> The latter problem could be solved easily by transcribing à as "dh," but > English speakers seem really terrified of the sequence "dh."
Not quite so fast. Where a d can end a syllable and an h can start one, then it can collide with dh representing Ã. The general issue is that whenever you use a sequence of letters in the target for transliteration/transcription, and the elements of that sequence can individually be targets, then you can get ambiguity. There are mechanisms to separate a sequence of letters that would otherwise be read as a unit: apostrophe: as in Japanese transliterations (When vowels or consonant y follow the syllabic nasal n, ng, m, add apostrophe (') after "n". Example: ren'ai / gen'in / sin'en / kon'ya -- Cabinet Order (Kunrei) No.1) hyphen: as in the Korean Ministry of Education transliteration, to distinguish "jeong-eum versus jeon-geum ")... diaeresis on second element: (doesn't work very well, since it only really sits well on vowels). > Transcriptions are another matter; the reader can read "Tchaikovsky" or > "Beijing" without knowing anything at all about Cyrillic or Chinese, and > still come close (theoretically) to the real pronunciation. Agreed about the distinction in meaning between 'transcription' and 'transliteration'. However, the two examples of transcriptions are not necessary good ones, at least for English speakers: the only reason that English speakers will read 'Tchaikovsky' reasonably is because they have learned the word, since it doesn't follow normal English orthographic rules. ÎÐrk ----- Original Message ----- From: "Doug Ewell" <[EMAIL PROTECTED]> To: "Unicode Mailing List" <[EMAIL PROTECTED]> Cc: "Mark Davis" <[EMAIL PROTECTED]>; "Mike Ayers" <[EMAIL PROTECTED]> Sent: Saturday, July 03, 2004 09:40 Subject: Re: Looking for transcription or transliteration standards latin- >arabic > RE: Looking for transcription or transliteration standards > latin->arabicMark Davis wrote: > > > In that case, we'd call it a transcription, since it doesn't roundtrip > > from source to target back to source. It is actually quite common for > > style guides for non-academic publications to have a restricted list > > of characters and character + accent combinations, and convert all > > others. For example, the Economist style guide, as I recall, > > recommends keeping accents in French, German, Italian, and Spanish > > names and words, but dropping them otherwise; and converting > > characters like à and à to nearest equivalents, "th". > > > > Note that the latter loses information in two ways; the obvious one is > > that the distinction between à and à are lost; the less obvious one is > > that the distinction between them and a *real* 't' followed by 'h' in > > the source is lost. So that loses the distinction in sounds between > > 'th' in 'cathode' and 'cathouse', as well as between 'thy' and > > 'thigh'. > > The latter problem could be solved easily by transcribing à as "dh," but > English speakers seem really terrified of the sequence "dh." > > The former problem is only a problem if "t" + "h" combinations (like > "cathouse") are actually used in the language. I don't know if this is > true for Icelandic. It is certainly true for Old English, where à and à > are also seen. > > -Doug Ewell > Fullerton, California > http://users.adelphia.net/~dewell/ > > > ----- Original Message ----- From: "Doug Ewell" <[EMAIL PROTECTED]> To: "Unicode Mailing List" <[EMAIL PROTECTED]> Cc: "Anto'nio Martins-Tuva'lkin" <[EMAIL PROTECTED]> Sent: Saturday, July 03, 2004 14:22 Subject: Re: Looking for transcription or transliteration standards latin- >arabic > Anto'nio Martins-Tuva'lkin <antonio at tuvalkin dot web dot pt> wrote: > > >> Only specialists can make sense of them, > > > > Pray tell, why so? Is the letter "Ã" an usuperable obstacle for those > > who know only the letter "a"?... > > > > Can't the "remove diacriticals" action be performed in the reader's > > brain, instead of in the typesetter's office? > > But if the reader merely removes the diacriticals, that destroys the > whole purpose of using a *transliteration* scheme, where 'a' and 'Ã' > represent different letters in the source writing system. > > Jony's point (I think) was that only specialists can keep track of which > target characters represent which source characters, especially when > obscure diacritics or digits or other symbols are used. At that point, > the specialist probably knows the source characters well enough to read > them directly, and the widespread use of Unicode enables document > producers to use them directly. > > Transcriptions are another matter; the reader can read "Tchaikovsky" or > "Beijing" without knowing anything at all about Cyrillic or Chinese, and > still come close (theoretically) to the real pronunciation. > > -Doug Ewell > Fullerton, California > http://users.adelphia.net/~dewell/ > > > >