above
in the later category. Words like [ļi):], [ko)e)], [si)a)] have nasalization
which cannot be assigned to any particular nasal. So the best way
is to abandon the attempt of assigning them to any paritcular phoneme,
since even if the nasalization is used in such cases expressly, it
is merely accidental and non-phonemic.
On the other hand, in situations given in No. [iii] above nasalization
occurs in complementary distribution with [N].
Therefore, the automatic conclusion in such a case will be to assign
nasalization to both of the phonemes /n/and /N/ and it will have special
status of being di-phonemic in character. This avoids even the necessity
of setting up a separate phoneme of `nasalization' which apart from
reducing the number of phonemes can also make nasalization predictable.
However, in phonemic writing the nasalization will be shown as suprasegmental. |
1.6.
Identifying the intervocalic [ ]: |
Let us
consider following set of words:
[se#:ć] `intelligent'
[dć:ć] `to search'
[pć:ć] `public
sacrificer' [d : ć]
`to overturn a vassel accientally'
[bć:]
`a rounded thread' [ :]
`to extinguish fire'
[cõ:õ] `the
act of omitting [hć:ćb]
`to put into fail'
[mõ:]
`five' [mć:]
`slowly'
[c:ć?] `corruption'
[r:]
`debt' |
Mudari
has hundreds of such words where an intervocalic retroflex []
occurs, making both preceding as well as succeeding vowels nasalized
and the preceding vowel at the same time extra long. Mundas differ
in phonetic assessment of this sound and even there are dialectical
variations. Some suspect that this should be identified as [ ]
sound.
Unfortunately phonetics does not help in correct identification in
this case. An intervocalic [ ]
does become [],
but strong nasalization invariably in all such cases cannot be accounted
for.
The occurrence like [ba:ø·a ] excludes the possibility of a [n·]
changing as []
intervocalic ally.
Distribution ally, []
and []
stand in contrast in intervocalic position, while it also contrasts
with [n] and [m]. Hence, the only choice is to consider []
as a separate phoneme with very restricted distribution. It occurs
between two nasalized vowels and before [].
In situations before [],
it also contrasts with /n/, as it will be evident from examples like
[man·i] and [ba:a].
The incidental occurrence of nasalizatin with the neighboring vowels
can also be explained by accepting this segment as []. |
| 1.7.
Non-phonemic status of dependent segments: |
| The segments
indicated in the list § 1.1 as [y] and [w] with an adjective `glide'
are not even non-syllabic sounds. They are simply transitional and
their pronunciation and existence depend upon the neighboring vowels.
Therefore, they are depended segments and as such they cannot have
a phonemic status. They combine with the adjacent vowels to make phonetically
a trip thong, in which situation they vary with zero. |
The environments
for the appearance of y-glide ar [i-a], [e-a] and [o-e]. Sometimes
it is also heard indistinctly in the contexts [a-i], [a-e], [e-o],
[i-u].
The environments for w-glide are [o-a], [u-a] and [a-o] and indistinctly
[a-u]. There is no example of contrast of y- glide with the non-syllabic
[ ]
and of w-glide with non-syllabic [ ]but
at the same time they are also non-allophonic by virtue of being simply
transitional and transitory. They function as off-glides for the preceding
vowel.
A segment which varies with zero and is dependent upon other segments
is redundant. |
|
