Linguistic Miscellany Thread

[Nachtswalbe]

Are there abugidas with specialized symbols for syllable-final consonants?

[zompist]

Are there abugidas with specialized symbols for syllable-final consonants?

Well, there's one and perhaps two in Devanagari. The visarga, which looks like a colon (:), is a separate symbol used for final h. There's also the anusvara, transliterated ṃ-- it's just a dot above the previous letter, so you could consider it a diacritic.

That there's not more can be attributed to historical luck: Prakrit, for which the ancestral Brahmi script was devised, had no final consonants except -m. Once they appeared, a vowel-canceling diacritic was introduced rather than new letters.

[Richard W]

Are there abugidas with specialized symbols for syllable-final consonants?

Yes, though there aren't large repertoires of specialised syllable final consonants. The alternative to the vowel killing diacritic was to write a consonant small, and such consonants could be indistinguishable from the 'normal' subscript consonants. There are also two basic cases that don't make major use of the vowel killer - scripts that are used for Sanskrit or Pali, and those that aren't.

In scripts that aren't used for Sanskrit or Pali, one frequently finds that one is dealing with a CRVC syllable structure in the main language, possibly with tone marks, both orthographically and finally. In Unicode, the initial consonant is the plain one, the 'R' one is named as 'MEDIAL' and the final one may be named as 'FINAL'. Cham, Limbu and Lepcha seem to be the typical systems of this type.

For the scripts supporting Sanskrit or Pali, the major examples were Khmer and the Lanna (= Tham) script. Cambodian Khmer went over to ambiguity, like Thai, just over a century ago, and I'm not sure of the state of Southern Thai Khmer. I'm not sure to which of the two types I should assign late 'Tai Noi' or 'Lao Buhan' if that's what it is - I've been intermittently skimming through a pair of mid-20th century textbooks on the Tham script written in Lao in this script. It's not encoded in Unicode yet.

Now this scheme may seem horribly ambiguous, but knowing the phonotactics of the language will usually enable one to deduce the correct pronunciation. There may also be cues in the precise positioning of other marks in the cluster.

Now, the Tai Tham script will write a few of the final consonants (K, NG, R, W) above the initial consonant if there is no room below, and indeed the superscript final K may be written above without any such need. Because Unicode needs to encode the position, one might see them as specialised forms, but only the K has a specialised form, which also serves as a vowel mark (for A in native words) and has graphically merged with a tone mark in much of the region. Tai Tham also has two forms of subscript R and subscript L, which one might see as being specialised to the onset and coda. However, some people appear to distinguish subscript L, in the vernacular, upon whether it is sounded or not.

As Northern Thai tends to preserve Sanskrit and Pali consonants in loanwords, and to drop the stem final -a, and Sanskrit or Pali consonant can be subscripted to mark it as a final consonant even though phonetically there are only 9 final consonants, one of which is implicit or written as visarga. When the vowel dropping leaves a final cluster, a diacritic is used to indicate that the final vowel has been silenced. It can also be used to silence a final I or U, though that is rare, and is sometimes used on a single consonant (most commonly L when it cannot be made subscript) and also on NG rather than superscripting it - superscript NG seems to have dropped out of favour in Northern Thailand and Burma. For single consonants, it is commoner to leave uncertainty as to whether the final vowel has been dropped.

Now, when R or NG starts an intervocalic consonant cluster, some styles write it above the next consonant (the traditional place) but with the next consonant at the level of most syllable initial consonants. Phonetically, it is the final consonant of the previous symbol, and this has occasionally become a specialised final consonant associated with the previous syllable - layar for R in Javanese and closely related scripts, and mai kang lai for NG in the Tai Khuen variant of the Lanna script (and also some varieties in other places).

[Richard W]

See also the Fremdzeichen of the Tocharian script.

[Nachtswalbe]

Are there logo-abugidas which combine characters for meaning and abugidas and how could they evolve?

[Raphael]

Serious question: Do people whose first language uses a different writing system than the Latin Alphabet find it weird that in languages that do use the Latin Alphabet, the less commonly used letter forms (uppercase) are usually treated as the standard, while the more commonly used letter forms (lowercase) are usually treated as the modification of the standard?

And, are there any languages that do use the Latin Alphabet whose speakers usually think of the lowercase letters as the standard letter shape?

[Richard W]

Are there logo-abugidas which combine characters for meaning and abugidas and how could they evolve?

Evolution may have gone pretty much the same way as alphabet-based logograms, such as Tironian ampersand, get used in Old Irish and 15th century German, along with the examples of English '3rd', 'l8t' and of course Akkadograms. There are examples with Northern Thai ᪓᩠ᨴ <THAM DIGIT THREE, SAKOT, LOW TA> '3 times' (the vowel of the word for 'time' has been shed), ᩓ᩠ᩅ <LAE, SAKOT, WA> 'already' (if you accept LAE as a logogram for 'and') and ᩁᩦ 'i.e.' read /anwaː/ but written as though /hiː/, an equivalent of an Akkadogram (I suppose a 'Khmerogram'), which is not so different from £sd for 'pounds, shillings and pence'.

[Richard W]

Do people whose first language uses a different writing system than the Latin Alphabet find it weird that in languages that do use the Latin Alphabet, the less commonly used letter forms (uppercase) are usually treated as the standard, while the more commonly used letter forms (lowercase) are usually treated as the modification of the standard?

I don't know, but a similar case is their treating the independent vowels as the primary forms. I suppose their custom is skewed by /a/ usually having the null glyph as its dependent form.

[bradrn]

And, are there any languages that do use the Latin Alphabet whose speakers usually think of the lowercase letters as the standard letter shape?

The Africa Alphabet was meant to be all-lowercase, but I don’t believe any modern African language continues this.

[Otto Kretschmer]

Proto Bantu is dated at 2000 BC. How come that all Bantu languages have nearly the same phonotactics? The time depth that separates Zulu from Lingala is similar to the one separating Italian from German

[Creyeditor]

What exactly do you mean by having the same phonotactics?

[Otto Kretschmer]

What exactly do you mean by having the same phonotactics?

Very similar syllabe structure and strong preference for open syllabes, to the degree that for a lay person they all look like the same language.

[Rounin Ryuuji]

Probably just an areal feature.

[zompist]

What exactly do you mean by having the same phonotactics?

Very similar syllabe structure and strong preference for open syllabes, to the degree that for a lay person they all look like the same language.

Some quick observations...

1. The timing of the Bantu expansion is highly debatable. Some Googling suggests dates anywhere from 5600 to 1500 years ago. There may have been several waves of expansion, and especially movements (or rearrangements within Bantu) due to the Iron Age, from about 500 BCE to 400 CE. For comparison, look at a map of Indo-European now and 2000 years ago: it may cover about the same area in Europe, but some subfamilies have greatly spread and some have greatly shrunk.

2. I wouldn't say Bantu has "the same phonotactics" all over. The northern languages have tone, the southernmost have clicks. You can look at my numbers page to see that plenty has been happening phonetically. Vague impressions of similarity by outsiders aren't worth much: as with generalizations about appearance, a lot depends on what you as an observer were brought up to notice.

3. As ever, be aware of areal effects and Sprachbunds. I don't know if these are considered to apply to Bantu, but they are always a possible explanation of similarities in language besides genetic descent.

[Otto Kretschmer]

Are there any texts on the web in any non-Bantu Bantoid language?

[Otto Kretschmer]

What exactly do you mean by having the same phonotactics?

Very similar syllabe structure and strong preference for open syllabes, to the degree that for a lay person they all look like the same language.

Some quick observations...

1. The timing of the Bantu expansion is highly debatable. Some Googling suggests dates anywhere from 5600 to 1500 years ago. There may have been several waves of expansion, and especially movements (or rearrangements within Bantu) due to the Iron Age, from about 500 BCE to 400 CE. For comparison, look at a map of Indo-European now and 2000 years ago: it may cover about the same area in Europe, but some subfamilies have greatly spread and some have greatly shrunk.

2. I wouldn't say Bantu has "the same phonotactics" all over. The northern languages have tone, the southernmost have clicks. You can look at my numbers page to see that plenty has been happening phonetically. Vague impressions of similarity by outsiders aren't worth much: as with generalizations about appearance, a lot depends on what you as an observer were brought up to notice.

3. As ever, be aware of areal effects and Sprachbunds. I don't know if these are considered to apply to Bantu, but they are always a possible explanation of similarities in language besides genetic descent.

As far as I know, southern Bantu languages also have tones. The only major Bantu language without tones is Swahili

[zompist]

As far as I know, southern Bantu languages also have tones. The only major Bantu language without tones is Swahili

Yes, my mistake— I was looking at transcriptions, and most of them don't mark tone.

(Though if I'm reading this paper correctly, it's mostly pitch-accent systems.)

[bradrn]

Many languages all over the word use auxiliary verb constructions. However, the Germanic languages seem to be unique in allowing multiple stacked auxiliaries (I will have been waiting and so on) — or at least, I can’t easily find other examples. Are there any non-IE languages which allow auxiliary stacking? And, since I strongly suspect this will turn out to be uninterestingly true: are there any languages outside the broader Western European language area (IE, Uralic, Kartvelian, Semitic etc.) which allow this? I’d be particularly interested in any examples from Altaic, Paleosiberian or Papuan languages.

[Creyeditor]

Papua Indonesian stacks at least two auxilliaries:

Sa mo dapat pukul.
I FUT PASS beat
I will get beaten up.

Maybe you can do more than that, idk. mo also means want and dapat also means get. Just out of curiousity: which languages don't allow any stacking here?

[Zju]

It depends on what exactly your definition of 'auxiliary' is - is there a single, language independant one? My hunch is that all languages have a way to express notions like 'might want to' and 'will have to be' (passive) - and for that you'll have to stack affixes if not words.