The CBB

I figured out, for the most part, the phonological development of Ábḫanni, which should span at least a thousand years. Unlike in some of my older posts I'm not really interested in the older forms of Ábḫanni but am simply developing the history in order to be able to have realistic and interesting morphophonology and irregularities.

So in this post I'm just showing off the very first results of the historical development I came up with, applied to the most basic active verb morphology, which shows off some of the complexities and irregularities of the agglutination system of Ábḫanni. Fundamentally, the system is still agglutinative as before, but is pervaded by certain morphophonological processes (notably, movement of the pitch accent away from the root) that make it sometimes hard to predict what the surface form of a given conjugation will be.

The endings that I'll show are as follows:

-Ø / -n / -sí / -n-sí / -tí / -n-tí / -ttámo / -n-ttámo / -riyá (see two posts above for the rough grammatical meanings)

Note that -n-sí always (to my knowledge) surfaces as -ssí (or -ssi, if stress is retracted).

Here's the (most basic) conjugation of the verb mére, "to give back/return". This root has little phonological change in its conjugation but gives the basic idea (note the front/back harmony of the central vowels with the -ttámo ending and the rVr > rVn dissimilation):

mére / méren / meresí / meressí / meretí / merentí / merettáme / merenttáme / mereniyá

Here's a conjugation of the hypothetical (i.e. I haven't given it a meaning) verb meré, which displays the stress retraction from the endings (following the phonological diachronics directly, the -riyá form should be mereniyá, without retraction, but I think regularizing verbs of this type to have static stress is pleasing. The rule for the stress contraction is, basically, that the stress must move forward (it only ever moves forward) by at least two syllables):

meré / merén / merési / meréssi / meréti / merénti / meréttame / merénttame / meréniya

Here's the conjugation of ḫáza, "to fall". This is an example of a verb with a fairly standard conjugation, except that the movement of the stress away from the root gives rise to a "weak stem" aḫza-:

ḫáza / ḫázan / aḫzasí / aḫzassí / aḫzatí / aḫzantí / aḫzattámo / aḫzanttámo / aḫzariyá

Here's a more complex conjugation, of the hypothetical verb ḫáz, which has two weak stems: a nonfuture weak stem aḫ- and a future weak stem aḫzo-. The epenthetic-looking -o- isn't truly epenthetic, since it is the remains of what used to be a root-final vowel. Note the degemination in the -ttámo form; geminates in clusters are restricted to nasal + long unvoiced plosive:

ḫáz / ḫázon / aḫsí / aḫzossí / aḫtí / aḫzontí / aḫtámo / aḫzonttámo / aḫriyá

Hope this was somewhat interesting. The phonological diachronics are still in the works so stuff will change and hopefully become more interesting but this is probably a good idea of how things will look.

Although this isn't the focus of my project, here's some idea of how the sound changes work and how the old form of Ábḫanni looks (unlike Ábḫanni, the proto-language is written with IPA):

ǝgǝʔsǝnsí > aḫzossí
míroriá > mereniyá

I would like at some point in the future to derive a whole language family from the proto-Ábḫanni. My idea is that Ábḫanni is a sort of "isolate" within the family, cut off at a very early from the rest of the family, and very innovative compared to the rest. But that's super far ahead and at this point in time I'm just happy with how Ábḫanni is progressing and excited to do more development.

Beautiful aesthetic do post more

DesEsseintes wrote: ↑28 Jan 2020 12:25 Beautiful aesthetic do post more

Thanks Des, I appreciate the compliment We'll see how well I can find time outside of work and classes but I really hope to post more soon!

It's been quite a while since I've posted but I've still been regularly thinking about and working on Ábḫanni when I have time. So in this post I'm going to talk about the revisions to the verbal system I've made.

——————

There are now six (main) aspects: perfective, progressive, imperfective, habitual, prospective, and retrospective. The first four sort of function similarly in that they modify the internal temporal structure of the event, sort of diffusing the event from "most eventive" to "least eventive". (Any suggestions on what to call these four aspects, as a group?) Probably, the retrospective will be able to be combined with the first four aspects as well, but I'm not sure the extent to which I want to do that.

Let's look at an example. Here's the example I gave above, ḫáza 'to fall', with the aspects in the order described above:

ḫáza, aḫzasí, aḫzatí, aḫzattámo, ḫázan, ahariyá

(The habitual is a more modern innovation, deriving from -tí + támo, or something like that.) Here's some diachronics on why these are like that. The word derives from *fǝ́sa. For e.g. the progressive, you have something like:

*fǝ́saʃí > *fǝsaʃí > *(ǝ)fsaʃí > *afsaʃí > axsaʃí (not being very precise with the phonemes here)

The retrospective is more interesting. The proto-language had a future tense -n (which I guess became the prospective) and past tense -l. When the retrospective was formed, it was built off the past tense. But the tense endings of the protolanguage had a tendency to metathesize whenever they preceded another affix, so the retrospective evolved as followed:

*fǝ́slariá > *fǝ́zzariá > *fǝzzariá > *(ǝ)zzariá > *azzariá > *aɦarijá

(although the -riá ending probably did not exist until around the second or third stage.) I'm debating on whether to use these past tense stems (like aha-) in other places, or the bare past tense (which, due to the lack of metathesization, would be ḫázao) but I'm not sure yet. The old future tense underwent the same metathesization, so what was the future progressive would have evolved to aḫmasí (< *fǝ́snaʃí) but I'm not sure if I want to use these either. I'm wary of having to many competing stems for each verb; with the retrospective I think it's fine since it will always be marked by the -riyá, and I think the "strong"/"weak" ḫáza/aḫza- is nice, but further than that I'm not sure...

——————

There are also two new irrealis moods, the volitive and the potential. I'll just describe the diachronics at this point as the semantics are still something I'm working out (although you can probably make an educated guess at it based on the names I've given them). In these moods we see the first glimpse of a morphophonological process that I've been enjoying a lot recently: stress-advancement. This is simply moving the stress forward to the final syllable of the stem, or replacing a falling tone with a rising tone. This is used also with nouns/stative verbs, but with active verbs, stress-advancement is characteristic of both of the irrealis moods.

The volitive is quite old, and probably the earliest form of stress-advancement in Ábḫanni's evolution. Hence, it surfaces rather irregularly. The example above in the volitive is as follows:

aḫzó, aḫzósi, aḫzói, aḫzóttamo, aḫzón, ahóriya

First think to note is that, with both of the irrealis moods, the imperfective is marked with -i. (I thought I had a good diachronic reason for this, but actually I don't. But I like it enough that I think I'll keep it around as some irregularity.) The age of the volitive manifests itself in the a - ó ablaut. We do not see this in the newer potential mood:

aḫzá, aḫzási, aḫzái, aḫzáttamo, aḫzán, aháriya

However, probably some verbs would end up with identical volitive and potential stems, so there might for some verbs be a potential-forming affix, but I haven't figured that out yet.

——————

Other things which will be developed more in the future are other (perhaps predominantly "spatial"?) aspects like the frequentative mpa- [ᵐba], for example in mpaḫáza 'to fall all over, to (fall and) spill/shatter/scatter'. There will also be a simple evidentiality-like system, e.g. zoiḫáza, 'i saw (it) fall', although this won't be a mandatory affix and will have uses other than just evidentiality (I really want to do something with mirativity...)

——————

Although I meant this post to be mainly about active verbs, I'll also briefly describe a little bit about nouns and stative verbs. I can't help but also mention the noun morphology: there are four cases (absolutive, ergative, dative, and partitive) and a possessive clitic/affix thing, and generally the declension is fairly regular. Unlike verbs, there's no stress shifting in the declension, and the main irregularities are irregular plural forms (e.g. ákon - ákini 'tree - trees', háḫ - hájhi 'person - people', as well as in collectives, e.g. sáoḫo - sáonni 'tribe - tribes' (the latter really is the collective of háḫ, but the diachronics to get there obscure that a little)) and an alternation between -ro/-ma and -ar/-am allomorphs of ERG/DAT.

But actually the main thing I wanted to talk about is stress-advancement in nouns/active verbs. Stress-advancement is sometimes used for forming stative verbs from nouns, e.g. áken -> akén 'wood -> (to be) wooden' (áken is just the inanimate counterpart to ákon; see the following paragraph). An example where this doesn't happen is bór -> bór 'hair -> (to be) hairy'. This process is relatively new, so that there usually isn't much morphophonology other than shifting the stress.

One example, though, which goes against this is rósmi -> hesmí 'fire -> (to be) on fire, to burn'. First, we have #r -> #h when the following vowel is unstressed (although, I believe, also not secondarily stressed; a similar rule is #n -> #m which would operate under the same conditions; both of these rules are productive, so we get things like zaresmí ákon, 'at the burning tree'). The second process is that stative verbs are genderless. In this case, rósmi derives from *rítmi, which would yield *résmi except that the animacy necessitates that all e become o. Historically, this comes from an animate noun classifier, something like *ó, which was eventually lost, but not before the mid-vowel harmonization occured; the corresponding inanimate classifier would likely be *é). But stative verbs never took these classifiers, so the stative verb did evolve to become résmi before the stress advancement occured.

Another thing about stress, although this generally does not occur through stress-advancement(?) is that earlier stress tends to correspond to more physical/literal things where as later stress tends to correspond to more abstract/metaphorical notions. For example, we have soró 'name' (this does not come from any sóro, however). The details on this are still fuzzy.

——————

Okay, that's it for now. Hopefully that was comprehensible despite how rambly it was. As usual, suggestions/feedback/ideas welcome!

This isn't really a big post, but I had a go at developing the Ábḫanni pronouns just now and I think it turned out rather nice. Ábḫanni just has first, second, and third persons, but the third person (and hence all the pronouns) is used for human referents, so determiners are used as non-human third person pronouns. I haven't figured these out yet though. The first and second person pronouns have the only true Ábḫanni duals (although some nouns have collectives which function as duals).

In the tables, I've indicated that the alignment for the pronouns is NOM/ACC, compared to the ERG/ABS I've been using for nouns. But, actually, I think this is what I will do: with pronouns, the cases are really AGT/PAT and Ábḫanni has a fluid-S-type system. Contrastively, with normal nouns, the system would just be a normal ERG/ABS system. However, I'll have to think about this more...

Finally, each pronoun has a "stressed" and "unstressed" version. Probably, the stressed pronouns would be used for contrastive focus sort of thing, but I haven't given it a whole lot of though. In addition, the unstressed pronouns also have enclitic versions for attaching to active verbs and possessed nouns which will probably have multiple surface forms, but here I'll just show the full-word forms.

Anyways, here are the first person pronouns. The first person has a simple inclusive/exclusive distinction in addition to the above. Note the syncretism in the NOM.sg/NOM.EXCL.dl/ACC.EXCL.dl and NOM.INCL.pl/ACC.INCL.pl; this is one of the rare (only?) places syncretism pops up in Ábḫanni:



The stressed NOM.EXCL.pl form is fun: it derives from *yóga, and the unstressed form from *yog. Similar things give rise to the other wierd-looking exclusive forms.

Here are the second person pronouns:



Note the occurrence of the rare d phoneme, as well as the amusing i-ó / í-a / i-á / í-o stress/vowel pattern in the dual and plural forms. This is due to a vowel shift, which only occurred on stressed vowels, which swapped some ó (specifically, ones which were originally *ɔ́) and á in certain environments. In fact, the same pattern can be seen in the plural of the third person human pronouns below.

And here are the third person human (even though it says "animate") pronouns:

Ope it's... verb time again? Verb morphology time again???

The strong stem

Aight, here's our proto-Ábḫanni root: *gǝ́sa 'to fall' (changed from the old *fǝ́sa). Evolving this root as is, we get ḫáza, which is the "strong" stem for the modern Ábḫanni word. The strong stem comes into play in one other aspect-mood(-voice) combination, the indicative prospective, where *gǝ́sa-n gives simply ḫázan.

The weak stem

However, many of the endings were stressed in proto-Ábḫanni, which messed with things after the tone system turned into more of a pitch-accent system. For example, the progressive evolved as follows:

*gǝ́sa-sí → *gǝsaʃí → ksaʃí ⟨kzasí⟩

(again, not being very precise with IPA here). This unstressed stem kza-, the "weak" stem, is used for the indicative progressive, imperfective kzatí, and habitual kzattámo.¹

The retrospective stem

There is one other stem used in the indicative, and is only used for the retrospective (hence, the 'retrospective' stem). The evolution of this stem I gave in a previous post: the old retrospective riɛ́ was formed with the old past tense -l, which had a tendency to metathesize.² So, the indicative imperfective evolved as follows:

*gǝ́sla-riɛ́ → aɦarijá ⟨ahariyá⟩,

where the particular steps are basically identical to what I gave previously. This aha- is the retrospective stem of ḫáza. Although, I'm still debating whether to allow -riyási, -riyáti, and -riyáttamo endings, which would make this stem a bit more useful.

The irrealis strong stem

I've reworked the two irrealis moods, and they now involve a feature of proto-Ábḫanni which I don't think I've mentioned before: ablaut. Proto-Ábḫanni had actually two forms of ablaut, i-ablaut and u-ablaut, but only i-ablaut is important for verbs. The older irrealis mood, the volitive, was formed by the past tense suffix plus the i-ablaut-triggering suffix *-ne. This latter suffix was sort of a nominalizer, like "the act of X-ing (in general, not a particular one)". For example, "I want him to eat" would be translated, in proto-Ábḫanni, to something like "I want him having eaten",³ and later the "him having eaten" was reinterpreted as some sort of irrealis construction which eventually came to be able to stand on its own.

Aaaaanyways, the bare volitive stem for ḫáza evolved as follows (note the *-ɛ ending of the proto form—this is the i-ablaut in action):

*gǝ́slɛ-ne → *gǝ́zzenʲa → xáɦeɲa ⟨ḫáheja⟩

This ḫáhe- is the irrealis strong stem of ḫáza. Strong because it is stressed, and irrealis, well, you can figure that one out. As with the indicative strong stem, the irrealis strong stem is also used with the retrospective, in this case ḫáhejan.

The irrealis weak stem

Just as with the indicative strong stem, we get a new stem when stress shifts away from the irrealis strong stem. In our case, the irrealis weak stem of ḫáza is ahe- (this evolves in basically the same way as the retrospective stem, except plus the ablaut).⁴ Hence we have things like -VOL-PROG ahejasí, the irregular -VOL-IPFV ahejí, -VOL-HAB ahejattáme, and (no need for a new stem) -VOL-RET ahejariyá.

I should briefly mention the potential mood as well, which is as I've mentioned a newer innovation, and is now formed (almost regularly) from the volitive by stressing the irrealis marker -ja. The one irregularity is -POT-IPFV ahejái.

The antipassive stem

Whew, one more stem to go! The antipassive was formed using an other nominalization suffix *-t, which also triggered i-ablaut. This stem was more like "a specific act of X-ing". The antipassive is also formed with -ab-, which derives from *ɛ́pal, 'to make a noise'. So, the old construction for an antipassive "He shot the bear" would have been something like "Him shooting made a noise at the bear". I have some ideas for the uses of the antipassive but that's for another time.

So, the old antipassive stem of ḫáza would have been *gǝsɛ- (unstressed, no past tense ending, i-ablaut) giving kze-. Fortunately, this one stem works for all aspect/mood combinations in the antipassive; e.g. -ANTIP-PROG kzetabsí, -ANTIP-RET kzetabriyá, -ANTIP-VOL kzetábja, -ANTIP-VOL-PROG kzetabjasí, etc.⁵

That seems pretty complicated

Yeah, it does. Three different morphophonological alterations interacting to give six different stems for each verb? Am I too obsessed with nonconcatenative morphology??? Here's the full aspect-mood-voice paradigm for ḫáza:



But fortunately many verbs are actually rather regular (but at the same time, I'm not sure any of them I would actually call "regular"). For example, let's look at támo 'to move (oneself)' (see ¹). The six roots are as follows:

Strong: támo
Weak: tamo
Retrospective: tamro
Irrealis strong: támra
Irrealis weak: tamra
Antipassive: tama

This one is about as uninteresting as it gets; stress shift does nothing but shift the stress, it has a normal o-a ablaut of the final vowel, and the stems deriving from the old *-l- just have an *-r- infix. Here's another one, ȃza 'to bring/take (something down)':⁶

Strong: ȃza
Weak: ạza
Retrospective: ạha⁷
Irrealis strong: ȃhe
Irrealis weak: ạhe
Antipassive: ạze

The irregularity here is the z-h alteration from the past tense suffix.

When strong stems end in a consonant, the past-tense-suffixed stems generally get a sort of suffix, e.g.:

Weak: abah, Retrospective: abanna 'to not make sense (of a person speaking), to say stupid/false things' ← *ɛ́pal
Weak: amek, Retrospective: amesra 'to bite' ← *agník
Weak: eppeḫ, Retrospective: eppeḫa 'to fart' ← *ippíf

Stems with irregular strong/weak forms (like ḫáza) seem actually to be relatively uncommon, and the irregularities generally surface as deleted vowels. Here's another example, kámata 'to have sex' ← *kómata:

Strong: kámata
Weak: ḫmata
Retrospective: ḫmasra
Irrealis strong: kámasre
Irrealis weak: ḫmasre
Antipassive: ḫmate

Okay, that's it for now

Hopefully I can move on from verbal morphology now. I do like the system I've come up with and I think it will stick around mostly intact, but of course we'll see. I've also reworked the pronouns too... it seems that the majority of the posts in this thread are becoming "Ignore my previous posts, here's how things REALLY work". But, maybe I'll save pronouns for some other time and talk about something new in the next post.

---

¹ As I've mentioned, the habitual is a newer innovation; what I did not mention is that it in fact derives from *tɛ́mɔ, 'to go'.

² Specifically, *-V́(C)V-l-x → *-V́(C)lV-x, where *-x is any nonempty suffix; the future ending -n was subject to the same metathesization, but this ended up not affecting modern Ábḫanni at all.

³ Here "him" is actually a possessive, so maybe I should have written "his", but proto-Ábḫanni didn't make this distinction, at least in this case.

⁴ Actually, that is a lie: the irrealis weak stem of ḫáza would be expected from sound change to come out as *ihe-, but this was probably leveled to ahe- pretty quickly.

⁵ The two irregularities in the antipassive endings are the bare -ANTIP kzetábah and the -ANTIP-PRSP kzetábran.

⁶ From *ɛ́ʔ- + *gǝ́sa; the *ɛ́ʔ- was a causative of proto-Ábḫanni and is no longer productive. Also, I've been using these new diacritics to mark tone + glottal stop combinations, e.g. ⟨ȃ⟩ /áʔ/ and ⟨ạ⟩ /aʔ/. I really like these...

⁷ Here is another example of leveling: the glottal stops here, and in the other stems deriving from *-l-, are not diachronically expected.

Instead of trying to write out grammatical stuff which will inevitably change, I think I'll just have fun working out some etymologies. This post will be the etymology of just one word, zéja; future posts will have etymologies for more words derived from the proto-language root considered here.


Etymology

---

The story begins with Proto-Abhannic (PA) tʰu, a stative verbal root meaning 'to see'.

The PA passive present nonfinite form is tʰuʔŋi, something like 'to be seen', which becomes tʰʊ́ɲ in Proto-Core Abhannic (PCA). By Old Ábḫanni, the passive has been generalized into a sort of anti-reflexive, hence this becomes the present nonfinite θéɲ, literally something like 'to see' or 'seeing'. However, in OA, stative verbs are starting to become obsolete and archaic, being used to derive active verbs, nouns, and adjectives. In particular, the OA nonfinite forms often become abstract nouns; in this case, θéɲ comes to mean 'eyesight' or 'vision'.

This word remains θéɲ in Middle Ábḫanni (MA), but begins to be used to refer to eyes in general (specifically, as tools for seeing rather than physical objects). Also, by the time of MA, the noun case system has solidified, with previously separate prepositions fusing to the noun as suffixes after a major stress/tone shift. In particular, we have

Absolutive: θéɲ
Ergative: θéɲalæ
Dative: θéɲamæ
Locative: θéɲitat


Definition

---

zéja [ˈθ̱é.ɲ̟ɐ] deictic inanimate noun. eyes (as tools for seeing rather than physical objects), eyesight, vision (i.e., the ability to see); attention


Declension

---

Characteristic of modern Ábḫanni, the plural has been generalized for a distributive-like use:

Absolutive: zéja / zéji
Ergative: zéjar / zéjimata
Dative: zéjam / zéjima
Locative: zíjtạ /zéjitạ


Some semantic details

---

It has been emphasized that zéja refers to the eyes as "tools for seeing rather than physical objects". This is in opposition to the words taá 'eye'/'eyeball' and tanttá 'pair of eyes' (probably the former being somewhat archaic, anatomical, or diverged in meaning, with the latter taking its place), referring to the actual physical objects.

In most cases, tanttá remains the word used to refer to eyes. For example, sentences like 'they have blue eyes', 'he hit my eye', etc., this word would be used. One would only use zéja when specifically talking about vision. Some of these uses are better translated with 'eyesight' or 'vision', such as in 'my vision is bad'. Other uses could be (using semi-literal translations):

"my vision is dark" = 'i can't see'
"they turned their eyes" = 'they looked away' (maybe "they turned their eyes (tanttá)" = 'they looked around')

By extension, zéja can refer to 'attention' as well (when it is "toward" or "away from" something).

Also, the locative form is the most neutral way to translate "see", and in this case there is no possible interpretation of 'attention'. For example:

í zíjtạ né kó
[í ˈθ̱íɲ̟.t̪ɐʔ n̪é ˈkɔ́]
PRS vision-LOC PROX 2sg
'i see you', lit. 'you are in my vision'


"Deictic"

---

What is meant by deictic in the definition is that zéja is an inherently possessed noun. In Ábḫanni, there are two types of inherently possessed nouns, deictic and personal. The former consist mainly of body parts, mental qualities, and things related to the body, such as vision or breath; the latter consists mainly of family members, personal characterics/possessions such as name, birthplace, birthday, age, home, and more abstract things like burdens or duties.

Anyways, deictic inherently possessed nouns are so-called since they are marked by an obligatory determiner, which also determines the possession. So, you have zéja né 'my eyes', lit. 'the eyes here'; zéja bá 'your eyes', lit. 'the eyes there (by you)'; and zéja bám 'their eyes', lit. 'the eyes (over) there'. In particular, the full range of spatial deixis isn't taken advantage of; zéja zé (the "emphatic" deictic), zéja mám (intermediate to bá and bám), and zéja bekkái (nonvisible) are ungrammatical. (Ábḫanni seems to be marked by being very anti-drop in a generalized sense: these deictic things, it doesn't like dropping pronouns, the obligatory copula-like words like í, probably something with resumptive-like pronouns in relative clauses...)

---

Hopefully this post was interesting! I'm hoping it can be a productive way for me to make posts about Ábḫanni while it's continually undergoing significant grammatical and diachronic changes.