Umbrean Script

[zelos]

With this thread getting questions about the huge numbers in the background story of the question itself I figured I will give a deeper explination of the script I am currently working on.

I had the idea for this scriptal structure someitme ago, how long ago I cannot remember but it is unlike the big five, logographic, syllabic, abugida, abjad and alphabet. Yet it shares simularities with a few of them.

The general structure phoneticly each grapheme represent is CVC, a consonant vowel and then lastly another consonant. If I recall right this is the structure of the general logogram in chinese.

But I have always had a delight for abugidas, the thought of a little diacritic to change the vowel just strikes me as so lovable so that is in my writing aswell. A little diacritic that changes the vowel.

So in an essence every grapheme stands for a given consonant pair with an inherent vowel that is modified with diacritics. But a thing to notice is that the grapheme for P-M has no similarity to P-L or M-P, so they are all independed and unrelated.

One gigantic drawback with this is for n different phonemes the script will demand n² graphemes to be cover all possibilities. And that makes the amount grow extremely large extremely fast.

The best description in terms of old systems I have found is bi-consonantal abugida or my own "logugida" logographic+abugida.

I have none to present now visually as I am working on the graphical stuff, I have only worked out the structure of the script and this is the base one, still working out the details to reduce the number of graphemes I need by having imperfections and ways to compensate them.

[Xing]

One gigantic drawback with this is for n different phonemes the script will demand n² graphemes to be cover all possibilities. And that makes the amount grow extremely large extremely fast.

How many consonants are there? Can any consonant appear in any position? How many of the possible syllables would actually be utilised?

I'd guess that the "worst" thing that could happen is that the script would be equivalent to a logographic script, in terms of how many characters need to be memorised.

Edit: If you total 15129 C-C combinations, and all consonants can appear both initially and finally, there must be 123 consonants (or perhaps consonant combinations??). That's quite many, which makes the question how many of these would actually be utilised even more crucial.

[rickardspaghetti]

Would semi-vowels count as consonants or vowels?

[zelos]

One gigantic drawback with this is for n different phonemes the script will demand n² graphemes to be cover all possibilities. And that makes the amount grow extremely large extremely fast.

How many consonants are there? Can any consonant appear in any position? How many of the possible syllables would actually be utilised?

I'd guess that the "worst" thing that could happen is that the script would be equivalent to a logographic script, in terms of how many characters need to be memorised.

Edit: If you total 15129 C-C combinations, and all consonants can appear both initially and finally, there must be 123 consonants (or perhaps consonant combinations??). That's quite many, which makes the question how many of these would actually be utilised even more crucial.

Correct deduction, the amount in placement and "types" aren't that made, if you exclude features like voicing, gemination, palatalization and labialization you get about 13 consonants, those 4 features increase the number afterward. Though checking with lists of natural conlangs I have made it sure it is within reason, a high number yes, but there are enough languages that has comparatively many. when I first saw the numbers a long time ago I got terrified, but having made some research atleast calmed my nerves in that.

Would semi-vowels count as consonants or vowels?

Semi-vowels would be tricky but I would say it depends on how they are used, if they are used like in english "yes" it is a consonant, but if it is like some, correct me if I am wrong, nucleus of a syllable like in slavic languages it would count toward a vowel then.

[Xing]

Correct deduction, the amount in placement and "types" aren't that made, if you exclude features like voicing, gemination, palatalization and labialization you get about 13 consonants, those 4 features increase the number afterward. Though checking with lists of natural conlangs I have made it sure it is within reason, a high number yes, but there are enough languages that has comparatively many. when I first saw the numbers a long time ago I got terrified, but having made some research atleast calmed my nerves in that.

How many of these possible C-C combinations would actually be used? I don't know how many vowels you have. But say you have five of them - that would result in more than 75000 CVC-glyphs, which seems to be more than enough to cover everyday, non-specialist language. (cf the ~4000 or so hanzis that are used in Chinese newspapers...)

Would semi-vowels count as consonants or vowels?

Semi-vowels would be tricky but I would say it depends on how they are used, if they are used like in english "yes" it is a consonant, but if it is like some, correct me if I am wrong, nucleus of a syllable like in slavic languages it would count toward a vowel then.

If a semi-vowel is the nucleus of a syllable, it is a vowel. There may be some problematic cases, though. (Should, for instance, the /j/ in English words like new and cute count as a part of the onset - as a part of a consonant cluster, or as a part of the nucleus - as a part of a diphthong?)

Do you have a phoneme chart for your lang?

[zelos]

Correct deduction, the amount in placement and "types" aren't that made, if you exclude features like voicing, gemination, palatalization and labialization you get about 13 consonants, those 4 features increase the number afterward. Though checking with lists of natural conlangs I have made it sure it is within reason, a high number yes, but there are enough languages that has comparatively many. when I first saw the numbers a long time ago I got terrified, but having made some research atleast calmed my nerves in that.

How many of these possible C-C combinations would actually be used? I don't know how many vowels you have. But say you have five of them - that would result in more than 75000 CVC-glyphs, which seems to be more than enough to cover everyday, non-specialist language. (cf the ~4000 or so hanzis that are used in Chinese newspapers...)

Oh heavens no! The vowel is treated like in an abugida, no diacritic and it is "pal" and with a diacritic it becomes "pel" and so on. But I am abondening the richness of that system, it is not feasable so I go for a less perfect system that needs between 169 (13²) and 1014 (6*13²) instead.

There may be some problematic cases, though. (Should, for instance, the /j/ in English words like new and cute count as a part of the onset - as a part of a consonant cluster, or as a part of the nucleus - as a part of a diphthong?)

If its a diphtong I would say treat it as a vowel, if its part of the consonant I would qualify it more as palatalization of the consonant

Do you have a phoneme chart for your lang?

Yes, but currently I am not posting it on here until I feel the conlang is suffiently done to be represented. This post was mostly to discuss the system I thought of rather than details of my own conlang.

[Xing]

Oh heavens no! The vowel is treated like in an abugida, no diacritic and it is "pal" and with a diacritic it becomes "pel" and so on. But I am abondening the richness of that system, it is not feasable so I go for a less perfect system that needs between 169 (13²) and 1014 (6*13²) instead.

If you have 15000 CC-graphemes, that can be combined with 5 vowel diacritics (or how many they are... and including a zero-digraph), I get that there will be 75000 glyphs in total. Or where does the calculation go wrong?

[zelos]

Oh heavens no! The vowel is treated like in an abugida, no diacritic and it is "pal" and with a diacritic it becomes "pel" and so on. But I am abondening the richness of that system, it is not feasable so I go for a less perfect system that needs between 169 (13²) and 1014 (6*13²) instead.

If you have 15000 CC-graphemes, that can be combined with 5 vowel diacritics (or how many they are... and including a zero-digraph), I get that there will be 75000 glyphs in total. Or where does the calculation go wrong?

I believe the mistake is mostly that a question of definition, if you mean glyph=grapheme+diacritic then you are correct.

But as said, thats far too many for me to even be arsed to make that many, so I go for an inherently imperfect system that is not as rich in details to minimize the quantity fo graphemes

[GBR]

If I were looking to cut down on the number of necessary characters here, I might decide to mark other features with diacritics as well, e.g. how Japanese uses two dots to mark voicing: さ /sa/ ざ /za/. It also uses a circle to turn /h/ onset syllables to /p/ onset syllables: へ /he/ ぺ /pe/.
And then, it uses smaller versions of /j/ onset syllables to show palatalization, e.g. きや/kija/ but きゃ/kja/.

Obviously Japanese only has one permissable coda, /n/, which they write separately, but still, I should imagine you could have a set of diacritics for vowels, and then another for 'initial consonant palatalized and vowel', or simply whack two diacritics on one character, one for palatalization and one for the vowel.

Also, there's nothing wrong with straight up ambiguity. English orthography has plenty of it (<c> can be /k, s, ʃ, tʃ/ in the right circumstances, etc etc). What if the characters for /pwat/ and /pat/ look exactly the same, but context can tell people which it is? Might be unwieldy if most of your words are monosyllabic, but if not I can see it working.

You keep saying the script is logographic - does that mean there's some semantic information in the character as well?

[zelos]

If I were looking to cut down on the number of necessary characters here, I might decide to mark other features with diacritics as well, e.g. how Japanese uses two dots to mark voicing: さ /sa/ ざ /za/. It also uses a circle to turn /h/ onset syllables to /p/ onset syllables: へ /he/ ぺ /pe/.
And then, it uses smaller versions of /j/ onset syllables to show palatalization, e.g. きや/kija/ but きゃ/kja/.

Obviously Japanese only has one permissable coda, /n/, which they write separately, but still, I should imagine you could have a set of diacritics for vowels, and then another for 'initial consonant palatalized and vowel', or simply whack two diacritics on one character, one for palatalization and one for the vowel.

Also, there's nothing wrong with straight up ambiguity. English orthography has plenty of it (<c> can be /k, s, ʃ, tʃ/ in the right circumstances, etc etc). What if the characters for /pwat/ and /pat/ look exactly the same, but context can tell people which it is? Might be unwieldy if most of your words are monosyllabic, but if not I can see it working.

You keep saying the script is logographic - does that mean there's some semantic information in the character as well?

I did consider doing that, though my attempts to do that and have it aesteticly pleasing have been nothing but failures it just looked aweful with shit all over the place

I kind of was thinking about that ambiguity, though more along the lines of "pat" and "pjatj" do exist as graphemes but there are no "pjat" and "Patj"

I don't keep sayign it because it isn't. What I did say is that it bears resamblance to chinese due to its syllable structure. Logugida is about as close as I get.