Language Log

Lisa Feldman Barrett, "Zombie ideas", Observer 10/2019:

It’s October, a month auspicious for All Hallow’s Eve and everything spooky. Accordingly, our topic for this month is … zombies. Not the charmingly decayed corpses you encounter in movies and books, but zombie ideas. According to the economist Paul Krugman (2013), a zombie idea is a view that’s been thoroughly refuted by a mountain of empirical evidence but nonetheless refuses to die, being continually reanimated by our deeply held beliefs. […]

If you think that formal science training will zombie-proof your mind, you’re out of luck, my friend. Hordes of zombie ideas flourish in science (Brockman, 2015). They also fester in our own field, quietly biding their time in peer-reviewed papers and textbooks, waiting to infect another generation of unsuspecting psychological scientists.

"Beijing reveals mascots for 2022 Winter Games", YouTube (9/19/19):

Yiming Wang et al., "Espresso: A fast end-to-end neural speech recognition toolkit", ASRU 2019:

We present ESPRESSO, an open-source, modular, extensible end-to-end neural automatic speech recognition (ASR) toolkit based on the deep learning library PyTorch and the popular neural machine translation toolkit FAIRSEQ. ESPRESSO supports distributed training across GPUs and computing nodes, and features various decoding approaches commonly employed in ASR, including look-ahead word-based language model fusion, for which a fast, parallelized decoder is implemented. ESPRESSO achieves state-of-the-art ASR performance on the WSJ, LibriSpeech, and Switchboard data sets among other end-to-end systems without data augmentation, and is 4–11× faster for decoding than similar systems (e.g. ESPNET)

Article in SCMP Magazine:

"How Hong Kong slang terms for ‘police’ have evolved over time", by Lisa Lim (9/28/19):

Back in the day, Hong Kong policemen were referred to in Cantonese as luhky ī  [sic; VHM: luk6ji1 綠衣]  (“green clothing”), for the green uniforms they had worn since the 19th century. Khaki drill became the summer uniform around 1920 while the current get-up of light-blue shirt and black trousers, worn year-round, was adopted in December 2004.

Ben Zimmer, "How Maguire Accidentally Made the Case for Singular ‘They'", The Atlantic 9/27/2019 (subhead: "The national intelligence director’s recent testimony inadvertently supported the argument against grammar purists"):

When the committee chairman, Adam Schiff, asked Maguire if he thought that the whistle-blower was “a political hack” as Trump had suggested, Maguire responded, “I don’t know who the whistle-blower is, Mr. Chairman, to be honest with you. I’ve done my utmost to protect his anonymity.” But if Maguire was seeking to protect the whistle-blower’s anonymity, why use the pronoun he to identify the person’s gender?

Schiff, in his questioning, was more circumspect, avoiding gendered references by relying on a time-honored strategy: deploying they as a singular pronoun. When Maguire said he thought the whistle-blower was “operating in good faith,” Schiff said, “Then they couldn’t be in good faith if they were acting as a political hack, could they? … You don’t have any reason to accuse them of disloyalty to our country or suggest they’re beholden to some other country, do you?”

Here's a simply titled article from China Daily:

"Opening up of financial market continues" (9/26/19).

The article may have a plain title, but it is full of gibberish.  The concluding sentence takes the cake:

Therefore, the securities market as the focus of the internationalization of the entire financial market must be targeted in order to truly realize the internationalization of the market and the internationalization of finance.

Yesterday, Merriam-Webster tweet-teased Donald Trump over a couple of glosses:

'Transcript': 🎙 a typed copy of dictated or recorded material

'Memorandum': 📝 an informal report or message

— Merriam-Webster (@MerriamWebster) September 25, 2019

[This is a guest post by Douglas Q. Adams]

Readers of this blog may remember the excitement generated a few months ago by the announcement that “Tocharian C,” the native language of Kroraina (Chinese Loulan) had been discovered, hiding, as it were, in certain documents written in the Kharoṣṭhī script ("Tocharian C: its discovery and implications" [4/2/19]). Those documents, with transcription, grammatical sketch, and glossary, were published earlier this year as a part of Klaus T. Schmidt’s Nachlass (Stefan Zimmer, editor, Hampen in Bremen, publisher).  However, on the weekend of September 15^th and 16^th a group of distinguished Tocharianists (led by Georges Pinault and Michaël Peyrot), accompanied by at least one specialist in Central Asian Iranian languages, languages normally written in Kharoṣṭhī, met in Leiden to examine the texts and Schmidt’s transcriptions.  The result is disappointing, saddening even.  In Peyrot’s words, “not one word is transcribed correctly.”  We await a full report of the “Leiden Group” with a more accurate transcription and linguistic commentary (for instance, is this an already known Iranian or Indic language, or do the texts represent more than one language, one of which might be a Tocharian language?). Producing such a report is a tall order and we may not have it for some little time.  But, at the very least, Schmidt’s “Tocharian C,” as it stands, has been removed from the plane of real languages and moved to some linguistic parallel universe.

"Teens charged with Qld arsenal 'completely despicable'", Sky News 9/11/2019.

For a natural demonstration of what diglossia is in the Chinese-speaking context, watch this 0:53 video.  The speaker begins in local Xi'anese (also called Guānzhōng huà 关中话 / 關中話), but at 0:20, when he suddenly realizes that he is talking to a television reporter, after hilariously sprucing himself up a bit, he abruptly switches to Modern Standard Mandarin (MSM):

https://www.youtube.com/watch?v=MKwjpGVaf4Q

During the 2015 presidential campaign, Geoff Pullum wrote about "Trump's aphasia", and I responded ("Trump's eloquence") that

[I]n my opinion, he's been misled by a notorious problem: the apparent incoherence of much transcribed extemporized speech, even when the same material is completely comprehensible and even eloquent in audio or audio-visual form.

This apparent incoherence has two main causes: false starts and parentheticals. Both are effectively signaled in speaking — by prosody along with gesture, posture, and gaze — and therefore largely factored out by listeners. But in textual form, the cues are gone, and we lose the thread.

Last Friday, an Australian journalist complained about the same sort of thing (Lenore Taylor, "As a foreign reporter visiting the US I was stunned by Trump's press conference", The Guardian 9/20/2019). The sub-head: "Despite being subjected to a daily diet of Trump headlines, I was unprepared for the president’s alarming incoherence."

She's talking about a recent tour of border-wall construction at Otay Mesa in California, and she summarizes her reactions this way:

In writing about this not-especially-important or unusual press conference I’ve run into what US reporters must encounter every day. I’ve edited skittering, half-finished sentences to present them in some kind of consequential order and repeated remarks that made little sense.

In most circumstances, presenting information in as intelligible a form as possible is what we are trained for. But the shock I felt hearing half an hour of unfiltered meanderings from the president of the United States made me wonder whether the editing does our readers a disservice.

From Fuchsia Dunlop’s Facebook page (taken in Xi'an):