The Click languages

[VaptuantaDoi]

For quite a while now I've had in the back of my mind an idea for a Click language family in Rodgezhua which is the font of a sprachbundt that extends to four different language families. Basically the only distinctive feature is that it has clicks. I know Click isn't a very imaginative name for the family, and I might change that to Gk!xogk!oic if I feel like it.

Proto-Click

Proto-Click had an areally large consonant inventory, in contrast to ur-East Rodgezhuan's */p b t d k ɡ/. It likely has a deeper connection with the Tumbleweed languages, which share features includine a uvular series, nominal gender and split-S alignment, forming part of the ur-West Rodgezhuan phylum. It is most likely reflective of a very ancient pre-East Rodgezhuan migration dating back at least 20,000 years, although proto-Click itself is probably no more than 10,000 years old. I have plans for about 30 descendants, most of which obviously won't be fleshed out. The influence of Click languages is seen through the direct or indirect borrowing of clicks into several members of the Nomadic family, two Kwreid Isthmus languages, the Lower Ećhicy branch of Trintinic and the isolates of Ánni and Qqań’ılła, as well as isolated instances like ergative constructions in Trintinic and central vowels in Kwreid Isthmus.


Phonology

Proto-Click had a large inventory of forty-three consonants, including twelve clicks, and seven vowels. There was a pharyngealisation contrast on a quarter of the consonants, as well as a three-way phonation contrast for obstruents (clicks are obstruents, right?) of voiceless, voiced and ejective. The descendants of proto-Click have tended to expand rather than reduce its consonant inventory, especially in the non-pulmonic section.

Consonants:
/b bˤ t tˤ d dˤ k q ɢ/
/t͡s’ t͡ɬ’ k͡x’ q͡χ’/
/f ð ðˤ s ɬ ɬˤ ɣ χ ʕ/
/s’ ɬ’/
/m mˤ n ŋ/
/w wˤ l j/

Click consonants:
/k͡ǀ k͡! k͡!ˤ k͡ǁ k͡ǁˤ/
/k͡ǀ͜χ’ k͡ǃ͜χ’ k͡ǁ͜χ’/
/ɡ͡ʘˤ ɡ͡ǀ ɡ͡ǃ ɡ͡!ˤ/

Vowels:
/i ɨ u e ə o a/

While proto-Click did not have any phonation contrast, many descendants innovated creaky voiced vowels. Stress was regularly word-initial, as in Tumbleweed.


Major sound changes

The two branches, Grassland Click and Northern Click, are located on either side of the Kwreid Isthmus, suggesting that historically their territory was much larger. The Grassland Click languages are characterised by a major sound change; the metathesis of clicks from the C₂ position to the start of a word, with subsequent repair. This led to the development of click voicing contours:

proto-Click *ðˤək͡ǀ͜χ’u
> *ðˤəɢ͡k͜ǀ͡χ’u (spread of voicing from vowel to onset of click)
> *ɢ͡k͜ǀ͡χ’ðˤəu (metathesis to word boundary)
> *ɢ͡k͜ǀ͡χ’əðˤəu (echo vowel to break up cluster)
> proto-Grasslands Click *ɢ͡k͜ǀ͡χ’oðˤo (vowel smoothing)

This explains how Grasslands Click languages such as Gḳǁhɨgḳǁhɨ and Łakg’akg’u have inventories of up to fifty clicks. Proto-Grasslands had three click voicing contours; voiced to fortis voiceless, voiced to lenis voiceless, voiced to ejective:

proto-Click *wak͡!ˤe → proto-Grasslands *ɢ͡k͜!ʷˤʰɛwɛ
*kək͡ǃ͜χ’a → *ɢ͡k͜!͡χ’ɐkɐ
*ɬiɡ͡ʘˤi → *ɢ͡ɡ̊͜ʘˤiɬi

In contrast, Northern Click underwent a chain shift exemplified by *t͡s’ k͡ǃ͜χ’ k͡! ɡ͡ǃ → *ʔ t͡ʃ’ k͡ǃ͜χ’ k͡!, i.e. ejectives were debuccalised to /ʔ/, glottalised clicks became ejectives, plain clicks were glottalised and voiced clicks were devoiced. In most descendants nasal clicks were innovated, and in a few voiced clicks were restored through lenition.

proto-Click *q͡χ’eje → proto-Northern Click *ʔeɟe
*ðˤək͡ǀ͜χ’u → *dˤət͡s’u
*k͡ǁˤuɣa → *k͡ǁ͜χˤ’uɡa
*ɡ͡ʘˤolə → *k͡ʘˤolə

Next post will be on more sound changes for each branch. I've told myself I'll stop procrastinating on this and post regularly, but in all honesty I probably won't, so don't be surprised if there's no more posts for a while.


Word structure

Word structure is quite rigid in the Click languages, which is interestingly similar to proto-Tumbleweed. Proto-Click roots could be one of two forms, either *C₁V₁C₂V₂ or *C₁V₁C₂V₁C₂V₂, for example *χɨbˤo "cockle shell", *ɡ͡!ˤəŋəŋi "fasten with a strap". Only some grammatical forms (*fe "1SG.MASC", *ɡ͡ʘˤu "near-hearer demonstrative") and borrowed words (*walɨtˤə "Roudan" (place name) < late proto-Trintinic *rᵒálr̥ᵉtᵃn̥, *befəɣi "bird sp." < proto-Kwreid Isthmus *bɛɸəɣi) did not conform to these structures, and even then they tended to assimilate (e.g. proto-Grasslands *beɣeɣi < *befəɣi). In Grasslands Click the two vowels of bisyllabic roots tended to assimilate unless they were of the same height, e.g. *s’aɢi > *s’eɡe vs. *deɡ͡!ˤo > *ɢ͡ɡ̊͜!ˤedo.
While reduplication was not productive in proto-Click, there are some vestiges that suggest it was at an earlier stage, and these comply to the trisyllabic structure e.g. *k͡ǁˤaqaqu "swim" <~ *k͡ǁˤaqu "kick", *t͡ɬ’ibibə "harvest" <~ *t͡ɬ’ibə "shave", *kok͡!ok͡!e "prepare food" <~ *k͡!oke "rub".
Morphologically complex words on the other hand did not follow any set structure, e.g. *-k͡ǃ͜χ’a "3SG.SUBJUNCT" could be applied to *k͡!oke-k͡ǃ͜χ’a or *ɡ͡!ˤəŋəŋi-k͡ǃ͜χ’a without requiring any modification. In some descendants this was changed to adhere more consistently to the word structure rules.


Morphology

Click displays a greater degree of concatenative morphology than most of its neighbours, in both verbs and nouns. Verbs were marked for tense, aspect and mood, subject and object agreement and valency, while nouns marked gender, case and number. In contrast, proto-Nomadic and proto-Kwreid were both almost isolating with very limited morphology, and early proto-Trintinic had almost zero nominal morphology, although extensive verbal morphology. Proto-Click was exclusively suffixing, as are most descendants although some Northern Click languages have innovated a proclitic series.

I'll leave the actual suffixes for a later date, but verbs have the following structure:

ROOT-valency-subject P/N/G-object P/N/G-tense/aspect-secondary aspect-mood

While nouns are as follows:

ROOT-case/number/gender-case

Click languages also have a full class of adjectives which agree for gender only with the noun.


Derivational morphology

Derivational morphology primarily consists of switching the part of speech of a root. The following shifts could be performed:

• verbal root => noun
• verbal root => adjective
• conjugated verb => noun
• nominal root => verb
• nominal root => adjective

Syntax

I'm absolutely shitarse at syntax, but briefly speaking:

- Proto-Click was most likely AVO/SV~VS, with an active-stative alignment based on the concept of control and being lexically determined by the verb.
- Syntax was flexible due to case marking on nouns
- Noun phrases are head-final, while verb phrases are head-initial
- Relative clauses are formed through nominalisation of a phrase with some kind of method.

---

I'll flesh this out a lot more. Right now I'm trying to work out some sound changes I'm happy with. Here's a spoiler of that:

Eastern Grasslands is characterised by a rightwards shift of the voicing contour on clicks, with proto-Grasslands pre-voiced clicks *ɢkǀʰ ɢkǀχ’ ɢɡ̊ǀ etc. becoming post-voiceless clicks *ɡǀχ ɡǀχ’ ɡǀq, which lost their click phonation in the easternmost descendants leaving plain velars /q͡χ q͡χ’ q/, e.g. Qgok!’o /q͡χˤ’o/ < proto-Grasslands *ɢkǀχ’oðˤo. Other Eastern Grassland languages retain the click clusters (G!ama /ɡ͡ǀqo̰lo̰/, Kǁqgwem‘e /ɡ͡ǀq͡χ’oɗo/) or insert an echo vowel (Gǁuħũbũ /ɡ͡ǀoqo̰də̰/).

[sangi39]

I really like this! The singular instance of a bilabial click, /ɡ͡ʘˤ/ stands out, but with the lack of /p/ in the pulmonic consonants it's not actually odd at all. The word stucture is pretty cool as well, and I like the sort of internal history of the trisyllabic roots without also relying on that as the sole origin of the structure

Beyond that, there's not much I can say. It's all just positive "I like this so much", haha

[VaptuantaDoi]

I really like this! The singular instance of a bilabial click, /ɡ͡ʘˤ/ stands out, but with the lack of /p/ in the pulmonic consonants it's not actually odd at all. The word stucture is pretty cool as well, and I like the sort of internal history of the trisyllabic roots without also relying on that as the sole origin of the structure

Beyond that, there's not much I can say. It's all just positive "I like this so much", haha

Thanks! This is one of the few languages I've made where I'm happy with the aesthetic (for now anyway).

As a quick update, I finally made us all a map! It's not the most artistic map in the world but it should give yous an idea of where all the languages I'm talking about are situated. The (small) continent shown is Rodgezhua, which is connected by the Kwreid Isthmus to a large swath of land inhabited by the Re, a species of giant sentient corvids who are pretty warlike and don't live alongside humans. Bear in mind that its small size is due to Kokhene (the planet) itself being quite a bit smaller than Earth at 0.591RE, which means that comparatively Rodgezhua's not actually that small.

[VaptuantaDoi]

Proto-Click Morphology

My diachronic conlanging always works best when I actually make the proto-lang before all of the descendants (whod've guessed?), so this time I'm gonna completely ignore sound changes until I've got at least basic morphology down.


Verbal morphology

As I mentioned previously, the proto-Click verbal template was like this:

ROOT-valency-subject P/N/G-object P/N/G-tense/aspect-secondary aspect-mood

Roots were either two syllables (*ɬˤɨma "enter", *fula "sing", *kǁχ’ɨtˤi "whistle") or three syllables (*ɡʘˤuŋuŋo "confront", *dˤɨɢɨɢe "pierce with a needle", *k!omomɨ "carve"), with only a couple of exceptions which tend to be onomatapoeia (*bˤɨbˤɨbˤɨkɨ "stammer", *ɡʘˤi "kiss", *kǁo "wallow") or borrowings from other languages (*bɨtololə "shuck" < proto-Kwreid *bɨtoɣəɾə, *qakakele "ride" < proto-Nomadic *kakadeke).


Valency marking

While functionally, the placement and semantics of the valency markers was very similar to proto-Tumbleweed, being placed immediately after the verb and marking an antipassive and a causative, they were formally completely different. This slot was obligatorily filled in proto-Click with single consonant suffixes which fell into four classes:

Class I: Default *-t-, antipassive *-ts’-, causative *-dˤ-
Class II: Default *-k-, antipassive *-kx’-, causative *-ʕ-
Class III: Default *-s-, antipassive *-s’-, causative *-ðˤ-
Class IV: Default *-kǀ-, antipassive *-kǀχ’-, causative *-ðˤ-

The default consonant was effectively unpredictable, but the antipassive and causative could always be determined from the default. Nevertheless it makes sense not to count the default consonant as part of the root because a) all other bisyllabic roots are CVCV-, b) even verbs which were clearly onomatapoeic or borrowed take these consonants, and c) verbs regularly derived from nominal roots also take these suffixes. However, in a lexicon I'd give roots with their default consonant e.g. *fula-s-, *k!omomɨ-t- etc.
A relatively large number of roots are only attested in non-default forms, for example *dˤaɬu-ʕ- "kill", *kədədi-ts’- "look at", *k!ˤetˤi-dˤ- "beckon". There are too many of these for them to simply reflect gaps in our knowledge of proto-Click; it seems that some roots were inherently causative or antipassives (with the number of inherent causatives far outweighing inherent antipassives).


Subject P/N/G

The next set of affixes were subject markers for person, number and gender. These suffixes were of the forms -V- or -VCV- with a repeated vowel, and there was only one set regardless of the valency class. Note that gender was distinguished in the third person, with masculine, feminine, animate and inanimate. The masculine was used for male humans only, the feminine for female humans or female animals, the animate for male animals or animals with indestinguishable gender, and inanimate for everything else. In the plural, the masculine and feminine merged into the common. Also bear in mind that these suffixes are only used with about 75% of intransitive verbs, the rest of which take "object" suffixes.

1sg. *-i
2sg. *-ɨ
3sg.m. *-a
3sg.f. *-əðə
3sg.anim. *-iði
3sg.inan. *-itˤi
1pl. *-e
2pl. *-u
3pl.com. *-o
3pl.anim. *-oʕo
3pl.inan. *-aqa

In this case they're far closer to the proto-Trintinic person/number infixes, 1sg. *-i- 1pl. *-e- 2sg/pl. *-u- 3sg. *-a- 3pl. *-o-, although given that they're all single vowels this could simply be a coincidence. The other possibility is that one of them borrowed from the other, as the two languages are certainly not related. However, compare 3sg.f. *-əðə to proto-Tumbleweed 3.I.SG -DHV- and 3pl.inan. *-aqa to 3.IV.PL -ngkV-.

Object P/N/G

These suffixes all have two distinct allomorphs depending on their position, although they are clearly closely related. When used immediately following valency markers, i.e. for stative intransitive verbs, they were all of the shape -VCV, again with a repeated vowel. Note that the masculine and feminine were also merged in the singular for these suffixes.

1sg. *-aja-
2sg. *-ɨðˤɨ
3sg.com. *-oŋo
3sg.anim. *-ɨðɨ
3sg.inan. *-ɨjɨ
1pl. *-eɡʘˤe
2pl. *-ɨtɨ
3pl.com. *-iɬi
3pl.anim. *-ini
3pl.inan. *-ɨmˤɨ

The second set of allomorphs was used when following the subject markers, the first vowel was dropped, hence *wˤimu-s-i-mˤɨ "I know it", *qakakele-k-a-ðɨ "she is riding it". In transitive sentences these suffixes were only obligatory when there was not a specified nominal object; in intransitive (i.e. stative) sentences they were still required.


Tense

Tense suffixes were optional and pretty simple. Unmarked verbs were assumed to be present tense, while past and future could both be marked:

Past: *-fɨ
Future: *-tˤɨ

Yep, that's about it.


Aspect

As suggested by the slot template, verbs distinguished two types of aspect. The first was a binary perfective/imperfective distinction, with the perfective used for events viewed as an indivisible whole, and the imperfective for events with internal heterogeneity. These were optional, but if they did occur they had to be combined with secondary aspect markers. As such, they could both be single consonant suffixes:

Perfective: *-s-
Imperfective: *-w-

Secondary aspect introduced a number of further distinctions. Underlyingly, there were four of these:

Durative: *-a - events which last for a long period of time
Punctual: *-i - events which last for a short period of time
Discontinuous: *-e - refers to an event which is not true at the present time
Nondiscontinuous: *-ə - refers to an event which is true at the present time (there's gotta be a better term for this)

The way these combine with the primary aspect distinction and tense is somewhat complicated, and allows for a number of fine distinctions:

*tewa-ts’-a-fɨ-s-e
eat-ANTIPASS-3.SG.M.SUBJ-PAST-PFV-DISCONT
"He has eaten (and he's since stopped)"

*tewa-ts’-a-fɨ-w-e
eat-ANTIPASS-3.SG.M.SUBJ-PAST-IPFV-DISCONT
"He was eating (and he's since stopped)"

*tewa-ts’-a-fɨ-w-ə
eat-ANTIPASS-3.SG.M.SUBJ-PAST-IPFV-NONDIS
"He has been eating (and he still is)"

The discontinuous cannot be combined with the present tense for obvious semantic reasons; however it can be combined with the future:

*kǀebə-kǀ-əðə-ðˤɨ-tˤɨs-e
hit-DEFAULT-3.SG.F.SUBJ-2.SG.OBJ-FUT-PFV-DISCONT
"She will hit you (for this first time)"

While the punctual and durative are similar in meaning to the perfective and imperfective respectively, there's a clear semantic distinction and they can be used with either primary aspect:

*məke-t-e-fɨ-s-a
swim-DEFAULT-1PL.SUBJ-PAST-PFV-DUR
"We'd been swimming for a while."

*mats’ɨ-kx’-a-w-i ɡ!ˤikǀo-n-e, kiwə-t-a-s-e bˤɨbɨbə-ŋ-o
close-ANTIPASS-3.M.SG.SUBJ-IPFV-PUNCT eye-ANIM.SG-DAT, see-DEFAULT-3.M.SG.SUBJ-PFV-PUNCT father-ACC.SG-ACC
"While blinking his eyes, he sees his father."


Mood

Mood suffixes are a bit weird in proto-Click. They marked a four-way distinction between irrealis, conditional, subjunctive, and negative (which behaves exactly like a mood) and also redundantly marked third person singularness - i.e. they had two distinct and seemingly unrelated forms, one set used in the 3sg and the other used for all other persons and numbers.

Irrealis: 3sg. *-sə, other *-kǀo
Conditional: 3sg. *-wi, other *-mu
Subjunctive: 3sg. *-kǃχ’a, other *-qχ’e
Negative: 3sg. *-no, other *-ʕi

The realis was the default unmarked case. The irrealis marked events which are not known to be true in a very general sense:

*tɬ’ibibə-kǀ-a-jɨ-sə
harvest-DEFAULT-3.M.SG.SUBJ-3.INAN.SG.OBJ-IRREAL
"He might harvest it."

While the conditional was only to set up a hypothetical scenario which another clause would explain the result of:

*tɬ’ibibə-kǀ-a-jɨ-no tewa-t-e-jɨ-tˤɨ
harvest-DEFAULT-3.M.SG.SUBJ-3.INAN.SG.OBJ-COND eat-DEFAULT-1.PL.SUBJ-3.INAN.SG.OBJ-FUT
"If he harvests it, we'll eat it."

The subjunctive also has an irrealis sense but is only used in dependent clauses, which I'll discuss more when (if) I get to syntax.

---

This post is long enough as it is, so I'll leave nominal morphology for the next post.

[Omzinesý]

I'm not very good at conlang families. I get messed up with many lang in the same thread.

But interesting stuff!

Is Proto-Clich so agglutinative because it is a protolang?
I think languages with more fusion are morphologically more interesting.

Is no primary or secondary aspect unmarked?
The aspect system is very interesting. I have never really understood the Navaho system. In a conlang the stuff must be more clearly presented.
Some extra explanation would still do but I understand that this is just a protolang.

[VaptuantaDoi]

Is Proto-Clich so agglutinative because it is a protolang?
I think languages with more fusion are morphologically more interesting.

Yeah, you've got a point. I tend to make agglutinative protolangs cos i dont like making fused affixes without having some idea of how they evolved. Anyway proto-Click is much more fusional than proto-Tumbleweed, and its descendants will be pretty fusional.

Is no primary or secondary aspect unmarked?
The aspect system is very interesting. I have never really understood the Navaho system. In a conlang the stuff must be more clearly presented.
Some extra explanation would still do but I understand that this is just a protolang.

Verbs can be unmarked for aspect, but if they are marked they have to take both primary and secondary suffixes. I'll elaborate on the specific semanitcs of aspects in a later post. Basically the underlying distinctions are:

Perfective vs. Imperfective - whether something changes within the time an action is happening
Durative vs. Punctual - how long on an absolute or relative scale an action goes on for
Discontinuous vs. Nondiscontinuous - whether the state resulting from an action is true at the present time

When I post about aspectual semantics I'll give more specific examples of how these interact with each other and with tense, and when verbs unmarked for aspect are used. Also this will be an important development in the more fusional descendants.

[Arayaz]

[snip]

Consonants:
/b bˤ t tˤ d dˤ k q ɢ/
/t͡s’ t͡ɬ’ k͡x’ q͡χ’/
/f ð ðˤ s ɬ ɬˤ ɣ χ ʕ/
/s’ ɬ’/
/m mˤ n ŋ/
/w wˤ l j/

[snip]

First of all, I love this langauge so far! I haven't made a click language in a really long time. Maybe I should...

It's peculiar in my opinion that you have /ɢ/ but no regular voiced velar. But maybe there's an explanation I didn't see?

[VaptuantaDoi]

First of all, I love this langauge so far! I haven't made a click language in a really long time. Maybe I should...

Thanks! I used to stay away from clicks for some reason but now I can't stop adding them to all my langs

It's peculiar in my opinion that you have /ɢ/ but no regular voiced velar. But maybe there's an explanation I didn't see?

Yeah that is weird, no I don't really have an explanation for it and tbch I don't urgently feel the need for one. Depending on whether I pursue the Tumbleweed connection it might be something like this:

*ᵐb ᵐbˤ ⁿd ⁿdˤ ᵑɡ ᵑɡˤ → *b bˤ d dˤ ŋ ɢ
*ᵐp ᵐpˤ ⁿt ⁿtˤ ᵑk ᵑkˤ → *f wˤ t tˤ k q
*p pˤ t tˤ k kˤ → *w wˤ ð ðˤ ɣ ʕ

That's a bit bludgey tho. I guess if I really feel the need I can stick *ɡ in there but proto-Click is less of a focus than all its descendants, most of which will either gain /ɡ/ or drop /ɢ/.

[VaptuantaDoi]

Nominal morphology

I'm trynna power through proto-Click so as I don't get distracted and procrastinate. This means I won't be going into much depth with anything but hopefully the descendants will be better for it (i.e. they'll actually exist).

As in basically all languages nouns display less morphological complexity than verbs. I mentioned in the first post that the nominal slot template is something like this:

​ ​ ​ ​ ​ ROOT-case/number/gender-case


Roots

Nominal roots are similar to verbal roots in being either CVCV (*wɨkǀə "body", *ɣoɬo "cooking pot", *fakə "rain") or CVCVCV (*ɬabˤabˤi "cloth", *kǃɨðɨðe "sweat", *ɡǃˤokǃokǃu "item of jewellery"), excluding borrowings (*ʕelaɡǃo "bone fishhook" < proto-Nomadic *edadkue, *ðaboko "sand dune" < Old Soraw *rāpǫk). These were then marked for case, number and gender.


First suffixes

The first set of suffixes consisted of single consonants, and marked gender, number and case. These suffixes only distinguished oblique vs. nominative, with the oblique being differentiated by the second set of suffixes. There was considerable syncretism amongst these.

​ ​ ​ ​ ​ Masc.sg.obl. *-ŋ-
​ ​ ​ ​ ​ Fem.sg.obl. *-ɢ-
​ ​ ​ ​ ​ Anim.sg.obl., Masc./Fem./Anim.pl.obl. *-n-
​ ​ ​ ​ ​ Inan.sg./pl.obl. *-w-
​ ​ ​ ​ ​ Masc.sg.nom. *-tˤ-
​ ​ ​ ​ ​ Fem./Anim.sg.nom *-m-
​ ​ ​ ​ ​ Inan.sg.nom. *-wˤ-
​ ​ ​ ​ ​ All genders.pl.nom. *-f-

(Boy I wish we could make tables on this board)

The gender marking in these suffixes was very rarely productive, i.e. the gender of a noun could not be changed through these suffixes alone. In this sense they were entirely redundant as the gender was inherrent on the noun already.


Second suffixes

The second set of suffixes only encoded case, and consisted of single vowels except for the locative. The nominative marker was originally *-i, although it seems that later the instrumental *-a was extended into the nominative, used with either the nominative or oblique consonants (reflexes differ between languages). This originally arose in the ergative sense (i.e. only with active verbs) by a relatively common semantic drift.

​ ​ ​ ​ ​ nom. *-i, later *-a in some branches
​ ​ ​ ​ ​ acc. *-o (also an absolutive in stative intransitive clauses)
​ ​ ​ ​ ​ dat. *-e
​ ​ ​ ​ ​ loc. *-ɨ or *-u
​ ​ ​ ​ ​ instr. *-a

The split between locative *-ɨ and *-u is a major division in the family, with the Northern Click languages reflecting *-ɨ and the Grasslands Click languages reflecting *-u (a few vestiges of *-ɨ appear in Grasslands, suggesting that it is the older of the two and *-u is an innovation; see also footnote 3).
The only instance in which roots occurred without these suffixes was in the object role of a transitive sentence, which may have carried some semantic implications. Some descendants in fact only preserve zero-marked accusatives and relegate the "marked accusative" into the absolutive role (subject of stative verbs). This is the citation form for nouns, given that it is the most morphologically basic.


Gender-based derivational morphology

While derivational morphology as a whole will be described in another post, here I'd like to mention the ways in which noun roots can be shifted between genders through the series of prefixes shown below. In most cases this came with a very direct semantic correlation, i.e. a change in nominal gender represented a genuine reassignment of said noun to the target gender; other cases were more idiomatic. This process was for obvious reasons most productive in the masculine and feminine forms.

​ ​ ​ ​ ​ masc. *ŋV-
​ ​ ​ ​ ​ fem. *ɢV-
​ ​ ​ ​ ​ anim. *nV-
​ ​ ​ ​ ​ inan. *ʕV-

Some degree of similarity between these prefixes and the first set of case/gender/number suffixes is apparent. Note that when these suffixes were applied, all words derived from C₁V₁C₂V₂-form roots which can be attested back to the proto-Click stage underwent repair to conform to the C₁V₁C₂V₁C₂V₂ root form. This entailed filling the vowel slot of the prefix with the first vowel of the root, and replacing the second consonant of the root and replacing the second consonant of the root with a copy of the first consonant, hence:

​ ​ ​ ​ ​ *ɬˤamˤɨ "goat ANIM" > *ŋaɬˤaɬˤɨ "billy goat MASC, *ɢaɬˤaɬˤɨ "she-goat FEM", *ʕaɬˤaɬˤɨ "goat meat"¹
​ ​ ​ ​ ​ *ts’ɨwa "fisherman MASC" > *ɢɨts’ɨts’a "fishermanwoman", *nɨts’ɨts’a "heron"
​ ​ ​ ​ ​ *s’ojo "stick INAN" > *nos’o’o "stick insect", *ŋos’o’o "baby with disease characterised by extremely skinny limbs"

For C₁V₁C₂V₁C₂V₂ roots, the prefixes simply overrode the first consonant of the word:

​ ​ ​ ​ ​ *wˤots’ots’ɨ "deer ANIM" > *ŋots’ots’ɨ "buck MASC", *ɢots’ots’ɨ "doe FEM, *ʕots’ots’ɨ "venison".
​ ​ ​ ​ ​ *χekeka "dust particles INAN" > *nekeka "midge ANIM"


Pronominal morphology

Also highly contrastive to the East Rodgezhuan languages,² Click shows a fully developed pronoun system even at the earliest stage. The gender distinctions were well maintained in the third person but did not feature in the second person, and only distinguished masculine vs. feminine in the first person. The table below shows the oblique and nominative roots:

​ ​ ​ ​ ​

The oblique stems took the same four oblique vocalic suffixes (acc. *-o, dat. *-e, loc. *-u,³ instr.*-a); for instance *fo "me ACC", *te "to you DAT", *moqu "by him LOC", *ʕa "using it INSTR". Note that the extension of the instrumental to the nominative did not occur in pronouns likely due to the semantic rarity of instrumentals in the non-third person and their irregular nominative forms.
The locatives are also used as a possessive series which in some descendants are treated as adjectives and in other descendants as nouns. It is possible that the ending *-u was originally a genitive in contrast to the locative *-ɨ, which became generalised to the locative first in pronouns then later in nouns.


Adjectival morphology

Adjectives form a somewhat separate class to nouns, although they are definitely related. They're in fact highly SAE, although that makes them typologically unusual for the region. They have two sets of prefixes which mirror the nominal ones but are more consistently distinguished for gender and number; the first set are shown below while the second set is identical to the nominal set (including the nominative *-i/-a and locative *-ɨ/-u alterations; in no languages are their different reflexes consistently seen between nouns and adjectives).


​ ​ ​ ​ ​


These agree with the noun they modify, or can appear bare when used with the copula.

---

1) As seen here the inanimate often had an association with mass nouns, cf. *ʕeɡǃeɡǃi "milk INAN" < *ɡǃeji "cow FEM"; in the Gǁuħunbunic branch the INAN/ANIM distinction was re-formed to one of mass vs. count nouns e.g. ɡʘoqoo "stone MASS" vs. noobo "a rock COUNT" < pClick *wekǃu, *ne-wewu.

2) Just to be clear, "East Rodgezhuan" is the ancient phylum representing the second wave of migration (c. 20000 BC) to Rodgezhua and contains Trintinic, Kwreid Isthmus, Bechsukchwan-Nomadic, Decééyinéeqi and Ánni, while "West Rodgezhuan" is an even older phylum from the first wave of migration (c. 30000 BC) consisting of just Tumbleweed and Click. From an far earlier stage (the first human arrival on Rodgezhua c. 200000 BC) is the indigenous isolate Qqeń'iłla which may have some connection with the Ąkuyook Island languages Ǫneebǫwhítw and Icááwənun.

3) In the pronominal system *-u is ubiquitous even in Northern Click, suggesting that in Grasslands Click nominal *-u may have arisen from pronouns.

[VaptuantaDoi]

Who wants to completely restart this family from scratch? I do! Also since I find it amusing I will be making proto-Click2 peoples unreasonably advanced, with unanalysable words for things like "earplug" and "download speed" and "ballpoint pen". Presumably modern Click2 peoples will be even more advanced, or perhaps they were just idiots who've been cruising on their ancestors' technology. Also technological levels are all super fucked up in Kokhene because of time travel (don't worry, Earth's will have been too once we've invented time travel).

PHONOLOGY

FEET
The proto-Click2 word is based around feet (not the stinky ones with toes, but the ones poetry is made of, without toes). There are two permissible foot types, trochaic (stressed-unstressed) and dactylic (stressed-unstressed-unstressed), both made up of (C)V syllables, with obligatory word-initial onsets. Dactyls have an interesting property of always having a repeated VC sequence in the middle of them.

Some trochees:
*kǃʰoβi "gear, equipment"
*pɛŋ̊ǁˀɛ "to pair"
*beɾi "zipper"

Some dactyls:
*ŋǂeβeβɛ "perfume"
*dinini "potash"
*soβoβe "fingertip"

Each foot has one of four tones; High, Low, Rising and Falling. The tone at the right edge of a word must always end up low.

Code: Select all

    ⋕===================‡===================⋕
    ǁ      TROCHEE      |      DACTYL       ǁ
    ⧺-------+-----------+-------+-----------⧺
    ǁ final | non-final | final | non-final ǁ
⋕===⋕===================‡===================⋕
ǁ H ǁ  H-L  |    H-H    | H-H-L |   H-H-H   ǁ
⧺---⧺-------+-----------+-------+-----------⧺
ǁ L ǁ        L-L        |       L-L-L       ǁ
⧺---⧺-------+-----------+-------+-----------⧺
ǁ R ǁ  L-F  |    L-R    | L-H-L |   L-L-H   ǁ
⧺---⧺-------+-----------+-------+-----------⧺
ǁ F ǁ        H-L        | H-L-L |   H-H-L   ǁ
⋕===⋕===================‡===================⋕

However, it turns out that dactyls are actually just reduplicated trochees with the two touching edges removed:

|soβe-soβe| → /soβoβe/
|dini-dini| → /dinini/

In morphologically complex words, syllables tend to rearrange themselves into neat bunches of feet. To achieve this, we first need to note that there are some monosyllabic affixes (actually there are some subsyllabic affixes too, but they turn into monosyllablic chunks so it's fine). These mess shit up real bad, but it's repairable.
We start with the underlying form, which means all the dactyls are replaced by their underlying bitrochees. First off, a morphologically complex word is chopped up into trochees, which is generally possible, but sometimes leaves a little fucker off the right edge:

Code: Select all

ˈdoɣɑL ˈpekǀʰɛF ˈmukǀʰɛF ˈmuL
 doɣɑ-       pe-      kǀʰɛmu.kǀʰɛmu
  R           L          F      F 
COMMISS 1SG.A/3M.SG.O     cook

We see here that tone is actually assigned to the initial syllable rather than the whole foot, which makes sense since that's the stressed one. The last tone is then taken as the tone of the foot, and anything left without a tone is given L.
All pairs of trochees then turn into dactyls, by deleting the right edge of the first and the left edge of the second, and also the second vowel assimilates to the first. In proto-Click2 itself, the resulting dactyl wasn't forced to have the repeated VC sequence, but this later happened in almost all descendants (which royally fucked up morphology).

Code: Select all

ˈdoɣokǀʰɛF ˈmukǀʰɛF ˈmuL
ˈdoɣɑL ˈpekǀʰɛF ˈmukǀʰɛF ˈmuL
 doɣɑ-       pe-      kǀʰɛmu.kǀʰɛmu
  R           L          F      F 
COMMISS 1SG.A/3M.SG.O     cook

This still leaves the fucker ˈmu, which really needs to be footed. To fix this, it's just tagged onto the end of the trochee before it (if there is one), making a dactyl, again spurring vowel harmony.

Code: Select all

ˈdoɣokǀʰɛF ˈmukǀʰumuL
ˈdoɣokǀʰɛF ˈmukǀʰɛF ˈmuL
ˈdoɣɑL ˈpekǀʰɛF ˈmukǀʰɛF ˈmuL
 doɣɑ-       pe-      kǀʰɛmu.kǀʰɛmu
  R           L          F      F 
COMMISS 1SG.A/3M.SG.O     cook

Finally, we have to fix up a few things, because there are heavy restrictions on root-medial consonants. Foot-initial /p b t d s x/ correspond to foot-medial /β m ɾ n ɣ Ø/. Now we can get the surface form (remembering the tone assignment procedures):

*Dóɣ-óʇh-ɛbuʇhumu

Code: Select all

ˈdóɣókǀʰɛ̀ˈbùkǀʰùmù
ˈdoɣokǀʰɛF ˈmukǀʰumuL
ˈdoɣokǀʰɛR ˈmukǀʰɛL ˈmuL
ˈdoɣɑR ˈpekǀʰɛL ˈmukǀʰɛL ˈmuL
 doɣɑ-       pe-      kǀʰɛmu.kǀʰɛmu
  R           L          F      F 
COMMISS 1SG.A/3M.SG.O     cook

Now this is all well and good, but what happens if the little fucker at the right edge is following a dactyl already? Well this can't be allowed, so the offending monosyllabic affix (or clump of affixes) gets a trochaic form, usually just reduplicated, but in some cases (such as polypersonal agreement) unpredictable. The trochaic form of pe- is paβe-:

*Pávɛ́-muʇhɛ́mu

Code: Select all

ˈpáβɛ́mù ˈkǀʰɛ́mù
ˈpaβɛmuF ˈkǀʰɛmuF
ˈpaβeL ˈkǀʰɛmuF ˈkǀʰɛmuF
    paβe-     kǀʰɛmu.kǀʰɛmu
      L          F      F 
1SG.A/3M.SG.O     cook

So to recap, we've seen two forms of "cook", one ɛˈbuʇhumu and one muˈʇhɛ́mu. In isolation, it's ˈʇhɛ́mɛmu. There might be some other surface forms, but I'll have to work that shit out later. Long and short of it is, people trying to learn proto-Click2 are fucked. Even more fucked when it comes to some of the descendants.


CONSONANTS

At a surface level, if feet are ignored, there are the following consonants:

Code: Select all

           ⋕=====‡=====‡=====‡=====‡=====⋕
           ǁ Lab | Den | Alv | Lat | Dor ǁ
⋕==========⋕=====‡=====‡=====‡=====‡=====⋕
ǁ Nas  +vc ǁ  m  |     |  n  |     |     ǁ
⧺----------⧺-----+-----+-----+-----+-----⧺
ǁ Stop -vc ǁ  p  |     |  t  |     |     ǁ
ǁ      +vc ǁ  b  |     |  d  |     |     ǁ
⧺----------⧺-----+-----+-----+-----+-----⧺
ǁ Cont -vc ǁ     |     |  s  |     |  x  ǁ
ǁ      +vc ǁ  β  |     |  ɾ  |     |     ǁ
⧺----------⧺-----+-----+-----+-----+-----⧺
ǁ     +asp ǁ     | kǀʰ | kǃʰ |     |     ǁ
ǁ Cli +nas ǁ     | ŋǀ  | ŋǃ  | ŋǁ  | ŋǂ  ǁ
ǁ     +glo ǁ     |     | ŋ̊ǃˀ | ŋ̊ǁˀ | ŋ̊ǂˀ ǁ
⋕==========⋕=====‡=====‡=====‡=====‡=====⋕

These can be very clearly split up into two groups, "oral" and "click". The former of these can be further divided up into "foot-initial" and "foot-medial"; clicks can occur in any position. There is a consistent correspondence between foot-initial and foot-medial consonants:

Code: Select all

 Foot-initial  p  t  b  d  s  x
 Foot-medial   β  ɾ  m  n  ɣ  Ø

This lets us draw up the following oral consonant inventory:

Code: Select all

           ⋕=====‡=====‡=====⋕
           ǁ Lab | Alv | Dor ǁ
⋕==========⋕=====‡=====‡=====⋕
ǁ Stop -vc ǁ p~β | t~ɾ |     ǁ
ǁ      +vc ǁ b~m | d~n |     ǁ
ǁ Fric -vc ǁ     | s~ɣ | x~Ø ǁ
⋕==========⋕=====‡=====‡=====⋕

I'll take |p t m n s x| to be the archiphonemes; |p t s x| then elide foot-internally, and |m n| denasalise foot-initially (both normal processes). Most descendants shifted /t n/ to /k ŋ/ to heighten the stop-click distinction, and in fact the following may be taken as proto-typical Grasslands Click:

Spoiler:

Code: Select all

 p     k
 b  d
 m  ɾ  ŋ
 ɸ  s  x

Returning to proto-Click2, what's left is the click consonants:

Code: Select all

              ⋕=====‡=====‡=====‡=====⋕
              ǁ Den | Alv | Lat | Pal ǁ
⋕=============⋕=====‡=====‡=====‡=====⋕
ǁ Aspirated   ǁ kǀʰ | kǃʰ |     |     ǁ
ǁ Nasal       ǁ ŋǀ  | ŋǃ  | ŋǁ  | ŋǂ  ǁ
ǁ Glottalised ǁ     | ŋ̊ǃˀ | ŋ̊ǁˀ | ŋ̊ǂˀ ǁ
⋕=============⋕=====‡=====‡=====‡=====⋕

In total proto-Click2 has 15 consonant phonemes, a far cry from the original 43, but I think a more interesting system.

VOWELS
Proto-Click had vowels, as many languages do.

Code: Select all

       ⋕=======‡=======⋕
       ǁ Front | Back  ǁ
⋕======⋕=======‡=======⋕
ǁ High ǁ   i   |       ǁ
ǁ Mid  ǁ   e   |   o   ǁ
ǁ Low  ǁ   ɛ   |   ɑ   ǁ
⋕======⋕=======‡=======⋕

I don't know if this system is attested in any natlangs, but it's not far off Onondaga's oral vowels (/i e o æ a/) and in the end it's just a square system with a corner missing so I'm keeping it.