LINGUIST List 2.306

Tuesday, 18 June 1991

Disc: Orthography

Editor for this issue: <>

Directory

1. Susan Ervin-Tripp, Re: Responses
2. Robert D Hoberman, Vowel-omitting writing systems

In reply to John Coleman re orthography:
> The comment about not being
> able to read Arabic aloud until you know what it means is equally true
> of alphabetic scripts, yes even Finnish.
Those of us who have had the experience of reading aloud to illiterates
letters in languages we don't speak or understand can attest that
we can then get translations. Obviously won't work when relevant
features (e.g. prosody) are inferred rather than systematically
represented.
Susan Ervin-Tripp

	Margaret Fleck asked (LL 2.275) about the functionality of writing
systems like those of Arabic and Hebrew and those, including Persian and Urdu,
which borrowed the Arabic writing system; all of these omit any indication of a
subset of the vowels. She wrote, "Is it just masochism, or are there
particular features of the phonology of these languages that makes it more
plausible than it sounds?" One answer has been suggested over and over in the
Semiticist literature: that the "particular feature" is Semitic discontinous or
nonconcatenative morphology, in which a word root is normally just a sequence
of consonants, and the vowels which are interdigitated with it are separate
inflectional or derivational morphemes. Here is a typical statement of this
idea, from the authoritative handbook SEMITIC WRITING: FROM PICTOGRAPH TO
ALPHABET, by G.R. Driver, revised ed., 1976, p. 178:
	"The Greeks, when they took over the Semitic alphabet, at the same time
adapted it to the needs of an Indo-European language and so made it to all
intents and purposes universal.
	"In the Semitic languages the fundamental element in the root of a word
is the consonants, while the vowels are accidental [=inflectional, RH]; they
are, of course, essential to its pronunciation but they serve merely to modify
its basic sense: for instance, while the idea of killing was inherent in Q-T-L
as the root, the distinction between QATAL(A) 'he killed' and QUTIL(A) 'he was
killed' was shown only by the changed vocalization.... Consequently, the
Semites could write only the consonants and leave the reader to supply the
vowels as the context and his own sense suggested. In the Greek language the
vowels were of equal value with the consonants and had therefore to be
represented in the written word...."
	The trouble with this explanation is that it can easily be turned on
its head. QATALA and QUTILA can easily fit into the same syntactic and
semantic slot, so it could be argued that they MUST be distinguished in
writing; in an Indo-European language, on the other hand, while there might be
sets like lid, leed, led, lead, load, lad, lewd, laid, laud, it would be rare
to find a pair that could be mistaken in context. This is because a pair of
words sharing the same consonants would, in an IE language, usually belong to
entirely different semantic domains. Though READ (past and present) is one
such, ambiguity like that of READ is, in a Semitic language, built into the
system. Consequently, it should be easier to omit vowels in writing an
Indo-European language than a Semitic one; specifically, the Arabic-origin
writing system should work better for Persian than for Arabic!
	I don't know how to test this hypothesis. One one hand, though in
adopting and adapting the Arabic script Persian and Urdu writers invented
several letters for consonants absent in Arabic, they did not abandon the
principle of omitting certain vowels. On the other hand, Yiddish writing,
using the Hebrew alphabet, has seen fit to represent all the vowels
unambiguously. Is there anyone out there who knows both Arabic and Persian or
Urdu and can offer at least a subjective judgment? Is there data about
children's learning of reading and writing in different linguistic communities?
Bob Hoberman