Zompist Bboard Again

Some languages, like English and German, have a complex ruleset for phonotactics. Their ruleset is too big.

Other languages, like Japanese, have a very simple ruleset for phonotactics. Their ruleset is too small.

But Classical Latin has the Goldilocks "just right" ruleset for phonotactics. Syllables in Classical Latin are just the right size and complexity to produce rich words while not producing ugly-sounding words. That's one reason why the Catholic Church has insisted that Latin be the universal language of the Church.

What are these rules?

HolyKnowing wrote: ↑Sun Apr 21, 2024 1:43 am Some languages, like English and German, have a complex ruleset for phonotactics. Their ruleset is too big.

Other languages, like Japanese, have a very simple ruleset for phonotactics. Their ruleset is too small.

What does it even mean for a language to be ‘too big’ or ‘too small’? It’s just… how the language works. If you dislike their phonaesthetic, fine, but please don’t pretend it’s some kind of objective judgement. (For instance, I suspect you’d greatly dislike Coptic and Dorig [see section 4.2], which are two of my favourites when it comes to phonotactics.)

What are these rules?

It’s surprisingly hard to find an answer for this, but the maximal structure would appear to be something like (s)(C)(r)V(r)(C)(s). For more detail see, for instance, Latin syllable structure in typological perspective (Lehman n.d.).

bradrn wrote: ↑Sun Apr 21, 2024 4:28 am It’s surprisingly hard to find an answer for this, but the maximal structure would appear to be something like (s)(C)(r)V(r)(C)(s). For more detail see, for instance, Latin syllable structure in typological perspective (Lehman n.d.).

I think (s)(C)(l,r)V(l,r,N)(C)(s) is better, with the caveat that in three-consonant clusters Cs had to be a stop and some other things. Valid onsets iirc are

/p t k kʷ b d g (gʷ) f s h m n w l r j/ (/gʷ/ only in intervocalic /n.gʷ/)
/pr tr kr br dr gr fr/
/pl kl bl gl fl/
/sp st sk skʷ (sm)/
/spr spl str skl skr/

Codas are more difficult

/p t k b d g s m n l r/ (voiced stops rare)
/rp rt rk rb rd rg lp lk lb lg/ (all rare and I don't know examples for most of them but they should be possible at least theoretically)
/mp nt nk mb nd ng/ (not so sure about these)
most /Cs/ probably
/mps nts nks lps lks rps rks/ (only in third declension nom.sg. forms)

But bear in mind phonotactic rules have fairly little to do with how a language sounds. Valid Latin syllables can be made to sound like English (/spraŋks hips skeps klor demps dips sklema/) or Japanese (/a riga to gosae̯ masu/). What makes Latin sound like Latin isn't the rules, it's the statistical distribution of syllable types and phonemes.

bradrn wrote: ↑Sun Apr 21, 2024 4:28 am

HolyKnowing wrote: ↑Sun Apr 21, 2024 1:43 am Some languages, like English and German, have a complex ruleset for phonotactics. Their ruleset is too big.

Other languages, like Japanese, have a very simple ruleset for phonotactics. Their ruleset is too small.

What does it even mean for a language to be ‘too big’ or ‘too small’?... please don’t pretend it’s some kind of objective judgement.

No, I'm with Holy on this. Some languages are incorrect in what they think is acceptable, and it's time we had the courage to admit it.

Who gets to say what's acceptable and what is correct?

Moose-tache wrote: ↑Sun Apr 21, 2024 9:49 pm No, I'm with Holy on this. Some languages are incorrect in what they think is acceptable, and it's time we had the courage to admit it.

I know you like to be sarcastic, but recently I’ve been dealing with enough opinions like this to feel that the encouragement really isn’t necessary…

Actually English phonotactics are almost identical to those of Latin, you just have to add some more sC onsets, a few more RC codas and then a indiscriminate post-coda /s~z θ t~d/. Suck it, Catholics.

While one cannot say that some languages are "too complex" or "too simple", I do find the middle attractive, and Latin is just right for me.

Darren wrote: ↑Mon Apr 22, 2024 6:29 am Actually English phonotactics are almost identical to those of Latin, you just have to add some more sC onsets, a few more RC codas and then a indiscriminate post-coda /s~z θ t~d/. Suck it, Catholics.

Ah, the strengths of English.

I would be remiss not to mention Piotr Gąsiorowski's paper and its analysis of English in particular and the Indo-European family more generally (including Latin) including a series of "presigmatised stops" (which give the paper its name: "SPecial STatus: Presigmatised Stops"). It argues the resulting analysis of the languages' phonotactics is simplified by avoiding having to appeal to otherwise irregular behaviour of one segment in particular.

Ketsuban wrote: ↑Mon Apr 22, 2024 4:33 pm I would be remiss not to mention Piotr Gąsiorowski's paper and its analysis of English in particular and the Indo-European family more generally (including Latin) including a series of "presigmatised stops" (which give the paper its name: "SPecial STatus: Presigmatised Stops"). It argues the resulting analysis of the languages' phonotactics is simplified by avoiding having to appeal to otherwise irregular behaviour of one segment in particular.

Appealing on language-internal grounds but IMO the typological case is overstated. Qiang has /ʂp st ʂk ʂq ɕtɕ ʐb zd ʐg ʑdʑ xk xtɕ xtʂ ɣz ɣdʑ ɣdʐ χq ʁd ʁdʑ ʁdʐ/; obviously none of these are units. Hiw has, assuming /w/ is treated as the labiovelar counterpart of the fricative /ɣ/, /βt ɣt wt jw/ as its only sonority-violating clusters. (François cites Wright 2004's alternate sonority model as "manag[ing] to explain the frequent unorthodox behaviour of /s/".) Japhug has, to give some examples, /ft βd st zd ltʰ ld ɕt rt rd jt jd xt ɣd χt ʁd mt md nt nd/, and actually disprefers plosives in C1 position. And so on.

The really unusual thing about PIE is that its non-Anatolian descendants shifted all inherited fricatives except *s to zero. If Qiang had unconditioned loss of dorsal fricatives, it could have presigmatized stops as well. The OE alliteration evidence seems reasonable, though.

Does alliteration exist as a poetic device in the languages of the Southeast Asian sesquisyllabic erosion area? If so, can C1- alliterate with P.C1-?

/s/ is always weird phonotactically. Some say it's a vowel.

Nortaneous wrote: ↑Mon Apr 22, 2024 6:11 pm Does alliteration exist as a poetic device in the languages of the Southeast Asian sesquisyllabic erosion area? If so, can C1- alliterate with P.C1-?

It does, and it depends.

Darren wrote: ↑Tue Apr 23, 2024 4:43 am /s/ is always weird phonotactically. Some say it's a vowel.

You could (might? should?) treat the <i> in Mandarin <si zi ci shi zhi chi> (pinyin) as sibilant vowels.

Although in Latin, /s/ is kinda the opposite: an extra element at the syllable margin. If you take out the syllable-margin /s/, Latin phonotactics calms down a lot.

“/s/ at the margins” does seem to be a fairly common IE thing, but it’s not just an IE thing. Muskogee (indigenous to SE north America) has this, and IIRC you also see it in some Iroquoian and Algonquian languages.

My guess is that /s/ (and other fricatives) are common as syllable- (or word-) margins due to their audibility. For /s/, at least, since it involves just the front of the tongue, it’s pretty easy to combine articulatorily with labials or dorsal, at least.

kodé wrote: ↑Tue Apr 23, 2024 10:42 pm My guess is that /s/ (and other fricatives) are common as syllable- (or word-) margins due to their audibility. For /s/, at least, since it involves just the front of the tongue, it’s pretty easy to combine articulatorily with labials or dorsal, at least.

An interesting example of fricatives as syllable margins is Dayang Pumi, which allows, in addition to the sibilant-sibilant initials /sʃ zʒ ɕʃ/, the clusters of fricative and homorganic plosive /ɸp ɸpʰ βb st stʰ zd ʃtʃ ʃtʃʰ ʒdʒ χq χqʰ ʁɢ/, and no other fricative-obstruent clusters.

Moose-tache wrote: ↑Tue Apr 23, 2024 7:03 pm

Nortaneous wrote: ↑Mon Apr 22, 2024 6:11 pm Does alliteration exist as a poetic device in the languages of the Southeast Asian sesquisyllabic erosion area? If so, can C1- alliterate with P.C1-?

It does, and it depends.

So the answer to the second part seems to be 'No'. Alliteration, at least for alliterative quasi-reduplication, requires both minor and major syllables to agree (absent minor syllables agree), and moreover to agree in the complete onsets. The same rule seems to apply to the Northern Thai quadrisyllabic intensives, with their ABAB onset pattern, though the A-syllable can have too much substance to be argued to be a minor syllable.

Darren wrote: ↑Tue Apr 23, 2024 4:43 am /s/ is always weird phonotactically. Some say it's a vowel.

Deep.

kodé wrote: ↑Tue Apr 23, 2024 10:42 pm You could (might? should?) treat the <i> in Mandarin <si zi ci shi zhi chi> (pinyin) as sibilant vowels.

Although in Latin, /s/ is kinda the opposite: an extra element at the syllable margin. If you take out the syllable-margin /s/, Latin phonotactics calms down a lot.

“/s/ at the margins” does seem to be a fairly common IE thing, but it’s not just an IE thing. Muskogee (indigenous to SE north America) has this, and IIRC you also see it in some Iroquoian and Algonquian languages.

My guess is that /s/ (and other fricatives) are common as syllable- (or word-) margins due to their audibility. For /s/, at least, since it involves just the front of the tongue, it’s pretty easy to combine articulatorily with labials or dorsal, at least.

Good explanation of mechanics.

HolyKnowing wrote: ↑Wed Apr 24, 2024 12:34 pm

Darren wrote: ↑Tue Apr 23, 2024 4:43 am/s/ is always weird phonotactically. Some say it's a vowel.

Deep.

That’s not a joke.

kodé wrote: ↑Tue Apr 23, 2024 10:42 pm

Darren wrote: ↑Tue Apr 23, 2024 4:43 am /s/ is always weird phonotactically. Some say it's a vowel.

You could (might? should?) treat the <i> in Mandarin <si zi ci shi zhi chi> (pinyin) as sibilant vowels.

For sure, phonetically.

Phonemically, like other aspects of Mandarin, it's a puzzle. By normal phonological principles I think we can argue that the 'vowels' in <si zi ci shi zhi chi> are all the same phoneme. But, is it the same vowel as in <pi bi di ni li> etc.? And what the heck is <ri>?

(FWIW Jerry Norman seems to think they are all /i/, but doesn't give detailed arguments. He has a nice five-vowel system /i y e u a/, but you end up with a shitload of allophony.)

kodé wrote: ↑Tue Apr 23, 2024 10:42 pmMuskogee (indigenous to SE north America) has this, and IIRC you also see it in some Iroquoian and Algonquian languages.

Source?