Re: A few phonetics-related q's

On 13 Sep 2004 Trebor Jung <treborjung@FRE..> wrote:

> In a word like /anta/, would it be more likely that it's pronounced [anda]
> or [an_0ta]?
  IMHO you should make first a more principal decision: Do you
prefer regressive or progressive assimilation? The actual
likelihood is based on this underlying liguistic preference.

  However, in Hungarian -- which is uses basically progressive
assimilation -- nasals are not subject of voice assimilation at
all, therefore /anta/ = [anta] is also possible. (And even [a~ta]
is likely, see later.)

  Btw you said that yout conlang is derived from English. In this
case, there must be a sudden change to adopt progressive
assimilation. (I guess eg. strong Slavic-Hungarian substratum
effect.)


> If a language has a rule (a) /s/ is [S] before /i/ and (b) /s/ is [z]
> intervocalically, would it be more likely that a word like /asi/ be
> pronounced [azi] or [aZi] or even [aSi]?
  English gives examples of [aS] in |Asia| and of [aZ] in |casual|.
In Hugarian |S| > |Z| voicing is tipical.  My preference would be
[aZi], less [aSi] and least [azi].

  However, probably, you might assign a time-scale to sound
changes, e.g. whether intervocalic voicing is happened earlier or
later than pre-iotic palatalization. If it was earlier (as AFAIK
e.g. in Romance), you should consider the possiblity of
intervocalic [S] > [Z] voicing. It it was later, the likelihood of
a pre-iotic [z] > [Z] change is in question.


> French nasal vowels can differ from their oral counterparts, cf. [i] ~
> [e~]. Is there an articulatory/acoustic precedence for this? What are some
> oral-nasal correspondances for /i/, /e/, /A/, etc.?
  There are also subphonemic nasal vowels in Hungary: orals are
usually pronounced as nasals before a nasal consonant, especially,
before a consonant cluster starting with a nasal.  Moreover, in
sequence VNS (where V=vowel, N=nasal consonant, S=sibilant) nasal
consonant is regularly assimilated into the vowel as a nasal
acoustic feature, e.g. VNS > V~S. During these nasalizations, the
original quality of the vowel is retained (they are allophones, not
phonemes, though).

  I think there was rather a phoneme merging in French. I am not
well-informed in French phonetics, but my textbooks mention a
present /2~/ > /e~/ change and that would be a later step in the
merging process.

  Similarly Old Slav /e~/ and /o~/ merged into one nasal phoneme
(let's say /@~/) in Old Polish, and that was split according to the
vowel length: short nasal vowel became front /E~/ and the long
variant became /O~/.

  However it seems that there is an articulatory precedence that
causes merging the nasal vowels.  Nasal vowels have less dispersion
of contrast than oral ones, therefore masals are merged more often.
Cf. a paper on Polish nasals
<http://wso.williams.edu/~nsanders/Papers/sanders-wccfl21.pdf>.

  Btw, if you use nasal vowels, /anta/ in the first question would
be realized also as [a~ta].


> How could vowel harmony (any type) develop in a language?
  In Uralic branch, the back ~ front alternation seems to be an old
derivational tool, and it is still vivid in Hungarian; e.g. root
pair |gur-| ~ |gör|: |gurít| 'to roll, trundle', |gördít| 'to
wheel, roll, push'; root pair |ker- ~ kör| ~ |kar-|: |kerék|
'wheel' ~ |karika| 'ring, circle' etc. (Not counting the obvious
near-distant contrast in demonstrative pairs like, |ez| 'this' ~
|az| 'that', |itt| 'here' ~ |ott| 'there' etc.)

  IMHO the opposition of back ~ front variants was emphasized that
suffixes (bound morphemes) were joined in their harmonic form.

----

For some of your previous questions:

> How could I represent /K/ and /tK)/ as non-digraphs?
  If you do not want to use modified Latin letters, I would propose
a compromise between previous proposals: |x| for /K/ (|x| is used
both for [S] in some languages, and [S] is a [-lateral] variant of
postalveolar [K]; moreover |x| is obvious for [x] which is also a
fricative and evokes the "hissing" characteristics of /K/), and |j|
for /tK/ (|j| conveys an affricate in many languages, the voice
variant of [tS]; and the latter is a non-lateral counterpart for
/tK/).


> Any other [Spanish] dialects have <y> as /dZ/?
  I heard Julio Iglesias singing on the TV yesterday. For my
Hungarized
ears, he pronounced |yo| after a pause with [J\j\], i.e. a voiced
palatal
affricate. (N.B. This sound is very familiar to me, since Hungarian
|gy| is
phonetically rather an affricate [J\j\] than a stop [J\].)