Chris' scratchpad (was: Ch'ubmin)

[chris_notts]

I've been a bit stuck recently on main clause word order. I put something together, but I'm not sure if it's too complex or not.

The long version is here:

https://chrisintheweeds.com/2023/04/22/ ... ndecision/

The short version of the current idea, which I'm not sure I'm completely happy with, is:

1. Null anaphora freely allowed, so most clauses are VS or VO, but when there are two core arguments...
2. Unmarked word order verb initial, broadly speaking verb - topic - focus (i.e. VSO unless the subject is no topical / focal / not backgrounded / information structure deficient compared to expections, in which case VOS is permitted)
3. Focal and non-topical subjects are marked by the ablative-instrumental preposition fe' or the suppletive determiner forms bel~ben~ber. This is normal when word order inversion happens (VOS), when a transitive verb has a subject but no object (VS, null O), when the subject is present but the object is extracted / clefted / head of a relative clause etc. It's more common with transitive subjects, but also possible with intransitive subjects when the clause is thetic or the subject is new.
4. As in many verb initial languages, fronting is allowed for both contrastive / switched topics and for argument focus
5. Topic extraction is via left dislocation, with an intonation break, marked with a demonstrative. There is no impact on the form of the verb
6. Focus extraction is via an ~inverted pseudo-cleft. The focus precedes the verb, which must be in the conjunct form also used for relative clauses and adverbial clauses.
7. These can be combined, in which case the order is topic, focus conjunct.verb
8. There is no pied-piping with topic or focus fronting, so in effect there is a no case before the verb rule. Topics can control a resumptive pronoun. Foci cannot, but either can control verb agreement or possessor agreement on oblique-marking relational nouns which remain in place. This means that neither can be marked nominative/ergative, but some form of oblique role marking can be retained in place.

The main reason I'm not sure about this is that it feels somewhat messy/complex, although I'm not sure if it's more complex than some natural languages.

[Imralu]

Messy/complex is good. This feels OK to me, although I admit I haven't got my head around everything.

[chris_notts]

Messy/complex is good. This feels OK to me, although I admit I haven't got my head around everything.

This is a revised attempt at writing it up for my grammar. I don't know if might help you get your head round it or not, since I've found it hard work to concisely explain the fuzzy pragmatic factors in a way that's low level enough to allow the reader to pick a word order. I think it's probably not enough, e.g. there's still some open questions about the trade-offs or boundaries between normal post-verbal arguments and fronted arguments.


https://chrisintheweeds.files.wordpress.com/2023/04/clausal-word-order-extract.pdf

[bradrn]

I've found it hard work to concisely explain the fuzzy pragmatic factors in a way that's low level enough to allow the reader to pick a word order.

Honestly, I think most reference grammars struggle with this. Very few are written in a way which allows a reader to actually use the language they describe.

[Torco]

I absolutely agree with that. especially true with isolating languages like english or chinese: sometimes if feels like just listing use cases and how the language handles them works better, some websites for learning chinese take this approach.

[chris_notts]

I've been debating for some time whether incorporated nouns should go where they currently do in the template, and reading through Valentine's grammar of Ojibwe has given me more thoughts on the issue and an idea about how to move forward.

https://chrisintheweeds.com/2023/06/21/ ... andwiches/

The idea:

One possible adjustment for Ch’ubmin would be to move in the direction suggested by the above vague speculation. Namely:

[*] Shift towards the coverbs / initials being the ‘root’ and making incorporated classifiers and the closed class verbs more suffix-like… this would be more economical in many ways, since currently most of my verb roots are CVC, but this feels a bit heavy as a firm requirement for a closed class with some high frequency semantically bleached members

[*] Make many of the coverbs describing stable or resultant state also function as adjectives within the NP (Ch’ubmin as written already kind of has this)

[*] Have a set of bound classifier suffixes (including repeaters as an explanation for more generic noun incorporation) which occur both as incorporated absolutives within verbs, as nominal derivational affixes and as adjectival agreement / licensing markers

That is, unify the equivalent of Algonquian verb medials and noun finals and treat [coverb-classifier] ~ [initial-medial] combinations as the incorporation of a fully inflected depictive or resultative adjectival modifier. Since Ch’ubmin also already has noun derived coverbs with causal or instrumental nouns, e.g. rain-hit “get rained on”, this would also imply the kind of selectional restrictions on incorporation suggested by Valentine’s grammar for Ojibwe. Coverbs which are stative~adjectival in origin would be compatible with a wide array of classifiers, whereas a number of noun derived coverbs would be more or less semantically incompatible with incorporated classifiers. Compare:

[split-CL:STRING.LIKE]-cut = cut it split-string = “cut it apart”
??? [rain-CL:PERSON]-hit = hit him/her rain-person = “get rained on (for a person)”

The ~issue with it is that, although it feels kind of elegant to me as a way to solve a number of issues with (1) why incorporated nouns go where they do, (2) which "nouns" can be incorporated, (3) how independent adjectives are formed (since there's already an attempt at a way to make standalone adjectives from some coverbs) etc.... it also breaks my original decision that this time I wasn't going to end up with a load of genders or noun classifiers within NPs and moves a bit in the direction of kitchen sinkery.

[bradrn]

I’m not sure how qualified I am to comment on this, but Aikhenvald’s Classifiers has a nice chapter on verb classification if you’re interested.

[chris_notts]

I’m not sure how qualified I am to comment on this, but Aikhenvald’s Classifiers has a nice chapter on verb classification if you’re interested.

Yeah, I have that book but the two chapters on multiple classifier languages are not that helpful. They're very high-level, mostly for languages that I don't have detailed grammars for, and given that she uses Tariana as an example in places of a multiple classifier language with verbal classifiers, which I don't think it really is... I'd need to see the details.

The issue I have with Tariana as a verbal classifier language is mentioned in my blog post: it's mostly limited to a topic advancing construction which looks a bit like a cleft with a nominalised verb, which is somewhat different to the S/O categorisation type classification of noun incorporation or suppletive verb stems.

[bradrn]

I’m not sure how qualified I am to comment on this, but Aikhenvald’s Classifiers has a nice chapter on verb classification if you’re interested.

Yeah, I have that book but the two chapters on multiple classifier languages are not that helpful. They're very high-level, mostly for languages that I don't have detailed grammars for, and given that she uses Tariana as an example in places of a multiple classifier language with verbal classifiers, which I don't think it really is... I'd need to see the details.

I was thinking more of chapter 6, on verbal classifiers.

[chris_notts]

I renamed the thread because I've been doing a substantial reworking of what I did before, and I don't want to keep creating new threads. The focus has mostly been on a reworked phonology, because I got to a point where the only decisions didn't seem to be quite working for what I was trying to achieve. I was trying for a coverb - independent verb or initial - final type structure (see Australian and Algonquian languages respectively), but the parts didn't feel like they fit together in a pleasing word kind of way. My hope was that going back and designing something with fewer phonemes but more allophonic variation and various adjustment processes (syncope, assimilations, ...) would generate morphologically complex words that felt more integrated.

Having done that, I've then tried to rebuild the verbal morphology of Ch'ubmin, within the new phonology, with fixes for things that felt like they didn't work before. I've retained much but not all of the morpheme ordering, and many of the shapes, with the biggest deviation being in the agreement paradigms and the shape of the agreement morphemes. The next step will be to rework other chapters from my previous grammar, adjusting as necessary.

Here's the current draft, complete with the fake linguistic institute title page I thought it'd be amusing to add...

https://chrisintheweeds.files.wordpress ... rammar.pdf

[chris_notts]

One thing I've realised with this is that the syncope rules definitely drive which suffix shapes work. There are four relevant rules here:

1. Hiatus is mostly resolved by deletion of V1 (with some complexities when V1 is long)
2. Syncope deletes alternating light syllables (light = open, short vowel)
3. Apocope deletes final short vowels
4. Consonants in internal coda position undergo various assimilation processes

If we compare:

CV - weak, vowel can be deleted by both hiatus resolution, apocope and syncope, if vowel deletes then consonant may undergo mergers in coda position
VC - slightly stronger because vowel is not subject to hiatus deletion, and if it syncopates then the consonant will be in onset position
CVC - middling, syncope can reduce to CC, but the cluster is then in intervocalic position so relatively distinctive
CVCV - little better than CVC, since final vowel can be lost to apocope and hiatus resolution
VCVC - stronger, because initial V cannot be targeted by hiatus resolution or apocope, so only risk is syncope producing CVC or VCC surface shapes
...

Basically, given the rules, suffixes that begin with a vowel and end in a consonant look "better" than the other way round in terms of maintaining a basic level of distinctiveness after the various deletion and assimilation rules have applied to them. And to remain distinctive, most non-final suffixes probably want to include two consonants instead of one. Or if not, a long vowel, because long vowels universally resist deletion.

[chris_notts]

I've ported over the TAM semantics chapter from the older Ch'ubmin grammar with basically no changes apart from reworking the examples and any morpheme references to match the new shapes. I still haven't reinvented many words/morphemes yet, beyond the verbal inflections... the trickiest part is getting a feel for what the most natural classification of an event would, in a way that's a bit messy and naturalistic and not overly literal.

Some of my examples in the old grammar used house building, and I've decided that the natural classifier for constructing buildings, walls etc. is probably -pāne "haul, drag, pull, act by force", because my own personal experience of building things is moving heavy materials backwards and forwards. Thus:

gujumpānin tēvos
gu-kuñ-pāne=in tēvos
[ŋʊjʊmpaːnɪn teːvɔʃ]
1SG.PFV-up-haul=DET house
"I built the house"

with a generic kuñ-pāne up-haul = build, or the incorporated version:

gurēvospān
gu-tēvos-pāne
[ŋʊɾeːvɔʃpaːn]
1SG.PFV-house-haul
"I built houses"

with tēvos-pāne house-haul = build houses. The form combining both, tēvos-kuñ-pāne / kuñ-tēvos-pāne, is not right, or at least marked, given the rules in the "Verb Stem Derivation" chapter in the PDF I posted yesterday.

Anyway, I expect that, like Kalam verbs, I need to develop a broad web of meanings for the finals, with the actual meaning narrowed down by the root the final is combined with.

[chris_notts]

Bradrn posting about this triggered me to post about an orthographic dilemma I've been pondering.

I won’t be focussing much on phonology here, but it is worth noting that the preverb as a unit is somewhat distinct from the verb complex proper. For instance, hiatus avoidance does not apply at the end of a preverb: e.g. to-asan ‘it fell’. Similarly, when they contain more than a single subject marker, preverbs tend to get final stress independently of the main verb. On the other hand, processes like nasal assimilation (on which I’ll say more later) do apply across the boundary.

In Qummin, as in Ch'ubmin as it was, there is also one or more phonological boundaries in the grammatical verbal word (or rather, verbal phonological phrase). The verb is minimally:

[auxiliary]-[verb stem]

where the aux marks subject agreement, most TAM categories, reflexive/middle voice and directionality / associated motion, whereas the verb stem hosts two aspectual affixes and object agreement. These show Qummin normal compound word phonology, i.e.:

1. they retain their syllable structure and mostly their stress patterns (although the first shows reduced primary stress / secondary stress only), and neither side can trigger or block normal apocope or syncope in the other
2. but consonant clusters at the boundary are subject to the same cluster assimilation processes as word internal clusters (i.e. a complex set of nasal assimilations, lenition, fortition, gemination...)

In a way, this is a bit like Celtic mutations, which also applied across word boundaries despite other changes being blocked (and of course the conditioning environment was then lost, which was how they got grammaticalised).

Additionally, nouns/classifiers can be incorporated/trapped between the auxiliary and stem, as can certain adverbial elements, in which case they also form part of the compound:

[auxiliary]-[classifier]-[verb stem]
[auxiliary]-[adverbial]-[verb stem]

The question is how to write this orthographically. The internal boundaries are relevant to stress (not written), since compounding resets the footing and rhythm, and to syllabification, since the boundary prevents one side from triggering or preventing vowel deletion on the other.

I currently write such compounds as one word, and show the results of boundary assimilation. I could choose to write them as multiple words, but then the orthography wouldn't really reflect pronunciation. A third option is to insert a - (as Ojibwe does with preverbs), either combined with writing the assimilated forms of the consonants on either side or not. For example:

gap + ñihcut
1sg-FUT + see-3SG
[ŋaɲːiχt͡ʃʊtʰ]
"I saw it"

Could be written as any of:

1. gaññihcut (single word)
2. gañ-ñihcut (dash, assimilation shown)
3. ga-ññihcut (dash, assimilation shown, but geminate grouped together after dash if the result of assimilation is a geminate)
4. gap-ñihcut (dash, no assimilation shown)
5. gap ñihcut (space, no assimilation shown)

Since elsewhere the orthography is very faithful to the pronunciation, I do favour showing the assimilation, which in my head rules out the final (independent orthographic word) solution because there's nothing at all to mark that there's a difference in pronunciation compared to a sequence of words that doesn't form part of the same phonological phrase.

So far I've gone for (1), but I'm tempted to migrate to (2) or (3).