The Prakrit of the Ashokan edicts had no syllabic vowels, so this letter, like the two r vowels, has no place in the earliest use of the script. It belongs to the later Sanskrit-writing Brahmi, and even there its occasions are countable. Its survival through two thousand years and across two dozen descendant alphabets is therefore evidence about how the order of letters was transmitted rather than about the language.
The order was memorised and recited before it was ever written out as a list, and a recited sequence resists loss in a way that a working inventory does not. A letter drops out of an inventory when nothing needs it; a letter does not drop out of a recitation, because the recitation is the thing being preserved. That is why the alphabets of South Asia agree on the position of a sign that almost nothing is spelled with, and why a script adapted for a language with no such sound still carries it. The same conservatism explains features of the order that make no sense from the sounds alone, including the long syllabic l that follows this letter and has no attested word behind it in any period of the language.
The same conservatism is visible in the descendant scripts of Southeast Asia, which carry the syllabic vowels forward into alphabets used for Khmer, Thai, Javanese and Burmese, none of which has ever had such a sound in its own vocabulary. Those letters are used there for writing Sanskrit and Pali loanwords and for the religious terminology that arrived with Indian religion, so they are not quite idle, but they exist because the order was received whole rather than because the languages required them.