Page 1 of 2

Constructing syllabaries for complex syllables?

Posted: 08 Mar 2017 15:06
by MoonRightRomantic
I am trying to construct a syllabary for phonotactic constraints more complex than CGVC in the fewest amount of graphemes. The idea is that a civilization that spoke a language allowing complex syllables devised a logo-syllabary to write their language.

So far I am aware of the following methods syllabaries have devised for dealing with complex syllables
  • Echo or dummy vowels; as in Mayan, Linear B or Cherokee
  • Digraphs for glides; as in Kana
  • Graphemes for onset and/or coda; as in Cree, Kana or Cherokee
  • Graphemes for body and rime; as in Akkadian or Sumerian
  • Graphemes for initial, medial and rime; as in Bopomofo or Kommodam's Khom script.
If anyone has other suggestions or workarounds I would love to know.

Re: Constructing syllabaries for complex syllables?

Posted: 08 Mar 2017 15:20
by clawgrip
Stacking is also a possibility. I did something like this for a vertical syllabary I made, though this was only for coda consonants and geminates. When you had a coda consonant, you would simply write two glyphs beside each other. This is basically the echo/dummy technique you mention, except with an explicit visual cue that a dummy vowel is being used.

Image
Sịmekkalit Ngwụddọ mul rupu chahaimat sẽpet Rụkụ̃t zẽ mul tuzọfọu shek tẽrĩ tẽri pasa mil mil maike ọghụ.

You can see the first word Sịmekkalit (top left) is written as sị / me / ka-ka / li-ti, with echo vowels (there is ambiguity between kka and kak, but that is part of the fun of it). Also this script uses signs for coda nasals, which also mark voiced geminates, but that's a different story.

Re: Constructing syllabaries for complex syllables?

Posted: 08 Mar 2017 16:32
by masako
I have something similar, but less complex and/or ornate.

http://www.omniglot.com/conscripts/kala.htm

Re: Constructing syllabaries for complex syllables?

Posted: 09 Mar 2017 02:54
by Man in Space
You could go the Mayan route and "underspell" certain syllables.

Re: Constructing syllabaries for complex syllables?

Posted: 09 Mar 2017 18:55
by MoonRightRomantic
clawgrip wrote:Stacking is also a possibility. I did something like this for a vertical syllabary I made, though this was only for coda consonants and geminates. When you had a coda consonant, you would simply write two glyphs beside each other. This is basically the echo/dummy technique you mention, except with an explicit visual cue that a dummy vowel is being used.

Image
Sịmekkalit Ngwụddọ mul rupu chahaimat sẽpet Rụkụ̃t zẽ mul tuzọfọu shek tẽrĩ tẽri pasa mil mil maike ọghụ.

You can see the first word Sịmekkalit (top left) is written as sị / me / ka-ka / li-ti, with echo vowels (there is ambiguity between kka and kak, but that is part of the fun of it). Also this script uses signs for coda nasals, which also mark voiced geminates, but that's a different story.
Writing Systems and Cognition refers to echo vowel sequences of <CV-CV> producing /CCV/ or /CVC/ as "progressive" and "regressive", respectively. It is impossible to determine which (or both!) of these is being used, which is what made Cypriot so difficult to read.

I am not sure how a pointing system would work, except to insert a vowel or diacritic next to the syllable nuclei whenever a sequence of the same vowel occurs. For example, <po-ro-o-ca-la-a-ru-u-ʃu> or <po-ró-ca-lá-rú-ʃu> would produce /proclaruʃ/.
Linguifex wrote:You could go the Mayan route and "underspell" certain syllables.
The Cypriot and Linear B syllabaries represented Greek inaccurately, often omitting inconvenient onsets and codas on top of using echo and/or dummy vowels. For example, /knōsos/ was represented as <ko-no-so> in Linear B; /anthrōpos/ as <a-to-ro-po-se> in Cypriot.

That is what I am trying to avoid.

Re: Constructing syllabaries for complex syllables?

Posted: 09 Mar 2017 20:32
by sangi39
MoonRightRomantic wrote:
clawgrip wrote:Stacking is also a possibility. I did something like this for a vertical syllabary I made, though this was only for coda consonants and geminates. When you had a coda consonant, you would simply write two glyphs beside each other. This is basically the echo/dummy technique you mention, except with an explicit visual cue that a dummy vowel is being used.

Image
Sịmekkalit Ngwụddọ mul rupu chahaimat sẽpet Rụkụ̃t zẽ mul tuzọfọu shek tẽrĩ tẽri pasa mil mil maike ọghụ.

You can see the first word Sịmekkalit (top left) is written as sị / me / ka-ka / li-ti, with echo vowels (there is ambiguity between kka and kak, but that is part of the fun of it). Also this script uses signs for coda nasals, which also mark voiced geminates, but that's a different story.
Writing Systems and Cognition refers to echo vowel sequences of <CV-CV> producing /CCV/ or /CVC/ as "progressive" and "regressive", respectively. It is impossible to determine which (or both!) of these is being used, which is what made Cypriot so difficult to read.

I am not sure how a pointing system would work, except to insert a vowel or diacritic next to the syllable nuclei whenever a sequence of the same vowel occurs. For example, <po-ro-o-ca-la-a-ru-u-ʃu> or <po-ró-ca-lá-rú-ʃu> would produce /proclaruʃ/.
(point out the Stargate reference first)

Well Devanagari and other abugidas sometimes use a special diacritic to mark the fact that a given syllabic sign did not have a vowel. I can't see any reason why a syllabary couldn't develop a similar marker. Start out with echo syllables, then have writers start to indicate which ones are pronounced in full, or not pronounced at all. There's nothing saying that there has to be a next step so it could just end there.

If the script were borrowed you could go down the route of:

PI PU PA > /pi pu pa/ vs. PA-I PA-U > /pe po/
SI-PI SU-PU SA-PA > /spi spu spa/ vs. SA-PI SA-PU > /spe spo/
SI-I-PI SU-U-PU SA-A-PA > /sipi supu sapa/ vs. SA-A-PI SA-A-PU > /sapi sapu/
SA-I-PI SA-U-PU > /sepi sopu/ vs. SA-A-PA-I SA-A-PA-U > /sape sapo/
... etc.

MoonRightRomantic wrote:
Linguifex wrote:You could go the Mayan route and "underspell" certain syllables.
The Cypriot and Linear B syllabaries represented Greek inaccurately, often omitting inconvenient onsets and codas on top of using echo and/or dummy vowels. For example, /knōsos/ was represented as <ko-no-so> in Linear B; /anthrōpos/ as <a-to-ro-po-se> in Cypriot.

That is what I am trying to avoid.
The "underspelling" in Linear B, and to a lesser extent the Cypriot Syllabary, partly comes from the script being a borrowed invention, as well as their use. Linear B was, as far as I can remember, mostly used for texts relating to tax (and other "state" stuff), so the vocabulary was relatively limited and supplemented with logograms. The chances for ambiguity were, IIRC, relatively low as a result and familiarity with Mycenaean Greek certainly would have aided in reading. From what I can tell, Cypriot was used in a wider range of contexts, which is probably why underspelling was somewhat remedied, but again familiarity with the spoken language probably would have aided in reading (much as familiarity with spoken Hebrew or Arabic aids in reading their respective scripts despite not writing vowels at all). (see your own example of Linear B A-TO-RO-QO vs. Cypriote A-TO-RO-PO-SE).

Underspelling isn't necessarily a bad thing and it might actually make the script more interesting [:)]

Re: Constructing syllabaries for complex syllables?

Posted: 10 Mar 2017 14:50
by MoonRightRomantic
I think I figured out a scheme to solve my consonant problem and make my syllabary stand out from other syllabaries.
  • Onset clusters would be represented with progressive spelling a la Cypriot: <CV-CV> would produce /CCV/ if the vowel is echoed.
  • Coda consonants would represented with rime syllabograms, or a diacritic indicating reversal, a la Sumerian: <CV-VC> or <CV-CVᴿ> would produce /CVC/ if the vowel is echoed.
  • Coda clusters, if necessary, would be represented using regressive spelling and rime syllabograms in exact reversal to onsets: <VC-VC> would produce <VCC> if the vowel is echoed.
A maximal syllable /CCVCC/ would be spelled <CV-CV-VC-VC> or <CV-CV-CVᴿ-CVᴿ>. If fictional history is necessary, perhaps the reversal diacritic developed later to distinguish progressive and regressive spelling and to spell /VC/ syllables in fewer letters.

Critique?

Re: Constructing syllabaries for complex syllables?

Posted: 20 Mar 2017 13:35
by MoonRightRomantic
Since syllabaries seem so rare as far as constructed scripts go, I might as well start making some for variety. I suppose it would be faster to make one set of syllabograms that may be written using one of multiple modes like Tengwar, at least to start with.

This depends on the phonotactic constraints of the language: any maximal syllable more complex than /CGVX/ (where C = consonant, G = glide, V = vowel, X = consonant or 2nd part of a long vowel or diphthong) would probably be better served by an alphabet or alphasyllabary.

The typical set of syllabograms consists of vowels and consonants followed by vowels, as exemplified by the Cherokee, Greek, Kana and Mayan syllabaries. Represented closed or complex syllables might rely on a system of vowel harmony of varying complexity, or follow the path of least resistance and approximate most of the underlying phonemes, rarely using distinct onset/coda graphemes.

The Aboriginal syllabary uses body graphemes and coda graphemes (which rarely double as onset for certain classes of words), whereas Bopomofo or Kommodam's script uses rime graphemes and onset graphemes (which rarely double as coda when writing, say, Cantonese). The Sumerian syllabary uses both body and rime syllabograms and a simple system of echo vowels.

The Paleohispanic languages seem to have allowed complex syllables, with the notable constraint that plosives must always be followed by a vowel (or semivowel?). So their younger scripts were alphabetic with the exception of syllabograms for plosives.

Some food for thought, I guess...

Re: Constructing syllabaries for complex syllables?

Posted: 20 Mar 2017 13:57
by clawgrip
Are you going to start designing the actual glyphs?

Re: Constructing syllabaries for complex syllables?

Posted: 21 Mar 2017 11:43
by elemtilas
MoonRightRomantic wrote:Since syllabaries seem so rare as far as constructed scripts go, I might as well start making some for variety.
Are they really so rare? Not arguing, just curious as to perceptions. Wonder if anyone's ever tried a census of invented script types.
Some food for thought, I guess...
Lots of good ideas in there!

Re: Constructing syllabaries for complex syllables?

Posted: 21 Mar 2017 11:54
by masako
elemtilas wrote:Wonder if anyone's ever tried a census of invented script types.
I don't know, but here would be a decent place to start.

Re: Constructing syllabaries for complex syllables?

Posted: 21 Mar 2017 13:53
by MoonRightRomantic
clawgrip wrote:Are you going to start designing the actual glyphs?
I have been scribbling in the margins for a couple of weeks trying to see what works. For the phonemes I decided to start with a fairly simple /p t k m n ŋ f s x w r j a i u/ giving 15×3=45 syllabograms.
elemtilas wrote:Are they really so rare? Not arguing, just curious as to perceptions. Wonder if anyone's ever tried a census of invented script types.
I am currently doing a census of Omniglot conscripts for conlangs and natlangs. Along the way I noticed that some conscripts are hidden because of incomplete indexing. But from my preliminary analysis it seems that most conscripts are alphabets, then alphasyllabaries, then everything else (syllabaries, consonantal, semanto-phonetic). The numbers may be skewed because the samples are mostly American and European in origin, but the smaller Asian samples demonstrate an analogous bias to emulate indigenous scripts (in this case, Brahmic scripts, Hangul and Bopomofo).

In conclusion, the vast majority of conscripts surveyed attempt to map phonemes to graphemes on an average 1:1 ratio.

Re: Constructing syllabaries for complex syllables?

Posted: 21 Mar 2017 15:01
by clawgrip
Syllabaries generally require a large number of glyphs, so I can imagine people avoid it in favour of something with a smaller inventory. Or they want to fake a larger inventory, so they go with an abugida/alphasyllabary.

Re: Constructing syllabaries for complex syllables?

Posted: 23 Mar 2017 00:57
by elemtilas
clawgrip wrote:Syllabaries generally require a large number of glyphs, so I can imagine people avoid it in favour of something with a smaller inventory. Or they want to fake a larger inventory, so they go with an abugida/alphasyllabary.
Indeed. Of course, "faking" it with too few symbols is a valid path to go if it's not really your plan to write out long / complete texts in the native script. The alphasyllabary is also a way to go.

As far as complex ones go, any ideas on how many symbols would be required to make a syllabary for (insert here your relatively standard value of) English? How many syllables do we even have?
MoonRightRomantic wrote:
elemtilas wrote:Are they really so rare? Not arguing, just curious as to perceptions. Wonder if anyone's ever tried a census of invented script types.
I am currently doing a census of Omniglot conscripts for conlangs and natlangs. Along the way I noticed that some conscripts are hidden because of incomplete indexing. But from my preliminary analysis it seems that most conscripts are alphabets, then alphasyllabaries, then everything else (syllabaries, consonantal, semanto-phonetic). The numbers may be skewed because the samples are mostly American and European in origin, but the smaller Asian samples demonstrate an analogous bias to emulate indigenous scripts (in this case, Brahmic scripts, Hangul and Bopomofo).

In conclusion, the vast majority of conscripts surveyed attempt to map phonemes to graphemes on an average 1:1 ratio.
Interesting. I'd have thought syllabaries would be more in evidence!

Re: Constructing syllabaries for complex syllables?

Posted: 23 Mar 2017 11:08
by masako
elemtilas wrote:As far as complex ones go, any ideas on how many symbols would be required to make a syllabary for (insert here your relatively standard value of) English? How many syllables do we even have?
Depends on the dialect, and very much on what constitutes a 'syllable'.

Re: Constructing syllabaries for complex syllables?

Posted: 23 Mar 2017 13:31
by clawgrip
There are tens of thousands or more possible syllables in English, and many of them are unused (e.g. sloorkth), so a true, basic syllabary would be impossible. You would need to incorporate various strategies that have been discussed in this thread, and the total number of glyphs in your English syllabary would depend very much on which and how many of those strategies you chose to use.

Re: Constructing syllabaries for complex syllables?

Posted: 23 Mar 2017 14:12
by MoonRightRomantic
clawgrip wrote:Syllabaries generally require a large number of glyphs, so I can imagine people avoid it in favour of something with a smaller inventory. Or they want to fake a larger inventory, so they go with an abugida/alphasyllabary.
I would contest that, at least if vowels are indicated with diacritics. Most of the alphasyllabic constructed scripts I surveyed are highly regular in comparison to natural scripts. It seems to be more difficult for someone not to unconsciously emulate or regularize the conventions of their native script unless they consciously defy it.
elemtilas wrote:As far as complex ones go, any ideas on how many symbols would be required to make a syllabary for (insert here your relatively standard value of) English? How many syllables do we even have?
clawgrip wrote:There are tens of thousands or more possible syllables in English, and many of them are unused (e.g. sloorkth), so a true, basic syllabary would be impossible. You would need to incorporate various strategies that have been discussed in this thread, and the total number of glyphs in your English syllabary would depend very much on which and how many of those strategies you chose to use.
There are exactly 15,831 syllables used in the English language (src). Someone devised an "emoji syllabary" for writing English, although it is structured more like hieroglyphics (src).
elemtilas wrote:Interesting. I'd have thought syllabaries would be more in evidence!
According to Peter T. Daniels, and corroborated by my preliminary survey, constructed scripts tend toward syllabaries if their creator is illiterate and alphasyllabaries if their creator is literate.

Re: Constructing syllabaries for complex syllables?

Posted: 23 Mar 2017 14:41
by clawgrip
MoonRightRomantic wrote:
clawgrip wrote:Syllabaries generally require a large number of glyphs, so I can imagine people avoid it in favour of something with a smaller inventory. Or they want to fake a larger inventory, so they go with an abugida/alphasyllabary.
I would contest that, at least if vowels are indicated with diacritics. Most of the alphasyllabic constructed scripts I surveyed are highly regular in comparison to natural scripts. It seems to be more difficult for someone not to unconsciously emulate or regularize the conventions of their native script unless they consciously defy it.
I am always slightly disappointed when I see constructed alphasyllabaries that are 100% regular in formation. As an example of what I mean, many letters in Devanagari have a vertical bar on the right which can be removed in order to form consonant clusters. If this were a conscript, I think it's very likely that every consonant would have this vertical bar in the exact same place, and every cluster would be transparently formed simply by removing it. However, in actual Devanagari, there are several letters that don't have a bar to remove and which must therefore form their ligatures irregularly, and other letters that form have the bar and yet form unintuitive and unexpected ligatures.

This step of introducing irregularity is what makes a script interesting and more realistic, which is why I am disappointed when I see a more systematic and universal approach, and this occurs more frequently than I wish it would.

Re: Constructing syllabaries for complex syllables?

Posted: 23 Mar 2017 16:54
by MoonRightRomantic
clawgrip wrote:I am always slightly disappointed when I see constructed alphasyllabaries that are 100% regular in formation. As an example of what I mean, many letters in Devanagari have a vertical bar on the right which can be removed in order to form consonant clusters. If this were a conscript, I think it's very likely that every consonant would have this vertical bar in the exact same place, and every cluster would be transparently formed simply by removing it. However, in actual Devanagari, there are several letters that don't have a bar to remove and which must therefore form their ligatures irregularly, and other letters that form have the bar and yet form unintuitive and unexpected ligatures.

This step of introducing irregularity is what makes a script interesting and more realistic, which is why I am disappointed when I see a more systematic and universal approach, and this occurs more frequently than I wish it would.
Most of the alphasyllabaries I have seen lack ligatures. They are generally written like you would write with the Latin alphabet, except vowels are appended to preceding consonants (if any) as diacritics.

It is not wholly unrealistic. This sort of regularity is normal for scripts invented by literate scribes to cover natural languages that lack their own script. Hangul and scripts devised by missionaries all demonstrate this.

Re: Constructing syllabaries for complex syllables?

Posted: 23 Mar 2017 20:04
by elemtilas
MoonRightRomantic wrote:According to Peter T. Daniels, and corroborated by my preliminary survey, constructed scripts tend toward syllabaries if their creator is illiterate and alphasyllabaries if their creator is literate.
Oh, interesting indeed! I must be semiilliterate, then...
[}:D]