"One big Donald Trump AIDS"

Jun. 25th, 2017 02:34 pm
[syndicated profile] languagelog_feed

Posted by Mark Liberman

As I've observed several times over the years, automatic speech recognition is getting better and better, to the point where some experts can plausibly advance claims of "achieving human parity". It's not hard to create material where humans still win, but in a lot of ordinary-life recordings, the machines do an excellent job.

Just like human listeners, computer ASR algorithms combine "bottom-up" information about the audio with "top-down" information about the context — both the local word-sequence context and various layers of broader context. In general, the machines are more dependent than humans are on the top-down information, in the sense that their performance on (even carefully-pronounced) jabberwocky or word salad is generally rather poor.

But recently I've been noting some cases where an ASR system unexpectedly fails to take account of what seem like some obvious local word-sequence likelihoods. To check my impression that such events are fairly common, I picked a random youtube video from YouTube's welcome page — Bill Maher's 6/23/2017 monologue — and fetched the "auto-generated" closed captions.


Here's an example that combines impressive overall performance with one weird mistake:

5:07 Mitch McConnell says he wants a vote
5:10 before the 4th of July when Trump voters
5:13 traditionally blow their hands off
5:19 oh the fourth of July hey summers here
5:24 boy it was real Beach weather in Phoenix
5:26 the other day did you see that it was
5:28 122 122 plains could not take off hey
5:34 climate deniers
5:36 if melting IceCaps and rising oceans and
5:40 pandemics aren't enough to scare you not
5:42 being able to leave Phoenix that should
5:50 work

I'll give the machine a pass on "summers" instead of "summer's", and we can ignore the issue of "oh" vs. "ah", and forgive the hallucinated "work" at the end — but "plains could not take off"? In Psalm 114:4 the mountains skipped like rams, but not even then did the plains take off.

A bit later:

6:32 but speaking of solar Donald Trump broke
6:36 some news at the rally that the wall you
6:39 know the wall between us and Mexico it's
6:41 going to have solar panels on he said it
6:43 was his idea solar battles okay so the
6:47 wall which is never going to be built
6:49 which Mexico is never going to be paying
6:52 for which now has imaginary so propels
6:56 on because if it's one big Donald Trump
6:59 AIDS it's fake news

So the system got "solar panels" right the first time, but then heard "solar battles" and "so propels". In fairness, Maher kind of garbles the last one into something like "solar pels":

But still, I don't think anyone in the audience heard "so propels".

And then at the end, "if there's one thing Donald Trump hates it's fake news" get turned into "if it's one big Donald Trump AIDS it's fake news":

In that case, I don't hear any acoustic phonetic excuses. And surely "one thing Donald Trump hates" is a priori a more probable word string than "one big Donald Trump AIDS"…

I don't know which generation of ASR Google is using to generate YouTube captions. But it's possible that this sort of thing is an example of the sometimes-peculiar behavior of RNN language models.

This week's upcoming events...

Jun. 25th, 2017 04:00 pm
[syndicated profile] vintage_ads_feed

Posted by misstia

A reminder will be posted a day before each event

28 Wednesday ONE DAY EVENT: Reptiles. Ads that have reptiles

30-2 Friday - Sunday Weekend EventS: Poisons and Fingers & toes. Ads for poisons and/or that mention poison AND ads featuring fingers and/or toes!!

Korean War Veteran's Day

Jun. 25th, 2017 02:26 pm
[syndicated profile] vintage_ads_feed

Posted by misstia

Many thanks to our guest host bradygirl_12 for our Korean War Veteran's Day. We do this event, now yearly, because as she said, the Korean War is quite often forgotten. Sandwiched between WWII and the Vietnam War, it's just a blip in many history books and the veterans are also overlooked. Whether it be 6 months, 6 years, or 3 years, as the Korean War was (1950-1953), it should not be overlooked, nor should any veteran who served be forgotten.

Ad from 1952

6e717fd76b227fb4aa4e053747df77f0--coke-ad-coca-cola-ad

That ad, for some reason, automatically reminded me of this scene from Dr. Strangelove.

One Day Event: The Korean War

Jun. 25th, 2017 01:30 pm
[syndicated profile] vintage_ads_feed

Posted by bradygirl_12

Good morning!

Today is the 67th anniversary of the start of the Korean War ("Isn't this where we came in?" "It's even Sunday morning!"). And it was a Sunday that it began, as the film's dialogue attests. I am your guest host for this event..

One Minute To Zero stars Robert Mitchum and Ann Blyth in a film released in 1952, with a year to go in the Korean War. A mix of flag-waving and gritty realism, the movie is most noteworthy for actual combat scenes and a controversial scene involving Korean refugees. Howard Hughes, the owner of the studio (RKO), refused to delete the scene when requested by the U.S. Army.

25 Sunday ONE DAY EVENT: Korean War Veteran's Day. Any ad from the period of the Korean War. This war, and it's veterans, is often overlooked.

Daily Happiness

Jun. 25th, 2017 01:06 am
torachan: a cartoon owl with the text "everyone is fond of owls" (everyone is fond of owls)
[personal profile] torachan
1. The city fireworks show was tonight and it was sooooooo loud and really scared the kitties, but they've all come out of hiding now and seem to be back to normal. I gave them lots of treats when they finally came out.

2. McDonald's has these really tasty blueberry cream pies right now. Carla loves the strawberry ones, too, and has gotten them a bunch, but while I thought those were okay, I wasn't super into them. But these blueberry ones are so good! And it's weird, because I usually like strawberry more than blueberry, but idk. The blueberry one is so much better.

3. Everybody loves this box so much. It's got nice flaps to make you feel hidden, and rustly paper inside to play with. Just the best box. Three out of three kitties recommend.

Renewal of the race / nation

Jun. 25th, 2017 02:53 am
[syndicated profile] languagelog_feed

Posted by Victor Mair

Jamil Anderlini in the Financial Times (6/21/17), "The dark side of China’s national renewal", writes:

To an English-speaking ear, rejuvenation has positive connotations and all nations have the right to rejuvenate themselves through peaceful efforts.

But the official translation of this crucial slogan is deeply misleading. In Chinese it is “Zhonghua minzu weida fuxing” and the important part of the phrase is “Zhonghua minzu” — the “Chinese nation” according to party propaganda. A more accurate, although not perfect, translation would be the “Chinese race”.

That is certainly how it is interpreted in China. The concept technically includes all 56 official ethnicities, including Tibetans, Muslim Uighurs and ethnic Koreans, but is almost universally understood to mean the majority Han ethnic group, who make up more than 90 per cent of the population.

The most interesting thing about Zhonghua minzu is that it very deliberately and specifically incorporates anyone with Chinese blood anywhere in the world, no matter how long ago their ancestors left the Chinese mainland.

“The Chinese race is a big family and feelings of love for the motherland, passion for the homeland, are infused in the blood of every single person with Chinese ancestry,” asserted Chinese premier Li Keqiang in a recent speech.

This is a highly perceptive, and troubling, article that merits reading in its entirety.

In this post, I will focus on some key terms.

First of all, front and center, what is this mínzú 民族?  It can mean lots of things:  nation, nationality, people, ethnic group, race, volk.  This is not the first time that mínzú 民族 has erupted on the international stage.  One of the most notable instances was four years ago, emanating right here from the University of Pennsylvania.  The incident is well recounted by R.L.G. in "Johnson" at The Economist (5/21/13), "Of nations, peoples, countries and mínzú:  Differing terms for ethnicity, citizenship and group belonging ruffle feathers":

DID Joe Biden insult China?  The American vice-president has a habit of sticking his foot into his mouth, and in this case, the recent graduation speech he gave at the University of Pennsylvania inspired a viral rant by a "disappointed" Chinese student at Penn, Zhang Tianpu. What was Mr Biden's sin? Was it Mr Biden's suggestion that creative thought is stifled in China?

You cannot think different in a nation where you cannot breathe free. You cannot think different in a nation where you aren't able to challenge orthodoxy, because change only comes from challenging orthodoxy.

No, that wasn't it.

The source of the insult is a surprising one: Mr Biden called China a "great nation", and a "nation" repeatedly after that. Victor Mair, the resident sinologist at the Language Log blog, translates Mr Zhang's complaint.

In this sentence, "You CANNOT think different in a nation where you aren't able to challenge orthodoxy", he used the word "nation". This is what really infuriated me, because in English "nation" indicates "race, ethnicity", which is different from "country, state". "Country, state" perhaps places more emphasis on the notion of the entirety of the country, even to the point of referring to the idea of government.

Mr Mair explains:

The weakness in Zhang's reasoning lies mainly in his confusion over the multiple meanings of the word mínzú 民族…. [M]ínzú 民族 can mean "ethnic group; race; nationality; people; nation".  Coming from the English side, we must keep in mind that "nation" can be translated into Chinese as guó 国 ("country"), guójiā 国家 ("country"), guódù 国度 ("country; state"), bāng 邦 ("state"), and, yes, mínzú 民族 ("ethnic group; race; nationality; people; nation").

It is clear that, when Biden said "China is a great nation", he was respectfully referring to the country as a whole.  Yet the sensitivity to questions of ethnicity in China, especially with regard to the shǎoshù mínzú 少数民族 ("ethnic / national minorities"), e.g., Uyghurs, Tibetans, and scores of others, caused Zhang to take umbrage over something that the Vice President never intended.

In a later post about smartphone zombies, Cant. dai1tau4 zuk6 / MSM dītóu zú 低頭族 (“head-down tribe”), "Tribes" (3/10/15), I wrote:

The first word I think of when I see 族 as a suffix is Mandarin mínzú, Japanese minzoku 民族 (“nation; nationality; people”), which is formed from 民 (“people; subjects; civilians”) + 族 (“family clan; ethnic group; tribe”).  The term is a neologism coined in the late 19th century by Japanese thinkers to match the Western (especially German) concept of “nation”.

… I have assembled a large amount of material concerning the absence of mínzú / minzoku 民族 as a lexical item corresponding to “nation” in China before it was introduced from Meiji [1868-1912] Japan.

When we prefix mínzú 民族 with shǎoshù 少数 ("few; small number; minority"), we have shǎoshù mínzú 少数民族 ("minority; national minority; ethnic minority").  Here it gets really tricky, because, as Anderlini points out in his article, there are officially 56 ethnic groups (mínzú 民族) in China, of which 55 are shǎoshù mínzú 少数民族 ("minorities; national minorities; ethnic minorities; ethnic groups"), with the 56th being the dominant, majority (over 90%) Hàn mínzú 汉民族 ("Han nationality; Han ethnic group").  Consequently, when Chinese politicians talk about the blood of the Chinese race, it's important to know whether they are are referring to Hàn mínzú 汉民族 ("Han nationality; Han ethnic group"), Zhōnghuá mínzú 中华民族 ("Chinese nation / people", where Zhōnghuá 中华 is understood as "Central cultural florescence"), or something else.  In each case, we need to judge carefully whether they meant to include all the ethnicities within the sovereign territory of the PRC or in the whole world, or whether they were referring specifically to individuals of Han ethnicity within the sovereign territory of the PRC or in the whole world.  Often, for politicians, as for poets, ambiguity is desirable, or at least convenient.

There are no less than half a dozen other words for "(the) people" that are in common use in Mandarin.  I won't go into all of them here, but will mention only one:  rénmín 人民, as in rénmínbì 人民币 ("RMB; people's currency") and Rénmín rìbào 人民日报 ("People's Daily").  This term, rénmín 人民, does not get involved with race, ethnicity, nation, and so on, but emphasizes the population as a whole.

As for "Zhongguo / China", that too is a huge can of worms, for which see this incisive paper by Arif Dirlik:

"Born in Translation: 'China' in the Making of 'Zhongguo'"

[h.t. John Rohsenow, Bill Bishop]

Bruria Kaufman

Jun. 24th, 2017 04:53 pm
[syndicated profile] languagelog_feed

Posted by Mark Liberman

The Annual Reviews have a tradition of featuring retrospective articles by or about senior figures, and the Annual Review of Linguistics has followed this pattern with pieces featuring Morris Halle in the 2016 volume and Bill Labov in 2017. For 2018, we'll be featuring Lila Gleitman.

As background, Barbara Partee, Cynthia McLemore and I spent the last couple of days interviewing Lila about her life and work. We've got more than 7.5 hours of recordings, which is more like a book than an article — and it may very well turn into a book as well, with edited interview material interspersed with reprints of Lila's papers. But what I want to post about today is one of the many things that I learned in the course of the discussions. This was just a footnote in Lila's life story, but it has its own intrinsic interest, and I'm hoping that some readers will be able to provide more information.

I learned that the founder of the Penn Linguistics Department, Zellig Harris, was married to a mathematical physicist named Bruria Kaufman. She worked with John von Neumann, wrote some widely-cited papers on crystal statistics in the late 1940s, published with Albert Einstein (Albert Einstein and Bruria Kaufman. "A new form of the general relativistic field equations", Annals of Mathematics, 1955), and later wrote papers like "Unitary symmetry of oscillators and the Talmi transformation", Journal of Mathematical Physics 1965, and "Special functions of mathematical physics from the viewpoint of Lie algebra", Journal of Mathematical Physics 1966.

The thing that interested me most was that Bruria Kaufman also worked for a while in the 1950s with Harris at Penn, at the same time as others including Lila Gleitman, Aravind Joshi, R.B. Lees, Naomi Sager, Zeno Vendler, and Noam Chomsky. And according to this 1961 NSF report, her contributions included Transformations and Discourse Analysis Papers (TDAP) numbers 19 and 20:

19. Higher-order Substrings and Well-formedness, Bruria Kaufman.
20. Iterative Computation of String Nesting (Fortran Code), Bruria Kaufman.

I've found a couple of citations to these works, but so far not the works themselves.

The 1961 NSF report says that

Paper 15 gives an information [sic — should be informal?] presentation of a general theory and method for syntactic recognition. Papers 16-19 give the actual flow charts of each section of the syntactic analysis program.

where 15-19 are

15. Computable Syntactic Analysis, Zellig S. Harris. (Revised version published as PoFL I, above)
16. Word and Word-Complex Dictionaries, Lila Gleitman.
17. Elimination of Alternative Classifications, Naomi Sager.
18. Recognition of Local Substrings, Aravind K. Joshi.
19. Higher-order Substrings and Well-formedness, Bruria Kaufman.

and "PoFL I" is Harris's String Analysis and Sentence Structure, 1962.

Aravind Joshi and Phil Hopely, "A parser from antiquity", Natural Language Engineering 1996, explains that

A parsing program was designed and implemented at the University of Pennsylvania during the period from June 1958 to July 1959. This program was part of the Transformations and Discourse Analysis Project (TDAP) directed by Zellig S. Harris. The techniques used in this program, besides being influenced by the particular linguistic theory, arose out of the need to deal with the extremely limited computational resources available at that time. The program was essentially a cascade of finite state transducers (FSTs).

More on the history from that source:

The original program was implemented in the assembly language on Univac 1, a single user machine. The machine had acoustic (mercury) delay line memory of 1000 words. Each word was 12 characters/digits, each character/digit was 6 bits. Lila Gleitman, Aravind Joshi, Bruria Kauffman, and Naomi Sager and a little later, Carol Chomsky were involved in the development and implementation of this program. A brief description of the program appears in Joshi 1961 and a somewhat generalized description of the grammar appears in Harris 1962.  This program is the precursor of the string grammar program of Naomi Sager at NYU, leading up to the current parsers of Ralph Grishman (NYU) and Lynette Hirschman (formerly at UNISYS, now at Mitre Corporation). Carol Chomsky took the program to MIT and it was used in the question-answer program of Green, BASEBALL (1961). At Penn, it led to a program for transformational analysis (kernels and transformations) (1963) and, in many ways, influenced the formal work on string adjunction (1972) and later tree-adjunction (1975).

The paper's bibliography cites

Transformations and Discourse Analysis Project (TDAP) Reports, University of Pennsylvania, Reports #15 through #19, 1959-60. Available in the Library of the National Institute of Science and Technology (NIST) (formerly known as the National Bureau of Standards (NBS)), Bethesda, MD.

So I'll ask my friends at NIST if these works are still there.

 

Weekend Event: Cheese

Jun. 24th, 2017 11:32 am
[syndicated profile] vintage_ads_feed

Posted by beaver67

Must have been a good recipe; it was cut out of the magazine from which I scanned the ad.

Weekend Event: Cheese

Jun. 24th, 2017 11:31 am
[syndicated profile] vintage_ads_feed

Posted by beaver67

Large, double-page spread. Click picture for close-up for easier reading.

Daily Happiness

Jun. 24th, 2017 01:11 am
torachan: tavros from homestuck dressed as pupa pan (pupa pan)
[personal profile] torachan
1. I finished another book tonight, which makes thirty-three for the year so far. I have moved my initial goal from twenty to thirty and then to forty, but I'm thinking I'm going to have to move it again at this rate, because the year's only halfway through.

2. I spoke with the insurance claim person today and she said that it sounded like since we were both backing up, it would be a 50/50 thing, which would mean our insurance wouldn't go up and we would be liable for half the deductible. So, possibly still have to pay a lot in car repairs, depending on how much it costs to repair the damage, but it could be worse.

3. Look at this fluffy Chloe tum! She just loves the tummy scritches. :)

Chinglish with tones

Jun. 23rd, 2017 07:57 pm
[syndicated profile] languagelog_feed

Posted by Victor Mair

4th tone – 3rd tone, it would appear:

Well, maybe not; the diacritics are probably meant to indicate vowel quality, but I don't know what system (if any) they are using.

Ben Zimmer writes:

The diacritics may be intended to evoke pinyin tone marks, but they're also reminiscent of dictionary-style phonetic respelling and stress marking. The grave accent on "ì" could be intended as an indicator of primary stress, though that's more typically marked with an acute accent. And the breve on the "ĭ" is a common enough way to represent /ɪ/ (the macron is used for long vowels and the breve for short vowels — see, e.g., Phonics on the Web). But this use of diacritics as typographical ornamentation is never very consistent — recall the styling of the play Chinglish as "Ch’ing·lish”.

The illustration appears at the top of this article:

It turns out that the image used by the People's Daily originally appeared as a promotion for the play Chinglish that Ben mentioned, specifically for its performance by the Singaporean theater company Pangdemonium in 2015. See the Pangdemonium website, as well as local coverage by PopSpoken and Today. So the People's Daily may have searched for a "Chinglish" image online and borrowed this one, without giving proper credit. (Credit should go to Olivier Henry of MILK Photographie.)

The six individuals in the picture seem to be aspiring to some idealized form of Chinglish in the sky above, overlying the cloud shrouded five star design of the Chinese flag, leading them on.  The thrust of the People's Daily article, however, is anything but adulatory of Chinglish:

Chinese authorities on June 20 issued a national standard for the use of English in the public domain, eradicating poor translations that damage the country’s image.

The standard, jointly issued by China’s Standardization Administration and General Administration of Quality Supervision, Inspection and Quarantine, aims to improve the quality of English translations in 13 public arenas, including transportation, entertainment, medicine and financial services. It will take effect on Dec. 1, 2017.

According to the standard, English translations should prioritize correct grammar and a proper register, while rare expressions and vocabulary words should be avoided. The standard requires that English not be overused in public sectors, and that translations not contain content that damages the images of China or other countries. Discriminatory and hurtful words have also been banned. The standard provided sample translations for reference, and warned against direct translation.

There are perpetual plans for eliminating Chinglish in China, but they are unlikely ever to materialize unless professional translators are sought after for their expertise and paid accordingly.

Earlier calls for the elimination of English more generally are no longer heard from responsible persons:

Now the goal is more reasonably just to get rid of Chinglish, but that will not happen on December 1, 2017 when the new standards go into effect.  Although it will take many years for their full implementation and realization, the standards are admirable goals to aim for.

See also:

[h.t. Jim Fanell, Toni Tan]

Smoking Guns

Jun. 23rd, 2017 07:31 pm
[syndicated profile] vintage_ads_feed

Posted by pigshitpoet

Does Cigarette Smoking Cause Cancer?

Smoking gun

Wouldn't you rather have a Marlboro ?


Harley Davidson and the Marlboro Man (1991)



" Harley, if you were shootin' for shit you wouldn't get a whiff! " - Marlboro

.

Life Magazine Miami Fashion

Jun. 23rd, 2017 01:42 pm
[syndicated profile] vintage_ads_feed

Posted by cuddyclothes

This isn't exactly advertising, but I thought it warranted being posted here:

LIFE magazine March 14, 1955 fashion feature: MIAMI MOVES UP IN STYLE—Breezy clothes for warm weather come from a likely source. “Newest of the Florida booms is in the field of fashion. With its fashion industry grown 10 times in size since 1945, Miami now has 125 firms manufacturing women’s apparel, last year sold $55 million worth of clothes for warm weather at any season of the year.”

1950s womens fashion

Ask Language Log: "assuage"

Jun. 23rd, 2017 11:41 am
[syndicated profile] languagelog_feed

Posted by Mark Liberman

Query from a reader:

Is it correct to use the word assuage to indicate a lessening of something? That is, it is often used in the realm of feelings, i.e. assuage hunger, assuage grief, etc. But would it be acceptable to use to indicate the lessening of something more tangible, such as assuage criminality, assuage the flow of water, assuage drug use.

I probably wouldn't use assuage to describe the lowering of flood waters or the amelioration of traffic jams. But I don't have any special standing to rule on such matters, so as usual, let's look at how others use the word.

The OED's entry for assuage, which is flagged as "not yet … fully updated (first published 1885)", has several senses marked as "arch. or Obs." that don't involve "angry or excited feelings", or beings in such a state.

There's the transitive form glossed "To abate, lessen, diminish (esp. anything swollen)", with examples like

1774   J. Bryant New Syst. II. 284   The Dove..brought the first tidings that the waters of the deep were asswaged.

There's the intransitive inchoative version of the same, glossed "To grow less, diminish, decrease, fall off, die away; to abate, subside", with examples like

1611   Bible (King James) Gen. viii. 1   And the waters asswaged .

COCA has 509 instances of "assuage", 134 of "assuaged", 46 of "assuaging", and 17 of "assuages". Looking at a random sample of 100, we find that all 100 are transitive, and that in 98 of them, what's assuaged is an negatively-evaluated emotion or feeling or concern ("the community's grief", "his guilt", "such mortal concerns", "the twitchy sensation in my cells", "white opposition to slave conversion", "my hunger", "Democratic anxieties", "India's complaints", "feelings of humiliation", the monarch's fears", "his own damaged pride", "the egos of movie stars", "my curiosity", …), or an person or group of people subject to such emotions or feelings or concerns ("his uneasy party", "the academic intellectual community", "the larger man", "international critics of the war", "his jittery passenger", "the chiefs", "the dealers", …).

The two exceptions in the sample are these:

In The Efficiency Trap, Steve Hallett claims that we will exhaust many of our resources by the 2030s, and violence and chaos will erupt as a result. Hallett proposes recycling and growing food locally as possible means of assuaging the damage.

The measure, which awaits Senate approval of a minor amendment next week, can not assuage the impending disaster that will kill virtually all the fish in the Dolores River this summer.

With respect to the specific examples in the query, Google finds

"assuage criminality": one example [link] Please reconsider your gig – don't play for a segregated audience in Israel and make of yourself a balm to assuage criminality.

"assuage the flow of water": no examples (though see biblical examples cited by the OED)

assuage drug use: one example [link] Becker's neoliberal drug policy presumes to assuage drug use and addiction by the instantiation of a highly regulated market as a system of control.

So the verdict of norma loquendi seems to be that applying assuage to things other than people and their feelings is out of fashion and currently marginal.

 

Daily Happiness

Jun. 23rd, 2017 01:06 am
torachan: nepeta from homestuck (nepeta)
[personal profile] torachan
1. Went out to lunch with my mom today for my birthday (which is actually Monday). We went to California Pizza Kitchen and I got the carne asada pizza, which was delicious.

2. I got in a car accident on the way home, which sucks a whole lot, but it was on the minor side of accidents, so I'm trying to keep positive about that. There's only a little bit of bumper damage to my car and the other car and no one was hurt. (I had just pulled in a parking space and was readjusting my parking, and the other person was pulling out of her space opposite me, and...crunch.) I'm hoping the repairs don't cost too much, are covered by the insurance, and don't make my payments go up too high. *crosses fingers*

3. Jasper!