So I thought I should start putting together the work I've been doing recently on the conlanging front and to put up a thread to show that I'm not totally useless in it. A lot of this is stuff has a pretty long history going back probably as far as early 2004. I'm not entirely sure of the actual date since I probably don't have most of my earliest notes left any more. The current work is heavily reworked however, and mostly represents what I've done during this year.
The setting is a geographically restricted conworld centred around a sound and with a humid boreal climate. This allows me to experiment with fun stuff such as language spread patterns and multiple loanword layers between both related and unrelated languages. It also leaves the wider world nicely mysterious as it would no doubt be experienced by the people living in the area. Technologically the most advanced cultures of the area are supposed to be preliterate but have trade connections outside their own area through which they are getting exposed to such things as glass and written language.
Linguistically the area is dominated by two unrelated language families one of which has established itself in the area during the last 1000 years or so. Thus this family also has less divergence between its members than the native family. To make things a bit more interesting, there are also signs of an even earlier language family that predates even the expansion the current "native" family. During the present day time of the description the family is only represented as one single small isolate language that's slowly dying out plus various toponyms and loans all over the place. There also has to be at least one external language that functions as a mediator for new cultural loans into the area (the language of the closest trading destination?) and I guess I have to work with it as well at some point.
The name of the area comes from Kišta, one of the central languages of the intruding family. In this language the sound defining the whole area is called Tenišše or simply Tenni. The first of the names includes a derivational suffix found in the names of many bodies of water and strictly refers to the sound itself. The latter name is the bare root of the name and is also used to refer to the larger area in general. This is an example of an old substrate toponym that is commonly found in the languages of the area and probably goes back to an ancient ancestor of the lone isolate language Nooníí kiskn. In this language the area around the sound is referred to as Tai, but it's unclear if this is an old retained name or has been transmitted as a loan via an unrelated language or even if the origin of the name seen in Kišta was a direct ancestor of Nooníí kiskn. Certainly the name hasn't retained any other meaning in Nooníí kiskn than what it has as a name of an area.
I'll be publishing stuff in this thread related to the project probably with a wildly irregular schedule. For the start you'll get to see the isolate language Nooníí kiskn as it at least started as a quicker project. I have much more ideas written down for Kišta but that language is much more of a victim of perfectionism and indecision so it'll need some more time for me to organise it all for presentation.
Languages of Tenni
Re: Languages of Tenni
Let's start then. In this post I'll describe the phonology of Nooníí kiskn as well as give a description of the most basic verb grammar.
Nooníí kiskn - Phonology
This is an isolate language that is slowly being left out of use and getting replaced by the languages surrounding it. It's spoken in only two villages by an order of 200 speakers all of whom are at least bilingual at this stage. As a result the language is full of recent loans. The thing is, these are from languages I haven't developed at all. Much of the vocabulary will be declared loaned later on, but in the case that providing a good loan etymology for some of the words I've used at some point proves too difficult I'll just state that what I'm using here is a very conservative lexicon.
The original idea behind the language was to go after the most minimal phonology that I could still feel comfortable with and be able to make have a naturalistic sound. I have posted a previous version of the same phonology in the post your phonology thread but that's already a bit outdated. In the end I settled to the following system of five consonants, three vowels and two tones:
/t k ʔ s n/
/a o i/
low vs. high tone
The places of articulation of the stops are dental, velar and glottal. The sibilant /s/ is prototypically an alveolar sibilant but when in contact with /t/ becomes dental as well. Also the nasal /n/ is underlyingly alveolar but it similarly has a dental allophone when preceding /t/ and a velar one when preceding /k/. When at the end of a word or preceding /ʔ/, the nasal is in free variation between the alveolar and velar realisations with the velar one gaining ground. It also dissimilates with the preceding /t k/ so that /tn/ is realised as [tŋ] and /kn/ as [kn].
Other consonant allophony consists of weakening of the obstruents under certain conditions. The dental stop /t/ never appears before the high vowel /i/. If it ends up in such a position due to morphology it undergoes a change into the sibilant /s/. The sibilant itself is pronounced increasingly weakly at the ends of words and often gets realised as a glottal fricative [h]. The velar stop is the weakest of all the obstruents and has two intervocalic approximant allophones. Between any vowel and /i/ it gets realised as [j] and between any vowel and /o/ as [w]. This is however the only allophony in the language that involves proper voicing and otherwise all the obstruents remain unvoiced, though lenis, in voiced environments.
The vowel system of the language consists of three vowels: a low vowel /a/, a back rounded vowel /o/ and a high vowel /i/. There's no conditioned allophony dealing with the vowel quality, but because of the sparsity of the system they have all pretty broad ranges of free variation around their nominal values. Besides this free variation the vowels /o i/ can loose their syllabicity and become [w j] respectively. This happens for /o/ following /a/ or the word boundary and preceding /a/ and for /i/ following /a o/ or the word boundary and preceding /a o/. Together with the allophony of /k/ this creates a healthy set of surface approximants into the language.
In addition to the vowels, also the nasal /n/ can become syllabic and is a valid syllable nucleus. This happens when it's not in contact with any of the vowels which would attract the syllabicity to themselves. In this position it's possible that the allophonic rules of assimilation to the POA of following /t k/ and dissimilation to the POA of preceding /t k/ can conflict. In such cases the progressive assimilation always wins so that /tnt/ is pronounced as [tn̩t] and /knk/ as [kŋ̍k].
The two surface tones are simply low vs. high and each vowel (plus the syllabic nasals) carries either of them on its own. Underlyingly there's also a third so called neutral tone that gets absorbed either into the low or the high tone. As only the vowels /a o i/ carry lexical tone the syllabic nasals are always treated as having the neutral tone.
Tone sandhi works so that a low tone always remains low and a high tone high except at the ends of words when preceded by two or more high toned morae. The neutral tone is low by default but become high when preceded by one or two high morae. When a vowel looses its syllabicity its tone is also lost.
The tone sandhi shows a somewhat different sense of word boundaries than what the allophony of the segmental phonemes obey. There are a number of particles that appear to be independent words but carry a neutral tone which has its value determined by the tones on the previous word as if they were all parts of one and the same word.
A few words about the phonotactics. Valid syllable nuclei consist of one or two vowels or a syllabic nasal. Bringing two identical vowels together creates a long vowel and two different ones a diphthong. Tones survive this treatment perfectly fine and can create contour tones in the long vowels and diphthongs. Triple identical vowels or double identical consonants are not allowed. If such thing were to form, they will simplify into long vowels and single consonants. Consonant clusters are tolerated fairly liberally. At the beginning and the end of a word clusters of up to three consonants are possible and between vowels up to five not counting the syllabic nasals, for example stkaa ("see"). Too complex clusters are broken up with a transition vowel /i/ (or /a/ after /t/ in order to avoid having to run through the change /t/ > /s/). The glottal stop can only appear preceding a vowel.
I'll use a mostly phonemic transcription for the language, e.i. using <t k ʔ s n a o i> for /t k ʔ s n a o i/ and acute accents for for the high toned vowels (<á ó í> for /á ó í/). I'll depart from this principle only in writing out the approximant allophones [w j] of /k o i/ as <w j> outside of strict phonemic notation. In the strict notation I'll write the vowels carrying the neutral tone with an under dot (<ạ ọ ị>). I may also use the under dot to highlight any syllabic nasals (<ṇ>) since these behave as if having the neutral tone.
Nooníí kiskn - Phonology
This is an isolate language that is slowly being left out of use and getting replaced by the languages surrounding it. It's spoken in only two villages by an order of 200 speakers all of whom are at least bilingual at this stage. As a result the language is full of recent loans. The thing is, these are from languages I haven't developed at all. Much of the vocabulary will be declared loaned later on, but in the case that providing a good loan etymology for some of the words I've used at some point proves too difficult I'll just state that what I'm using here is a very conservative lexicon.
The original idea behind the language was to go after the most minimal phonology that I could still feel comfortable with and be able to make have a naturalistic sound. I have posted a previous version of the same phonology in the post your phonology thread but that's already a bit outdated. In the end I settled to the following system of five consonants, three vowels and two tones:
/t k ʔ s n/
/a o i/
low vs. high tone
The places of articulation of the stops are dental, velar and glottal. The sibilant /s/ is prototypically an alveolar sibilant but when in contact with /t/ becomes dental as well. Also the nasal /n/ is underlyingly alveolar but it similarly has a dental allophone when preceding /t/ and a velar one when preceding /k/. When at the end of a word or preceding /ʔ/, the nasal is in free variation between the alveolar and velar realisations with the velar one gaining ground. It also dissimilates with the preceding /t k/ so that /tn/ is realised as [tŋ] and /kn/ as [kn].
Other consonant allophony consists of weakening of the obstruents under certain conditions. The dental stop /t/ never appears before the high vowel /i/. If it ends up in such a position due to morphology it undergoes a change into the sibilant /s/. The sibilant itself is pronounced increasingly weakly at the ends of words and often gets realised as a glottal fricative [h]. The velar stop is the weakest of all the obstruents and has two intervocalic approximant allophones. Between any vowel and /i/ it gets realised as [j] and between any vowel and /o/ as [w]. This is however the only allophony in the language that involves proper voicing and otherwise all the obstruents remain unvoiced, though lenis, in voiced environments.
The vowel system of the language consists of three vowels: a low vowel /a/, a back rounded vowel /o/ and a high vowel /i/. There's no conditioned allophony dealing with the vowel quality, but because of the sparsity of the system they have all pretty broad ranges of free variation around their nominal values. Besides this free variation the vowels /o i/ can loose their syllabicity and become [w j] respectively. This happens for /o/ following /a/ or the word boundary and preceding /a/ and for /i/ following /a o/ or the word boundary and preceding /a o/. Together with the allophony of /k/ this creates a healthy set of surface approximants into the language.
In addition to the vowels, also the nasal /n/ can become syllabic and is a valid syllable nucleus. This happens when it's not in contact with any of the vowels which would attract the syllabicity to themselves. In this position it's possible that the allophonic rules of assimilation to the POA of following /t k/ and dissimilation to the POA of preceding /t k/ can conflict. In such cases the progressive assimilation always wins so that /tnt/ is pronounced as [tn̩t] and /knk/ as [kŋ̍k].
The two surface tones are simply low vs. high and each vowel (plus the syllabic nasals) carries either of them on its own. Underlyingly there's also a third so called neutral tone that gets absorbed either into the low or the high tone. As only the vowels /a o i/ carry lexical tone the syllabic nasals are always treated as having the neutral tone.
Tone sandhi works so that a low tone always remains low and a high tone high except at the ends of words when preceded by two or more high toned morae. The neutral tone is low by default but become high when preceded by one or two high morae. When a vowel looses its syllabicity its tone is also lost.
The tone sandhi shows a somewhat different sense of word boundaries than what the allophony of the segmental phonemes obey. There are a number of particles that appear to be independent words but carry a neutral tone which has its value determined by the tones on the previous word as if they were all parts of one and the same word.
A few words about the phonotactics. Valid syllable nuclei consist of one or two vowels or a syllabic nasal. Bringing two identical vowels together creates a long vowel and two different ones a diphthong. Tones survive this treatment perfectly fine and can create contour tones in the long vowels and diphthongs. Triple identical vowels or double identical consonants are not allowed. If such thing were to form, they will simplify into long vowels and single consonants. Consonant clusters are tolerated fairly liberally. At the beginning and the end of a word clusters of up to three consonants are possible and between vowels up to five not counting the syllabic nasals, for example stkaa ("see"). Too complex clusters are broken up with a transition vowel /i/ (or /a/ after /t/ in order to avoid having to run through the change /t/ > /s/). The glottal stop can only appear preceding a vowel.
I'll use a mostly phonemic transcription for the language, e.i. using <t k ʔ s n a o i> for /t k ʔ s n a o i/ and acute accents for for the high toned vowels (<á ó í> for /á ó í/). I'll depart from this principle only in writing out the approximant allophones [w j] of /k o i/ as <w j> outside of strict phonemic notation. In the strict notation I'll write the vowels carrying the neutral tone with an under dot (<ạ ọ ị>). I may also use the under dot to highlight any syllabic nasals (<ṇ>) since these behave as if having the neutral tone.
Re: Languages of Tenni
Nooníí kiskn - Basic verbs
The verbs in Nooníí kiskn show both subject and object agreement. For transitive verbs the subject agreement is shown by pronominal proclitics
and the object agreement by object prefixes
as in
Wistkaa nani.
o=ị-stkaa n=ani
SG1.S=SG3.O-see REF=fox
"I saw a fox."
The subject pronominals can be seen to be clitics as they are omitted when an explicit third person subject is present,
Nani tastkaa.
n=ani ta-stkaa
REF-fox SG1.O-see
"A fox saw me."
and they also appear only once in complex predicated involving serialisation while the object prefixes get repeated on each verb,
Snaʔískí naʔa anini.
s=na-ʔískí na-ʔa ani-ni
SG3.S=PL3.O-hunt PL3.O-go fox-PL<RED>
"He hunts foxes. / He's a fox hunter."
Note that the SG3.O prefix carries a neutral tone. This means that it gets a high tone when combined with the PL3.S clitic,
Jósístkaa nani.
jós=ị-stkaa n=ani
PL3.S=SG3.O-see REF=fox
"They saw a fox."
Note also that the paucal subject pronominals behave as completely independent words. They are special also in being the only place within the language where the distinction between paucal and plural is made. Paucal is understood to refer to groups of three to five.
Looking at intransitive verbs shows that Nooníí kiskn has split ergativity. In the first and second persons the language follows a nominative alignment for most verbs and uses the subject clitics to mark person,
Okakíks ʔa naats.
o=kakíks ʔa naats
SG1.S=sleep go be.under
"I typically sleep soundly."
In third person the alignment switches to ergative and person marking is shown with the object prefixes,
Nakakíks nakakíks nanaats.
na-kakíks na-kakíks na-naats
PL3.O-sleep PL3.O-sleep PL3.O-be.under
"They are sleeping soundly."
Static verbs such as the copula differ from this pattern in being completely ergative,
Kik kaʔa.
ki-k kaʔa
SG2.O-COP old
"You are old."
Negation is done with the predicate final particle tạ and imperative and prohibitive with the particles kọ and ʔị while keeping the person marking intact,
Kik kaʔa ta.
ki-k kaʔa tạ
SG2.O-COP old NEG
"You are not old."
Skon ʔi tatók.
sk=on ʔị ta=tók
SG2.S=go.up PROHB DEF.PROX=tree
"Don't climb the tree!"
As can be guessed from the above examples, the word order of simple sentences is strictly SVO. The complement of the copula follows the copular verb and is considered an integral part of the predicate thus making the predicate final particles follow them. Other grammar bits that can be picked up from the examples is indicating frequentativity by serialising the verb with ʔa ("go") and progressivity by serialising it with itself. You can also see that the plural of nouns is indicated by reduplication using one of the many available patterns and that there are proclitic articles of which the referential article n= and the proximal definite article t(a)= are shown.
The verbs in Nooníí kiskn show both subject and object agreement. For transitive verbs the subject agreement is shown by pronominal proclitics
Code: Select all
SG DU PAUC PL
1 o= skik= skinʔa ʔan=
2 sk(i)= sik= jósínʔa kna=
3 s(i)= sik= jósínʔa jós=Code: Select all
SG DU PL
1 t(a)- ʔos- to-
2 k(i)- in- na-
3 ị- in- na-Wistkaa nani.
o=ị-stkaa n=ani
SG1.S=SG3.O-see REF=fox
"I saw a fox."
The subject pronominals can be seen to be clitics as they are omitted when an explicit third person subject is present,
Nani tastkaa.
n=ani ta-stkaa
REF-fox SG1.O-see
"A fox saw me."
and they also appear only once in complex predicated involving serialisation while the object prefixes get repeated on each verb,
Snaʔískí naʔa anini.
s=na-ʔískí na-ʔa ani-ni
SG3.S=PL3.O-hunt PL3.O-go fox-PL<RED>
"He hunts foxes. / He's a fox hunter."
Note that the SG3.O prefix carries a neutral tone. This means that it gets a high tone when combined with the PL3.S clitic,
Jósístkaa nani.
jós=ị-stkaa n=ani
PL3.S=SG3.O-see REF=fox
"They saw a fox."
Note also that the paucal subject pronominals behave as completely independent words. They are special also in being the only place within the language where the distinction between paucal and plural is made. Paucal is understood to refer to groups of three to five.
Looking at intransitive verbs shows that Nooníí kiskn has split ergativity. In the first and second persons the language follows a nominative alignment for most verbs and uses the subject clitics to mark person,
Okakíks ʔa naats.
o=kakíks ʔa naats
SG1.S=sleep go be.under
"I typically sleep soundly."
In third person the alignment switches to ergative and person marking is shown with the object prefixes,
Nakakíks nakakíks nanaats.
na-kakíks na-kakíks na-naats
PL3.O-sleep PL3.O-sleep PL3.O-be.under
"They are sleeping soundly."
Static verbs such as the copula differ from this pattern in being completely ergative,
Kik kaʔa.
ki-k kaʔa
SG2.O-COP old
"You are old."
Negation is done with the predicate final particle tạ and imperative and prohibitive with the particles kọ and ʔị while keeping the person marking intact,
Kik kaʔa ta.
ki-k kaʔa tạ
SG2.O-COP old NEG
"You are not old."
Skon ʔi tatók.
sk=on ʔị ta=tók
SG2.S=go.up PROHB DEF.PROX=tree
"Don't climb the tree!"
As can be guessed from the above examples, the word order of simple sentences is strictly SVO. The complement of the copula follows the copular verb and is considered an integral part of the predicate thus making the predicate final particles follow them. Other grammar bits that can be picked up from the examples is indicating frequentativity by serialising the verb with ʔa ("go") and progressivity by serialising it with itself. You can also see that the plural of nouns is indicated by reduplication using one of the many available patterns and that there are proclitic articles of which the referential article n= and the proximal definite article t(a)= are shown.
Last edited by gach on Thu Dec 12, 2013 8:12 am, edited 1 time in total.
-
roninbodhisattva
- Avisaru
- Posts: 568
- Joined: Sun Mar 05, 2006 11:50 pm
- Location: California
Re: Languages of Tenni
This language seems really cool. Question though. In the two examples involving `sleeping soundly', there are different agreement patterns on the word meaning 'be under'. Why does it should 3pl marking in the second case but not 1sg marking in the second.
Also, I'm a fan of the minimal inventory. I've been trying to do something similar and failing.
Also, I'm a fan of the minimal inventory. I've been trying to do something similar and failing.
Re: Languages of Tenni
I love it. I'm a big fan of minimal phonologies and I like that this lacks any lip closure at all. I would like it a little bit more if /s/ had a little more allophony, such as becoming pronounced as something like [ʃ] or [ɕ] when preceding /i/ (or even when not preceding /a/ or /o/), but that's just my aesthetic taste.
Glossing Abbreviations: COMP = comparative, C = complementiser, ACS / ICS = accessible / inaccessible, GDV = gerundive, SPEC / NSPC = specific / non-specific
________
MY MUSIC
________
MY MUSIC
Re: Languages of Tenni
Thanks for the reactions guys.
In the second example, nakakíks nakakíks nanaats, the subject is in the 3rd person which triggers the ergative agreement pattern and the use of the "object" prefixes. These are now true prefixes on the verb and appear on every single verb in the predicate (again highlighted).
There are still unresolved things with this. Firstly, I've now treated the local verbs such as naats ("go/be under") and on ("go/be over") as intransitives even if they have a direct local complement as in the last example of my previous post (kon ʔi tatók, "don't climb the tree"). I'm not sure if I'm going with this indefinitely of if I should change the cases with local complements into transitive. There's also the line to be drawn between the static and non-static verbs. As you can see kakíks ("sleep") isn't static for the purposes of this split. I don't think I'm making many more verbs than the copula and the existential verbs obey fully ergative "static" agreement.
This has to do with the switch between a nominative agreement pattern for the 1st/2nd persons and an ergative one for the 3rd person in non-static verbs. In the first of the examples, okakíks ʔa naats, the subject is first person meaning that person marking gets done with the "subject" clitics. These come only once per complex predicate (highlighted in above) so the latter two verbs of the chain get no person marking at all.roninbodhisattva wrote:In the two examples involving `sleeping soundly', there are different agreement patterns on the word meaning 'be under'. Why does it should 3pl marking in the second case but not 1sg marking in the second.
In the second example, nakakíks nakakíks nanaats, the subject is in the 3rd person which triggers the ergative agreement pattern and the use of the "object" prefixes. These are now true prefixes on the verb and appear on every single verb in the predicate (again highlighted).
There are still unresolved things with this. Firstly, I've now treated the local verbs such as naats ("go/be under") and on ("go/be over") as intransitives even if they have a direct local complement as in the last example of my previous post (kon ʔi tatók, "don't climb the tree"). I'm not sure if I'm going with this indefinitely of if I should change the cases with local complements into transitive. There's also the line to be drawn between the static and non-static verbs. As you can see kakíks ("sleep") isn't static for the purposes of this split. I don't think I'm making many more verbs than the copula and the existential verbs obey fully ergative "static" agreement.
I might consider something like this as a lately acquired pronunciation depending on the phonologies of the surrounding languages. For now I'm happy though with [j] being the only thing that falls between alveolar and velar.Imralu wrote:I would like it a little bit more if /s/ had a little more allophony, such as becoming pronounced as something like [ʃ] or [ɕ] when preceding /i/ (or even when not preceding /a/ or /o/), but that's just my aesthetic taste.
Re: Languages of Tenni
Lovely. No case marking on nouns? Are there adpositions? More serial verbs? Can you explain the articles a bit more?
vec
Re: Languages of Tenni
No case nor adpositions. The core arguments are marked with word order and the so called "subject" and "object" pronominals listed above. Adverbial information is indicated by various adverbs and serial verb constructions so yes, there will be more serialisation. Serialisation is also a major strategy for indicating modal and aspectual distinctions.vec wrote:Lovely. No case marking on nouns? Are there adpositions? More serial verbs? Can you explain the articles a bit more?
In addition to the two sets of pronominals connected to the predicate there's also a set of independent pronouns:
Code: Select all
SG DU PAUC PL
1 win skika skinʔa ʔani
2 skin sika jósínʔa ikna
3 sii sika jósínʔa jósThere are three proclitic articles in Nooníí kiskn, t(a)=, ski= and n=. The first two are definite articles. Their formal distinction is one of deixis, t(a)= is used for definite referents relatively near and ski= for the ones further away. The distinction gets analogised into pragmatic salience so that more pragmatically important definite referents get the proximal article while the less important ones get the distal one. Furthermore, the difference between t(a)= and ski= is very relative. Usually one of the definite NPs within a sentence gets marked by t(a)= while the others, if any, receive ski=. Using only ski= within a sentence signals strong distancing from the speaker.
The definite articles have a more restricted use in Noníí kiskn than for example in English. Only referents previously introduced to the discourse are understood to be definite. The sole exception to this are ʔanks ("sun"), tsiní ("moon") and a few place names that are commonly referred to with definite articles. Other NPs with definite referents, recoverable to the listener or not, are marked by the referential article n=. As you might expect, this is a pretty common sight in the language. Besides marking NPs it's used to construct complement clauses. I still have to think this critically through, but the idea is that you get a complete finite clause, prefix it with the referential article and out pops a complete complement clause ready for use. Non-referential NPs, i.e. ones that don't have a specific referent, don't receive any article. Many place names and all personal names also allow dropping the articles.
That's more or less it for the articles. I'll give you some actual example once I get to describing the NPs in more detail.
Re: Languages of Tenni
Nooníí kiskn - Quirks in person and number marking
As we saw above, the person marking of the intransitive verbs has a few twists in it. This isn't all and there are a two notable complications on the transitive side as well.
The modern pattern of participant marking on transitive verbs involves object marking with old prefixes and subject marking with much newer pronominal clitics. A group of verbs provides evidence that this in its prehistory the language had a system genuine agreement prefixes for subject as well as object marking. Not much of this is left and only a SG3 subject and general plural subject marker remain on the following 10 common verbs:
ʔóí - "give" (SG object)
kiji - "give" (non-SG object)
natsin - "take" (SG object)
í - "make" (SG object)
sí - "make" (non-SG object)
naí - "hold" (SG object)
as - "make noise"
sto - "hit"
sísk - "divide"
kots - "eat"
These verbs receive both the subject clitics and special subject prefixes that come between the object prefixes and the verb stem. For SG3 subjects the extra prefix is -s- and for plural subjects either -ʔ- or rising the tone of an immediately preceding vowel into a high tone. The -ʔ- allomorph of the PL.S prefix appears only on the vowel initial verbs í and as. The rest of the verbs take the floating tone allomorph except for ʔóí which doesn't take the plural subject prefix at all.
These prefixes are subject to a few more phonological restrictions. Firstly, the subject prefixes never use any transition vowel to separate themselves from the verb stem or the preceding object prefix and are thus subject to regular simplification of double consonants. In other words the SG3.S prefix never surfaces on sto or sísk or after DU1.O prefix ʔos-. The absence of any PL.S prefix on ʔóí can also be understood rising historically from this same process. Secondly, the floating high tone can only surface if it directly follows a lexical vowel (transition vowels don't count), i.e. when it comes after the SG3.O or PL[1-3].O prefixes.
Some conjugation examples of bare verbs displaying these features with SG1, SG3 and PL1 subjects are
wikots
o=ị-kots
SG1.S=SG3.O-eat
"I ate it"
siskots
s=ị-s-kots
SG3.S=SG3.O-SG3.S-eat
"he ate it"
ʔaníkots
ʔan=ị-[high]-kots
PL1.S=SG3.O-PL.S-eat
"we ate it"
ʔaniʔí
ʔan=–ị-ʔ-í
PL1.S=SG3.O-PL.S-make
"we made it"
stasto
s=ta-s-sto
SG3.S=SG1.O-SG3.S-hit
"he hit me"
In addition to this there are pairs of verbs of handling which have otherwise the same semantics but differ in the number of the implied object. For these pairs the singular object stem gets used with singular object prefixes and the non-singular stem with the the dual and plural prefixes. For examples I'll just complete the pairs that already appeared in the above list. There are more of these verb pairs than this.
ʔóí ~ kiji - "give" SG ~ non-SG
natsin ~ natsnis - "take" SG ~ non-SG
í ~ sí - "make" SG ~ non-SG
naí ~ ksak - "hold" SG ~ non-SG
As we saw above, the person marking of the intransitive verbs has a few twists in it. This isn't all and there are a two notable complications on the transitive side as well.
The modern pattern of participant marking on transitive verbs involves object marking with old prefixes and subject marking with much newer pronominal clitics. A group of verbs provides evidence that this in its prehistory the language had a system genuine agreement prefixes for subject as well as object marking. Not much of this is left and only a SG3 subject and general plural subject marker remain on the following 10 common verbs:
ʔóí - "give" (SG object)
kiji - "give" (non-SG object)
natsin - "take" (SG object)
í - "make" (SG object)
sí - "make" (non-SG object)
naí - "hold" (SG object)
as - "make noise"
sto - "hit"
sísk - "divide"
kots - "eat"
These verbs receive both the subject clitics and special subject prefixes that come between the object prefixes and the verb stem. For SG3 subjects the extra prefix is -s- and for plural subjects either -ʔ- or rising the tone of an immediately preceding vowel into a high tone. The -ʔ- allomorph of the PL.S prefix appears only on the vowel initial verbs í and as. The rest of the verbs take the floating tone allomorph except for ʔóí which doesn't take the plural subject prefix at all.
These prefixes are subject to a few more phonological restrictions. Firstly, the subject prefixes never use any transition vowel to separate themselves from the verb stem or the preceding object prefix and are thus subject to regular simplification of double consonants. In other words the SG3.S prefix never surfaces on sto or sísk or after DU1.O prefix ʔos-. The absence of any PL.S prefix on ʔóí can also be understood rising historically from this same process. Secondly, the floating high tone can only surface if it directly follows a lexical vowel (transition vowels don't count), i.e. when it comes after the SG3.O or PL[1-3].O prefixes.
Some conjugation examples of bare verbs displaying these features with SG1, SG3 and PL1 subjects are
wikots
o=ị-kots
SG1.S=SG3.O-eat
"I ate it"
siskots
s=ị-s-kots
SG3.S=SG3.O-SG3.S-eat
"he ate it"
ʔaníkots
ʔan=ị-[high]-kots
PL1.S=SG3.O-PL.S-eat
"we ate it"
ʔaniʔí
ʔan=–ị-ʔ-í
PL1.S=SG3.O-PL.S-make
"we made it"
stasto
s=ta-s-sto
SG3.S=SG1.O-SG3.S-hit
"he hit me"
In addition to this there are pairs of verbs of handling which have otherwise the same semantics but differ in the number of the implied object. For these pairs the singular object stem gets used with singular object prefixes and the non-singular stem with the the dual and plural prefixes. For examples I'll just complete the pairs that already appeared in the above list. There are more of these verb pairs than this.
ʔóí ~ kiji - "give" SG ~ non-SG
natsin ~ natsnis - "take" SG ~ non-SG
í ~ sí - "make" SG ~ non-SG
naí ~ ksak - "hold" SG ~ non-SG
Last edited by gach on Fri Dec 13, 2013 2:38 am, edited 1 time in total.
-
roninbodhisattva
- Avisaru
- Posts: 568
- Joined: Sun Mar 05, 2006 11:50 pm
- Location: California
Re: Languages of Tenni
This is what I figured was going on. That's also a nice argument for the morphosyntactic distinction between clitics and agreement in this language. Also, the prefixes on every verb in the intransitive cases is reminiscent of person marking in Tariana serial verbs.gach wrote:Thanks for the reactions guys.
This has to do with the switch between a nominative agreement pattern for the 1st/2nd persons and an ergative one for the 3rd person in non-static verbs. In the first of the examples, okakíks ʔa naats, the subject is first person meaning that person marking gets done with the "subject" clitics. These come only once per complex predicate (highlighted in above) so the latter two verbs of the chain get no person marking at all.roninbodhisattva wrote:In the two examples involving `sleeping soundly', there are different agreement patterns on the word meaning 'be under'. Why does it should 3pl marking in the second case but not 1sg marking in the second.
In the second example, nakakíks nakakíks nanaats, the subject is in the 3rd person which triggers the ergative agreement pattern and the use of the "object" prefixes. These are now true prefixes on the verb and appear on every single verb in the predicate (again highlighted).
A question, though. Is it that the transitivity of the entire complex predicate gets determined by some specific verb in the construction, or do certain constructions have certain transitivities?
This is an interesting question. I don't see why you couldn't treat them as a certain class of intransitives that take locative complements like you're doing without a problem.There are still unresolved things with this. Firstly, I've now treated the local verbs such as naats ("go/be under") and on ("go/be over") as intransitives even if they have a direct local complement as in the last example of my previous post (kon ʔi tatók, "don't climb the tree"). I'm not sure if I'm going with this indefinitely of if I should change the cases with local complements into transitive. There's also the line to be drawn between the static and non-static verbs. As you can see kakíks ("sleep") isn't static for the purposes of this split. I don't think I'm making many more verbs than the copula and the existential verbs obey fully ergative "static" agreement.
Re: Languages of Tenni
I've been thinking about this for a while. Turning a predicate transitive by adding a transitive verb into it will impose transitive agreement in the hole of it. Then again, I think that all of the transitivity altering serialisation patterns that the language has will be very formalised and part of the standard grammar of the verbs. Thus it's equally correct to say that the change in transitivity is brought by the use of these constructions and not any sing verb in them.roninbodhisattva wrote:A question, though. Is it that the transitivity of the entire complex predicate gets determined by some specific verb in the construction, or do certain constructions have certain transitivities?
This is what I'm now going for. It makes it much easier to handle the possibility of either dropping or including the local complement for these verbs as well as adding local information to transitive clauses.This is an interesting question. I don't see why you couldn't treat them as a certain class of intransitives that take locative complements like you're doing without a problem.
Re: Languages of Tenni
Nooníí kiskn - Final things to add to the verb morphology
Thus far we've seen verb morphology that's related to the marking of the person and number of the core arguments. In addition to this there isn't much else morphology on them and even the remaining bits have to deal with coding the argument structure.
The verbs in Nooníí kiskn vary a bit in how strictly they follow their default transitivity but in general you should never assume that a verb could be used both as a transitive and an intransitive verb without any additional modifications. In particular transitive verbs are very picky in having to be used with transitive conjugation. If a truly intransitive version of the verb makes sense and is needed, it's formed by reduplicating the final syllable of the verb root. Compare
Siskots nkoo.
s=ị-s-kots n=koo
SG3.S=SG3.O-SG3.S-eat REF=bread
"He ate bread."
Ikotsots ʔawóó.
ị=kotsots ʔawóó
SG3.S=eat<INTR> much
"He eats a lot."
where the the latter example refers to the process of eating itself instead of an action preformed to some object. The reduplication pattern consists in general of the vowels of the last syllable plus the final consonants. With verbs ending in a vowel also the consonant immediately preceding the final vowels is included as in stkaa ("see.TR") ~ stkakaa ("see.INTR, look"). Shortening long vowels before the reduplication is also a common pattern as seen here, though it only happens on certain more ancient roots.
The reduplication to get intransitives from transitives doesn't apply to all verbs and some drag object marking everywhere they go,
Skinaí kó tans!
sk=ị-naí kọ tans
SG2.S=SG3.O-hold IMP firm
"Hold tight!"
Here no object is mentioned and not necessarily even referenced anaphorically and the object prefix ị- just tags along as dummy agreement. The reason why this pattern is grammatical instead of using reduplication with no object marking is that the action of "holding something" is though to be nonsensical without having it refer to some object.
Intransitive verbs are more liberal in how they are used and in many cases it's possible to add an object to an intransitive verb and treat it transitively. Even still, there are two derivational suffixes that are used to derive transitives from intransitives. The more common one is -nạ. This is a very general transitivising affix that appears both in adding objects to intransitive verbs as well as in creating a bunch of applicatives from already transitive verbs,
ʔaníka.
ʔan=níka
PL1.S=sing
"We are singing."
ʔaniníkana nkokóna.
ʔan=ị-níka-nạ n=kokóna
PL1.S=SG3.O-sing-TR REF=story
"We are singing a story."
Wiknakana ʔai.
O=ị-knaka-na ʔai
SG1.S=SG3.O-shoot arrow
"I'm shooting (with) arrows."
Using -nạ in the applicative sense always overrides the original object slot.
The other transitivising affix is -ạst which is occasionally used to derive comitative applicatives referring to human company,
Ntanisk tníkaast.
n=ta-nisk t-níka-ạst
REF=SG1.POSS-sister SG1.O-sing-COM
"My sister is singing with me."
Neither of the transitivising suffixes are very productive and both whether they are used and what's the resulting effect of -nạ has to learned separately for every verb.
The final part of verb morphology is actually inflectional and fully productive. This is the -(ị/ạ)s same subject suffix which is used to signal that a verb is coreferential with the subject of a previous verb. The (ị/ạ) in the affix is just the regular transition vowel which in the case of this affix is used always when following a consonant. More of this affix will come later, but one of its uses is binding a dislocated verb to a serial verb complex when it's cut away from it by a full NP,
Ostkakaa on tatók.
o=stkakaa on ta=tók
SG1.S=look be.up DEF.PROX=tree
"I looked up to the tree."
Okstkaa skin onis tatók.
o=k-stkaa skin on-is ta=tók
SG1.S=SG2.O-see you be.up-SS PROX.DEF=tree
"I saw you up in the tree."
In case you are wondering, the verb on is marked with the SS suffix because as a local/directional verb it's primarily modifying the main verb stkaa.
Thus far we've seen verb morphology that's related to the marking of the person and number of the core arguments. In addition to this there isn't much else morphology on them and even the remaining bits have to deal with coding the argument structure.
The verbs in Nooníí kiskn vary a bit in how strictly they follow their default transitivity but in general you should never assume that a verb could be used both as a transitive and an intransitive verb without any additional modifications. In particular transitive verbs are very picky in having to be used with transitive conjugation. If a truly intransitive version of the verb makes sense and is needed, it's formed by reduplicating the final syllable of the verb root. Compare
Siskots nkoo.
s=ị-s-kots n=koo
SG3.S=SG3.O-SG3.S-eat REF=bread
"He ate bread."
Ikotsots ʔawóó.
ị=kotsots ʔawóó
SG3.S=eat<INTR> much
"He eats a lot."
where the the latter example refers to the process of eating itself instead of an action preformed to some object. The reduplication pattern consists in general of the vowels of the last syllable plus the final consonants. With verbs ending in a vowel also the consonant immediately preceding the final vowels is included as in stkaa ("see.TR") ~ stkakaa ("see.INTR, look"). Shortening long vowels before the reduplication is also a common pattern as seen here, though it only happens on certain more ancient roots.
The reduplication to get intransitives from transitives doesn't apply to all verbs and some drag object marking everywhere they go,
Skinaí kó tans!
sk=ị-naí kọ tans
SG2.S=SG3.O-hold IMP firm
"Hold tight!"
Here no object is mentioned and not necessarily even referenced anaphorically and the object prefix ị- just tags along as dummy agreement. The reason why this pattern is grammatical instead of using reduplication with no object marking is that the action of "holding something" is though to be nonsensical without having it refer to some object.
Intransitive verbs are more liberal in how they are used and in many cases it's possible to add an object to an intransitive verb and treat it transitively. Even still, there are two derivational suffixes that are used to derive transitives from intransitives. The more common one is -nạ. This is a very general transitivising affix that appears both in adding objects to intransitive verbs as well as in creating a bunch of applicatives from already transitive verbs,
ʔaníka.
ʔan=níka
PL1.S=sing
"We are singing."
ʔaniníkana nkokóna.
ʔan=ị-níka-nạ n=kokóna
PL1.S=SG3.O-sing-TR REF=story
"We are singing a story."
Wiknakana ʔai.
O=ị-knaka-na ʔai
SG1.S=SG3.O-shoot arrow
"I'm shooting (with) arrows."
Using -nạ in the applicative sense always overrides the original object slot.
The other transitivising affix is -ạst which is occasionally used to derive comitative applicatives referring to human company,
Ntanisk tníkaast.
n=ta-nisk t-níka-ạst
REF=SG1.POSS-sister SG1.O-sing-COM
"My sister is singing with me."
Neither of the transitivising suffixes are very productive and both whether they are used and what's the resulting effect of -nạ has to learned separately for every verb.
The final part of verb morphology is actually inflectional and fully productive. This is the -(ị/ạ)s same subject suffix which is used to signal that a verb is coreferential with the subject of a previous verb. The (ị/ạ) in the affix is just the regular transition vowel which in the case of this affix is used always when following a consonant. More of this affix will come later, but one of its uses is binding a dislocated verb to a serial verb complex when it's cut away from it by a full NP,
Ostkakaa on tatók.
o=stkakaa on ta=tók
SG1.S=look be.up DEF.PROX=tree
"I looked up to the tree."
Okstkaa skin onis tatók.
o=k-stkaa skin on-is ta=tók
SG1.S=SG2.O-see you be.up-SS PROX.DEF=tree
"I saw you up in the tree."
In case you are wondering, the verb on is marked with the SS suffix because as a local/directional verb it's primarily modifying the main verb stkaa.
Last edited by gach on Fri Jan 10, 2014 2:35 pm, edited 1 time in total.
Re: Languages of Tenni
I'm still not getting why it's onis – since the subject of stkaa is 1SG, not 2SG...
vec
Re: Languages of Tenni
You have to understand the use of main verb + directional verb as do in/towards the location of as in "see in ...". In other words it's an adverbial extension to what the subject is doing. The directional verb is cut off from the main predicate by the subject NP so it has to be connected to it using some special means. This is done with the same subject suffix as if the directional verb were still part of the same predicate and controlled by its subject. The name might be a misnomer for the suffix but what can you do.
Re: Languages of Tenni
Back on the lathe on this work. Let's also try out the new formatting.
Nooníí kiskn - existential verbs
Apart from the directional verbs another class of verbs important for indicating location and related categories are the existential verbs. These appear in existential and locational clauses as well as in possessive constructions.
There are six existential verbs in Nooníí kiskn and they form a classificatory set based of the posture and quality of the subject:
tón, for "compact" not too extended subjects, also for "sit"
káá, for vertically standing subjects, also "stand"
tai, for horizontally extended subjects, also "lie"
ʔís, for very small subjects with no regard of their shape
naka, for uncountable substances
tsít, for compact subjects coming in coherent groups
Like the copula, these verbs obey fully ergative alignment and take the object prefixes to mark their subjects. Location is specified by serialising a directional verb to the existential verb. Predicative possession is essentially a form of the locational clause using the directional verb kis ("go beside"). The location and possessor are nominal adjuncts (not core arguments) of the directional verbs and directly follow them. Some examples are,
Nsnisk iton.
"There is a stone."
Takáá taa.
"I'm standing here"
(Note that demonstratives can indicate location without a directional verb.)
Sʔisís natsít najis tanatans ʔíí.
"The merchant has beads."
Maybe a more natural way to construct a possessive clause would be to topicalise the possessor. This is easily achieved by fronting the possessor NP to the beginning of the clause and filling its regular position directional position by a corresponding pronoun,
Tanatans ʔíí sʔisís natsít najis sii.
"The merchant has beads."
The last of the existential verbs, tsít, is typically restricted to coherent groups of objects that are easily handled. However, it also appears now and then as a generic non-singular stem for all of the existentials for countables as can be seen in the following example pair,
Ikaka natón naas nost.
Ikaka natsít naas nost.
"There are fish in the water"
Nooníí kiskn - existential verbs
Apart from the directional verbs another class of verbs important for indicating location and related categories are the existential verbs. These appear in existential and locational clauses as well as in possessive constructions.
There are six existential verbs in Nooníí kiskn and they form a classificatory set based of the posture and quality of the subject:
tón, for "compact" not too extended subjects, also for "sit"
káá, for vertically standing subjects, also "stand"
tai, for horizontally extended subjects, also "lie"
ʔís, for very small subjects with no regard of their shape
naka, for uncountable substances
tsít, for compact subjects coming in coherent groups
Like the copula, these verbs obey fully ergative alignment and take the object prefixes to mark their subjects. Location is specified by serialising a directional verb to the existential verb. Predicative possession is essentially a form of the locational clause using the directional verb kis ("go beside"). The location and possessor are nominal adjuncts (not core arguments) of the directional verbs and directly follow them. Some examples are,
Nsnisk iton.
| n=skisk |
| REF-stone |
| ị-ton |
| SG3.O-exist.COMP |
"There is a stone."
Takáá taa.
| ta-káá |
| SG1.O-exist.VERT |
| taa |
| PROX |
"I'm standing here"
(Note that demonstratives can indicate location without a directional verb.)
Sʔisís natsít najis tanatans ʔíí.
| sʔisí-s |
| bead-PL |
| na-tsít |
| PL3.O-exist.GROUP |
| na-kis |
| PL3.O-go.beside |
| ta=natans ʔíí |
| DEF.PROX-merchant |
"The merchant has beads."
Maybe a more natural way to construct a possessive clause would be to topicalise the possessor. This is easily achieved by fronting the possessor NP to the beginning of the clause and filling its regular position directional position by a corresponding pronoun,
Tanatans ʔíí sʔisís natsít najis sii.
| ta=natans ʔíí |
| DEF.PROX-merchant |
| sʔisí-s |
| bead-PL |
| na-tsít |
| PL3.O-exist.GROUP |
| na-kis |
| PL3.O-go.beside |
| sii |
| SG3 |
"The merchant has beads."
The last of the existential verbs, tsít, is typically restricted to coherent groups of objects that are easily handled. However, it also appears now and then as a generic non-singular stem for all of the existentials for countables as can be seen in the following example pair,
Ikaka natón naas nost.
| ika-ka |
| fish-PL |
| na-tón |
| PL3.O-exist.COMP |
| na-as |
| PL3.O-go.in |
| n=ost |
| REF-water |
Ikaka natsít naas nost.
| ika-ka |
| fish-PL |
| na-tsít |
| PL3.O-exist.GROUP |
| na-as |
| PL3.O-go.in |
| n=ost |
| REF-water |
"There are fish in the water"
Re: Languages of Tenni
Well, if you mean the possession is location metaphor then maybe, but otherwise I'd say there's no resemblance at all. And possession as location is hardly restricted to Uralic.
The choice for this structure for predicate possession came from the idea of building inalienable attribute possession on the existentials as well. That's a very marked feature so having the existentials appear also in predicate possession was a natural way to mild things down. I also considered constructing predicate possession as an existential phrase, which would have fit as well, but that's going to be a feature of the Kištaic languages and I don't want there to be a connection here.
The third language family, still annoyingly lacking a proper name, will sport a lot of have. This family is in close contact with Nooníí kiskn so I might add a similar construction of the form X does Y for X has Y as an intrusive strategy for forming possessive clauses. This would be a symptom of the language death Nooníí kiskn is heading towards. Compare the traditional construction
Sʔisís natsít najis tanatans ʔíí.
"The merchant has beads."
with the intrusive one
Tanatans ʔíí iʔas sʔisís.
"The merchant has beads."
Note that the object like modifiers the verb ʔas gets are just post verbal nominal adjuncts that don't induce agreement on the verb. The verb is strictly intransitive.
I side thought that occurred to me while writing this is that I'd very much like to have bare verb roots in the language. That means getting rid of the 3rd person agreement morphology of intransitive verbs or more specifically changing all verbs except the copula and the existentials into a fully accusative pattern. I still have to think about this for a while but it would make the language a bit more flowing.
The choice for this structure for predicate possession came from the idea of building inalienable attribute possession on the existentials as well. That's a very marked feature so having the existentials appear also in predicate possession was a natural way to mild things down. I also considered constructing predicate possession as an existential phrase, which would have fit as well, but that's going to be a feature of the Kištaic languages and I don't want there to be a connection here.
The third language family, still annoyingly lacking a proper name, will sport a lot of have. This family is in close contact with Nooníí kiskn so I might add a similar construction of the form X does Y for X has Y as an intrusive strategy for forming possessive clauses. This would be a symptom of the language death Nooníí kiskn is heading towards. Compare the traditional construction
Sʔisís natsít najis tanatans ʔíí.
| sʔisí-s |
| bead-PL |
| na-tsít |
| PL3.O-exist.GROUP |
| na-kis |
| PL3.O-go.beside |
| ta=natans ʔíí |
| DEF.PROX-merchant |
"The merchant has beads."
with the intrusive one
Tanatans ʔíí iʔas sʔisís.
| ta=natans ʔíí |
| DEF.PROX-merchant |
| ị-ʔas |
| SG3.O-do |
| sʔisí-s |
| bead-PL |
"The merchant has beads."
Note that the object like modifiers the verb ʔas gets are just post verbal nominal adjuncts that don't induce agreement on the verb. The verb is strictly intransitive.
I side thought that occurred to me while writing this is that I'd very much like to have bare verb roots in the language. That means getting rid of the 3rd person agreement morphology of intransitive verbs or more specifically changing all verbs except the copula and the existentials into a fully accusative pattern. I still have to think about this for a while but it would make the language a bit more flowing.
Re: Languages of Tenni
Nooníí kiskn - Plural formation
The core nominal morphology of Nooníí kiskn consists of plural formation and bound pronominal affixes for inalienable possession. You might also consider the articles here. Although they are NP initial clitics and not bound to the noun (other determiners come between the two), they are always phonologically part of the following word whatever it is.
We've already seen some plurals in the above examples. They are all formed through reduplication, though there are several patterns in which to do this. The most productive one merely involves the reduplication of the last consonant of the noun root with a possible preceding epenthetic vowel, -(i)Cn. This is also the pattern that gets used for most recent loans. Nouns using this pattern include kowóna "story" > kowóna-n "story-PL", sónánka "dog" > sónánka-k "dog-PL" and tstak "year" > tstak-ik "year-PL". The epenthetic vowel appears on consonant final stems using this plural pattern to break up a double consonant and prevent the plural formant from reducing to zero.
For shorter nouns there are two other common patterns. Vowel final nouns can form plural using the pattern -CnV(V), i.e. reduplicating the last consonant of the stem and all following vowels. Such words are ani "fox" > ani-ni "fox-PL" and ika "fish" > ika-ka. For consonant final nouns the pattern doesn't include the consonant preceding the last vowels of the stem but includes all the coda consonants following them, -V(V)C(C...). One example of this pattern is snisk "stone" > snisk-isk "stone-PL".
A complication for these patterns is that some times the original final syllable shortens its long vowel or looses an off-glide as in koo "bread" > ko-woo and ʔai "arrow" > ʔa-ʔai "arrow-PL". This is all due to diachronic simplification but it's not possible to predict which words have this feature and on which anything similar has been levelled away. Note the regular change /k/ > [w] in kowoo. Similar both regular and historical alternations can complicate the surface plural patterns and make them appear even less regular.
Lastly also full reduplication is attested on some monosyllabic nouns such as tók "tree" > tók-tók "tree-PL".
The core nominal morphology of Nooníí kiskn consists of plural formation and bound pronominal affixes for inalienable possession. You might also consider the articles here. Although they are NP initial clitics and not bound to the noun (other determiners come between the two), they are always phonologically part of the following word whatever it is.
We've already seen some plurals in the above examples. They are all formed through reduplication, though there are several patterns in which to do this. The most productive one merely involves the reduplication of the last consonant of the noun root with a possible preceding epenthetic vowel, -(i)Cn. This is also the pattern that gets used for most recent loans. Nouns using this pattern include kowóna "story" > kowóna-n "story-PL", sónánka "dog" > sónánka-k "dog-PL" and tstak "year" > tstak-ik "year-PL". The epenthetic vowel appears on consonant final stems using this plural pattern to break up a double consonant and prevent the plural formant from reducing to zero.
For shorter nouns there are two other common patterns. Vowel final nouns can form plural using the pattern -CnV(V), i.e. reduplicating the last consonant of the stem and all following vowels. Such words are ani "fox" > ani-ni "fox-PL" and ika "fish" > ika-ka. For consonant final nouns the pattern doesn't include the consonant preceding the last vowels of the stem but includes all the coda consonants following them, -V(V)C(C...). One example of this pattern is snisk "stone" > snisk-isk "stone-PL".
A complication for these patterns is that some times the original final syllable shortens its long vowel or looses an off-glide as in koo "bread" > ko-woo and ʔai "arrow" > ʔa-ʔai "arrow-PL". This is all due to diachronic simplification but it's not possible to predict which words have this feature and on which anything similar has been levelled away. Note the regular change /k/ > [w] in kowoo. Similar both regular and historical alternations can complicate the surface plural patterns and make them appear even less regular.
Lastly also full reduplication is attested on some monosyllabic nouns such as tók "tree" > tók-tók "tree-PL".
Last edited by gach on Sat Jan 11, 2014 4:41 pm, edited 1 time in total.
-
ObsequiousNewt
- Avisaru
- Posts: 434
- Joined: Fri Oct 18, 2013 5:05 pm
- Location: /ˈaɪ̯əwʌ/
Re: Languages of Tenni
Wait, "tree" or "three"?gach wrote:Lastly also full reduplication is attested on some monosyllabic nouns such as tók "tree" > tók-tók "three-PL".
퇎
Ο ορανς τα ανα̨ριθομον ϝερρον εͱεν ανθροποτροφον.
Το̨ ανθροπς αυ̨τ εκψον επ αθο̨ οραναμο̨ϝον.
Θαιν. Θαιν. Θαιν. Θαιν. Θαιν. Θαιν. Θαιν.
Ο ορανς τα ανα̨ριθομον ϝερρον εͱεν ανθροποτροφον.
Το̨ ανθροπς αυ̨τ εκψον επ αθο̨ οραναμο̨ϝον.
Θαιν. Θαιν. Θαιν. Θαιν. Θαιν. Θαιν. Θαιν.
Re: Languages of Tenni
Fixed, thanks for noting it, it should be "tree". I was in hurry for having to leave to get my skis from a friend's place. That's also why I didn't write anything more about possession and the articles in that post.ObsequiousNewt wrote:Wait, "tree" or "three"?gach wrote:Lastly also full reduplication is attested on some monosyllabic nouns such as tók "tree" > tók-tók "three-PL".
I hate typos that produce valid English words and thus get unnoticed by spell checking. Especially when you leave some of those in a research paper and get comments from your referee implying that you are stupid because you can't even write.
Re: Languages of Tenni
Nooníí kiskn - Alienable possession and basic structure of NPs
A while ago we got familiar with the existential verbs and saw that they are closely connected to expressing possession. In the case of alienable possession this connection carries also over to attributive possession. For future reference, there are also some nouns that can or must receive inalienable possession via pronominal prefixes, but this group is very limited and I'll return to it later.
The idea behind the alienable attributive possessive phrases is a continuation of the existential possessive clauses. For attributive possession we just relativise the possessive clause and attach it following the possessed noun. Relativisation is done simply by prefixing the finite clause with the referential article n=. For attributive possession there's an additional simplification as the directional verb kis is dropped from the relative clause. As a result we get from the previous predicative example
Tanatans ʔíí sʔisís natsít najis sii.
"The merchant has beads."
the attributive counterpart
nsʔisís natsít tanatans ʔíí
"the merchant's beads"
A possessed noun always has a specific referent so also the full possessed NP must be preceded by one of the articles. Unfortunately here we don't see the referential article introducing the relative clause as on the surface it merges together with the PL3.O prefix of the existential verb. It remains visible in the case of singular or dual possessed nouns,
nsónánka nitón win
"my dog"
Comparing this to the previous example also demonstrates the classificatory function of the existential verbs.
So this already demonstrates quite nicely different aspects of the NP structure. Both adjectives and relative clauses follow nouns. Complex attributes tend away from the head noun so relative clauses follow adjectives,
nsónánka kaʔa nitón win
"my old dog"
If the NP has a specific referent, definite or not, it must be prefixed by one of the three articles, referential n=, proximal definite t(a)= or distal definite ski=. Lack of an article always signals generic reference. Sʔisí-s (bead-PL) in the first example above is one such case as the speaker has no specific beads in mind, just the idea that the merchant has some of them available.
Besides the articles there are some other determiners or determiner like words that also precede the noun. These include the demonstratives,
taa, PROX
kaa, MEDIUM PROX, same level
ʔot, MEDIUM PROX, below
ját, MEDIUM PROX, above
nan, DIST, same level
tao, DIST, below
sin, DIST, above
skín, DIST, invisible/anaphoric
numerals (giving here the cardinals 1-10),
1 jók
2 tiní
3 wan
4 ʔisí
5 kiis
6 onaasoas
7 sonasi
8 tniʔiis
9 tnsi
10 ksis
as well as some other determiners,
tanʔi, some
tó, basic interrogative for WH-questions
etc.
None of these override the need for an article, they just come between the article and the noun. Thus we have,
ntaa sónánka
"this dog"
tataa sónánka
"this dog (previously mentioned)"
skinan sónánka
"that dog (previously mentioned)"
ntintí sónánka
"two dogs"
ntó sónánka
"what/which dog"
Note how the demonstrative can affect the choice of the definite article between the second and third examples and also how numerals take singular nouns.
I'm not yet absolutely certain how I should analyse this NP structure. Is it safe to assume that all the determinative words coming before the noun are actually heads of the NPs? I'm also not sure at all what to do with the ordering in compounds. For example natans ʔíí ("merchant") is a compound of "sell" (natans) used as a noun and "man" (ʔíí), i.e. "a male merchant". Another compound would be the name of the language itself, Nooníí kiskn. This consists of "coast" (nooníí) and the verb "speak" that's suffixed with an unproductive nominaliser (kisk-n), i.e. "coast speech". Do these orderings make any sense with what I've written about the non-compound NP structure or are they in need of readjusting?
A while ago we got familiar with the existential verbs and saw that they are closely connected to expressing possession. In the case of alienable possession this connection carries also over to attributive possession. For future reference, there are also some nouns that can or must receive inalienable possession via pronominal prefixes, but this group is very limited and I'll return to it later.
The idea behind the alienable attributive possessive phrases is a continuation of the existential possessive clauses. For attributive possession we just relativise the possessive clause and attach it following the possessed noun. Relativisation is done simply by prefixing the finite clause with the referential article n=. For attributive possession there's an additional simplification as the directional verb kis is dropped from the relative clause. As a result we get from the previous predicative example
Tanatans ʔíí sʔisís natsít najis sii.
| ta=natans ʔíí |
| DEF.PROX-merchant |
| sʔisí-s |
| bead-PL |
| na-tsít |
| PL3.O-exist.GROUP |
| na-kis |
| PL3.O-go.beside |
| sii |
| SG3 |
"The merchant has beads."
the attributive counterpart
nsʔisís natsít tanatans ʔíí
| n=sʔisí-s |
| REF-bead-PL |
| n=na-tsít |
| REF-PL3.O-exist.GROUP |
| ta=natans ʔíí |
| DEF.PROX-merchant |
"the merchant's beads"
A possessed noun always has a specific referent so also the full possessed NP must be preceded by one of the articles. Unfortunately here we don't see the referential article introducing the relative clause as on the surface it merges together with the PL3.O prefix of the existential verb. It remains visible in the case of singular or dual possessed nouns,
nsónánka nitón win
| n=sónánka |
| REF-dog |
| n=ị-tón |
| REF-SG3.O-exist.COMP |
| win |
| I |
"my dog"
Comparing this to the previous example also demonstrates the classificatory function of the existential verbs.
So this already demonstrates quite nicely different aspects of the NP structure. Both adjectives and relative clauses follow nouns. Complex attributes tend away from the head noun so relative clauses follow adjectives,
nsónánka kaʔa nitón win
| n=sónánka |
| REF-dog |
| kaʔa |
| old |
| n=ị-tón |
| REF-SG3.O-exist.COMP |
| win |
| I |
"my old dog"
If the NP has a specific referent, definite or not, it must be prefixed by one of the three articles, referential n=, proximal definite t(a)= or distal definite ski=. Lack of an article always signals generic reference. Sʔisí-s (bead-PL) in the first example above is one such case as the speaker has no specific beads in mind, just the idea that the merchant has some of them available.
Besides the articles there are some other determiners or determiner like words that also precede the noun. These include the demonstratives,
taa, PROX
kaa, MEDIUM PROX, same level
ʔot, MEDIUM PROX, below
ját, MEDIUM PROX, above
nan, DIST, same level
tao, DIST, below
sin, DIST, above
skín, DIST, invisible/anaphoric
numerals (giving here the cardinals 1-10),
1 jók
2 tiní
3 wan
4 ʔisí
5 kiis
6 onaasoas
7 sonasi
8 tniʔiis
9 tnsi
10 ksis
as well as some other determiners,
tanʔi, some
tó, basic interrogative for WH-questions
etc.
None of these override the need for an article, they just come between the article and the noun. Thus we have,
ntaa sónánka
| n=taa |
| REF-PROX |
| sónánka |
| dog |
"this dog"
tataa sónánka
| ta=taa |
| DEF.PROX-PROX |
| sónánka |
| dog |
"this dog (previously mentioned)"
skinan sónánka
| ski=nan |
| DEF.DIST-DIST.LEVEL |
| sónánka |
| dog |
"that dog (previously mentioned)"
ntintí sónánka
| n=tiní |
| REF-two |
| sónánka |
| dog |
"two dogs"
ntó sónánka
| n=tó |
| REF-Q |
| sónánka |
| dog |
"what/which dog"
Note how the demonstrative can affect the choice of the definite article between the second and third examples and also how numerals take singular nouns.
I'm not yet absolutely certain how I should analyse this NP structure. Is it safe to assume that all the determinative words coming before the noun are actually heads of the NPs? I'm also not sure at all what to do with the ordering in compounds. For example natans ʔíí ("merchant") is a compound of "sell" (natans) used as a noun and "man" (ʔíí), i.e. "a male merchant". Another compound would be the name of the language itself, Nooníí kiskn. This consists of "coast" (nooníí) and the verb "speak" that's suffixed with an unproductive nominaliser (kisk-n), i.e. "coast speech". Do these orderings make any sense with what I've written about the non-compound NP structure or are they in need of readjusting?