Most languages in the world use some system of alphabetical characters: Latin, Greek, Cyrillic, Hebrew, Arabic, Hindi and so on. A foreigner that does not know the language can get some information from a written text provided he knows the alphabet and disposes of a dictionary. When there is no longer an alphabet, but only pictorial characters, the problem becomes at first unsolvable. Chinese is the main language where an ignorant foreigner is completely analphabet. Fortunately there are methods that after some training allow the recognition of pictorial characters. In our university some twenty pupils of the Excellence School participated to an experiment of Chinese alphabetization gluing a traditional practical Chinese first course with information theory methods for dealing with image data bases. In this article first we discuss both the theoretical foundations. Then we give a report of the merging of the two conceptual schemes as it was performed at the excellence school. Finally we draw some conclusions about improvements of the method.
