Thai does not separate words by spaces, so I figured out it should be in 'reliable characters regexp' that denotes languages that do the same. Related #13891.