Well, the assumption was based on the fact that they chose base 26, and that the "26" came from the use of a 26 character alphabet. The 10 Arabic numerals are then a convenient character set to expand to 36, which is a much nicer number base than 26.

Japanese could be combined with the Hindi character set to yield base 96, which is fairly convenient. Cyrillic would be harder -- perhaps the best options there would be to drop a character to yield base 32, or perhaps 3 characters to yield base 30.

I'd argue that base 60 is probably the optimal number base for nearly any use (with base 16 or 64 as close second and third for working with binary data). Hindi's 50 characters combined with our 10 Arabic numerals could indeed be a great way to get there.