Page 1 of 1
conlang speech segmentation...
Posted: Sun Jun 02, 2024 3:25 pm
by xxx
in writing, the spaces between words are enough
to produce the self-segmentation of the language flow...
unless you import natural languages,
it's hard to imagine how ambiguities can be avoided,
and whether they would paralyze the transmission of meaning...
conlangs, which are mostly silent languages,
have little data on their prosodies...
and tonic accents are often doubled by stereotyped syllables,
as in Esperanto...
in 3SDL, where the word limit is irrelevant
(I see it more as a one-word language),
I got around it with triple S equality (1Sense=1Sign=1Sound)...
and you, have you thought about it...
Re: conlang speech segmentation...
Posted: Sun Jun 02, 2024 9:52 pm
by keenir
xxx wrote: ↑Sun Jun 02, 2024 3:25 pm
in writing, the spaces between words are enough
to produce the self-segmentation of the language flow...
Greek called...couldn't stop laughing long enough to leave a message.
unless you import natural languages,
it's hard to imagine how ambiguities can be avoided,
I'm sorry - what?? Are you saying that natlangs...have no ambiguities?
conlangs, which are mostly silent languages,
have little data on their prosodies...
I would bet hard money that there is, in fact,
lots of data about the sounds of conlangs.
and you, have you thought about it...?
as we used to say in my day, "fixed it for you"
Re: conlang speech segmentation...
Posted: Mon Jun 03, 2024 1:09 am
by bradrn
keenir wrote: ↑Sun Jun 02, 2024 9:52 pm
xxx wrote: ↑Sun Jun 02, 2024 3:25 pm
in writing, the spaces between words are enough
to produce the self-segmentation of the language flow...
Greek called...couldn't stop laughing long enough to leave a message.
And Chinese, and Japanese, and Thai, and Lao, and Khmer, and Burmese, and…
For that matter, even in speech it’s not always easy. I’ve been having great difficulty learning to understand French, and word segmentation is a big part of that. My understanding is that spoken French has basically no phonetic cues to word boundaries (insofar as one can define wordhood for French anyway, which is something I’m not sure about).
Re: conlang speech segmentation...
Posted: Mon Jun 03, 2024 2:30 am
by Ares Land
Oh, I think prosody is something I'd like to work on; I should try and get some resources on that.
I think there's something valid in the issue xxx is bringing up; my conlanging work is mostly in written form, which means some issue probably don't get worked out.
Re: conlang speech segmentation...
Posted: Mon Jun 03, 2024 6:02 am
by xxx
keenir wrote: ↑Sun Jun 02, 2024 9:52 pm
bradrn wrote: ↑Mon Jun 03, 2024 1:09 am
Dudes, we're in "conlangery"
in natlangs, usage and reality allow us to determine,
how to get past any ambiguities in segmentation.
For an L2, as you said, it's more complicated,
but in "conlangery",
how can we segment the flow of a language without real usage that wouldn't be based on natlangs...
in writing, could you do without spaces between words,
and when speaking, what markers do you use...
Re: conlang speech segmentation...
Posted: Mon Jun 03, 2024 7:14 am
by Zju
obligatory xkcd
Re: conlang speech segmentation...
Posted: Mon Jun 03, 2024 11:05 am
by keenir
xxx wrote: ↑Mon Jun 03, 2024 6:02 ambut in "conlangery",
how can we segment the flow of a language without real usage that wouldn't be based on natlangs...
And why should we not be allowed to base anything on natlangs? Your conlang uses the same script as several natlangs, however mutable it seems to be.
Re: conlang speech segmentation...
Posted: Mon Jun 03, 2024 1:02 pm
by Travis B.
xxx wrote: ↑Sun Jun 02, 2024 3:25 pm
in writing, the spaces between words are enough
to produce the self-segmentation of the language flow...
unless you import natural languages,
it's hard to imagine how ambiguities can be avoided,
and whether they would paralyze the transmission of meaning...
conlangs, which are mostly silent languages,
have little data on their prosodies...
and tonic accents are often doubled by stereotyped syllables,
as in Esperanto...
in 3SDL, where the word limit is irrelevant
(I see it more as a one-word language),
I got around it with triple S equality (1Sense=1Sign=1Sound)...
and you, have you thought about it...
Remember, we here at the ZBB typically create
artlangs, which typically seek naturalism as a goal, where in another reality someone very well could speak them. Trying to eliminate naturalism for the sake of clarity would in and of itself go against this goal.
Re: conlang speech segmentation...
Posted: Mon Jun 03, 2024 2:03 pm
by xxx
it's not a problem of naturalism,
as seen with the bradrn's post about French...
unless you assume that it's spoken “elsewhere”,
without testing it yourself (and why not),
my questions are valid for all conlangs :
in writing, could you do without spaces between words,
and when speaking, what markers do you use...
even if it's less important with a posteriori language,
or if you borrow from natural languages
or with relex...
Re: conlang speech segmentation...
Posted: Tue Jun 04, 2024 2:35 am
by Creyeditor
I have given Kobardon prosody some thought and I am pretty happy how the interaction of word-level pitch accent, intonation and rhythmic lengthening plays out. I will poat more information in due time in the thread over at the CBB.
Re: conlang speech segmentation...
Posted: Tue Jun 04, 2024 1:03 pm
by Travis B.
Natlangs rely on a variety of different features, ranging from word, morpheme, and syllable boundary-sensitive phonological rules, word structure-dependent stress and pitch placement, word-formation rules (e.g. fixing the order of elements in verb complexes while allowing freer orders outside them), and so on.
Re: conlang speech segmentation...
Posted: Tue Jun 04, 2024 1:51 pm
by xxx
Travis B. wrote: ↑Tue Jun 04, 2024 1:03 pm
Natlangs rely on a variety of different features, ranging from word, morpheme, and syllable boundary-sensitive phonological rules, word structure-dependent stress and pitch placement, word-formation rules (e.g. fixing the order of elements in verb complexes while allowing freer orders outside them), and so on.
what's about your conlang...
Re: conlang speech segmentation...
Posted: Tue Jun 04, 2024 4:37 pm
by Travis B.
xxx wrote: ↑Tue Jun 04, 2024 1:51 pm
Travis B. wrote: ↑Tue Jun 04, 2024 1:03 pm
Natlangs rely on a variety of different features, ranging from word, morpheme, and syllable boundary-sensitive phonological rules, word structure-dependent stress and pitch placement, word-formation rules (e.g. fixing the order of elements in verb complexes while allowing freer orders outside them), and so on.
what's about your conlang...
For Fad'ami I chose word boundary-dependent stress, where the leftmost heaviest syllable out of the last three syllables of a word gains stress. In this it phonologically encodes word boundaries.
Re: conlang speech segmentation...
Posted: Tue Jun 04, 2024 5:42 pm
by keenir
xxx wrote: ↑Mon Jun 03, 2024 2:03 pmunless you assume that it's spoken “elsewhere”,
without testing it yourself (and why not),
How do you test if its spoken elsewhere?
my questions are valid for all conlangs :
what questions? it felt more like you were listing objections, such as not being able to make conlangs without referencing natlangs. (boy are you gonna be shocked by Dritok)
Re: conlang speech segmentation...
Posted: Wed Jun 05, 2024 2:01 am
by xxx
Travis B. wrote: ↑Tue Jun 04, 2024 4:37 pm
For Fad'ami I chose word boundary-dependent stress, where the leftmost heaviest syllable out of the last three syllables of a word gains stress. In this it phonologically encodes word boundaries.
It's a bit irregular to place the accent in such a way as to recognize a word, isn't it,
have you tried to understand sentences orally...
for my part, I admit that I have the greatest difficulty understanding spoken words,
although there's no segmentation difficulty, quite the contrary,
but the amount of information to recognize is high
with a meaning on each phoneme,
and my ear isn't trained for it...
as much as I use 3SDL fluently in writing,
speaking poses a problem...