Eiríkur Rögnvaldsson

A Mixed Method Lemmatization Algorithm Using a Hierarchy of Linguistic Identities (HOLI)

We present a new mixed method lemmatizer for Icelandic, Lemmald, which achieves good performance by relying on IceTagger [1] for tagging and The Icelandic Frequency Dictionary [2] corpus for training. We combine the advantages of data-driven machine learning with linguistic insights to maximize performance. To achieve this, we make use of a novel approach: Hierarchy of Linguistic Identities (HOLI), which involves organizing features and feature structures for the machine learning based on linguistic knowledge. Accuracy of the lemmatization is further improved using an add-on which connects to the Database of Modern Icelandic Inflections [3]. Given correct tagging, our system lemmatizes Icelandic text with an accuracy of 99.55%. We believe our method can be fruitfully adapted to other morphologically rich languages.

Morphosyntactic Tagging of Old Icelandic Texts and Its Use in Studying Syntactic Variation and Change

We describe experiments with morphosyntactic tagging of Old Icelandic (Old Norse) narrative texts using different tagging models for the TnT tagger [3] and a tagset of almost 700 tags, originally developed for Modern Icelandic. It is shown that by using a model that has been trained on both Old and Modern Icelandic texts, we can get 92.7% tagging accuracy which is considerably better than the 90.4% that have been reported for Modern Icelandic. Although our tagging is morphological in nature, the tags carry a substantial amount of syntactic information and the tagging is detailed enough for the syntactic function of words to be more or less deduced from their morphology and the adjacent words. We show that the morphosyntactic tags can be very useful in locating certain syntactic constructions and features in a large corpus of Old Icelandic narrative texts. We demonstrate this by searching for—and finding—previously undiscovered examples of a number of syntactic constructions in the corpus.We conclude that in a highly inflectional language, a morphologically tagged corpus can be an important tool in studying syntactic variation and change, in the absence of a fully parsed corpus which of course gives more possibilities.

Some Comments on Reflexivization in Icelandic

In this paper I will make some comments on Anderson’s and Maling’s papers in this volume on reflexivization in Icelandic. I will show how that Anderson’s theory makes some predictions which are not borne out by the facts, while Maling’s notion of predication often makes it difficult to see what the predictions of her theory actually are. Occasionally, I will hint at other explanations, but these comments do not present any coherent theory of Icelandic reflexivization.