LINGUIST List 3.935

Sat 28 Nov 1992

Sum: Subphonemic Writing

Editor for this issue: <>


Directory

  1. Joseph C. Salmons, Sum: Subphonemic writing

Message 1: Sum: Subphonemic writing

Date: Wed, 25 Nov 92 16:02:24 ESSum: Subphonemic writing
From: Joseph C. Salmons <salmonsmace.cc.purdue.edu>
Subject: Sum: Subphonemic writing

The considerable response to the query about subphonemic writing warrants
a summary to the list. Examples and discussion came from Gene Buckley,
Tucker Childs, Leo Connolly, John Cowan, Will Dowling, Alice Faber, Thomas
Field, Brett Kessler, John Koontz, Mark Mandel, Geoff Nathan, Mary Niepokuj,
David Stampe, Allan Wechsler, Neal Whitman. Thanks to all. This summary
is very abbreviated but I hope not to have done violence to anybody's views.

While the cases sent do involve writing of allophones, most are problematic
on other fronts. For example, the best known case is that of the Sanskrit
palatal nasal, which was written although it occurred only with palatals.
Possible interfering factors include the role of phoneticians (Stampe), desire
to preserve an archaic pronunciation for religious purposes (Nathan) and the
phonemicization of the distinction in spoken varieties (Stampe). Parallel
problems can be found elsewhere, e.g. Tiberian Hebrew marking of stop vs.
fricative allophones by the dagesh, which may reflect efforts to preserve
'authentic' pronunciation for liturgical purposes and may have been introduced
by non-natives (Faber).

The most common complication, including above, is bilingualism (isn't it
always?), e.g. where non-natives imported distinctions phonemic in their own
languages into the writing of languages where these distinctions were allo-
phonic. Such cases include Kisi (<v>~<w>, Childs), Fox (obstruent voicing,
Stampe) and the 'romaji' system for Japanese (Cowan), etc. Similar but more
complicated may be the case of Old High German, where scribes were native
speakers of OHG, but trained/practiced in writing Latin. In some dialects,
voicing was not distinctive but some lenis stops were marked for voice based
on the scribes' Latin habits (Connolly). [My query arose from a set of claims
by J. Voyles (in his '76 Phonology of OHG, '91 PBB article, etc.) that OHG
scribes variably recorded a whole array of allophonic distinctions.] A
system developed for Omaha by LaFlesche (a native speaker) includes some
allophonic distinctions which may be connected to this same problem, with
the additional twist that he was apparently influenced by earlier ortho-
graphies developed by non-natives (Koontz).

Two cases are considerably different. The first is Tigrinya writing (adapt-
ed from the Geez syllabary), where [k]~[x] and [k']~[x'] are marked dia-
critically although the fricatives derive from stops by regular postvocalic
spirantization (Buckley). 'The different symbols do...serve a function in a
particular context: since gemination is not marked, and geminate stops fail
to spirantize, any [k] written after a vowel must be geminate.' This marking
is an innovation vis a vis Geez. Based on what I've seen, this is the strong-
est case for allophonic writing, although foreign influence apparently can't
be entirely ruled out here.

The second draws on Patricia Donegan's work (In press: 'On the phonetic basis
of phonological change' Historical Linguistics, ed. Ch. Jones. Longmans) and
moves beyond orthography to the definition of phoneme. Here's part of David
Stampe's summary: 'Naive speakers from Baltimore say that _write_ and _ride_
"have different vowels, [6y] and [ay]". ...She suggests that in general an
obligatory strengthening is phonemic, regardless of contrast.' Donegan
follows Baudouin de Courtenay and Sapir in accepting 'that a phoneme is a
speech sound that is perceptible and pronounceable on its own terms, rather
than in terms of some other sound. ...Swadesh's 1934 article on how to use
complementary distribution as an analytic technique begins by defining the
phoneme perceptually, then gives the distributional method, but doesn't ask
why the two should coincide. Donegan thinks they don't. If so, Stampe would
expect that certain strengthened alternants would be among the "allophones"
that have found a place in writing systems.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue