in some
dialects. The intrusive consonants may or may not be written in the
source files, but the difference is never distinctive.
In most languages the native syllable final stops are checked and
voiceless, as in Mon-Khmer, but they are pronounced as voiced stops
before vowels, and may be identified lexically as voiced. (In fact,
some Juang dialects no longer check final stops, and pronounce them as
voiced.) Citations of stops in syllable-final position may be
morphophonemic <> or phonemic , the latter with or without the glottalization diacritic.
Historically speaking there was only one series in final position,
exactly as in Mon-Khmer, so the transcriptions would not matter,
except that in many languages Indo-Aryan words have been borrowed
with final unchecked voiced and voiceless (and in Kharia and the
Kherwarian group of North Munda, borrowing from Hindi, and in Korku
borrowing from Marathi) even final aspirated voiced and voiceless
stops, all contrasting with the native final checked stops. In all
these cases we have followed the source transcriptions, except for
words from Kh(B), Biligiri's Kharia Grammar, which used a coexistent
phoneme-systems type of transcription that had arguable merits but
was difficult to reconcile with other sources for Kharia, and was
therefore normalized to conform to their type of transcription.
In Gorum checked stops have in some cases become checked nasals and
are written as such. (Some checked nasals may however be from stop
plus nasal.) In several languages the point of articulation of a
checked stop is lost before a homorganic stop or nasal, and that may
or may not be shown explicitly. In all these cases we have followed
the source transcriptions.
In Kharia (and often in North Munda, and Hindi) there is flapping of
to , to , and probably Hindi loans with
as ; they may written thus. In Juang, is flapped
to instead, while syllable-final becomes .
Juang in some dialects has changed some 's to or zero, but this
is lexically diffused and therefore represented explicitly; due to
entering , , and zero variants in distinct alphabetical orders,
the same entry may have been found three times in search files. This
is true of alphabetically distinct variants in any of the languages,
unless the search results have been hand-edited to eliminate the
repetitious entries.
Most Munda languages vowels have a palatal (fronted, raised) glide
before tautosyllabic palatals, i.e. . In some languages
the vowel is then transcribed or . The palatal may itself be
lost, leaving a glottal in the case of the checked stops, nasalization
in the case of the nasal stop, and nothing in the case of the glide.
In some of the languages the resulting forms have been reinterpreted
as phonemic.
Different Sora sources may disagree on vs , vs , and
vs . This is gradually being corrected in the Sora materials, but
in the meantime we have followed the earlier sources. Also before
the palatals is usually pronounced [+] or [e], and may be
found transcribed as <+> or .
----------------------------------------------------------------
Please address questions and comments to
Patricia J. Donegan
and David Stampe
University of Hawai`i Department of Linguistics
Honolulu, Hawai`i 96825 USA