Brahmi Ddha 𑀠

This letter is the clearest place to see what the Indian phonetic table is actually made of. It is not a sound with a sign attached but the intersection of three independent choices: the tongue curled back, the vocal folds vibrating, the release accompanied by breath. Each of those three varies on its own, and the table is the product of them, five places by two values of voice by two values of aspiration, giving twenty stops, with a nasal added at each place to make twenty-five. A learner who has the three features can predict any cell without ever having met a word that uses it.

That economy is what allowed an inventory of nearly fifty signs to be held in memory and recited in a fixed sequence by people who might never write. It is also why the order has never needed reorganising. An arrangement built on features stays correct as the language changes, because the features do not move even when the frequency of a particular cell does. Every script descended from Brahmi inherited the table intact, and the alphabets of South Asia still agree on the position of every one of the twenty-five classed consonants despite two thousand years of independent development.

One further consequence follows from building a table this way, and it is the reason the arrangement is worth calling an analysis rather than a list. Because the features are independent, a sound the compilers had never heard can still be placed: given its point of articulation, its voice and its aspiration, it has exactly one cell. That is what allowed the same table to absorb the retroflex series, and it is why the descendant scripts could file borrowed sounds under existing letters rather than rebuilding the order around them.

Middle Indo-Aryan resolved a large class of inherited clusters by assimilating them into doubled consonants, and where the cluster contained a retroflex element and an aspirate the outcome was frequently this sound written twice over. Brahmi writes a doubled consonant as a conjunct like any other cluster, the second copy stacked beneath the first, so the sign occupies two levels of the writing space rather than appearing twice side by side.

That vertical stacking is one of the defining habits of the script and it has consequences a reader notices immediately. A line of Brahmi is not a strip of equal-sized signs but a band with material hanging below it and marks sitting above it, and a cluster of three consonants with a vowel mark can occupy a good deal of vertical room. The descendant scripts handle this in different ways: some kept the stacking and became famously three-dimensional on the page, some flattened the clusters into horizontal ligatures, and some, in Southeast Asia, developed subscript forms of every consonant as a regular part of the system. Which route a script took is one of the ordinary ways of grouping the family.

Doubling has a second consequence, in metre rather than in writing. A syllable followed by a cluster is heavy whatever its own vowel, and a doubled consonant is a cluster, so the Middle Indo-Aryan assimilation that produced these doubles also changed the metrical shape of the words involved. Prakrit verse and Sanskrit verse therefore scan differently even where the vocabulary corresponds word for word, and a reader working out the metre of an inscription has to know which stage of the language is in front of them.