David Boxenhorn suggested that Setswana could calque existing terminology from other Bantu languages. I agree. Why reinvent the wheel? "Towards Strategies for Translating Terminology into all South African Languages: A Corpus-based Approach" shows how isiZulu and Sepedi cope with the challenge of translating terms like accreditation.

Setswana makes a cameo appearance in Table 1 which shows a paucity of direct borrowings from English. Only 6% of English terms in the corpus were borrowed with English spelling kept intact, and only 7% were modified to fit Setswana phonology. I presume a large percentage of the remainder may already consist of calques of other Bantu languages or English.

Note that the language names in that table other than English and Afrikaans have the prefix se- or the presumably cognate prefixes si-, isi-, xi- (pronounced like English she in Xitsonga), and tshi- (pronounced 'chee' in Tshivenda). The Zulu names for English and Afrikaans (isiNgisi and isiBhunu [< Boer]) have the isi-prefix.

airforcewife got me a little interested in Setswana. I found an entire blog devoted to it. This post jumped out at me:

"Hello, I am Karabo," she loudly purrs out her name the way a culture shocked African-American would, winning a few amused stares from an army of bored public cellular phone operators lining the space of the mall. "Akere ne rra we meet kgantele ke chaisa? Call me ga o sa kgone," she continues in that overly nasal vein, relishing as much every word she utters as those she swallows in her passion.

A few seemingly cathartic purrs and nasal-affected syllables later, she leaves a legacy of scorn and amused gossip behind.

I doubt Karabo sounds African-American. What's important is that she thinks she does. Apparently AA speech strikes her - or at least the writer - as nasal.

I wonder how rr- and kg- are pronounced.

A fluent Setswana sentence has fallen into the realm of the nostalgic, that which elicits an awed response when heard.

This has to be an exaggeration unless almost all Setswana sentences contain loanwords or code-mixing. Heavy borrowing hasn't hurt languages from Persian to Japanese - or even English.

This linguistic movement impresses not everyone. Some see the cultural phenomenon as an ominous bell that points to the extinction of the indigenous Setswana language, together with the cultural identity and knowledge duly encapsulated in the sonic breadth of words ...

"People who mix up Setswana with English wish that they could be fluent in English. If they had great fluency in English they would only use that language."

Is that merely speculation, or is it really true?

"We learn from an early age that English confers more of a sense of self worth and pride than our local languages. English is seen as wealth."

That's also the case in other former Anglosphere colonies, but Hindi, Tagalog, etc. are not dying out any day soon. Hinglish and Taglish have yet to supplant them.

"I think it’s time that jobs were applied for in Setswana. Imagine how it would be if we wrote our curriculum vitae in Setswana," he contends.

It's possible. The challenge is to adapt the language to CVs and other modern contexts.

Kekwaletswe contends that Setswana and other indigenous languages, as Third World languages, are not developed in a way that can accommodate new lexical icons, especially those found in information technology such as ‘mouse’ and ‘cursor’.

If Chinese can develop native IT terminology (e.g., 鼠标 shubiao 'mouse marker' and 光标 guangbiao 'light marker'), so can Setswana. Or Setswana speakers could just keep the English terms more or less as is: cf. Japanese mausu and kaasoru.

He further laments that Radio Botswana seems to no longer have personnel that can ably add new words to the Setswana vocabulary as they did with terms such as ‘kgotlha-more-o-itirise’ for ‘machinegun’.

That is a mouthful compared to Mandarin 机关枪 jiguanqiang or Japanese 機関銃 kikanjuu which are both trisyllabic like English.

A lecturer in the department of African Languages at the University of Botswana, who would not divulge her name because she "presents a synopsis of various opinions solicited from a number of linguistic and literary persons in the faculty of Humanities", is adamant that given its own cultural context, Setswana is perfectly adequate to express any concept or knowledge.

I obviously agree with her, and I would add that it can expand to accomodate other cultural contexts.

The UB linguist further argues that Setswana has the potential to be a language of business and trade. "Isn’t it that business is done with Batswana [Setswana speakers]? Covers and tags for locally produced goods should be in both Setswana and English," she argues.

This reminds of what Brian Deutsch wrote about cosmetics:

I'm glad I'm not the only one who doesn't understand why Korean cosmetic companies insist on using French and English on their labels, when most potential customers have no idea what they say ...

... [U]sing a language your customers actually understand isn't such a bad idea ...

It is, as one of the women quoted in the press release article says, because companies and customers want to feel more sophisticated through these foreign languages ...

Language is meaningful even when it can't be understood.

[The UB linguist] adds that language is made stronger by its users. According to her, it is the speakers who make English superior to Setswana and it is them who can also place it in a positive light by not placing it below other languages. She concurs with Kekwaletswe that there is a need to restore confidence in indigenous languages on the part of government by translating public texts, including laws and policies because they are after all made for Batswana.

Imagine not being able to read the laws of your own country. (Not that I can navigate legalese in English!)

"Let us put in place language use structures in our administrations, communications, businesses and schools. Let us write HIV/AIDS messages in Setswana and other indigenous languages of Botswana, where possible.

One out of four adults in Botswana has AIDS. AIDS education in Setswana is more important than any of this:

Let us reward good use of Setswana for instance, the best Setswana speaker and user in the media, at work or at school. Let us invite the general public to coin new Setswana expressions for new technologies and equipment," she advises.

Getting the public to actually adopt these creations is another matter. Suppose you are a Motswana engineer trained in English. Do you really want to relearn all your terminology in Setswana?

(Motswana is the singular of Batswana.)

Keep the Batswana alive first and worry about Setswana later. (Note how the prefixes distinguish the people from the language and the country.)

ADDENDUM: How can this be?

In 2002, the gross primary enrollment rate [in Botswana] was 103 percent ...

Were some children simultaneously initially enrolled at more than one school?

09.4.16.00:48: DRAWING ON DONKEYS

I never had a schildersezel 'painter's donkey'.

09.4.16.00:46: 超人前傳 SUPERMAN: THE PROLOGUE*

What show is that? And what is 異世奇人 Different World Strange Person?

See other ATV translations of Anglosphere TV show titles here.

Google translates 前傳 'prologue' (lit. 'before story') as ... 'Batman'! Is this because of the movie 蝙蝠俠：俠影之謎 Bat Hero: Mystery of the Hero's Shadow which some regard as a prequel? It's really a reboot; its sequel is 蝙蝠俠—黑夜之神 Bat Hero - God of the Black Night.

*Literally 'before story'.

09.4.16.00:40: WHY DOES MAC OS X HAVE "A STRANGE ARABIC KEYBOARD LAYOUT"?

asks aktub.com. I know Macs are supposed to be different, but that might be going too far. I'd like to know how that layout differs from other ones I've seen, but I can't find a graphic.

Why do some Arabic keyboards I've seen (e.g., these two) have initial forms of some letters instead of isolated forms?

is Google's translation of 玉皇朝 'Jade Dynasty', a major Hong Kong comics publisher. Oddly, 玉 'jade' and 皇朝 'dynasty' (never 'North Korea'!) are translated correctly in isolation. Is this mistranslation due to statistically-based translation based on texts mentioning both North Korea and its Kim dynasty?

I just discovered that Jade Dynasty is now ... Sparkle Roll. What a ridiculous name. Who would want to buy "ultra-luxury cars" from them? Answer: People who don't care about the English name. The Chinese name of the company - the name that matters most - is 耀萊 'shining pigweed (or wild grass)'. Doesn't sound so great either. What's the reasoning for these names? Their URL (hk970.com) is hard to remember.

Jade Dynasty (and its predecessor Jademan) were named after the Stan Lee of Hong Kong, 黃玉郎 Wong Yuk Long (Yuk Long is 'yellow jade man'), the pen name of 黃振隆 Wong Chun Loong ('yellow raise prosperity').

09.4.16.00:16: RAR PLUS TWO STILL EQUALS RAR

In "The Passage of the Night", I found that many tangraphs pronounced raʳ contained the element

Tonight, I counted 10 out of 17 raʳ-tangraphs with that element.

Three more raʳ-tangraphs

have that element plus two horizontal lines:

(The "Two" in the title refers to Chinese 二 'two' which reminded me of these two added horizontal lines. Note that the extra lines in the tangraphic component are of equal length, unlike the lines of 二 'two'.)

Why wasn't the same element used in all thirteen tangraphs? I don't know. There is a minimal pair

But there are no minimal pairs for

so the two 'extra' horizontal strokes could be omitted and the phonetic elements would remain functional.

The absence or presence of the two 'extra' lines does not correlate with tonal differences. Although

is only in tangraphs with the rising tone, it is not the only raʳ rising tone phonetic because

is in tangraphs for raʳ with both the level and rising tones.

The presence of either element does not guarantee a reading raʳ: e.g.,

kəɨ, kəɨ, ziʳ, tʃhwɨʳ, reʳ

mɨʳ, liẹ

Moreover, there are four raʳ-tangraphs without either element - and without anything in common!

This situation is reminiscent of sinography:

- a syllable X can be represented by different phonetics A and B: e.g., ma = 麻 and 馬 in

嘛痲犘嬤 ... (all ma)

媽嗎螞碼 ... (all ma)

- A and B can also appear in non-X sinographs: e.g., 麾 hui and 駟 si

(In fact, 麻 ma < Old Chinese *mraj is phonetic in 麾 hui < Old Chinese *hmraj, but this is not obvious to modern Chinese speakers. 馬 ma 'horse' is semantic in 駟 si 'team of four horses', a semantic extension of 四 si 'four'.)

- a syllable X can also be written without A and B: e.g., 抹孖蟆 (all ma).

(The examples are based on modern standard Mandarin without reference to tones and earlier readings. In Old Chinese, 麻 *mraj and 馬 *mraʔ did not rhyme.)

refers to an etymology I came up with for Tangut

na raʳ 'tomorrow'

Initially I thought this could be an unanalyzable disyllabic word since neither of its halves are attested in isolation. However, na is homophonous with

na 'night; darkness'

and raʳ is homophonous with

raʳ 'pass through; passage; go away; depart; retire'

so I wondered if 'tomorrow' was originally 'night passage' (i.e., the time after the night goes away). The Tangut may have devised two distinct tangraphs for 'tomorrow' instead of simply using the existing tangraphs for 'night' and 'passage' because they no longer regarded 'tomorrow' as a compound 'night-passage'. Both tangraphs for 'tomorrow' contain the radical

'go beyond'

(derived from the right side 戉 of Chn 越?)

<

na (with the right side omitted)

and

raʳ, found in many raʳ-tangraphs

including the one for the second half of Sarah's name:

as well as

'passage'

Tonight I confirmed my guess when I found that the Tangraphic Sea defined na of na raʳ 'tomorrow' as

the passage [raʳ!] of today's night; therefore [it is] called 'tomorrow'.

... of nation-building, with strong political and weak linguistic nationalism?", asked Pål Kolstø, who's not fooled by hype about language (emphasis mine):

Large parts of the titular population had a better command of Russian than of their own ‘mother tongue’, or they could not speak their putative mother tongue at all. This was the situation in vast swathes of Soviet Central Asia, even though the official statistics did not show it. The regular censuses taken roughly every tenth year included a question about ‘mother tongue’ or rodnoi iazyk. In Central Asia more than 90 per cent of the indigenous population in all censuses in the Soviet period stated the language of their ethnic group as their rodnoi iazyk. Many Western researchers (including renowned authorities such as Alexandre Bennigsen, S. Enders Wimbush and Richard Pipes) were misled by these statistics to believe that the position of the various local languages was very strong. The statistics, however, while not deliberately doctored by anyone, were quite deceptive. When interviewed by a pollster many linguistically Russified Central Asians seem to have stated the traditional language of their ethnic group as their rodnoi iazyk. To themselves they may have justified this in various ways. The adjective rodnoi is derived from the word ‘rod’ which means family, kin or clan, and some might have reasoned that, well, Kazakh is the language of my kin and clan, even if I do not speak it myself. Those who reasoned in this way, however, have hardly acted in good faith, as most of them no doubt understood quite well what the pollster wanted to find out. So the main reason behind such imprecise or deceptive responses, I believe, was reluctance to admit that one had ‘let down’ the language of one’s forbears.

Kazakhstan was one of the republics where the discrepancy between actual and stated mother tongue was exceptionally large. While 98.5 per cent of ethnic Kazakhs in 1989 declared Kazakh as their rodnoi iazyk, most sober analysts today estimate that as much as 25-40 per cent of them at the time were unable to express themselves fluently in this ['native'] language. This was particularly true with regard to the Kazakh intellectual and political elites ...

An article in the Kazakhstani social science journal Saiasat in 1996 noted that ‘today, it is the Russian language that creates the basis for the unity of all Kazakhstanis. It functions as a means of communication all over the republic and enables social interaction among all citizens in the country, across social, demographic, territorial and professional group boundaries.’

Kolstø's answer to his own question appears to be no, since he sees the Kazakhstani situation as triadic rather than dyadic like Ireland and Norway:

... the Irish test of strength was therefore dyadic, with only two main combatants, the Irish nationalists on the one hand, and the British authorities and the British settlers on the other ...

Also in Norway the national struggle has been dyadic rather than triadic, but for different reasons. In this country there was indeed a strong challenger group among the nationalists rallying under the banners of the vernacular language, but in this country the third element that complicates the Kazakhstani linguistic picture was lacking. In Norway after independence there were hardly any Danish settlers, and the few that were had long ago ‘gone native’. Thus, for all practical purposes, there was no demographic group in this country who spoke the old colonial language as their ‘ethnic’ or ‘indigenous’ language. The triangular power struggle that characterises the Kazakhstani language situation, therefore, seems to be quite unique.

09.4.14.00:41: WHAT'S THE NUMBER ONE NUMBER TWO FOREIGN LANGUAGE?

... on Korean college admission exams (via Brian Deutsch)? I would never have guessed! (Emphasis mine.)

However, no schools [in Seoul] are running Arabic language classes, which some 300 million people in the world use. Ironically, Arabic was the most selected [second foreign] language by Korean students for the college admission test last year. ...

And the reason why is a surprise, too:

According to the Ministry of Education, Science and Technology, 29,278 high school students out of 99,693, or 29 percent, adopted Arabic language for the second foreign language test of the Scholastic College Ability Test last year because the Arabic test was easier than other foreign language tests.

I agree with sonagi92 (sonagi is Korean for 'rain shower'):

These [language] exams, like so many others, are a waste of time and money. Most students study the language only to prepare for the exam and then do nothing with the language afterwards.

John B added:

Which is pretty much just like the 1-2 year foreign language requirements for undergraduates at American universities.

Or American high schools. Or a lot of school subjects.

(Insert obligatory sanctimonious speech about opening the minds of The Children here.)

True or not, this is an argument I've heard for learning Korean before Japanese:

Apparently a lot of parents in Bulgaria for example choose to have their children focus on German because English is so widespread and easy to learn after German that you can pretty much get it for free afterward, whereas going from English to German (less people use it, grammar takes a long time to master) is more of an uphill struggle.

Going from Japanese to Korean was "more of an uphill struggle" for me.

09.4.14.00:33: A MAS-Q-ULINE Q-ONSONANT?

According to Haeri (1991),

Women's speech [in Cairo] has frequent and advanced palatalization, while men's does not. On the other hand, men's speech contains qaf lexical items more frequently than women's in all educational levels and social classes.

Palatalization and uvular q are at opposite ends of a spectrum. Why is each associated with a different gender? My initial guess was that male speech is more like Classical Arabic which lacks palatalization and has a q that has shifted to ʔ in Cairene. However, the abstract indicates no correlation between these Classical-like features and "educational levels and social classes". Even uneducated, poor men have these Classical-like features and conversely even educated, wealthy women tend to lack them (relative to men in their stratum, or men in general?).

Sagart (2007) proposed that Old Chinese nonemphatic uvular *ɢ- (but not emphatic uvular *ɢ-) became Middle Chinese *j-. I objected to this for two reasons:

1. Does any extant language have a distinction between nonemphatic and emphatic uvulars? According to Islam Youssef (2006), Cairene Arabic has a consonant system similar to OC with a nonemphatic/emphatic distinction for all consonants with the exception of its single uvular q which behaves like an emphatic.

2. g. is closer to j (i.e., fronter) than ɢ. Hence I might expect a chain shift like

ɢ > g > j

Yet in Sagart's system

OC*ɢ- > MC *j- (fronter than *g-)

OC*g- > MC *g- (no change!)

How did OC*ɢ- 'gjump' to MC *j-? Perhaps there was a fricative bridge:

OC*ɢ- > *ʁ- > *ɣ- > MC *j-

ʁ occurs in 4.88% of UPSID's sample but its stop counterpart ɢ only occurs in 3.1%. I suspect that some ʁ were originally *ɢ.

This takes care of my second objection, but my first remains.

Then again, I just discovered Rutul, which has both nonemphatic and emphatic ɢ(ʷ). But it is the only language in the UPSID sample to have that distinction. However, not even Rutul has a distinction like Sagart's

OC *q : *q

OC *qʷ : *qʷ

Rutul only has *q and lacks *qʷ and *qʷ. Yet it does have both nonemphatic and emphatic uvular fricatives (unlike Sagart's OC) and aspirated stops (like Sagart's OC):

χ : χ

χʷ : χʷ

ʁ : ʁ

ʁʷ : ʁʷ

qh : qh

qʷh : qʷh

And it has nonemphatic and emphatic uvular ejectives (unlike Sagart's OC):

qʼ : qʼ

qʷʼ : qʷʼ

So maybe I should withdraw my first objection as well, though I am still wary of making OC and Rutul the only two languages in the world with lots of uvulars with nonemphatic/emphatic distinctions.

For comparison, my OC reconstruction has nonphonemic emphasis for all uvulars:

*q *qʷ *qh *qhʷ *ɢ *ɢʷ

/q qʷ qʰ qʰʷ ɢ ɢʷ/

[qˁ qʷˁ qʰˁ qʰʷˁ ɢˁ ɢʷˁ]

is what

na 'night; darkness' (from "Have a Great Day - and a Great Night!")

looks like to me. It contains

'surround'

ironically surrounded by two

'people'.

In theory, 'night' could be analyzed as

1. A + B + A

2. A + B + C (with the 'people' having different functions)

3. AB + C

4. A + BC

5. (A + C) + B with the 'people' as a single surrounding unit

cf. how Chinese 行 and 干 (no relation to the tangraph) combine into 衎

6. ABA as an indivisible whole happening to resemble elements A and B

But the actual analysis in Tangraphic Sea is

=+

na 'night; darkness' =

left of na (first half of na raʳ 'tomorrow' - more on that ... tomorrow) +

left of kõ (second half of na kõ 'gloom')

The left and center of na 'night' look like

pi 'wall (hence 干 'surround' on the right?); name of an ancestor (hence 'person' on the left?); majestic, glorious'

but seem to be an abbreviated phonetic na.

Is 'person' from kõ really enough to suggest darkness? Is kõ attested anywhere but in the disyllabic word

na kõ 'gloom'

=+

kõ (second half of na kõ 'gloom') =

left of na 'night' +

all of nɨaa 'black'

Was 'person' added to 'black' so that kо̃ would share a left side with na 'night' in

na kõ 'gloom'?

Disyllabic words with bound morphemes are often written with matching radicals in Chinese: e.g.,

蝴蝶 'butterfly' (with repeating 虫)

珊瑚 'coral' (with repeating 王)

09.4.13.23:49: I STILL DON'T KNOW WHY THE OCS WIKIPÉDIA IS ÁCCENTED

I can't find any reference to suprasegmentals in Comrie and Corbett's The Slavonic Languages or the Wikipedia article on OCS.

Although the article on the early Cyrillic alphabet describes a Greek-like system of diacritics for OCS, I don't see the full range of accents in the OCS Wikipedia and am not sure if these diacritics are reliable guides to OCS phonology.

"Old Church Slavonic accent" and "OCS accent" have only 2 and 6 hits respectively in Google. Am I overlooking something obvious?

Was OCS normally written with accents? In Comrie and Corbett's The Slavonic Languages, the OCS transliteration has no accents, even though the Russian transliteration contains them (in spite of the absence of written accents in normal Russian orthography). Earlier tonight I was looking at Kortlandt's "From Proto-Indo-European to Slavic" in which OCS is also transliterated with accents. So if these linguistics books lack accents, what are accents doing in http://cu.wikipedia.org/? The Glagolitic script logo for the OCS Wikipedia has no accents, yet the OCS for 'Wikipedia' appears with an accented є

Википє́дїѩ

in the Cyrillic script. Obviously there was no word 'Wikipedia' in OCS until very recently, so I'm hoping the accent was determined by analogy with some other word like Russian Википедия which is accented on е. Similarly, 'Unicode' appears with an accented о

Юнико́дъ

(I am guessing the nominative singular which I can't find at cu.wikipedia.org)

even though it too cannot be ancient; its accent may be by analogy with Russian Юникод which is accented on о.

(The Russian links lead to Wiktionary entries containing full paradigms for 'Wikipedia' and 'Unicode' with accents.)

On Ostern, as German speakers call Easter, I wondered if German Ost (and English east, Dutch oost, etc.) could be cognate to Russian vostok 'east' as in Vladivostok (once 海 參崴 'sea cucumber cliffs' in Chinese!). But I checked and concluded that they are coincidental lookalikes. The Germanic words are from a Proto-Indo-European root *aus- 'shine' (east being the direction where the sun first shines) without a *w- corresponding to Russian v-.

Although prothetic v- can develop in Slavic -

'window':

Ukrainian vikno

Belarusian vakno (unstressed o became a)

colloquial Czech vokno

cf. Russian, Polish, and literary Czech okno

'he':

Ukrainian vin

colloquial Czech von

(but Belarusian has jon, not von!)

cf. Russian, Polish, literary Czech, and Serbo-Croatian on

(I'd like to examine Ukrainian -i- corresponding to -o- in other languages later. More differences between colloquial and literary Czech here.)

- I don't know of any such cases in Russian, and the v- is quite old, judging from Old Church Slavonic въстокъ vŭstokŭ 'east'. Does any other Slavic language have a cognate for these words? I've found two types of non-vostok forms for 'east' elsewhere in Slavic:

1. South Slavic only: Serbo-Croatian istok, Bulgarian iztok

2. South, West, and East Slavic: Slovene vzhod, Czech and Slovak východ, Polish wschód, Ukrainian sxid (with i for o), Belarusian usxod

(Slovene h and Czech, Slovak, and Polish ch are [x].)

While talking to Guard Wife about Paas, I verified that its name is indeed based on a p-word for 'Easter': Pennsylvania German Passen (cf. Dutch Pasen and Afrikaans paasfees; the standard German name is Ostern).

Along the way, I found some Slavic names for 'Easter' which mean 'great day' and 'great night'. The Slavic n-words for 'night' are cognate to English night, but the Slavic d-words for 'day' are not cognate to English day, though they are cognate to Latin dies (as in Dies Irae).

I could translate the Slavic names into Tangut as

liẹ nɨɨ

'great day'

liẹ na

'great night' (is na 'night' cognate to nɨaa 'black'?)

In various East Asian languages, 'Easter' is the 'resurrection (< 'again-live') celebration':

Note that Vietnamese has 'X of Y' order (rite of resurrection) whereas the others have 'Y X' (resurrection festival) order.

The Thai name used at Wikipedia is simply

อีสเตอร์

ʔiittəə