Language Log

Read here now: the fine profile of my friend and research collaborator Julie Washington in the April issue of the Atlantic magazine. It’s been out for a while but you might not have seen it if, as in Madison WI where I live, it’s still February (we had the biggest snowstorm of the season this week). Julie is a professor at Georgia State University and the head of their program in Communication Sciences and Disorders. She’s an expert on the structure, acquisition, and use of African American English (AAE), and her research focuses on how use of the dialect affects reading achievement and educational progress, the assessment of children’s language and reading, and the identification of developmental language and reading disorders. The article describes her view that children who speak AAE in the home and community will make better progress in learning to read, and in school, if they can code switch between AAE and the mainstream dialect, often termed (though not by her) "standard" American English.

No, not polyvinyl chloride. This is the (first run of the) Prosody Visualization Challenge, organized by Agnieszka Wagner, Anne Cutler, and  Grażyna Demenko, in association with Speech Prosody 2018.

From Perry Link:

Jeff DeMarco saw this sign in Chengdu:

The International Phonetic Alphabet is a useful invention, which everyone interested in speech sounds should learn. But it's much less useful for actually doing phonetics than you might think. Whenever this comes up in discussion, I'm reminded of the Dr. Seuss classic On Beyond Zebra:

In the places I go there are things that I see
That I never could spell if I stopped with the Z.
I'm telling you this 'cause you're one of my friends.
My alphabet starts where your alphabet ends!

In "More Cohen Businesses Coming to Light," on Talking Points Memo, Josh Marshall writes:

The biggest taxi operator in New York, Evgeny “Gene” Friedman, now manages Cohen’s 30+ NYC medallions or at least did the last time we spoke to him. Friedman has been struggling for the last year to keep his taxi businesses out of bankruptcy and himself jail.

The final three words of the boldfaced clause present a weird, and dare I say unusual, case of double ellipsis. The semantic content communicated by those three words (in the context of the sentence) is richer than you'd think could be expressed by only three words, especially given that one of them is merely the conjunction and. That content can be represented as follows, with the struck-through text standing for the content that the reader must infer:

Friedman has been struggling for the last year to keep his taxi businesses out of bankruptcy and to keep himself out of jail.

There's nothing unusual about the first omission; I don't see anything wrong with the clause to keep his taxi businesses out of bankruptcy and himself out of jail. But the omission of out of strikes me as very strange, and what's even stranger is that to my ear, the clause is worse if to keep is put back:

* Friedman has been struggling for the last year to keep his taxi businesses out of bankruptcy and to keep himself jail.

[h/t Three Quarks Daily via Bob Shackleton]

Lessons here.
 

By coincidence, today's email brought two contributions of links to remarkable examples of PP-attachment ambiguity.

The first one was the lede from this story — Jason Rosenbaum & Marshal Griffin, "Hawley: Evidence exists to charge Greitens for obtaining charity fundraising list", St. Louis Public Radio, 4/18/2018:

Attorney General Josh Hawley is asking the St. Louis circuit attorney to file criminal charges against Gov. Eric Greitens for allegedly illegally obtaining a fundraising list from a charity he founded for political purposes.

It took me a couple of re-reading to clarify the point that Mr. Greitens obtained the list for political purposes, not that he founded the charity for political purposes.

And in this headline, it's the man who was charged, not the woman he shot: "Man out of jail after 16 months for shooting Nashua woman charged with vicious beating of new girlfriend", NHangle.com 4/16/2018

[h/t John Lawler and Mark Mandel]

Zeyao Wu sent in this video of French politicians pronouncing Xi Jinping's name:

Zeyao tells me that her Chinese friends who hear them have no idea what they're saying.

From Charles Belov:

My Apple Music subscription served me a folk-pop hip-hop song "Yibin BBQ" by Yishi Band at the tail end of a playlist mostly made up of rock from the former Yugoslavian republics.

Googling this band reveals that they sing in a dialect called Yibin.

I thought I heard a final consonant stop at 0:57-58 and 1:10 but I imagine that's a mishearing as the Wikipedia entry for Sichuan dialect does not list any consonant stops as possible finals. Also, as someone who doesn't know Mandarin, I fear this could be standard Mandarin without my knowing it. That said, when I try to match the first few words, what they rap doesn't quite match the printed lyric, and in particular, the character for the number one appears in the printed lyric and I'm hearing something that sounds like the number one in Cantonese and not in standard Mandarin.

(I took three semesters of Cantonese but never became fluent.)

I couldn't find this on YouTube and hope you either have streaming or know someone who can stream this for you.  Hope you can find and enjoy this.

That's the abbreviated title of a popular webcomic by Lin Chao 林超.  The full title in Chinese is Nà nián nà tù nàxiē shì 那年那兔那些事 (lit., "that year that rabbit those affairs"; i.e., "The story of that rabbit that happened in that year")

From the beginning of the Wikipedia article:

The comic uses animals as an allegory for nations and sovereign states to represent political and military events in history. The goal of this project was to promote nationalistic pride in young people, and focuses on appreciation for China's various achievements since the beginning of the 20th century.

The First DIHARD Speech Diarization Challenge has  results!

"Diarization" is a bit of technical jargon for "figuring out who spoke when". You can read more (than you probably want to know) about the DIHARD challenge from the earlier LLOG post ("DIHARD" 2/13/2018) the DIHARD overview page, the DIHARD data description page, our ICASSP 2018 paper, etc.

This morning's post presents some evidence from the DIHARD results showing, unsurprisingly, that current algorithms have a systematically higher error rate with shorter speech segments than with longer ones. Here's an illustrative figure:

For an explanation, read on.

Readers of Language Log will be thoroughly familiar with "topolect", since it is one of our regular categories (see, for example, here, here, here, here, here, and especially here).  Imagine my delight when I received from Neil Kubler the following photograph of a label in an ethnographical museum in China: