Talk:Proto-Dumic

Am I right to assume this was canonised somewhere on the appropriate secret board of the forum and is just now being extracted and published from there?

(P.S. the addition of a manner adverb suffix -mi I find... disappointing. Morphologically distinct manner adverbs are one of those crosslinguistically rare features which are however familiar in SAE and therefore too prevalent in conlangs. Surely Proto-Dumic could have gotten by with its already ample palette of deverbalisers.) 4pq1injbok (talk) 06:02, 5 February 2015 (PST)

Yes, this is all coming from the secret forum, which I believe is to be made public Real Soon Now; to minimise waiting I've also posted a link to Serafín's PDF grammar sketch. I didn't have much of a hand in the writing, and I am aware that a number of features were debated (e.g. the participles had support from Basilius but not really anyone else); some of these can probably be changed retrospectively, though. thedukeofnuke (talk) 07:39, 5 February 2015 (PST)

Isn't it public already? :P the Devilcat (talk) 08:28, 5 February 2015 (PST)

Ah, so it is. Forum delving to do, then, which is a little intimidating.

As those watching the changelog will know, I have gone and started a Dumic language. It actually sprang into mind in 2011 when I was pitching into the T1 reconstruction effort a little (and "sprang into mind" is correct; if not I wouldn't've chosen to add another member to an over-replete family). It was Cedh encouraged me to make it official / write it up publicly.

At any rate, that biasses me in the direction of wanting to retain the things that actually got reconstructed. The participles were among these (and DLNAF attests *-kaga, though not *-tini); *-mi e.g. was not. 4pq1injbok (talk) 09:30, 5 February 2015 (PST)

I will remove the *-mi suffix in that case; I can't remember if it shows up in any daughters, but if it does we can treat it as an innovation. thedukeofnuke (talk) 02:24, 13 February 2015 (PST)

RE: Participles. I forget if I made it clear back then, but the reasons for me to support them were roughly the following: (1) they were already used in one of the daughterlangs; (2) WeepingElf's "Subordinator" -ki also appeared (in some of his examples) as equivalent of complement clauses, i. e. looked indeed like a generic marker of secondary predication, so it seemed appropriate to have a specifically adnominal form beside it (perhaps with a difference in usage in adnominal position). --Basilius (talk) 13:10, 17 February 2015 (PST)

RE: Adverb markers. I wouldn't drop them for the following reasons.

(1) There are some adverbial-oids with this marker in the protolanguage:

kimi : adv. in no way (also emphatic negation)

mumi : adv. thus, like that

rammi : adv. how?

simi : adv. thus, like this

They shouldn't be taken as a direct attestation of the productive model (since they aren't derived from predicate stems), but they show that the language did use some morphologically distinct adverbs, with the marker in question.

(2) Alternatives (e. g. case forms of infinitives) would look clumsy.

(3) Straightforward revisions of "official" descriptions of this type are IMO a Bad Thing. Perhaps just my own obsession, but anyway.

--Basilius (talk) 13:25, 17 February 2015 (PST)

(1) is a fair point, though also a possible locus of analogical extension. Perhaps there are only a few more forms than the "correlatives" with -mi kicking around Proto-Dumic, and Kataputi (e.g.) heavily generalised the pattern.

I disagree with (2) as a good reason. Plenty of natlangs give no dedicated expression to manner adverbs; if the result seems "clumsy" one of my first impulses would be to wonder if that's a result of looking at the problem through SAE-coloured glasses. Case forms of nominalisations, or (just as likely) such caseforms with a lexical scattering of other methods, make a wholly reasonable strategy.

(3) I'm not going to dispute. 4pq1injbok (talk) 11:31, 3 March 2015 (PST)

---

Did anyone ever make a Proto-Dumic word for 'all'? I want one. (As a base for DLNAF 'both'; my current word kovita is untenable.) 4pq1injbok (talk) 07:31, 16 October 2015 (PDT)

If not, and I got to propose one, I'd propose mumsam = mumma + ram, formed as per WALS chapter 56. 4pq1injbok (talk) 12:21, 16 October 2015 (PDT)

I believe that if a lexical item isn't in this thread:

http://akana.conlang.org/forum/viewtopic.php?f=13&t=75

- then it hasn't been officially adopted.

The lexicon (including root lexicon) needs to be expanded anyway, so I'd suggest simply adding a new root, since the meaning is very basic. It was intended to be an open process; you pick a blank root here from the second message in the aforementioned thread, assign a meaning to it and post to the same thread. If no objections are posted in reasonable time, it's official. Since the thread has been silent for years, it may be a good idea to send notifications to other people who used to be involved. This may awaken a few dormant participants like this one, who may even feel an urge to resume work on their creations ;)

(Conversely, introducing a brand-new type of morphological processes in the language, e. g. morpheme truncation, without prior discussion - is probably NOT the right way to work on a common protolanguage of several other people's creations...) --Basilius (talk) 14:06, 23 October 2015 (PDT)

Ah, secret things. I still haven't actually gone and read that forum. I'll go undertake that process, I suppose.

But to answer you here: I'm certainly not proposing that Proto-Dumic show a truncation process in... well, in any way that would be sensible to document in a grammar sketch. I'm merely proposing that it show a relic of grammaticalisation, which I explicitly do want here, having the chance, because it's underexploited in conlangdom as a whole -- e.g. three natlangs in eight show a process like this one, but can you name any conlang that does? Grammaticalisation is well-known to be associated with irregular losses of phonetic material, such as may not be exampled in any morphological process that was ever productive. And as this is a basic meaning, I felt it was compelling to give it a bisyllabic form, like everything else basic, and not the trisyllabic ordinary compound. Synchronically it would seem to be a basic root. 4pq1injbok (talk) 02:19, 24 October 2015 (PDT)