Print

Print


Hi!

After reading a lot about Portuguese recently, I started to think
about Terkunan again.  One of my favorite sound shifts, most prominent
in Romance in Portuguese, Catalan, and Southern Italian dialects like
Sicilian and Calabrese, had not been implemented yet in Terkunan,
because I faced some problems.  I am talking about the reduction of
unstressed vowels to /a i u/, meaning e>i and o>u in unstressed
syllables.

So, the problems in Terkunan were the following: first of all, I do
not want Terkunan stems to ever change spelling, and they should be
written phonetically, but because the stress shifts when endings are
attached, the /a i u/ rule would change the pronunciation of same stem
vowels:

    solu    + (a)t  > solut     'solved'  (participle)
    /'solu/ + /t/   > /su'lut/

OK, so with this rule, the conflict is unavoidable, I cannot have
both: same spelling and same pronunciation. I have to live with it,
otherwise I cannot have the /a i u/ reduction.  The compromise is to
write _e_ or _o_ whenever there is a form of the stem where they are
actually pronounced as /e/ or /o/.  All vowels that are *always*
reduced will be written as _i_ or _u_.  Since stress is determined
from the end of the word, only endings can shift the stress and the
stress can only move towards the end, so all vowels *before* the
stress of the plain stem will be written with only _a i u_.

The stress rule is simple: if the ultimate is closed, it is stressed,
otherwise the penultimate is.  Some examples:

   OLD       NEW
   'solu     'solu    'to solve'
   bro'kan   bru'kan  'volcano'

So, my decisions stands: no spelling changes for stems; the compromise
means writing of Terkunan requires knowledge of the stems and is not
purely phonetic anymore.  I must live with that.

Second problem: usually, all original inflection endings drop, usually
including the thematic vowel (e.g. _not_ < NOCTEM, _prot_ < PORTA,
_rot_ < HORTUM).  The current language requires an epenthetic _e_ from
an earlier /@/ on stems ending in certain voiced consonants and most
clusters, which cannot end a word.  Some examples:

   mange  < *mang@   'to eat
   deve   < *dev@    'to have to'

These vowels drop before another vowel:

   mange + (a)t  >  mangat 'eaten' (participle)
   ave   + (a)t  >  avat   'had'   (participle)

OK, yes, 'avat' sounds strange to Romance afficionados, I know.  Could
be fixed by retaining a final -i.  Different problem, please ignore
that for now.  :-)   Some verbs do end in -i because of a retained glide
or because of -i conjugation:

   odi + (a)t  >  odit   'heard' (participle)

This is where the problem arises: -e and -i collapse to /i/ with the
new /a i u/ rule.  And I do not like stems being not written as they
are pronounced.  So my urge would be to change spelling from -e to -i
(this particular -e will *never* be stressed, because it will rather
drop for an alternative vowel), but then, I either get participle
forms that even I find too strange due to the collapse with the original
thematic -i:

   *mangit 'eaten' ??

Or I get forms that cannot be derived from the stem, which I also to
not want (some of you would love it, I know, but it is against my
design goals here), e.g.

      mangi +(a)t  > mangat
but   odi   +(a)t  > odit

Nah!  So this is the main reason why I did not implement the /a i u/
shift before.  I had no solution.


Today, I talked to a friend and he suggested a very simple and elegant
solution I had not thought of before: make the epenthetic vowel an
/a/.  I think I remember that at the beginning of Terkunan planning,
this thought once occurred to me, but I then quickly discarded it,
because -a- sounds so thematic to Romance ears, and as an epenthetic,
it sounded weird.  But in the meantime, Terkunan has become quite
untypical/weird anyway in some areas so that it feels like the perfect
solution now. :-)

I think this solves all my problems and this will be the next major
change in Terkunan.  You get:

  manga > mangat  'eaten'
  ava   > avat    'had'      (maybe avi/avit later: different problem)
  solu  > solut   'solved'
  odi   > odit    'heard'
  am    > amat    'loved'
  dik   > dikat   'said'
  ntres > ntresat 'interested'

This -a (among many other vowels) drops if a vowel follows:

  Mi mang' api.   'I eat celery.'
  M' av' oli.     'I have oil.'

Some interesting changes:

  OLD         NEW
  boral       bural      'northern'
  brebe       breba      'verb'
  bride       brida      'green'
  brokan      brukan     'volcano'
  felik       filik      'happy'
  fimbre      fimbra     'woman'
  genes       gines      'genesis'
  etren       itren      'eternal'
  Karle       Karla      'Karl'
  kokodril    kukudril   'crocodile'
  kervis      kirvis     'beer'
  Lak Leman   Lak Liman  'Lake Geneva'
  mbrekat     mbrikat    'market'
  ombre       ombra      'man'
  oskur       uskur      'dark; obscure'
  patre       patra      'father'
  present     prisent    'present'
  rotim       rutim      'ultimate'
  sambe       samba      'blood'  (how poetic: the blood dance! :-))

BTW, some 'small words' and prefixes that never get primary stress
will change, too:

  pe          pi         'for'  (just like in Southern Italian)
  o(b)-       u(b)       the prefix 'ob-'

Thoughts?  Suggestions?


Mbrut ublegat pi tus atensiun!    [m='bru tUbli'gat pi tu satIn'sjUn]

  **Henrik

PS: No decision yet about the language name 'Terkunan'.  Maybe it will
    be kept, maybe it will change to 'Tirkunan'.  I said earlier that
    'Terkunan' rings the 'ter'-bell to native speakers (_ter_ =
    'earth, land'), so there might be possibility to construct a
    reason for keeping the language name. :-)
To: [log in to unmask]
Subject: Terkunan: revision: /a i u/ in unstressed syllables
FCC: ~/Mail/Sent
--text follows this line--
Hi!

After reading a lot about Portuguese recently, I started to think
about Terkunan again.  One of my favorite sound shifts, most prominent
in Romance in Portuguese, Catalan, and Southern Italian dialects like
Sicilian and Calabrese, had not been implemented yet in Terkunan,
because I faced some problems.  I am talking about the reduction of
unstressed vowels to /a i u/, meaning e>i and o>u in unstressed
syllables.

So, the problems in Terkunan were the following: first of all, I do
not want Terkunan stems to ever change spelling, and they should be
written phonetically, but because the stress shifts when endings are
attached, the /a i u/ rule would change the pronunciation of same stem
vowels:

    solu    + (a)t  > solut     'solved'  (participle)
    ['solu] + [t]   > [su'lut]

OK, so with this rule, the conflict is unavoidable, I cannot have
both: same spelling and same pronunciation. I have to live with it,
otherwise I cannot have the /a i u/ reduction.  The compromise is to
write _e_ or _o_ whenever there is a form of the stem where they are
actually pronounced as /e/ or /o/.  All vowels that are *always*
reduced will be written as _i_ or _u_.  Since stress is determined
from the end of the word, only endings can shift the stress and the
stress can only move towards the end, so all vowels *before* the
stress of the plain stem will be written with only _a i u_.

The stress rule is simple: if the ultimate is closed, it is stressed,
otherwise the penultimate is.  Some examples:

   OLD       NEW
   'solu     'solu    'to solve'
   bro'kan   bru'kan  'volcano'

So, my decisions stands: no spelling changes for stems; the compromise
means writing of Terkunan requires knowledge of the stems and is not
purely phonetic anymore.  I must live with that.

Second problem: usually, all original inflection endings drop, usually
including the thematic vowel (e.g. 'not' < NOCTEM).  The current
language requires an epenthetic _e_ from an earlier /@/ on words with
some voiced consonants and most clusters, which cannot end a word.
Some examples:

   mange  < *mang@   'to eat
   deve   < *dev@    'to have to'

These vowels drop when a vocalic ending is added:

   mange + (a)t  >  mangat 'eaten (participle)
   ave   + (a)t  >  avat   'had   (participle)

OK, yes, 'avat' sounds strange to Romance afficionados, I know.  Could
be fixed by retaining a final -i.  Ignore that for now, please.  Some
verbs do end in -i because of a retained glide or because of -i
conjugation:

   odi + (a)t  >  odit   'heard' (participle)

This is where the problem arises: -e and -i collapse to /i/ with the
new /a i u/ rule.  And I do not like stems being not written as they
are pronounced.  So my urge would be to change spelling from -e to -i
(this particular -e will *never* be stressed, because it will rather
drop for an alternative vowel), but then, I either get participle
forms that even I find too strange due to the collapse with the original
thematic -i:

   *mangit 'eaten' ??

Or, I get forms that cannot be derived from the stem, which I also to
not want,e.g.

      mangi +(a)t  > mangat
but   odi   +(a)t  > odit

Nah!  So this is the main reason why I did not implement the /a i u/
shift before.  I had no solution.


Today, I talked to a friend and he suggested a very simple and elegant
solution I had not thought of before: simply make the epenthetic vowel
an /a/.  I think I remember that at the beginning of Terkunan
planning, this thought once occurred to me, but I then quickly
discarded it, because -a- sounds so thematic to Romance ears, and as
an epenthetic, it sounded weird.  But in the meantime, Terkunan has
become quite untypical/weird anyway in some areas so that it feels
like the perfect solution now. :-)

I think this solves all my problems and this will be the next major
change in Terkunan.  You get:

  manga > mangat  'eaten'
  ava   > avat    'had'
  solu  > solut   'solved'
  odi   > odit    'heard'
  am    > amat    'loved'
  dik   > dikat   'said'
  ntres > ntresat 'interested'

This -a (among many other vowels) drops if a vowel follows:

  Mi mang' api.   'I eat celery.'
  M' av' oli.     'I have oil.'

Some interesting changes:

  OLD         NEW
  boral       bural      'northern'
  brebe       breba      'verb'
  bride       brida      'green'
  brokan      brukan     'volcano'
  felik       filik      'happy'
  fimbre      fimbra     'woman'
  genes       gines      'genesis'
  etren       itren      'eternal'
  Karle       Karla      'Karl'
  kokodril    kukudril   'crocodile'
  kervis      kirvis     'beer'
  Lak Leman   Lak Liman  'Lake Geneva'
  mbrekat     mbrikat    'market'
  oskur       uskur      'dark; obscure'
  patre       patra      'father'
  present     prisent    'present'
  rotim       rutim      'ultimate'
  sambe       samba      'blood'  (how poetic: the blood dance! :-))

BTW, some 'small words' and prefixes that never get primary stress
will change, too:

  pe          pi         'for'  (just like in Southern Italian)
  o(b)-       u(b)       the prefix 'ob-'

Thoughts?  Suggestions?


Mbrut ublegat pi tus atensiun!    [m='bru tUbli'gat pi tu satIn'sjUn]

  **Henrik

PS: No decision yet about the language name 'Terkunan'.  Maybe it will
    be kept, maybe it will change to 'Tirkunan'.  I said earlier that
    'Terkunan' rings the 'ter'-bell to native speakers (_ter_ =
    'earth, land'), so there might be possibility to construct a
    reason for keeping the language name. :-)