ROMLEX - the lexical database of Romani varieties

The aim of the project is a comprehensive and representative documentation of the Romani lexicon, which i.a. allows cross-variety comparison as well as in-depth intra-variety studies of the characteristics of Romani. Ten selected varieties of Romani representing the manifold of diversity among Romani will be linked in a custom-made database on the basis of different linguistic parameters. The individual glossaries of the ten selected varieties will each display vocabulary representative of a number of selected basic domains, which guarantees the cross-variety comparability. While homogenising the lexical coverage of the individual glossaries, as well as the orthography, a major aim of the project is to allow for the greatest variation with respect to individual lexical entries. Thus, instead of implementing a unified "high variety" of Romani through the back door, the current approach proposes a new model of language documentation for non-standardised, heterogeneous, stateless languages, allowing for all the variation necessary to reflect the reality of such a language. Furthermore this approach takes into consideration the sociolinguistic situation of speech communities, which consist of a large number of heterogeneous groups that lack political power and/or economic resources. The current project proposal is based on the output of three previous working phases. In these earlier project phases the development and programming of the database, using most recent standards of electronic documentation of language and up-to-date open source database technologies, was mostly completed. With respect to contents, at present ROMLEX is a collection of 25 electronic glossaries, edited in a unified writing system, with basic morphosyntactic categorisation and translations into various European languages (for details see: http://romani.uni- graz.at/romlex/). While the individual 25 varieties are already homogenised with respect to codification, i.e. the writing system implemented, lexical coverage in ROMLEX at present is very heterogeneous and makes cross- variety comparison a difficult matter. With the project phase described in this proposal, current problems of ROMLEX, such as cross-variety comparison, will be solved. At the end of the project phase applied for, ROMLEX will represent not only the most comprehensive, but also the most elaborated source of lexicographic information on Romani to date. ROMLEX will be accessible to the scientific community, and in an adapted way in a future step also to the general public, via an online web-interface with custom-made, easy to handle, search functions, which also will make ROMLEX fertile for complex linguistic questions. Beyond its importance for Romani linguistics - providing the most comfortable way imaginable to conduct linguistic studies on Romani, saving both time and work effort, without sacrificing the highest scientific standards - ROMLEX will represent an innovative model of language documentation and codification without standardisation, which suffices the demands of a heterogeneous, stateless language of a diaspora minority. Furthermore, it is an example of sensible use of up-to-date technology in the humanities. As such ROMLEX illustrates the potential of such technologies for linguistic research in particular, and the humanities in general. Finally, beyond its scientific contribution, the project will have a significant and unprecedented impact on the state of literacy in Romani, making for the first time accessible such a broad range of lexical data, with such extensive professional editing. ROMLEX is not only a valuable tool for scientific research, but provides the speech community with a resource, that will, and already has, proved very useful in text production by professional translators, literary authors and political activists.

ROMLEX documents the vocabulary of Romani. The vocabulary of more than 25 different Romani varieties is currently freely accessible online (http://romani.uni-graz.at/romlex/) . The range includes translations into English, German and other relevant majority languages such as Hungarian, Romanian, Serbian, etc. ROMLEX moreover offers grammatical and etymological information on individual words. The project stage of ROMLEX supported by the FWF focussed on ten selected Romani varieties (Arli Romani, Bugurdži Romani, Burgenland Romani, East Slovak Romani, Gurbet Romani, Kalderaš Romani, Lovara Romani, North Russian Romani, Sinte Romani and Ursari Romani). The most important results concerning these varieties can be summarised as follows: 1. the lexical expansion and homogenisation of the vocabulary of the chosen varieties based on a query list comprising approximately 5,000 words 2. the linking of individual words to form word families within varieties 3. the etymologising of these word families 4. the linking of these word families spanning the varieties based on their etymology The data for the lexical expansion were collected through queries of competent speakers of the ten varieties and not only close the gaps in the basic vocabulary, but moreover also document the use of Romani in public-formal domains. Grammatical information which had as yet been lacking was also added. ROMLEX is thus also a useful instrument for text production. The expansion based on a standardised list of words allows for a comparison of the strategies for expanding the lexicon which are applied within the functional expansion (borrowing vs. derivation vs. periphrasis). The internal linking connects lexical entries to form word families. This visualises the actual selection from potential morphological word formation strategies as well as their frequency within each variety. The word families are finally linked to form super families spanning all varieties and the respective derivational base is provided with etymological information. The database thus structures the lexical material for research comparing the varieties. Zusammenfassung für die Öffentlichkeitsarbeit ROMLEX dokumentiert das Vokabular des Romani. Gegenwärtig kann man über eine frei zugängliche Webseite (http://romani.uni-graz.at/romlex/) das Vokabular von über 25 verschiedenen Romani-Varietäten konsultieren und erhält Übersetzungen ins Englische, Deutsche und andere relevante Mehrheitssprachen wie z.B. Ungarisch, Rumänisch, Serbisch, etc. Darüber hinaus bietet ROMLEX grammatische und etymologische Informationen zu den Einzelwörtern. In der durch den FWF geförderten Projektphase von ROMLEX lag der Arbeitsschwerpunkt auf zehn ausgewählten Romani-Varietäten (Arli-Romani, Bugurdži-Romani, Burgenland- Romani, East Slovak-Romani, Gurbet-Romani, Kalderaš-Romani, Lovara-Romani, North Russian-Romani, Sinte-Romani und Ursari-Romani). Die wichtigsten Ergebnisse in Bezug auf diese Varietäten lassen sich wie folgt zusammenfassen: 1.) die lexikalische Erweiterung und Homogenisierung des Vokabulars der ausgewählten Varietäten auf Basis einer ca. 5000 Wörter umfassenden Abfrageliste 2.) die varietäteninterne Verlinkungen von Einzelwörtern zu Wortfamilien 3.) die Etymologisierung dieser Wortfamilien 4.) die varietätenübergreifenden Verbindungen dieser Wortfamilien auf Basis ihrer Etymologie. Die Daten für die lexikalische Erweiterung wurden mittels Abfragen von kompetenten SprecherInnen der 10 Varietäten gesammelt und schließen nicht nur Lücken im Grundwortschatz, sondern dokumentieren darüber hinaus auch den Gebrauch des Romani in öffentlich-formellen Domänen. Auch bisher fehlende grammatikalische Informationen wurden ergänzt. Somit wird ROMLEX auch zu einem nützlichen Instrument für Textproduzenten. Durch die Erweiterung auf Basis einer standardisierten Wortliste werden die im Zuge der funktionellen Expansion angewandten Strategien zur Erweiterung des Lexikons (Entlehnung vs. Ableitung vs. Periphrase) vergleichbar. Bei der internen Verlinkung werden lexikalische Einträge zu Wortfamilien verknüpft. Somit wird die tatsächliche Auswahl aus den potenziell möglichen morphologischen Wortbildungsstrategien sowie deren Häufigkeit innerhalb einer Varietät sichtbar. Die Wortfamilien werden schließlich zu varietätenübergreifenden "Superfamilien" verknüpft und die jeweilige Ableitungsbasis wird mit etymologischen Informationen versehen. Somit strukturiert die Datenbank das lexikalische Material für varietätenvergleichende Untersuchungen.