Ranjana Ha

Ha is the thirty third and last consonant of the Ranjana varnamala, a voiced sound made at the glottis, and it stands alone at the end of the chart because it belongs to none of the groups before it. It is not a stop, so it cannot enter the grid; it is not a semivowel, since it involves no narrowing in the mouth itself; and it is not a sibilant, since the friction is made at the vocal folds rather than at a point in the mouth. The tradition therefore lists it last, as a class of one, and the recitation of the order ends on it.

In Ranjana it is written as any consonant, carrying the inherent short a and hanging from the head stroke, and it takes vowel marks and enters joined forms like the rest. Its position at the close of the order gives it a particular place in the teaching of the script, since a learner who reaches it has completed the varnamala, and a recitation that reaches it has finished. Its place at the end also gives it a role in the writing the script is best known for. Ha is the final letter of the syllable that closes the six syllable mantra written in Ranjana on prayer wheels and carved stones, and it is the base of several of the seed syllables that Tibetan traditions write in this script under the name Lantsa. What is public there is the writing.

Ha is the letter that the unresolved question about Ranjana's inventory turns on. Nepal Bhasa, the Newar language for which the script was used alongside Sanskrit, has six aspirated sonorants that Sanskrit does not, and the question is whether the Ranjana tradition counts them as letters of its own varnamala or writes them as combinations of a base letter with this one.

The atlas does not answer it. There is no character standard to appeal to, because Ranjana is unencoded; the openly licensed typeface used for the outlines here ships one build that can reach these six shapes and one that cannot, depending on which script's code points it is mapped onto; and neither the font nor the sources consulted state what the tradition itself holds. Six identifiers have been reserved in case the answer is that they are letters, and the question is recorded with what would settle it: a reading of the alphabet charts and primers used to teach Ranjana in the Kathmandu Valley, or of the encoding proposals put to the Unicode committee. Leaving it open is the point. Deciding it by assumption would put a wrong count of a living community's letters into a reference work, which is a worse outcome than an admitted gap.