A semi-unsupervised language independent morphological analyzer useful for stemming unknown language text, or getting a rough estimate of possible parses for morphemes in a word. Input: a corpus. Uses compression, maximum entropy and fieldlinguistics.