Update: * Video description now contains sources for the full audio clips, including Rouen, which I forgot to show in the video. * I want to apologise for my mischaracterisation of Thiruvananthapuram. What originally struck me as entertaining was how much we suck at correctly interpreting the audio, rather than Malayalam itself; in my attempt to entertain I ended up ridiculing the language instead, which is not appropriate. * The Korean county explanation needs more context: Jeungpyeong is indeed evidently a self-contained urban centre (you can see that on a map), but to the administrative system of South Korea there doesn't seem to be legal entity corresponding to just the urban centre, only the entire county. As a result it also doesn't have a "town" Wikipedia article. The same applies to most other mid-size towns in korea.
@@Kire1120 I didn't have that feeling because I'm a foreign speaker who trying to speek in a standard accent for my university. You could tell I'm not a Brit but unless you would know Polish and English comparative phonology you wouldn't know I'm Polish (my pronunciation teacher, on the other hand, can spot regional Polish features in my English).
Hey! As someone who is from the state of kerala in India, and has family in Thirunananthapuram, I’d like to say that most people still call the city Trivandrum, cause no-one wants to say that mess in the middle of a conversation. Just hearing it, I’d tokiponise to Tiwanatapujo perhaps I’m biased though for being able to understand that word at all😅
This was really interesting, although I'd have a few caveats about whether this is necessarily the best approach towards adapting loanwords. There is an issue with the fact that the primary language of poll takers could affect their perception of foreign words. For instance, consider how English speakers are far more likely to interpret the true-mid [e̞] in Spanish "tres" [tɾe̞s] as sounding closer to the [ei̯] in English 'trace' [tɹei̯s] (which has nearly the same vowel, although with an -i̯ glide) than to the [ɛ] in English 'tress' [tɹɛs] (which has a different vowel quality). Native Spanish speakers, on the other hand, would claim the opposite, as they don't perceive the [ɛ] in 'tress' as being significantly different from [e̞] but consider that the [ei̯] in 'trace' sounds completely different (closer to the Spanish diphthong [e̞i̯]). It could well be that English speakers taking the poll would judge the front-ness in [ø] to be more relevant than its roundness and adapt it as [e] where speakers of the local language might consider it more relevant to keep the roundness and adapt it as [o]. It would also be worth pondering whether phonetic realization (the actual sounds pronounced by the speakers) should really take priority over the underlying phonemes as perceived by natives. For instance, the first vowel in Russian's Россия [rɐˈsʲijə] is definitely closer to an /a/ than to an /o/, but Russian speakers should be aware that [ɐ] is a realization of unstressed [o] (as shown in alternations such as он [on] and они [ɐˈni]), so it's not so obvious whether Lasija would be a more faithful adaptation than Losija. And, of course, there might be all sorts of practical considerations. I'm thinking of something like how Barcelona is [bər.səˈlo.nə] in the local Catalan variety, which might justify an adaptation like Pasalona, but it's underlying phonetic representation is arguably /bar.seˈlo.na/ (as shown by Eastern Catalan [baɾ.seˈlo.na]), which might point us towards adapting it as Paselona. Then there's the fact that Spanish [baɾ.θeˈlo.na] is basically as prevalent within the city itself as the Catalan version, which should also favor Paselona (which, when combined with the previous argument, make for a strong case imo), but that could also open a nasty can of worms about the usage of local vs national varieties, regional and minority languages and so on. Then, again, there's no real reason why we'd need to pick just _one_ adaptation; multiple versions could easily coexist without causing too much trouble.
@taimunozhan I appreciate the feedback! The point about native speakers could be augmented by splitting the responders into native and non native groups for each of the questions, but with a low sample size (n ~ 100) and low frequency languages (Arakanese, Malayalam, Farsi) it becomes too hard to get representative data. Someone else can try, though! This particular poll was, indeed, built on the premise that shallow phonetic representation matters more than the underlying representation. This is consistent with the example given by Sonja about preferring Towano over Tolonto. (Its possible to argue that Toronto is an example of a nt > n sound change affecting the phonemic structure of a word while the ɔ > ɐ change in Rosija is allophonic, but the distinction is quite superficial.) Whether or not this understanding is preferable is ultimately out of scope for the experiment, and up to the community at large. I agree with the sentiment that multiple names can coexist; in fact, the segmentation method shown in this video allows us to document not only the plurality option but also the distribution of deviations from it. Unfortunately, there are circumstances where the community will likely have to present just one option at a time (or few of them instead of 20+), such as when naming a Wikipesija article or submitting toki pona language name translations to Debian's iso-codes package (something im hoping to tackle some day). For those cases, its useful to explore the mechanisms behind tokiponisation, while ofc always acknowledging that any such name shouldn't be final.
I've been kid of annoyed by the official translation for Hungary (Magyar-) being ma Mosijo / toki Mosijo / jan Mosijo etc, as gy -> dj -> dij -> tij -> sij seems like an unreasonable amount of steps compared to gy -> j, which even sounds a lot closer in the end. It might be a bias stemming from orthography, but to me hungarian "a" also sounds closer to toki pona "a" than "o", so my translation would be ma/toki/jan Maja
So cool ! As a french speaker i'd probably tokiponise Rouen as Luwan or Wawan. To me french 'r' is too unlike /k/ as it can be a uvular trill or an approximant in some context and is still/used to be an alveolar trill in some dialects so the choice to make it into a /k/ is weird to me. I can completely conceed that this is due to orthographical bias though. The /r/ in the clip you chose sound to me like a uvular trill and this probably caused people to hear a cluster like br or something due to the multiple "explosions" there are in a trill.kj
The people taking the survey were only given the audio snippet that was "cropped" to the word itself, right? In many cases hearing a bit of context around the word (after knowing which word to keep my ears open for) actually made it easier to tell what the word was supposed to be. And especially where exactly it starts and stops.
I imagine having more than one audio would make spellings more consistent Other tests that I think might be interesting: - having audio of someone from the place saying its name and someone who speaks the same language with another accent - if locals commonly refer to a place by a "nickname" testing if the full name or the nickname is chosen
Even for the average french, the pronounciation of Rouen is weird. It sounds like a very local way of saying it, which is definitively how we should tokiponise words, but the /ɑ̃/ sounds almost closer to a /œ̃/.
Here was my attempt while listening to this video before seeing other results (which is absolutely fascinating! There's so much room for tokiponization in ways I didn't expect, with lots of choices that feel really unexpected or make sense in hindsight but of which I would never have thought!) - Lakan - Kenke - Jole - Pemeli - Witon - Pen - Ipajan - Sinson - Kuwan - Lentonlen (I burst out laughing when I saw someone had suggested Kijetesantakalu. I was thinking about it.) I'm French, btw. I'm not sure if it changes anything but it probably comes with its own bias (I knew Rouen, if anything). In any case, thank you so much for this study, it was wonderful! Also it makes me want to study phonetics now. I only had a rough idea of most of the phonology terms used.
wikipedia transcribes it as [a] but the recording is more [ə]; from there the responses are distributed approximately how youd expect, with a e o being common and i u being rarer
mi jan pi toki Kanse la mi sona ala e tan pi kalama "PO" lon nimi "Rouen". ni li musi a. Nice work, and I hope there will be other experiments in the same vein. Actually I might even submit some town or people names from my region on my own to see how people would tokiponize them based on their sound only (with local accent)
Here are mine (native English speaker): lakan kentasi - I was not expecting palatal stops and heard something like [gjøndjɐʃ] ensole sanesi witan pemi - surprisingly the same as you esawan - didn't even hear the F sinson - thought it was qingchong (Chinese province) for a moment, did not hear the P kuwa - I very clearly hear a B at the beginning ("brouen") but ignored it when I caught a glimpse of the transcription. I can see how some might hear this B as a distinct syllable kilantalan - I only even heard 5 syllables (maybe?) but wrote 4 because I didn't know what to do with the middle
my take on this: Wakan (Rakhine) Jentese (Gyöngyös) Ensole (Ndjolé) Leneli (Llanelli) Wetin (Weirton) Pelin (Perm) Osowan (Isfahan) [why did i choose for the first one] Sinpon (Jeunpyoeng) [at first i thought those were lateral fricatives, but i couldnt just keep that in] Kulo (Rouen) Telentolaten (Thiruvananthapuram)
I would have tokiponized these as: (1) Lakan, (2) Jenki, (3) Nisole, (4) Łanełi (IMO x[K] is not really mappable to either or ), (5) Witon, (6) Pen, (7) Pesawan, (8) Sinson, (9) Puku, and (10) Kulentolulen.
@@cmyk8964 for sure but the overwhelming majority of the time this town is mentioned has nothing to do with the locals so im okay disregarding that practice for the joke
My attempt: keeps # of syllables, topiponizes Welsh /ʊ/ as /o/ and Welsh /ɡw/ as /k/ sanwapokinkikokeliwintoposantesilijokokoko (sanwapokinki kokeliwintopo santesilijo kokoko) (Llanfairpwllgwyngyll gogerychwyrndrobwll llantysilio gogogoch)
My results: ma Lakawin ([aj] → "awi" not frequent enough) ma Jenjusi (Rank 2, 1, 1, 2, 3, 2) ma Ensule (Rank 3, 1, 1, 1, 1) ma Sanesili (Rank 1, 1, 1, 1, 5, 1) ma Wijetun (Rank 1, 3, 1, 6, 1) ma Pelen (Rank 1, 1, 1, 2, 1) ma Esowan (Rank 1 only) ma Kinkon (Initial "k" not frequent enough) ma Powa ([u] → "o" not frequent enough) ma Kelejontopolen (In general, ???)
As an english speaker who isn't super familiar with toki pona (so forgive me if i don't get the phonology/phonotactics right!), here's what I came up with watching the video - Lakena (heard some sort of schwa-like sound at the end but not sure where it came from) - Kenkusi - Unsole - Lanesi - Wilekin (wanted *Wiletin here) - Pije - Esawan - Sinpon - Luwas (*Luwos) - Kulemtojalem (i got Kulem- and then had to slow down the video for the second half, hahaha) I think a few of these were swayed by audio quality, i couldn't hear the /f/ in isfahan and the /rm/ in perm at all, and i think the speaker for rouen breathed out at the end? Very interesting video concept, it was a joy to watch and play along!
Usually, *ti is transformed into si; using ki is a bit unusual, because the reason *ti is forbidden is because some languages don't distinguish ti and si. As for *Luwas, the only consonant allowed to end a syllable in Toki Pona is n, so that isn't allowed. Luwa and Luwan are good options though. Similarly, *Kulemtojalem should be Kulentojalen. Everything else is good though
this is so cool bro. I wanna do something like this, and make another (longer) google form, with both place names and name names, do I have your permission to do that? (will definitely give credit)
I've been forever wondering how rounded front diphthongs, like those of Finnish (e.g. /ˈpøy̯tyˌæ/ or /ˈsæy̯nætˌsalo/ or /ˈsy̯øteˣ/) but I'm biased by orthography and recognizing them. I'm not sure the knowledge learned from Gyöngyös necessarily applies, but again, I'm biased so I can't say
also thiruvananthapuram is surprisingly simple to adapt to toki pona when compared to the other place names, it's just long and was pronounced really fast
@@wildstarfish3786 phonemically it doesnt really matter if you write it /t/ or /d/. Phonetically I suppose it could be a nasal release of d, rather than a glottal stop?
It might be more reasonable to let native speakers of that language decide how their names should be adapted, because in this way they will tend to pick (non-standard) allophones and avoid options that are classified as different phonemes. This is arguably a more accurate representation given the phonetic constraints.
@@treborhuang233reasonable? wdym? are you that type of people who watch can ______ speak pure ______? and believe that the original language where word got loaned get to say what they can do with the word?
(was gonna comment this 10 hour ago but got distracted so i forgot lol) i dont know about other people but for me tokiponization make no sense robwords turn into jan waputo when it shouldve been jan lawo even tho toki pona have /l/ they still loan english /ɹ̠/ as /w/ even tho /l/ is closer in both how they sound and how theyre pronounced now i know it **KINDA** forgivable because english /ɹ̠/ is usually pronounce as [ɹ̺̠ʷ] [ɻ̺ʷˤ] and [ɹ̟̈ʷ] but doesnt matter still closer to toki pona /l/ than /w/ (atleast for me)