Layout

You are using an outdated browser. Please upgrade your browser to improve your experience.

Dictionary: Overview

The dictionary was not only built on data from the corpus of spoken language that was compiled in the same project, but also on a range of additional sources: data elicited from complementary interviews with young Tunisians and lexical material taken from various published historical sources dating from the middle of the 20th century and earlier. The most important of these is Hans-Rudolf Singer’s monumental grammar (1984; almost 800 pages) of the Medina of Tunis. Singer’s data was systematically evaluated and integrated into the dictionary, all the material being indicated by reference to the book. Additionally, other resources including (Nicolas 1911, Marçais/Guîga 1958-61, Quéméneur 1962, Abdellatif 2010) were also consulted in order to verify and to complete the contemporary data. The diachronic dimension will help to better understand processes in the development of the lexicon (for more details see Moerth, Prochazka, & Dallaji 2014).

The first layer of lexical data to be incorporated into the digital lexicographical system was extracted from Veronika Ritt-Benmimoun’s textbook titled Skriptum zu den Lehrveranstaltungen Tunesisch-Arabisch Kurs A und B (Vienna, 2012/13). This data is referenced in the dictionary as Ritt-Benmimoun 2014.

Please keep in mind that the dictionary will remain work-in-progress. We do not claim any degree of completeness, neither with respect to lemmas nor with respect to data relevant to the entries’ microstructure. This holds particularly true for etymologies and usage labels indicating regimen of verbs.

The project was embedded in the activities of the two large-scale pan-European research infrastructure consortia in the humanities, CLARIN (Common Language Resources and Technology Infrastructure) and DARIAH (Digital Research Infrastructure for the Arts and Humanities) (Budin, Moerth, & Durco 2013).

Data

The TUNICO project was conducted in the spirit of open access. You can download the dictionary data by clicking here here.

Query Interfaces

The compilation of the dictionary was accompanied by the ACDH-OeAW’s efforts to build an integrated Virtual Research Environment for lexicographers. This has had an influence on work practise as well as the presentation of the results. In the following explanations, you will find links to three different dictionary web-frontends which all build on the same lexical data but make use of different technologies. This has got to do with the experimental nature of the technical parts of the project. As the functionality of these tools is not exactly the same, we leave it open to our users to choose by themselves.

Interface I

The data of our dictionary were first made public on the VICAV Website. The Vienna Corpus of Arabic Varieties (VICAV) is an international project aiming at the collection of digital language resources documenting varieties of spoken Arabic. It provides a wide range of materials such as language profiles, dictionaries, annotated texts, bibliographies and more. There the Tunis dictionary can be accessed together with other dictionaries (Damascus, Cairo). The VICAV dictionary interface was designed to allow users to compare Arabic varieties.

Interface II

The main access point to the dictionary is part of this web site. Technologically it is part of the ACDH-OeAW's corpus_shell framework, a modular service-oriented architecture for a distributed and heterogeneous landscape of digital language resources. The principle idea behind the architecture is to decouple the modules serving data from the user-interface components. To achieve this end, a number of basic requirements are imposed on the system: dynamic configuration of data sources, dynamic configuration of front-end layout, support for different protocols and support for different data formats. The backbone of the system is the SRU/CQL search protocol which supports FCS (Federated Content Search), an evolving standard increasingly used in the pan-European CLARIN (Common Language Resources and Technology Infrastructure) community.

Interface III

The most recent achievement was an innovative interface that - by contrast to the above mentined ones - builds entirely on XML and cognate technologies. The web frontend interacts with Basex, an XML database. Queries are executed in pure xQuery,data is delivered via a REST interface, making use of XQuery and XSLT. This part of the Academy's Dictionary-in-a-box project was finished during the last weeks of the TUNICO project and is fully integrated with the Viennese Lexicographic Editor (VLE). The database is easy to install, can also run on local machines (i.e. does not necessarily need a web server) and allows researchers to easily visualise and publish their lexicographic work. VLE comes with a wizard that allows to create a new interface at the push of a button making use of ready-made templates.

Encoding

The dictionary are encoded in TEI (P5). While the data model applied for the corpus was easy to implement, the model of the lexicographic data needed some work as comparable data was not available. Major topics that were discussed in conferences and publications were the issue of diachrony in TEI encoded dictionaries, modelling of statistical data in lexical resources, the issue of language identifiers (in particular in research that needs a high degree of granularity) as well as the dictionary-corpus interface.

The encoding documentation of the dictionary are part of the VICAV dcomentation which has been enhanced and refined through data and experiences gathered during the TUNICO project. The coumentation can be found on the website of the ACDH-OeAW's DictGate.

Tools

Several tools were further developed as part of the TUNICO project.

Viennese Lexicographic Editor (VLE)

The main tool for editing the lexicograph data of the project was the Viennese Lexicographic Editor (VLE). VLE is a standalone Windows application that allows lexicographers to process standards-based lexicographic and terminological data in basically any XML-based format such as Lexical Markup Framework (LMF; ISO 24613:2008), TermBase eXchange (TBX; ISO 30042:2008), Resource Description Framework (RDF) or TEI. This general purpose XML editor provides a number of functionalities to streamline lexicographic editing procedures. It allows collaborative dictionary editing on the internet and is built entirely on XML and related technologies (XPath, XQuery, XSLT, XML Schema). While the tool for several years was functioning as part of a client-server-based architecture making use of MySQL in the backend, most recent VLE versions can be used in combination with the free and easy-to-use XML database BaseX (http://basex.org/).

tokenEditor

The tokenEditor loads corpus texts vertically, i.e. every token on a separate line. It furnishes methods to add labels to the tokens and to manually review automatically applied annotations. The tool allows users to filter, aggregate, visualise and – most importantly – edit word-level annotations in digital texts. The tool was implemented in several applications and integrated in our semi-automatic NLP pipeline. Although it was initially an independent Windows application, the development has until recently been tightly integrated with the development of the lexicographic editor. Most recent VLE versions all feature a built-in tokenEditor component which allows for the seamless integration of corpus and dictionary. This module takes as input XML texts and/or text collections and creates verticalised views of the data. This process is accomplished on the basis of freely configurable XPath expressions which create lists of XML elements on which a wide range of operations can be performed. The tokenEditor can create and display subsets of the list items through the application of regular expressions. It allows the manual correction of the linguistic information. PoS and lemma information can be changed for single tokens or in batch for multiple occurrences. What is key for the usability of the tool is that users have access to the context, can browse through the filtered lists and still can access the underlying corpus. While the tool was primarily designed to review part-of-speech tags and lemmas, it is actually fully customizable and capable to support any number of annotation layers.

Other Tools

In the process of creating the dictionary, several other tools were also used. The most important of these are oXygen and Audacity.