We propose a language-independent word normalization method exemplified on modernizing historical Slovene words. Our method relies on character-based statistical machine translation and uses only shallow knowledge. We present the relevant lexicons and two experiments. In one, we use a lexicon of historical word--contemporary word pairs and a list of contemporary words; in the other, we only use a list of historical words and one of contemporary ones. We show that both methods produce significantly better results than the baseline.