kskʼɩroijø̂u — an antilanguage

[bradrn]

Like many conlangers, my languages can suffer from some amount of… how should I put it? monotony of style. This thread is an experiment: to produce an antilanguage with as many features as possible which I haven’t used before. To do this in a systematic way, I have gone through WALS, found all the features where I haven’t yet explored one or more options, then attempted to create a conlang from those.

(And no, kskʼɩroijø̂u is not the name of this language. It means ‘an antilanguage’ in throijø̂uzh kʼaezh.)

Segmental phonology

2A: Vowel Quality Inventories — Large (7-14)
4A: Voicing in Plosives and Fricatives — In fricatives alone
6A: Uvular Consonants — Uvular stops and continuants
7A: Glottalized Consonants — Ejectives only
8A: Lateral Consonants — Laterals, but no /l/, no obstruent laterals
9A: The Velar Nasal — No velar nasal
10A: Vowel Nasalization — Contrast present
11A: Front Rounded Vowels — High and mid

I’ve made consonant inventories of most sizes, from 8 consonants (Hlʉ̂) to 27 (unnamed protolanguage 20). Vowel inventories, on the other hand, tend to remain smaller: I’m most comfortable with 4-vowel inventories, and only my most recent conlangs have had as many as 7. Furthermore, I tend to be fairly boring with my vowel systems, which are either /a i u/ + one mid vowel or /a e i o u/ + two central vowels; none have diphthongs or indeed anything more interesting than a length contrast. My consonant systems are less boring, but still tend to be symmetrical and stop-heavy, with minimal or no voicing distinctions, no consonants further back than velar, and no non-pulmonic consonants.

Consonants:

/m n ɲ/ ⟨m n nj⟩
/p t k q kʼ qʼ/ ⟨p t k q kʼ qʼ⟩
/t͡θ t͡s t͡ʃ/ ⟨tc ts ch⟩
/f θ s ʃ χ h/ ⟨f th s sh kh h⟩
/v ð z ʒ/ ⟨v dh z zh⟩
/ɺ ʎ r j w/ ⟨l lj r j w⟩

Vowels:

/i y u/ ⟨i y u⟩
/ɪ ʊ/ ⟨ɩ oe⟩
/e ø o/ ⟨e ø o⟩
/ɛ ɑ/ ⟨ae a⟩
/ĩ ũ ẽ õ ɑ̃/ ⟨ĩ ũ ẽ õ ã⟩
/ɑɛ̯ ɑɪ̯ ɑʊ ɛɑ̯ ɛɪ̯ eo̯ øʊ̯ yi̯ yu̯ ʊɪ̯ ou̯/ ⟨aae ai ao ea ei eo øu yi yu oi ou⟩
/ɑ̃ĩ̯ ɑ̃ũ ẽɑ̯̃ ẽõ̯ ẽĩ̯ õũ̯ ũĩ̯/ ⟨ãi ão ẽa ẽo ẽi õu ũi⟩

The orthographies are chosen to use as many digraphs as possible — ambiguity is ignored, unlike my usual orthographies. Especially unusual for me is the use of digraphs for vowels, though I also use some particularly obnoxious letters.

Suprasegmental phonology

WALS is particularly poor in this area. Mostly my languages have very simple syllable structure, so this one uses what Easterday calls ‘highly complex’ syllable structure. I’ve used both fully tonal and fixed stress systems, so this one has a flexible pitch-accent system.

Syllable structure is (C₁)(C₂)(C₃)(C₄)(C₅)V(C₆)(C₇)(C₈), where C₁–C₄ may be any obstruents agreeing in voicing (though note that the tenuis stops have voiced allophones), C₅ may be any sonorant or /h/, C₆ may be any continuant except a semivowel, and C₇–C₈ may be any obstruents agreeing in voicing. Additional restrictions apply to /h/, which cannot occur as any of C₂–C₄ or in the coda, to /ɺ/, which may not occur at the start of a word, and to ejectives, which may not appear at the end at a word. Geminates, as well as stops followed by homorganic ejectives, are disallowed; if one is produced in the morphology it is degeminated.

Stress is ‘diacritic’ and right-weighted, in that one syllable of each morpheme may be marked for stress, and primary stress occurs on the rightmost stressed syllable. The stressed syllable takes either high tone /ˊ/ or falling tone /ˆ/; the tone used is lexically determined by the root. Thus for instance [TODO]

(Nominal) Morphology

21A: Exponence of Selected Inflectional Formatives — Case + number
25A: Locus of Marking: Whole-language Typology — Dependent marking
28A: Case Syncretism — Inflectional case marking is syncretic for core and non-core cases
30A: Number of Genders — Two
31A: Sex-based and Non-sex-based Gender Systems — Sex-based
32A: Systems of Gender Assignment — Semantic and formal
37A: Definite Articles — Definite affix
38A: Indefinite Articles — Indefinite affix
41A: Distance Contrasts in Demonstratives — Two-way contrast
43A: Third Person Pronouns and Demonstratives — Third person pronouns and demonstratives are related to remote demonstratives
45A: Politeness Distinctions in Pronouns — Binary politeness distinction
49A: Number of Cases — 10 or more cases
50A: Asymmetrical Case-Marking — Additive-quantitatively asymmetrical

Most of my languages — all of them, now that I think about it — have been heavily head-marking, or at least double-marking with a low number of noun cases. Obviously the best thing to do in this language, therefore, is to have as big a case system as I can manage. I am famously fond of non-accusative alignments; thus, nominative–accusative alignment seems good here — with minimal syntactic case-marking even, restricted to pronouns (so just like English). I also strongly tend to avoid gender systems, especially non-semantic ones, so a two-term formal gender system seems right out of my comfort zone.

There are two genders: male and female. Gender assignment is largely formal, in that the gender of humans is more or less random. Gender is marked on demonstratives, third person pronouns, the definite prefix, case suffixes, and adjectival modifiers (via case concord). All nouns are assigned a gender: either masculine or feminine, though some human nouns may take either. Adjectives may take either gender marking, depending on their head noun.

Definiteness and plurality are marked via prefixes, in that order, as follows:

Indefinite	k-
Definite	a- (M) th- ~ dh- (F)
Plural	õ-

Thus for instance we have ktsɩvŷ ‘a table’, atsɩvŷ ‘the table’, kõtsɩvŷ ‘tables’, ãotsɩvŷ ‘the tables’; or kʼéop ‘hill’, thkʼéop ‘the hill’, kõkʼéop ‘hills’, thõkʼéop ‘the hills’. (Note minor fusion in the masculine definite plural.)

A very small number of otherwise feminine nouns take the masculine definite article. These mostly relate to plant and animal products: achélaae ‘the leaf’, aqnǿur ‘the tuber’, apŷlks ‘the wool’.

The following cases are distinguished:

• Nominative — transitive and intransitive subject; unmarked case
• Accusative — transitive object
• Dative — recipient or beneficiary
• Locative — location
• Instrumental — instrument
• Comitative — accompaniment
• Privative — without, atelic object
• Equative — copula complement
• Comparative — index of comparison
• Ablative — source, location from
• Allative — destination, location to
• Vocative — addressing a person

There is also a genitive -ch(-ɩ), for possessors. However, this suffix cannot really be considered a ‘case’, since it does not alternate with other cases or engage in case concord, but rather co-occurs with other cases in the same word.

The inventory of cases differs slightly between pronouns and nominals. The personal pronouns have no equative or vocative forms; first and second persons additionally have no instrumental. On the other hand, only personal pronouns have a distinct accusative; other nominals merge the nominative and accusative into a single ‘direct’ case.

Nouns and adjectives take distinct case-markers depending on whether they are male or female. The case-markers further differ between those words ending in consonants, and those ending in vowels. Some consonant-following case-markers require a lexically determined thematic vowel, always a short non-nasalised vowel. Case-markers are as follows:

Case	M Noun -V	M Noun -C	F Noun -V	F Noun -C	M Adj -V	M Adj -C	F Adj -V	F Adj -C
DIR	thâle-∅	koist-∅	thâle-∅	koist-∅	chnjé-∅	hõkʼâdh-∅	chnjé-∅	hõkʼâdh-∅
DAT	thâle-p	koist-p	thâle-f	koist-ɩ-f	chnjé-zp	hõkʼâdh-ae-zp	chnjé-f	hõkʼâdh-ae-f
LOC	thâle-zh	koist-ɩ-zh	thâle-∅	koist-∅	chnjé-z	hõkʼâdh-ae-z	chnjé-j	hõkʼâdh-ae-j
INSTR	thale-qâe	koist-qáe	thâle-k	koist-k	chnje-zkáe	hõkʼadh-kâe	chnjé-ch	hõkʼâdh-ch
COM	thâle-st	koist-ɩ-st	thâle-s	koist-ɩ-s	chnjé-zt	hõkʼâdh-ae-zt	chnjé-sh	hõkʼâdh-ae-sh
PRIV	thâle-lø	koist-ø	thâle-woe	koist-oe	chnjé-lø	hõkʼâdh-ø	chnjé-joe	hõkʼâdh-oe
EQU	thâle-k	koist-k	thâle-kh	koist-ɩ-kh	chnjé-zk	hõkʼâdh-k	chnjé-j	hõkʼâdh-ae-j
COMP	thâle~je (~jV)	koist~oi (~V)	thâle~we (~wV)	koist~woi (~wV)	chnjé-z	hõkʼâdh-ae-z	chnjé-j	hõkʼâdh-ae-j
ABL	thâle-m	koist-ɩ-m	thâle-p	koist-p	chnjé-n	hõkʼâdh-ae-n	chnjé-p	hõkʼâdh-p
ALL	thâle-lj	koist-ɩ-lj	thale-lɩ̂	koist-ɩ-lɩ́	chnjé-lø	hõkʼâdh-ø	chnje-jɩ́	hõkʼadh-jɩ̂
VOC	thâle-wo	koist-o	thâle-∅	koist-∅	chnjé-∅	hõkʼadh-∅	chnjé-∅	hõkʼâdh-∅

(thâle ‘parent’, koist-ɩ ‘person’, chnjé ‘red’, hõkʼâdh-e ‘big’.)

Pronouns and demonstratives exhibit some irregularities in case-marking, inflecting as follows:

Case

1SG

1PL

2SG

2SG (polite)

2PL

3SG.M

3SG.F

3PL (M & F)

PROX.SG.M

PROX.SG.F

PROX.PL.M

PROX.PL.F

REM.SG.M

REM.SG.F

REM.PL (M & F)

NOM

tai

nai

soe

noe

nũ

pe

fe

mẽ

kʼae

khae

õkʼae

õkhae

pe

fe

mẽ

ACC

tae

nae

zø

noe

nũ

pɩ

fɩ

mĩ

kʼe

khe

õkʼe

õkhe

pɩ

fɩ

mĩ

DAT

tap

nap

soep

noep

nũm

pep

fep

mẽp

kʼaep

khef

õkʼaep

õkhef

pep

fef

mẽp

LOC

tazh

nazh

soezh

noezh

nũnj

pezh

fezh

mẽzh

kʼaezh

khaezh

õkʼaezh

õkhaezh

pezh

fezh

mẽzh

INSTR

—

—

—

—

—

peqâe

feqâe

mẽqâe

kʼaeqâe

khaek

õkʼaeqâe

õkhaek

peqâe

feqâe

mẽqâe

COM

tasi

nasi

soesi

noesi

nũsi

pesi

fesi

mẽsi

kʼaest

khaes

õkʼaest

õkhaes

pest

fes

mẽst

PRIV

tao

nao

sø

nø

nõ

peo

feo

mẽo

kʼaelø

khaewoe

õkʼaelø

õkhaewoe

pelø

fewoe

mẽlø

COMP

taji

naji

soewoe

noewoe

nũwoe

peje

feje

mẽje

kʼaejae

khaejae

õkʼaejae

õkhaejae

peje

fewe

mẽje

ABL

tam

nam

soem

noem

nũm

pem

fem

mẽm

kʼaem

khaep

õkʼaem

õkhaep

pem

fep

mẽm

ALL

tao

nao

soe

noe

nũ

pei

fei

mẽi

kʼei

khei

õkʼei

õkhei

pei

fei

mẽi

An important part of grammar here is case concord: all nominal modifiers take the same case inflections as the head. This applies to the genitive as well, which can result in extensive Suffixaufnahme (case stacking):

Jaae kroitcépek ahoekálavchɩkh Nêjakchɩchɩkh Skafárpozhchɩk.

jaae

it.is

k-roi-tcépe-k

INDEF-NMLZ-mark-EQ

a-hoe-kálav-ch-ɩ-kh

DEF-F-temple-GEN-THM-EQ

Nêjak-ch-ɩ-ch-ɩ-kh

Nêjak-GEN-THM-GEN-THM-EQ

Skafárp-o-zh-ch-ɩ-k

Skafárp-THM-LOC-GEN-THM-EQ

This is a drawing of the temple to Nêjak at Skafárp.

My languages also tend to be rather lacking in derivational morphology. Therefore, to avoid this, I will immediately create numerous derivational affixes which can be applied to nouns:

• skʼɩ- (n→n) — nominal negation (‘un-’, ‘anti-’)
• ljǿ- (n→n) — nominal past tense (‘ex-’)
• -shoi (n→n) — augmentative, intensifier (‘arch-’, ‘over-’, etc.)
• es- (n→n) — collective noun
• -fo (n→n) — object related to
• ei- (n→adj) — similarity (‘-like’)
• -njeo (n→adj) — quality (‘-ness’)
• soen- (n→adj) — having (’-ful’)
• -ale (n→v) — turn into (‘-ify’)
• qʼu- (n→v) — enclose in (‘en-’)
• fkhej- (n→v) — pass through

Also of note is the patient nominalisation prefix roi- (v→n), used in one or two examples above. Thus we have for instance ljøktsɩvŷ ‘ex-table’, eikʼéop ‘hilly’, roitcépeale ‘convert to drawing’, eskálav ‘cluster of temples’.

[Richard W]

How would it pronounce Przewalski? After all, English doesn't require adjacent obstruents to agree in voicing.

[Zju]

That's an interesting experiment. I wonder if the choice of some options is just a secondary artifact of them being rare in languages overall, e.g. 4A or 49A.

[bradrn]

That's an interesting experiment. I wonder if the choice of some options is just a secondary artifact of them being rare in languages overall, e.g. 4A or 49A.

Perhaps, but for those ones at least I had solid reasoning: 4A is because I’m fond of voicing distinctions in stops but not in fricatives, and 49A is because I tend towards smallish case systems, if any.

How would it pronounce Przewalski?

If it’s adapted from the English, probably pørzhøválzki. If it’s adapted from the Russian, probably pørzhiváljskɩj. If it’s adapted from the Polish, probably pshaeváljski. (Going from the pronunciations on Wikipedia.)

After all, English doesn't require adjacent obstruents to agree in voicing.

Well, this language isn’t English. Besides, it’s easier here, since stops aren’t marked for voicing, so clusters like zk [zɡ] become allowable.

[Ares Land]

That's an interesting approach -- though I wonder how to handle diachronics that way

I'd be very happy to see more of this!

[bradrn]

That's an interesting approach -- though I wonder how to handle diachronics that way

With great difficulty. I’m not even going to attempt it. (Though feel free to try if you want!)

I'd be very happy to see more of this!

Thanks!

[Travis B.]

How would it pronounce Przewalski? After all, English doesn't require adjacent obstruents to agree in voicing.

In the English here all obstruent sequences must agree in phonetic voicing, and by default obstruent sequences are voiceless even when the original phonemes were not all fortis. Note that this has nothing to do with phonemic fortisness versus lenisness, because that is still reflected in preceding vowel length, aspiration, and (pre)glottalization.

[Vardelm]

Finally getting around to going through this...

Like many conlangers, my languages can suffer from some amount of… how should I put it? monotony of style. This thread is an experiment: to produce an antilanguage with as many features as possible which I haven’t used before. To do this in a systematic way, I have gone through WALS, found all the features where I haven’t yet explored one or more options, then attempted to create a conlang from those.

I think this is an excellent exercise! I seem to remember someone on the ZBB doing something similar quite a while ago, maybe in the 2010-2012 range? Most of us probably have similar tendencies in our conlanging, so specifically going against the grain should help get one out of their rut. To some extent, the fact I'm working on 4 conlangs at once which are specifically designed to be different from one another has done that same thing for me. Even there, I think there are some grooves that I fall back into.

Anyway....

I’ve made consonant inventories of most sizes, from 8 consonants (Hlʉ̂) to 27 (unnamed protolanguage 20). Vowel inventories, on the other hand, tend to remain smaller: I’m most comfortable with 4-vowel inventories, and only my most recent conlangs have had as many as 7. Furthermore, I tend to be fairly boring with my vowel systems, which are either /a i u/ + one mid vowel or /a e i o u/ + two central vowels; none have diphthongs or indeed anything more interesting than a length contrast. My consonant systems are less boring, but still tend to be symmetrical and stop-heavy, with minimal or no voicing distinctions, no consonants further back than velar, and no non-pulmonic consonants.

I wonder if simply looking at WALS and analyzing your tendencies like this will do the same trick to break out of ruts without having to make an entire conlang.

Consonants:

/m n ɲ/ ⟨m n nj⟩
/p t k q kʼ qʼ/ ⟨p t k q kʼ qʼ⟩
/t͡θ t͡s t͡ʃ/ ⟨tc ts ch⟩
/f θ s ʃ χ h/ ⟨f th s sh kh h⟩
/v ð z ʒ/ ⟨v dh z zh⟩
/ɺ ʎ r j w/ ⟨l lj r j w⟩

Not that this matters for this lang, but how stable would /t͡θ/ be? I imagine it would merge into /t͡s/ very quickly. Still, I don't think I've seen a conlang with it, so +5 points to Gryffindor for blazing a bold trail.

The orthographies are chosen to use as many digraphs as possible — ambiguity is ignored, unlike my usual orthographies. Especially unusual for me is the use of digraphs for vowels, though I also use some particularly obnoxious letters.

That actually doesn't strike me as obnoxious. There's a fair number of vowels, so you're not going to get away with just <a e i o u>.

Syllable structure is (C₁)(C₂)(C₃)(C₄)(C₅)V(C₆)(C₇)(C₈),

I realize some natlangs do this, so it's not completely bonkers. I wanted one of my langs to have more complex syllables like this, but I just couldn't push myself this far.

Most of my languages — all of them, now that I think about it — have been heavily head-marking, or at least double-marking with a low number of noun cases. Obviously the best thing to do in this language, therefore, is to have as big a case system as I can manage. I am famously fond of non-accusative alignments; thus, nominative–accusative alignment seems good here — with minimal syntactic case-marking even, restricted to pronouns (so just like English). I also strongly tend to avoid gender systems, especially non-semantic ones, so a two-term formal gender system seems right out of my comfort zone.

I used to have a strong preference for conlangs with cases, but I find now that the design space without them is fun to work with as well.

Gender seems to be one of those things that most conlangs lack, maybe since it can add a fair bit of complexity with - arguably - minimal benefit.

A very small number of otherwise feminine nouns take the masculine definite article. These mostly relate to plant and animal products: achélaae ‘the leaf’, aqnǿur ‘the tuber’, apŷlks ‘the wool’.

The inventory of cases differs slightly between pronouns and nominals. The personal pronouns have no equative or vocative forms; first and second persons additionally have no instrumental. On the other hand, only personal pronouns have a distinct accusative; other nominals merge the nominative and accusative into a single ‘direct’ case.

Nice touches!

---

So.... question for you: in working on this, how do you like the result? Is it aesthetically or intellectually appealing to you, just "meh", or do you even not like it? What's your plans on developing it further? I'm just curious to see if making a language that goes against your tendencies results in a conlang you like or not.

[bradrn]

Finally getting around to going through this...

Like many conlangers, my languages can suffer from some amount of… how should I put it? monotony of style. This thread is an experiment: to produce an antilanguage with as many features as possible which I haven’t used before. To do this in a systematic way, I have gone through WALS, found all the features where I haven’t yet explored one or more options, then attempted to create a conlang from those.

I think this is an excellent exercise! I seem to remember someone on the ZBB doing something similar quite a while ago, maybe in the 2010-2012 range?

Yes, I remember this also, though I think it was a bit later than 2012. (It might have been Curlyjimsam, maybe? I recall their name started with a C.) This thread, though inspired by that one, is slightly different, in that they attempted to create a language which they would actively hate, whereas I’m just trying to make something different, not necessarily hateful.

Most of us probably have similar tendencies in our conlanging, so specifically going against the grain should help get one out of their rut. To some extent, the fact I'm working on 4 conlangs at once which are specifically designed to be different from one another has done that same thing for me. Even there, I think there are some grooves that I fall back into.

I’m in the same boat. I’ve posted here about 3 quite different conlangs, but they’re still similar in some subtle ways.

I’ve made consonant inventories of most sizes, from 8 consonants (Hlʉ̂) to 27 (unnamed protolanguage 20). Vowel inventories, on the other hand, tend to remain smaller: I’m most comfortable with 4-vowel inventories, and only my most recent conlangs have had as many as 7. Furthermore, I tend to be fairly boring with my vowel systems, which are either /a i u/ + one mid vowel or /a e i o u/ + two central vowels; none have diphthongs or indeed anything more interesting than a length contrast. My consonant systems are less boring, but still tend to be symmetrical and stop-heavy, with minimal or no voicing distinctions, no consonants further back than velar, and no non-pulmonic consonants.

I wonder if simply looking at WALS and analyzing your tendencies like this will do the same trick to break out of ruts without having to make an entire conlang.

I doubt it: actually making something does wonders for your understanding of a concept. (Not just in conlanging, I should note.)

But also, why would I choose not to make a conlang if I had the opportunity?

Consonants:

/m n ɲ/ ⟨m n nj⟩
/p t k q kʼ qʼ/ ⟨p t k q kʼ qʼ⟩
/t͡θ t͡s t͡ʃ/ ⟨tc ts ch⟩
/f θ s ʃ χ h/ ⟨f th s sh kh h⟩
/v ð z ʒ/ ⟨v dh z zh⟩
/ɺ ʎ r j w/ ⟨l lj r j w⟩

Not that this matters for this lang, but how stable would /t͡θ/ be? I imagine it would merge into /t͡s/ very quickly.

You’re probably right, not that I’d know.

Still, I don't think I've seen a conlang with it, so +5 points to Gryffindor for blazing a bold trail.

I’m pretty sure I’ve heard remarks from conlangers that [t̪] is likely to affricate to [t̪͡θ], so I’d be surprised if no other conlang had these.

The orthographies are chosen to use as many digraphs as possible — ambiguity is ignored, unlike my usual orthographies. Especially unusual for me is the use of digraphs for vowels, though I also use some particularly obnoxious letters.

That actually doesn't strike me as obnoxious. There's a fair number of vowels, so you're not going to get away with just <a e i o u>.

Oh, I’d say that ⟨ɩ ø⟩ are pretty obnoxious, as letters. Just be glad I didn’t use ⟨ɑ þ ĸ ƞ ḫ ꜫ ꜭ ʒ ǯ⟩ as well — or consonantal ⟨ɛ⟩. (Þroijø̂uɛ ꜭaeɛ, ksꜭɩroijø̂u, jɑe kroiʒþepek ɑhoekɑ́lɑvǯɩḫ Nêjɑkǯɩǯɩḫ Skɑfɑ́rpoɛǯɩk… now that really is obnoxious!)

Syllable structure is (C₁)(C₂)(C₃)(C₄)(C₅)V(C₆)(C₇)(C₈),

I realize some natlangs do this, so it's not completely bonkers. I wanted one of my langs to have more complex syllables like this, but I just couldn't push myself this far.

I’m in a similar position to you — I’ve been wanting to explore large vowel systems and/or large syllable structures for a long time, but I just couldn’t force myself to do it. (Now, of course, I have both.) I based the syllable structure off Easterday’s book on Highly complex syllable structure, an excellent open-access resource. Georgian and Itelmen were particular influences. It’s worth noting that this sort of thing isn’t as rare as you might think — even English has a more complex coda structure than this language (e.g. texts has four obstruents in a row).

Most of my languages — all of them, now that I think about it — have been heavily head-marking, or at least double-marking with a low number of noun cases. Obviously the best thing to do in this language, therefore, is to have as big a case system as I can manage. I am famously fond of non-accusative alignments; thus, nominative–accusative alignment seems good here — with minimal syntactic case-marking even, restricted to pronouns (so just like English). I also strongly tend to avoid gender systems, especially non-semantic ones, so a two-term formal gender system seems right out of my comfort zone.

I used to have a strong preference for conlangs with cases, but I find now that the design space without them is fun to work with as well.

Same here. Most of my first conlangs consistently had 4–5 cases (Yaazhgigzhoomnem is the only one I’ve posted here), but now I don’t use them nearly as often.

Gender seems to be one of those things that most conlangs lack, maybe since it can add a fair bit of complexity with - arguably - minimal benefit.

In general, I think conlangers tend to avoid any irregular or lexically determined features — conlang grammars are uniquely regular compared to natlang ones.

A very small number of otherwise feminine nouns take the masculine definite article. These mostly relate to plant and animal products: achélaae ‘the leaf’, aqnǿur ‘the tuber’, apŷlks ‘the wool’.

The inventory of cases differs slightly between pronouns and nominals. The personal pronouns have no equative or vocative forms; first and second persons additionally have no instrumental. On the other hand, only personal pronouns have a distinct accusative; other nominals merge the nominative and accusative into a single ‘direct’ case.

Nice touches!

Thanks! The second one is really just an animacy-based alignment split in disguise — identical to what we have in English, in fact. But I’m rather proud of the first, which I haven’t seen in any natlang but which seems plausible nonetheless.

So.... question for you: in working on this, how do you like the result? Is it aesthetically or intellectually appealing to you, just "meh", or do you even not like it? What's your plans on developing it further? I'm just curious to see if making a language that goes against your tendencies results in a conlang you like or not.

I’m not sure I love it, but I certainly don’t hate it, and it’s at least fun to say aloud. (And surprisingly easy, too — compare Wēchizaŋkəŋ, which seems basically impossible to speak despite having only seven vowels, no non-pulmonics and few clusters at all.) When I have time, of course I’m going to keep on working on it, though… I haven’t even done the verbal morphology yet!