Zompist Bboard Again

I've noticed over the weekend my older nephew, who is now 6, with non-rhotic pronunciation (specifically eliding the /r/ in /VrC/) even though he is from the Chicago area, which clearly is not home to non-rhotic varieties, which was very curious.

There are languages with fixed stress, but are there languages with fixed pitch accent? Like maybe last two syllables are always H and the rest are always L: LLLHH
Or are there languages with weight-based pitch accent, just as there is weight-based stress?

One thing I have had questions about is what are the sources of non-presigmatized /p/ in early Germanic languages, considering that *p in Germanic ostensibly comes from PIE *b except in the case of PGmc *sp, where PIE *p > PGmc *f does not occur, but PIE *b is very rare in practice? I know that one source of /p/ in early Germanic was loans, such as Latinate/Romance loans, but then there are words such as PGmc *upp which are not loans (but rather PGmc *upp comes from PIE *upo). So what explains PGmc *p here; does gemination prevent PIE *p > PGmc *f? And why did it geminate in the first place? (I note that PIE *upo also gives OHG oba, NHG ob, which is the more expected outcome.) (No, the expected outcome would be ofa - why did I write the previous...)

Zju wrote: ↑Thu Apr 07, 2022 1:27 pm There are languages with fixed stress, but are there languages with fixed pitch accent? Like maybe last two syllables are always H and the rest are always L: LLLHH
Or are there languages with weight-based pitch accent, just as there is weight-based stress?

This might not be quite what you're looking for but Japanese has specific accent patterns. A word can be unaccented (no pitch downstep at all), head accent (downstep after the first mora), "middle" accented (stepdown after the penultimate mora when there are 3 or more syllables), or tail accented (downstep after the last syllable, if there is a particle or other word following). So I think you could have a language with pitch accent with even more restrictive.

Well, I already knew about Japanese. Are there even languages that allow for arbitrary pitch patterns? They'd be tonal then. I was wondering specifically about fixed or predictable pitch patterns.

Travis B. wrote: ↑Thu Apr 07, 2022 1:32 pm One thing I have had questions about is what are the sources of non-presigmatized /p/ in early Germanic languages, considering that *p in Germanic ostensibly comes from PIE *b except in the case of PGmc *sp, where PIE *p > PGmc *f does not occur, but PIE *b is very rare in practice? I know that one source of /p/ in early Germanic was loans, such as Latinate/Romance loans, but then there are words such as PGmc *upp which are not loans (but rather PGmc *upp comes from PIE *upo). So what explains PGmc *p here; does gemination prevent PIE *p > PGmc *f? And why did it geminate in the first place? (I note that PIE *upo also gives OHG oba, NHG ob, which is the more expected outcome.)

IIRC there was a then-popular now-not-so hypothesys that pre-PGmc *Dn yielded PGmc *TT, which does seem to explain some of the cases. What is really puzzling, then, are the cases of word-initial non borrowed PGmc *p.

Zju wrote: ↑Thu Apr 07, 2022 2:07 pm IIRC there was a then-popular now-not-so hypothesys that pre-PGmc *Dn yielded PGmc *TT, which does seem to explain some of the cases.

You're referring to Kluge's law. As far as I can see, it's actually on its way to become mainstream.

Zju wrote: ↑Thu Apr 07, 2022 2:07 pm What is really puzzling, then, are the cases of word-initial non borrowed PGmc *p.

What would those cases be?

Travis B. wrote: ↑Thu Apr 07, 2022 1:32 pm (I note that PIE *upo also gives OHG oba, NHG ob, which is the more expected outcome.) (No, the expected outcome would be ofa - why did I write the previous...)

No, the part you struck out is correct. PIE *upó -> PGmc. *ufá -> *uvá with Verner's law -> OHG oba (actually, OHG final -a must go back to a PGmc / West Gmc. long vowel, so there must have been some additional lengthening or suffixation going on)..

Travis B. wrote: ↑Thu Apr 07, 2022 1:32 pm I know that one source of /p/ in early Germanic was loans, such as Latinate/Romance loans, but then there are words such as PGmc *upp which are not loans (but rather PGmc *upp comes from PIE *upo). So what explains PGmc *p here; does gemination prevent PIE *p > PGmc *f? And why did it geminate in the first place?

The usual explanation is by way of Kluge's law as metioned by Zju: *up-nó -> *uppá with assmilation resulting in geminated /p/, which would not become spirantised -> *úppa when Germanic switched to initial stress -> up(p) with loss of short final vowel.

IIRC there were about one or maybe two dozen cases. The ones I can recall on top of my head are pack, peak and pike.

Edit: apparently also poke. Curiously enough, they're all pVk.

There are actually quite a few p-initial words reconstructible for Proto-Germanic. Especially since Pokorny, it has been fashionable to pretend "these ones are explained by borrowing," and "these ones are not." But the fact is that nearly every single one meets the criteria of "maybe it's a borrowing, who knows." I can't think of a single one that has any cognates outside of Europe, except in the case of some overly clever etymologies (i.e. Bh deaspirating to B for no clear reason so that it yields P in Germanic), and of course the lending language is almost always extinct so we're just guessing where these forms came from. There really is no difference between words like "pack" (treated as an anomoly) and words like "pamper" (treated as a known borrowing). They both have European cognates based on phonologically unusual PIE forms, and they both lack cognates outside of Europe. Issue solved.

Zju wrote: ↑Thu Apr 07, 2022 1:27 pm There are languages with fixed stress, but are there languages with fixed pitch accent? Like maybe last two syllables are always H and the rest are always L: LLLHH
Or are there languages with weight-based pitch accent, just as there is weight-based stress?

The Lesbian Aeolic dialect of Ancient Greek had a predictable pitch accent in content words... a "recessive" type of accent, as they say in Greek linguistics, if you know the term (basically the accent is placed on the earliest mora possible, up to the fourth-to-last mora). Function words could be unpredictably accentless or accented, though.

linguistcat wrote: ↑Thu Apr 07, 2022 1:47 pm This might not be quite what you're looking for but Japanese has specific accent patterns. A word can be unaccented (no pitch downstep at all), head accent (downstep after the first mora), "middle" accented (stepdown after the penultimate mora when there are 3 or more syllables), or tail accented (downstep after the last syllable, if there is a particle or other word following). So I think you could have a language with pitch accent with even more restrictive.

That definition of middle-accented words seems inappropriate... I mean, what is arìgatō in this classification?

Moose-tache wrote: ↑Sat Apr 09, 2022 7:54 pm There are actually quite a few p-initial words reconstructible for Proto-Germanic. Especially since Pokorny, it has been fashionable to pretend "these ones are explained by borrowing," and "these ones are not." But the fact is that nearly every single one meets the criteria of "maybe it's a borrowing, who knows." I can't think of a single one that has any cognates outside of Europe, except in the case of some overly clever etymologies (i.e. Bh deaspirating to B for no clear reason so that it yields P in Germanic), and of course the lending language is almost always extinct so we're just guessing where these forms came from. There really is no difference between words like "pack" (treated as an anomoly) and words like "pamper" (treated as a known borrowing). They both have European cognates based on phonologically unusual PIE forms, and they both lack cognates outside of Europe. Issue solved.

That's more or less the picture I have as well - there are no p-initial words that have good IE etymologies, except for words that are assumed to be borrowings from other IE languages (like path from Iranian).

Do any other English-speakers often have trouble with non-native English-speakers who do not have proper vowel length allophony, whether by failing to shorten vowels before fortis obstruents or failing to lengthen vowels when not before fortis obstruents, or who fail to (pre)glottalize coda fortis plosives? I find that this very commonly results in my hearing obstruents wrong, as for me voicing is not a good indicator of obstruent fortisness/lenisness, and in general is a major barrier to understanding, much moreso than one might expect offhand.

For me, the difficulty comes with the failure of some non-native speakers to aspirate voiceless initials. While I do have voicing on unaspirated initials, I think it's weaker than the voicing of a resonant or fricative, so I could perceive pig as big (though context would probably disambiguate in the case of a common monosyllable). The odd thing is, when speaking other languages, I can usually perceive the lesser aspiration in Japanese, and the unaspirated voiceless stops of French, as distinct from their voiced counterparts.

Rounin Ryuuji wrote: ↑Wed Apr 13, 2022 8:32 am For me, the difficulty comes with the failure of some non-native speakers to aspirate voiceless initials. While I do have voicing on unaspirated initials, I think it's weaker than the voicing of a resonant or fricative, so I could perceive pig as big (though context would probably disambiguate in the case of a common monosyllable). The odd thing is, when speaking other languages, I can usually perceive the lesser aspiration in Japanese, and the unaspirated voiceless stops of French, as distinct from their voiced counterparts.

I am the same way - in English a lack of aspiration for an initial/non-presigmatized-stressed plosive indicates a lenis plosive for me, whereas in other languages I oftentimes can still hear voiced and voiceless initial plosives distinctly (even though in one case I do remember hearing a recording of Italian pasta that distinctly sounded like it started with /b/, but that may have been due to interference from English pasta).

A random English thing:

A while ago I told my friend about an opera I recently listened to, Koanga (1896) by Frederick Delius.

When I said the name of the opera, based on how I had read it in my head, I intended it to be /koʊˈɑŋgə/, with the first syllable adapting a foreign "cardinal o", like how we use /oʊ/ for Spanish /o/ in Spanish loans into English. (In this case, it was not really a loan, but a fictional "African-looking" name.)

Anyways, what I actually said was apparently more like [kəˈwɑŋgə], since my friend assumed the title of the opera was spelled Kawanga. I feel like this is (anecdotal) evidence for the analysis of the GOAT vowel /oʊ/ as genuinely being a concatenation of /ə/ and /w/, with the /w/ being resyllabified into the onset of the next syllable in [kəˈwɑŋgə].

Another random English thing:

I'm a typical younger American with a complete low back merger, or so I thought, but recently I found myself thinking that I seem to preserve a distinction in open monosyllables - the word law sounds wrong if I pronounce it with a truly low and unrounded [ɑ], and the filler syllable la sounds odd if I pronounce it with a rounded vowel. Likewise the word raw seems to be slightly higher and rounded, in comparison to the name of the Egyptian god Ra. Also paw vs. pa, possibly Shaw vs. shah. Or maybe I'm just talking myself into this? It's hard to tell. And in closed monosyllables there is no question that, for example, caught is pronounced exactly the same as cot.

missals wrote: ↑Mon Apr 18, 2022 8:24 pm A random English thing:

A while ago I told my friend about an opera I recently listened to, Koanga (1896) by Frederick Delius.

When I said the name of the opera, based on how I had read it in my head, I intended it to be /koʊˈɑŋgə/, with the first syllable adapting a foreign "cardinal o", like how we use /oʊ/ for Spanish /o/ in Spanish loans into English. (In this case, it was not really a loan, but a fictional "African-looking" name.)

Anyways, what I actually said was apparently more like [kəˈwɑŋgə], since my friend assumed the title of the opera was spelled Kawanga. I feel like this is (anecdotal) evidence for the analysis of the GOAT vowel /oʊ/ as genuinely being a concatenation of /ə/ and /w/, with the /w/ being resyllabified into the onset of the next syllable in [kəˈwɑŋgə].

No, I just think that [ʊ̆] and [ŭ~w] are practically impossible to distinguish if you’re not a trained phonetician.

missals wrote: ↑Mon Apr 18, 2022 8:24 pm A random English thing:

A while ago I told my friend about an opera I recently listened to, Koanga (1896) by Frederick Delius.

When I said the name of the opera, based on how I had read it in my head, I intended it to be /koʊˈɑŋgə/, with the first syllable adapting a foreign "cardinal o", like how we use /oʊ/ for Spanish /o/ in Spanish loans into English. (In this case, it was not really a loan, but a fictional "African-looking" name.)

Anyways, what I actually said was apparently more like [kəˈwɑŋgə], since my friend assumed the title of the opera was spelled Kawanga. I feel like this is (anecdotal) evidence for the analysis of the GOAT vowel /oʊ/ as genuinely being a concatenation of /ə/ and /w/, with the /w/ being resyllabified into the onset of the next syllable in [kəˈwɑŋgə].

In the dialect here**, (ignoring vowel length and nasalization) stressed /oʊ/ before a consonant is [o̞]~[ɵ̞]* and unstressed /oʊ/ before a consonant outside of careful speech is [ə] but stressed /oʊ/ before a vowel is [o̞w]~[ɵ̞w]* while unstressed /oʊ/ before a vowel outside of careful speech is [əw]. Just from reading the name of that opera, ignoring the rest of your post, I would assume that word would be /koʊˈɑŋgə/, which I would pronounce as [kʰəːˈwãːŋɡə(ː)]. Note that for me the [w] syllabifies as part of the following syllable.

* The dialect here has considerable central-back allophony of mid and high phonemically-back vowels.

missals wrote: ↑Mon Apr 18, 2022 8:24 pm Another random English thing:

I'm a typical younger American with a complete low back merger, or so I thought, but recently I found myself thinking that I seem to preserve a distinction in open monosyllables - the word law sounds wrong if I pronounce it with a truly low and unrounded [ɑ], and the filler syllable la sounds odd if I pronounce it with a rounded vowel. Likewise the word raw seems to be slightly higher and rounded, in comparison to the name of the Egyptian god Ra. Also paw vs. pa, possibly Shaw vs. shah. Or maybe I'm just talking myself into this? It's hard to tell. And in closed monosyllables there is no question that, for example, caught is pronounced exactly the same as cot.

I can't help you there - the dialect here** has a full distinction between /ɑ/ and /ɔ/, realized as (ignoring vowel length and nasalization) [a] and [ɒ] respectively.

** Here is southeastern Wisconsin.

The whole Shaw vs Shah thing is probably just borrowings subsequent to the COT-CAUGHT merger. The merger doesn't mean that speakers of those dialects may never again say an unrounded ah. But they'll have to look outside their inherited English vocabulary to do it. Similarly, Bostonians dressing up as pirates for Halloween are not forced by their non-rhoticism to shout "ah" while brandishing their plastic cutlasses.

In the Dutch language area there are a number of organizations that have a "Word of the Year" competition. The most well known are those organized by the Van Dale dictionary and the Genootschap Onze Taal.

When looking at the words that won (overview here, it is evident practically all of them fall into two categories. They are either loan words (e.g. tiktok, brexit, boomer, twitteren) or they are compound words (anderhalvemetersamenleving "one-and-a-halve-meter-society", laadpaalklever "charging station clinger, i.e. someone who unnecessarily occupies an charging station for an electric vehicle", yogasnuiver "yoga snorter, i.e. someone who lives a healthy lifestyle during the week, but uses party drugs in the weekend", treitervlogger "harassment vlogger" explained (here)).

Rarely a word of more opaque provenance is elected Word of the Year. There are a couple that spring to mind:

Wappie: "conspiracy nut (pejorative)". This definition has come into prominence during the pandemic. An adjectival use is, as far as I know, older and would mean "high, drugged", as in mijn buurman blowt veel en is 's ochtends al wappie = 'my neighbour smokes a lot of weed and is high by morning'. The root verb wapperen means "to flutter, to wave" and how this came to have two derivations that both have the sense of "to be mentally unsound" is quite interesting to me.

Swaffelen: "to rub ones dick onto someone". I have no clue how this word became high-profile, and frankly I am too much of a prude to care.

Tokkie: "lower class, anti-social person (pejorative)". I am unsure whether this was an official Word of the Year-winner, but it has definitely become a crowd favorite. It is used primarily as a noun, but can also be used as an adjective or adverb. For example, deze buurt is erg tokkie = 'this is a very marginal neighbourhood', or zich tokkie gedragen = 'to behave antisocially'. The origin of this word is the surname of a family that once featured in a television series. Dialectally I have also heard boender, with an meaning equivalent to tokkie, but coming from the surname of an infamous, local family.

Raholeun wrote: ↑Fri Apr 22, 2022 5:16 amyogasnuiver "yoga snorter, i.e. someone who lives a healthy lifestyle during the week, but uses party drugs in the weekend",

I think out of the ones that you've mentioned, that's my favorite.