-clause structure.

(Oh, and I’ve just realised I made a mistake in my previous post: I meant to say that the article would fuse with the noun, not the verb.)

Hmm, I thought you were right in saying that, but I wasn't quite sure what you meant and now I'm even less sure. The articles essentially are the only nouns and they need to be completed with an attached verb which functions as more or less as a relative clause, e.g. oož is essentially "the (c-ranked) entity that is me" (i.e. "I" in the c rank).

So for the question words such as who, I was either thinking of having a specific nominaliser/subject marker that means "entity known to you but not me" attached to a generic class verb such as "be.human", or simply using a definite nominaliser/subject marker with a verb that explicitly contains the wh-meaning, such as "be.who". I think possibly combining both would be a good solution. Explicit wh-verbs (or at least being able to do that in a verb phrase), would be good as it would allow them to be used in the predicate (without that kind of imbedding I discussed above), but a specific nominaliser for wh-words would also be good (A) for extra redundancy, something which is often lacking in my often very compact languages) and (B) to avoid the issue of having to choose a rank for an unknown entity.

This is a very interesting point — at what level of loaning can you say a class is closed? But I suppose you can answer this by comparison to other languages. Here, Kalam provides a relevant contrast — it has ~130 verbs and none of them are borrowed (though speakers reanalyse SVCs as single verbs occasionally), whereas Wutung has 27 containing one loan. Percentage-wise at least, the Wutung verb class is far more open than the Kalam one. And yes, it’s extremely small, but Papuan languages generally tend to have small verb classes.

I guess there's no absolute. Some classes might be completely closed but haven't always been (thus accounting for old loanwords) and all languages evolve. You could say that verbs in Wutang are currently a closed class but that they were a slightly open class at the time that the loanword entered the language. That's more or less just playing with semantics though and at the end of the day, closed and open are just labels. The question is how complete do you want the closure to be for it to count as "closed" and how small a crack would cause you to say it's an open class. Perhaps there are diagnostic tests for that and a clear-cut, established answer, but either way, the level of openness of a word class to knew words is clearly more or less on a sliding scale.

gos tmey nŋ- thought bad perceive ‘dislike, hate something’

mnm ag ask ay- talk say avoid stabilize ‘leave or avoid (a topic)’

ap tan ap yap g-p-ay tam come ascend come descend do-HAB-3PL junction ‘crossroads’ (i.e. ‘junction where they come up and go down’; a nominalization)

ñag jw yok- shoot withdraw displace ‘rout (the enemy) in a war’

pwŋy pag yk- impinge disturb open ‘prise something open or free’

d nŋ- touch perceive ‘feel something (by touching)’

kwy ap-INFL nŋ- odour come perceive ‘smell something’

It helps that most Kalam verbs are really vague: tk-, for instance, means ‘cross (a divide), sever, separate, interrupt, cut off, transform, suddenly become (night, day, clear), have a child’, amongst other things.

Damn, these are cool! Thanks!

[Imralu]

I’m not sure I understand your problem. Why would you need to embed a noun phrase within a verb phrase to apply a definite nonspecific article?

I've already written about this above, but I just wanted to illustrate this a bit better and give more examples of the copula/verbaliser in Iliaqu to give examples of it really not looking anything like a copula and also to show why it's really quite indespensible in Iliaqu. As I said before, it's function is to allow a nominal to function as a verb and prevent it from signalling the beginning of a new adjunct*. All of the examples I've had so far have only had one adjunct* per sentence, meaning perhaps their signalling role was not clear.

*One thing that's unusual about Iliaqu is that even subject and object are never obligatory in Iliaqu, and as far as I've understood the distinction between the words "adjunct" and "argument", that means there are, strictly speaking, no arguments in Iliaqu, only adjuncts, so here I'm using the word "adjunct" simply as a catch-all term for arguments, adjuncts, whatever: nouny-boys in the sentence that relate to the predicate. Another unusual thing is that adjuncts always hang syntactically from the predicate and (without subordination by means of the copula/verbaliser) never from neighbouring adjuncts. Even genitive (association) and possessive (legal ownership) don't simply modify a neighbouring adjunct as they generally do in most languages with similar cases but always modify the predicate.

Here are some sentences showing that there are no obligatory arguments. A bare predicate with no adjuncts is generally translated as "There is/are [noun]" (e.g. iio = there's a fish; kau ?"there's an eater") or "Somebody or something is [verb]ing" (e.g. kau "somebody is eating"; iio ?"something is a fish"), Adjuncts, with their case, provide more information about the entity/action in the predicate.

[Iio]_pred.

iio

be.fish

There's a fish. / There are fish.

[Iio]_pred [xuue]_nom.

iio

be.fish

xuue

NOM.3S.DEF.INAN.PROX

This is a fish.
(There is a fish and it's this.)

[Iio]_pred [unu]_pos.

iio

be.fish

unu

POS.1S.ICS

I have a fish. / I legally own a fish.
(There is a fish and it's mine.)

[Gunu]_pred [xu iio]_nom.

N-

cop-

unu

POS.1S.ICS

xu

NOM.3S.DEF.INAN

iio

be.fish

The fish is mine. / The fish is my legal property.
(Something is mine and it's the fish.)

[Iio]_pred [nua]_loc.

iio

be.fish

nua

LOC.1S.ICS

I have a fish (with me).
(There's a fish and it's at me.)

[Iio]_pred [unu]_pos [nua]_loc.

iio

be.fish

unu

POS.1S.ICS

nua

LOC.1S.ICS

I own a fish and I've got it with me.
(There is a fish, it's mine and it's at me.)

[Iio]_pred [unu]_pos [jua muja ]_loc.

iio

be.fish

unu

POS.1S.ICS

jua

LOC.3S.DEF.ICS

muja

be.man

I've got a fish and the man has it.
(There is a fish, it's mine and it's at the man.)

I could of course translate the latter as "the man has my fish" as, pragmatically, that's exactly what it means, however, syntactically, it's more like making two claims about the fish, or potentially three: fish exists, it's my legal possession and it is located at the man. To say "the man has my fish", the most literal translation would be:

[Iio [gunu]]_pred [jua muja ]_loc.

iio

be.fish

N-

COP

unu

POS.1S.ICS

jua

LOC.3S.DEF.ICS

muja

be.man

.
The man has my fish (on him/with him).
(There is a fish of mine and it's at the man.)

Now, my has been subordinated to within the predicate phrase (iio gunu). Unu means "my" as a noun phrase, indicating that the predicate is the speakers legal possession. Gunu is a verb phrase meaning "is mine", but because it is, in this sentence, a modifier within the predicate verb phrase, it doesn't really "feel" like it means "is mine", but rather, simply modifies the preceding word(s) within the verb phrase. Pragmatically, in this particular sentence, it makes little difference to this sentence whether you say unu or gunu except that unu places greater focus on my ownership of the fish. Another couple of sentence should hopefully make it clear why this structure is frequently necessary and can make a big difference.

[Ila]_pred [ju muja]_nom [vexi iio [gunu]]_top

ila

speak

ju

NOM.3S.DEF

muja

be.man

vexi

TOP.3S.DEF.INAN

iio

be.fish

N-

COP-

unu

POS.1S.ICS

The man talks about my fish.
(Someone talks, it's the man, their topic is the fish of mine.)

(N.b. TOP here doesn't mean a topical in the sense of a grammatical topic, but rather simply a case that basically means "about".)

If you don't use the verbalised form of unu, gunu in this sentence, but instead use the bare nominal form unu, it relates back to the predicate, as all adjuncts do, and thus gives a very different meaning.

[Ila]_pred [ju muja]_nom [vexi iio]_top [unu]_pos

ila

speak

ju

NOM.3S.DEF

muja

be.man

vexi

TOP.3S.DEF.INAN

iio

be.fish

unu

POS.1S.ICS

The man talks about the fish and he's my speaker that I own.
(There is a speaker, it's the man, his topic is the fish and he's my property.)

Using unu without nominalising it, it relates to the predicate and asserts that I own the speaker. The word order is also a bit unusual as "unadorned nominals" (those that are not followed by a subordinated verbal phrase) tend to gravitate towards the beginning of the sentence and it would be natural for the unadorned nominal unu to appear before the adorned nominal phrases ju muja and vexi iio (not because it relates to the predicate -- all the adjuncts* (marked in red) relate to the predicate -- but simply because it is unadorned and thus, "lighter"). The word order, with unu occurring after vexi iio is not enough for the possessive adjunct to modify the preceding topical adjunct; it needs to be verbalised with a nasal prefix (in this case g- /ŋ/) in order to be subordinated within the topical adjunct's own subordinated verb phrase.

Here are another few examples that should make it clear why verbalisation of nominals is useful. There is a verbal-derivation that essentially adds a locative meaning to verbs: the prefix a-

[mala]_pred [nu]_nom

mala

be.house

nu

NOM.1S.ICS

I am a house. (Generally pretty nonsensical.)

[amala]_pred [nu]_nom

amala

be.at.home

nu

NOM.1S.ICS

I am at home.

Amala is understood idiomatically as being at home, but we cannot give much more information about the house with this. There are also equivalent verbs for a whole range of different locations:

loie 'be (a) bed(s)' → aloie 'be in bed'
utigia 'be (a) forest(s)' → autigia 'be in (a/the) forest(s)'
lubu 'be (a) box/container/cage' → alubu be contained/boxed/caged/imprisoned

As with all verbs, there is little possibility for telling much about the definiteness, and in these cases, the house, the bed, the forest etc. cannot be directly described. Adding, for example, tta 'be large' to these does not describe the size of the house, bed, forest or the container but rather the entity that is at home, in bed, in a forest, caged etc. because it modifies the whole word, not simply everything after a-

lubu tta 'be a large cage' → alubu tta 'be a large prisoner'

So to say "I am at the big house" or "I am at a big house", the whole nominal phrase "at a/the big house" is verbalised with the nasal prefix and that allows it to appear anywhere any verb can, whether as the head of the predicate, a modifier within the predicate, the head of a subordinated verb phrase within a noun-phrase or a modifier within a subordinated verb phrase within a noun phrase.

[[nxua mala [guma muja tta]]_pred [nu]_nom

N-

COP-

xua

LOC.3S.DEF.INAN

mala

be.house

N-

COP-

uma

POS.3S.SPEC.ACS

muja

be.man

[gloss]be.large[/gloss]

nu

NOM.1S.ICS

I am at the house of a large man.

[[nzua mala tta]]_pred [nu]_nom

N-

COP

zua

LOC.3S.SPEC.INAN

mala

be.house

tta

be.large

nu

NOM.1S.ICS

I am in a large house.

As you can see, if you've followed me this far, subordination of nominals within verb phrases (by verbalising them) is a vital process in Iliaqu. Without it, the only way to indicate location or ownership or a multitude of other grammatical functions indicated by the case-number-definiteness-specificity-accessibility marking nominals would be through separate adjuncts, which always refer back to the predicate. That would make it quite hard to say sentences like "My flag boy said to your flag boy: 'I'm gonna set your flag on fire.'"

I don't yet think that such a system would be necessary in Balog, but I haven't yet got a full idea of the whole grammar and it may turn out that I do feel the need for something like that; I just don't want to rush it yet and copy Iliaqu because maybe I'll instead come up with a completely new work-around for things like that. From the character of Balog, I think it would be somehow more fitting to do it by juxtaposition rather than explicit, unequivocal subordination as that seems to be more of the theme of Balog. Balog and Iliaqu function similarly in how the lexicon is comprised of a large verb-class that can be nominalised by essentially relativising them, but whereas Iliaqu is full of cases and can also verbalise nominatives, Balog, so far, has a predominantly Verb-subject, verb-subject, verb-subject structure. Aside from the subordination of verb phrases by the subject markers and the subordination of entire clauses (and groups of clauses) by the "propositional" subject markers (i.e. the complementisers ä=, äqä= and ähä=) I think I will try to avoid adding much else in the way of subordinating machinery if I can get away with it. Oh, there are already equative modifiers for use within verb phrases. Hmm. Anyway, I don't want these two languages to be too similar to each other because I guess the main point of most of my conlangs is testing out "Can this work?" and Iliaqu works. (I mean it works structurally at least; I feel as though most of my conlangs would be impractical as real languages because of too little redundancy.)

Yeah, it’s really nice when that happens!

Well, it's nice when it gives you something that you're happy with. I can't think of any specific examples, but I've definitely had decisions that have then ended up with me going "... but then that would mean ... NO! URGH! HORRIBLE! SCRAP THAT! BURN COMPUTER! LEAVE COUNTRY!"

[Qwynegold]

Hi! Hmm, not sure what you mean as I had a look at his Rkou and I honestly can't see any similarities other than the fact that they are both spoken by orcs. Or is that all you meant?

Yeah, and he made them not wear shirts because he's into muscles. Whereas you've said that you like strong guys...

The unspecific/specific/definite distinction is interesting. I think I've seen it in some other conlang as well, but does it exist in any natlang?

What if the orcs were completely illiterate? Or could the dominance problem in writing be somehow solved if there was a special scribe occupation for very few people, and they'd seldom write on their own behalf?

[Qwynegold]

It looks like we all joined in 2018, but I joined the original site in 2006 (in May, maybe, from memory??) and I mostly can't remember who's been around longer than me and who hasn't. Like, I joined the new site a few days before Qwynegold, but I can't remember if Qwynegold is an oldbie who's been around longer than me.

Hmm, I joined either in the fall of 2006 or spring of 2007.

[Imralu]

Yeah, and he made them not wear shirts because he's into muscles. Whereas you've said that you like strong guys...

Well, they have to photosynthesise
I probably have said that somewhere, but not in this thread. Shhhhh!

The unspecific/specific/definite distinction is interesting. I think I've seen it in some other conlang as well, but does it exist in any natlang?

Yeah, well, quite a few mark specificity rather than definiteness and a three-way distinction must be out there. The only natlang I can think of right now that has a three-way distinction is Tongan. Here I've compared it with two of my conlangs:

Code: Select all

                                         Tongan       Balog      Iliaqu
NSPC.INDEF: "a(ny) house"               ha fale     ihibben      a mala
SPEC.INDEF: "a (particulary) house"   (h)e fale     iqibben     zu mala
       DEF: "the house"               (h)e falé       ibben     xu mala

I literally don't know how to find other language examples because the topic of specificity has been horribly neglected and misunderstood by linguists. There isn't even an official abbreviation for specific or non-specific given in the Leipzig glossing rules. Lots of languages with a specificity distinction have been mistakenly explained in terms of a definiteness distinction. This sentence from the Wikipedia article on Samoan is pretty typical (emphasis mine):

The definite article is le: ʻo le Atua, God; indefinite e.g., ʻo le matai o Pai, (the) chief (named) Pai. It is sometimes used where English would require the indefinite article: Ua tu mai le vaʻa, a canoe appears.

The specific article le has been called the definite article because whoever wrote that (the Wikipedia article or whatever source was used) was unaware of the concept of specificity. (I also have no idea why the word "indefinite" appears after ʻo le Atua and why the "e.g.," appears AFTER the first example.

More: show

I looked back through the history and on the 21st of November, 2010 it looked like this:

The article le is both definite and indefinite; at least as it is constantly used in Samoan, whereas the English would require the indefinite article; definite e.g., ʻo le Atua, God; indefinife e.g., ʻo le aliʻi Pai, such as, one is a chief. On looking into such cases, it will be found that there is something definite, from a Samoan standpoint, which makes them use le rather than se, as Ua tu mai le vaʻa, a canoe appears.

Se is always indefinite; ta mai se laʻau, cut me a stick.

That's really terrible. It has been improved somewhat. I think I might do some work on it later.

It's really hard to actually search for more examples of languages specificity marking rather than definiteness, let alone languages with a three-way distinction (or even a four-way distinction, if it exists out there. I came across Tongan a few years ago and at the time, the Wikipedia article was also worded terribly as if it was a definiteness distionction. I have an old grammar of Tongan which also gives (h)e as a "definite article" even though it is unequivocally a specific article and it contains the sentence: "When it is not reinforced by the definitive accent, a definite article, so-called, is only semi-definite, and hence, at times, may seem to an English mind to be indefinite."

What if the orcs were completely illiterate? Or could the dominance problem in writing be somehow solved if there was a special scribe occupation for very few people, and they'd seldom write on their own behalf?

Well, most of them are and that's how I've envisaged it, that writing is a specialised activity performed by the upper echelons. In fact, probably they'd get non-orcs to write for them since they don't have such fine motor control. The person who had the writing done, however, is still responsible for it. In any case, whether or not the language is written at all in-world, if I'm using it on the forums here, that still poses these kinds of dilemmas ... dilemmae ... dilemmata.

[bradrn]

I’m not sure I understand your problem. Why would you need to embed a noun phrase within a verb phrase to apply a definite nonspecific article?

Because the article is a nominaliser and signals the subject and therefore as soon as you add it to a verb, it is not a predicate, so, currently, in Balog, I have no way of distinguishing between "The house is (a) big (one)" and "The house is the big one." …

OK… I’ve read through this several times now, and I think I understand it now. (As opposed to the Iliaqu post, which is still thoroughly confusing me.) So a content word with an article cannot act as a predicate, which means you can’t equate two definite arguments, which means you can’t distinguish definiteness in predicates.

First question: Why would you even need to make this distinction? Plenty of languages have no definiteness distinction at all, and get along just fine without a distinction between ‘the house is big’ and ‘the house is the big one’.

Second question: Why can’t a content word with an article act as a predicate? A simple way of solving this problem would be to give such a clause an equative interpretation.

You’ll note that I’ve carefully avoided using terms like ‘noun’ and ‘verb’ above. I think that for a language like Balog, such terms tend to be confusing rather than helpful — I analyse it with a single class of ‘content words’, which may occur in both predicate and argument positions. In the latter they may take an article (I can’t remember if this is obligatory or not).

(Oh, and I’ve just realised I made a mistake in my previous post: I meant to say that the article would fuse with the noun, not the verb.)

Hmm, I thought you were right in saying that, but I wasn't quite sure what you meant and now I'm even less sure. The articles essentially are the only nouns and they need to be completed with an attached verb which functions as more or less as a relative clause, e.g. oož is essentially "the (c-ranked) entity that is me" (i.e. "I" in the c rank).

So for the question words such as who, I was either thinking of having a specific nominaliser/subject marker that means "entity known to you but not me" attached to a generic class verb such as "be.human"

Sorry, I had forgotten that you were calling the content words ‘verbs’. But this is exactly what I was trying to say.

The unspecific/specific/definite distinction is interesting. I think I've seen it in some other conlang as well, but does it exist in any natlang?

Yeah, well, quite a few mark specificity rather than definiteness and a three-way distinction must be out there. The only natlang I can think of right now that has a three-way distinction is Tongan. Here I've compared it with two of my conlangs:

Code: Select all

                                         Tongan       Balog      Iliaqu
NSPC.INDEF: "a(ny) house"               ha fale     ihibben      a mala
SPEC.INDEF: "a (particulary) house"   (h)e fale     iqibben     zu mala
       DEF: "the house"               (h)e falé       ibben     xu mala

I literally don't know how to find other language examples because the topic of specificity has been horribly neglected and misunderstood by linguists. There isn't even an official abbreviation for specific or non-specific given in the Leipzig glossing rules. Lots of languages with a specificity distinction have been mistakenly explained in terms of a definiteness distinction.

I think Hausa may have a three-way distinction, but I’m not sure. I know I’ve seen one somewhere though.

[Imralu]

OK… I’ve read through this several times now, and I think I understand it now. (As opposed to the Iliaqu post, which is still thoroughly confusing me.) So a content word with an article cannot act as a predicate, which means you can’t equate two definite arguments, which means you can’t distinguish definiteness in predicates.

First question: Why would you even need to make this distinction? Plenty of languages have no definiteness distinction at all, and get along just fine without a distinction between ‘the house is big’ and ‘the house is the big one’.

Of course, but that's like saying why have definiteness at all. Why have specificity at all. Many languages do without a lot of things. Balog could, at this stage, easily get along without definiteness/specificity marking in the predicate and, as I've said, I may simply find a biclausal solution to situations like that.

In any case, I don't know of a language that marks definiteness but cannot mark it in predicate positions. In English, for example, the strategies are different for different parts of speech. With nouns it's the most simple because there's already a slot for articles.

Indefinite: This is a house.
Definite: This is the house.

With adjectives, indefiniteness is default. To mark definiteness, we need a dummy noun or pronoun like "one" as well, or alternatively a relative clause.

Indefinite: You're crazy.
Definite: You're the crazy one. / You're the one who's crazy.

With verbs, again, indefiniteness is default and relativisation is needed for definiteness.

Indefinite: I apologised
Definite: I'm the one who apologised.

In Balog, this could mostly be achieved simply by swapping what is subject and what is predicate as the subject can be marked for definiteness, but then what is pushed into the predicate is interpreted as indefinite. For example, if you were talking about a movie with someone and they said "The man apologised" and you contradict them and say "No, the woman was the one who apologised" or, with more focus on the woman, No, it was the woman who apologised, you're equating two definite noun phrases. Currently, in Balog, you could only have definiteness in the subject of the sentence, either: "No, the woman apologised" (definite subject, indefinite predicate) or "No, the one who apologised was a woman" (definite subject, indefinite predicate).

Of course it's certainly possible for a language not to have a way around this, but I suspect I may want to come up with a way around it. I just don't know which way yet.

Second question: Why can’t a content word with an article act as a predicate? A simple way of solving this problem would be to give such a clause an equative interpretation.

Because the article marks a word as a subject and thus not a predicate.

Of course, I could change that rule and allow simple equative phrases where two articled phrases could constitute a clause, but since sentences can already consist of paratactic arrangements of clauses simply juxtaposed without any obligatory conjunctions, doing away with the marking that shows explicitly what is a predicate and what is a subject is bound to lead to some sentences that are difficult to parse. Having unmarked predicates followed by marked subjects makes parsing very easy. Of course, clause boundaries could be shown by intonation, but I simply don't want that, and, logically, the predicate and the subject of a clause, if both marked for status, would always show the same rank (as they refer to the same referent), so pairs of like-ranked articled phrases could be interpreted as clauses, but still, I don't think I really like that as a solution. The articles are first and foremost an indicator of a subject (contrasting with predicates which are zero-marked) The definiteness, specificity and rank distinctions, although enjoyable, were more or less a secondary addition to go along with the marking of a subject. Since there has to be something there, what other stuff can I do with it.

Also, in Iliaqu, two articled phrases in the same case are interpreted as being linked by "and" I really find it very fun and practical and might use that in Balog as well. In Iliaqu, this kind of juxtaposition is even the normal way of forming inclusive "we". The first person plural ina (accessible) or inu (inaccessible) is explicitly exclusive.

Amo

be.where

ua

NOM.2S.ACS

na

NOM.1S.DEF.ACS

?

?

Where are we (you and I)?

Ti

RET

bozu

find.coincidentally

nu

1S.ICS

zi

ACC.3S.SPEC.INAN

kuaga

coconut

zi

ACC.3S.SPEC.INAN

kutaki

key

.

.

I found a coconut and a key (and I wasn't looking for them).

You’ll note that I’ve carefully avoided using terms like ‘noun’ and ‘verb’ above. I think that for a language like Balog, such terms tend to be confusing rather than helpful — I analyse it with a single class of ‘content words’, which may occur in both predicate and argument positions. In the latter they may take an article (I can’t remember if this is obligatory or not).

Yep. Good choice. Content word is clunky though, so maybe CW.

I think Hausa may have a three-way distinction, but I’m not sure. I know I’ve seen one somewhere though.

Yeah, they must be out there.

Māori is a funny case that's relevant in a few ways here, but basically nothing is simple, everything is complicated. Usually, learning materials give the articles as

Code: Select all

        SG. | PL.
      |-----|------|
  DEF:| te  | ngā  |
      |------------|
INDEF:|     he     |
      |------------|

There is also tētahi, which I have seen glossed as "a certain", "a particular" and I've also seen it called specific and he as non-specific, so we could think of Māori as having a three way distinction:

Code: Select all

             SG.    | PL.
           |--------|-------|
       DEF:| te     | ngā   |
           |--------|-------|
SPEC.INDEF:| tētahi | ētahi |
           |----------------|
NSPC.INDEF:|        he      |
           |----------------|

BUT, I don't think it's anywhere near that simple. He is used in the predicate a great deal. It gets complicated by the fact that he is pretty restricted in where it can be used. For example, it can't be used after prepositions such as i (which is used to introduce a direct object, among many other things) and due to the funny sentence transformations required to even negate a sentence, such as with ehara, which more or less seems to function as a transitive verb, it means that a reliable definiteness distinction is often destroyed by negation.

he

INDEF.NSPC

kaiako

teacher

te

DEF.SG

wahine

woman.SG

The woman is a teacher.

If the predicate should be definite (or a proper noun etc.), the word ko is needed (in most dialects at least) and it can be thought of as a copula that indicates that the subject and predicate are the same individual (not simply a subset relationship as in the sentence above) and very often, it seems pretty blurry which noun phrase is regarded as the subject and which the predicate as they both refer to the same referent(s) rather than a set and a subset. Often, the translation seems to work either way.

ko

COP.EQU

te

DEF.SG

kaiako

teacher

te

DEF.SG

wahine

woman.SG

The woman is the teacher. (Or possibly: The teacher is the woman.)

To negate both of these sentences, because i is needed, he cannot be used, so generally te is. (I think in this case, it would always be te rather than ngā for plural, but I'm really not sure):

ehara

NEG

te

DEF.SG

wahine

woman.SG

i

ACC

te

DEF.SG

kaiako

teacher

The woman is not a/the teacher.

I'm also not sure if tētahi can be used there instead. I've found quite a few examples of ehara [subject] i tētahi ... but generally what follows is o ("of") and then often a plural noun phrase, so I think you could say Ehara te wahine i tētahi o ngā kaiako, which I suppose would mean "The woman is not one of the teachers," but I'm not sure if you can say ?Ehara te wahine tētahi kaiako.

Where I am sure that tētahi can be used is in the object phrase and locational phrases etc, and there they are often used because he cannot be.

i

PST

kite

see

ahau

1S

i

ACC

tētahi

SPEC.INDEF

manu

bird

I saw a bird.

I find it really quite funny when people act as if morphological complexity means a language is difficult, all those memes showing endless different forms of Finnish nounds or Spanish verbs. I've found Swahili's grammar to be quite simple to learn, but Māori just confuses me all the time and I honestly think it's the hardest language I've tried to learn. (Like, locative prepositions have tense in the predicate: hei = FUT.LOC, kei = PRES.LOC, and i = PST.LOC, but outside of the predicate, in adjunct positions, only i is used and ... I don't remember how to negate anything ever. I don't even know if I can say he X he Y for "An X is a Y". In any case, I don't think things are quite as simple as saying this is just a three way definiteness-specificity distinction. If I had another point with this, I can't really remember it right now and I'm falling asleep.

[Qwynegold]

Shhhhh!

Well, most of them are and that's how I've envisaged it, that writing is a specialised activity performed by the upper echelons. In fact, probably they'd get non-orcs to write for them since they don't have such fine motor control. The person who had the writing done, however, is still responsible for it.

I was thinking what if they left the writing for the nerds, who would have to rank themselves low? But then you'd still have the problem what if they rank two third persons wrong relative to one another.

I found glossing abbreviations on Wikipedia; SP, SPEC or SPECFC for specific, and NSP or NSPEC for non-specific.

[Imralu]

I was thinking what if they left the writing for the nerds, who would have to rank themselves low? But then you'd still have the problem what if they rank two third persons wrong relative to one another.

Yeah. I guess that could make being a scribe a pretty unappealing career since, whatever you write, power imbalances can shift later. Perhaps scribes would not be held accountable, but the people they write for would be.

I found glossing abbreviations on Wikipedia; SP, SPEC or SPECFC for specific, and NSP or NSPEC for non-specific.

Oh, yeah, that looks like it has been overhauled since I last looked at it. Much better! There is now no mention of C being used for complementisers now, as there use to be, and I've now being doing that for years and, urgh, gone.

[bradrn]

(sorry for taking so long to reply, your posts are always interestingly mind-bending and it takes a while for me to wrap my brain around them…)

OK… I’ve read through this several times now, and I think I understand it now. (As opposed to the Iliaqu post, which is still thoroughly confusing me.) So a content word with an article cannot act as a predicate, which means you can’t equate two definite arguments, which means you can’t distinguish definiteness in predicates.

First question: Why would you even need to make this distinction? Plenty of languages have no definiteness distinction at all, and get along just fine without a distinction between ‘the house is big’ and ‘the house is the big one’.

Of course, but that's like saying why have definiteness at all. Why have specificity at all. Many languages do without a lot of things. Balog could, at this stage, easily get along without definiteness/specificity marking in the predicate and, as I've said, I may simply find a biclausal solution to situations like that.

OK, I suppose that’s as good a reason as any other. I’d note that there’s nothing wrong with a biclausal construction — as you noted, English also uses highly different strategies for marking definiteness on predicates vs arguments. (Indeed, I’d suggest that a unified notion of ‘predicate’ is entirely inappropriate for English; there’s no reason Balog can’t be the same.)

Second question: Why can’t a content word with an article act as a predicate? A simple way of solving this problem would be to give such a clause an equative interpretation.

… Of course, I could change that rule and allow simple equative phrases where two articled phrases could constitute a clause …

This is exactly what I was suggesting.

… but since sentences can already consist of paratactic arrangements of clauses simply juxtaposed without any obligatory conjunctions, doing away with the marking that shows explicitly what is a predicate and what is a subject is bound to lead to some sentences that are difficult to parse.

I don’t see how this follows. The parsing rule is simple: if you see two marked phrases in a row, it’s an equative sentence. Otherwise, it isn’t.

And in any case, even if it ends up a bit ambiguous, well, most languages have no shortage of ambiguous parses. I don’t see it as a show-stopping problem.

The articles are first and foremost an indicator of a subject (contrasting with predicates which are zero-marked) The definiteness, specificity and rank distinctions, although enjoyable, were more or less a secondary addition to go along with the marking of a subject. Since there has to be something there, what other stuff can I do with it.

In that case, here’s an idea: make two series of articles, one of which can go on arguments, and another of which can go on predicates (and includes a zero term for the most common indefinite case). This should maintain a predicate/subject distinction while also allowing definiteness marking on the predicate.

Also, in Iliaqu, two articled phrases in the same case are interpreted as being linked by "and" I really find it very fun and practical and might use that in Balog as well.

Hmm… this gives me an idea. In some languages, predicative possession is expressed using conjunction: ‘X has Y’ is rendered as ‘X is with Y’, or even just ‘X and Y’. The obvious extension here is to nominal equation: you could simply declare that a clause consisting of only two conjoined articled phrases is given an equative interpretation. (Or, if you’re uncomfortable with argument-only clauses, you could add an existential: ‘there is X and Y’ = ‘X is Y’.)

BUT, I don't think it's anywhere near that simple. He is used in the predicate a great deal. It gets complicated by the fact that he is pretty restricted in where it can be used. For example, it can't be used after prepositions such as i (which is used to introduce a direct object, among many other things) and due to the funny sentence transformations required to even negate a sentence, such as with ehara, which more or less seems to function as a transitive verb, it means that a reliable definiteness distinction is often destroyed by negation.

Now this is fascinating. Are you saying that accusative arguments cannot be indefinite? If so, that would be a pretty interesting violation of the animacy hierarchy. I’d be curious to know how that situation came about.

I find it really quite funny when people act as if morphological complexity means a language is difficult, all those memes showing endless different forms of Finnish nounds or Spanish verbs. I've found Swahili's grammar to be quite simple to learn, but Māori just confuses me all the time and I honestly think it's the hardest language I've tried to learn. (Like, locative prepositions have tense in the predicate: hei = FUT.LOC, kei = PRES.LOC, and i = PST.LOC, but outside of the predicate, in adjunct positions, only i is used and ... I don't remember how to negate anything ever. I don't even know if I can say he X he Y for "An X is a Y". In any case, I don't think things are quite as simple as saying this is just a three way definiteness-specificity distinction. If I had another point with this, I can't really remember it right now and I'm falling asleep.

Oh, I completely agree. I’ve seen even experienced linguists (McWhorter comes to mind, though I may be misremembering) define linguistic complexity based purely on morphology. We don’t have a complete analysis of English grammar, even! This is something which irritates me immensely.

[Imralu]

(sorry for taking so long to reply, your posts are always interestingly mind-bending and it takes a while for me to wrap my brain around them…)

No worries, and thank you!

… but since sentences can already consist of paratactic arrangements of clauses simply juxtaposed without any obligatory conjunctions, doing away with the marking that shows explicitly what is a predicate and what is a subject is bound to lead to some sentences that are difficult to parse.

I don’t see how this follows. The parsing rule is simple: if you see two marked phrases in a row, it’s an equative sentence. Otherwise, it isn’t.

But next to another clause, there wouldn't just be two marked phrases. As yet, the structure of a clause is

[ CW | ART-CW]

... so if I allow clauses that are

[ART-CW | ART-CW]

that means I can have sentences that are

CW ART-CW ART-CW ART-CW

and more. With "and" indicated by juxtaposition of subjects and, that means clauses will not always be pairs and I'm also thinking of allowing clauses that are just [CW] with no subject, so if they can also be [ART-CW] it will be completely impossible to parse where clauses begin and end without some other kind of marking (intonation and commas for example). I simply don't like that strategy at all because I already have a perfectly good way to mark where clauses begin and end. Because sentences are already relying so much on the implied context between unlinked clauses, something which no-natlang does, as far as I know, to the point of not even having transitivity, and without having lexical word classes to tell you what's what syntactically, I can't see it doing anything other than falling apart into salad.

And in any case, even if it ends up a bit ambiguous, well, most languages have no shortage of ambiguous parses. I don’t see it as a show-stopping problem.

No, but there is always a way to disambiguate and, I mean, I've already got one. The marking is there to show what's the subject, not the predicate. Allowing the predicate to be marked the same way as a subject defeats the purpose of marking subject and predicate. Sure, it might be naturalistic to have kinks and lots of ad hoc solutions and disambiguation strategies, but, in case it's not obvious, I'm really not motivated that much by naturalism and I just prefer a system that is interesting and elegant to my own tastes.

In that case, here’s an idea: make two series of articles, one of which can go on arguments, and another of which can go on predicates (and includes a zero term for the most common indefinite case). This should maintain a predicate/subject distinction while also allowing definiteness marking on the predicate.

That's essentially my idea and the strategy I have in Iliaqu, except the predicate forms are formed regularly from the argument/adjunct forms by prefixation. The prefix can be regarded as more or less a copula. It's actually fairly similar to how ko works in Māori, although ko also depends on the definiteness, not just the presence of an article. (I don't think it was inspired by ko as I don't think I understood ko back when I came up with N-

Ko te whare te whare.
Nxu mala xu mala.
The house is the house.

???He whare he whare. *
(Ga) mala a mala. (Non-specific) **
A house is a house.

* I actually have no idea if this sentence is grammatical, as I mentioned above. I haven't come across sentence like this in Māori and textbooks are mysteriously silent on indefinite subjects in verbless clauses. Google translate can't be trusted (and doesn't use macrons ... it's complete lack of any diacritics in Romanian completely enrages me), but it pretty reliably makes the subject definite when equating it with an indefinite noun phrase in the predicate and ... urgh, I still just don't get it. It seems like articles are very restricted in how they can be used syntactically and that gets in the way of definiteness marking a lot of the time.

** The article in the predicate in this sentence is completely unnecessary because it's indefinite and nominative, but if it is there, it must be marked for the predicate (i.e. ga, not just a, which would instead mean "a house and a house")

Of course, for Balog, I could make the relationship between the subject articles and the predicate articles not simply a matter of adding a morpheme to the beginning that could be regarded as a copula, but something more irregular.

Also, in Iliaqu, two articled phrases in the same case are interpreted as being linked by "and" I really find it very fun and practical and might use that in Balog as well.

Hmm… this gives me an idea. In some languages, predicative possession is expressed using conjunction: ‘X has Y’ is rendered as ‘X is with Y’, or even just ‘X and Y’. The obvious extension here is to nominal equation: you could simply declare that a clause consisting of only two conjoined articled phrases is given an equative interpretation. (Or, if you’re uncomfortable with argument-only clauses, you could add an existential: ‘there is X and Y’ = ‘X is Y’.)

Yeah, that's not too far from strategies I'm thinking of. I thought of something like:

exist DEF(c)=be.man | be.same DEF(c)=apologise
The man is the one who apologised. / The one who apologised is the man

If the CW "be.same" gets seen as unnecessary and dropped, the sentence could, ordinarily, be understood as

exist DEF(c)=be.man DEF(c)=apologise
There is a man and an apologiser. / There's a man and someone apologised.

But "exist" there could very, very easily not be thought of as simply asserting existence of the two articled phrases but as equating them, in which case, it essentially just becomes a copula, to be, and that is, again, essentially exactly the same as the prenasalising copular prefix in Iliaqu.

COP DEF(c)=be.man DEF(c)=apologise
The one who apologised is the man.

BUT, I don't think it's anywhere near that simple. He is used in the predicate a great deal. It gets complicated by the fact that he is pretty restricted in where it can be used. For example, it can't be used after prepositions such as i (which is used to introduce a direct object, among many other things) and due to the funny sentence transformations required to even negate a sentence, such as with ehara, which more or less seems to function as a transitive verb, it means that a reliable definiteness distinction is often destroyed by negation.

Now this is fascinating. Are you saying that accusative arguments cannot be indefinite? If so, that would be a pretty interesting violation of the animacy hierarchy. I’d be curious to know how that situation came about.

Yeah, I think so, although (t)ētahi is a workaround that allows an accusative argument to be marked indefinite, although I get the impression that it's only used where it's seen as really necessary and simply using the definite article is the normal way it's done. I've seen he explained as only being able to be used as the first word of its phrase, so that means no prepositions before it.

Tētahi pretty transparently derives from te (DEF.SG) + tahi ("one"). The vowel lengthening is normal when it's compounded to other things. For example "this house" can be expressed as either te whare nei or tēnei whare. Demonstratives and possessives are formed from the article. What is interesting is that the plural of these is indicated by dropping the "t", which is not true for the plain old article use.

te whare nei → tēnei whare = this house
ngā whare nei → ēnei whare = these houses
te whare o Rāwiri → tō Rāwiri whare = Rāwiri's house
ngā whare o Rāwiri → ō Rāwiri whare = Rāwiri's houses
he/tētahi whare = a house
he/ētahi whare = houses

So, yeah, I've seen it indicated that he is non-specific and (t)ētahi is specific, but I'm not sure any of the articles can be definite purely in terms of definiteness, specificity and number since they get swapped around for each other depending on syntax as well.

I haven't looked it up, but I'm assuming that ngā is probably a more recent innovation to fill in a gap and the dropping the t of the determiners based on the article is an older strategy. I think cognates of ngā are restricted to Eastern Polynesian (which Māori belongs to in spite of being in the far southwest of Polynesia), although I only know of a cognate in Hawaiʻian. Tahitian and Rarotongan (Cook Islands Māori) have a different strategy.

ka hale → nā hale = the house → the houses (Hawaiʻian)
te fale → te mau fale = the house → the houses (Tahitian)
te fare → te au fare = the house → the houses (Rarotongan / Cook Islands Māori)

The specific article le is only used in the singular in Samoan, which is more distantly related, and the specific plural is marked by a null article (and the dual by la?). (The non-specific articles are se in singular and ni in plural and I'm assuming that se and le are cognates with Māori he (INDEF.SG+PL) and te (DEF.SG) respectively. Looking fast through a grammar of Samoan, I can't find anything that indicates restriction of the distribution of se.

I'm rambling now. Time to stop.

[Vardelm]

I can't see it doing anything other than falling apart into salad.

What did salad ever do to deserve such disrespect?

[Qwynegold]

Balog could, at this stage, easily get along without definiteness/specificity marking in the predicate

I just wanted to say that if you did decide to do this, you don't have to worry about it being a weird random feature (=bad conlanging). It would be weird, but there's a perfectly logical explanation for why that is.

[quinterbeck]

Balog could, at this stage, easily get along without definiteness/specificity marking in the predicate

IMO this is an interesting limitation that could lead to a creative solution. I haven't checked to see if you mentioned relative clauses above at all, but a predicate 3p pronoun modified by a relative clause whose subject would have been the predicate might be a possible construction that allows the speaker to state its defiteness/specificity.

[Raholeun]

Just chiming in to say I have been lurking this thread, like the discussion, but have little to contribute at this point.

[Imralu]

I can't see it doing anything other than falling apart into salad.

What did salad ever do to deserve such disrespect?

... prevented me from making friends

Balog could, at this stage, easily get along without definiteness/specificity marking in the predicate

I just wanted to say that if you did decide to do this, you don't have to worry about it being a weird random feature (=bad conlanging). It would be weird, but there's a perfectly logical explanation for why that is.

Yeah, thanks. I wouldn't mind if it's a weird, random feature that others regard as bad conlanging. I've never really been overly concerned with realism. I just like making syntactic structures that I find elegant and interesting. I'll only get frustrated if it turns out I've painted myself into a corner and made it hard to express certain concepts without some inelegant work-arounds that I find unsatisfying.

IMO this is an interesting limitation that could lead to a creative solution. I haven't checked to see if you mentioned relative clauses above at all, but a predicate 3p pronoun modified by a relative clause whose subject would have been the predicate might be a possible construction that allows the speaker to state its defiteness/specificity.

Thanks. It took me a while to understand what you mean, but now I think I get it. You mean, like, "The tree is it that falls," (contrasting with the indefinite predicate in "The tree falls,"), yeah?

The problem with that is that subject pronouns are formed of two elements: (1) the subject marker, which indicates a subject conveys rank, definiteness and specificity, and (2) a verbal proform conveying other information such as grammatical person and number, or (2b) the absence of a verbal proform, leaving the rank and definiteness of the subject marker to suffice.

2b cannot be used in the predicate as it's a null-morpheme (marked only by vowel lengthening of the subject marker, which also occurs before a verb/CW of a single consonant), there are verbal proforms, 2, that can appear in the predicate (although, I may complicate that - see below), but these don't convey definiteness other than in the first and second person where definiteness is a pragmatic implication. I'll backtrack a bit and get to that in a moment:

As far as relative clauses go, the subjects themselves (underlined) can be regarded as relative clauses:

Magaz iddauz. magaz fall.over i= DEF(e) dauz be.tree The tree(s) fall(s) over. / That/those which is/are (a) tree(s) fall(s) over.

Dauz immagaz. dauz be tree i= the (e-ranked) individual(s) that magaz fall over What falls over is a tree / are trees.

Basically, a subject consists of a subject marker (which can be thought of as a relativiser) + a verb phrase.

The verb phrase can be absent for pronominal cases, being instead replaced by a lengthening of the vowel: Magaz ii. "It falls over. / They fall over." Dauz ii. "It is a tree. / / They are trees."

These lengthened pronominal forms (which are most commonly used in the third person but can also be used in the first and second persons because of the assignment of rank, which should make it unambiguous) cannot occur in the predicate, but other pronominal forms that consist of verbs/CWs can, such as ž "be me", l "be you (sg)" etc.

As far as other available third person pronouns, there are the verbal/CW forms for number: the singulative n and the plurative m (only used when number information is regarded as necessary by the speaker), but these are derived from singulative and plural verbal/CW suffixes -an and -im respectively and I think I might decide that n and m are bound morphemes that are always attached to the subject markers, giving iin and iim for example, as an explicitly singular it and an explicitly plural they, but not able to occur as the head of predicates. (I also think I will probably do the same with other pronominal forms such as ž and l and give them longer forms that must be used in the predicate. E.g. Ž'oož "I am me", would, under this proposal, no longer be grammatical as ž is a bound morpheme occuring only on subject markers, but instead a verbal/CW derivation of it, using the singulative suffix, žan would be used instead, thus žan oož = "I am me", or, if using the full form in the subject as well: žan ožžan.)

There are also attributive means of subordinating verbs/CWs inside other verb clauses, such as with the equative we=. It only embeds a verb phrase within another verb phrase. The embedded verb phrase cannot take a subject. The implied subject is identical to what it attaches to in its matrix clause. They're underlined below.

Magaz iddauz weŋŋaž. magaz fall.over i= DEF(e)= dauz be.tree we= EQU= ŋaž be.tall The tall tree(s) fall(s) over.

Dauz immagaz weŋŋaž. dauz be.tree i= DEF(e)= magaz fall.over we= EQU= ŋaž be.tall The tall faller-overer(s) is/are tree(s). / The tall thing(s) falling over is/are tree(s).

Although very often, in sentences, there is little reason to subordinate and instead parataxis can be used with little difference in meaning. The following example simply contains two independent clauses. The rank of i= makes it unambiguous that the trees are what is being described as tall (unless there has been doubling-up of ranks in the discourse). Without any TAM information in this sentence, there is also no issue with the fact that the trees are no longer tall after having been cut down.

Magaz iddauz ŋaž ii. magaz fall.over i= DEF(e)= dauz be.tree ŋaž be.tall i= DEF(e) The tall tree(s) fall(s) over. / The tree(s) fall(s) over; it/they is/are tall.

So, as far as I understand your suggestion, it could be something like this, forgetting for a moment the fact that I think I'm going to make n be a bound morpheme. Let's pretend it's not and that it's an independent verbal proform:

? N'weddauz immagaz. N' be.SG [glossEQU]we=[/gloss] dauz be.tree i= DEF(e) magaz be.tree ? The thing(s) that fall(s) over are/is the tree(s).

I like that structure and I have previously very often conflated definiteness/specificity markers and 3rd person pronouns, however, in this lang, the pronominal (pro-verb(i?)al) forms such as n (assuming it is not a bound morpheme anyway), convey no information about definiteness/specificity. All of that is achieved with the subject marker, e.g. iin "it", iqiin "something (specific)", ihiin "something/anything (non-specific)". The n itself is completely neutral there and conveys only singularity. Even in the third-person proforms, definiteness is conveyed by the distinction on the subject marker between the definite i= and the indefinite iqi= and ihi=, not by the verbal part of the proform that can appear in the predicate. For first and second-person forms, for pragmatic reasons, "I" and "you" are by default definite, even when in the predicate, because ... well, isn't it obvious who I am if I'm speaking, or who you are if I'm speaking to you, however somewhere up-thread I think I mentioned the possibility of idiomatic uses for indefinite forms of first and second persons. These would only really able to be used in subjects, not in predicates, however, but that is unlikely to be an issue I need to find much of a work-around for.

Just chiming in to say I have been lurking this thread, like the discussion, but have little to contribute at this point.

Thank you! Lurking is welcome.

[bradrn]

After some research on Māori, I managed to turn up Chung, Mason & Milroy (1995), ‘On Maori He and the Uses of Indefinites’, which may or may not be of interest to you.

[Imralu]

After some research on Māori, I managed to turn up Chung, Mason & Milroy (1995), ‘On Maori He and the Uses of Indefinites’, which may or may not be of interest to you.

Thanks, but I can't get the full text. (Can you?) I found this one that I can of: MAORI HE REVISITED - Maria Polinski, 1992.

[bradrn]

After some research on Māori, I managed to turn up Chung, Mason & Milroy (1995), ‘On Maori He and the Uses of Indefinites’, which may or may not be of interest to you.

Thanks, but I can't get the full text. (Can you?)

I can see the text. (I sometimes forget how many things my university gives me access to.) But I managed to find what looks like a free version: http://www.jps.auckland.ac.nz/document/ ... 429-460/p1.

I found this one that I can of: MAORI HE REVISITED - Maria Polinski, 1992.

Looks interesting, thanks for the link!

[Raholeun]

I hope this language is not dormant?