Script preference

A forum for discussing linguistics or just languages in general.
Post Reply
User avatar
CarsonDaConlanger
cuneiform
cuneiform
Posts: 167
Joined: 02 Nov 2017 20:55

Script preference

Post by CarsonDaConlanger » 18 Sep 2018 00:35

I often see reference to the fact that Hànzì (Or semanto-phonetic scripts in general) wouldn't work well for synthetic languages. Why not? Wouldn't it be trivial to just have an accusative symbol etc?

User avatar
Pabappa
cuneiform
cuneiform
Posts: 199
Joined: 18 Nov 2017 02:41
Contact:

Re: Script preference

Post by Pabappa » 18 Sep 2018 01:38

One problem is the size of the glyphs. if you have a large number of glyphs in the language, most word roots will consist of just one or two glyphs. But what size are you going to use to indicate the inflections? Suppose your word for "slap" is a single glyph, but you have dozens of verb inflections that are indicated by a few glyphs apiece .... you may find that your verb inflections take up more space than the actual verbs. imagine if in english the whole word "Slap" was forced into a single square, while the ending "-PED" was written at four times that font size, with each letter taking up as much space as the entire root. Thus it takes four times as much space to write a simple inflected form of the root as it does to write the root itself, and all writing in this language will be forced to use smaller font sizes. This will lead to people squinting to read the text, as the semantically crucial root is forced into a small font size to accommodate the massive waste of space afforded to the inflections. Nouns would be similar, though the problem is less severe.
Image

User avatar
sangi39
moderator
moderator
Posts: 3141
Joined: 12 Aug 2010 00:53
Location: North Yorkshire, UK

Re: Script preference

Post by sangi39 » 18 Sep 2018 01:47

CarsonDaConlanger wrote:
18 Sep 2018 00:35
I often see reference to the fact that Hànzì (Or semanto-phonetic scripts in general) wouldn't work well for synthetic languages. Why not? Wouldn't it be trivial to just have an accusative symbol etc?
Zompist wrote about using a logographic script for writing English here which covers inflections very briefly. As far as I understand it, he suggests using a character with a similar pronunciation as the inflection (with some modification to show that it is in face an inflectional morpheme being represented). So for English you might use characters for "is", "eat", and "ink" for the plural, the past tense and the present participle.

IIRC, this is roughly what happened in the history of Japanese writing, where sometimes characters were used for there sounds, regardless of their actual meaning (Mongolian was transcribed almost exclusively in this manner at one point from what I can remember).

There's also the option to use characters for their meanings, but again modify them to show they're inflections, e.g. using a character meaning "many" to indicate the plural or "before" to indicate the past tense.

And then there's the Ancient Egyptian, Mayan, Sumerian and Akkadian languages which all employed semanto-phonetic scripts and represented inflection.
You can tell the same lie a thousand times,
But it never gets any more true,
So close your eyes once more and once more believe
That they all still believe in you.
Just one time.

User avatar
Omzinesý
runic
runic
Posts: 2452
Joined: 27 Aug 2010 07:17
Location: nowhere [naʊhɪɚ]

Re: Script preference

Post by Omzinesý » 18 Sep 2018 09:57

I think somebody used logographs for their consonant-radical conlang. I dont know how far it went.
I also tried to copy the idea but my consonant-radical conlang had very few inflection. I used just phonetic vowel singns as inflections, marking all sandhi etc.

But in principle, Semitic languages have quite few - too many to learn though :) - consonant-radical roots. So logographs could do quite well for them. Of course the grammatical signs should be developed for every language specifically. Standard Arabic, for example, has ten verb inflections, all of which should have a marker. Aspect can be marked quite easily by positioning the person marker either before or after the root.

Arabic: tuDaRRiSiin 'You sg.F teach', Pattern II from root DRS 'learn' Would have three signs: [SG2.F][DRS][pattern II] Passive and subjunctive would need fourth and fifth signs if they appear, though they are rare and the adyad Arabic now uses also marks them incompletely. If pattern I is zero-marked, only two signs would be needed for taDRiSiin 'you learn' [sg2.F][DRS]

Not impossible at all! Noun derivations are much more unproductive though.

User avatar
lsd
roman
roman
Posts: 896
Joined: 11 Mar 2011 21:11
Contact:

Re: Script preference

Post by lsd » 18 Sep 2018 19:07

Even if synthetic languages are not favorable to invention of writing ...all historic logographies were re-used for transcription of synthetic languages...

I use pure logography for my olygosynthetic conlang, some signs have inflection meanings, but they are not empty words...

Salmoneus
MVP
MVP
Posts: 1538
Joined: 19 Sep 2011 18:37

Re: Script preference

Post by Salmoneus » 19 Sep 2018 12:06

Omzinesý wrote:
18 Sep 2018 09:57
I think somebody used logographs for their consonant-radical conlang. I dont know how far it went.
I also tried to copy the idea but my consonant-radical conlang had very few inflection. I used just phonetic vowel singns as inflections, marking all sandhi etc.

But in principle, Semitic languages have quite few - too many to learn though :) - consonant-radical roots. So logographs could do quite well for them. Of course the grammatical signs should be developed for every language specifically. Standard Arabic, for example, has ten verb inflections, all of which should have a marker. Aspect can be marked quite easily by positioning the person marker either before or after the root.

Arabic: tuDaRRiSiin 'You sg.F teach', Pattern II from root DRS 'learn' Would have three signs: [SG2.F][DRS][pattern II] Passive and subjunctive would need fourth and fifth signs if they appear, though they are rare and the adyad Arabic now uses also marks them incompletely. If pattern I is zero-marked, only two signs would be needed for taDRiSiin 'you learn' [sg2.F][DRS]

Not impossible at all! Noun derivations are much more unproductive though.
A great idea logically - but the problem would be development. The grammatical signs wouldn't be able to pattern themselves after an existing word phonetically, and many don't lend themselves to obvious, drawable semantics either. I suppose you could do things like 'past' coming from 'yesterday' and the like, but it doesn't feel very natural - writing usually at least starts out a representation of spoken speech, rather than a logical analysis of grammar.

You could have a Koreanesque "linguistically-minded king invents a new script based on that of a neighbouring culture" or the like, though. Say, a semiticesque language near a 'China', whose scribes use 'Chinese' for their writing, until their king decides they need their own writing now and creates a new script modeled on 'Chinese' but suited to their grammar.

Actually, that would solve the problem quite directly: if 'Chinese' isn't triconsonantal, they could use 'Chinese' words for their vowel patterns markers.

User avatar
lsd
roman
roman
Posts: 896
Joined: 11 Mar 2011 21:11
Contact:

Re: Script preference

Post by lsd » 19 Sep 2018 18:31

Salmoneus wrote:
19 Sep 2018 12:06
writing usually at least starts out a representation of spoken speech
not so on, with pictographic beginning, the writing started as a memory aid of the sense more than of the sound ... it is not for nothing that the beginning was ideographic ... it is only much later that in second-hand use, the phonemes were cut out...
We must believe that the meaning is simpler than the sound to perceive, no offense to our alphabetic habits, and especially to our audiovisual drift...
back to the origins, let's use a pure ideography...

User avatar
sangi39
moderator
moderator
Posts: 3141
Joined: 12 Aug 2010 00:53
Location: North Yorkshire, UK

Re: Script preference

Post by sangi39 » 19 Sep 2018 19:30

lsd wrote:
19 Sep 2018 18:31
Salmoneus wrote:
19 Sep 2018 12:06
writing usually at least starts out a representation of spoken speech
not so on, with pictographic beginning, the writing started as a memory aid of the sense more than of the sound ... it is not for nothing that the beginning was ideographic ... it is only much later that in second-hand use, the phonemes were cut out...
We must believe that the meaning is simpler than the sound to perceive, no offense to our alphabetic habits, and especially to our audiovisual drift...
back to the origins, let's use a pure ideography...
I'm not sure, but I think you may have misunderstood Sal's point. When he said "spoken speech", he's not necessarily talking about the individual sounds of the language, but more "language as is" as opposed to some grammatically analysed abstract, e.g. writing "amo" as LOVE-"o" (where LOVE is written with an ideogram, and the "o" is a phonetic reading of some other character), rather than LOVE-1st.person-singular-present-indicative.
You can tell the same lie a thousand times,
But it never gets any more true,
So close your eyes once more and once more believe
That they all still believe in you.
Just one time.

User avatar
lsd
roman
roman
Posts: 896
Joined: 11 Mar 2011 21:11
Contact:

Re: Script preference

Post by lsd » 20 Sep 2018 21:36

I think ideography don't really bother with 'language as is' but is a second language... where semantic is the first purpose...
the adding of sound is just a weaker and less elegant manner of increasing number of signs or importing foreign ideas (but also a way to play on words...)
'the language as is' is a goal for re-use of a prior foreign writing that needs sound or grammar adding (in the same time: most re-uses were syllabic transcriptions...)

With a constructed language, the possibility of maintaining pure ideography is a target in se, and a way to obtain a first version of the con-lang... on this base we can add, in a second pass, a spoken speech (and even a synthetic one)... which can be presented as the language as is...

these influences of writing on speech, draw amazing resemblances with ideographic writing nat-cultures... but any language was first a con-language

User avatar
GrandPiano
runic
runic
Posts: 2673
Joined: 11 Jan 2015 23:22
Location: Ohio, USA

Re: Script preference

Post by GrandPiano » 20 Sep 2018 23:20

Don’t forget that even a logographic writing system with no phonetic indicators still follows the syntax and semantics of the spoken language it represents.
:eng: - Native
:chn: - B2
:esp: - A2
:jpn: - A2

Salmoneus
MVP
MVP
Posts: 1538
Joined: 19 Sep 2011 18:37

Re: Script preference

Post by Salmoneus » 21 Sep 2018 11:43

It should be pointed out for those who are not aware, no ideographies actually exist, nor is there any evidence that they might ever have existed; they are purely an ideological hypothesis no longer generally believed to have a basis in reality. They were invented when people didn't understand logographies.

What do exist are mnemonic pictographs - a series of pictures of things, used as an aid to memory in the recitation of long texts, usually in ritual contexts. But as these can only reliably be read by people who already know the text (as they are ambiguous and generally are used very differently by different 'writers'), they cannot be considered ideographic. The Naxi sometimes use pictographs with the trickier bits spelled out in writing - but both their writing and their pictographs postdate their exposure to writing. Indeed, AIUI, mnemonic pictographs are generally an imitation of writing, rather than a predecessor of it. Writing instead probably generates from labels and markers, as in tax records and contracts and the like, only later assembled to represent speech.

User avatar
lsd
roman
roman
Posts: 896
Joined: 11 Mar 2011 21:11
Contact:

Re: Script preference

Post by lsd » 21 Sep 2018 16:37

Salmoneus wrote:
21 Sep 2018 11:43
It should be pointed out for those who are not aware, no ideographies actually exist
obviously, you speak here of pure natural ideographies, even if they exist in natural environments in a partial way, for reduced lexicons and as a writing mechanism... but in laboratories their existence is obvious ...
Salmoneus wrote:
21 Sep 2018 11:43
What do exist are mnemonic pictographs - a series of pictures of things, used as an aid to memory in the recitation of long texts, usually in ritual contexts. But as these can only reliably be read by people who already know the text (as they are ambiguous and generally are used very differently by different 'writers'), they cannot be considered ideographic.
but they are, however, so primitive...
in any case, any ideography is by nature read in different ways by each reader...
moreover, all logographic scripts had been adopted, for a time, by neighboring peoples and used to code their language sometimes very distant...
(even if the alphabetic writings also are pronounced or understood in a differentiated ways, according to the distance of their readers in the space or time...)
Salmoneus wrote:
21 Sep 2018 11:43
AIUI, mnemonic pictographs are generally an imitation of writing, rather than a predecessor of it.
this is contrary to the oldest listed epigraphic documents, and to the first remains of the recorded writings ...
Even if the origin is forever lost, many paths seem leading to it...

in conlangology, especially in a priori languages, the passage from pasigraphy to spoken language, by pasilaly is attested...

Any way some deny the name of writing to ideography because of the poor link to spoken language, as the name of language for a priori languages, because of the lack of double articulation... that stops all discussions...

User avatar
k1234567890y
runic
runic
Posts: 2967
Joined: 04 Jan 2014 04:47
Contact:

Re: Script preference

Post by k1234567890y » 21 Sep 2018 19:44

Hanzi alone may not work well for synthetic languages, but they can work on synthetic languages with a set of additional purely-phonemic symbols. Two examples are Japanese and pre-1950 Korean.

Outside of Chinese, there are synthetic languages using a logographic writing system, such as Sumerian(cuneiform), Egyptian(Hieroglyphs) and Mayan languages.

Besides, there's a possibility that Old Chinese was actually synthetic and some if not most of the inflections that were not important for distinguishing meanings were not expressed in writing.
私のアツい人工言語活動!言カツ!始まります!!

User avatar
GrandPiano
runic
runic
Posts: 2673
Joined: 11 Jan 2015 23:22
Location: Ohio, USA

Re: Script preference

Post by GrandPiano » 22 Sep 2018 02:49

lsd wrote:
21 Sep 2018 16:37
but they are, however, so primitive...
in any case, any ideography is by nature read in different ways by each reader...
moreover, all logographic scripts had been adopted, for a time, by neighboring peoples and used to code their language sometimes very distant...
(even if the alphabetic writings also are pronounced or understood in a differentiated ways, according to the distance of their readers in the space or time...)
This is off topic, but lsd, I think you would sound more credible and people would find it easier to take you seriously if you didn't end every sentence with an ellipsis.
:eng: - Native
:chn: - B2
:esp: - A2
:jpn: - A2

User avatar
lsd
roman
roman
Posts: 896
Joined: 11 Mar 2011 21:11
Contact:

Re: Script preference

Post by lsd » 22 Sep 2018 17:14

GrandPiano wrote:
22 Sep 2018 02:49
This is off topic
I did not think, when it came to choice of writing according to the type of language that the fact that a pure ideography does not describe a single spoken language, was without interest, imha...
people would find it easier to take you seriously if you didn't end every sentence with an ellipsis.
I do not believe in full stop for conlanging... (ok I'm not serious...)

User avatar
GrandPiano
runic
runic
Posts: 2673
Joined: 11 Jan 2015 23:22
Location: Ohio, USA

Re: Script preference

Post by GrandPiano » 22 Sep 2018 18:15

lsd wrote:
22 Sep 2018 17:14
GrandPiano wrote:
22 Sep 2018 02:49
This is off topic
I did not think, when it came to choice of writing according to the type of language that the fact that a pure ideography does not describe a single spoken language, was without interest, imha...
I meant what I was saying was off topic.
lsd wrote:
22 Sep 2018 17:14
people would find it easier to take you seriously if you didn't end every sentence with an ellipsis.
I do not believe in full stop for conlanging... (ok I'm not serious...)
A joke is fine, but I was kind of hoping for an actual response as well.
:eng: - Native
:chn: - B2
:esp: - A2
:jpn: - A2

clawgrip
MVP
MVP
Posts: 2399
Joined: 24 Jun 2012 06:33
Location: Tokyo

Re: Script preference

Post by clawgrip » 18 Oct 2018 02:45

k1234567890y wrote:
21 Sep 2018 19:44
Hanzi alone may not work well for synthetic languages, but they can work on synthetic languages with a set of additional purely-phonemic symbols. Two examples are Japanese and pre-1950 Korean.

Outside of Chinese, there are synthetic languages using a logographic writing system, such as Sumerian(cuneiform), Egyptian(Hieroglyphs) and Mayan languages.

Besides, there's a possibility that Old Chinese was actually synthetic and some if not most of the inflections that were not important for distinguishing meanings were not expressed in writing.
I've mentioned it before, but pre-modern Japanese could often get by with very few purely phonetic symbols, using Chinese characters for conjunctions, inflections, etc., or just entirely dropping certain inflections that are easily guessed, e.g.

御趣意相弁小前末々迄茂不取締之儀無之様可致事

Breaking this down:

御趣意相 1 小前末々 迄茂2 不取締 34 5 6 7 8

御趣意相 わきまえ1 小前末々 までも2 不取締 3これ5 無き4 よう6 致す8 べき7
  • までも made mo is written as 迄茂;
  • 之 is used both for の no and これ kore;
  • よう yō is written in kanji as 様;
  • easily guessed inflections are dropped on 無き naki (→無), 致す itasu (→致), and わきまえ wakimae, which also has its stem written in kanji (→弁);
  • the verb inflection べく -beku is written as 可.
For inflections, it's just a matter of adding a character for each morpheme in the chain. You don't need to represent all the little details of the phonetics for combining the morphemes properly, because native speakers already know these things, e.g.

取 toru, 不取 torazu, 取候 torimasu, 不取候 torimasen, 不被取候 toraremasen, etc. (Japanese has the additional nonsense of putting them out of spoken order, but it still works). The changing of the thematic vowels is not represented, nor is the change of the form of the negative morpheme indicated.

Some characters are being used phonetically, and depending on the writer and the text, more hiragana will appear, but still, it's definitely possible to write a synthetic language with mostly logograms.

User avatar
k1234567890y
runic
runic
Posts: 2967
Joined: 04 Jan 2014 04:47
Contact:

Re: Script preference

Post by k1234567890y » 18 Oct 2018 08:48

clawgrip wrote:
18 Oct 2018 02:45
k1234567890y wrote:
21 Sep 2018 19:44
Hanzi alone may not work well for synthetic languages, but they can work on synthetic languages with a set of additional purely-phonemic symbols. Two examples are Japanese and pre-1950 Korean.

Outside of Chinese, there are synthetic languages using a logographic writing system, such as Sumerian(cuneiform), Egyptian(Hieroglyphs) and Mayan languages.

Besides, there's a possibility that Old Chinese was actually synthetic and some if not most of the inflections that were not important for distinguishing meanings were not expressed in writing.
I've mentioned it before, but pre-modern Japanese could often get by with very few purely phonetic symbols, using Chinese characters for conjunctions, inflections, etc., or just entirely dropping certain inflections that are easily guessed, e.g.

御趣意相弁小前末々迄茂不取締之儀無之様可致事

Breaking this down:

御趣意相 1 小前末々 迄茂2 不取締 34 5 6 7 8

御趣意相 わきまえ1 小前末々 までも2 不取締 3これ5 無き4 よう6 致す8 べき7
  • までも made mo is written as 迄茂;
  • 之 is used both for の no and これ kore;
  • よう yō is written in kanji as 様;
  • easily guessed inflections are dropped on 無き naki (→無), 致す itasu (→致), and わきまえ wakimae, which also has its stem written in kanji (→弁);
  • the verb inflection べく -beku is written as 可.
For inflections, it's just a matter of adding a character for each morpheme in the chain. You don't need to represent all the little details of the phonetics for combining the morphemes properly, because native speakers already know these things, e.g.

取 toru, 不取 torazu, 取候 torimasu, 不取候 torimasen, 不被取候 toraremasen, etc. (Japanese has the additional nonsense of putting them out of spoken order, but it still works). The changing of the thematic vowels is not represented, nor is the change of the form of the negative morpheme indicated.

Some characters are being used phonetically, and depending on the writer and the text, more hiragana will appear, but still, it's definitely possible to write a synthetic language with mostly logograms.
nice (: thanks for the examples

btw, I have a feeling that if English were using a Chinese-like logographic(that is, logographic with little if any purely phonemic symbols), there's a good chance that the English 3rd present singular -s would be omitted and people would not know its existence without actually speaking English
私のアツい人工言語活動!言カツ!始まります!!

clawgrip
MVP
MVP
Posts: 2399
Joined: 24 Jun 2012 06:33
Location: Tokyo

Re: Script preference

Post by clawgrip » 18 Oct 2018 23:22

That would likely be the case. If all personal endings were distinct, like Spanish, they would probably be indicated, but with only the s, it's a fully predictable inflection (as long as the plurality of the subject is indicated. And because of the importance of plurality in English, I suspect that plurality might be indicated even on words that don't change form.

Post Reply