This is the second of a series of 12 blog posts published in 2019, discussing open problems in computational diversity linguistics. It presents the problem of automatic morpheme segmentation.