xkcd 2682 for linguists

Comments

• lvxferre@mander.xyz ⁨3⁩ ⁨days⁩ ago

  I’d say pretty accurate:

  top left

  English verbal paradigm is rather barebones; because of that, the content verb of a sentence often “ditches” meaningful distinctions into the rest of the sentence. Sometimes an aux verb, or even a conjunction. That’s the case here; you got a distinction between realis and irrealis, that plenty languages would convey through the verb, but English doesn’t.

  Note the “counterfactual timeline” (irrealis past, unreal time etc.) deals with events the speaker wishes that would have happened in the past.

  mid left

  Ah, here’s a paper about this. I didn’t read the paper fully, but: apparently it is computable but NP-hard.

  bottom left

  Indo-European. Germanic branch.

  top centre

  There’s a proposed language family called Dené-Yeniseian; the languages in question are spoken in Siberia (Yeniseian) and a chunk of North America (Na-Dené).

  Trivia: remember the Huns? Likely Yeniseian speakers.

  schwa

  The current mainstream hypothesis is unvoiced vs. voiced vs. breathy voiced. There’s also a bunch of alt hypotheses including glottalisation; for example “stiff” unvoiced (i.e. [p t k]) vs. pre-glottalised vs. “slack” voiced; Javanese has the stiff/slack contrast, and pre-glottalised consonants are somewhat common.

  bottom centre

  Two main pressures:

  • the speaker wants things to be easy to pronounce; e.g. if an articulatory gesture is unneeded, it might get ditched.
  • the hearer wants things to be easy to distinguish; e.g. if two sounds associated with different phonemes are a bit too similar, and the distinction is meaningful, they might drift apart from each other.

  Those two pressures are in a tug-o-war, and that tug-o-war drives sound changes.

  top right

  I feel like the spread of Afro-Asiatic might have to do with this period, as it probably allowed people to migrate further than through drier periods. But past that? I have no idea, and I hate that I have no idea.

  mid right

  Eh… it’s complicated. It seems, for most authors, that Tibetan and the Sinitic languages are in different branches of the family; and usually Burmese is placed in Tibetan’s branch. Everything else, though? No consensus at all.

  bottom right

  The problem starts with the definition of a language. I’ll illustrate it with the Romance languages:

  • If you speak Portuguese you’ll probably understand Galician just fine. So let’s count them as a single language. It’s reasonable, right? Portuguese is basically a Galician dialect.
  • If you speak Galician you’ll probably understand Asturian just fine. It makes sense - both originated from Latin dialects spoken right next door to each other. Let’s count both as the same language. Alongside Portuguese, as per the step above.
  • Asturian and Castilian/Spanish are really similar, so let’s lump them together. Alongside Portuguese and Galician.
  • Castilian and Aragonese, too. Same language as Portuguese, Galician, Asturian.
  • Catalan is really similar to Aragonese. One more into the bag!
  • [one thousand steps later…]
  • Sicilian and Calabrian are really similar, right? Same language then. They get into the same bag as the others.

  So you reach the conclusion that none of those varieties “counts” as a language. Then you proudly put in some paper "number of Romance languages: three (Italo-Western, Sicilian, Romanian). Italian is now an Italo-Western dialect, French is an Italo-Western dialect, everything else is a dialect.

  Except that most of those so-called “Modern Latin speakers” can’t understand each other. And the speakers don’t consider them the same language, they consider it as different things.

  But this isn’t just with the Romance languages. Cue to English and Scots, or the Germanic varieties in the continent. Or the Sinitic varieties spoken in China. The Bantu family. The Slavic branch. I think Quichua has the same issue, too.

  Yeah, nah, you aren’t “counting” them - you’re placing arbitrary divisions here and there to make the number bigger or smaller.

  source

  • Tar_alcaran@sh.itjust.works ⁨3⁩ ⁨days⁩ ago

    “how many languages are there” is the same kind of question as “how many colours are there”. It doesn’t work like that, you can group stuff into as many or as few categories as you like.

    source

    • stabby_cicada@lemmy.blahaj.zone ⁨3⁩ ⁨days⁩ ago

      It’s interesting to note that every person who knows a language defines its words just a tiny bit differently than everyone else, based on their personal experiences with the objects and concepts and actions the words refer to.

      It is, therefore, arguable that there are at least eight billion different languages spoken on Earth. And at least twice that, when you take bilingual and multilingual speakers into account.

      source

      • -> View More Comments
  • stelelor@lemmy.ca ⁨3⁩ ⁨days⁩ ago

    “schwa” is true neutral aaaaand now I need to see the vowels’ alignment chart

    source

    • lvxferre@mander.xyz ⁨3⁩ ⁨days⁩ ago

      We could simply map front/back = law/chaos and closed/open = good/evil… but that’s booooring.

      Instead, let’s say:

      • Cardinal vowels are lawful, glides are chaotic. Everything else is neutral.
      • Good vowels are only defined by height, backness, and roundness. Give them an additional contrast and they become neutral, two and they’re evil, three and they’re the sounds uttered by Satan when you just reach Hell.

      So for example the schwa strictu sensu is true neutral - it isn’t just a central vowel, it’s also reduced (shorter). Something like [ɜ̃͜ʊ̃] is chaotic evil. But most languages have a backbone of lawful good vowels.

      source
  • oxideseven@lemmy.ca ⁨3⁩ ⁨days⁩ ago

    Schwa is such a crappy inside joke lol. 10 points!

    source
• FundMECFS@quokk.au ⁨3⁩ ⁨days⁩ ago

  I’m so fucking jealous of Randall Monroe. They’ve got the coolest job.

  source