Extending the tool, or how to annotate historical language varieties

We present a general and simple method to adapt an existing NLP tool in order to enable it to deal with historical varieties of languages. This approach consists basically in expanding the dictionary with the old word variants and in retraining the tagger with a small training corpus. We implement this approach for Old Spanish. The results of a thorough evaluation over the extended tool show that using this method an almost state-of-the-art performance is obtained, adequate to carry out quantitative studies in the humanities: 94.5% accuracy for the main part of speech and 92.6% for lemma. To our knowledge, this is the first time that such a strategy is adopted to annotate historical language varieties and we believe that it could be used as well to deal with other non-standard varieties of languages.