Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.1 6/24/83; site sphinx.UChicago.UUCP Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxn!ihnp4!gargoyle!sphinx!mmar From: mmar@sphinx.UChicago.UUCP (Mitchell Marks) Newsgroups: net.nlang Subject: Re: about diacritical marks (danish dynamite) Message-ID: <1037@sphinx.UChicago.UUCP> Date: Thu, 22-Aug-85 06:29:59 EDT Article-I.D.: sphinx.1037 Posted: Thu Aug 22 06:29:59 1985 Date-Received: Sat, 24-Aug-85 15:55:50 EDT References: <1065@diku.UUCP>, <492@talcott.UUCP> Organization: U Chicago -- Linguistics Dept Lines: 193 [Naw, I don't really believe there's a line-eater.] really believe there's a line-eater. [] I've already had my say on the issue involved here (I don't think it's unreasonable for people to want computers and terminals that implement their national alphabets). So in the following I am not pursuing that top-level argument; but I do think it's time to attend to some of the facts and claims that have come up. > NO. THE GERMAN UMLAUT IS *NOT* A LETTER WITH A FUNNY SIGN ON TOP OF IT. > The German umlaut is a combination of a vowel and an 'e', which was > contracted to two little lines on top of the vowel in handwriting and > later in print. Don't ever just drop the two dots/lines. Write the 'e' > out if you must, i.e. 'A"' is 'Ae', 'o"' is 'oe', &c. I'll take your word for it, if you're talking about the history of the orthography. But if you mean something phonological, the umlaute do not indicate diphthongs. They are single vowel sounds. In general there _is_ a relation between an umlaut and (one of) the sound(s) represented by the corresponding letter without the umlaut mark, namely that it is fronted. For example, u is a high back rounded vowel, and ue (I'll write it your way, in the traditional alternate spelling) is a high _front_ rounded vowel. Only in an obscure metaphorical way is it a combination of the plain vowel and e. Even if the writer is talking about orthography only, it's still a little odd to say that an umlaut vowel *is* the plain vowel + e. So we have the following: 1. Phonologically, an umlaut is not a combination of vowel + e, nor indeed any `combination' at all. 2. Orthographically, the umlaute were once written exclusively as the sequence vowel + e. 3. Now they are primarily written as single characters. These characters could be regarded, for graphic purposes, as combinations of a plain vowel and a diacritic; the writer above urges that conceptually they should be so considered. His point is reasonable insofar as it is based on matters like alphabetization sequence; but it does not have a basis in points 1 and 2. 4. The sequence of plain vowel + e has been retained as a traditional alternate way of spelling the umlaute. I suggest that the writer should concentrate on point 4 as the basis for his claim that there is no need to implement graphic umlaute on terminals, and avoid fuzzy mystic claims that the umalute ``really are'' sequences. > 2. Take 2 letters that, when pronounced very fast after each other, > have some similarity to the wanted sound, and decree that, when > seen together, they sound different from usual. Example: au,ou,eu > in Dutch, ng in Dutch and English. This has the advantage of not > needing more letters, but the disadvantage that it creates > ambiguities: 'engrave' is pronounced as en-grave, not eng-rave. I basically agree with this writer about why we use the orthographic sequence `ng' for this sound (voiced velar nasal consonant, henceforth referred to by the name for its symbol, `engma'). Pronounced quickly as a sequence, it does have a resemblance to the desired sound. Indeed, sometimes it does represent a sequence of sounds -- but that's engma+g, not n + g. (Just as orthographic `nk' in English usually indicates the sequence of sounds engma + k.) You can hear the difference by comparing `finger' and `singer'(the former is a sequence, the latter is plain engma). In the particular example the writer mentions, in my speech `engrave' has a sequence engma + g, not n + g; it's hard to find a real n + g in English, as we almost always will velarize an /n/ in the presence of a following velar stop (/k/ or /g/). Another way of explaining why we find that sequence a reasonable way of spelling that single sound is that the sound combines features of the two sounds most commonly indicated by the letters involved: it takes the nasality of the /n/ and the backness of the /g/. A similar claim could be made about the umlaute: they combine the height and rounding of the base vowel and the frontness of the `e'. > Nasalised 'n' is > written as 'ng'. Pardon me? What is 'nasalised n'? /n/ is already a nasal consonant, you can't make it more nasal. If you're talking about changing the point of articulation, you might want the terms 'palatalised' (but not for /n/ in English or German -- this is the `gn' in French `agneau') or 'velarised' (for engma), or the more general `backed' or `retracted'. I don't mean to be a fussbudget over terminology as much as this looks; I'm just as happy with a clear description in any terms, but nasalising a nasal can only confuse people. > > English is one of the feuu languages that can get by uuithout adding to > > that alphabet (and one of the feuu that uses all of it), but only > > because uue're prepared to put up uuith such a loose connection > > betuueen sound and symbol. I iust don't belieue it uuould euen be > > possible to deuise *usable* orthographies for the many languages of the > > uuorld that relied on combinations of letters rather than special > > letters and diacritics. > > Given how complicated orthography is in English, people are doing very > well. In fact, English is one of the easiest languages to learn. > Therefore, even if a spelling reform that eliminates all national > characters would complicate the orthography slightly (which I strongly > doubt), it would probably not harm the language too much. But if a > government undertook the task of a spelling reform with the goal of > eliminating national characters, they would at the same time probably > also correct some unrelated spelling problems, which would improve > rather than worsen matters. Ah, but this is an element that wasn't evident in the previous postings. We didn't quite reach this question when some discussion of spelling reform went around a few weeks ago, but I agree with the second writer that using letter sequences under fixed conventions would be a much more reasonable approach than striving for a one-sound-one-symbol method. But why call upon languages that already have a wealth of symbols to give that up? Surely not just for the convenience of computer technology. Wait a few years until technology has worked its way around to handling these symbols more easily than at present, and then look for a spelling reform based on the real needs and resources of your language. In the meantime, in the absence of general spelling reform throughout a country, don't ask the people who use computers to adopt a different alphabet and spelling from the one in general use by their compatriots. [Sorry, it seems I _am_ commenting on the top-level issue after all.] > Why am I arguing about this at all? The existence of national > characters is a problem: it requires special equipment and impedes > trade and information exchange. I have experienced these problems > myself (being German), and I believe that the most reasonable solution > is to eliminate national characters rather than to live with the > burden, unless such an elimination is linguistically unacceptable, as > in the case of Chinese or Japanese. If there are such linguistic > reasons in the case of the Scandinavian languages, I would like to hear > about them. Mere flaming or insistence is not going to help anyone. I don't know if the writer would consider them `linguistic reasons', but some pretty good reasons have come through the net. Let me try to reconstruct the argument I was relying on in short form above. 1. There may or may not be a real need for spelling reform in a given language. 2. If you think spelling reform should aim at a one-sound-one-symbol matchup, you are then committed to an enlargement of the symbol-set, for most languages. Even if you agree (as I do) that one-sound-one- symbol is not a reasonable basis for spelling reform in some languages (and that sequences fixed by convention are one group of useful tools), still you have little reason to diminish the resources of the extant symbol-set. 3. There will be many interest groups involved in spelling reform, and it's quite premature to assume that the convenience of computer users should be the highest-priority factor. 4. In the absence of general spelling-reform in a country, it is reasonable for users of computer equipment to want to have available the same alphabet and spelling as their compatriots use for all other purposes. 5. Nonetheless, there is and would remain a real problem of compatibility in international exchange. A partial solution is to embrace an extended character-set to include many or all of the `extra' characters. (That is, characters extra to the basic Latin alphabet. This suggestion does not imply that one and the same system should try to handle radically different entire alphabets.) 6. The proposal in point 5 doesn't solve the ordering problem; but then neither does the proposal to use only sequences of the 26-letter alphabet. To clarify point 6, consider Spanish. I hope you will agree that spelling in that language is in pretty good shape; you might not like some aspects, but there is no burning need for general reform. If we don't consider accented vowels as distinct characters (they don't affect ordering), there is only one character (~n) outside the 26-letter alphabet. (The ~n could be formed, like the accented vowels, by overprinting ASCII characters; but it's still a distinct letter for alpabetisation.) Even so, there's an ordering problem. Some two-letter sequences are treated (at least in some countries) as distinct letters for ordering purposes. As best I recall, the order goes like this: a b c ch d e f g h i j (k) l ll m n ~n o p q r rr s t u v (w) x y z The ordering applies not just at the beginning of words, but wherever the pair comes up within a word or name. Now, what seems most reasonable?-- A. Expect everybody throughout the Spanish-writing world to adopt a uniform single-letter-at-a-time ordering principle, for all purposes and contexts. B. Expect them to use some listings etc ordered on the traditional basis, and others (generated by computer) ordered on a single- letter-at-a-time basis. C. Expect software sources to provide sorting methods sophisticated enough to handle the traditional order. B is pretty damn inconvenient, but I suppose it's not out of the question. I think that A _is_ out of the question. The language is not in general need of spelling reform. And the ordering principle, frustrating as it may be for computer users, reflects an underlying reality; it's not just strange tradition. So basically we're left with C; and I don't think it's asking too much. -- -- Mitch Marks @ UChicago ...ihnp4!gargoyle!sphinx!mmar