Comment by umanwizard

Because then it would be impossible to tell from looking at a byte whether it is the beginning of a character or not, which is a useful property of UTF-8.