Marshallese's phonemic SOCCER BALL vowel

Today we talk about the time a linguist used emojis to transcribe the unique vowels of a poor language: Mark Hale and his 2000 paper on optimality theory which uses Marshallese as a main example and point of discussion. I tried messing with my audio a bit and I think it sounds better in this video, so please let me know what you think!
__________________
RESOURCES & REFERENCES
• [Mark Hale's original paper] - www.degruyter.com/document/do...
• [A 2007 paper where it looks like he ditched his emojis. lol] - www.researchgate.net/publicat...
• [Practical Marshallese by Peter Rudiak-Gould] - www.peterrg.com/Practical%20M...
• [Peace Corps Marshallese] - www.livelingua.com/peace-corp...
• [English to Marshallese dictionary] - s3.amazonaws.com/scschoolfile...
• [Intro to optimality theory] - linguistics.ucla.edu/people/z...

[@alextemplemusic]

A transition between two identical vowels is a cisition

[@Sundrobrocc]

LMAO

[@SkyTheHusky]

Wow, that's an interesting word, and your vocabulary is truly~ Ohhhhhh!

[@shilaylee11]

no way composer alex temple is here commenting on linguistics videos :o

[@alextemplemusic]

@@shilaylee11 If I hadn't gone into music I almost definitely would have gone into linguistics!

[@pinky_pepper]

linguists try not to develop cursed orthography challenge (super hard)

[@LuNa_097]

cant believe we got emojis as phonemes before emojis as tone indicators

[@Sundrobrocc]

Dude that's a wicked idea

[@wordart_guian]

he just saw wingdings and thought "i should use that as a font for a scientific paper, let's make up a reason why"

[@Sundrobrocc]

Nah I think it was a cool argument but it's definitely very meme-y lol

[@meeb_consumer]

Live life with your own objective, as that is the path to happiness

[@chesqen]

2:30 Marshallese vowel jumpscare (very startling)

[@Sundrobrocc]

LOL

[@betos-08]

marshallese already has barely any resources to learn, wtf is this emoji stuff

[@WSCKR]

my day is going better now that i'm aware of this

[@Liggliluff]

(1:40) The IPA is already misleading in so many cases. You can just use /i u a o/ and use the exact sound within [ ] and it'll be fine.

[@Sundrobrocc]

YEAH thats exactly the argument I made

[@NotUselessProductions]

Or /i u e o/ because “a” is a back vowel in many languages

[@vogel6554]

@@NotUselessProductions True. I get so annoyed seeing high shool-level Swedish textbooks transcribing our long and short as /a/ and /a:/ respectively.

[@vogel6554]

This is the kind of content I want to see more of on the linguistics side of youtube. Six minutes about a 20+ year old obscure paper with questionable reasoning, no more and no less.

[@Fenditokesdialect]

Who needs to transcribe allophones when you have footballs

[@devofficialchannel]

🧍tʲɛtʲ🧍

[@Sundrobrocc]

yES

[@GreatJaggia]

and people told me the celtic languages' grammar was "nightmare difficulty"

[@MaeBlythe]

Using emojis to invent new IPA symbols is brilliant! Nearly all modern devices support emoji. Sure, it is obtuse and memorizing them would be annoying, but it's still the most accessible thing I can think of.

[@Sundrobrocc]

YEAH that's a pretty cool take. My only disagreement is that you could just use /a, e, o, u/ for the vowels (or some other analysis) which are already sufficiently accessible, but I could see a point for more complex vowel systems. Given that some devices have different ways of rendering emojis (think iPhone vs Android), would you think this would impact accessibility?

[@MaeBlythe]

@@Sundrobrocc I feel like some emojis are standardized and unambiguous enough that they could be used without issue. If it's not allowed to have similar emojis (like ☺️ and 😊) and it's things that are rather universal (♥️ will always look like a heart) then there's not too many issues! I'm a vocal synth user. There's two camps (with me in the smaller one!) Some people switch between different systems for different languages. For English they'll use Arpabet, for Mandarin they'll use pinyin... But that bothers me 😝 so outside of being forced by phonemizers, I exclusively work with SAMPA. A project I'm working on even differentiates between t and t_d! For people like me, it's not about the actual language but the phonemes itself. So being specific is more helpful in that context.

[@Sundrobrocc]

I see. Very cool stuff!

[@xwtek3505]

What system supports emoji but not ɨ ə ɜ a

[@MaeBlythe]

@@xwtek3505 I'm on mobile 😅 my system displays ipa characters with no issues, but I would need to download a keyboard to use them. Whereas my emoji keyboard is right on the comma key. But the issue isn't a system not supporting ipa characters and instead what to do when you need to represent something novel that doesn't have a set symbol already. You're constrained to what is within Unicode already. You could just say that π is now a phoneme name. That seems more problematic than saying that 🥋 is now a phoneme name.

[@xavierreichel8254]

I did that NACLO problem recently! Not American, just love doing linguistics problems wherever I can get them.

[@Sundrobrocc]

It was definitely something. I got destroyed lol

[@themurpleman802]

I got that question at a competition but I had to skip it cause I couldn’t figure anything out💀

[@Sundrobrocc]

@@themurpleman802 LMAO it was crazy

[@devonoknabo2582]

I kinda disagree about your argument about /a/ being just as unspecified as those emojis. while I think you're mostly correct. You can at least make an assumption with the ipa symbol /a/ being [a] or something in that range. With a soccer emoji I have no clue what to assume.

[@Sundrobrocc]

That's fair, but assuming /a/ as [a] is still an assumption that can't be verified without context which is the main idea of using emojis (IMO)

[@rateeightx]

True, but it's also very common to transcribe [ä] with /a/, so by that regard it's probably actually one of the vaguest phonemic transcriptions, Alongside the likes of /ə/. Plus it's decently common that the actual phonetic realisation of a vowel will vary somewhat based on context, for example in English I generally pronounce vowels further back when they're followed by /l/. So yeah, you can definitely make a decent guess about pronunciation when a vowel is transcribed /a/, But it's definitely not guaranteed to be accurate. /⚽/ in isolation would definitely be unclear, But in the context of a paper where he specifically tells you its pronunciation, I don't think it's too different from using a standard IPA symbol would be.

[@treeps637]

​@@rateeightxBut how would he decide which surface realization should be used to represent the underlying phoneme?

[@rateeightx]

​@@treeps637 It doesn't matter, So long as he described the actual phonetic realisation in the paper. /⚽/ would work just as well as /ɔ/ or /ɛ/, Or just about anything else, Provided you specify what it actually means.

[@shilaylee11]

to (possibly) lend some clarity on the bracket notation: i think 2:07 is the key point here, and while i don't have access to the article to confirm this (damn you de gruyter), what i imagine Hale is saying is that Marshallese vowel allophony is better analyzed as part of "phonetics as the output of the body", i.e. a result of the phonetic implementation and not the phonology that precedes it. i like to think of this in terms of articulatory phonology, a model in which the "basic units" of phonology aren't abstract features but rather (abstractions of) gestures of the tongue, lips, etc.; in this sense, the output of the phonology is basically a list of instructions for the mouth, which then moves accordingly. for a fully specified vowel like [i], these instructions include (roughly) tongue placement and lip rounding. for an underspecified vowel like the "soccer ball" vowel, though, these instructions would be incomplete. what i'd expect Hale to argue, if he's adopting this framework, is that this is actually fine: the tongue and lips don't need those extra instructions! given the nature of the allophony we see here, it could just be that gestures associated to consonants are carried over into their neighboring vowels, so that e.g. a labialized consonant spreads some lip rounding to the vowel next to it, making part of it rounded. what bugs me about the above approach is that the allophony here seems to be systematic; we *know* that a vowel will end up partly rounded because it's next to a labialized consonant, at least enough to merit transcribing the vowels as rounded, etc. while vowel-consonant coarticulation is by no means uncommon, something this regular is more likely to emerge from the phonology, which, being an abstract system, tends to deal in categories and absolutes. i hope there's been some articulatory (or even acoustic) study of Marshallese vowels, because i'd really love to see what the evidence is, if any, that this is a phonetic rather than phonological process. i do love me some underspecification and vowel coarticulation (i've done a little work on vowel harmony and AP and plan to dig more into each in grad school) anyway, thanks for the video! cool to see people boosting engagement in the field :)

[@Sundrobrocc]

Thanks for the awesome comment! While I see your point, I still think these instructions can be carried out with only the two set of brackets, even if perhaps drifting away from their typically conventional uses, _especially_ because like I said, he didn't provide any examples where // and [] notation have differing information. In a sense, // already can represent more basic / general instructions that can be fully produced between brackets, but this is all IMO. Funny guy and funny work, who really knows lol ¯\_(ツ)_/¯

[@MarshallLevin]

0:52 that's not a rotary telephone 🙂

[@Sundrobrocc]

It's a rotary telephone on the original paper, try zooming in :)

[@MenaceGallagher]

The emojii shown in the video is a touch tone telephone (I'm the only one on your AM radio)

[@arkonnKeb]

question unrelated to dhe video (which is still very interesting) but what fonts are you using? Espacialy dhe one for big text. Thanks

[@Sundrobrocc]

Big text - Brixton Sans Most body text - Lora :)

[@SkyTheHusky]

Can I steal dhat use of "dh"? I love it

[@arkonnKeb]

@@SkyTheHusky yes! Spread dhe word

[@christopherellis2663]

Romanian has seven vowels. Six of which are found in Marshallese.

[@Sundrobrocc]

.w.hhhaaarrrr

[@bazyl_ia6425]

i don't really have anything to say i just love it when you post :3

[@Sundrobrocc]

HAHA WHAT AW

[@catomajorcensor]

I wanted to make a conlang with a similar (though simpler) vowel system before ever learning about this, but never got to it. Glad to know it's still naturalistic!

[@Sundrobrocc]

Bro if Arabic exists anything can be naturalistic. You can literally do anything given you have some sort of reasoning for it

[@SkyTheHusky]

ANADEW A Natlang Already Did it Even Worse My principle for conlanging

[@bruhistantv9806]

If all vowels in Marshallese are modified by the consonants near them, then perhaps it would benefit from an abjad-type script

[@Sundrobrocc]

brooo

[@CerealBowlSystem]

emoji are cool and unicode is pretty much *the* standard on electronic devices nowadays, but having to draw emoji on paper sounds hellish. We have so many writing systems, I don't think it would've been too wild to borrow some arbitrary symbols from existing language scripts

[@Starguy256]

If we had to regularly write emojis on paper they'd eventually be simplified over time. Like, instead of an ox's head, maybe just an acute angle with a line through the middle, something like "A". Or instead of a drawing of a fish, maybe just a curvy line and a straight line, something like "D". (my point is this has already happened)

[@camelcaseco]

4:30 it's true that arabic vowel phonemes (especially /a/) can vary a good bit. I'm not a native speaker so take this with a grain of salt, but I'm not sure if your transcriptions are correct for any dialect. [hna] sounds Moroccan but I don't believe that it sounds like [ɑ] in that position, and likewise the vowel in كلاب would be lower, perhaps closer to [æ]. again, however, I could be totally wrong. In modern standard arabic, the main allophony is of /a aː/ is that they back to [ɑ] around emphatic (pharyngealized) consonants. And in Jordanian/Palestinian dialects this also applies, though the quality of long /aː/ is more front and slightly higher. And in Lebanese and Syrian dialects this can raise to the extent that long /aː/ merɡes with lonɡ /eː/ (related to modern standard arabic's /aj/ diphthong). Anyhow, stellar video, big fan of this.

[@Sundrobrocc]

It's in fact Moroccan Arabic :) That's basically my default unless I say otherwise-I'll try to start clarifying this more in videos. /a/ word-finally in Darija is very commonly realized [ɑ], so that's where that comes from. As for the /a/ in كلاب, I'm pretty confident I hear [ɛ] most of the time in Moroccan Arabic, though the vowels vary a lot and the distinction between [ɛ~æ] is probably not too important within the context of the example I was giving, but thanks for the comment nonetheless! I love learning about other dialects

[@rateeightx]

While I can understand using irregular symbols, Such as emoji, to transcribe phonemes that vary drastically in realisation based on context, trying to draw a distinction between the brackets and his human markers for "Phonetics as result of the body" seems quite ridiculous to me. Sure, you could try to draw a distinction between broader and more specific phonetic transcription, The English word "Force" for example, you might transcribe phonemically as /fors/, broadly phonetically as [fɔɹs], and then specifically phonetically, Regarding one person's specific realisation, as [ʋ̥ɔ̝ɹs] or something, But it doesn't seem like that's what he's actually doing, if anything the realisation of the Marshallese vowels (Based on my highly limited understanding of the language) seems to be more based on grammar (Or I suppose phonotactics) than on the body, Especially when compared to something like the English tendency to realise /d/ as [ɾ] when speaking quickly, because those two sounds are made with basically the same tongue movements, [ɾ] is simply a more fluid motion, so it's unsurprising that one could become the other when speaking quickly, Because it's simply slower/more effort to actually distinguish them.

[@salt407]

Babe wake up. Sundro posted

[@Sundrobrocc]

what did you say

[@lavender_verandah]

I'd rather depict the Marshallese vowels with additional imaginary axis and it would still look more unironic and feasible than this whole emoji thing

[@Sundrobrocc]

DUDE YES when we bringing φ into this

[@robkabar303]

pretty cool

[@thurm64]

toki pona cameo! wawa

[@Sundrobrocc]

hell yeah

[@choqi29]

Im interested in purchasing One

[@Sundrobrocc]

Dude I swear

[@YamamotoTV2021]

Purchasing what? A soccer ball? I have a soccer ball at home.

[@casualmajestic9223]

Hi, could you make an in-depth dive into ithkuil video c:

[@Sundrobrocc]

While I absolutely love Ithkuil (it's probably my favorite conlang by far), I don't know if I feel qualified to explain the grammar just yet, but hopefully one day I will be enough so to be able to make that a possibility :)

[@interbeamproductions]

agma schwa's dad

[@Sundrobrocc]

wHAT

[@interbeamproductions]

🚹ŋ⚽️🚹

[@Sundrobrocc]

@@interbeamproductions dude that goes so hard

[@Sqone1]

good video, but the volume is a little low

[@Sundrobrocc]

thanks! i'll keep in mind

[@wlwgwlwgnomesarereal]

SIUUUUUUUUUU

[@vanessagarcia248]

I like your fish :)

[@Sundrobrocc]

ur fish 😭😭😭😭😭😭😭😭😭😭😭

[@lawrencecalablaster568]

What was the fish for?

[@Sundrobrocc]

the word in the thumbnail means _Lutjanus Flavipes_ which is the fish you see lol

[@choqi29]

Set necessary do to deer do no pat consume air due chocolate teen🙏

[@cherrycolareal]

What is lil bro yapping about

[@Sundrobrocc]

علاش الدري كيياپ💀💀💀💀

[@aureole6383]

do to deer do no pat consume air🗣️🗣️🫁🫁👄👄💥💥💥

[@Sundrobrocc]

dude i been saying this

[@user-id9bn1ic9v]

o͜o

[@equilibrum999]

io'kwe

[@Sundrobrocc]

no u also nice profile picture 💀

[@equilibrum999]

@@Sundrobrocc picture profile 阴阳 is, unknowable 人

[@choqi29]

Fell off

[@wigwagstudios2474]

3:03 MMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMM'p TYUp gyothj;['up'[og; yjitfhupyh ['goytur7'h[lp;ruyhp't[g oty[hp;'/tuhj'/;lp[i8up.['o;tji[ ]/ji7ku /p['jtikhy [}"PUYILjk[]p

[@Sundrobrocc]

WHAT

[@wigwagstudios2474]

8)

[@user-if1gj2wh5c]

Just started casually learning about linguistics this week. I learned the ipa and the difference between phonemes and phonetics and I feel like a god. Any recommendations on stuff to do next? Like interesting books or youtube channels or something. Also I got this video in my recommended and thought u might have a strong opinion on it kzbin.info/www/bejne/bGe2pKJ5nJuij9Esi=EWOe_XjyqiK7nWrO

[@Sundrobrocc]

An in-depth response to that video I'd say is beyond the scope of this comment since I'd have a lot to say, but in general he makes really interesting points (I absolutely love his videos lol) that do have merit. In terms of recommendations on stuff to do next, try your hand at conlangs! They're how I learned a huge chunk of what I know about linguistics-learn by doing.

[@cmyk8964]

Wow, the j-ɣ-w analysis is interesting and makes sense

[@Poghuiro]

o͜o