ɯ

much like the entire abstract can be parsed with a bit of effort, e.g. "hot change machine" is probably somehow coming from "thermodynamics" 2/2

an asterisk for the linguist types: I unironically love this meme because it is an "ungrammatical" sentence which nonetheless makes perfect sense with some extra parsing time, and the underlying joke is actually pretty funny! 1/2

don't know what everyone is up in arms about, from my prior readings on thermodynamics (see attached) this is fine

Just read the abstract 🫠 via Alexander Magazinov. I don't believe he is on Bluesky.

Here is the spreadsheet DOGE had ChatGPT produce, to determine which grants were too DEI www.historians.org/wp-content/u...

Picture of DOGE guy Nate Cavanaugh

Screenshot of my DOGE letter “Dr. Joseph Rezek Dear NEH Grantee, This letter provides notice that the National Endowment for the Humanities (NEH) is terminating your federal grant (Grant Application No. FEL29509824) effective April 3, 2025, in accordance with the termination clause in your Grant Agreement. Your grant no longer effectuates the agency's needs and priorities and conditions of the Grant Agreement and is subject to termination due to several reasonable causes, as outlined in 2CFR§200.340. For instance, NEH has reasonable cause to terminate your grant in light of the fact that the NEH is repurposing its funding allocations in a new direction in furtherance of the President's agenda. The President's February 19, 2025 executive order mandates that the NEH eliminate all non-statutorily required activities and functions. See Commencing the Reduction of the Federal Bureaucracy, E.O. 14217 (Feb. 19, 2025). Your grant's immediate termination is necessary to safeguard the interests of the federal government, including its fiscal priorities. Any objections or appeals to this termination will be managed in strict accordance with the President's Executive Orders,

Last year, this guy (left) from DOGE used ChatGPT to find NEH grants that were too “DEI” for Trump, and canceled them, including mine, as shown by the letter I received last April (right). Huge new NYT article on the back story link below

I realise this is like focusing on kitchen hygiene in the Titanic at this point, but:

the LLM the NEH asked "Does the following relate at all to DEI?" (!) usually - though inconsistently - cited language documentation and preservation as grounds for grant termination.

("Yes" => "Defund")

Matthew McConaughey smoking a cigarette and looking absolutely possessed. He offered to take this photo in place of the author, as bluesky user @strange.website does not smoke.

hashtag linguistics

Secretary Hegseth announced this afternoon that DOD will cease ALL graduate study by military personnel at universities including MIT, all the Ivies, Stanford, etc because such schools teach "the enemy's wicked ideology" to officers.

thehill.com/policy/defen...

I regret to report that I am going to have to go full Luddite in my teaching and assessment

Creating a sign language out of everything and everywhere Mariana Martins: Creating a sign language out of everything and everywhere - An example from the deaf people of Bissau

It was a great pleasure and a honor to act as the co-supervisor for Mariana Martins’ dissertation about the emergence of a new Sign Language in Guinea Bissau, for which she received the doctorate yesterday. The book is available open-access.🧵

www.lotpublications.nl/creating-a-s...

I want my money back

for 50 seconds, I thought there was monsters on the world

🗃️ Tyler Austin Harper’s insane Andrew Mellon Foundation hit piece, which of course @theatlantic.com was happy to print, has rightfully pissed off everyone in the Humanities.

As someone who works in higher education funding, here is a 🧵 on why this article is even more infuriating than you think.

Intended: "Morphology doesn't determine language complexity", etc.

Actual answer: "Indo-European doesn't even have the most morphology, my guy, check out Turkish or Swahili"

I really appreciate when you can tell a student wrote their own work and definitely not an LLM

Problem: identify what's wrong w/ the statement "Indo-Eur. languages are the most complex because they have lots of morphology"
🐦🐦
1/2

I'd say yes in this case, especially if the student is also leading on the writing - the last author is often taken to be the advisor in an advisor-student pair. but if I'm working with an RA on my idea/data, and doing most of the writing, they would be second author

it's already hard enough to deal with students cheating; this is potentially a *much* worse problem. why have we not nipped this in the bud by (say) mandating watermarking in LLMs?

some *conferences* do have shorter turnaround times like this, mainly for brief proceedings papers (4-6 pages). for a journal I think I'd agree that this is not a good sign

Review: A concise summary of the work's core contributions to demonstrate understanding, and an objective assessment of its strengths and weaknesses. Reviews should also check for clarity, proper formatting, and ensure the methodology and conclusions are valid, ...

unrelatedly, check out this paper I found by a guy named John O'Hala

The 12th World Congress of African Linguistics (WOCAL12) will be held at the University of Vienna, 7th - 10th July 2027. Abstracts due 30 May 2026

univie.eventsair.com/wocal2027/

#WOCAL12 #WOCAL2027 #linguistics

Don't give anything an Emmy until it's as good as this.

Very high density of brand new sentences in this piece

One of every thousand American residents has been taken into custody by ICE in the last year.

Some knowledge that was new to me while writing this review with my colleagues:
1. The respiratory system, i.e. how thorax and abdomen work together is very different in young infants
2. Empirical data on the coordination between motion and breathing are very limited

The current regime took a week to abolish USAID, which is three times as old and was not in the news every day for brutalizing children

baby's first paper acknowledging an NSF grant that I (baby) willed into existence