LINGUIST List 9.1529

Mon Nov 2 1998

Disc: German Spelling Reform

Editor for this issue: Karen Milligan <>


  1. bwald, Re: 9.1520, Disc: German Spelling Reform
  2. Friedrich Neubarth, German Spelling

Message 1: Re: 9.1520, Disc: German Spelling Reform

Date: Sat, 31 Oct 1998 23:29:45 -0800 (PST)
From: bwald <bwaldHUMnet.UCLA.EDU>
Subject: Re: 9.1520, Disc: German Spelling Reform

I have tried to abstain from the "spelling reform" controversy, but I
finally can't resist any longer. Inevitably the discussion has been
generating a lot of heat, and some of the good points (sometimes
inadequately made) are being lost in politically motivated vituperations
(from all sides). In general, if the discussion reflects more generally on
linguists' positions on spelling reform, and on the knowledge base and
argumentation that they have at their disposal for their positions, THEN I
think it is a good thing that (certified) linguists don't play a prominent
role in "spelling reform" POLICY- making (and should do their best to avoid
blame for whatever policy is decided on). [Instead, invest your energy in
soliciting funds from interested parties on relevant RESEARCH, along the
lines vaguely hinted at below.]

The subject is truly complicated, and it is no wonder to me that the
Akkadians were slow to relinquish Sumerian as their written language, long
after it died. In fact, the deader the better.
Non-spoken languages stop changing phonologically and disaligning
pronunciation from spelling. They can be *fixed* once and for all, instead
of constantly needing RE-fixing..

Vaguely similarly, former European scientists (aka philosophers) persisting
in using Latin until a few centuries ago. And a host of other examples,
e.g., written French vs. spoken French, and last but not least (official)
Chinese, readable in mutually unintelligible pronunciations (or none at
all). By the way, we can now get another perspective on why the ancient
Indian grammarians were so frantic to preserve and nail down Sanskrit
pronunciation. It wasn't only because the gods only recognise one
pronunciation; it was also because it helped the literate remember how to
spell words.

(Uhm, I have to acknowledge from previous discussions, that the "death" of
spoken Latin did not keep it from evolving variant pronunciations,
depending on the first language of the European scholars who wrote -- and
spoke/read out loud -- in it. But, then, all other readers of Classical
Greek learn to pronounce it somewhat more authentically than current Greek
speakers do, and probably have less trouble remembering how to spell many

One theme has been to reform spelling to more closely and more consistently
match spoken pronunciation via biunique phonemicisation. The critics ask:
phonemicisation of which dialect? Implicit among supporters of biunique
spelling, but worthy of further discussion, seems to be that spelling
reform applies to the "standard" language. Well, then, don't we move from
one form of elitism to another, i.e., from jumping through the hoops of
(partially) unmotivated spellings in order to demonstrate a generalised
educational "competence" (= degree of obedience) -- to elitism in
knowledge of the "standard" pronunciation? And then, in American English,
for example, there are crucial gaps in "standardisation" of pronunciation
where it counts for spelling. Should we spell "cot" and "caught" the same?
The standard does not insist on either pronouncing them the same or
different. How are speakers who maintain a difference supposed to guess
that they are pronounced the same in some dialects (or in the reformed
spelling)? What about "merry", "Mary", marry", "send", "sinned", etc.,
etc. And so on for other dialect differences that affect phoneme
inventory, and so on for other languages where various mergers and
distinctions differentiate dialects.

(BTW, you really want to spell "find" and "fined" the same? OK, just
asking. You see, there is actually a difference in how likely both these
words are to be pronounced like "fine" by ANY English speaker, and the
difference is quite large for some English dialects. Same with "mist" and
"missed" as "miss", etc. But then the notion of "likely" pronunciation
hardly enters into spelling reform.)

How controlled is the standard (or call it the "written"? language) for
pronunciation? How controlled SHOULD it be (when it comes to mergers and
distinctions)? How much can you accomplish if you alienate the written
language from the spoken language of the learner? Doesn't such alienation
go against the grain of what aligning spelling and ("current")
pronunciation intends to accomplish in the first place? (Or does it just
pretend to accomplish such an alignment, ignoring any, maybe even the
majority, of the spoken dialects upon whose speakers it will be foisted as
the common written variety?)

One deduced solution might be: let 'em spell it the way they pronounce it.
OK. But then assure me that the various dialects detectable according to
dialectal spelling variation will not prejudice the gate-keepers. How can
you give such assurance? At least if we all have to learn to read and
write in Sumerian then we're all (presumably) at the same initial
disadvantage (overcome by the privileged procuring extra tutoring in
Sumerian -- no, there's no way out of this advantage of the privileged
short of equitable distribution of wealth and faith ojn the educational
process -- which is hardly a linguistic matter, and it is hard to see what
linguistics can do to motivate such a redistribution. "Subversive"
spelling reforms, though well-intentioned in this direction seem
pathetically inadequate to me for the above stated reasons, e.g., they
cannot destroy the influence of privileged dialects, but, in fact, only
enhance it. You can't discriminate against me according to my
pronunciation by the way I write; I'm not dismissed until I have the
face-to-face interview -- or leave a message on voice-mail.)

Similarly, for those who foam at the mouth when hearing about "holistic"
written word recognition (as in Chinese), can you assure me that an
advanced reader will not trip (even in silent reading) over variant
spellings reflecting dialect variation, and prefer a consistent spelling
for a particular word. (It's hard for us to judge since current
conventions INSIST on a single spelling per word, with a few exceptions,
e.g., some Brit vs. US conventions, e.g., "tire" vs. "tyre" -- the
consistency of spelling norm may be a prejudice that confounds experiments
which try to test fluent comprehension with variant spellings the way the
ear often succeeds in understanding variant pronunciations from different
dialect speakers -- i.e., by IGNORING the differences.)

Opposed to the biunique phonemicisation theme is the theme that spelling
should reflect "underlying" phonology. This at once raises the theoretical
linguistic question popular since the advent of generative phonology of
whether "underlying" phonology or biunique phonemes are more reflective of
(or at least more accessible to) speakers (or something like that). A
sometimes proposed virtue of "deep" spelling is that to a large extent it
apes the (historical) unity underlying much current dialect diversity and
thus avoids the "superficial" pronunciation = spelling problem. Similarly,
it links conditioned changes in the pronunciation of base morphemes, e.g.,
sphere-spherical, etc etc etc. My favorite twisted example in English is
that I can remember to spell "repEtition" instead of "*repAtition" or
"*repItition" by remembering that "repEat" features the "same" (graphic =
morphophonemic) /e/. (Without intending immodesty, I am proud that my long
term irrational memory is equal to spelling "repeat" instead of "repete" or
"repeet". "repet" is out of the question. That could only spell "pet

Well, as I mentioned above, arguments for such an approach to spelling,
unsurprisingly more conservative to current spelling (for ANY language with
a sufficiently long written history), depends crucially on how valid access
to "deep" phonology is to would-be readers. Chomsky & Halle's famed Sound
Patterns of English is hardly a plausible basis for an introduction to
spelling, until the post-graduate level of reading has already been
attained anyway, at which point it is only of theoretical, not practical.
interest. The issue remains moot (though often less fantastic) for more
recent "deep" analyses of English and other (morpho)phonologies.

Clearly "deep" vs. "surface" spelling (reform) awaits resolution of the
issues of phonological "reality" (whatever that turns out to be, and
probably a little of both themes, constantly in need of up-dated spelling
reforms as the language continues to change and dialects continue to

To conclude with something more empirical, it did seem to me from some
research I once did that some languages are "luckier" than others with
respect to the match between spelling and dialect diversity. Spanish, for
example, generally maintains a uniform surface phonology across most of its
dialects, varying phonetically much more than phonemically. The glaring
exceptions are various consonantal mergers that affect most varieties, but
less the Castiliian varieties reflected in the spelling, e.g., the
distinction between /s/ on one hand and /z/ (also /c/ before front vowels)
on the other, totally opaque to Latin Americans, and the /ll/ and /y/
distinction opaque to almost everybody. I found that Mexican Spanish
speakers who dropped out of school at the third grade did not have accurate
spelling when there were merges in their dialects, but they were able to
decipher words from conventional orthography, and even understand written
instructions in Spanish with some effort. However, even here it might be
that it is easier for speakers with mergers to ignore distinctions in
deciphering writing than for speakers with distinctions to extract meaning
from an under-differentiated writing system. (Note that in Hebrew and
Arabic orthography, vowel diacritics are a crutch to
decoding/word-recognition in early reading instruction, but withdrawn from
orthography, as if "superfluous", as readers become more experienced and
fluent. I am not aware how much trouble absence of tone marking gives
initial readers whose languages are tonal but whose orthographies are not,
e.g., many African languages.)

In contrast to what I observed for Spanish reading, speakers of languages
with orthographies as deviant from biunique phonemic spelling as English,
were often still having trouble in fifth and sixth grade (same for
drop-outs from these and even later grades) decoding English words, not
able to any more than make wild guesses quite oblique to orthography after
stuttering at length over the first spelled syllable of a word, and, before
becoming violently ill, they often gave up on the whole word with wild
guesses which were not English words known to ANYONE. Anyone who has
witnessed acquisition of reading skills in English recognises the painful
failed decoding process alluded to here, e.g., puh-puh-puhsi-say-saykramani
(a wild guess at the spelling "psychology" or whatever). The most striking
part of this process is that the frustrated reader often does not even try
to FAKE a real word. This is extremely common among poor English readers.
(I mean they read poorly, not that they or their parents are necessarily

PS: In first grade I read ahead in my reader and wondered why Dick and
Jane gave the horse a cigar, spelled "sugar" (not "shoogger"?). When we
actually came to that passage in class, I was disappointed before the fact
but grateful after the fact that I was not the one chosen to read that
passage out loud. (I haven't misspelled either of those words since --

Unfortunately, English, and various other Germanic languages, are cursed
with complicated and (thus) dialectally unstable VOWEL systems, where it
is evident that vowels are much more crucial to "phonetic" decipherment
than consonants, and Germanic dialect variation in vowel mergers and
distinctions (within languages), conditioned or not, are formidable
impediments to reading (and spelling) acquisition.

I end by submitting that we have a lot more to learn about the decipherment
process in reading acquisition at various stages, how it relates to reading
comprehension at various stages, and how it relates to specific KINDS of
languages before we go this route in changing the world to advance social

(On the other hand, developing orthography for unwritten languages is fair
game for us, even now. Our only problem is to decide what constellation of
speech varieties to include and exclude from some particular written
variety we concoct) -- Benji
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: German Spelling

Date: Sun, 1 Nov 1998 19:14:20 +0100
From: Friedrich Neubarth <>
Subject: German Spelling

There is some discussion going on about the letter "" in the German
orthography, since the rules according to which it has to be used or not are
quite complex and not really insightful. Not even historical.

Although it could be argueable to keep it in the German writing system
because it is one of its specialities and reminds people in international
(or worse - national) standardization committees of the German language, I
would rather take this argument for a perverted one. The very reason why it
is more than ridiculous to stick to that character as a part of the German
writing system is that it was born by a hardly-to-believe misunderstanding:
Since German (or German writers) had chosen reduplication of consonants as
one of the means to indicate that the preceding vovel is short instead of
long, this device was also used with an "s". That is the end of the story.

Meanwhile, in handwriting the s at the end of a word was fine to write, but
an s not at the end was not. Therefore a similar grapheme was used but it
was stretched in length to comfort hands and eyes (the long vs. round s in
"Current-" and "Fracturschrift". At the end of a word two ss would be
written with a long and a round s. And because that combination occured
quite frequently, a kind of ligature was conventionalized, meaning nothing
more than two "s"s (at the end of a word). But since the second part of this
ligature at some point resembled the (old) z, people started to think that
this ligature was some special character and even today you can very often
see double "s"s writen as "sz" which indicates that noone has been aware of
that coincidence for more than 150 years.
In latin writing that peculiar sign is more than obsolete (since it is not
violating any rules to write two "s"s in a row, why should it be). For that
reason I would not even think of prolonging that idiosyncratic sign with no
meaning and dubious graphic qualities and go on discussing whether and/or
with what devices short and long vovels should be marked in orthography.

And now lets go on to the real stuff...

Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue