R> An example I usually see in Semitic Transliteration: The Arabic R> letters kha and ghain are usually romanized by underbarred digraphs R> kh and gh, respectively [...] to indicate laryngeal(-ized) phonemes R> (laryngeal fricatives kh, gh in Hebrew, Arabic, and languages using R> Arabic script;
As I understand it, the point of double underbarring these is to distinguish "kh" from separate "k" + "h"; for example, the Encyclopaedia of Islam also transliterates Arabic sheen with underbarred "sh" to distinguish it from "s" + "h", where there is no laryngealization at all R> however, laryngealized sibilants s,/i>, z occur in Farsi, R> Urdu, and most others of those languages [in Classical Arabic, they R> appear as interdental fricatives th, dh]); They are not laryngealized in these languages; instead, Arabic interdentals get pronounced as ordinary [s] or [z] respectively in Arabic loans to Farsi, Urdu etc. The transliteration comes from the same ZDMG system where the Arabic interdentals in Arabic are transliterated as t and d with combining macron below, U+1E6F and U+1E0F. In order to be able to reconstruct the Arabic orthography in these nouns, but to also reflect the Persian pronunciation, the ZDMG recommends to use s and z with combining macron below (U+0073 U+0331 and U+1E95). Laryngealization (pharyngealization, actually) gets represented with a combining dot below in ZDMG transliteration. There's a short paper on Arabic transliteration by yours truly at http://www.orientasia.uni-bonn.de/downloads/arab_trans.pdf, with an overview of the most common systems. ZDMG gets explained in detail because it's what we use at our university :-) R> a laryngeal resonant h also exists (please tell me in what R> languages that laryngeal 'h' would be found as a distinct phoneme}. Well, laryngeal "h" is a problem because our common [h] is laryngeal already. Arabic has an additional epiglottal, "hard" h-sound: "ha" (U+062D) of course, usually transliterated by U+1E25 (LATIN SMALL LETTER H WITH DOT BELOW). Arabic "kha" is transliterated by U+1E2B (LATIN SMALL LETTER H WITH BREVE BELOW) in the ZDMG transliteration system used in Germany and most of Eastern Europe; on systems without the BREVE BELOW is not available, this often gets transliterated with a macron below: U+0068 U+0331, maybe you mean this. As far as the "distinct phonemes" are concerned, there appears to exist at least one Caucasian language that has contrasting pharyngeal, epiglottal and laryngeal "h" sounds (IPA U+0266, U+029C and U+0068, respectively, IIRC). Mail me off-list if you need a reference. On a side note, this e-mail really convinced me that I need a Unicode-capable mail client. Philipp
