Mongolian Sha ᠱ

Sha writes the voiceless palato alveolar sibilant, the sound spelled sh in English, and it is the seventeenth letter of the alphabet. Mongolian has both sibilants natively and keeps them apart, and the alphabet supplies a separate letter for each, inherited from the same Aramaic ancestry that gave it the plain s. Unicode encodes sha at U+1831. The two sibilants behave differently in the history of the language, and one of the differences is visible in the vocabulary: the palatal sibilant is common before certain vowels in native words and turns up constantly in words taken from Tibetan and Sanskrit, where it renders more than one distinct source sound. That is the point at which two letters stop being enough. Sanskrit distinguishes three sibilants, and a Mongolian scribe writing a Sanskrit name accurately needs to mark distinctions the Mongolian alphabet was never built to carry.

The extended signs encoded under the name Ali Gali exist for this, alongside the ten letters added at the end of the alphabet for other foreign sounds. A reader meeting an unfamiliar sibilant shape in a Buddhist text is usually looking at one of those additions rather than at a native letter. In the traditional recitation the two sibilants stand next to each other, which reflects both their sound and their shared descent from a pair of letters that were already adjacent in the alphabets Mongolian inherited from.

Transcribing Sanskrit into a script that was not designed for it is an old problem with a standard shape, and the Mongolian solution belongs to a family of solutions found across Buddhist Asia. Tibetan built a script from an Indian model and could represent Sanskrit almost directly. Chinese, with no alphabet at all, transcribed Sanskrit syllable by syllable using characters chosen for their sound and accepted heavy losses. Mongolian sat between the two, holding an alphabet that could represent most consonants and vowels but could not mark the aspirated and retroflex series that Indian phonology turns on, nor keep three sibilants apart, nor write the clusters Sanskrit permits.

The response was to extend the alphabet with additional signs rather than to abandon it, producing a transcription system capable of writing Sanskrit and Tibetan in full while remaining recognisably the same script. The signs are encoded in Unicode under the name Ali Gali, separately from the thirty-five letters of the alphabet proper, which is the right arrangement: they are an apparatus for writing other languages, not letters of the Mongolian alphabet, and counting them into it would misstate what the alphabet is.