Linguistic Miscellany Thread

[zompist]

Wikipedia has a good article on brioche. It's basically bread enriched with butter and eggs.

The phrase is pretty straightforward:

Qu'ils mangent de la brioche.
sub they.m eat-subj.3p of the.f brioche
Let them eat some brioche.

This is a standard way to express a wish or command; it can't be taken as a question.

(This doesn't mean the German translator was wrong. Literalness usually makes for a stiff translation, and maybe that would be the case here. Even "Let them..." sounds maybe too imperious in English; a more idiomatic translation might be "So they should eat brioche.")

[Kuchigakatai]

Romance third-person imperatives, man.

Actually, I think "Why don't they eat cake, then?" is a fairly good translation, especially in terms of getting the intention across. But "So they should eat cake" is a bit better still.

I swear I'm never all that happy about how to translate this kind of construction. "He/she/they should [VP]" is about as good as it gets. Sometimes I use "make sure that he/she/they [VP]", especially when it's said in an angry tone. ¡Que se vayan! 'Make sure they go away!'

[Raphael]

Thank you!

[Raphael]

Over the last couple of weeks, I have several times - sometimes in books, sometimes on the internet - come across the spelling "deejay" for "disc jockey" in English-language texts.

WHY?

Why not write "DJ"?

[Linguoboy]

Over the last couple of weeks, I have several times - sometimes in books, sometimes on the internet - come across the spelling "deejay" for "disc jockey" in English-language texts.

WHY?

Why not write "DJ"?

Ok.

[WeepingElf]

Because a deejay is not the same as a DJ! In hip-hop, the DJ is the one who puts the records on the turntables, and the MC (for "Master of Ceremony") is the rapper. In Jamaican sound system culture, the deejay is the one who does the "toasts", i.e. the announcements, thus more or less the equivalent to the hip-hop MC, while the one who plays the records, i.e. what in hip-hop is the DJ, is the selector. Of course, hip-hop music was strongly influenced by Jamaican sound system culture, and "deejay" is from the abbreviation "DJ" (for "disk jockey", originally a music presenter at a radio station), but their meanings are different.

[Vilike]

Is it the same story for MC=emcee?

[Raphael]

Ooops, that explains that.Sorry, my bad.

[Pabappa]

I wonder how languages with Suffixaufnahme or other types of case stacking handle things like https://www.pitt.edu/~dash/type2035.html and "There's a Hole in the Bottom of the Sea". Would you end up with a word with twelve locative case endings stacked on? Im guessing not .... even if people speaking those languages deliberately chose to construct a sentence like these, there's probably some understood limit, perhaps two or three, beyond which you stop adding. Poswa doesnt have Suffixaufnahme, but I might try a translation in which I assume it does, perhaps for poetic effect.

edit:I should add perhaps that i meant more particularly the second link, where there are no coordinating conjunctions like "that" involved, ... the whole thing is one single clause.

[Pabappa]

https://en.wiktionary.org/wiki/%E0%A4%A ... 0%E0%A4%B8

Sanskrit dvācatvāriṃśat to Hindi bayālīs in (I think) about 2500 years. indivisdual steps spelled out at https://en.wiktionary.org/wiki/Wiktiona ... 0%E0%A4%B8

[Vijay]

Isn't it beyālīs?

And that's how every modern Indo-Aryan numeral that isn't a multiple of ten is. That's why in all of those languages, you have to memorize every single number from 1 to 100 or you'll get them wrong.

[Pabappa]

fifteen languages ranked according to how good their spelling systems are ("orthographic depth", as they call it). English scores the worst overall, but beats French on one of the two metrics. The two metrics are basically "how easy is it to spell a word, once Ive heard it?" and "How easy is it to pronounce a word, once Ive seen it?" The only surprise to me is that they said German is only a tiny bit easier to spell than English. Turkish scored the best, followed closely by French (the "fro" in the database is a radical reworking of French, not an independent language).

[Richard W]

fifteen languages ranked according to how good their spelling systems are ("orthographic depth", as they call it). English scores the worst overall, but beats French on one of the two metrics. The two metrics are basically "how easy is it to spell a word, once Ive heard it?" and "How easy is it to pronounce a word, once Ive seen it?" The only surprise to me is that they said German is only a tiny bit easier to spell than English. Turkish scored the best, followed closely by French (the "fro" in the database is a radical reworking of French, not an independent language).

One of the conclusions should be that some rules need to be taught rather than abstracted from examplesː

As a reminder, a phonemic reading of an English word often does not work because of its high number of grapheme-to-phoneme possibilities.
For instance the grapheme <u> can either correspond to /ʌ/ (as in "hug"), to /ju:/ (as in "huge"), to /3:r/ (as in "cur") or /jU@:/ as in "cure".

The writing of French /ʒ/ is also a good example.

One has to wonder how much damage was done by stripping out the length mark <ː> when very significant.

[Linguoboy]

One has to wonder how much damage was done by stripping out the length mark <ː> when very significant.

It would help explain why German does so surprisingly poorly. The author notes that, in the German case, "[t]he most common error was not to guess if a /t/ should be written as a single <t> or as a double <t>"--a choice determined almost entirely by whether the preceding vowel is long or short.

[Kuchigakatai]

A post about the development of the French rounded vowels, front and back. As with my previous post on Mandarin‑Cantonese tone correspondences, this is not all that original in content but in succinctness and what I perceive to be clarity of presentation, particularly in the use of examples. To discuss this, I think it's easiest to look at early Old French and then look both backwards to Classical Latin and forwards to late Old French, when the modern system of rounded vowels was basically established (with only a few changes happening since then).


In Early Old French, some time around the 10th and 11th centuries, we begin with a back vowel system of /ʉ o ɵ ɔ/. Besides this, we'll also consider the diphthong /wɛ/ and words with /al ɛl jɛl el ɔl ol/ (in Late Latin, so this may include Classical /ɪl/, etc.) followed by a consonant (typically inside the last syllable).

In general:
- /ʉ/ descends from Classical Latin /u:/ (lūna 'moon' > lune /ˈlʉnə/, nūllum 'not any X' > nul /nʉl/)
- /o/ from Classical /ʊ o:/ inside closed syllables (curtum 'trimmed down' > cort /kort/ 'short'), /ʊ o/ in unstressed sylables (*dupl‑ā‑tis > doblez /doˈblets/ 'you guys double sth'), and also /ɔ/ in unstressed open syllables (colōrem /kɔˈloːrɛm/ 'colour' > color /koˈlɵr/)
- /ɵ/ from Classical /ʊ o:/ inside open stressed syllables (calōrem 'heat' > chalor /tʃaˈlɵr/) (Note that "ɵ" is a bit notational: considering that it's the counterpart at the back of the vowel space of Classical /ɪ e: ɔɪ/ > /ei/, as in tēla > teile [ˈteɪlə], "ɵ" may have been a diphthong such as [ɵw], or a flattened form of it, [o:], even if that would've been the only long vowel... Naturally I personally favour the [ɵ] conjecture, paralleling /ʉ/ for the developments below.)
- /ɔ/ from Classical /ɔ/ in closed syllables (tostum 'roasted' > tost /tɔst/ 'fast') or /aw/ (aurum 'gold' > or /ɔr/, causa 'reason' > chose /ˈtʃɔzə/ 'thing')
- /wɛ/ from Classical /ɔ/ in openstressed syllables (dolium 'trick' > dueil /dwɛʎ/)

Early Old French had a pretty ambiguous writing system for the mid‑rounded vowels. Although /ʉ/ could only be <u> and /ɔ/ could only be <o>, both /o/ and /ɵ/ could be written either <u> or <o>, e.g. <chalor/chalur> /tʃaˈlɵr/. You can actually find manuals of Old French that recommend looking up both the modern French descendant and the Latin etymology, besides medieval rhymes, to find what vowel phoneme a word had.

Examples for the lateral groups: caballōs 'horses' > chevals /tʃəˈvals/, caelōs 'heaven' > ciels /tsjɛls/, capillōs 'someone's hair on the head' > chevels /tʃəˈvels/, colaphum 'blow, hit' > colp /kɔlp/, multum 'a lot' > molt /molt/.



By the 12th century, the lateral coda of /alC jɛlC elC ɔlC olC/ began to vocalize as [w], also merging the height distinction of the mid‑vowels: [awC jewC ewC owC owC]. /ɛlC/ became something like [jaw] before quickly losing the [j]-like first segment.

Then, as the 13th century went on, a new system was created by means of a great vowel shift where /ʉ ɵ/ were further fronted to the new /y ø/, /ew/ joined /ø/ (including /jew/ > /jø/), and /wɛ/ became /œ/. The big vowel space left at the high back corner was then simultaneously filled in by /o/ and /ow/ which merged into the new high /u/, and the gap left by that was filled by /aw/ which became the new /o/ (including /(j)aw/ > /o/). (Meanwhile, /ɔ/ stayed there peacefully, not moving an eighth of an inch.)

Or putting it another way:
- /ʉ/ > /y/
- /ɵ el/ > /ɵ ew/ > /ø/, and /jɛl/ > /jɛw/ > /jø/
- /wɛ/ > /œ/
- /o ol ɔl/ > /o ow ɔw/ > /u/
- /al ɛl/ > /aw jaw/ > /o/
- /ɔ/ > /ɔ/

The writing system also settled with /y/ = <u>, /ø œ/ = <eu> (except for occasional odd spellings for /œ/: cœur /kœr/, cueillir /kœʎir/ > modern /kœjiʁ/), /u/ = <ou>, /o/ = <au>, /ɔ/ = <o>. The change of /ɛl/ > /jaw/ > /aw/ > /o/ explains the orthographic changes seen in bel > beau > beau (plural bels > beaus/beaux > beaux).

Since then, things haven't changed much. /y/ and /u/ have been very stable. /ø/ and /œ/ have undergone a near‑complete merger, with /ø/ appearing in open syllables and before coda /z/, and /œ/ in closed syllables not ending in /z/, except for a few instances of /ø/ before other codas. /o/ and /ɔ/ have also merged into /o/ before coda /z/ (and coda /s/ in Middle French, with the /s/ usually disappearing afterwards) and word‑finally in an open syllable (except in Belgium: peau [po], pot [pɔ], which are both [po] elsewhere).

These rules about phonotactics involving coda consonants also apply today now that word‑final -e /ə/ is gone (phonemically), so chose Late OF [ʃɔzə] > [ʃɔz] > [ʃoz] > [ʃo:z] (with an /ɔ/ > /o/ change due to coda /z/), and chaleur Late OF [ʃaˈlør] > [ʃaˈlœr] > [ʃaˈlœːʁ] (with an /ø/ > /œ/ change due to the syllable being closed with coda /r/). These rules have also had effects in the writing system, since the effects of coda /z s/ mean that /o/ is sometimes spelled <os> (where <s> = /z/) or <ô> now, even though <o> = /ɔ/ overall.



Thus we end up with the following in late Old French:

Classical Latin ca. 1st c. BC > Early Old French ca. 11th c. > Late Old French ca. 14th c.

/u:/ > /ʉ/ > /y/
- lūna > lune [ˈlʉnə] > lune [ˈlynə] (> modern [lyn])
- nūllum > nul [nʉl] > nul [nyl] (> modern [nyl])

/ʊ o:/ (some syllable types) and /ɔlC {ʊ,o:}lC/ > /o ɔlC olC/ > /u/
- curtum > curt/cort [kort] > court [kurt] (> modern [kuːʁ], Quebec [kʊuʁ])
- *dupl‑ā‑tis > dublez/doblez [doˈblets] > doublez [duˈbles] (> modern [duˈble])
- colōrem > colur/color [koˈlɵr] > couleur [kuˈlør] (> modern [kuˈlœːʁ], Quebec [kuˈlaœ̯ʁ])
- colaphum > colp [kɔlp] > coup/coulp [kup] (> modern coup [ku])
- multum > mult/molt [molt] > mout/moult [mut] (the word is now obsolete, replaced by beaucoup [boku]; modern spelling pronunciation moult [mult])

/ʊ o:/ (some syllable types) and /aɪlVC {ɪ,e:,ɔɪ}lC/ > /ɵ/ and /ɛlC elC/ > /ø/ (...and also /ɛlVC/ > /jɛlC/ > /jø/)
- calōrem > chalur/chalor [tʃaˈlɵr] > chaleur [ʃaˈlør] (> modern [ʃaˈlœːʁ], Quebec [ʃaˈlaœ̯ʁ])
- caelōs > ciels [tsjɛls] > ciels/cielx [tsjɛws] > cieux [sjøs] (> modern [sjø])
- capillōs > chevels [tʃəˈvels] > chevels/chevelx [tʃəˈvews] > cheveux [ʃəˈvøs] (> modern [ʃəˈvø], les cheveux [leˈʃfø])

/ɔ/ (some syllable types) and /aw/ > /ɔ/ > /ɔ/
- tostum > tost [tɔst] > tost [tɔst] (> Middle French [tost] > modern tôt [to])
- aurum > or [ɔr] > or [ɔr] (> modern [ɔːʁ], Quebec [ɑɔ̯ʁ])
- causa > chose [ˈtʃɔzə] > chose [ˈʃɔzə] (> modern [ʃoːz], Quebec [ʃouz])

/ɔ/ (some syllable types) > /wɛ/ > /œ/
- dolium > duel/dueil/doel/doeil [dwɛʎ] > dueil/deuil [dœʎ] (> modern deuil [dœj])

/alC ɛlC/ > /alC ɛlC/ > /o/
- caballōs > chevals [tʃəˈvals] > chevals/chevalx [tʃəˈvaws] > chevaux [ʃəˈvos] (> modern [ʃəˈvo], les chevaux [leˈʃfo])
- bellōs > bels [bɛls] > beaux/biaux [bjaws] (or similar: [be̯aws]) > beaux [bos] (> modern [bo])

Some of the above phonemes are fed words through a lot of borrowings from Latin too, e.g. Medieval Latin auctorizare was borrowed with /o/ due to the spelling <au>: autoriser [otɔriˈze(r)], and trucidare with /y/: trucider [trysiˈde(r)].

In the Middle Ages and the Renaissance, the convention to borrow Latin words involved getting rid of the -Vm endings. In the early modern period, this is sometimes not followed, and in such words -m is [m], and -um specifically is pronounced /ɔm/ (not /ym/). So Latin auditorium was borrowed as auditorium [oditɔʁiˈ(j)ɔm] 'auditorium' (if the medieval convention had been followed, we'd now have auditoire [odiˈtwaːʁ] instead, which does happen to exist as well meaning 'the audience').

[Estav]

By the 12th century, the lateral coda of /alC ɛlC elC ɔlC olC/ began to vocalize as [w], also merging the height distinction of the mid‑vowels: [awC ewC ewC owC owC].

I think [ɛlC] actually merged with [alC], not [elC], except for when the [ɛ] was originally in an open syllable (where it usually diphthongized to [jɛ]). Hence beaux, peaux, nouveaux, etc.

[Ryusenshi]

As a reminder, a phonemic reading of an English word often does not work because of its high number of grapheme-to-phoneme possibilities.
For instance the grapheme <u> can either correspond to /ʌ/ (as in "hug"), to /ju:/ (as in "huge"), to /3:r/ (as in "cur") or /jU@:/ as in "cure".

It seems their IPA dataset for English isn't very consistent, as I can't imagine how a consistent set would have /ɜːr/ in "cur" but /jʊəː/ in "cure": the first one fits a rhotic accent and the second a non-rhotic one. Especially considering their four examples are completely regular.

[Ryusenshi]

@Ser: thank you, that was very instructive.

- /o/ from Classical /ʊ o:/ inside closed syllables (curtum 'trimmed down' > cort /kort/ 'short'), /ʊ o/ in unstressed sylables (*dupl‑ā‑tis > doblez /doˈblets/ 'you guys double sth'), and also /ɔ/ in unstressed open syllables (colōrem /kɔˈloːrɛm/ 'colour' > color /koˈlɵr/)
- /ɵ/ from Classical /ʊ o:/ inside open stressed syllables (calōrem 'heat' > chalor /tʃaˈlɵr/)

This explains a few vowel variations that are still visible in modern French: pouvoir -> elle peut, vouloir -> elle veut, se mouvoir -> elle se meut. Also, prouver -> une preuve though the verb itself has been regularized to elle prouve; for trouver, the form elle treuve is attested as late as the XVIIth century but is now regular, elle trouve.

[Travis B.]

As a reminder, a phonemic reading of an English word often does not work because of its high number of grapheme-to-phoneme possibilities.
For instance the grapheme <u> can either correspond to /ʌ/ (as in "hug"), to /ju:/ (as in "huge"), to /3:r/ (as in "cur") or /jU@:/ as in "cure".

It seems their IPA dataset for English isn't very consistent, as I can't imagine how a consistent set would have /ɜːr/ in "cur" but /jʊəː/ in "cure": the first one fits a rhotic accent and the second a non-rhotic one. Especially considering their four examples are completely regular.

People forget that regular orthographies need not be one-phoneme-per-grapheme. Sure, English is not a regular orthography by any means, but this is not an example of that.

[Kuchigakatai]

By the 12th century, the lateral coda of /alC ɛlC elC ɔlC olC/ began to vocalize as [w], also merging the height distinction of the mid‑vowels: [awC ewC ewC owC owC].

I think [ɛlC] actually merged with [alC], not [elC], except for when the [ɛ] was originally in an open syllable (where it usually diphthongized to [jɛ]). Hence beaux, peaux, nouveaux, etc.

Thank you! I see it is [jɛlC] that becomes a front rounded vowel later, not [ɛlC]. I'll edit my post with a correction later.

However, why do you say that it merged with [alC]? I mean, it does merge with it eventually, but I'd think the spellings <eau iau> it gets suggest something like [eaw jaw] before getting rid of the first segment and then [aw] proceeding to become [o]. Godefroy's dictionary of Old French mentions the spelling <biel> in the 14th-century Chronicles of Jean Froissart too.

One has to wonder how much damage was done by stripping out the length mark <ː> when very significant.

It would help explain why German does so surprisingly poorly. The author notes that, in the German case, "[t]he most common error was not to guess if a /t/ should be written as a single <t> or as a double <t>"--a choice determined almost entirely by whether the preceding vowel is long or short.

Inadequately carried-out experiments: the second bane of the progress of knowledge. (The first one would be not wanting to learn anything at all, and not allowing others to do so.)