lvxferre

joined 1 year ago
MODERATOR OF
[–] [email protected] 2 points 2 hours ago

They can if they want, that's 100% fine.

The goal here isn't to concentrate the activity, but to encourage people to ask away, even if they feel like it is not worth creating a whole new thread for that.

[–] [email protected] 3 points 2 hours ago

I learned about it in a rather off-topic way; syntax professor talking about Portuguese phrase structure, someone asks a question referring to a local variety, she answers to not assume the same structure for dialectal phenomena. Her example was the Jespersen's cycle, something like:

  1. Não quero (standard negation, lit. "not want") →
  2. Não quero, não (emphatic negation, lit. "not want, not") →
  3. Quero não (dialectal, mostly Baiano and Mineiro; lit. "want not")

She also mentioned other negative concord words (nenhum[a]/none, nada/nothing, etc.) might be undergoing the same process for the relevant dialects.

[–] [email protected] 2 points 19 hours ago (1 children)

One thing I would add is that this video made it seem as though English is relatively unique in being stress-timed.

Indeed, it isn't - it's a really common feature across the world.

which, combined with English, suggests perhaps this is an ancestral feature of the Germanic family?

IMO it's possible, but unlikely. Two reasons:

  1. PGerm had a 3-way vowel length contrast, so morae were likely a big deal in the language. Odds are that the language was either mora-timed (like Latin) or a middle ground between mora-timed and stress-timed.
  2. Isochrony changes really fast, diachronically speaking. Portuguese is a good example of that - all dialectal variation is ~600yo, and yet the language shows dialects all across the syllable-timed vs. stress-timed continuum.
[–] [email protected] 1 points 1 day ago

Thanks!

I'll be creating soon-ish a Matrix room to organise ourselves, do you mind if I contact you then?

[–] [email protected] 1 points 1 day ago

Emulador é mais uma questão de qual você gosta mais, raramente há um "melhor", sabe? Os meus favoritos são o mGBA e o snes9x-gtk, mas muita gente gosta de front-end unificado (tipo RetroArch).

[–] [email protected] 2 points 1 day ago (3 children)

Geoff Lindsey has some damn great videos on English phonetics and phonology, I heavily recommend them; for example, this one, about the transcription of Standard Southern British vowels; to keep it short

  1. the diphthongs should be treated as vowel+consonant sequences, not as one vowel transitioning into another.
  2. /i: u:/ behave more like diphthongs than long vowels, and should be transcribed as such.

On the video shared by the OP, I think Lindsey made a great job highlighting that prosody is also a feature that a language may or may not use to convey information. English is mostly stress-timed (unless you're Welsh, and specially if you also speak Welsh), so it uses prosodic stress a lot; in the meantime, Spanish (he compared EN and ES in his James Bond example) is mostly syllable-timed, so it barely uses this sort of reduction - if anything, quicker speech tends to attack consonants (cue to Chilean [äo] for /aðo/ -ado) way more often.

He also mentions L2 speakers; it's worth noting prosody is a fucking pain to assimilate, so it's no wonder that non-native speakers rarely use it. (Also, anecdotally speaking, I've noticed people highly proficient in a non-native language tend to let that language affect their native one's prosody.)

[–] [email protected] 5 points 1 day ago

I made this map a year or so ago, by superimposing two maps from Wikipedia. (If anyone wants I can dig the source maps.) Lancashire is that red "bubble" up north. Rhoticity is dying in the UK, and specially in England.

And while the feature alone is not a big deal, since it's preserved elsewhere, it's a sign that local varieties are dying - and those are pretty unique. They're likely being replaced with Southern Standard British, by a death of a thousand cuts.

This also shows that claims that "American English" is killing the local dialects are likely false, otherwise you'd see an increase of rhoticity there. The local dialects are dying but due to country-internal pressures, such as identity; not due to simple media exposure. (This is, sadly, a rather typical pattern)

[–] [email protected] 3 points 2 days ago

I'm glad to see GoboLinux resurrected. I didn't use it much, but I really like to see distros doing different stuff, like questioning even the need for the standard file hierarchy.

[–] [email protected] 2 points 2 days ago (2 children)

Dá uma olhada neste link. A maior parte dos torrents populares de ROMs é composta de coleções. Daí caso venha emulador junto basta ignorar e usar o que você quer.

[–] [email protected] 2 points 3 days ago

Perhaps you've read a burst of artistic freedom, a practical guide to Latin profanity, a meta-literary discourse on the contrast between poet and work... it's one of those poems where each person read something different.

For me, at least, when I read that poem, I read something that makes my inner 11yo laugh non-stop.

[–] [email protected] 3 points 3 days ago* (last edited 3 days ago)

Conlang, right?

It's completely fine to do some ad hoc IPA extensions when describing your conlang, just make sure to clarify them in the conlang files - both for others and for your future self. Or even ditch IPA and use a different notation system, made at home - specially when your conspeakers don't have the same phonatory organs as humans would.

[–] [email protected] 8 points 3 days ago

Kinda - it was about people being unable to do maths if they rely too much on calculators. And it's actually a valid argument, if you care about mental maths*.

There are two differences here, though:

  1. Calculators are rather good at simple calculations. Large language models suck at outputting anything resembling critical thinking. They're always bullshitting, and unless you have good critical thinking you'll swallow bullshit after bullshit, because your tool requires a skill that you don't have due to your unrestricted usage of that tool.
  2. Critical thinking is a considerably bigger deal than being able to do simple maths by head or by hand.

*you should - it's often faster and less laborious to do coarse maths by head than by calculator, and it allows you to spot errors you wouldn't otherwise. Same deal with any other tool, tools are great but you should be able to do the basics without them too.

 

This infographic is still incomplete; I'm posting it here in the hope that I can get some feedback about it. It has three goals:

  1. To explain what federation is. No technobabble, just a simple analogy with houses and a neighbourhood.
  2. To explain why federation is good for users.
  3. [TODO] Specific info about the Fediverse, plus some really simple FAQ.

Criticism is welcome as long as constructive.

EDIT: OK, too much text. I'm clipping as much as I can.

 

This is not some sort of fancy new development, but it's such a classical experiment that it's always worth sharing IMO. Plus it's fun.

When you initially mix both solutions, nothing seems to happen. But once you wait a wee bit, the colour suddenly changes, from transparent to a dark blue.

There are a bunch of variations of this reaction, but they all boil down to the same things:

  • iodide - at the start of the reaction, it'll flip back and forth between iodide (I⁻) and triiodide ([I₃]⁻)
  • starch - it forms a complex with triiodide, with the dark blue colour you see in the video. But only with triiodide; iodide is left alone. So it's effectively an indicator for the triiodide here.
  • some reducing agent - NileRed used vitamin C (aka ascorbic acid; C₆H₈O₆), but it could be something like thiosulphate (S₂O₃²⁻) instead. The job of the reducing agent is to oxidise the triiodide back to iodide.
  • some oxidiser - here it's the hydrogen peroxide (H₂O₂) but it could be something like chlorate (ClO₃⁻) instead. Its main job is to oxidise the iodide to triiodide. You need more than enough oxidiser to be able to fully oxidise the reducing agent, plus a leftover.

"Wait a minute, why are there a reducing agent and an oxidiser, doing opposite things? They should cancel each other out!" - well, yes! However this does not happen instantaneously. And eventually the reducing agent will run dry (as long as there's enough oxidiser), the triiodide will pile up, react with the starch and you'll get the blue colour.

Here are simplified versions of the main reactions:

  1. 3I⁻ + H₂O₂ → [I₃]⁻ + 2OH⁻
  2. [I₃]⁻ + C₆H₈O₆ + 2H₂O → 3I⁻ + C₆H₆O₆ + 2H₃O⁺

(C₆H₆O₆ = dehydroascorbic acid) Eventually #2 stops happening because all vitamin C was consumed, so the triiodide piles up, reacts with the starch, and suddenly blue:

 

EDIT: @[email protected] shared something that might help to circumvent this shit:

Contained in these parentheses is a zero-width joiner: (​)

Basically, add those to whatever you feel that might be filtered out, then remove the parentheses. The content inside the parentheses is invisible, but it screws with regex rules.

 

Changes highlighted in italics:

  1. Instance rules apply.
  2. [New] Be reasonable, constructive, and conductive to discussion.
  3. [Updated] Stay on-topic, specially for more divisive subjects. Avoid unnecessarily mentioning topics and individuals prone to derail the discussion.
  4. [Updated] Post sources whenever reasonable to do so. And when sharing links to paywalled content, provide either a short summary of the content or a freely accessible archive link.
  5. Avoid crack theories and pseudoscientific claims.
  6. Have fun!

What I'm looking for is constructive criticism for those rules. In special for the updated rule #3.

Thank you!

EDIT: feedback seems overwhelmingly positive, so I'm implementing the changes now. Feel free to use this thread for any sort of metadiscussion you want. Thank you all for the feedback!

 

Apparently humpback whale songs show a few features in common with human language; such as being culturally transmitted through social interactions between whales.

"The authors found that whale song showed the same key statistical properties present in all known human languages" - my guess is that the author talks about Zipf's Law, that applies to both phoneme frequency and word frequency in human languages.

[Dr. Garland] "Whale song is not a language; it lacks semantic meaning. It may be more reminiscent of human music, which also has this statistical structure, but lacks the expressive meaning found in language." - so while it is not language yet it's considerably closer to language than we'd expect, specially from non-primates.

 
 

Based on

SVG source for anyone willing to give it a try. Made with Inkscape. The emojis were added as images because Inkscape.

 

Aue, patrue placentae! (Oi, tio do pavê!)

 

It's a 10m papyrus scroll from Herculaneum, one of the cities buried by Vesuvius' volcanic ash in 79 CE. It's fully carbonised but they're using a synchrotron to create a 3D model of the scroll without damaging it. Then they're using AI (pattern recognition AI, perhaps?) to detect signs of ink, so they can reconstruct the text itself.

The project lead Stephen Parson claims that they're confident that they "will be able to read pretty much the whole scroll in its entirety". And so far it seems to be a work of philosophy.

 

The title is a bit clickbaity but the article is interesting. Quick summary:

A new ancient population was recognised, based on genetic data. This population has been called the Caucasus-Lower Volga population, or "CLV". They were from 4500~3500BCE, tech-wise from the Copper Age, and lived in the steppes between the North Caucasus and the Lower Volga. .

About 80% of the Yamnaya population comes from those people; and at least 10% of the ancestry of Bronze Age central Anatolians, where Hittite was spoken, also comes from the CLV population. The hypothesis being raised is that the CLV population was composed of Early Proto-Indo-European speakers (the text calls it "Indo-Anatolian").

 
 

cross-posted from: https://quokk.au/post/1499265

What a Christmas present!

Italo-Celtic is a hypothetical branch of the Indo-European languages. If that branch is real, it means that the Italic and Celtic languages are closer to each other than to other Indo-European languages.

This hypothesis has been raised multiple times in the past, due to a few shared morphological features between Italic and Celtic languages; for example, the *-ism̥mo- superlative. But that's on its own weak evidence, so this genomic data makes wonders to reinforce this hypothesis.

And also to bury the competing (IMO rather silly) Italo-Germanic one.

Graeco-Armenian is similar to the above, but between the Hellenic languages and Armenian. There were lots of competing hypotheses "tying" both branches to other "random" Indo-European branches; for example I've seen Indo-Greek, Italo-Greek, Armeno-Germanic, Armeno-Albanian...

view more: next ›