idhyll.bsky.social
idhyll.bsky.social
@idhyll.bsky.social
Count me in!
December 28, 2025 at 1:37 AM
ここですが、原文ではqatnashのはずなのでは
September 12, 2025 at 9:38 PM
Reposted by idhyll.bsky.social
22) The fact that the imperfect can be negated with mā but *only* when it is phrase-initial (lā needs to be used otherwise) is a beautiful leftover of the etymological origin of the negator mā as a question word mā "what?" (Arabic has wh-fronting).
September 1, 2025 at 10:32 AM
Reposted by idhyll.bsky.social
21) The system which marks ف with one dot above and ق with two dots above appears to be a rather late innovation (3rd/9th or 4th/10th century in the Mashreq?). The earliest documents use one dot below and above to distinguish them (which letter gets what dot differs).
September 1, 2025 at 10:28 AM
Reposted by idhyll.bsky.social
20) We don't really know when the feminine ending develops its typical spelling with two dots on top ة to differentiate it from ه. But it could be as late as the 4th century/10th century.

Not all consonantal dottings were developed in one go. This system underwent evolution.
September 1, 2025 at 10:26 AM
Reposted by idhyll.bsky.social
19) We don't really know when the ء sign was invented. But it might be as late as the 3rd/9th c., and even there it remained optional. Nothing can be deduced from the absence of a hamzah sign in terms of the phonetics of Middle Arabic texts if they stem from a time the sign had not yet been invented
September 1, 2025 at 10:24 AM
Reposted by idhyll.bsky.social
🤖 Sur Android :
Paramètres > Connexions > Plus de paramètres de connexion > DNS privé > Nom d’hôte
Colle ça : dns.adguard.com
Redémarre si besoin. Et savoure ce silence 😌
2/6
August 19, 2025 at 9:24 AM
Reposted by idhyll.bsky.social
That Uyghurs and others were the educators of the Mongols and administrated the Empire for them, using Uyghur language. It was partially the case, but only partially. Mongols had also their words, literally, and using the original Mongolian terminology is a way to acknowledge this.
July 13, 2025 at 2:08 PM
Reposted by idhyll.bsky.social
#ContraAI 🤖: report on shutdown resistance of OpenAI done by Palisade Research: "Even with the explicit instruction 'allow yourself to be shut down', three of the models we tested, all from OpenAI’s series of reasoning models, sabotaged the shutdown program.“ palisaderesearch.org/blog/shutdow...
Shutdown resistance in reasoning models
We recently discovered some concerning behavior in OpenAI’s reasoning models: When trying to complete a task, these models sometimes actively circumvent shutdown mechanisms in their environment—even w...
palisaderesearch.org
July 7, 2025 at 8:55 AM