{"items": [{"author": "Paul", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922544900502", "anchor": "fb-922544900502", "service": "fb", "text": "The wide range of alien monosyllabic commands used here illustrate this https://youtu.be/8SkdfdXWYaI", "timestamp": "1516684172"}, {"author": "Alan", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922554900462", "anchor": "fb-922554900462", "service": "fb", "text": "I did not know that.", "timestamp": "1516690975"}, {"author": "Andy", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922555374512", "anchor": "fb-922555374512", "service": "fb", "text": "Meh!", "timestamp": "1516691742"}, {"author": "Alan", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922555374512&reply_comment_id=922555619022", "anchor": "fb-922555374512_922555619022", "service": "fb", "text": "→ I found out that meh is illegal in Words With Friends ;)", "timestamp": "1516692109"}, {"author": "Jim", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042", "anchor": "fb-922557106042", "service": "fb", "text": "/b s/ is an illegal coda in American English; your example words containing it are mistranscribed and should be /b z/. In general s/z after a stop does not have phonemic voicing. (There are a whole bunch of codas in the list that are illegal under this rule.)", "timestamp": "1516693848"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042&reply_comment_id=922579700762", "anchor": "fb-922557106042_922579700762", "service": "fb", "text": "→ Thanks! That looks like another category of errors in the syllables list I forgot to remove. Fixes:
* Removed /b s/, /g s/, /l d s/, /ng s/, /ng s t/, /m s/, /n d s/, /r b s/, /k z/, /f t z/, /ng k z/, /p z/, and /f z/
* Left /n s/ as in \"ounce\", /r n s/ as in \"Clarence\", /m s k/ as in \"Omsk\", and /m s t/ as in \"Zemst\"
* Changed \"midsts\" from /d s t s/ to /d z t s/, and \"angsts\" from /ng s t s/ to /ng k s t s/
Commit is: https://github.com/.../261a845a8137a719c0253c88bfbb7224bf...
(Weirdly, it looks like many of these errors, such as /b s/, were introduced by the syllable list. The BEEP dictionary it was derived from doesn't even have 'FORBS')", "timestamp": "1516714179"}, {"author": "BDan", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042&reply_comment_id=922610164712", "anchor": "fb-922557106042_922610164712", "service": "fb", "text": "→ I think ounce and Clarence are generally pronounced with a final /n ts/, not with /n s/. And for me Clarence has two syllables, though I guess if I try I can pronounce it with one.", "timestamp": "1516726585"}, {"author": "Bruno", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042&reply_comment_id=922612085862", "anchor": "fb-922557106042_922612085862", "service": "fb", "text": "→ I think 'Cairns' is a better example of the /rns/ ending.", "timestamp": "1516727433"}, {"author": "BDan", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042&reply_comment_id=922613098832", "anchor": "fb-922557106042_922613098832", "service": "fb", "text": "→ Cairns is /\u0279nz/, though. You could theoretically have /\u0279nt\u0361s/, but I can't think of any examples. I don't think you would ever get /\u0279ns/ without the affricate or voicing.", "timestamp": "1516727946"}, {"author": "Bruno", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042&reply_comment_id=922632509932", "anchor": "fb-922557106042_922632509932", "service": "fb", "text": "→ Oh. You're right, it gets voiced. The thing is, I don't really parse 'Clarence', even when pronounced as only one syllable, as having an affricate. (I'm not a native speaker of English.) /\u0279nt\u0361s/ sounds to me more like 'parents'.", "timestamp": "1516736786"}, {"author": "BDan", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042&reply_comment_id=922632884182", "anchor": "fb-922557106042_922632884182", "service": "fb", "text": "→ For me, and I suspect for most native speakers, Clarence and parents are perfect rhymes; similarly, prince and prints are homophones.", "timestamp": "1516736998"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042&reply_comment_id=922634356232", "anchor": "fb-922557106042_922634356232", "service": "fb", "text": "→ BDan: Looks like \"CLARENCE\" is another mistake in the syllable dict. BEEP has it as \"k l ae r ax n s\" which seems much more reasonable to me. I'll remove \" r n s\".", "timestamp": "1516737659"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042&reply_comment_id=922634466012", "anchor": "fb-922557106042_922634466012", "service": "fb", "text": "→ BDan: For \"OUNCE\" BEEP has \"aw n s\". Are you saying you think it should be \"aw n t s\"?", "timestamp": "1516737700"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042&reply_comment_id=922635578782", "anchor": "fb-922557106042_922635578782", "service": "fb", "text": "→ /r n s/ removed: https://github.com/.../0313b3902f1aaf34485886a4426f2ad333...", "timestamp": "1516737981"}, {"author": "BDan", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042&reply_comment_id=922639106712", "anchor": "fb-922557106042_922639106712", "service": "fb", "text": "→ Jeff: Yes. But more importantly, even if it is realized as /a\u028ans/ for some speakers, I strongly suspect that there are no dialects that make a distinction between final /nts/ and final /ns/; 'prince' and 'prints' are always homophonous.", "timestamp": "1516739489"}, {"author": "Bruno", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042&reply_comment_id=922640164592", "anchor": "fb-922557106042_922640164592", "service": "fb", "text": "→ Many American dialects have a glottal stop for post-vocalic /t/ (a glo'l stop, if you wish). I wouldn't be entirely surprised if some people somewhere made that difference between \"prince\" and \"prints\". This seems to be corroborated by https://forvo.com/word/prints/#en .", "timestamp": "1516739978"}, {"author": "Bruno", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042&reply_comment_id=922640369182", "anchor": "fb-922557106042_922640369182", "service": "fb", "text": "→ And a few of the pronunciations here have nothing between /n/ and /s/: https://forvo.com/word/prince/#en", "timestamp": "1516740080"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922557106042&reply_comment_id=922647484922", "anchor": "fb-922557106042_922647484922", "service": "fb", "text": "→ I feel like I distinguish \"prints\" and \"prince\" but probably sneaking up on me with a recording device would be more accurate than my introspecting here.", "timestamp": "1516743462"}, {"author": "Frederic", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922577041092", "anchor": "fb-922577041092", "service": "fb", "text": "Q: Why are those sheep eating that other sheep?
A: Because they are woolves", "timestamp": "1516712709"}, {"author": "Marcus", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922578084002", "anchor": "fb-922578084002", "service": "fb", "text": "In Mandarin Chinese, each legal sound, including tones, has on average 7 characters that are pronounced that way. So it's actually possible to ridiculously overload the sound system of a language and still have it be intelligible.
(An important caveat being that while each character does have a distinct meaning, many words, especially nouns, are two characters long.)", "timestamp": "1516713402"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922578084002&reply_comment_id=922581432292", "anchor": "fb-922578084002_922581432292", "service": "fb", "text": "→ Neat! Chinese has about the same number of vowels as English after you include tones, but much more restrictive syllables. It's CV with an optional n or ng at the end. The V can be two vowels, but only three of the six vowels can go first.
So I count [23 consonants plus null] * [i u \u00fc null] * [a o e i u \u00fc] * [five tones] * [n ng null] which is just 8,640.", "timestamp": "1516715518"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922578084002&reply_comment_id=922581442272", "anchor": "fb-922578084002_922581442272", "service": "fb", "text": "→ But I suspect some are unused, so the crowding is even worse.", "timestamp": "1516715557"}, {"author": "Bruno", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922578084002&reply_comment_id=922583328492", "anchor": "fb-922578084002_922583328492", "service": "fb", "text": "→ The number of valid syllables in Mandarin is actually just over 400 (not counting tones). The wiki article \"Pinyin table\" has an exhaustive list. And while there are a few syllables with a distinct neutral-tone meaning, I suspect the majority don't, so the number of possible syllables including tones is closer to 1600 than 2000.
On the other hand, most words are two syllables, and the Beijing dialect makes significant use of 'er' as a modifier for other syllables.", "timestamp": "1516716924"}, {"author": "Marcus", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922578084002&reply_comment_id=922585713712", "anchor": "fb-922578084002_922585713712", "service": "fb", "text": "→ The seven characters per sound is also a little deceptive because Chinese has some very obscure characters that are basically never used in modern writing, much less in common speech.", "timestamp": "1516718311"}, {"author": "Bruno", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922578084002&reply_comment_id=922585948242", "anchor": "fb-922578084002_922585948242", "service": "fb", "text": "→ Yes, I was thinking about this. Something like: for a subset of all characters that represents how everyday Mandarin is actually spoken, what's the *median* number of characters per syllable (with tone)? I wouldn't be surprised if it was 1.", "timestamp": "1516718461"}, {"author": "Marcus", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922578084002&reply_comment_id=922603907252", "anchor": "fb-922578084002_922603907252", "service": "fb", "text": "→ Randomly scanning through my concise Chinese dictionary it seems like the median is closer to two, though I suppose it depends on what you consider an \"everyday\" word. Also whether your count the neutral tone in the average, since most phonemes never use it even once.", "timestamp": "1516724119"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922578084002&reply_comment_id=922609915212", "anchor": "fb-922578084002_922609915212", "service": "fb", "text": "→ Marcus: scanning the dictionary won't include syllables that aren't used, though, right?", "timestamp": "1516726450"}, {"author": "Marcus", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922578084002&reply_comment_id=922610429182", "anchor": "fb-922578084002_922610429182", "service": "fb", "text": "→ Yes, but unless it isn't used at all in any tone combination I'll notice that there's no character for third tone, for instance. This wasn't super scientific, I was just trying to get a sense.", "timestamp": "1516726679"}, {"author": "Danner", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922602310452", "anchor": "fb-922602310452", "service": "fb", "text": "In b4 worst possible words get picked up for slang. I think they're lovely.", "timestamp": "1516723099"}, {"author": "Jim", "source_link": "https://www.facebook.com/jefftk/posts/922544256792?comment_id=922626042892", "anchor": "fb-922626042892", "service": "fb", "text": "Out of a random sample of 20 words from the missing list, 8 seemed like invalid words, 4 seemed like they would be valid only as loanwords because of their onsets, and two were rare surnames. Of the six remaining words, two had questionable consonant clusters, two were pluralizations of words that also don't exist, and the last two were \"shimp\" and \"thranch\".
Observations:
* ax (schwa) isn't exactly a real vowel - all of the words that use it are either function words, or substitute a different vowel when the word is stressed.
* Some consonant clusters are only allowed if they're generated by adding an affix to a root word, such as /s t s/. They'll be missing if the root word doesn't take that affix.
* Some onsets are only allowed in loanwords, eg /b y/ and /s r/.
* /oy r/ splits syllables
* /m t/ and /m p t/ are the same coda
Sample words:
1879020 z l ow n s t loanword only
5910560 th r oh th s false plural
242440 g y ax s p t illegal - /ax/
5431772 b y er d z loanword only, false plural
38480988 sh ih m p
653540 s r uw th t loanword only, false conjugation
2963070 th r aa n ch
697620 zh er dh d illegal - no real words have /er dh/
38550 v l uh l n
1782648 s m ih l ch rare surname (Smilch)
104060 t oy r th s illegal - /oy r/->/oy.er/
144948 v uh n s t s false plural
1004220 s k w uw dh d illegal - /s k w uw/ -> /s k r uw/
10275650 g y ah k loanword only
1497376 s w ae l p s rare surname (Swalp)
2883573 s y aa f s illegal - /s y aa/
420200 t y ey n th s illegal - /t y ey/
152800 k w ey m t
84588 f ua s th illegal - /s th/ can only follow /ih/ or /ih k/
699600 zh oh l s illegal - not phonemically distinct from Joel's", "timestamp": "1516733390"}]}