What the Dot Does, and What It Does Not

The dot written beneath a Devanagari letter is called the nukta, and it is worth being clear about its scope. It does not modify the letter it sits under in any general way; each nukta letter is a separate character with its own value, established by convention for the particular language that uses it. Hindi uses a set of them for sounds borrowed with Persian and Arabic vocabulary, and Hindi orthography has been debated for a century over how consistently they should be written, since many speakers do not distinguish the borrowed sounds from the native ones and many printers omit the dots. The Unicode Standard encodes both the combining nukta and, for several of the commonest cases, precomposed characters, which means a text can represent the same word in more than one way and software has to normalise between them.

None of this touches the classical letter itself, whose value in Sanskrit is the voiced unaspirated retroflex and whose place in the varnamala is third in the third class. The atlas records the classical value as the letter's value and notes the modern use as modern use, because collapsing the two would misdescribe both. For a reader the practical consequence is that a Devanagari letter can carry two different values depending on whether a dot is present beneath it, and that the dot is easy to lose at small sizes, in a scanned page or in a font that positions it badly. Where a letter in this atlas has a dotted variant with a distinct value, that variant is stated on the letter's own page rather than left to be inferred.