Re: Two seperate questions: Rhoticity/Topic-Comment

This went only to Dana Nutter instead of to the list.  Sorry!

li_sasxsek@nutter.net skrev:
> li [Behalf Of Adam F.] mi mi tulis la
>> These questions I have are things I've always been
>> curious about. Does anyone know why some languages have
>> the tendency to have weak rhoticity or to drop so-called
>> "r" sounds?
[snip]
>
> There is also a bit of the same thing in Swedish and
> Norwegian, but R isn't just dropped. There is usually some
> effect on the following sound unlike English where there
> is usually some sort of change to the previous vowel.
That is actually a whole nother thing, namely a matter of
post-dental consonants /t d n l s/ assimilating to a
preceding apico alveolar /r/ -- probably actually a trill at
the time --, eventually becoming subapico-alveolar(*). Later
the /r/ dropped, not by becoming vocalized but because it
had become auditorily redundant or indistinct before the now
homorganic following consonant -- or by becoming assimilated
to it;, there was probably a good deal of variation and
mixture on the local and idiolectal level, and the net
result is that you sometimes get a geminate and sometimes a
simple subapico-alveolar, with morpheme structure playing
some rôle for the outcome.

(*) Actually the subapico-alveolars, which I propose to
symbolize at leas ad_hoc with [t_r\] etc., certainly aren't
retroflexes of the kind found in Dravidian languages. They
are in some Swedish dialects in contrast with true (sub)apico-
cacuminal consonants derived from a similar assimilation of
[r`] and a following (post)dental. This [r`] originally
arose out of Old Swedish /rD/. Later Old Swedish /l/ for
some reason merged into [r`] in most dialects that had the
latter sound, and so [r`]+C sequences arose. NB that in this
parallel case vocalization is out of the question, since
[r`] isn't a sound expected to vocalize. It is BTW probable
both that /l/ > [r`] and "l-coalescence" were formerly more
widespread, and that Old Swedish /rl/ also became [r`]. The
evidence for this is that many lects which otherwise have
r-coalescence have biphonic [rl] (with some variation in the
actual articulation of /r/) or plain postdental [l] for
underlying /rl/, so that you'll hear [,svat_r\: 'p&:rla] for
_svart pärla_ 'black pearl'. The reason why [r`] -- whether
as a realization of /l/ or as a realization only of former
/rD/ -- has been receding is that it has been regarded as
(too) rustic. A 17th century author seeing things from the
opposite angle claimed that the people of the capital had
weak tongues and were unable to pronounce [r`] -- probably
as a result of a lot of people in the capital descending
from Low German speakers. The distinction between /v/ from
earlier /w/ and /w/ or /W/ from earlier /hw/, and the
distinction between /3\/ from earlier /O/ and /o/ went the
same way. IMNSHO the language didn't get any better
(certainly not any richer!) for that!

Finally there are dialects (including traditional
Gothenburg dialect) where subapico-alveolars and
postdentals later merged with each other as plain apico-
alveolars. My 8-year-old was very confused by what to him
was the homonymy between _mord_ 'murder' and _mod_
'courage'. I simply had to point out to him that one of the
words is spelled with an _r_, and *may* be pronounced
differently, while the other mightn't.


li_sasxsek@nutter.net skrev:
> li [Philip Jonsson] mi tulis la
>
>> FWIW the only 'non-rhotic' Romance lang I know of is
>> Catalan, but there probably are others on the
>> dialectal level.
>
> What about the French final "-er" which is pronounced
[e]? Although it
> doesn't affect the other combinations like "-ir", "-eur",
etc.

It appears that some 16th-17th century French dialects
(presumably including lower-class Parisian) had r-
vocalization -- and it was [r\] at the time, as shown by
the later tendency for /r/ to merge into /z/ in Low
Parisian --, but this was later reversed by influence from
more prestigeous lects, except in final _-er_, presumably
because it had gone as far as total loss in this
particular position.


Eric Christopherson skrev:
> On Dec 9, 2006, at 7:05 AM, Benct Philip Jonsson wrote:
>> Among sonorant consonants the scale actually looks
>> something like:
>>
>>    1. [n]   > [n=] > [A~] > [A]
>>    2. [m]   > [m=] > [u~] > [u] or > [m=] > [A~] > [Á]
>>    3. [D]   > [j] > [i] or > [0]
> [...]
>
> Are you using > to mark sound changes, or to show ranking
> on the sonority hierarchy? If it's the latter, you have
> the direction backwards (except that I think *any* sound
> would be more sonorous than 0, i.e. no sound, so [D] > 0
> would be correct).
It is the former (i.e. sound change), otherwise I would have
used -> rather than > ! I should have, and intended to,
point it out in words, but I forgot to, due to cronic hurry.
Sorry for that!

>
>> FWIW the only 'non-rhotic' Romance lang I know of is
>> Catalan, but there probably are others on the
>> dialectal level.
>
> I know at least some varieties of Portuguese delete final
> /r/, and I think I heard about some varieties of Spanish
> that do as well.
Aha. But AFAIK Catalan is the only Romance lang where it is
not seen as substandard.

daniel prohaska skrev:
> Carsten,
>
> There are many areas in the German speaking countries
> where [r] or [4] is still heard. It was the dominant
> realisation until the late 19th century, so you will find
> it in the typical “Rückzugsgebiete” (areas of
> withdrawal). These are usually rural areas with a high
> frequency of local dialect speakers, as opposed to
> urbanised or suburbanised areas with close-to-standard
> regional dialects. There are many areas in the North where
> conservative varieties of Low Saxon realise /r/ as
> [r] or [4]. Also in the South-West, especially rural
> Swabia, large parts of Switzerland, and the Upper
> Franconian and many Bavarian dialects as you mentioned
> already, but also in the Hesse and Palatinate. The
> moribund dialects of the areas lost after the second
> world war predominantly had [r].
FWIW my German-born mother and grandparents have/had apical
/r/ *and* vocalization! Probably due to dialect mixture. My
grandfather and his children were all born in Pomerania,
although he had Polish as L1 (or was a from-birth
bilingual), while my grandmother was Ukrainian-born, but
only eight when she came to Pomerania (in 1917, no wonder).
After 1944-45 the family lived in Berlin and then Swabia
from the mid-1950s. My mother and aunt have lived in Sweden
since the early 1960s, but my aunt who remained more in
touch with Germany due to working for a German firm at some
point shifted to [R]. I don't know about my uncle who
remained in (southern) Germany.

Andreas Johansson skrev:
> Quoting taliesin the storyteller <taliesin-
>> To be exact, the r and the following consonant fuse into
>> a retroflex in some dialects (including mine), meaning
>> some Norwegian and Swedish dialects have a full series of
>> retroflex consonants. Neat, huh?
>
> "Some" dialects is rather an understatement - it's a
> feature of rikssvenska, and of the speech of I'd think
> most Swedes.
I would contend that _rikssvenska_ is a norm of
orthography, grammar and vocabulary rather than of
pronunciation. In particular there have always been
academics from Scania who have spoken _rikssvenska_ with
[R] and without r-coalescence. NB that in grammar and
vocabulary their speech is markedly different from Scanian
dialect, and also in pronunciation, but only as regards the
absent or milder diphthongization of long vowels, and not
in the realization of /r/, which is always some kind of
uvular for these people.  The fact that most people heard
in radio and TV speak more-or-less _rikssvenska_ with a
more-or-less mid-Swedish pronunciation is another matter.

Vive la difference!

(My .sig is especially relevant today, methinks! :-)


/BP 8^)>
--
Benct Philip Jonsson -- melroch at melroch dot se

    a shprakh iz a dialekt mit an armey un flot

                                 (Max Weinreich)

Reply