LINGUIST List 4.948

Fri 12 Nov 1993

Disc: Rules that refer to spelling

Editor for this issue: <>

Directory

1. , Rules that Refer to Spelling (Some Clarifications)

I have received a ton of mail from people regarding my recent
posting on this subject. I will post a summary of the examples
later, since I assume there will be more. So far, I have mostly
received complaints about my analysis of the Polish and French
facts cited in my query.

I would, therefore, like to add the following clarifications:

(a) In response to several questioners, yes, there is clearcut
evidence that nonliterate speakers (and indeed many literate ones)
do not follow the spelling-based rules I alluded to, in both
languages. In the case of French, this is abundantly described
in the literature on French verse.

(b) Bert Peeters points out that I may have given the impression
that I was claiming that the final -e in foie counts a syllable
in classical French verse, which is at best true in a very strange
sense I won't go into. What happened was that I tried not to
give all the details of the rules (which treat consonant and vowel
final words differently and then only gave an example involving
vowel finals). The crucial fact is that foie and foi are NOT
equivalent according to the standard rules of French versification
(which, of course, are not followed in popular verse but which are
taught and enforced in "official" poetry). The most striking
fact is that words like foie are never allowed in this kind of
poetry to appear before a word that begins with a consonant,
while words like foi are. So the phrase foie gras, for example,
cannot be used in standard French verse.

(c) In response to several writers who claim that Polish final /w/
is "underlyingly" a different segment when spelled with slashed l
than when it is spelled with w or u or who claim that foi and foie
are "underlyingly" different in French BECAUSE they are treated
differently by the rules of versification, I would like to say
that this is no way to defend a theory. The theory advanced
by several generative phonologists has been that there exists
a so-called underlying level of representation at which various
strange things exist, such as a French final schwa. There is
nothing a priori good or bad about this theory, and we cannot
a priori decide that it is right or wrong. We CAN make such
decisions on the basis of the data. Some of the data that have
been cited have been that some languages supposedly exhibit
rules of versification sensitive to this level of representation
(or some level close to it). These claims could again be correct
or not. What I point out is that in the case of French anyway
they are NOT, because there is no PHONOLOGICAL basis for giving
foi and foie different underlying representations (or to be
precise to give them SUCH URs AS WOULD be required to account
for the facts of versification. Instead, there is a perfectly
well-known, well-documented, and correct theory which does
account for these facts and which says that the rules were
artificially designed by people who did not understand the way
that we do the relation of speech and writing, that these rules
are explicitly taught and enforced (i.e., they are not learned
subconsciously the way that linguistic rules are), and that
they in particular refer to spelling.

Similar considerations apply I think to the Polish case, although
here the issues are more subtle because the rule is not, at least
not always, artificially taught.