Computational morphology is the study of capturing the morphology of a language in a way that can be understood by computer systems. Due to the wide variety of [[derivational morphology]] and [[inflectional morphology]] of different languages, computational morphology relies heavily on probabilistic models. Approaches that work in one language do not work in others. For that reason, even basic applications like spelling correction have not been developed for many underrepresented languages. There are 7,097 languages in the world and only 3,500 have written forms, but most of those do not have large corpora for the purposes of training models. Further, there are 350 writing systems in use around the globe (e.g., Latin, Chinese, Arabic, Cyrillic, Hebrew, Devanagari, Bengali). [[Unicode]] supports most of the writing systems.