Zompist Bboard Again

quinterbeck wrote: ↑Fri Jan 15, 2021 12:57 pm
I love this aesthetic.

The sample has a surprising number of English words in it (I count 21 out of 150)

Yeah. I started working on the Foundation series, where these where I imagine supposed to be worn out English, but it's funny how much it matches a lot of SF. (I tried it on names from Iain Banks' Culture series, and it fits a lot of names!)

I did manage to get a reasonable phonology out of it.

In one of the older boards a piece of software was once mentioned, that could look through your words and tell you which minimal pairs you have. Does anyone know what it was called and if it's available anywhere?

Yalensky wrote: ↑Thu Dec 31, 2020 10:21 pm Congrats to the Board Lord for receiving David J. Peterson's Smiley Award. A very nice write-up of Verdurian and Zompist's overall influence on conlanging.

Kudos! Very cool.

I'm beginning to feel that all my conlangs are ripoffs of Burushaski.

Well, not really, but Burushaski does contain a really implausibly high number of my favorite features. Honestly I'm surprised it doesn't get more discussion here, it's exactly the sort of language (isolate, typologically unusual, etc.) I'd assume would generate a lot of interest from conlangers.

I hope this isn't against board etiquette, but in case anyone is interested, I'm running a "typological voting game" on the CBB where we're gradually building a sketch grammar by voting on features from the World Atlas of Language structures. If anyone has an account over there, feel free to drop in and cast a vote. This week we're doing stress systems and reduplication. Here's the link.

dɮ the phoneme wrote: ↑Sun Feb 07, 2021 4:36 pmI'm beginning to feel that all my conlangs are ripoffs of Burushaski.

Well, not really, but Burushaski does contain a really implausibly high number of my favorite features. Honestly I'm surprised it doesn't get more discussion here, it's exactly the sort of language (isolate, typologically unusual, etc.) I'd assume would generate a lot of interest from conlangers.

Bad advertisement; should've listed some sample features.

Someone told me about mood prominence and I thought that would be a neat idea for my conlang instead of just nonfuture vs future. Although I barely understand the concept from what I've read and what the resulting implications would entail. Still, I like the idea of combining futurity, intentionality, potentiality, and conditionality all in one thing.

Anyone got any info on the so-called "theme" vowels of Semitic roots? I know there was some kind of resource being like "roots with i in them indicate stative while those with a or u are dynamic blah blah"

After working for a couple of months on my other hobby, converting the source code of the PC version of Rick Dangerous 2 to C++ and SDL, having more or less finished that I felt a sudden need to go conlanging again. A while ago I played with the idea of creating a proper Germanic conlang, reading up on Proto Germanic grammar and vocabulary and such, but due to time constraints and my quest for perfection it never really took off. That's why I decided to make a Germanic bogolang: it looks like Germanic, it feels like Germanic, but it's in no way actual Germanic.

So I've now created the first sketches of a language that has 3 genders (m/f/n), 6 cases (nom/gen/dat/acc/loc/voc), 2 conjugated tenses (present/past), 2 aspects (perfective/imperfective) and 2 moods (declarative/subjunctive). The vocabulary I took from the oldest stages of various Germanic languages (mostly West-Germanic ones) and reconstructed proto-Germanic, the declension and conjugation patterns fully made up but created to look vaguely Germanic. Here's an example sentence:

Hildaskon was skonar maget, synar hera skones so sy guldas waste, unda synar anskona was wysas skonar, so sy ganges den alawer yr skajwanar.

[ˈhildɐˌskon wɐs ˈskonɐr ˈmɐget ˈsiːnɐr ˈherɐ ˈskones so siː ˈguldɐs ˈwɐste ˈundɐ ˈsiːnɐr ˈɐnskonɐ wɐs ˈwiːsɐs ˈskonɐr so siː ˈgaŋes den ˈɐlɐwer iːr ˈskajwɐnɐr]

Hildaskon

Hildaskon

was

be.PST.IMP

skon-ar

beautiful-3.F.SG

maget

maiden.F.SG.NOM

syn-ar

3.SG.POSS-F.NOM

hera

hair.F.NOM

skon-es

shine.PST-3.SG.PST.IMP

so

as_if

sy

3.F.SG.NOM

guld-as

gold-GEN

waste

be.3.SG.PST.SBJV

unda

and

syn-ar

3.SG.POSS-F.NOM

anskona

appearance.F.NOM

was

be.3.SG.PST.IMP

wysas

so

skon-ar

beautiful-F.SG.NOM

so

when

sy

3.F.SG.NOM

gang-es

walk-3.SG.PST.IMP

den

then

alawer

everyone

yr

3.F.SG.ACC

skajw-anar

look_at-3.PL.PFV

Certain words can be transparently derived from other words, e.g. wysas is the genetive of wysa, manner, and alawer is the prefix ala-, "all", with the noun wer, "man", "person".

“Hildaskon was a beautiful maiden, her hair shone as if it were golden, and her appearance was so beautiful, when she walked (by) then everyone looked at her.”

EDIT: Forgot the translation.


JAL

Are there any conventions for glossing intruding non-morphemic material? I looked through the Leipzig rules but didn't find anything relevant. I don't how prevalent such a process is in natlangs (or even if it's an acceptable analytical approach).

Some examples in Leima of phonological rules for affixation:

1.) When an open syllable with a close vowel precedes a syllable beginning with a vowel or glide, an epenthetic consonant is inserted as follows:

	_vowel	_glide
i_	n	d
u_	m	g

The action auxiliary u (gloss: ACT) plus the singular third person pronoun -we (3s) surfaces as ugwe (ACT-3s)

ugwe youm

u-we

ACT-3s

youm

sing

they_SG sing

2.) When an open syllable with the vowel /a e o/ precedes a syllable beginning with a glide, the vowel becomes one in the 'combined' class

	orthographic	IPA
a	aa	ɑː
e	ea	ɛ͡ə
o	oa	ɔ͡ə

deanaayor
[dɛ͡əˈnɑːjɔɺ]

deana-yor

good-least

worst

quinterbeck wrote: ↑Sat Mar 27, 2021 1:38 pm Are there any conventions for glossing intruding non-morphemic material

Yeah, you don't gloss it. It's like you said non-morphemic and doesn't have any meaning, it's just allophony or sandhi etc.


JAL

I liked the approach of Mihas in her writings on Alto Perené, where she glossed epenthesis with EP:

No=sai-t-aty-e=ro nihaa
1SG.A=pour-EP-PROG-IRR=3n.m.O water

I will pour the water.

quinterbeck wrote: ↑Sat Mar 27, 2021 1:38 pm Are there any conventions for glossing intruding non-morphemic material? I looked through the Leipzig rules but didn't find anything relevant. I don't how prevalent such a process is in natlangs (or even if it's an acceptable analytical approach).

Yes, you just put E or EPEN(TH) to indicate an epenthetic "morpheme"

jal wrote:Yeah, you don't gloss it. It's like you said non-morphemic and doesn't have any meaning, it's just allophony or sandhi etc.

Sometimes it's important to indicate epenthesis so people don't get confused about the random insertion of a phoneme and/or erroneously think the epenthetic morpheme is part of this or that actual morpheme.

Ahzoh wrote: ↑Sat Mar 27, 2021 9:56 pm Yes, you just put E or EPEN(TH) to indicate an epenthetic "morpheme"

jal wrote:Yeah, you don't gloss it. It's like you said non-morphemic and doesn't have any meaning, it's just allophony or sandhi etc.

Sometimes it's important to indicate epenthesis so people don't get confused about the random insertion of a phoneme and/or erroneously think the epenthetic morpheme is part of this or that actual morpheme.

Aha! Thanks, avoiding confusion was exactly my concern

I've been frustratingly torn on many aspects of my conlang.

It is a rather verb-heavy language, so I'm not sure whether my adjectives should behave more noun-like (participles) or more verb-like (verb with relative suffix). And there's the matter of what shape the verbal adjective should take at all, whether C₁aC₂C₃- or C₁aC₂C₂aC₃- Alternatively, I must decide what taking the verb stem and adding the noun affixes does, whether it turns it into an adjective, or a noun of action. Then there's deciding what to do about making biliteral roots with short vowels best fit the tricon paradigm. It's easy to say a biliteral with a long vowel is just a tricon with a middle weak root.

I also worry about words being too "long" for the commonness or simpleness of the information conveyed (e.g. a 4- or 5-syllable word meaning "red"). I don't wanna pull a Navajo where some nouns are basically a dictionary definition.
I also feel like I've limited myself because it's hard to come up with affixes because they either accidentally copy Semitic or they don't harmonize well with other consonants. I also feel like choosing against subject prefixes has limited me somewhat as well (because my lang is tricon and I don't wanna copy Semitic).

I'm stuck, this is frustrating.

Ahzoh wrote: ↑Tue Mar 30, 2021 3:33 am I've been frustratingly torn on many aspects of my conlang.

Been there, done that… you’re hardly alone!

It is a rather verb-heavy language, so I'm not sure whether my adjectives should behave more noun-like (participles) or more verb-like (verb with relative suffix).

You may be interested to know that there is a rather strong set of statistical universals — covered quite well in Stassen’s article on Black and white Languages — depending in part on whether adjectives are nouny or verby. For instance, nouny adjectives correlate with VO word order, non-tensedness and casedness, while verby adjectives correlate with OV word order, tensedness and non-casedness. You may find this helps somewhat to make a decision.

I also worry about words being too "long" for the commonness or simpleness of the information conveyed (e.g. a 4- or 5-syllable word meaning "red"). I don't wanna pull a Navajo where some nouns are basically a dictionary definition.

I sometimes end up with this problem (mostly when trying to come up with numbers), so I’d also like to know what other people have to say about this.

I also feel like I've limited myself because it's hard to come up with affixes because they either accidentally copy Semitic or they don't harmonize well with other consonants. I also feel like choosing against subject prefixes has limited me somewhat as well (because my lang is tricon and I don't wanna copy Semitic).

Perhaps try deriving them diachronically if you aren’t already? (If you do derive them diachronically, then copying Semitic is absolutely fine, since you’ve justified it as a coincidence.)

bradrn wrote: ↑Tue Mar 30, 2021 3:59 am You may be interested to know that there is a rather strong set of statistical universals — covered quite well in Stassen’s article on Black and white Languages — depending in part on whether adjectives are nouny or verby. For instance, nouny adjectives correlate with VO word order, non-tensedness and casedness, while verby adjectives correlate with OV word order, tensedness and non-casedness. You may find this helps somewhat to make a decision.

The distant ancestor of Vrkhazhian started off as a head-initial VSO language, but Vrkhazhian has become strongly SOV and displays some elements of other head-final properties (somehow), like being primarily postpositional. I do imagine that its primary distinctions, the realis and irrealis mood, have developed into non-future and future and whatever cases it may have had has collapsed into a mere Nominative-Oblique system and it has obivation markers on the verb.
Despite it's VSO origins it has a large sweet of attributive verbs.

I got my verbs down pat and even a means of turning nouns into verbs by way of a causative lexical derivation. Still I'm uncertain where to derive adjectives. The relativizer suffix is a good way but that should be active in meaning but I also want passive adjectives but I have no passive voice.

Perhaps try deriving them diachronically if you aren’t already? (If you do derive them diachronically, then copying Semitic is absolutely fine, since you’ve justified it as a coincidence.)

I do some level of diachronics but not too far because that's the path toward frustration and dissatisfaction if you're not a diachronics wizard. That's another source of frustration. No matter how many books of diachronic developments in various languages I read I can never think of them myself and the words will barely look like they changed or shorten. Or they will break stuff and lay waste to my paradigms undesirably.

Ahzoh wrote: ↑Tue Mar 30, 2021 4:36 am I got my verbs down pat and even a means of turning nouns into verbs by way of a causative lexical derivation. Still I'm uncertain where to derive adjectives. The relativizer suffix is a good way but that should be active in meaning but I also want passive adjectives but I have no passive voice.

I honestly have no idea what you mean by this paragraph. Why would you need to ‘derive’ adjectives from something else?

Perhaps try deriving them diachronically if you aren’t already? (If you do derive them diachronically, then copying Semitic is absolutely fine, since you’ve justified it as a coincidence.)

I do some level of diachronics but not too far because that's the path toward frustration and dissatisfaction if you're not a diachronics wizard. That's another source of frustration. No matter how many books of diachronic developments in various languages I read I can never think of them myself and the words will barely look like they changed or shorten. Or they will break stuff and lay waste to my paradigms undesirably.

I also had the same problem a while ago. What fixed it was reading more about diachronics — not general books about historical linguistics, that is, but about the specific sound changes which specific languages have gone through. I’d recommend Blust’s The Austronesian languages as a starting point; it has a nice chapter (Chapter 9) about the sorts of sound changes found in Austronesian. In recent days, I also found it useful to participate in the Phrase evolution game — my suggestions were pretty horrible at first, but even after just a week I feel a lot more confident now. It’s particularly helpful to look at the sort of changes other people have been doing.

(Also, I should point out that for triliteral systems, you want to start with a normal system and ‘lay waste’ to it, as you say. Past that point, you can maintain regularity even with massive sound changes, since analogy immediately restores any paradigms which might have been broken. Deutscher’s The Unfolding of Language has a nice explanation if you want one.)

bradrn wrote: ↑Tue Mar 30, 2021 5:46 amI honestly have no idea what you mean by this paragraph. Why would you need to ‘derive’ adjectives from something else?

There are very few adjectives that just exist and they're usually only interrogatives, demonstratives, or numbers. The rest have to be derived, either from nouns or verbs. That's my entire problem, I don't know what the "pattern" should look like or if it's just verb stem + gender-case ending (which would work better in a lang where adjectives can behave like nouns and vice-versa)

I also had the same problem a while ago. What fixed it was reading more about diachronics — not general books about historical linguistics, that is, but about the specific sound changes which specific languages have gone through.

Oh no, I wasn't talking about generalist books, but books of specific languages, like Akkadian and Tigre, and some other languages of other families.

(Also, I should point out that for triliteral systems, you want to start with a normal system and ‘lay waste’ to it, as you say. Past that point, you can maintain regularity even with massive sound changes, since analogy immediately restores any paradigms which might have been broken. Deutscher’s The Unfolding of Language has a nice explanation if you want one.)

Don't take this the wrong way, but I'm far from a noob regarding triliteral systems or the languages's pervasive vowel mutations and analogizing, and also their limitations. In fact, that book is quite old hat to me now.

There's just no way I can simulate thousands of years of changes without losing my mind or being burnt out. I do have a general idea of how this or that thing would operate or turn into something else within a triconsonantal system. But I just don't have the imagination to turn something like kurpaneddakazen "rostrum, dias" or dapsenadkazen "registry" into something shorter.

But the diachronic wizards on this board and others make such changes seem too easy and thoughtless.

quinterbeck wrote: ↑Sun Mar 28, 2021 4:59 amAha! Thanks, avoiding confusion was exactly my concern

Then mention it once, and gloss without the inserted phones. Having it in each and every gloss would become bothersome quickly.


JAL