ICEMorph: Toward a Second Generation Analyzer
ICEMorph is a second generation morphological analysis and look-up tool for the study of Old Icelandic / Old Norse, the most morphonologically complex of the Germanic
languages.
The analysis tool uses a second generation formal language, FM/Haskell, to tackle the problem of this complexity.
The look-up tool is based on two of the most important dictionaries for Old Icelandic / Old Norse study: Cleasby-Vigfusson
An Icelandic-English Dictionary (1874), and Johan Fritzner's Ordbog over det gamle norske Sprog ( 1883-1896).
In its pilot phase, the look-up tool is based on Zoega's subset of Cleasby-Vigfusson, and this dictionary supplies the lexical set for
the analysis tool.
The Analyzer
The analyzer currently exists in several forms. Our first generation analyzer, written in Perl, and based on revised notions of cascading rewrite rules, is still operational on our developer server. A more detailed discussion of the architecture of the first-generation analyzer, and the revised architecture behind the second generation morphological analyzer can be found on the architecture page.
The analyzer is designed to take as input any lemma or any form, and return the dictionary definiton of the lemma, and the inflectional paradigm for that word. So, for example, inputting either the word ketill or any of its inflectional forms, such as the gen. sg. ketils, the analyzer will return the following dictionary entry, "ketill, noun, gender: m, a-stem", followed by the paradigm:
|
The tool can be used to tag texts, and those texts in turn can be explored using various word-study tools or visualization tools such as those incorporated into Greenstone Digital Library 3.0.
 |
The goal is to incorporate the analyzer into a rich markup and study environment, in which scholars can work with both manuscript images, diplomatic transcriptions, and normalized versions of the texts. Such an environment might look as follows:
 |
The first generation morphological analyzer can be found here. Although it returns forms with an accuracy of 90% across the entire lexical set that it recognizes, it is constrained by two problems: (1) debugging the Perl system is complex and is a barrier to increased accuracy and extensability of the system to other Indo-European languages and (2) the lexical set that it recognizes is currently limited to Zoega.
The first generation analyzer has also been connected to the Legendary Sagas and the word study tools at Perseus.
The second generation analyzer has only been implemented for weak nouns to date, and is running as a stand alone application. Once we reach the level of accuracy and coverage currently achieved by the first generation analyzer, we will enable it for the web, and turn off the first generation analyzer.