An introduction to Acranasian

Hi! It’s been a long time–very nearly a full two years–since I’ve last posted here. A lot has happened in the interceding time; I’ve been dealing with a lot of mental and physical illness, and I’ve been largely far too stressed to conlang seriously beyond the short composition of texts in South Eresian, but on the other hand, I’ve certainly matured as a conlanger and a linguist as I’ve progressed towards my linguistics bachelor’s degree. I’ve been learning Ojibwe for two years now, and some of the features of that have made their way into my own conlangs. I will, hopefully, be graduating at the end of December, and then I’ll be off to grad school (provided anyone accepts me!). Right now it is summer and I have loads of free time, so in lieu of the assorted things I ought to be doing, I’ve been conlanging pretty furiously over the last few weeks and finally have something to show for my work on Acranasian (South Eresian’s mother language). An overview of that is the topic of this post.

Acranasian was the language of the Acranasian civilization, which was based out of the city of Acránas (the endonym was Ak’ulānassu) and lasted from 1500 years ago until its collapse around 1100 years ago. It is an Eastern Noruic language of the Noro-Eresian family. This post will cover the phonology and grammar of Classical Acranasian in particular, and give an idea of the changes that occurred between Early Acranasian (~1500-1400 ya), Classical Acranasian (1400~1150 ya) and Late Acranasian (1150~1100 ya).



 NASAL  m  n  ŋ  ŋʷ
 STOP  p pʼ  t tʼ  k kʼ  kʷ kʷʼ  q qʼ  qʷ qʷʼ  ʔ
 AFFRICATE  t͡s t͡sʼ
 LAT. AFFRICATE  t͡ɬ t͡ɬʼ
 FRICATIVE  ɸ ɸː  s sː  x xː  xʷ xʷː  χ χː  χʷ χʷː

/ŋ ŋʷ/ would more accurately be described as dorsal than velar, since they assimilate to uvular as well as velar consonants (whereas the other nasals, until the Late period, did not).

Someone might suggest that such a large set of dorsal fricatives would perhaps not be particularly diachronically stable, and they would be correct; the system of fricatives has collapsed by South Eresian, which has the very modest fricative inventory of /s ʃ x/.


/a i u a: i: u: ai au/


/m n ŋ ŋʷ/ <m n ŋ ŋw>
/p pʼ t tʼ k kʼ kʷ kʷʼ q qʼ qʷ qʷʼ ʔ/ <p p’ t t’ k k’ kw kw’ q q’ qw qw’ ‘>
/t͡s t͡sʼ t͡ɬ t͡ɬʼ/ <c c’ ƛ ƛ’>
/ɸ ɸː s sː x xː xʷ xʷː χ χː χʷ χʷː/ <f ff s ss h hh hw hhw x xx xw xxw>
/ɬ ɬː l/ <ł łł l>

/a i u a: i: u: ai au/ <a i u ā ī ū ai au> (plus acutes for stress)


The syllable structure of the language is (C)V(S), where C stands for consonant, V for vowel and S for sonorant (any of the nasals or /l/). Geminates are very much capable of being word-initial, as in ssailā́ŋka /sːaiˈlaːŋka/ “blood.”

There are no sequences of sonorant+fricative.

Content morphemes tend to have one, and no more than one, lexical stress, although they don’t always. Function morphemes frequently occur completely unstressed.


Although Acranasian is an isolating language and as such largely lacks inflectional morphophonology, some of the many derivational processes that occur do cause some morphophonological processes to take place.

Morpheme-final fricatives delete in all positions except prevocalically and cause following obstruents to become fortis (stops become ejective, fricatives become geminate): Ał-sułłā́łu “the star god(s),” from the deity honorific ał- plus sułłā́łu ‘star’ is pronounced [ɐsːʊˈɬːaːɬʊ].

Final sonorants cause (a different kind of) fortition on following fricatives: /ɸ ɸː s sː x xː xʷ xʷː χ χː χʷ χʷː/ become [p p ts ts k k kʷ kʷ q q qʷ qʷ] respectively. The only exception to this is clusters of /l/ plus a following lateral fricative /ɬ ɬː/, which both just become [ɬː]

The unstressed vowel /a/ deletes preceding any long vowels: thus the sequence /ai:/ just becomes [i:].

There is a process in the language wherein grammatical phrases that do not have any lexically-marked stress become stressed on their initial syllable. This will show up at least once in the grammar section below, and is also responsible for a number of disparate South Eresian etymologies: óxta ‘house,’ for instance, is from the Acranasian uhhi ta ‘we dwell there,’ which became initially stressed (úhhi ta) because the phrase has no lexical stress.


The Early Acranasian phase had, rather than ejectives, gemination and preglottalization in free variation with each other. They had become full ejectives by Classical Acranasian.

Early Acranasian had geminate sonorants as well: /mː nː ŋː ŋʷː lː/, which could not occur in syllable codas. By the Classical period, these had merged with non-geminate sonorants. Compensatory lengthening (and also some diphthongization of sequences of /am:/ and /aŋʷː/) occurred with their deletion on preceding short vowels, so words like ínnu ‘shape’ and ‘issammá ‘hole’ became ī́nu and ‘issaumá, respectively. Word-initially, vowels were epenthesized before their shortened forms (/u/ for /mː  ŋʷː/, /a/ for /mː ŋː lː/ , so words like mmín ‘boy’ became umín.

Classical Acranasian short vowels, which I have listed in the chart as /a i u/, were really more like [ɐ ɪ ʊ].

By Late Acranasian, /l/ was being pronounced [ɾ~ɺ] intervocalically, but still [l] elsewhere.

Also by the Late phase, all nasals assimilated in place to following stops, and short unstressed /u/ was quickly centralizing to [ʉ].


As  mentioned above, Acranasian is an isolating language, with very little inflectional morphology. It is pro-drop and generally (though not completely) head-final, and has a mostly strict SOV word order with optional clause-final topic marking. It, like its descendant South Eresian, has a noun class distinction down animacy lines and lacks any sort of grammatical tense marking.


I don’t really know quite where to start, so here’re four very basic sentences and I’ll explain what’s going on in them afterwards.

Ai ‘iláma
1SG sleep.
“I am asleep”

Ma-xwíła ‘iláma
ABS-woman sleep.
“The woman is asleep”

Xwíła ma-ŋwā́ka ‘ulítap’u.
woman ABS-man hit
“The woman hits the man.”

Xwíła na-ŋwā́ka nák’u.
woman INDIR-man see
“The woman sees the man.”

The prefixes m(a)- and n(a)-, along with the interrogative prefix q(a)- (more on that later), are effectively the only inflectional affixes in the language. They are descended from Proto-Eastern-Noruic proximal and distal demonstratives, and in Acranasian have come to be an absolutive marker and an indirect object marker, respectively.  The absolutive marker m(a)- does not generally occur on pronouns, although it may. These markers only became fully grammaticalized in Late Acranasian, and something like

Xwíła ŋwā́ka ‘ulítap’u.
woman man hit
“The woman hits the man.”

 was still a grammatical sentence, albeit quite archaic-sounding, in Classical Acranasian.

Onto some more complex stuff:

K’iat’ái nī́, ƛáp’a axxula c’a nī́.
stand 1SG, house away go 1SG
I stood up and I walked away from the house.

Xwíła n-a p’áil łłī́ ī ‘umā́kw’a ‘u łłī́.
woman INDIR-3SG speak 3SG MAN know DM 3SG
Because the woman told him, he knew.

These sentence have narrative-type topic marking in it–the nī́ at the end of each clause in the former, which specifies that the topic is the first person singular, and the łłī́ in the latter, which specifies that the topic is third person singular. This started out, as has probably been surmised, in storytelling as a stylistic feature (repetition, like the repetition of the topic here, is a very important part of the style of all Eresian storytelling traditions), and was becoming very common in the Classical phase. By Late Acranasian, this clause-final topic marking was obligatory in all speech.

Also featured is the discourse marker ‘u, the usage of which, like every other discourse marker, is not easy to describe but which frequently occurs with the matrix verb of sentences with subordinate clauses. It always occurs directly after the verb. It can help to disambiguate structure discussed below.

Note the use of the postposition axxula ‘away’ here. This is part of a complex system of postpositions in Acranasian, most of which ultimately turned into the locative system discussed in the last post (relatedly, the verb c’a ‘go’, in the first example, became the inceptive suffix -tz), but which also includes some of the things that ultimately became South Eresian prepositions and a few of the other derivational suffixes. Several postpositions also serve as subordinating conjunctions for adverbial clauses; in the second example, you see ī, which indicates manner, being used in such a way, specifying how the topic knew whatever the information was. More on this:

Amakiā́lqwa axxula c’a nī́ łłā qúlāk’i nī́.
Amachálco away go 1SG because angry 1SG

This contains another postposition functioning as a subordinating conjunction, and this sentence is actually ambiguous: it can mean either “I left (the town of) Amachálco because I was angry” or “I was angry because I left Amachálco.” This ambiguity here arises because of conflict between the postpositional usage of łłā, which essentially treats amakiā́lqwa axxula c’a as a relative clause, and its usage as a clausal conjunction, which, aside from these postpositions slash adverbial clause conjunctions, always precede clauses. By the Classical period, the former was very much the preferred reading (due to analogy with other conjunctions), and clause-final conjunctions eventually fell entirely out of use.

Xxū łłáhhi a’ā ma siā ƛáp’a āp’usa c’a ‘u hi’ī.
IRR rain all.around CONJ N house outside go HORT
If it’s raining, then I shouldn’t leave the house.

Xxū łłáhhi a’ā qa?
IRR rain all.around Q
Is it raining?

Kwī ‘umā́kw’a qa ma xxū łłáhhi a’ā?
2SG know Q CONJ IRR rain all.around
Do you know if it’s raining?

All of these make use of the irrealis marker xxū. I suppose this is perhaps closer to being subjunctive than anything else, although it was in part inspired by the Ojibwe irrealis prefix ji-.

The negative particle siā occurs in the first sentence. This mostly occurs clause-initially.

The particles hi’ī and qa are both modal–the first one marks that the speaker hopes something will happen or thinks it should, and which I generally call hortative (there’s another particle that I call optative, and I won’t go into the nuances of that distinction here). The latter marks questions (not just polar questions, but polar when it occurs without other question marking; see below). Both of these are postverbal particles; the only modal particle that can be classified as preverbal, at least in the Classical period, is xxū. By the Late period, qa was shifting towards being preverbal, influenced by such constructions as in the third sentence above, where it occurs prior to the clause of the question that is being asked pragmatically, if not semantically. The particle a’ā, another postverbal thing, indicates that the verb is occurring all around the deictic center and is very frequently used for weather verbs.

Also seen in the first and third examples above is the coordinating conjunction ma. This occurs in certain constructions, particularly to mark the consequent of a conditional (as in the first) and to mark that the matrix verb of the sentence takes a proposition as an argument (as in the third).

And speaking of clauses…

Ŋwā́ka na-qúlāk’i na-xwíła nák’u łłī́.
man INDIR-angry INDIR-woman see 3SG
The man sees the angry woman.

Ŋwā́ka n-a m-ai ‘ulítap’u iā na-xwíła nák’u łłī́.
man INDIR-3SG ABS-1SG hit REL INDIR-woman see 3SG
The man sees the woman who hit me.

These are relative clauses. As you can tell, they precede their heads–one of the few head-final features that survived into South Eresian. The indirect prefix na- occurs both on the beginning of the relative clause (qúlāk’i ‘angry’ in the first and a (3SG) in the second) and on the head of the relative clause (xwíła, the woman). The second includes the relativizer , which in the Early phase only occurred when the ergative argument of a transitive verb was relativized but in the Classical phase became generalized to all transitive verb relativization. Acranasian relative clauses can be said to be gapped; neither xwíła nor any pro-forms which refer to xwíła occurs within the relative clause proper.

Q-ā́ŋ n-ai qúlāk’i qa?
Q-person INDIR-1SG angry Q
Who is angry at me?

N-ai qa-ŋínū ī qúlāk’i qa?
INDIR-1SG Q-reason MAN angry Q
Why are you angry at me?

Siā ai na-kwī n-ai qúlāk’i ī́ na-ŋínū ‘umā́kw’a.
N 1SG INDIR-2SG INDIR-1SG angry MAN INDIR-reason know
I don’t know why you’re angry at me.

Here are some wh-questions and also how the language handles a statement (the third) that is marked with wh- words in English but not in Acranasian. The first two use the interrogative prefix q(a)- on the element being questioned, with the derivational agentive suffix -āŋ in the former and the noun ŋínū ‘reason’ in the latter. No wh-movement occurs, as is evident in the second sentence. The third sentence makes use of a relative clause kwī n-ai qúlāk’i ī́ “you are angry at me in such a way” on ŋínū, plus the double indirect marking already discussed in the bit on relative clauses (I should note that the verb ‘umā́kw’a ‘know’ takes arguments with n(a)-, not the absolutive m(a)-, for the known thing).

The postposition ī here, seen before in this post a few times, is stressed in the third sentence–this occurs in cases like this when the object of the postposition here is null, as relative clauses in Acranasian use a gapping strategy. The morphophonological principle discussed way up above that applies stress to the initial syllable of phrases that don’t have any predetermined lexical stress thus applies.

That’s all I can remember of the grammar for now. I’m sure there’s more running around in notebooks or frolicking with dust bunnies in the dark corners of my mind.


Acranasian makes use of a fairly large amount of compounding to form new words. This is largely pretty simple stuff. Compounds are head-final, such as in pā́ŋqala-qála, ‘flower-time,’ a poetic term for “spring.” Verb-noun compounds are frequently lexicalized relative clauses:

“clothing worn during mourning”

but for the most part, verb-noun compounds are dispreferred compared to noun-noun compounds. By Late Acranasian, the lexical stress on the second element of most compounds was being dropped.

There are a number of derivational suffixes in the language. One, -ūssu, can be seen above–this one takes a noun or a verb and makes it more general and abstract for the purposes of talking about the general properties of things, in ways roughly equivalent to -ness or -ity in English, eg ƛā́k-ūssu, ‘write-ūssu’, is roughly equivalent to ‘orthography.’ These suffixes are mostly derived from Proto-Eastern-Noruic nouns that had become defunct aside from compounding by Early Acranasian.

Acranasian society was fairly stratified, and this was marked linguistically through the common use of honorific prefixes. Like the derivational suffixes mentioned in the last paragraph, these are derived from lexicalized PEN elements–but in the case of the honorific prefixes, these elements were relative clauses. There is a long list of these (and I keep changing them), but one of the ones that survived most into South Eresian was the prefix ał-, which is affixed to the name of natural phenomena and some abstract concepts to mean “the god of (phenomenon)”– thus do we get Ał-łauhhiā́ ‘sea goddess’, Ał-łā́cu ‘sun goddess,’ Ał-xxānuqā́ ‘god of victory.’

That’s what I’ve got for now. The next post will probably be either an examination of South Eresian morphophonology or an overview of the changes, both phonological and morphosyntactic, that occurred between Late Acranasian and Modern South Eresian. These topics are, of course, pretty closely related. If anyone wants to read one more than the other (in the unlikely scenario that anyone reads this at all), do let me know. Otherwise, feel free to ask questions about the material or correct any stupids I’ve probably made, and thanks for reading!


South Eresian phonology

Okay, I should just give up entirely on writing things I plan to write and focus on writing things I want to write.  Stuff will get done eventually anyway.

I figured I might as well do a writeup of South Eresian phonology, since I’ve got a very good idea of how it works (although I keep changing it, but it’s nearing perfection I think).  At some point I’ll probably give an optimality theory account of it, but for my purposes here I’ll just use boring features.




Note that I listed the lateral continuant as a fricative /ɬ/ rather than an approximant /l/, even though in a large majority of situations it is realized as an approximant.  This is because calling it a fricative simplifies things, as it patterns exactly like the fricatives /s ʃ/, especially in the contexts of phonoactics morphophonology.  I’ll explain this later on.  However, there are some situations (especially allophony) where it is better to refer to it as an approximant as well, so it’s a bit of a toss-up as to which is better in the phoneme chart.  I opted for calling it a fricative, but it would also be very well justified to call it an approximant and I will treat it as an approximant which sometimes becomes a fricative in the allophony section.  Whenever I get an optimality theory account of this up, this will become a non-issue anyway.

The phonemic status of /kʷ kʷʼ/ is somewhat iffy, as they occur only in certain positions and never contrast with /kw kʼw/.  I opted to consider them phonemes only because they can occur in a syllable-initial position, where no consonant+glide clusters may occur.


The orthography of South Eresian is based off of various orthographies invented by the Spanish for Mesoamerican languages, especially Nahuatl.  As South Eresian is highly reminiscent of Mesoamerican languages in many aspects, such an orthography suits it fairly well.

/p pʼ t tʼ k kʼ kʷ kʷʼ q qʼ ʔ/ <p p’ t t’ c/qu c’ cu/uc/c_u c’u q q’ h>
/t͡s  t͡sʼ t͡ʃ  t͡ʃʼ/ <tz tz’ ch ch’>
/t͡ɬ  t͡ɬʼ/ <tl tl’>
/s ʃ x/ <s x j>
/ɬ/ <l>
/ɾ/ <r>
/j w/ <y u/hu/uh>

/a e i o/ <a e i o>
+stress: <á é í ó>

/k/ is written <qu> before /i e/, <c> elsewhere.

/kʷ/ is written <uc> syllable-finally, <cu> elsewhere.  The sequence /kʷɾ/ is written <cru>

/w/ is written <hu> word-initially, intervocalically and after /j/, <uh> word-finally, and <u> elsewhere.  It is also written <hu> following the glottal stop, so that an intervocalic /w/ is written <hu> but a cluster /ʔw/ is written <hhu>.

When /k/ must be distinguished from the cluster /qw/ before /i e/, a diaeresis <ü> is applied to the <u> in the cluster.

I’ve only just changed the transcription of /t͡s  t͡sʼ/ to <tz tz’> from <z z’>, both for æsthetic reasons and because I wanted to bring the orthography a little bit more in-line with Nahuatl.

Stressed vowels are written with acute accents as marked, except when occurring after an ejective; ejectives may only occur in South Eresian at the onset of a stressed syllable, so marking the vowel with a diacritic is redundant and I am lazy.

Where I am positing underlying geminate consonants (to neatly explain some weirdities in certain paradigms), monographs are written doubled andpolygraphs are written thus:

/kː kʷː t͡sː  t͡ʃː  t͡ɬː wː/ → <cqu ccu ttz tch ttl gu>

There are no situations in which ejectives should be usefully posited to occur as underlying geminates.


Phonotactics in South Eresian seem very intuitive to me, although they’ve proven surprisingly difficult to define concisely.  Here are the results of my efforts, with a few notes:

  • For the purposes of phonotactics, /l~ɬ/ is treated as a fricative; this is because it patterns exactly like the other coronal fricatives in the language, and not at all like the sonorants.  I will, as such, be using the glyph <ɬ> for it in this section.
  • The glottal stop /ʔ/ does its own thang instead of acting like the other plosives.
  • I will be ignoring the ejectives in my analysis of the syllable structure; ejectives have extremely limited distribution in the language (they may only occur in the onset position of a stressed syllable), and where they occur in clusters they act exactly like their pulmonic counterparts.

Onset parameters:

Onset parameters in South Eresian are fairly simple and easily defined:

  • [C/G/[p/k/kʷ/q]ɾ]

What this means is that an onset may consist of any single consonant or glide (which fall under G), or one of the set of noncoronal, nonglottal stops /p k kw q/ followed by /ɾ/.

Nucleus parameter:

This is so simple it is barely worth mentioning:

  • [V]

A nucleus can be a vowel.  Easy mode!

Coda parameters:

This is where things become upfucked; this is largely due to the facts that a) word codas are governed by entirely different parameters than govern codas within clusters, and b) codas that occur in clusters are constrained by the onset of the following syllable (so whatever follows them).

All codas, regardless of placement, are limited to being only one segment.

Word codas:

A word coda may be any non-ejective, [+consonantal] thing that is either [-sonorant] or [+nasal] that is both [-labial] and [-back], or it may be both [-consonantal] and [-syllabic] (i.e., a glide).  In practice, this means that word codas are limited to the following set:

  • [n/t/k/t͡s/t͡ʃ/t͡ɬ/ʔ/s/ʃ/ɬ/j/w]

This is pretty straightforward.  However…

Cluster codas:

As mentioned above, the rules governing codas in clusters are complex.  Here is a breakdown of what is going on:

Nasals may occur before any of the following:

  • Homorganic (non-glottal) plosives or affricates
    (It is useful to pretend that /n/, in this case, is completely underspecified for place here, as a homorganic nasal may occur before any non-glottal plosive or affricate)
  • Heterorganic fricatives
    (/n/ here is still somewhat underspecified as it is considered heterorganic with all of /s ʃ ɬ/, even though /s ɬ/ are at a different PoA than /ʃ/.  However, here it is definitely specified as [+coronal], as it may occur before /x/.)
  • Glides

Non-glottal plosives may occur before any of the following:

  • Heterorganic coronal obstruents
    (This rule functions very similarly to the rule regarding nasals and heterorganic fricatives above, with /t/ being somewhat underspecified for place but still marked as [+coronal].  This rule ultimately means that any of the set of /p k kʷ q/ may occur before any of the set of /t t͡s t͡ʃ t͡ɬ s ʃ ɬ/.)
  • Glides

The glottal stop may occur before the following:

  • Sonorants
    (This is the set of /m n ɾ j w/)

Affricates may not occur in any syllable coda position preceding any consonant or glide.

Fricatives may occur before any of the following:

  • Non-glottal stops
    (The category of “stop” notably includes nasals [which are featurally [-continuant]], unlike the category of “plosive” [which I am using to mean things that are both [-continuant] and [-sonorant]]. This rule means that any of the fricatives /s ʃ ɬ x/ may occur before any of the set of /m n p t k kʷ q t͡s/t͡ʃ/t͡ɬ/)
  • Glides

The tap /ɾ/ may occur before any of the following:

  • Glides

Glides may occur before any of the following:

  • Consonants
  • Glides
    (Geminate glides may be posited to occur underlyingly, although they do not surface as such.)

Additional phonotactic details:

  • Non-glottal stop + /ɾ/ clusters should only be analyzed as onset clusters instead of medial clusters when they occur medially.  This is because a) codas, within the parameters specified above, may occur before one of these clusters and b) the allophonic rule listed below that lengthens stressed open syllables applies to syllables preceding these clusters.
  • An underlying content root must contain one, and only one lexical stress.  This is likely to occur on the first syllable of the root, though that is far from being an absolute rule.
  • Underlying forms don’t necessarily need to conform to the phonotactics of the language; these ultimately end up being ironed out by epenthesis and allophonic rules.


I’ve recently been fucking around with South Eresian allophony in order to get something I like a bit better that epenthesizes fewer glottal stops and does not sound like the speaker is choking as much as its previous incarnation.  This explains any discrepancies between what I have posted on the ZBB in the past and what I am posting here now.

I will be writing out these rules in modified featural notation since that is comfortable for me, but I will also explain what they mean in parentheses.  Obligate rule ordering will be found at the bottom of the list.  Don’t expect the rules to necessarily be in order in the list.

Some notational stuff: S denotes any segment.  C denotes any consonant.  V denotes any vowel.  G denotes any glide.

  1. S → [-high] / [-high, +consonantal]_
    (This makes any segment become non-high after non-high consonants.  The only consonants that are specified as [-high] are the uvulars.  In this language’s phonology, this means that /i e o j w x/ become [ɛ ɛ ɔ ɛ̯ ɔ̯ χ] after uvulars.)
  2. S → [-high] / _[-high, +consonantal]
    (This is the exact same thing as the above rule, except it applies before uvulars.)
  3. [+low] → [+back] / [+stress]
    (/a/, which is generally central to central-back [see the Vowels chart above] goes all the way back to [ɑ] when it is stressed.)
  4. [+back] → [+round] / [+labial]_
    (back vowels become rounded after labial things)
  5. [-low, -consonantal] → [-syllabic, +high] / V_]σ
    (This rule means that any non-low, syllable-final vowel becomes nonsyllabic and high after another vowel.  Functionally, this turns the vowels /i e o/ into [j j w].)
  6. [-low, -consonantal] → [-syllabic, +high] / [(V)C/∅]_[-high]
    (This turns /i e o/ into /j j w/ before non-high vowels, provided this does not break the phonotactic rules on consonant clusters.  This was kind of a screwy way to write a rule, and I’d welcome any suggestions to make it cleaner…)
  7. [αPLACE, -low] → [αHEIGHT] / _σ[[αPLACE, αHEIGHT, -consonantal]
    (This rule turns /e o/ into /i u/ respectively before syllable-initial /j w/, via some alpha-notation trickery that I probably screwed up extraordinarily because it is fuck o’clock AM)
  8. [+syllabic] → [αPLACE, -low] / _σ[[αPLACE, -low, -consonantal]
    (This basically turns /a/ into /e o/ before /j w/)
  9. [-glottal, -sonorant] → [+aspirated] / σ[_[+stress]
    (non-glottal obstruents become aspirated in the onset of a stressed syllable.  Ejectives do not aspirate because they are [+glottal].)
  10. [+lateral] → [-sonorant, -voice] / _[+aspirated]
    (/l/ becomes [ɬ] before aspirated consonants)
  11. [+sonorant] → [-vce] / [+aspirated]_
    (the sonorants /ɾ l/ devoice after aspirated things; /m n/ would also devoice here but they never occur in this position)
  12. [+dorsal, +aspirated] → [+delrel]
    (dorsal aspirated things [kʰ kʷʰ qʰ] affricate to [kxʰ kxʷʰ qχʰ])
  13. S: → S
    (any geminate or long segment shortens)
  14. [+coronal, +continuant] → [+delrel] / [+nasal]_
    (the coronal continuants /s ʃ l/ become the affricates [t͡s t͡ʃ t͡ɬ] after nasals)
  15. [+coronal, +continuant] → [+delrel] / [+consonantal, +continuant]_
    (the coronal continuants /s ʃ l/ become the affricates [t͡s t͡ʃ t͡ɬ] after consonantal continuants)
  16. [+dorsal] → [-continuant] / [+nasal]_
    (/x/ becomes /k/ after nasals)
  17. [+dorsal] → [-continuant] / [+consonantal, +continuant]_
    (/x/ becomes [k] after consonantal continuants)
  18. [+nasal] → [αPLACE] / _[αPLACE, -continuant]
    (nasals assimilate to the place of a following stop)
  19. V → [+nasal] / _[+nasal, +dorsal]
    (vowels nasalize before the dorsal nasals [ŋ ɴ])
  20. [-nasal, +sonorant] → [+lateral] / _]σ
    (/ɾ/ becomes [l] syllable-finally)
  21. [+labial, -dorsal, -sonorant] → [+glottal] / _#
    (/p/ glottalizes word-finally.)
  22. [-sonorant] → [-labial] / _#
    (the glottalized [p] created by the last rule becomes [ʔ], and /kʷ/ becomes [k] word-finally)
  23. [+consonantal, -high] → [+high] / _#
    (uvulars become velars word-finally)
  24. [+syllabic, +stress] → [+long] / _]σ
    (stressed vowels lengthen when they are in open syllables)

Rule orderings:

3 < 4

13, 6 < 5 < 1, 2

9 < 10, 11

18 < 19

24 < 13



South Eresian has some epenthesis rules that work in close harmony with the morphophonology of the language, which will be talked about in a different post.    For now, I’ll just say that vowel epenthesis favors keeping consonant clusters together, wherever possible, rather than splitting them apart; /stá/ would be epenthesized as /astá/, whereas ást would be epenthesized as /ásta/.  Additionally, vowel epenthesis is harmonic: if the preceding vowel is one of /i e o/, the epenthesized vowel is /e/, whereas if the preceding vowel is /a/ the epenthesized vowel is also /a/.  If there is no preceding vowel, the epenthesized vowel is /a/.

Cultural note:

The Tlaqoyan dialect of South Eresian, which is the one I am documenting here, has recently had a fair amount of contact with some other regional languages through increased trading.  This has caused it to develop two phonemes that only occur in loanwords: /f/ and /u/.  Neither of these has any real unusual properties, but their increasing presence should be noted.

I’m sure there are fuckups in this post.  I would very much appreciate anyone pointing them out to me.

Cwindoià: Phonemes, syllable structure and prosody

Cwindoià is a language I started coming up with a few days ago.  I intend to make it a polysynthetic language, as I’ve never really played with constructing one of them before.  It’ll probably be based heavily on Aymara.

Phoneme Inventory:

Nasal (N): /m n ŋ/ <m n nh>
Stop (T): /p t k ʔ/ <p t c ‘>
Prenasalized (D): /mb nd ŋg/ <mb nd ng>
Fricative (S): /f θ s ɬ ʂ ʃ ç xʷ h/ <f th s hl hr sh hy hw h>
Liquid (L): /r l/ <r l>
Glide (Y): /j w/ <y w>

Short vowel (V): /a ɛ ə ɔ e o i u/ <a è à ò e o i u>
Long vowel (V:): /aː ɛː ɔː eː oː iː uː/ <á ê ô é ó í ú>
Diphthong (V:) : /aɪ ɛɪ ɔɪ oɪ aʊ ɛʊ ɔʊ eʊ eə oə iə uə/ <ai èi  òi oi au èu  òu eu eà oà ià uà>

A note: I split long vowels and diphthongs up here for space concerns; they are treated identically both phonotactically and prosodically, so they got the same letter.

Syllable Structure:

The maximal syllable structure in Cwindoià is:


What this means is that within a syllable, a long vowel or a diphthong cannot be followed by a coda consonant, and that the only real consonant clusters allowed in a syllable are in the onset position and consist of a voiceless stop followed by a sonorant consonant.

Syllable Weight:

In Cwindoià, there are three syllable weights, that are determined by onset weight and rime weight added together.  Weight may never exceed 2.

Light onsets (+0 weight): T,  N, S, L, Y
Heavy onsets (+1 weight): TL, TY, D

Light rimes (+0 weight): V
Heavy rimes (+1 weight): V:, VT, VN, VS, VL
Superheavy rimes (+2 weight): VD


The language has a pitch accent system.

The assignment of high pitch is based on syllable weight.  The rules for assignment are as follows:

  1. Weightless (0-weight) syllables cannot take high pitch.
  2. On weighted syllables (1 or 2), stress occurs on alternating syllables starting at the word onset.  Two stressed syllables may not occur contiguously in any circumstance.
  3. In any sequence of weighted syllables with unequal weights, the heavier syllable takes pitch.

There is, in addition, a low pitch to be assigned:

  1. Low pitch is assigned to any unassigned syllable that occurs between two unassigned syllables, starting at the word onset.

After high and low pitch are assigned, all syllables with neither high nor low assignments are assigned middle pitch.

Next stop: allophony.