Language Log

http://languagelog.ldc.upenn.edu/nll/?p=23175

For Bob Dylan connoisseurs, the release of The Cutting Edge 1965-1966: The Bootleg Series Vol. 12 is a momentous occasion. It encompasses the studio sessions that gave us the albums Bringing It All Back Home, Highway 61 Revisited, and Blonde On Blonde, and it's available as a 2-CD sampler, a reasonable 6-CD version, and an ultra-comprehensive 18-CD collector's edition for the true Dylan obsessives. The collector's edition, which compiles every outtake from those crucial 1965-66 sessions, may have been released by Columbia primarily for copyright reasons, but for those willing to slog through the 19-hour runtime, there are some unexpected pleasures.

For a Billboard review, Chris Willman listened to the whole 18-CD set in a marathon session. Here's how he describes one track:

Dylan grows increasingly frustrated by how he feels the Hawks are mangling "She's Your Lover Now." "Aw, it's ugly," he says. "I can't. I can't even." Did Bob Dylan just invent the 21st century catchphrase "I can't even"? I think he did!

I managed to find the track in question: "She's Your Lover Now – (1/21/1966) Rehearsal" (Disc 11, Track 3). Here's a clip of the relevant bit:

Some background… The Jan. 21, 1966 session took place as Dylan was putting together the songs that would become Blonde on Blonde. He was recording in Columbia's New York studio with a group that included members of The Hawks, who backed him up on his first "electric" tour and would soon find fame on their own as The Band. Dylan was dissatisfied with how "She's Your Lover Now" was turning out — although a nearly complete full-band version of the song, first compiled on The Bootleg Series Vol. 1-3, sounds awfully good. After the rehearsal in which Dylan "can't even," he dismissed the band and tried doing the song on his own, accompanying himself on the piano. That sounds awfully good too. (Elvis Costello says it's his favorite track on The Cutting Edge, telling Esquire, "The whole clambake is worth that one performance.") But Dylan would ultimately abandon the song, leaving it off Blonde on Blonde, which was mostly recorded by studio musicians in Nashville after Dylan moved down there in Feb. 1966. (See the Sean Wilentz essay "Mystic Nights" for more on these sessions.)

So was Dylan half a century ahead of his time because he couldn't even? For more on "I can't even" and recent elaborations like "I've lost the ability to even," see:

• Tia Baheri, "Your Ability to Can Even: A Defense of Internet Linguistics," The Toast, Nov. 20, 2013
• Rebecca Cohen, "In Defense of I Can't Even," Lexicon Valley (Slate blog), Mar. 12, 2014
• Michael Reid Roberts, "Life Sentences: The ABCs of 'I Can’t Even.'" The American Reader, May 16, 2014
• Clive Thompson, "I Can't Even." The Message (Medium blog), July 21, 2014
• Gretchen McCulloch, "Why Is It That You 'Can't Even' But You Never Find That You 'Can Even'?" Mental Floss, Mar. 24, 2015

I like Gretchen McCulloch's characterization of "I can't even" and its kin as "stylized verbal incoherence mirroring emotional incoherence." When Dylan complained, "I can't… I can't even…", he was certainly experiencing emotional incoherence, and in his verbal expression of it, he was perhaps trying to convey something like, "I can't even get through this song (or this sentence)." Of course, that's a long way from the stylized incoherence of modern-day can't-eveners, as spoofed not too long ago on Saturday Night Live.

http://languagelog.ldc.upenn.edu/nll/?p=22954

One of Kohei Jose Shimamoto's photos on Facebook:

Fó tiàoqiáng 佛跳墙
("Buddha jumps over the wall")

This is the name of a famous dish with a long history:

"Buddha jumps over the wall" –> "God use VPN".  Ingenious!

[h.t. Ki-Tsìng; thanks to Maiheng Dietrich and Yixue Yang]

http://languagelog.ldc.upenn.edu/nll/?p=22975

This is a photograph of a sign above a urinal at the Tokyo University of Foreign Studies taken by Joseph Williams who was there for a Japanese test.  Besides the Japanglish, it's interesting that spaces are added between the words.  And there are no kanji.

Toire wa kirei ni tsukatte kudasai
トイレは  きれいに  つかって  ください

The correct translation will emerge in the course of the following discussion.

First of all, although the Japanese may look strange because it is all in kana (no kanji) and has spaces between words, there is nothing unusual about the wording.  While this may not be the most exemplary Japanese, one can perfectly well state the request as it appears.

A more traditional and elegant way of saying the same thing would be:

Toire wa kirei ni tsukaimashou
トイレは綺麗に使いましょう

"Kirei ni tsukau 綺麗に使う" is standard wording to express "to use something cleanly".  The sentence might well have been translated as "Please keep the restroom clean."

The sign is targeted at language learners (Japanese-Language Proficiency Test [JLPT]), which explains the spaces. Picture books and even textbooks used in Japanese elementary schools have spaces for early readers.  I do not know when word separation like this began, but it was probably fairly early into public education in Japan, since it may be seen in prewar and wartime textbooks:

I don't think that there is any rule against putting spaces between words, and providing them makes it easier to understand, especially because there are no kanji in this sentence.

The only mistake in the translation is the result of choosing "beautiful" instead of "clean" from the two main meanings of kirei きれい.

You see signs like this all the time (Google Images).

So this is just a case of misunderstanding that, while cleanliness is next to beauty, it's not the same thing as beauty.

To sum up, the spaces between words and the lack of kanji on this sign are clearly for non-Japanese speakers who come to take the lower levels of the Japanese-Language Proficiency Test.  All levels of the JLPT are administered on the same day.  The kanji requirement for the lowest level (Level 5 or N5) is quite light:  tsukau 使う (probably) and kirei 綺麗 (definitely) are not included.  It's interesting that, if the same sentence had been posted with kanji and no spaces, it is assumed that students in the lower levels of Japanese would not have been able to understand it.  Written all in kana and with word spacing makes the sentence understandable to students who are testing for all five levels of the exam.  Writing in rōmaji and with spaces would make it understandable to an even greater range of Japanese language learners.

[Thanks to Nathan Hopson, Cecilia Segawa Seigle, Hiroko Sherry, and Miki Morita]

http://languagelog.ldc.upenn.edu/nll/?p=23178

All Synthèse books published before 2005 appear to be free to download in .pdf form from Springer. I haven't verified that this is true for IP addresses outside of universities with a subscription, but I think it is.

This include the series Studies in Linguistics and Philosophy, but there are likely to be other titles of interest to some LLOG readers.

[h/t Kai von Fintel]

http://languagelog.ldc.upenn.edu/nll/?p=23151

This one's been on my to-blog list for a while. "The Ascent: Political Destiny And The Makings of a First Couple," a "sponsored content" piece for Netflix's series House of Cards prepared by The Atlantic, won both the Judge's Selection and the People's Choice awards for "Best Sponsored Content Editorial" at the Native Creatives:

What are the "Native Creatives"? According to the web site,

The Native Creatives honors creativity and innovation in native advertising. Native creatives are publishers and brands who have established themselves as visionaries in content creation, creative execution, and user engagement.

What's the relevance to Language Log? Well, to start with, this award introduces a term that you might not know, though by now you're thoroughly familiar with its referent: "native advertising". According to Wikipedia,

Native advertising is a type of advertising, usually online but feasibly elsewhere, that matches the form and function of the platform upon which it appears. In many cases, it manifests as either an article or video, produced by an advertiser with the specific intent to promote a product, while matching the form and style which would otherwise be seen in the work of the platform's editorial staff. The word "native" refers to this coherence of the content with the other media that appears on the platform.

And a second reason is that I'm listed as a source of "Linguistics Expertise" for the winning piece, The Ascent:

In fact, aside from a bit of script-hacking, my only role was to advise Sam Rosen to turn away from the idea of analyzing real-world exchanges of letters, towards an analysis of the scripts and performances in the House of Cards series itself; and to connect him with Jamie Pennebaker, who in turn recruited his student Ryan Boyd to do the linguistic analysis that led to the segment of "The Ascent" called "Frank and Claire: Patterns of Power", which included these figures, along with an explanation of the methods used:

Native Advertising is an interesting hybrid case — journalists don't normally pay for technical help on stories, but maybe pieces of this kind should be different. I treated this case like any other interesting question from a journalist — and I think that the results were more accurate and interesting than usual, linguistically and otherwise — but perhaps I should have asked for a consulting fee. The subhed of an Advertising Age story tells us that the funds were there: "Netflix Goes Native on The Atlantic to Promote 'House of Cards': Pays Six-Figure Sum for Lengthy Multimedia Article on First Couples".

Anyhow, I sent a note to Jamie Pennebaker on another topic recently, and added "Congratulations on the House of Cards awards!" Jamie's reply:

What House of Cards awards?
Did I win the "Best Text Analysis" Emmy again?

Not yet, I'm afraid. But I'll take this as another piece of evidence that linguistic analysis has mass appeal.

http://languagelog.ldc.upenn.edu/nll/?p=23131

If you look up the pronunciation of poinsettia in the dictionary, you'll find two versions, one that follows the spelling in a regular way (/ˌpɔɪnˈsɛ.tɪə/) and one that would more naturally correspond to the spelling "poinsetta" (/ˌpɔɪnˈsɛ.tə/).

A couple of days ago, a journalist contacted me about this. I knew that the word was formed by adding the usual pseudo-Latin -ia to the last name of Joel Roberts Poinsett, just as Clarke Abel's name gave us abelia and William Forsyth's name generated forsythia. And I knew that there is a common (and even dictionary-sanctioned) alternative pronunciation for poinsettia. But why the i-less version of poinsettia and not (for example) a similar version of forsythia?

A quick web search didn't turn up anything specific about the history. But I suspect that this is a sort of orthographic/phonological neighborhood effect. There are lots of common traditionally-female first names that end in -etta — the U.S. Census Bureau list of the 5,000 commonest female first names from the 1990 census has 54 of them, from

to

as well as some familiar words like beretta, operetta, vendetta.

None of the 5,000 commonest female (or male) names ends in -ettia, and after poinsettia, the commonest -ettia words in an English Wikipedia snapshot are weird enough that I wasn't familiar with any of them (each word is preceded by its count of occurrences in the Wikipedia snapshot that I tested:

392 poinsettia
142 millettia
28 cettia
27 manettia
12 comparettia
8 lettia
7 hulettia
6 vettia
6 pugettia
6 silhouettia
5 quekettia

And -ttia isn't any better:

643 mattia
417 zakarpattia
392 poinsettia
142 millettia
119 attia
76 lottia
75 neottia
64 perittia
64 prykarpattia
64 melittia
55 harttia
36 orcuttia
28 cettia

It's true that -ythia yields only a single first name, Cythia (in 2058th place), and no familiar (to me) words other than forsythia and Scythia— thus again from the Wikipedia snapshot:

565 scythia
247 pythia
121 forsythia
105 paramythia
25 stichomythia
19 alaythia
7 skythia

But on the other hand,  -ytha is worse — no names, and slim pickings in Wikipedia:

91 gytha
88 cassytha
35 clytha
35 waytha
16 aaytha
10 paramytha
8 edgytha
6 mildgytha

In contrast, -thia yields a reasonable harvest of names — besides Cythia, there's  Cynthia (in 28th place), plus the variant spellings Cinthia / Synthia, and Alethia. And there are a few other perhaps-familiar place names ending with -thia in the Wikipedia snapshot, like Parthia and Carpathia.

So all in all, I think the neighborhood-effect explanation is plausible, though there might well be some additional factors unknown to me.

The journalist also wondered whether the "poinsetta" pronunciation is a mistake, and whether people who use it should be corrected. My response was that there are lots of similar cases of variants with a phoneme or two missing — february, surprise, etc. — and the fact that such variants are listed in dictionaries is a good reason not to correct people who prefer them. And there are other cases where pronouncing the lost phonemes is actually a mistake — wednesday (at least in the U.S.), worcester, etc.

What Horace said about word usage also applies to word pronunciation:

multa renascentur quae iam cecidere cadentque
quae nunc sunt in honore vocabula, si volet usus,
quem penes arbitrium est et ius et norma loquendi.

Many words shall revive, which now have fallen off;
and many which are now in esteem shall fall off, if it be the will of custom,
in whose power is the decision and right and standard of language.

http://languagelog.ldc.upenn.edu/nll/?p=23102

Never mind that it doesn't work, this is the supreme pipe dream for inputting Chinese characters on electronic communication and information processing devices.  Of the many thousands of Chinese character inputting systems (see also here and here) that have been devised, some work fairly well and some barely function at all, but this one has to take the cake for being the most ridiculous of all.  It is all the more preposterous that initially it was intended for smartwatches with their tiny glass surfaces.

The name of the system gives it away, that is, yībǐyīzì 一筆一字 ("one stroke one character").

Since the average Chinese character has twelve strokes, and many characters have twenty or more strokes, it would be utterly impossible to input the thousands of different characters with just one stroke.

You can find an index to the characters by total stroke count here.

Here's a bilingual introduction to "Ibeezi".  Note that it doesn't really tell you how the system works, and it doesn't give any examples.

The official site of "Ibeezi" contains a video purporting to show you how to use the application.

Here's a comment by a gullible graduate student from China who almost got snookered:

I tried the application myself, but haven't been able to master it.  Some words can be really hard to find.

It seems it's a combination of pinyin input and radical input, and I haven't found a way to input the characters by one stroke.

'Nuff said.

http://languagelog.ldc.upenn.edu/nll/?p=23097

[This is a guest post by Jichang Lulu]

The usual Chinese name for the Lena River is 勒拿河 Lèná hé. That's not a particularly felicitous transcription. Lèná rhymes with 圣赫勒拿 Shèng Hèlèná i.e. St Helena; it fails to reflect the palatalisation of the l in the Russian name. An alternative name transcribes the syllable ле with 列 liè, following the usual practice.

Lèná might not be particularly faithful to the Russian pronunciation, but at least it should immunise the Chinese public against the belief that Lenin 列宁 Lièníng named himself after the river. (The idea is widespread, but Lenin apparently was already using the pseudonym well before the events supposed to have motivated the choice.)

An earlier name is attested in Qing documents, namely 里雅那江 Lǐyǎnà jiāng. The name is not without historical significance: Langtan 郎坦, a Manchu official, began the negotiations that led to the treaty of Nerchinsk by claiming all territories up to the Lena river, something the Russians didn't appreciate. A Qing source that narrates the incident, the 1739 Baqi tongzhi chuji 八旗通志初集 or "First Edition of the History of the Eight Banners", uses the name Liyana in Chinese. I haven't been able to consult the Chinese original, but there's a Russian translation that shows that form the Chinese name (in its Pallady Cyrillisation лияна цзян)[1] (yes, that leads to a footnote; even blog posts can have footnotes).

The iya in Liyana makes the form look Manchu-mediated, and that's because it is. The talks at Nerchinsk were conducted in Latin through Jesuit interpreters, and the primary language of pretty much everyone in the Qing delegation (such as Songgotu (q.v. at Dartmouth), who signed the treaty, and indeed Langtan himself) was quite likely Manchu. So we would have Russian Лена Lena > (a Latinised form, oral or written) > Manchu Liyana > Chinese 里雅那 Lǐyǎnà.

The Lena river isn't mentioned in the text of the Nerchinsk treaty (because in the final agreement all its course ended up on the Russian side), but it appears in other Manchu documents. A Manchu map connected to the Nerchinsk negotiations is the "Map of the Nine Rivers of Jilin" (吉林九河圖), which does include the Lena. The map is reproduced on the website of Taiwan's National Palace Museum (國立故宮博物院); the Lena is visible on the upper left. Unfortunately the labels are unreadable at that resolution, but scholarship on Qing maps[2] confirms the Manchu name was indeed Liyana bira.

So the Manchu and Chinese name is transparently derived from Russian. While the name of the Lena river in Russian and Yakutian (a Turkic language spoken in the area) is assumed to be of ultimately Tungusic origin, the lack of a vowel before the 'l' means that the Manchus had to take it from the Russian rather than directly from, say, Yakutian or Evenki. Despite Langtan's vague claim that the Qing empire should extend all the way to the Lena because that's how far their state had reached in the past (unclear which state, and in which past), Qing officials weren't able to come up with a better Chinese or Manchu name for the river than a transcription of what the Russians called it.

That's relevant to a larger issue. Plenty of places in what is now Far Eastern Russia (though usually not so far inland) do have Chinese names, something many on both sides of the border are well aware of. Many non-Russian, often Chinese, toponyms in the Amur basin remained in common use more than a century after the Qing ceded a large swath of land to Russia at the Treaty of Aigun in 1858. China has denounced Aigun as an "unequal treaty" (不平等条约), and, although the Chinese government isn't claiming those lands back, anecdotal evidence suggests many in China feel they should have them. Those feelings are in turn seen with suspicion among some Russians now living in the region. Although a USSR Council of Ministers resolution Russified Far-Eastern toponyms wholesale in 1972, many of the older names remain in use among the local population.

An example I've written about is the location of the 'Tigre de Cristal' (水晶虎宫殿), the largest casino in Russia, opened a couple of months ago near Vladivostok to cater to a largely Chinese clientele. The casino is on a bay now officially called Muravyinaya (бухта Муравьиная; бухта is a loan from the German cognate of bight). Myrmecological though that sounds ('ants' cove'), the name is more likely to honour Count Nikolay Muravyov-Amursky than an insect infestation. It was Muravyov who signed the 'unequal' Aigun treaty on behalf of Russia, earning the lands Vladivostok now lies on for the Empire, and the comital title 'Amursky' for himself.

The name Muravyinaya is obviously not older than the Treaty of Aigun, and many locals keep using the older name, Tavayza Тавайза, said to come from Chinese Dawaizi (or Daweizi)[3]. In characters, that would be most likely 大崴子; 崴 wǎi and 崴子 wǎizi, attributed a range of meanings from 'mountain bend' (山湾) to 'bay', is common in place names in Jilin as well as across the border. The (older) Chinese name for Vladivostok is 海参崴 Hǎishēnwǎi; Posyet Bay (залив Посьета) used to be called 摩阔崴 Mókuòwǎi. There's also the idiom 跑崴子 pǎo wǎizi, originally meaning 'to go to Vladivostok' (to trade, possibly in sea cucumbers).

The word 崴(子), at least in this sense, seems in turn to be a Manchu loan. My ignorance of Manchu is appalling, but after rummaging through dictionaries and bilingual texts[4] I found a noun wai and a derived adjective waiku, also occurring (reduplicated?) in waiku daikū and meaning 'askew' (could that in turn be a loan from 歪 wāi?).

Tavayza or Dawaizi could be a more auspicious name for visitors to the new casino than the current official name, which points either to undesirable insects or to the even less desirable count who helped deprive China of those lands. (Dialectically enough, the loss of those lands to Russia has meant that Chinese people can now legally gamble on them.)

So while the abundance of Chinese toponyms is a vivid reminder of Qing rule beyond the Amur, the apparent lack of a non-Russian-mediated name for the Lena (or at least of one known to Qing officials) might deny one motivation to those inclined to formulate a historical claim further into Siberia. Some are indeed so inclined. While, as I said, such views aren't reflected in an official position, some people have tried to articulate the idea that much of Eastern Siberia had been under Chinese rule since time immemorial, with arguments not unlike those used to back Chinese territorial claims elsewhere. Qi Jun 齐钧 of the Institute of Law of the Chinese Academy of Social Sciences has written an article[5] that defends the Qing claim to the lands up to the Lena (and indeed goes on to assert that China exerted "effective administration" over large swaths of Siberia as early as the Tang, and that Yuan explorers reached the Arctic).

The merits of such claims and their reception among Chinese academics go beyond the linguistic issues I wanted to bring up, so let me just say one more thing about the Lena river. It also has a Mongolian name, Зүлгэ Zülge (Buryat Зүлхэ). That looks quite different from the Russian and Tungusic names, and I have no idea where it might come from, or how early it might be attested. All I can say is it looks similar to the word зүлэг züleg (ǰülge in traditional script), meaning 'lawn' or 'turf'.

So I don't know if (a) the words fall into two groups, with the Lena name related to the front-grass group; (b) the Lena name is related to the back-ravine group, but just switched to front vowels; (c) the Lena name has nothing to do with any of the other words; (d) all the words are related, or come from unrelated etyma but the semantics got mixed over time.

[1] Translated as "Биография Лантаня (Biography of Langtan)", in Davydova M., Myasnikov B., eds., Русско-китайские отношения в XVII веке. Материалы и документы (Russian-Chinese relations in the 17th century. Materials and documents), vol. 2. Moscow, Nauka, 1972. Available online. The Baqi tongzhi also has a Manchu version that had been translated into Russian before the Chinese one.

[2] Kicengge 承志, "尼布楚條約界碑圖的幻影—滿文《黑龍江流域圖》研究 (The illusion of the Nerchinsk treaty boundary-stone: The Map of the Amur Region in Manchu)", The National Palace Museum Research Quarterly vol. 29, no. 1 (Autumn, 2011), pp. 147-236. Online here.

[3] Solovyov F., Словарь китайских топонимов на территории советского Дальнего Востока (Dictionary of Chinese toponyms in the territory of the Soviet Far East). Vladivostok, 1975. Online here.

[4] An occurrence online claims to be from a modern edition of the Qingwen zhiyao 清文指要 or Manju gisun-i oyonggo jorin-i bithe, a Manchu-Chinese phrasebook whose first version seems to date to 1789. Waiku 'crooked' also occurs in the translation of the Book of the Nishan Shaman serialised in the Echoes of Manchu blog.

[5] Qi Jun 齐钧, "《尼布楚条约》所涉以雅库为界初考 (A preliminary study of 'Yaku as the border' alluded at the Treaty of Nerchinsk)", in  Han Yanlong 韩延龙 ed., 法律史论集 (Studies on legal history), Beijing, 法律出版社, 2006. Available online in abridged form.

http://languagelog.ldc.upenn.edu/nll/?p=23057

It's the season of political speeches, and so I've been listening to a few. One thing that sticks out is Donald Trump's rhetorical style, which has some characteristics that I haven't observed in other politicians. In "Donald Trump's repetitive rhetoric" (12/5/2015) I noted his tendency to repeat words and phrases. This repetition means that many phrases are entirely predictable well before their end, and perhaps for that reason, he often leaves the last bits unspoken. And finally, he has an almost Pirahã-like ability  — or perhaps I should say Elmore Leonard-like ability — to express complex thoughts in paratactic form, with very little clausal embedding.

There are plenty of examples of these three characteristics from Mr. Trump's 12/21/2015 speech in Grand Rapids MI.

He starts the rally this way:

And he ends it like this:

In between, he tells a long story about various imagined interactions with the Ford Motor Company over their plan to build a new multi-billion-dollar plant in Mexico. Presumably because of NAFTA, which he doesn't mention, Ford will (he says) be able to import cars and trucks and parts from that plant without any tariff payments. His story presupposes that Ford needs the president's permission to build that plant, and that the president has the power to impose arbitrary company- and product-specific import duties as part of making a deal to give such permission. In this context, Trump presents his personal wealth and ability to finance his own campaign as an advantage.

You'll see some further repetitions; some examples of omitted redundant phrase-endings, and plenty of paratactic constructions.

After introducing the topic, Mr. Trump imagines Jeb Bush and Hillary Clinton interacting with the company:

I find it interesting, by the way, that there was a lot of reaction to Trump's use of the word "shlonged" in another part of the same speech, but little or no reaction to the accusation of corruption in that part that you've just heard.

Mr. Trump then goes on to describe his own imagined interaction with Ford — he of course deals directly with the president of the company:

http://languagelog.ldc.upenn.edu/nll/?p=23059

Listen to what the Chinese scout in this video says at :43.  My first impression was that it sounds like he is speaking Cantonese, not Mandarin.

He seems to be speaking some kind of non-standard Cantonese, maybe Toishan, and is saying something like this:

"I don't know why I hit an old man when I went back (to school?) today, I have no idea why the old man is like this, I don't know why he hit me when I went back (to school?), then I hit him, then I…."

Of course, that doesn't make much sense, but the language is fast and compressed, and it's hard to make out every syllable, not to mention that my Cantonese is not very good, much less my non-standard Cantonese.

I asked about thirty Cantonese speakers, and none of them could understand all of it, though they could roughly make out that it had something to do with hitting and an old man.

I then showed the cartoon to Bob Bauer, and here are the bits and pieces he could pick out after watching and listening to it carefully several times:

打咗個老人
daa2 zo2 go3 lou5 jan4
hit the old man

個老人              唔知點
go3 lou5 jan4 m4 zi1 dim2
the old man I don't know why/how

打噉樣
daa2 gam2 joeng6/2
hit like that

佢打我          我打佢
keoi5 daa2 ngo5 ngo5 daa2 keoi5
he hit me I hit him

Bob remarks:  "If my interpretation is accurate, then he could hardly be said to have done a good turn for the day. This must have been an inside joke for the Cantonese speakers."

I'm told that the sound track was made in the 30s in New York city.  At that time, nearly all Chinese there would have been Cantonese speakers.  It was also a time when there were few strictures on racial stereotypes.

———–

Last minute arrivals:

From Ashley Liu:

It is Cantonese, but it's also almost gibberish. It seems to be about the boy witnessing some 台山人 randomly beating up an old person on the street.

From Abraham Chan:

我講畀你聽吖先生，我唔知點喎，嗰啲番鬼人打個路人，呢個路人唔知點呀咁樣，啲番鬼人咁樣打佢，佢打我，我打佢，收尾我……

[h.t. John McWhorter; thanks to Nelson Ching, Ranting Jiang, Fangyi Cheng, and Alan Chin]