Language Log

Following up on our previous post, "Sakas, Kushans, and Hephthalites: the sources in Greek, Latin, Persian, and Chinese" (9/24/25) by Taishan Yu, we turn now to Étienne de La Vaissière's "A Military Origin for New Persian?", which was published lightning fast by Acta Orientalia Academiae Scientiarum Hungaricae.

Received: 26 April 2025 • Accepted: 3 July 2025
Published Online: 5 August 2025

From Astral Codex Ten:

Something is off about this Bay Area House Party. There are . . . women.

“I’ve never seen a gender balance like this in the Bay Area,” you tell your host Chris. “Is this one of those fabled ratio parties?”

“No – have you heard of curtfishing? It’s the new male dating trend. You say in your Bumble profile that you’re a member of the Dissident Right who often attends parties with Curtis Yarvin. Then female journos ask you out in the hopes that you’ll bring them along and they can turn it into an article.”

“What happens when they realize Curtis Yarvin isn’t at the party?”

“Oh, everyone pools their money and hires someone to pretend to be Curtis. You can just do things. Today it’s Ramchandra.”

Sino-Platonic Papers is pleased to announce the publication of its three-hundred-and-sixty-sixth issue:

“Relations between Persia and Central Asia in Antiquity: An Examination of the Written Sources,” by Yu Taishan. (pdf)

A couple of decades ago, I reviewed the argument between Paul Ekman and Margaret Mead about whether facial expressions are universal or socially constructed:

"Political correctness, biology and culture", 10/31/2006
"The Cabinet of Dr. Birdwhistell", 11/2/2006

Ekman won that argument, at least as judged by most of subsequent intellectual history — though not everyone is convinced, and his own methods have been criticized.

The Japanese word for "spinach", "hōrensō", has many different graphic forms and meanings:

(Wiktionary)

Are nonlinear vocal phenomena as distracting as people think?
Andrey Anikin, ORCID Icon Bioacoustics The International Journal of Animal Sound and its Recording (18 Sep 2025)

Keywords

Roughness    vocal communication    attention    infant cries    auditory salience    nonlinear vocal phenomena

Today's For Better or For Worse:

April's "comftrable" is not dictionary-sanctioned — but maybe it should be?

Sino-Platonic Papers is pleased to announce the publication of its three-hundred-and-sixty-eighth issue:

“Demetrios of Bactria as Deva Gobujo and Other Indo-Greek Myths of Japan,” by Lucas Christopoulos. 

We've been inconsistent recently in our annual celebrations of Talk Like A Pirate Day, so here's our favorite image, with a couple of past post links below the fold:

There's been some buzz recently about how syntactic structures are implicit in Large Language Models — most recently, the Liu et al. paper noted yesterday by Victor, and an accepted ms by Futrell and Mahowald at Behavioral and Brain Sciences, "How Linguistics Learned to Stop Worrying and Love the Language Models". Futrell and Mahowald recognize something that Liu et al. mostly ignore, namely that constituent structure is obviously implicit in statistical patterns of sequential data, at least if the sequences were generated by a constituency-sensitive process — and that algorithms taking advantage of that fact have been Out There for 70 years or more.

"Active Use of Latent Tree-Structured Sentence Representation in Humans and Large Language Models." Liu, Wei et al. Nature Human Behaviour (September 10, 2025).

Abstract

Understanding how sentences are represented in the human brain, as well as in large language models (LLMs), poses a substantial challenge for cognitive science. Here we develop a one-shot learning task to investigate whether humans and LLMs encode tree-structured constituents within sentences. Participants (total N = 372, native Chinese or English speakers, and bilingual in Chinese and English) and LLMs (for example, ChatGPT) were asked to infer which words should be deleted from a sentence. Both groups tend to delete constituents, instead of non-constituent word strings, following rules specific to Chinese and English, respectively. The results cannot be explained by models that rely only on word properties and word positions. Crucially, based on word strings deleted by either humans or LLMs, the underlying constituency tree structure can be successfully reconstructed. Altogether, these results demonstrate that latent tree-structured sentence representations emerge in both humans and LLMs.

The most famous example, of course, is the 1945 "Who's on first?" dialogue:

Mark Metcalf wrote: