I think that's a good argument against using character count as a measure of program length. i.e. translating 印 to print does not improve the program in any way, so it shouldn't score better in any measure of complexity.
'print' is 5 characters long, 印 is 1 character long. Though in most fonts, CJK characters are twice as wide as Latin, so better make that 2 chars long. 5 chars vs 2 chars over an entire program adds up. In non-indentation languages, that can mean 3 lines can be put onto 1 line. Shorter programs are more readable, have less errors.
