moyogo
commented
3 years ago @projectshifter k̴ , ð̴ and θ̴ are used in Olivier Durand, 2017, “Remarques sur le consonantisme de l’arabe”, in Folia Orientalia, vol. 54, k̴ is also used in Ashraf Hassan, 2020, Stereotyped Representation of the Foreigner in Egyptian Cinema: A Phono-Morpho-Syntactic and Lexical Study and Corpus, itself citing Manfred Woidich, 2006, Das Kairenisch-Arabische: eine Grammatik which also uses it.
ʔ̴ and h̴ appear in Arne Krause, 2018, Transkription Deutsch-Arabischer Diskurse.
v̴ is used, along with ð̴, in http://www.eurasianphonology.info/listview?lang=Tillo+Siirt+Arabic
So that's v̴ ʔ̴ k̴ ð̴ h̴ θ̴ that are not encoded as single characters.
pauldhunt
projectshifter
An overlaid tilde is used in the IPA to represent velarization or pharyngealization. In many varieties of English, this is the difference between the light [l] sound in leaf and the dark [ɫ] sound in feel. Given that its usage overlaps with that of the superscripts ˠ and ˤ, it is somewhat rare, but it’s still officially part of the IPA.
Unicode proposal N2632 from 2004 made the case that representing phonetic characters with middle tilde using U+0334 COMBINING TILDE OVERLAY is inappropriate, given that the number of symbols that could reasonably take a tilde overlay is relatively small, and that the variations in size, shape and position of the tilde for each base glyph cannot be easily automated. The 11 characters from this proposal were approved for Unicode 4.0: U+1D6C–U+1D76. One more was added in Unicode 13.0: U+AB68 LATIN SMALL LETTER TURNED R WITH MIDDLE TILDE. However, very few fonts support these precomposed characters, and sequences with U+0334 are far more common. For historical reasons, Unicode doesn’t treat these sequences as canonically identical, but in practice, for example, U+0073 U+0334 s̴ is essentially a stand-in for U+1D74 ᵴ.
Here are the glyphs I’m requesting, in both their composed and decomposed forms.
N2632 identifies four more base glyphs that could potentially take a tilde overlay (v ʔ k ð), but the author couldn’t find any evidence for them. I have found several instances of an eth with middle tilde published since then; they are encoded as U+00F0 U+0334. Since most fonts don’t have anchor points for overlaid diacritics, the glyphs shown are probably ad-hoc, but evidence from UCLA’s handwritten field notes suggests that the tilde should go through the ascender rather than the bowl of the eth.