Acronyms in non-suffixing languages

[alice]

Maybe it's a continuum?

[bradrn]

Maybe it's a continuum?

Most things seem to be.

[WeepingElf]

the more i learn about them, the more i feel that the divisions between alphabet/abjad/abugida/syllabary are too blurry to stand up to any real scrutiny

I’d say that alphabet/abjad is quite blurry, as is abugida/syllabary, but the difference between those two supercategories is arguably quite useful — essentially whether glyphs are structured in terms of phonemes (alphabet/abjad) or syllables (abugida/syllabary). There are edge cases between these, but not many.

However, an abugida is not far from an abjad with obligatory vowel marking (in fact, both the Ethiopic and the Indian abugidas evolved from that), so even that boundary is blurry.

[bradrn]

the more i learn about them, the more i feel that the divisions between alphabet/abjad/abugida/syllabary are too blurry to stand up to any real scrutiny

I’d say that alphabet/abjad is quite blurry, as is abugida/syllabary, but the difference between those two supercategories is arguably quite useful — essentially whether glyphs are structured in terms of phonemes (alphabet/abjad) or syllables (abugida/syllabary). There are edge cases between these, but not many.

However, an abugida is not far from an abjad with obligatory vowel marking (in fact, both the Ethiopic and the Indian abugidas evolved from that), so even that boundary is blurry.

An abjad with obligatory vowel marking (as in Kurdish and Uyghur) isn’t quite an abugida, but rather something lying midway between an alphabet and an abugida. The key step to get from an abjad to an abugida is to associate consonantal letters with an inherent vowel, which indeed seems to be what happened in Ethiopic and Brahmi.

[WeepingElf]

I’d say that alphabet/abjad is quite blurry, as is abugida/syllabary, but the difference between those two supercategories is arguably quite useful — essentially whether glyphs are structured in terms of phonemes (alphabet/abjad) or syllables (abugida/syllabary). There are edge cases between these, but not many.

However, an abugida is not far from an abjad with obligatory vowel marking (in fact, both the Ethiopic and the Indian abugidas evolved from that), so even that boundary is blurry.

An abjad with obligatory vowel marking (as in Kurdish and Uyghur) isn’t quite an abugida, but rather something lying midway between an alphabet and an abugida. The key step to get from an abjad to an abugida is to associate consonantal letters with an inherent vowel, which indeed seems to be what happened in Ethiopic and Brahmi.

Of course. I didn't say they were the same, only that they were not far from each other, and that abugidas evolved from abjads with vowel marking.

[bradrn]

abugidas evolved from abjads with vowel marking.

Now this claim I hadn’t seen before. In fact, it seems obviously false to me, in that neither progenitor script had diacritical vowel marking in the first place. Aramaic only got it around the time of the Masoretes (in the form of Hebrew), while the South Arabian script from which Ethiopic is descended never had vowel diacritics at all. Both had matres lectiones, but that’s quite different to the style of vowel marking in current abjads. So I’m not sure how Brahmi or Ethiopic could possibly have been inspired by ‘abjads with vowel marking’ — rather, I suspect they were separate innovations.

[WeepingElf]

Thank you for correcting me. It was merely my own idea. But apparently, the Ethipians (despite speaking a Semitic language) and the Indians (who didn't speak a Semitic language) found the abjad to be deficient and wanted to indicate the vowels in writing, so they invented (independently from each other and the later inventions of vowel diacritics in Hebrew and Arabic) vowel marks to add to the consonant letters.

[bradrn]

Thank you for correcting me. It was merely my own idea. But apparently, the Ethipians (despite speaking a Semitic language) and the Indians (who didn't speak a Semitic language) found the abjad to be deficient and wanted to indicate the vowels in writing, so they invented (independently from each other and the later inventions of vowel diacritics in Hebrew and Arabic) vowel marks to add to the consonant letters.

Indeed, this is my understanding of the evolution.

[Nortaneous]

the more i learn about them, the more i feel that the divisions between alphabet/abjad/abugida/syllabary are too blurry to stand up to any real scrutiny

I’d say that alphabet/abjad is quite blurry, as is abugida/syllabary, but the difference between those two supercategories is arguably quite useful — essentially whether glyphs are structured in terms of phonemes (alphabet/abjad) or syllables (abugida/syllabary). There are edge cases between these, but not many.

Surprisingly few, really - the idea of compressing an alphabet by creating characters for common syllables or sequences comes up sometimes, e.g. in Latin (PDF), and technically every consonant letter of the Deseret alphabet has an inherent vowel that may inhere either before or after the consonant, so you could write "being" as <𐐺𐑍> with the two letters /biː/ and /ɪŋ/. But generally alphabets are alphabets and don't do that, except for words like "b&" and "v&".

[Richard W]

The key step to get from an abjad to an abugida is to associate consonantal letters with an inherent vowel, which indeed seems to be what happened in Ethiopic and Brahmi.

More precisely, they marked exceptional vowels. Brahmi and Kharoshthi at least were developed for languages in which /a/ is exceptionally common. Moreover, vowel length marking is a later development in Kharoshthi, and is a systematic modification in Brahmi. /aː/ was the second commonest vowel.

[Richard W]

Thank you for correcting me. It was merely my own idea. But apparently, the Ethipians (despite speaking a Semitic language) and the Indians (who didn't speak a Semitic language) found the abjad to be deficient and wanted to indicate the vowels in writing, so they invented (independently from each other and the later inventions of vowel diacritics in Hebrew and Arabic) vowel marks to add to the consonant letters.

Is independence certain? Meroitic, which is more alphabet-like in terms of glyph status, and Brahmi/Kharoshthi arose at the same time and ideas could have spread. Meroitic seems to have still been around when the Ethiopic system was developed. Now, the Meroitic system was not so dissimilar to Persian cuneiform, especially if you remove redundant vowel marking from the latter. Persian cuneiform was at the very least inspired by Mesopotamian cuneiform. Apparently the Assyrian form of the latter had lost some of the vowel distinction in CV signs, encouraging plene writing in which the vowel made a formerly redundant appearance. So the origin of abugidas might be the cuneiform CV-V-VC structure.

[bradrn]

The key step to get from an abjad to an abugida is to associate consonantal letters with an inherent vowel, which indeed seems to be what happened in Ethiopic and Brahmi.

More precisely, they marked exceptional vowels. Brahmi and Kharoshthi at least were developed for languages in which /a/ is exceptionally common. Moreover, vowel length marking is a later development in Kharoshthi, and is a systematic modification in Brahmi. /aː/ was the second commonest vowel.

Now that’s interesting — I didn’t realise /a/ was quite so common.

Thank you for correcting me. It was merely my own idea. But apparently, the Ethipians (despite speaking a Semitic language) and the Indians (who didn't speak a Semitic language) found the abjad to be deficient and wanted to indicate the vowels in writing, so they invented (independently from each other and the later inventions of vowel diacritics in Hebrew and Arabic) vowel marks to add to the consonant letters.

Is independence certain? Meroitic, which is more alphabet-like in terms of glyph status, and Brahmi/Kharoshthi arose at the same time and ideas could have spread. Meroitic seems to have still been around when the Ethiopic system was developed. Now, the Meroitic system was not so dissimilar to Persian cuneiform, especially if you remove redundant vowel marking from the latter. Persian cuneiform was at the very least inspired by Mesopotamian cuneiform. Apparently the Assyrian form of the latter had lost some of the vowel distinction in CV signs, encouraging plene writing in which the vowel made a formerly redundant appearance. So the origin of abugidas might be the cuneiform CV-V-VC structure.

I don’t understand this, though. In what way is Meroitic similar to Persian cuneiform? And what do you mean by ‘the cuneiform CV-V-VC structure’?

[Richard W]

I don’t understand this, though. In what way is Meroitic similar to Persian cuneiform? And what do you mean by ‘the cuneiform CV-V-VC structure’?

If we ignore words with Old Persian short diphthongs or Meroitic '/e/', whatever it was, the process to convert the language expressed alphabetically to the respective script goes:

1. For Old Persian, collapse the pairs /i/ and/i:/ and /u/ and /u:/. (We don't know about Meroitic vowel length - it probably had it.)
2. Strike out coda consonants that won't be represented.
3. Strike out A if not word-initial, and change all old Persian /a:/ to A. At this point, Old Persian has non-initial vowels A, I and U, and Meroitic has non-initial I and U, and we have an abugida.
4. For Old Persian only, change consonants immediately before I and U to special forms redundantly indicating the presence of that vowel. The majority of consonants don't change at all.
5. For cursive Meroitic, ligate (some?) sequences of consonants plus I.

(I've omitted a minor complication with Meroitic /to/ or whatever).

The general structure of the phonetic representation of a syllable in cuneiform was a character sequence CVC or (CV)-(V)-(VC), where having VC only implied the presence of a glottal stop or a morpheme boundary. Moreover, the 3 V's represent the 'same' vowel, counting U, Ú and Ù as being the same vowel. A syllable is written 'plene' if it has CV and V; normally being written plene indicated the presence of a long vowel. However, by late Assyrian, some of the CV contrasts had been lost, and so false plene was needed to show the quality of the vowel. Old Persian cuneiform was composed of (CV)-(V) units, where many of the CV units for the same consonant were the same, thus tending to (C)-(V).

Step 4 did not apply to Old Persian short diphthongs, so what may be transliterated da-i and di-i distinguish /dai/ on one hand from /di/ and /diː/ on the other, where <da> is the starting glyph for /d/. Because there was no distinction of <ta> and <ti>, the script cannot distinguish /tai/ and /ti/.

[bradrn]

I don’t understand this, though. In what way is Meroitic similar to Persian cuneiform? And what do you mean by ‘the cuneiform CV-V-VC structure’?

If we ignore words with Old Persian short diphthongs or Meroitic '/e/', whatever it was, the process to convert the language expressed alphabetically to the respective script goes:

1. For Old Persian, collapse the pairs /i/ and/i:/ and /u/ and /u:/. (We don't know about Meroitic vowel length - it probably had it.)
2. Strike out coda consonants that won't be represented.
3. Strike out A if not word-initial, and change all old Persian /a:/ to A. At this point, Old Persian has non-initial vowels A, I and U, and Meroitic has non-initial I and U, and we have an abugida.
4. For Old Persian only, change consonants immediately before I and U to special forms redundantly indicating the presence of that vowel. The majority of consonants don't change at all.
5. For cursive Meroitic, ligate (some?) sequences of consonants plus I.

This doesn’t feel very similar. The only common steps seem to be ‘get rid of coda consonants and the inherent vowel’.

(Though then again, you did merely say ‘not so dissimilar’, which I guess is fair enough.)

The general structure of the phonetic representation of a syllable in cuneiform was a character sequence CVC or (CV)-(V)-(VC), where having VC only implied the presence of a glottal stop or a morpheme boundary. Moreover, the 3 V's represent the 'same' vowel, counting U, Ú and Ù as being the same vowel. A syllable is written 'plene' if it has CV and V; normally being written plene indicated the presence of a long vowel. However, by late Assyrian, some of the CV contrasts had been lost, and so false plene was needed to show the quality of the vowel. Old Persian cuneiform was composed of (CV)-(V) units, where many of the CV units for the same consonant were the same, thus tending to (C)-(V).

So, if CV-V-VC is a CVC syllable, and CV-V is a CVː syllable, then how was a simple CV syllable expressed?

[Richard W]

The general structure of the phonetic representation of a syllable in cuneiform was a character sequence CVC or (CV)-(V)-(VC), where having VC only implied the presence of a glottal stop or a morpheme boundary. Moreover, the 3 V's represent the 'same' vowel, counting U, Ú and Ù as being the same vowel. A syllable is written 'plene' if it has CV and V; normally being written plene indicated the presence of a long vowel. However, by late Assyrian, some of the CV contrasts had been lost, and so false plene was needed to show the quality of the vowel. Old Persian cuneiform was composed of (CV)-(V) units, where many of the CV units for the same consonant were the same, thus tending to (C)-(V).

So, if CV-V-VC is a CVC syllable, and CV-V is a CVː syllable, then how was a simple CV syllable expressed?

The phonetic symbols in general cuneiform were CVC, CV, VC and V. CV-V-VC would have represented a closed syllable with a long vowel. I don't think CVC was ever used for a syllable with a long vowel, but I could be wrong. This is the ancestor of the system that inspired Persian cuneiform. Towards the end, certain CV syllables could not be distinguished using CV signs as the signs had merged, so the distinction in these cases had to be made by using false plene to spell out the vowel using CV-V. I haven't seen the details, just a reference to 'late Assyrian cuneiform'.

Persian cuneiform only had CV and V. In Persian cuneiform, a CV syllable would be written with a CV sign if the vowel was /a/ and as CV-V otherwise (except that a long diphthong would be written as CV-A-I or CV-A-U). A phonetic CVC syllable would be handled by dropping the coda, or by using CV for the final consonant. (A few Austronesian languages' Indic writing systems always drop the coda consonant!) For example, /imaːm/ 'this' was written i-ma-a-ma. One can't tell from the writing whether that word had /i/ or /iː/, or ended in /m/ or /ma/. (The Thai writing system has that second ambiguity, except word finally.) /imaivaː/ 'or these' was written i-ma-i-va-a, where writing <ma> rather than <mi> reveals the short diphthong, and is a point at which Persian cuneiform is not yet an abugida. One couldn't distinguish short diphthongs after /p/, for there was no distinction between <pa> and <pi> (or <pu>); Daniels and Bright's WWS transliterates them all as <p>.