Language Log

Jenny Kleeman, "‘You tried to tell yourself I wasn’t real’: what happens when people with acute psychosis meet the voices in their heads?" The Guardian, 10/29/2024: "In avatar therapy, a clinician gives voice to their patients’ inner demons. For some of the participants in a new trial, the results have been astounding."

I learned about early trials of this idea about 15 years ago from Mark Huckvale, who developed the voice-morphing technology that allows a therapist to sound like (one of) the hallucinated speakers, through a dashboard that looks like this:

And Mark is one of the authors of the 2018 paper that  the Guardian article leads with: Tom Craig et al., "AVATAR therapy for auditory verbal hallucinations in people with psychosis: a single-blind, randomised controlled trial". The Lancet January 2018.

An outstanding Chinese scholar of Central Asian art history and archeology told me that any mention of Tūjué / Tújué 突厥 online or on social media would be subject to censorship by the authorities in the PRC.  Since Tūjué  突厥 is an important early name of the Turks, that makes it hard to do serious, honest research on the history of the Turkic peoples in Chinese.

Tūjué  突厥

Etymology

Ultimately from a form which also gave rise to the name Türk (cf. (Türük)), but the phonetics are difficult to reconcile.

It has been suggested that this is a transcription of Rouran *türküt, a plural of the Mongolic type, composed of *türk +‎ *-üt (cf. Khalkha Mongolian -үүд (-üüd)) (Pelliot, 1915). Pulleyblank (1965) proposed that this is a direct transcription of Türk.

Middle Sinitic (ca. 600 AD):  thwot kjut

(Wiktionary)

Reading through Étienne de La Vaissière's massive magnum opus, Asie Centrale 300-850:  Des routes et des royaumes (2024), I came to a screeching halt when my gaze alighted on this photograph (III.6, p. 71):

Limestone relief of Saluzi ("Autumn Dew"), one of the Six Steeds of Zhao Mausoleum, along with an unknown human general. The general switched horses with the emperor and cared for Saluzi; he is seen here pulling an arrow out of Saluzi's chest. On display at the Penn Museum in Philadelphia, Pennsylvania. (source)

Please do not check in a dictionary or online before you try to pronounce the name just by looking at it.

We have regional varieties of English:  Australian, American (with many subvarieties), Indian (South Asian), and so forth.  Cantonese is spread all around the world, especially in Southeast Asia, so it is not surprising that it has also developed its own regional variants.  In this post, we will concentrate on a comparison of Hong Kong and Malaysian Cantonese.

"Lost in communication:  Just because we speak Cantonese doesn’t mean we can understand each other", by Mandy Li, The Hong Konger (16 October 2024)

Mandy Li remembers the first time she worked with a Malaysian colleague:

In Malaysia, a sizeable portion of the population have Cantonese heritage so can speak the language. They also enjoy watching Cantonese dramas. So, when my colleague learned I was from Hong Kong, she naturally switched to Cantonese when speaking to me. I was astonished to find that I could not understand everything she said.

Whisper is a widely-used speech-to-text system from OpenAI — and it turns out that generative AI's hallucination problem afflicts Whisper to a surprisingly serious extent, as documented by Allison Koenecke, Anna Seo Gyeong Choi, Katelyn X. Mei, Hilke Schellmann, and Mona Sloane,"Careless Whisper: Speech-to-Text Hallucination Harms", In The 2024 ACM Conference on Fairness, Accountability, and Transparency,  2024:

Abstract: Speech-to-text services aim to transcribe input audio as accurately as possible. They increasingly play a role in everyday life, for example in personal voice assistants or in customer-company interactions. We evaluate Open AI’s Whisper, a state-of-the-art automated speech recognition service outperforming industry competitors, as of 2023. While many of Whisper’s transcriptions were highly accurate, we find that roughly 1% of audio transcriptions contained entire hallucinated phrases or sentences which did not exist in any form in the underlying audio. We thematically analyze the Whisper-hallucinated content, finding that 38% of hallucinations include explicit harms such as perpetuating violence, making up inaccurate associations, or implying false authority. We then study why hallucinations occur by observing the disparities in hallucination rates between speakers with aphasia (who have a lowered ability to express themselves using speech and voice) and a control group. We find that hallucinations disproportionately occur for individuals who speak with longer shares of non-vocal durations—a common symptom of aphasia. We call on industry practitioners to ameliorate these language-model-based hallucinations in Whisper, and to raise awareness of potential biases amplified by hallucinations in downstream applications of speech-to-text models.

Today's Tank McNamara:

According to the NFL, a "hip drop tackle" "occurs when a defender wraps up a ball carrier and rotates or swivels his hips, unweighting himself and dropping onto ball carrier’s legs during the tackle". And I would have more or less guessed that meaning, before getting the authoritative definition.

A Sandwich Helix, on the other hand…

George Kingsley Zipf is famous for his work on the power-law distribution of word frequencies, which has come to be known as Zipf's Law. And he's also known for the related "Law of Abbreviation", and the hypothesized balance between effort and efficacy.

In his 1945 paper "The repetition of words, time-perspective, and semantic balance", Zipf looks at a different distribution, which is much less famous:

In the present study we shall attempt to show in preliminary outline how the rate of repetition of words in the stream of speech may be useful not only in indicating what we shall presently define as "time-perspective" but also in elucidating what we shall presently refer to as "semantic balance" – two terms of potential significance in the understanding of personality variants.

"Personality variants?" Wait for it…

Sign for a store that just opened in Mark Swofford's neighborhood in Banqiao, New Taipei City:

In the first part of this post, we came across the notion of "bèi zìshā 被自殺" ("be suicided").  Since, for many people, this idea (of somebody being "suicided") is hard to comprehend, I asked several graduate students from the PRC if they could explain how it and the related expressions "bèi tiàolóu 被跳楼" ("was jumped off a building"), "bèi shīzōng 被失蹤" ("be disappeared"), and so forth work.  One of them responded thus:

For details, see "Gabriella Cázares-Kelly, Pima County Recorder, unveils new ‘I Voted’ sticker".

As is wont for The New Yorker, this article is long, and it is particularly fascinating, so it is hard to resist quoting many of its more breathtaking revelations:

"How Scientists Started to Decode Birdsong:  Language is said to make us human. What if birds talk, too?"  By Rivka Galchen, The New Yorker (October 14, 2024)

Of course, we've been through the business of animal communication countless times on Language Log, but where this article differs from previous discussions is that it concentrates on content and consciousness rather than vocables and sounds.

On a drizzly day in Grünau im Almtal, Austria, a gaggle of greylag geese shared a peaceful moment on a grassy field near a stream. One goose, named Edes, was preening quietly; others were resting with their beaks pointed tailward, nestled into their feathers. Then a camouflaged speaker that scientists had placed nearby started to play. First came a recorded honk from an unpartnered male goose named Joshua. Edes went on with his preening. Next came a honk that was lower in pitch than the first, with a slight bray. Edes looked up. As the other geese remained tucked in their warm positions, incurious, Edes scanned the field. He had just heard a recorded “distance call” from his life partner, a female goose whom scientists had named Bon Jovi.

This is apparently from X in February of 2023, though it can now be found elsewhere:

So is ass an intensifier in "super mario level ass geological formation", or has it just been bleached into a formative for turning a phrase into a modifier?