The latest project I've been working on when my other conlangs start to bore me is Azdûgan. With it, I want to do everything I've become interested in that have not been compatible with previous conlangs (Uscaniv, Imutan, Kiassa Turasta, Kin-Saŋ), as well as a strong aesthetic impulse.
PROCESS
The way I've approached the design is similar to how Team 1 of the latest Akana relay approached the creation of Proto-Ronquian (yet to be fully unveiled). Namely, starting off with a list of desired features in Google Docs, and an aesthetic goal statement, I then wrote an outline (or Table of Contents if you will) and then began filling in chapter by chapter, using the comments feature to note issues that need to be resolved at a later date. It's been working fairly well and I want to show off some of the stuff so far.
AESTHETIC STATEMENT Each word is to feel comprehensive yet compact and sound like buzzing tiny soft pebbles in the mouth, like the rumble of a desert king. Every sentence should sound like an incantation and every proper name like the appellation of some ancient demon.
Phonologically, I've taken some inspiration from ancient near eastern languages: Elamite, Hittite, Akkadian and Sumerian. Grammar-wise I am mostly indebted to Bantu, Kartvelian and Eskimo lanaguages.
FEATURE LIST
Agglutinative: morphemes are monoexponential
Polysynthetic: polypersonalism paired with argument incorporation
Animacy, gender and noun-class: complex animacy-based noun class system which encodes gender for human nouns
Both differential subject marking and differential object marking, i.e. expected roles are unmarked, unexpected roles are explicitly marked accusative or ergative
Direct-inverse alignment in argument indexing on verbs
Differential locative marking: When acting as oblique locative phrases, human nouns require cushioning before locative marking, animates less so, inanimates even less, and specifically place-related nouns such as proper place names and words like 'hill' or 'cave' require no marking at all
Dative-marking for indirect objects
Asymmetrical TAM system with gaps in aspect and mood coverage per tense
Slight tendency towards double marking, i.e. head and dependent marking
Last edited by vegfarandi on Thu Jul 25, 2019 12:57 pm, edited 3 times in total.
PHONOLOGY Consonants
Azdûgan has 28 consonantal phonemes. Stops, prenasal stops and fricatives contrast in voice; nasals and approximants are only voiced.
p t c k
b d j g
mp nt nc ŋk
mb nd nj ŋg
m n ŋ
s ś ḫ h
z ź
l y
r
Vowels
There are 4 vowel qualities which appear in short and long varieties:
i î • u û
e ê • a â
There is a slight quality difference between long and short vowels. Long vowels are î /iː/ û /uː/, ê /æː/ and â /ɑː/. Short vowels are commonly reduced to i [ɪ], u [ʊ], e/è [ɛ] and a [ə].
Short a is [ɛ] when the next syllable contains i or î (i-umlaut). When this occurs, it is indicated by spelling it è in the transliteration. In addition, a is usually pronounced [ɔ] when the following syllable contains u or û (u-umlaut). As there is no phonemic o-like sound, this does not cause a phonemic merger, and is not indicated in transliteration.
Two-Vowel Sequences and Diphthongs
The language does not have any phonemic diphthongs in the traditional sense, although vowels may be preceded or followed by the semivowel glide y. The semivowel y may only occur in contact with the close front vowels i, î between vowels. This means, for instance that iya and eyî are legal sequences, but y disappears after a consonant preceding i, and likewise after i preceding a consonant so Cyi > Ci and iyC > iC.
When two vowels appear in sequence, the close vowels can usually appear in hiatus before other vowels but with other combinations, certain collapses occur:
Two identical vowels merge into one long vowel.
The sequences ea/ae become â
The sequences ai/ei become ê
The sequence au becomes â
The sequence eu becomes û.
Ay and ey are legal, although ey is commonly pronounced like î.
Y is always allowed before a vowel, except see note on i
Stress
Stress falls on the final syllable unless the penultimate syllable is long, and the ultimate short, in which case it falls on the penultimate.
Secondary stress falls on the third previous syllable or second previous syllable if it is long, or if there are no more syllables. If only one syllable precedes the primary stress syllable, no secondary stress applies.
Phonotactics
Syllables have the following structure:
(C)(r/y)V(C)(C)
where C is a consonant, N is a nasal, and V is vowel.
Onset consonants may be followed a medial, r or y can be stops or prenasalized stops:
I like it. Polysynthetic languages are my favorite type, and there's a lot of variety within that category, too. One question though, you didnt put up the IPA's of a consonants .... is this a scratchpad thread where you post a lot of posts with separate info in each?
Pabappa wrote: ↑Sun May 05, 2019 6:39 pm
I like it. Polysynthetic languages are my favorite type, and there's a lot of variety within that category, too. One question though, you didnt put up the IPA's of a consonants .... is this a scratchpad thread where you post a lot of posts with separate info in each?
Yeah, I've amended the title accordingly
As far as consonants, I haven't written any of the phonetic rules yet, leaving it for later. But roughly, they are (few surprises):
p t c k /p t ʧ k/
b d j g /b d ʤ g/
mp nt nc ŋk /m͡p n͡t n͡ʧ ŋ͡k/
mb nd nj ŋg /m͡b n͡d n͡ʤ ŋ͡g/
m n ŋ /m n ŋ/
s ś ḫ h /s ʃ x h/
z ź /z ʒ/
l y /l j/
r /r/
Before we delve into particular parts of speech, we need to examine indices, which are morphemes used across nouns, pronouns, verbs and adjectives to mark noun class and cross-indexing between constituents. These are somewhat similar to to "concords" and Bantu languages and "noun class markers" in Caucasian languages.
An Azudûgan index is a single-consonant morpheme used across that refers to nominals and gives information about their animacy, class, countability and number, and marks its agreement across syntactic constituents.
Nominal Indices
There are 8 nominal indices: m-, d-, g-, z-, r- in the singular, b-, y- and ź- in the plural. The nominal indices appear in 10 pairs as well as 3 possible groups of uncountable nouns for 13 noun classes in total.
Since it's hard to do tables here, I'm writing out in a list, using this format:
GLOSSING: singular / plural (semantics)
— in the plural slot marks uncountable classes
Notes: In glosses, the class abbreviations (M, F, A1 etc.) are used for singular forms. For human plural b-, simply H is used as it becomes ambiguous as to gender. Other plurals append .PL so A1.PL, I4.PL etc. When no direct reference is given in the clause for non-human referents, I pick a likely unspoken referent as opposed to glossing as I3–5.PL for a ź- index.
Pronominal Indices
In addition, there are 5 pronominal indices: n- (1s), ś- (2s), k- (1p.EXCL), niś- (2p.INCL), and h- (2p) and the reciprocal/reflexive index źuz-.
Index Use
Indices are used with all major parts of speech: nouns, pronouns, adjectives and verbs. With nouns, they have the primary role of encoding noun class and number, as well as marking possession. With adjectives, they mark agreement and with verbs, they mark (polypersonal) arguments. A different connecting vowel is used to differentiate between these uses.
Referential Indices
Some—though not all—nouns, such as dideg ‘finger’ and gipâ ‘path’, have index prefixed overtly marking their class. Unless the noun’s root starts with a vowel, these always use the referential connector -i-. Other nouns are headless, having no singular class marker, e.g. ziŋ ‘day’. However, in the plural, all nouns have a index and connector even if their singular is unmarked, e.g. biziŋ ‘days’.
Most pronouns and determiners are also composed of indices and have referential semantics but the vowel varies based on pronominal function:
[mana ‘he’, di ‘it (A1 or I1), ebe ‘who’ etc.]
Possessive Indices
An index may be used on a noun phrase using the possessive connector -u-:
[nuhadâl ‘my house’ duzisâbag ‘her sword’ (note the zi- referential index on the noun zisâbag)]
If the noun starts with u, there is a collapse:
[bûŋru ‘their dog’ (< b-u-uŋru)]
Attributive Indices
The attributive connector -a- is used for adjectival agreement (in the example, it is è due to the following i):
[gèdibu igên g-a-dibu igên
A2-ATTR-small bird
small bird]
And relative agreement:
[gazegasui igên – chirping bird]
The attributive connector may be swapped for the referential one to substantivize such phrases:
[gizegasui – the chirping one zazumi nuzisabâg – my long sword zizumi – the long one]
Argument Indices
Indices are used without connectors to mark arguments on a verb, often multiple at a time:
[nidazmegembyê – he made me give it to her]
More on this later.
F and I1 have the same indice pair d- / b-. Does that mean that they can never be distinguished through agreement, or does the opposition human/inanimate come into play otherwise e.g. different verb uses?
Vilike wrote: ↑Fri May 10, 2019 12:47 pm
F and I1 have the same indice pair d- / b-. Does that mean that they can never be distinguished through agreement, or does the opposition human/inanimate come into play otherwise e.g. different verb uses?
Good catch and question. Yes, morphologically the are identical, but when it comes to agentivity hierarchy, which affects both case and verb marking, they are distinct.
I'm finding maintaining this in two places a bit unwieldy and converting to BBCode sucks. So instead, I've decided to share my working Google Doc. Comments are enabled for now, so feel free to post questions/comments there, or here in the thread.
You can read this in the Google Doc under NOUNS > Case. I'm trying to describe the case system which is a bit convoluted but below is it:
----
Case
The language has between 8-10 cases. They can be split into term, oblique argument and adverbial cases.
The three term cases are the direct (-0) ergative (-e) and accusative (-ad). The oblique argument case is the dative (-za) and lastly, the adverbial cases are the locative (-ka) instrumental (-amma) and ablative (-ube).
The direct case marks is used for expected roles (agent for human nouns, patient for inanimate nouns, either for animate nouns depending on context), as well as locatives of place names: âtig śitiś midûz – the man stole a spear (not *âtige śitiśad midûz) Kudar nêza – I’m in Kudar (not *Kudarka nêza)
The ergative and accusative mark exceptions from expected roles: when an inanimate is agent, or when a human is patient, respectively; and animates are assigned case based on context. The dative primarily marks indirect objects. The remaining cases are used for various adverbial adjuncts.
The basic case ending morphs are ergative -e, accusative -ad, dative -za, locative -ka, instrumental -amma, ablative -ube. Human nouns need to be suffixed with the generic pronoun -na- in order to take any adverbial case endings. Inanimate nouns cannot take the accusative or dative case endings, but reappropriate the accusative ending for a distinct allative case, which in human and animate nouns is formally syncretic with the locative.
----
Is there a clearer way to describe all of this? I'm having a hard time getting the right words to this.
So what's the allative case for? That's really not clear. As far as I can see inanimate nouns don't take the accusative, because you expect that role and so it is direct case. Similarly you don't expect them to be indirect objects so you skip the dative, but then what's left for the allative?
Yeah, it'd be helpful if you went into detail about what the different cases of location do and what locative is if it is separate from allative.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him! kårroť
Some highlight changes:
- I've resolved (for the most part) my case description issue, I hope. I've identified the syntactic vs. morphological cases. Essentially, my current analysis is that the syntactic cases outnumber the morphological ones, and that syntactic cases may be filled with different morphological cases depending on animacy.
- I've also added a chapter on relational nouns and in turn nixed postpositions entirely.
Next up:
- I've decided to alter the TAM system, going fully tenseless (taking some inspiration from Mayan) and simplifying the amount of marked TAMs while diversifying how they get applied morphologically a bit.