Zompist Bboard Again

So after ages and ages I've finally got back round to that Algo-Lang project that's been driving me mad for ages, and I think I've hit upon something that I'm fairly satisfied with, so I'm gonna present some of the information I currently have on the language in this thread. I'm gonna give a general overview of the language and how I'm currently thinking about its extra-linguistic context at this point in time, so here goes.

The Aquecutta people (natively called the Nuwa‘hracku) are an Algonquian-speaking people located in the lower part of the Yellowstone River (in OTL's Eastern Montana), comprising about 12% of the population of the modern nation of Dakota. Notably, though falling under Bloomfield's "Plains Algonquian" branch, unlike their closest relatives the Arapaho and Cheyenne, the Aquecutta are more closely tied culturally to both the northern Caddoan peoples (in particular the Arikara and Pawnee) and the Crow-Hidatsa, which shows some reflection in the structure of the language. This seemingly new-found allegiance was found in the Indian Civil War, where the Aquecutta joined the Arikara and Crow in challenging the Sioux's political dominance in the Mid-Plains Region.

In view of this historiy the Aquecutta language is a typical Algonquian language in many respects. Nouns are marked for number and in animate nouns obviation as well, and there is a paradigm of person inflection for possessed nouns, while verbs are polysynthetic, with polypersonal marking as well as animacy-based stem alternations.

Phonologically, as indicated above, the language bears strong resemblences to Arapaho and Cheyenne, leading some scholars to argue that thogether they form a single subbranch of Algonquian. Notable shared isoglosses include:

• PA *we and *oː merging with PA *i and *iː respectively (shared with both Ar. and Ch.)
• PA *w merging with *y after consonants (conditioning shared with Ch.)
• Raising and rounding of *a(ː) (shared with both)
• Lowering of *e(ː) (shared with Ch.)
• Widespread loss of *k (shared with both)
• Frequent loss of *p (shared with Ch.)
• Occasional retention of both *p and *k with unclear condtioning (shared with Ch.)
• PA *s > h (shared with both)
• *š backs to x, which then undergoes palatalisation rules back to becoming a sibilant (shared with both)
• *m splits into w and a labial stop (shared with Ar. with different conditioning)
• Loss of *h in clusters (shared with both, though see below for additional details)
• Falling together of *n and *ʔ in clusters to ʔ (shared with Ar.)
• Merger of *š and *θ in clusters (shared with Ar.)
• Collapse of all of *s, *r, *š and *θ to ʔ when the first member of a cluster (shared with Ch.)
• Original post-vocalic *y > t (shared with Ch.)
• Merger of unclustered *θ with *t (shared with Ch.)
• Merger of *r and *n (shared with Ar.)
• Yodation of *k before *e(ː) (shared with Ch.)
• Loss of final vowels plus a preceding glide (shared with both)
• Disyllables are prevented from undergoing said apocope (shared with Ch.)
• Syncope of short vowels when not before a cluster (shared with Ar., though see below)

However, all that considered, there are some notable differences in other isoglosses which prevent Aquecutta from being assigned to either of these two branches.

• Initial and post-vocalic *w > k (vis-à-vis Ar. n and Ch. v)
• Merged postconsonantal *y becomes a sibilant s, merging with palatalised x (compare Ar. y and Ch. n)
• Retention of *č as c /ts/ (compare Ar. 3 /θ/ and Ch. s)
• Denasalisation of merged *n to r when not word-initial or geminated via syncope.
• Denasalisation of *m to p word-initially and after another consonant
• Long vowels are shortened before all original clusters (vs. Ar. all clusters bar *hC)
• *Hm > p (vs. Ar. b~w and Ch. m)
• Syncope of short vowels only occurs before originally unclustered true consonants (indicating a relatively early occurrence vis-à-vis Ar.)

As a taster here are some example sentences.

kuhti a‘taa‘uu a‘tuksuu
ABS IC-night.II IC-cold.II
It was a cold night

u‘tu‘harhituur aahua‘suutuur akaaru‘
IC-shoes-put.on.AI-1s IC-exit.AI-1s tent
I put on my shoes and left the tent

huucta‘ u‘xaahku kutuuwui
tree-LOC wolf-PL IC-see.TA-1s-PL
I saw some wolves among the trees

piihiih ikpuucihci
cow-OBV PROG-IC-eat.AI-3p
They were eating a buffalo carcass

EDIT: due to changes in the verbal syntax the above examples are out of date, current versions below the cut.

This is giving me nostalgia for our collab

I'm sad I never finished French Gaelic or Pamplonese I had so much fun with them

This is fun--it's Plains Algonquian, to be sure, but aesthetically it looks more Northern Caddoan than Algonquian. What's the name from--an Anglicization of akwikara or some such?

dhok wrote: ↑Tue Oct 08, 2019 6:49 amWhat's the name from--an Anglicization of akwikara or some such?

It's a shameless Anglicisation of a Arikara name for the Yellowstone River aahkawirahkatá [aːhkəwir̥ə̥hkətá] (yes I did delete the voiceless syllable dont @ me).

Whimemsz wrote: ↑Thu Oct 10, 2019 10:30 pm

Frislander wrote: ↑Mon Oct 07, 2019 11:42 amAs a taster here are some example sentences.

kuhti a‘taa‘uu a‘tuksuu
ABS IC-night.II IC-cold.II
It was a cold night

u‘tu‘harhituur aahua‘suutuur akaaru‘
IC-shoes-put.on.AI-1s IC-exit.AI-1s tent
I put on my shoes and left the tent

huucta‘ u‘xaahku kutuuwui
tree-LOC wolf-PL IC-see.TA-1s-PL
I saw some wolves among the trees

piihiih ikpuucihci
cow-OBV PROG-IC-eat.AI-3p
They were eating a buffalo carcass

So I'm curious which of these morphemes are inherited from PA. I recognize a few mainly because I already know what they look like in Arapaho or Cheyenne (plus just applying the relevant sound changes to *-a:n- in my head) but most I don't, without going through the trouble of actually, like, spending any energy on looking anything up or applying the sound changes you've given. AS IF!

They're all inherited, but the origins might be a little hard to spot. The <t> in the 1s is from the *y, but there's also a lot of cases of stem-final glides that show up as consonants in inflections (so the (proximate) singulars of "tree", "wolf" and "cow" are huuci, u‘xaahi and piihsu). Similarly, the progressive ik- is essentially a much-reduced form of *apiwa (which partly answers the question below, since the progressive has essentially evolved from an auxilliary construction which took the conjunct IC participle).

Anyway I'm also curious what function initial change serves in Aquecutta? From these few examples it looks like Aquecutta went the Arapaho route and changed conjunct verbs are now the neutral main clause verbs? (And if that's the case, what other verbal orders and modes are there, and what are they used for?)

Year you're pretty much correct in that it's gone down the Arapaho route, though it's more of a combination, in that while the function is very Arapaho-like, the actual morphological exponence is much more like Blackfoot (e.g. the non-IC roots of the verbs in these examples are taa‘-, tuks-, i‘tu‘harh-, ua‘s-, kuuw- and piicih-). In terms of the other modes, the iterative is similarly intact, and the weird "combined preterite" *-sapan is now found with conjunct inflection (minus initial change) as a reportative evidential.

The old indicative, on the other hand, has essentially turned into an irrealis, though unlike Arapaho this is independent of negation (which, just for interest, is also Algonquian-derived - it takes the form of a prefix (-)hii-, which is derived from the same form as the Ojibwe negative, which has done a jump to prefix position thanks to forms like the progressive, for example ikhiipuucihci "they are not eating [it]").

So let's talk phonology (I don't think I'm gonna have a post on phnological history, mainly because it's a a bit vague and in my head, and also because I've covered most of it in the original post, but I'll cover a bit more in this post).

In terms of consonants, like many mid-Plains languages it's not got that many, in this case only 10.

/p t t͡s k ʔ/ <p t c k ‘>
/s x h/ <s x h>
/w ɾ~n/ <w r~n>

Of these some contrasts are much more robust than others. In particular /w/ frequently alternates with /p/ in word-initial and post-consonantal position, e.g.

pa‘ta "bow" > nawa‘tiiw "my bow"
nikiw "my (maternal) grandmother" > nikpuru "our (maternal) grandmothers"

However there are some (near-)minimal pairs in intervocalic position that necessitate a distinction between the two, e.g. uupi "mouse" vs. uuwaci "iron, metal" (the latter a Crow-Hidatsa loanword). The reason for this conditional absence of contrast is simply that it represents a split in PA *m after the loss of *p from most positions (though "mouse" is an irregular example of the retention of *p).

In a somewhat similar vein is the r~n alternation. in most contexts r is found, however when it would appear word-initially it is instead nasalised, and similarly in the rare instances of gemination of the same consonant (e.g. nuwa‘h~naruwa‘haw ("fish~my fish"), aahannuwuur "I put it down"). In all other instances r is found.

Somewhat stronger are the t/c (e.g. a‘tuahi "it is cold" vs. a‘cuahi "it burns") and s/x (e.g. suut "enemy" vs. xuu "skunk") contrasts, though with some marked distributional skews. Firstly c is significantly more common that t before the high front vowels and diphthongs, and x is absent entirely from the same position. Secondly, due to its dual origin from post-consonant glides and *š, s is exceedingly common as the second or third member of a cluster, whereas x is much less common as the second and vanishingly rare as the third. Similarly, t universally shifts to c whenever it would be a member of cluster other than the final one, e.g. utaw~ucpii "dog (sg.~pl.)". On the flipside, x is found to the exclusion of s word-finally (though no such differentiation is found for t vs. c, since c as an affricate was already found is PA whereas the s/x contrast is a later innovation).

Additionally ‘ is fairly restricted in its ocurrence. Word-initially it is rather weak in its articulation to the point of being rendered as creaky voice on an otherwise word-initial vowel, and otherwise it appears to function as a prop for a vowel-initial word, since it is lost whenever a prefix is added (though unlike in other Algonquian languages the epenthetic -t- of the possessive series has been reanalysed as part of the possessive prefix due to the deletion of word-initial short vowels resulting in the loss of *we- entirely from the language). Similarly in clusters it it not found as the second or third member of a cluster (the reason being that it originated from original clusters which blocked the vowel syncope which produced most of the clusters of the modern language). On the flipside, it is the only consonant that can occur at the start of a three-member cluster, e.g. i‘psuur "ceremonial pipe".

In terms of vowels there appears to be a fairly simple system:

/i iː u uː/
/a aː/
/ia ua iu ui/

However, there are some notable phonetic complexities which complicate this picture. Notably short vowels are somewhat centralised [ɪ~ɪ̈ ɐ ʊ], while the diphthongs depart radically from their phonological patterning [ɪə ʊə jy øj]. As is common in languages of the region, some speakers show devoicing of short vowels word-finally and before h, e.g. aahuicaahithuuri [aːhøjt͡saːhɪthuːɾɪ̥]. This devoicing is blocked by the accent (see below).

In terms of phonotactics, as hinted above there is some freedom for the formation of clusters, though mostly this is in intervocalic position. In word-initial position clusters are exceedingly uncommon [meta note - the only item in the language so far that contains such a cluster is psaati "dung"]. In intervocalic position, on the other hand, there is near complete freedom (subject to the constraints noted above), though by far the most common first member of a cluster is ‘, since most PA clusters (i.e. all of them other than hC) have reflexes with this as the first member, e.g. xu‘xui "heron"; sa‘ri "vulture"; naksu "my mother", uucpikar "story", aspit "tree frog" (a Caddoan loan), utuukhit "he makes a low noise", pataakkar "dance". Similarly ‘ is also the only available first member of a three-term cluster, e.g. u‘ptuhir "cane", na‘xpah "my cross-niece", i‘ksu "nightjar". Word-finally clusters are restricted to original PA clusters with ‘ as the first member, e.g. iisi‘x "sun, moon".

In terms of accentuation, there is a system of pitch-accent with a rather low functional load (hence why the native writing system can get away with not representative). The system is probably best explained relative to historical factors so here goes.

Each word typically has one accent. In words with only original short vowels with no deleted consonants, the accent is found on the first vowel, e.g. á‘ix "fly". Original long vowels gained a rising accent, which overrides the initial accent and any preceding rising accents (barring analogical developments), e.g. taa‘iísi‘x "moon", but see útaw~úcpii "dog". This accent placement remains even after the shortening of a long vowel before a cluster, e.g. kiisú‘ "birch", paxú‘ "(piece of) firewood", saksá "sparrowhawk" (the latter forming a near-minimal pair with sáksu "you mother"). When a consonant has been deleted leading to vowel syncope, when the second vowel is short the accent of the resulting long vowel/diphthong is falling, while if it is long it is rising, and in either case this accent overrides any other accent from an original single vowel and any preceding similar accent, e.g. cíu "ghost", úi "crow" vs. xuú "skunk", nuá "man". Notably as well when a long vowel in a loanword would be eligible to receive the accent it also receives a fall, e.g. akáaru‘ "house" (from Caddoan), páasu "plume, quill" (from Siouan).

So, now we've talked the mess of the phonology, we'll move onto the mess of the nominal morphology! For simplicity's sake we won't tackle the possessive morphology yet, and we'll just focus on animate nouns (though admittedly inanimates don't differ much in this regard). Once again there will be reference to PA morphology for explication.

As in Proto-Algonquian, animate nouns distinguish the categories of both numbers and obviation, plus having a locative form. Like many languages in the family, the number contrast is levelled in the obviative, and like in Cheyenne it is the obviative plural form that has won out, helping distinguish the animate obviative and the inanimate plural. In a similar manner, the locative form does not distinguish number.

In terms of actual exponence there are numerous subclasses. The simplest group of nouns derive from bisyllabic roots which lacked a stem-final glide, which notably lack a separate plural form.
A slightly larger class consists of those bisyllabic roots which did have a post-consonantal glide. These nouns show the contraction of *wa to *oː between consonants, which as per Arapaho has been spread to all the contexts it could appear regardless of etymology. In the case of the bisyllabic nouns the original post-consonantal glide shows up in the proximate singular.
Things get more complicated when one leaves the realm of historically bisyllabic roots. The simplest classes show either an -u or and -ii in inflections as per the presence vs. absence of glides.
A significant number, perhaps indeed a majority, of noun however show epenthetic consonants in inflections. This is always one of t, k or s, which descend from an original *y, *w and postconsonantal *y/w respectively.
Another factor that can complicate inflection is the presence of contraction, where a short vowel before a single consonant deletes in inflected forms. These contractions can then further interact with other sounds changes to alter the resulting cluster.
This frequently combines with consonant epenthesis like so.
Finally there is one functionally irregular noun, xuu "skunk", which has the following inflection.

I like.

So just a quick discussion of the inanimate classes. They are pretty much the same as the animate ones, with some minor differences, as well as the more significant difference of the lack of obviation as a category and the fact that the plural has a coda -r.

Most of the different declension patterns found in animates are also found in inanimates, including inflections both in -u and -ii, as well as constracting stems and epenthetic consonants, though the frequencies are differently distributed, with -u forms being substantially more common relative to -ii forms for non-epenthesising stems (the opposite being true of animates) and t and k being more common than s as epenthetics.
There are of course minor differences in a couple of the declension classes. For one, a few inanimate nouns show vowel deletion before -ii endings, a class which appears to have been levelled out of animates if it ever existed at all.
Secondly, the class of original bisyllabic roots has a different vowel in the singular, reflecting the PA *-i of the inanimates rather than the *-u of the animates.
(if you're wondering why mass nouns such as "grease" and "firewood" have plural forms, they're treated as if they were accompanied by an appropriate count noun, e.g. "drops of grease" or "pieces of firewood".)

EDIT: Oh yeah, and one final thing: all cases of t~c alternation that would have arisen from *-t- stem nouns in the singular have been levelled, mainly in favour of the singular, e.g.

So, possession. As in other Algonquian languages there are two varieties - alienable and inalienable. Alienable nouns have reflexes of the possessed theme marker *-em- while inalienable nouns don't. The person of the possessor is marked by prefixes, while number of the possessor is marked by suffixes. While the alienable nouns are entirely regular in their prefix sets, many inalienable nouns exhibit irregularities in their prefixes, as well as boasting an "indefinite" form. Additionally, alienable animate nouns must be mark as obviative when possessed by a third person. Finally, it is interesting to note that the formerly epenthetic *-t- of vowel-initial stems has been reanalysed as a "possessed prefix" which appears in the third person (sans the original *we- prefix which was lost be regular sound law) as well as the first and second person.

So for regularity's sake let's start with the alienable paradigms. As indicated above, these follow the pattern of person-root-aw-number, with the realisation of the -aw component following the class of the noun as per the above posts, in particular mirroring the locative form. The examples given below are i‘psuur~i‘psuuru "pipe" and pu‘iih~pu‘iihii "horse" and huuci~huuctu "tree" Note that the 3rd person forms with the obviative also indicate the rexpective stem classes of the singular and plural possessor forms with regards to plurality/obviation of the possessed noun (namely, singular possessors take simple -u inflections while plural ones epenthesise -k-).

And similarly with the inanimates, here exemplified by uucpikar~uucpikrur "story" and paxu‘~paxu‘iir "(piece of) firewood".
Inalianable nouns are much the same, with the following caveats. Firstly the absence of the -aw formative results in the plural possessor suffixes being affcted by the stem class of the root in the same manner. Secondly, there is an indefinite form p(a)- which is added to represent an unspecified possessor - in animates this does not require obviative inflection like the normal 3rd-person marker. Finally there are notable differences in the prefix forms. In particular, when a zero-3rd-person prefix would result in a word-initial cluster or a monosyllabic form, an i- is found instead as a peg vowel, and when the root begins with a vowel (in contradistinction to alienable nouns) the following alterations are found - the 2nd person is marked by a zero instead of an s- and the 3rd person is marked by a k- rather than a zero. These facts will be exemplified by nasi‘h~nasi‘hu "my son", nikiw~nikpu "my maternal grandmother" and na‘kac~na‘kaciir "my buttocks/anus"
In particular one should note the paradigms of "mother" and "father", which show some additional irregularities.

OK, time for some pronouns.

Personal pronouns are continued somewhat intact, however there has been some notable reshaping, in particular of the plural forms. Clearly inspired by the Caddoan model, there has been some move to move the pronominal paradigm closer to the verbal/possessive paradigm, along the lines of pronouns being represented by an inflected copula, though retaining some historical irregularities in the prefix forms. Thus instead of the expected 1st person plural forms *niiruur/*iiruur we find the forms niinnu/iinnu, which are much more in line with the corresponding verbal/possessive conjugations.
Demonstrative pronouns (which occur both as free pronominal forms and combined with nominals) mark for three distance distinctions - proximal, distal and absent. There are also separate animate and inanimate forms.
In terms of inflection, the proximal and distal forms conjugate as contracting u-stems (e.g. ikpu "these (animate)"), the animate absentive conjugates as an animate bisyllabic u-stem (e.g. kuhtuh "that/those (obv.)") while the inanimate absentative conjugates like a bisyllabic inanimate (e.g. kuhtur "those (inanimate)").

Finally there are three indefinite pronouns - kaaru for animates, kaasi for inanimates and tusi for locations. The animate and inanimate forms conjugate like the corresponding absentative forms.

A short word before we continue. There are also a handful of corresponding clitics for some of these forms - two deictic clitics ti= and ni= (proximal and distal respectively) and three indefinite/interrogative clitics kaar=, kaas= and tus= (which are obviously related to/derived from their free counterparts). The deictic clitics frequently appear on verbs but also may appear on nouns, and are generally equivalent to demonstrative adverbial forms. The indefinite/interrogative clitics on the other hand are somewhat different. Whereas the free forms typically represent indefinite arguments, the clitic forms typically attach to conjunct verbs to mark WH-questions (reflecting their origin are predicative WH-words), e.g.

kaasi nirakuu‘t
IND.INA DIS=1-see.TI
I see something over there

kaasrikutuu‘cpuur?
INT.INA=DIS=IC-see.TA-1s
What do I see over there?

So I think I should perhaps start with this massive edifice that is the verbal morphology. As is to be expected from a north American language the verbs are a bit of a mess, so I'm gonna tackle them a piece at a time starting I think with a general overview.

The majority of the general points Algonquian verb system have survived fairly intact. Verbs are inflected for both their subject and in transitives their object as well in a hierarchical system of marking. Verbs come in four stem classes based on the animacy of the "absolutive" argument - animate intransitive (AI), inanimate intransitive (II), transitive inanimate (TI) and transitive animate (TA). The actual system of verbal inflection comes in two sets, independent and conjunct, the former of which uses both prefixes and suffixes while the latter concentrates all the marking in suffixes (though there are some historical complications of course as will be discussed below). Conjunct verbs frequently exhibit the phenomenon of initial change, which sees some verbs ablauting their first stem vowel and some prefixing aa(h)-~a‘-.

In terms of general distribution of the two orders, the following points should be made from a diachronic perspective. The historical pattern as reconstructed for PA has the independent occurring in main clauses and the conjunct in subordinating contexts such as participles. This distribution has been altered in Aquecutta, in that conjunct participles plus auxilliaries/predicative interrogatives have been grammaticalised into main clauses, e.g. the progressive, derived from *apiwa "to sit" plus participle. On the flipside, the general levelling of the syntax away from subordination has resulted in a greater frequency of independent verbs standing in parataxis, though notably the old conjunct iterative is retained with some of the same functions.

tisawiiwuu nara‘xuu
PROX=2-bring.TA 1-kill.TA
You bring him here and I'll kill him/If you bring him here I'll kill him

kutuuwtuurir nutaawtaaktuur
IC-see.TA-1s>2s-IT 1.PROG-IC-dance-1s
I'm [always] dancing whenever I see you

The mechanics of initial change have become somewhat frozen since PA. Generally only (historically) long vowels have preserved the original process of initial change, with an -ut- < *-ay- infixing before the non-front vowels and i~ii ablauting to u~uu, e.g. hu‘uut~hutu‘u‘ "rain" (II), piicihi~puucihit "eat" (AI). The exceptions to this come from verbs with an initial *we plus some verbs with initial *e, which mutate to ka(a)- and ta(a)- respectively, e.g. i‘pa‘uu~ka‘pa‘uut "go up" (AI), iku~taat "say" (AI). Note that in a handful of verbs these etymological short vowels were deleted word-initially, resulting in an apparently prefixing pattern, e.g. hiikhi~kaahiikhit "be yellow" (AI). All other verbs exhibit initial change by prefixing the original initial change preverbs aa(h)-, or a‘- before the coronals t and c, e.g. nuhui~aaruhuit "sit (down)" (AI), uuki~aahuukit "stand (up)" (AI), tuahi~a‘tuahit "be cold" (AI). Note that this prefixation sometimes feeds vowel syncope, resulting in alternations tuwiiruw~a‘cpiiru‘ "press, squeeze" (TI). Finally there is the entirely irregular verb "die", which alone retains the short vowel lengthening pattern: nasu~naasu.

Another point to note about initial change in Aquecutta is that it is a property solely of verb roots, unlike in PA where is could apply to preverbs as well when one was present. The reason for this seems mainly to have been due to the above mentioned patterns of grammaticalisation, where the independent auxilliary grammmaticalised as a prefix to a conjunct verb with initial change, resulting in an inconsistency between preverbs which showed initial change and those which left it on the verb root, which was resolved by making all preverbs invariant and the verb show the initial change in all circumstances where it would be present (this is likely partly an effect of contact, since the particular pattern of initial change seen in Algonquian has no parallel in Caddoan or Siouan).