Numbers from 1 to 10 updated

[zompist]

No, nothing here about Janko. I have a numbers page, did you know? More importantly, it's been updated. And here's the story behind the revamp.

The major change is that it now extensively uses Unicode, so all the wacky ways the sources wrote the numbers can be represented.

As the revamp story notes, I haven't actually been able to make corrections or additions for a couple of years. But now I can, and far easier than before. Which means this would be a good time to suggest corrections or additions! Check your favorite language families! Or even send me your favorite conlang's numbers. (Better to post them here than PM me.)

[hwhatting]

The following numbers are wrong for Kazakh:
3 is üš ( үш ), not uš; 4 is tört ( төрт ), not tort; 5 is altı ( алты ), not alti; 7 is žeti ( жеті ), not žetti, 9 is toɣız ( тоғыз ) not toɣiz.
If you want to cross-check, this is the online dictionary I normally use.

[clawgrip]

For Japanese, you should change the following:

ku → kyū
juu → jū

ku is a variant form of kyū that only occurs in limited contexts, and juu should have the macron because that's standard Hepburn Romanization.

[finlay]

"juu" is fine, but only if you're consistently never using the macron - but you already used it for "tō". i'd also find a way to add in that the 'sino-japanese' 4 and 7 are often replaced by "yon" and "nana" for most contexts.

[clawgrip]

Yes, the long vowels in the Japonic section are consistently marked with macrons except for juu, which is the main reason I went with it. Of course "ku" does occur even by itself, so it could remain, but it should be subordinate.

[Vijay]

It looks to me like there are a lot of problems with the Dravidian numbers. For starters, what's with all the cedillas?

EDIT: Also, it looks like you're saying Elamite is a branch of Indo-European.

EDIT2: Even the logo has mistakes. Five in Chinese (and Japanese) is 五, not 无, and in Hindi, it's पाँच, not पा.

[zompist]

The problems pointed out above should be corrected. Thank you!

(Except: Elamite still has the wrong classification level; I'll fix this next round.)

(Also, Photoshop for some reason can't display पाँच correctly, so I just left it out.)

[kodé]

It might just be my tired eyes, but it seems like a bunch of Penutian languages/language families are missing, e.g., Miwok and Yokuts. I've got a grammar of (Southern?) Sierra Miwok hiding around somewhere, and I have quite a few resources on Yokuts (my colleagues and I have done quite a bit of work on Yokuts languages). If you give me a few days, I should be able to provide a few more sets of numbers.

[zompist]

They're there— select Mesoamerican from the dropdown. But I'd be happy for you to check them over when you find them!

[Nortaneous]

https://mpi-lingweb.shh.mpg.de/numeral/

[kodé]

Ah, didn't realize they were under "Mesoamerica". I guess it makes sense that you wouldn't want to split up the Penutian macro-family, since there are members all along the Pacific coast. I guess dividing non-Na-Dene/Eskimo-Aleut indigenous Western Hemisphere languages intro geographical groups would be pretty tough.

As far as the numbers go, the Yokuts ones look fine. I'm most familiar with Valley Yokuts (which is moribund but not extinct; there are a few speakers left!), and it checks out. The transcription took a few seconds to adjust to, but that's kind of a general issue with trying to put up vocabulary lists from a ton of languages. For instance, I wasn't sure what you meant by <c> in <coopin>; however, you might as well leave it, since it has different values in different dialects (/s`/ in Chukchansi, /S/ in Yowlumne).

[zompist]

https://mpi-lingweb.shh.mpg.de/numeral/

Yes, that Eugene Chan's site; I've corresponded with him quite a bit and he provided me with his huge Austronesian database. My site was up first.

[kodé]

https://mpi-lingweb.shh.mpg.de/numeral/

Yes, that Eugene Chan's site; I've corresponded with him quite a bit and he provided me with his huge Austronesian database. My site was up first.

Any good database of Austronesian must be huge. That language family continues to impress me the more I look at it. IIRC one-fifth of languages are Austronesian, and I once measured that they are spread literally over half the globe (from Madagascar to Rapa Nui). In many ways, Austronesian is comparable to Indo-European (size, diversity, timeline, colonization); we should see more Austronesianlangs, to make up for all the IE-clones out there! /austronesianfanboy

[Salmoneus]

https://mpi-lingweb.shh.mpg.de/numeral/

Yes, that Eugene Chan's site; I've corresponded with him quite a bit and he provided me with his huge Austronesian database. My site was up first.

Any good database of Austronesian must be huge. That language family continues to impress me the more I look at it. IIRC one-fifth of languages are Austronesian, and I once measured that they are spread literally over half the globe (from Madagascar to Rapa Nui). In many ways, Austronesian is comparable to Indo-European (size, diversity, timeline, colonization); we should see more Austronesianlangs, to make up for all the IE-clones out there! /austronesianfanboy

Coincidentally, I've just been making some progress with Rawàng Ata verbs...
[not that Rawàng Ata is a real austronesianlang, as it's neither a posteriori nor a direct imitation, and is certainly non-austronesian in many ways. However, it's inspired more by austronesian than by anything else...]

[Whimemsz]

Per vlad, the Nahuatl numbers should be: cē, ōme, ēyi, nāhui, mācuilli, chicuacē, chicōme, chicuēyi, chiucnāhui, mahtlactli - with macrons added, and: ""chiconahui" is wrong. The word is /tʃikʷnaːwi/, which could be written "chicunahui" or "chicuhnahui" but not "chiconahui" -- that would be /tʃikonaːwi/. The current conventional spelling is "chiucnahui""; also: "changed "ma’tlactli" to "mahtlactli". Using an apostrophe for the saltillo is weird, especially when the spelling is otherwise traditional"

I have a number of updates for the Algic numbers (and a few other ones), but don't have time to give them tonight -- will try to post tomorrow.

[Jonlang]

Welsh also has the forms dwy, tair and pedair for two, three and four, as well as deng for ten. I don't know if you want to include these.

[hwhatting]

You have the right Cyrillic for Kazakh "7", but the transcription is wrong - Kazakh "ж" is /ʒ/, so it ought to be žeti, not *jeti.

[Vijay]

Thanks for updating the lists, zompist! That being said, here are some other issues I found:

1. The Persian word for 'one' is yek, not *yak. Yak is Tajik.
2. In Sanskrit, 'six' is ṣáṣ, not *s`as`, and 'eight' is aṣṭa, not *as`ṭá.
3. The word for 'six' in Hindi is usually pronounced che and spelled छह. छः is a spelling variant, and chai is a pronunciation variant that does not correspond to छः.
4. 'Two' in Gujarati is be, not *bɛ; 'three' is traṇ, not *trañ. Ñ is not a character used for transliterating Gujarati AFAIK.
5. In the interests of consistency (and following IAST), *<a> in all the Gujarati examples should probably be replaced with <ā>, *<ã> should probably be replaced with <ã̄>, and *<ə> should probably be replaced with <a>.
6. The Telugu numerals should surely be written out in Telugu script as well, considering that the Tamil, Kannada, and Malayalam numerals are, and Telugu has more speakers than any of those languages. They should be: ఒకటి ‎(okaṭi), రెండు ‎(reṇḍu), మూడు ‎(mūḍu), నాలుగు ‎(nālugu), ఐదు ‎(aidu), ఆరు ‎(āru), ఏడు ‎(ēḍu), ఎనిమిది ‎(enimidi), తొమ్మిది ‎(tommidi), and పది ‎(padi).
7. 'Two' and 'eight' in Kannada should be transcribed without macrons: eraḍu and eṇṭu, respectively. (The written forms of these two numbers are correct).
8. 'One' in Tamil should be written oṉṟu, not *oḷṟu, and oṉpatu 'nine' should not be capitalized.
9. ṇṭṭ is not a possible consonant cluster in Malayalam, so 'two' in Malayalam is raṇṭu, not *raṇṭṭu.
10. 'Five' in Malayalam should be añcu, especially since all the other ns in the transcribed list of Malayalam numbers are dental, whereas the nasal in that word is palatal.
11. 'Six' and 'seven' in Malayalam should be transcribed the same way as in Tamil (just like 'eight' is): āṟu and ēḻu. 'Nine' should also be oṉpatu.
12. The numerals in (Paraguayan) Guarani are usually spelled as follows: peteĩ, mokõi, mbohapy, irundy, po, poteĩ, pokõi, poapy, porundy, pa, i.e. with tildes instead of circumflexes and y instead of ï. Note that popa means 'fifty', not 'ten'.

Also, AFAICT, 'one' through 'three' in Brahui are asiṭ, irāṭ, and musiṭ. (Asi etc. are used in compounds, e.g. asi sad 'one hundred', and distributives, e.g. irā irā 'two by two').

Finally, I just have a few questions:

1. Why are acute accents used for Old English instead of macrons?
2. Why is 'ten' in Sanskrit written as dáça instead of dáśa?
3. Where do these "Vernacular" forms for Tamil come from? Which dialect are they in? Why is the transcription used in that row so markedly different from the other rows?

[Jonlang]

Under Welsh there is listed a set for "Cardiganshire". After doing some asking around these are simply Irish numbers written using Welsh orthography and they're not used in Cardiganshire. They were used there around the 17C by Irish workers but everyone there today use the Welsh un - deg.

[Richard W]

A lot of the answers are a matter of history, and the limited rendering technology available when the list of numbers was initially created. It is an old Internet resource.

2. In Sanskrit, 'six' is ṣáṣ, not *s`as`, and 'eight' is aṣṭa, not *as`ṭá.

It's a choice of transcription. The spacing grave accent is the XSAMPA diacritic for retroflexion.

What function does the acute accent serve in your recommended emendation? I suspect it is entirely predictable.

5. In the interests of consistency (and following IAST), *<a> in all the Gujarati examples should probably be replaced with <ā>, *<ã> should probably be replaced with <ã̄>, and *<ə> should probably be replaced with <a>.

Consistency is tricky - should the list follow phonetics or orthography for non-Roman alphabets?

Finally, I just have a few questions:

2. Why is 'ten' in Sanskrit written as dáça instead of dáśa?

<ç> is an old alternative to <ś>, and had the enormous advantage of being in the Latin-1 character set.

1. Why are acute accents used for Old English instead of macrons?

They're justified by being an alternative convention in Old English scholarship, and again has the advantage of being in the Latin-1 character set.

[Vijay]

It's a choice of transcription. The spacing grave accent is the XSAMPA diacritic for retroflexion.

Which isn't used for any of the other languages in the list besides Sanskrit.

What function does the acute accent serve in your recommended emendation? I suspect it is entirely predictable.

Pitch accent, surely.

Consistency is tricky - should the list follow phonetics or orthography for non-Roman alphabets?

It follows the orthography in all the other languages, so why not this one?

<ç> is an old alternative to <ś>, and had the enormous advantage of being in the Latin-1 character set.

There's daśa listed under "Fars."

[Tropylium]

A few bugs in the Uralic data:
– Finnish is not listed under languages with 1M+ speakers;
– "E Saami", as entered, should be "Inari Saami";
– "Lovozero" is more commonly known as "Kildin Saami";
– "Permyat" should be "Permyak";
– "Gornomari" should be "Hill Mari" (not "High Mari")

[edit] Several Turkic languages have missing 1M+ status too, at least Turkmen, Uzbek, Uyghur and Tatar; per most stats I've seen also just barely Bashkir.

And some transcription notes:
– South Saami: numerous transcription issues due to Finno-Ugric transcription idiosyncrasies, e.g. ‹ɔ› is preaspiration /ʰ/, and ‹ɯ› is central /ʉ/
– Permic:
• 1 should be Permyak /ötik/;
• 3 should be Udmurt /kwinʲ/, Zyrian /kujim/, Permyak /ku.im/;
• 4 should be Zyrian and Permyak /nʲolʲ/;
• 6 should be Udmurt /kwatʲ/ (though written куать);
• 7 should be /sʲizʲim/ in all three;
• ‹ы› in Permic varieties stands for /ɨ/, though I suppose you might be still using /ɪ/ as broad English-friendly transcription.
– Mari: ‹ы› in the Mari varieties, however, stands for a reduced vowel, approx. /ə/ (to my ears fairly close to [ɐ] in standard Mari), not anything in the /ɪ ~ ɨ/ region.
– Mordvinic:
• 4 should be Erzya /nʲilʲe/, Moksha /nʲilʲæ/
• 5 similarly should have /-tʲ-/ and 7 should have /sʲ-sʲ-/ in both;
• 7 should have Erzya retracted [-ə] (not lowered [-ɛ])

I could look up the numerals in some additional Sami, Mansi, Khanty and Selkup varieties, if you're interested. They have a fair bit of variation; and while a division in 10 Samic languages is now commonly used, the others have also been suggested to be dividable in 3-4 languages each.

[Xephyr]

Some of your Siouan examples use pretty awful and (I'd guess) unhelpful orthographies, which I'm guessing you must've gotten from some older sources. Here are some better ones:

Code: Select all

Chiwere  	iyáŋki	núwe	  dáñi	  dówe	thátaⁿ	 šágwe	  šáhma	   grerábrį	šánke	  grébraⁿ
Kansa		 míⁿxci	noⁿbá	 yábliⁿ	tóba	sátaⁿ	  shápe	  péyoⁿba 	kiadóba	 sháⁿka	 glébla
Quapaw		mį́xti	 nǫpá	  dábnį	 tówa	sáttą	  šáppe	  ppénǫpa	 ppedábnį	šą́kka	  kdébną
Tutelo		nǫ́:sa:	nǫ́:pa:	lá:ni	 tó:pa  kisą́:ha	aká:spe:  sa:kó:mį:  palá:ni	 ḳasą́:hka  pú:čka

Sources:
A Grammar and Dictionary of Tutelo by Giulia R.M. Oliverio
A Quapaw Vocabulary by Robert L. Rankin
Iowa, Otoe-Missourie Language Dictionary: English/Báxoje-Jiwére-Ñutʔachi -- Maʔúⁿke by Jimm G. GoodTracks (the web version, from the Ioway-Otoe-Missouria Language website)
English to Kanza Dictionary by Robert L. Rankin and Linda Cumberland (which has "DO NOT DISTRIBUTE" written on it, but I won't tell nobody if you don't)

The Osage also looks suspect, but I don't have access to the dictionary for that one (Linguoboy might know). The Lakota looks to be in a popular orthography which isn't quite the Ben Black Bear orthography (which is probably as close to ~standard Lakota~ there is) but which works I guess, except that you use <ś h́> in place of the correct <š ȟ>. I'm guessing that was a technical compromise you had to make.

('Round where I live you often see Lakota bumper stickers, tattoos, license plates, etc. And these tend usually to be in some kind of compromise orthography that combines elements from the Black Bear and White Hat orthographies + a general jettisoning of all diacritics, which resembles the orthography your page is using. So I'd say it's fine except for the acute accents. Sometimes people use overdots instead of haceks-- but never acutes.)

EDIT: Oh shit, and you don't even have the numbers for Stoney. Here they are:

Code: Select all

wiži   nũba    iyamnĩ    ktũθa    ðaptã    sakpé    sagowĩ    saħ noɣã    nãpčuwĩk    wĩkčé mnã

, according to the "official" Rocky Mountain Nakoda website. The orthography is a wee bit unusual for a NAmerican language, but trust me: it's WAY better than the one used by the John Laurie dictionary.


EDIT2: And while I'm at it, here they are for Proto-Siouan according to the Comparative Siouan Dictionary [codeboxes for some reason screw with the diacritics when I try to do it now, even though they were working before...]

1 *rų·-sa / ?
2 *rų́·pa
3 *rá·wrį
4 *tó·pa
5 *isá·ptą
6 *aká·we
7 *ša·kú·pa
8 ?
9 ?
10 *piraka / ?

Eight is listed as "unreconstructible"; nine and the second etyma for one and ten aren't said to be unreconstructible,but aren't given either.

[Richard W]

Under Welsh there is listed a set for "Cardiganshire". After doing some asking around these are simply Irish numbers written using Welsh orthography and they're not used in Cardiganshire. They were used there around the 17C by Irish workers but everyone there today use the Welsh un - deg.

Have you read the paper describing them - "An old system of numeration found in South Cardiganshire" in Transactions and archaeological record, Cardiganshire Antiquarian Society Vol. 3 (1924), p. 9-19?

While the numbers do look Irish, the preferred belief is that they hark back to when the area was under Irish control. Of course, the system was moribund by the time it was recorded.

I can't find any sign of the English sheep-counting numbers of Brythonic origins in the file. I believe these should be recorded under English, just as Japanese numbers of Chinese origin are arranged under Japanese (as 'Sino-Japanese').

[Vijay]

I can't find any sign of the English sheep-counting numbers of Brythonic origins in the file.

But are these any more notable than numbers in any non-standard dialect of English/Anglic variety/whatever indigenous to Britain? So far, all there is there is English and Scots (and that too apparently only one variety of Scots); there's no Anglo-Cornish numbers or Geordie numbers or whatever.