Talk:Akachenti

9/25/18

Notes

Lianamir (talk) 14:00, 25 September 2018 (CDT)

Akachenti

Akachenti appears to be a register-tone language with breathy-low, glottalic-high, and plain-mid registers.

Vowel Raising

1. patient markers experience vowel-shift in certain environments, likely triggered by some consonant distinction loss
2. patient markers lost their consonants, leaving behind subparadigms with vowel-shifted patient markers
3. high tone blocking rule arose
4. paradigms where it operated merged accented and unaccented forms
5. vowel-shifted markers were then generalised to all positions next to low-tone vowels, to reintroduce the contrast

Compare all diphthong locations to voicing/register of onset consonants. Transitioning between registers can cause perception of a glide and break the vowel. Diphthongized vowels contrasting with pure vowels may lose the register distinction.

Something notable here:

• a -> e = raised vowel
• o -> u = raised vowel
• i -> i /= raised vowel and i -> e = lowered vowel

Which is probably why it's rarely indicated and the e sounds have probably merged. In short, you can't raise higher than i.

possible template before synthesis

[ topic / question particle ] + [ recipient (dative / benefactive) / addressee (imperative) ] + [ patient ] + [ inflected verb ] + [ agent ]

General

lost onset consonants:

• voiced -> low tone
• murmured -> low tone
• aspirated -> high tone
• voiceless -> high tone
• nasal -> ? tone/raised vowel

lost coda consonants:

• glottal stop -> high/rising
• fricative -> low/falling
• murmured -> high tone

It has long been recognized that only an extremely limited set of postvocalic consonants contribute directly to pitch generation, specifically the postvocalic consonants that involve a distinctive laryngeal adjustment—glottal stops and -h.

Khmer devoiced initial voiced stops and transferred the distinction to the vowels instead: tense/high and lax/low. This is a classic sound change split and there had to be something that triggered a vowel split on the patientive "affix" in Akachenti.

Degree of breathiness is variable, with bilabial and forward consonants producing more breathiness and also lower/more open vowels will tend toward more breathiness than close ones.

In diphthongization:

1. high lax vowels and low tense vowels tend to remain stable
2. tenseness produces lowered onsets in high vowels
3. laxness produces raised onset in low vowels
4. mid vowels may participate in either pattern

vowel length

"Not too surprisingly, when the tense-voiced vowels are shorter, the tenseness is often derived historically from a final glottal stop."

vowel quality

"The literature documents two correlations between voice quality and vowel quality, one universally attested and the other more marginally attested. The most accurate statement of these correlations is found in Bradley (1982:120), who describes the vowels deriving from the older voiced and, as is obvious from the modern phonetics, breathy-voiced register as “higher” and“more fronted”, while describing the vowels deriving from the older creaky-voiced register as“lower” and “more backed”. The correlation between voice quality and vowel height, that is, F1 is widely attested and uncontroversial: countless scholars have observed that breathy-voiced vowels tend to be relatively higher (e.g., /i/ versus /I/), while tense-voiced vowels tend to be relatively lower (Henderson (1952, 1977), Huffman (1976), Denning (1989), Hombert (1978), Bradley (1982) and so on). The correlation of tenseness (laryngealization) with lower vowels is seen in Mpi, a language in the Lolo-Burmese branch of Tibeto-Burman; Denning (1989:29-33) examined the two-way phonation contrast between modal and tense voice in Mpi and noted that Mpi the tense-voiced vowels showed lowered vowels, that is, the tense-voiced vowels had a higher F1.

"The second correlation, between voice quality and vowel fronting, that is F2, is neither as often observed, nor as clearly designated: in Bradley’s terms (1982:120), breathy-voiced vowels tend to be more backed, while tense-voiced vowels tend to front. Henderson (1952, 1977:259) describes what seems to be the same correlation but in different terms; the breathy-voiced vowels have a “tendency to diphthongize”, while the tense-voiced vowels have a “tendency to centralize”.23 Similar tendencies observed by other authors (e.g. Huffman 1976) tend to parallel either Bradley’s or Henderson’s characterization." - https://pdfs.semanticscholar.org/3880/a6d440431413675afbdf2d37243e291a0f77.pdf

---

1. raising in open syllables
2. pre-rhotic lowering
3. pre-sonorant lowering (exceptions for frequency, 1st syllables in trisyllabic words, not conditioned by affixation)

Huffman (1976), Registrogenesis

• Stage 1: onset distinction, same vowel
• Stage 2: onset distinction, redundant register split
• Stage 3: optional onset distinction, distinctive registers
• Stage 4: full loss of onset distinction

Comments on Conlang Group

Comment Link

So watched a vid yesterday here: David Peterson’s The Art of Language Invention, Episode 27: The Evolution of Tone.

It gave me some direction I think I can actually work with to figure out how Akachenti got it’s tonal accents and possibly enough to backtrack how those standalone pronouns reduced down to the pronominals. In short, breathiness leads to low tone and tense consonants (frequently including glottals) lead to high tone, which definitely correlates with what Akachenti has going.

Also potentially helpful was his video on vowel quality change, which should help me figure out how the high vs. low vowels figure into the tonal accent as well (because they do).

I might actually start conlanging again after a massive block on that front, so looking forward to revisiting Akachenti now.

Today's Reading/Resources

• Vietnamese and tonogenesis: revising the model and the analysis
• Monosyllabicization: patterns of evolution in Asian languages
• Registrogenesis in Khmer: A phonetic account
• THE HISTORICAL PHONOLOGY OF KRIANG, A KATUIC LANGUAGE
• VOWEL HEIGHT AND REGISTER ASSIGNMENT IN KATUIC
• Stages of Tonogenesis: Evidence from Modern Burmese
• THE PHONETICS OF ATHABASKAN TONOGENESIS
• Sonorant-conditioned mid vowel lowering in Turkish

Backreading

Vowel Raising/Lowering in Akachenti

Tbh, I haven't figured out whether vowel raising or lowering is involved because of stress being retracted and glottalized, which is likely to cause vowel-lowering (and i/e/a/u followed by lowering is attested in an excellent natlang example), but then i/e/a/o would let the e displace to let a rise and split and send e into the i region, which could explain ae / e / i showing weird overlaps in their histories. But I have figured out that it's one of them and that figuring out diachronic prosody is my answer to most problems pronoun.

Right now, I'm thinking those elided consonants were possibly uvular, as that's in keeping with the general phonology pattern, the k that shows up in the agentive suffix used with short-fixed-vowel-final verbs, and with vowel lowering, but if vowel-raising turns out the better analysis (in keeping with primary person markers being the lower versions, not the raised versions), then it's likely something else, like a palatal, which is also plausible with the phonology. I'm kind of stuck really digging into both possible avenues into the current vowel set because there's no way I can work with diphthongs in sister languages without nailing down the vowel shifts that led to "modern" Akachenti.

Conversation(s) with Alex Fink

>There's the complication of the associative first person ae / é / ae seeming to be tied to the original 3rd, as evidenced by the vestigial 3rd person e's floating around everywhere and it's continuing to be used as a third person generic when the speaker is "associating" themselves with the generic group referenced. "We should always bide our time." vs. "One should always bide their time."

Hm, the semantic distinctions 1 :: 1.assoc :: 3.generic are very fine, on the grounds of your examples. I think they're easily the sort of thing that could have an ex-nihilo origin. Suppose that for some reason (to be invented) the "associative first person pronoun" used to take third-person agreement, but also that there were two allomorphs of the third-person agreement markers bopping around. One of them might just by some chance (being used in a proverb? phonetic resemblance? etc.) become fixed as the first person associative marker, leaving the other to be the third person marker.

>Third person pronouns don't decline and the associative first pronoun has taken over the regular first altogether.

i.e. free rein to invent the form of the older "regular" first person pronoun. But do try to give it some reason to have been eliminated. Pernicious homophony? A difference of politeness or social connotation?

---

>Which basically means, you can go from mid tone to low tone, but you can't go from high tone to a low tone. So it just switches to the alternate vowel set. Which may not have made a lot of sense if I thought too much about it, but it felt right and worked beautifully in practice, so I just have to learn to understand prosody now that I want to analyze it. And there is also that the the alternate vowel markers can be used for oblique, noncore arguments but stress is pretty much always accusative and sometimes genitive (head-marked).

Neat. So the vowel-changed versions appear only when the accented versions can't, as a surrogate. That to me says that their distribution will be recent. I.e. originally when the high tone blocking rule arose, the paradigms where it operated just merged accented and unaccented forms. But some subparadigm somewhere had these vowel-shifted patient markers (a vanished consonant could still be the culprit, etc.) and they were then generalised to all positions next to low-tone vowels, to reintroduce the contrast.

---

My suspect is that some other morph will be involved in the story. For example, if the patientive was marked by vowel raising, maybe the story is that there used to be a backgrounded agent suffix -C which appeared after the object morphs, which at this stage were identical with the subject morphs. Now assume a sound change that raises vowels before -C, and then lose the agent suffix.

Personal Pronominals

Case Agreement Markers

	Agentive	Patientive		Dative- Benefactive
		Glottalic	Low
1st Excl.	a	á	e	e
1st Incl. (aemen(t)e(r))	ae	é	e	e
2nd Anim. Prox. (usha(r))	a	á	u	u
2nd Obv. (usha(r))	o	ó	u	u
3rd Anim. Prox. (ih)	o	ó	i	i
3rd Obv. (ih)	i	í	i	i

9/26/18

Notes

Lianamir (talk) 11:37, 26 September 2018 (CDT)

Akachenti

Vowel Raising Possibilities

• raising (shortening) long vowels: a: -> e, o: -> u
• lowering (lengthening) short vowels: i -> e, u -> o

---

• a | e -> raised/fronted
• ae | e -> raised/fronted
• o | u -> raised
• (i | i)
• e | i -> raised/fronted

Order of Vowel Quality/Phonation Change

1. breathy register redundantly marks vowels following lenis consonants and possibly fricatives
2. a series of raised vowels (on oblique/patientive pronominals) in breathy register, while the majority continue to be original lower vowels on modal register
3. pronominals lose their consonants, leaving behind vowel-shifted person markers

Uncertain:

1. patientive pronominals lose glottalic consonant and are distinguished by higher pitch

Working on the Verb Template

immediately previous template:

topic / topic-oblique + patient + verb prefix + verb stem + verb suffix + agent

General

English low vowels are lower before voiceless obstruents. The second elements of English closing diphthongs are raised before voiceless obstruents, e.g. rice, house, relative to lower pronunciation before voiced obstruents, e.g. to rise, to house.

breathy-voiced vowels tend to be more backed, while tense-voiced vowels tend to front. tense-voiced vowels tend to be relatively lower and breathy tend to raise

Higher vowels are shorter than lower vowels.

Monophthongs are raised and diphthongs are monophthongized, because higher monophthongs and less wide diphthongs sound longer than, respectively, lower monophthongs and more diphthongal vowels.

Comments on Conlang Group

Comment Link

So been doing a ton of reading on the interactions between tone registers, phonation, pitch, consonant sonorance, and vowel quality: (article list of yesterday and today)

This has been useful in digging into Akachenti, which has a robust and complex personal pronominal system that has been difficult to untangle, primarily I suspect because it’s obviously vestigial of a once more robust pronoun case system that mostly no longer exists. I’m almost certain the reason I’ve been unable to create a full verbal paradigm is because the paradigms have to have come down diachronically from a combination of sound change, pronominal reduction, and person-marking slots in the verbal template trying very hard to merge.

On the bright side, I’ve also learned a lot from all this reading, starting with the fact that breathy phonation is a super simple and common addition to voiced consonants that often leads to redundant breathy phonation on the vowels, which in turn can lead to vowel raising and/or a tone register split in the language (Registrogenesis in Khmer), which is likely exactly what happened to Akachenti.

The two initial registers usually created in this process (or one coming from a glottalic “tense” side instead) are tense and lax. These are relative though, so modal is whichever the other isn’t. In this case, I know I end up with a tense (glottalic), normal (modal), lax (breathy) register system, but am not sure which came first. It’ll have to shake out though because both were clearly triggered by consonant or consonant distinction loss, as they do.

Right now I’ve got:

1. breathy register redundantly marks vowels following lenis consonants and possibly fricatives
2. a series of raised vowels (on oblique/patientive pronominals) in breathy register, while the majority continue to be original lower vowels on modal register
3. pronominals lose their consonants, leaving behind vowel-shifted person markers

Uncertain:

1. patientive pronominals lose glottalic consonant and are distinguished by higher pitch

But this is theory one. Separately we have theory two, which gets at the fact that there’s this old table which points out agreement markers don’t always line up perfectly with the original freestanding morphemes they agreed with.

Case Agreement Markers

	Agentive	Patientive		Dative- Benefactive
		Glottalic	Low
1st Excl.	a	á	e	e
1st Incl. (aemen(t)e(r))	ae	é	e	e
2nd Anim. Prox. (usha(r))	a	á	u	u
2nd Obv. (usha(r))	o	ó	u	u
3rd Anim. Prox. (ih)	o	ó	i	i
3rd Obv. (ih)	i	í	i	i

and also this table:

a / ae / o / i	-ar / -er	e / u / i	-at / -et / -ot / -it
agentive	genitive-accusative	oblique	genitive
agent	patient	patient	genitive
instrument/causee	possessor	benefactive	ablative ( re: )
posssessee		comitative
		dative
		causer

These two old tables led me to these notes today:

So the differences of the personal pronominals could be indicative of lost case.

If the original persons were:

• 1st excl.
• 1st incl.
• 2nd prox.
• 2nd obv.
• 3rd prox.
• 3rd obv.

There’s pretty much only one natlang family that has shown 2nd person obviation, but this is the only thing that makes sense to me with the pronominal markers on animate, proximate second person, even when combined with the usual freestanding pronoun as used for obviate.

And let’s posit there were three cases:

• agentive (assoc. with a / ae / o / e)*
• patientive (assoc. with á / e / ó / é) – caused by (glottalic?) consonant loss
• oblique (assoc. with e / e / u / i) – caused by (breathy?) consonant loss

*likely the -k(V) suffix that surfaces on small, old, high-frequency verbs, e.g. da(ke)

Of note in here is that I listed the agentive and patientive 3rd person markers as e and é, despite the fact that in Modern Akachenti, they are i and í. This is because there is a marked pattern of vowel raising from the agentive/patientive vowel quality to the oblique. In Modern Akachenti, the third person vowels have merged, but there are traces of /e/ being used where I would expect /i/. It makes more sense that the original vowel quality was then /e/, which raised to /i/ where expected and then began merging both markers, rather than the original quality being /i/ and having unexpected and inexplicable lowering to /e/.

Additionally, there is some confusion that comes from cases shifting meaning over time. In German, there are good examples of dative being used for some functions of the accusative case. In Akachenti, this sort of case movement has definitely occurred and will have to be looked at closer later.

Today's Reading/Resources

• VOICING AND VOWEL RAISING IN SUNDANESE - immediate download
• A vowel height split explained: Compensatory listening and speaker control

9/27/18

Notes

Lianamir (talk) 13:44, 27 September 2018 (CDT)

Akachenti

Sound Changes

So somewhere along the way before we got to Modern Akachenti, there were sound changes. Right now, I've got some conditioned sound changes with no final rules as yet of what the conditions are:

1. a₂ > aɛ (unconditioned?)
2. e₂ > e₁ (visible in 2nd incl pronominals)
3. a₁ > e (visible in 1st oblique pronominals and in medial syllables, i.e. "construct" form)
4. o > u (visible in 2nd oblique pronominals)
5. e₁ > i (visible in 3rd person pronominals)

Current ongoing sound changes are:

1. aɛ > {a,aɪ,e} (unconditioned?)

9/28/18

Notes

Lianamir (talk) 13:09, 28 September 2018 (CDT)

Akachenti

Agreement markers currently agree with case and person, but only arguably hits topicality (covered by order) and doesn't seem to handle number. It feels odd that number would be such a non-thing in Akachenti.

General

Mechanisms of Syntactic Change edited by Charles N. Li

Agreement markers usually arise from forms that were independent pronouns at an earlier stage of the language, those forms later attaching as clitics or affixes to the verb.

In the Iroquoian case, if this reconstruction is correct, the aggreement markers arose instead from a reinterpretation of affixes that were already being used for other purpose. Some originated as markers of number, then extended to mark person, then subsequently to mark gender.

In the Northern languages, agreement is for number, person, case, and gender (in the third person). Number includes singular, dual, and plural (sometimes "nonsingular" dual/plural). Person is 1st, 2nd, 3rd with clusivity marked in dual/plural on 1st. Case is agentive (doers, perceivers, subjects of epistemic verbs) and patientive (experiencer, certain recipients and beneficiaries). Gender includes masculine, feminine, and nonhuman.

Iroquoian have three classes of pronominal prefixes: agent, patient, and transitive (both).

Proto-Iroquoian-Caddoan had agreement for 1st and 2nd with number and clusivity but 3rd person had only number agreement. There was however a nonspecific or indefinite prefix which agreed in case. This in Caddoan remains unspecfic, but in Iroquoian languages was extended to general third person, then restricted to human, then further to feminine. Three other third person markers seem to have arisen from number markers. ka- was began as specific third person agent, later restricted to nonhuman, then extended to feminine as well in nonsingular or everwhere. Agents picked up dual/plural first but patients still make do with nonsingular.