Re: word derivation in sabyuka (some principles)

En réponse à Christophe Grandsire :


> Sorry not to have replied before, but I wanted to wait until I had time enough
> to give thoughful comments.
So did I : I could not answer yesterday ;).

> How are the vowels pronounced? And are there gradation effects that justify the
> classification in "weak", "medium" and "strong" vowels, or just phonetic
> phenomena like the Spanish classification of a, e and o as strong vowels and i
> and u as weak because when a weak vowel and a strong vowel meet they make a
> diphtongue, while when two strong vowels meet there is a syllable break between
> them?
I should have given pronouciation (I wondered if so for a while, but it
was already very long, and I had already given it in another post).
So here we are :
Weak vowels : 'à' = /@/ ; 'i' = /i/ ; 'u' /u/
medium vow. : 'a' = /a/ ; 'e' = /e/ ; 'o' /o/
strong vow. : 'â' = /a@/ ; 'ê' = /ei/ ; 'ô' /ou/

There is some variation in the pronouciation, especially for 'e' and 'o'
which can be mid-close or mid-open depending on position, dialect and so
on (but I'm not yet fixed with it ;))
I'm not sure for the names of those vocalic degrees (in french réduit,
plein and long, borrowed from indo-european terminology), but they
represent some phonological common features.  For instance
  - weak V + stronger vowel =  the first one become a consonant (/@/
originally turned to /?/, then /h/ and now zero)
  - strong V + any V : the second element of the diphtongue consonantize
(respectively 'ah', 'ey' and 'ov')
  - medium V + strong V : an epenthetic 'n' is inserted (/Sa+ou+ni/ 'to
be+PROSP+1SG' > /Sanouni/ 'xanôni' "I will/would be", where PROSP is
"prospective" and 1SG "first person singular" )



> > 'r' because *sr is an illegal cluster which turns to
> > 'sl'
> > (just because I can't pronounce it !!!)).
> >
>
> How do you pronounce 'r' in Sabyuka? Because I cannot even think of a
> pronunciation of 'r' that would be impossible to pronounce after 's'. And if I
> understood correctly 'v' marks /w/? Neat!
Actually 'r' is flapped or trilled 'r'. I have a true problem in pronouncing 'r',
so that my flapped 'r' is a retroflex one (no problem with sanskrit LOL). I cannot
do a proper alveolar/dental 'r' (that's why I gave up studying spanish
and I now study linguistics ;)). And 'sr', though not impossible, is
really hard for me (but easier than 'sxrawbo' in esperanto, the hardest
word in the world !!!).So, I didn't need that sound, but my girlfriend
loves it :(.


> > Examples will make it clearer :
> >
> > V :
> > CV : 'xa' "to be" (an exeption because of final 'a')
> > VC : 'ic' "to go"
> > VCV : 'urà' "water"
> > VCCV : 'aktà' "to lead"
> > CVC : 'leth' "to believe"
> > ClVC : 'smak' "to want"
> > CVCV : 'sabi' "to know, kowledge", manu "man"
> > and so on...
> >
>
> I'd like to see pronunciations :)) . (especially for the difference between 'c'
> and 'k' and the pronunciation of 'x' and 'th' :) )
Yes, you're right I give you the consonantic system :
'p' = /p/
't' = /t/
'z' = /t^s/
'tl' = /t^l/
'c' = /t^S/
'k' = /k/
'q' = /k^w/
'b' = /b/
'd' = /d/
'j' = /d^Z/
'm' = /m/
'n' = /n/
'f' = /f/
'th' = /T/
's' = /s/
'x' = /S/
'h' = /h/ for sabyajà (but pronounced as [x]) : merged from *h and *?
'v' = /v\/ as in sanskrit (there is a lot of variations depending on
dialets and position (post-consonantic, intervocalic)).
'r' = /r/
'l' = /l/
'y' = /j/


> >
> > Then, sabyuka is a nominative language with 6 cases. Referential roots
> > (for instance 'manu' can be directly declined : the declination basis
> > is
> > found by :
> > - deleting every final 'à' *urà > ur-
> > - consonantizing into 'y' final 'i' *sabi > saby-
> > - consonantizing into 'v' final 'u' *manu > manv-
> >
>
> Am I right to think that 'à' is the schwa? :)))

> > NOM : urà, sabyà, manvà
> > ACC : urna, sabyàna, manvàna
> > GEN : uri, sabyi, manvi
> > DAT : uru, sabyu, manvu
> > INST : ullu*, sabyàru, manvàru
> > LOC : ure, sabye, manve
> >
> > * from illegal *urru, that I can't pronounce :).
>
> Why? If 'r' is a flap, why don't you do like Spanish which takes 'r' as flap
> but 'rr' as trill, which sounds near enough to a geminated flap :)) . Any other
> pronunciation of 'r' I can think of can easily be made longer.
So, I think you know understand 'rr' is really undoable for me (and I do
swear I did my best to pronounce it !!!).

> But it doesn't
> mean you have to change anything, this r>l change is rather neat and adds a bit
> of irregularity which is nice and naturalistic.
That's exactly what I'm looking for, naturalisticity.:) And keep in mind
this is standard sabyukà, but there will be dialects where /rr/ is a
legal geminate ;).


> You normally put the stress mark in phonetic marking before the whole syllable,
> not the vowel only, because stress is a prosodic phenomenon, not a phenomenon
> linked only to vowels. So you should write ['sabiru]. Also, you should have
> written "['sabiru] for /'saby@ru/" since the second one, besides being what
> some people may say, is also the actual phonemic representation of the word,
> even for people who pronounce it ['sabiru]. Anyway, I already have my answer
> about the 'à' ;))) .
It depends, I saw both notations, as the vowel is the syllable nucleus.
But you're right, yours is better ;). However, what I meant was
['sabiru] *instead of* ['saby@ru] (and not *for*), as I was refering to
the pronounciation of the word. Thank you to have noticed it :).

>
> > As you can see, words are always accentuated on the root basic vowel :
> > such words are called "first derivation words". You can add for
> > instance
> > a suffix to the root ('sabyàjà' > "the-one+who-knows" >>
> > "grammairian")

> You should put your explanatory glosses (e.g. "the-one+who-knows") in the order
> of the morphemes in the word itself. For instance here the root |sabi| is
> followed by a suffix |àjà|: "the one who" (at least if my analysis is
> correct :)) ). So your gloss should have been "know+the-one-who": "grammarian".
> It's just to make things clearer so that we know where to look for the morpheme
> limits.
Yes, that's a true mistake. Excuse me. The problem is that the glosses
are not good : I should have either written "the one who knows" or
"know+human-reference-marker", but not a mixed system. I'll use the
second one. Thanks again.

>
> > There are also "second derivation words", which can use, for instance,
> > a
> > particular case of the root (I borrow this feature to Euskara).
>
> It's a neat feature isn't it? :))) I have it in Moten, which is slightly
> inspired by Euskara indeed :) .
I do love this feature :))) It came up as I was wondering what 'sabyukà'
could mean in the language (I had the word but not its meaning yet).

>
>  The
> > accent moves to the final vowel of the root to make a new root.
> > Example
> > : /s'abyu/ > /saby'u+kà/ = "that+which-leads-to-the-knowledge" -->
> > "language".
>
> Well, wouldn't |kà| rather mean "that-which" rather than "that" alone? Because
> |sabyu| to me only means "to-the-knowledge". Your analysis adds things in the
> wrong places in my opinion, unless for some reason you have a zero relative
> pronoun which is meant always with such derivations.
You're right, it comes from my bad gloss-system. And  also my bad
english ! So here is the good gloss :
/s'abyu/ > /saby'u+kà/ = "to the knowledge + reference marker" -->
"language".

The inanimate reference marker (-kà) is a suffix intended to nominalize
the root meaning.
But nominal roots (like 'urà' "water") can give pairs such as :
'urà' = "water"
'urkà' (ur+kà) = "lake", "sea".("water+RM", "that which +(is) water" )

'-jà' is an animate reference marker, with most of times an agentive
meaning. A conqueror is a "smakjà" ['smag.d^Z@] ("want+RM", "the one who
wants").

>
> > 'Ure' > 'urekà' = "that+in-the-water" --> "fish"
> > And so on.
> >
>
> You see, if you gloss |-kà| as "that-which", you can simply gloss both examples
> as:
> |sabyukà| = "to-the-knowledge+that-which" >> "language"
> |urekà|   = "in-the-water+that-which" >> "fish"
>
> It makes in my opinion the analyses clearer and more consistent between words.
> Note also the main rule of interlinear: follow the morpheme order of the
> *source* language.
So I corrected it, hope it's a little bit clearer.


> > There is also a few infixes, such as the metaphoric (sometimes called
> > poetic) infix "-ek", placed before the (first) vowel. The meaning is
> > really hard to define and there are plenty of uses. Compare :
> >
> > 'yol' "to cry" > 'yekol' "to rain"
> > 'teq' "to tell" > 'tekeq' "to sing"
> > 'mat' "to see" > 'mekat' "to desire"
> > "dem" " to do" > 'dekem' "to build"
> >
>
> That's a neat feature. Is it a productive feature (i.e. you can apply it to any
> verb and even to new ones) or just used with some verbs? (well, the number of
> different meanings for the infix lets me think it's the latter, but I may be
> wrong)
This is a really good question. Actually, -ek- will be the only still
productive (but decreasing) infix. I'll have fixed forms with other
infixes, with more irregularities as they will be more anciant.


> Well, my comments had more to do with the form rather than the language itself,
> but as you may know I don't consider myself to be able to comment on linguistic
> features themselves, except for asking more questions and maybe propose some
> features :)) .
Once again, thank you. Your comments were really straightforward and
helpful. A lot of things seem evident for me (of course, I have it all
in mind ) whereas they are not (glosses for example).

Bye,

Julien.

Reply