Journal for Language Technology and Computational Linguistics (JLCL)

Cramer, Irene

Finthammer, Marc

Kurek, Alexander

Sowa, Lukas

Wachtling, Melina

Claas, Tobias

English

Experiments on Lexical Chaining for German Corpora: Annotation,
Extraction, and Application

Sporleder, Caroline

Manually vs. Automatically Labelled Data in Discourse Relation Classification: Effects of Example and Feature Selection

Leopold, Edda

Kindermann, Jörg

Paaß, Gerhard

Analysis of E-Discussions Using Classifier Induced Semantic Spaces

Mönnich, Uwe

Kühnberger, Kai-Uwe

iiiLDV FORUM - Band 22(2) - 2007
1 Introduction
The rise of the world-wide-web in connection with the tremendous increase of electro-
nically available textual data of all kinds, types, genres, and forms make the scientific
study of text resources a trend-setting research endeavor. The joint work of researchers
trained in different disciplines and research traditions encompassing the theoretical
study of properties of texts, text transformations, markup languages, query languages,
and text structures, as well as the practical pursuit of archiving textual information,
retrieving background knowledge from texts, and adapting dynamically ontological
knowledge to new information can be considered as the birth of text technology as a
scientific discipline.
    It is not very astonishing that the triumphal elevation of hypertexts – powered by
the success of the world-wide-web – as a data structure, did play an important role in
making text technology a widely recognized scientific subject. Text technology as a
scientific discipline has a rather short history, although its origins have their roots in
classical academic research traditions like (computational) linguistics, computer 
science,artificial intelligence, literary sciences, and text sciences. Despite its recent emer-
genceas a coherent body of research text technology can easily be distinguished from 
theseneighboring areas and can claim to have become an autonomous discipline of its 
ownright. Text technology differs from classical computational linguistics and natural
language processing in focusing on text as a means in itself, not on text as a container
for language expressions (sentences) or text as a representation (or coding mechanism)
of utterances. Moreover, text technology considers structures and layouts of texts
contrary to literary history or classical linguistics and does not concentrate on finding
generative principles for the question of what constitutes a text. Last but not least, it
uses and develops markup standards (like XML, RDF, OWL etc.) and algorithms for
statistical and symbolic computations on texts, very similarly to computer science and
artificial intelligence in the area of the semantic web. But unlike computer science, text
technology directs its attention on texts instead of data structures in general, therefore
it attempts to structure data rather differently in comparison to, for example, the
semantic web tradition, and combines ideas from structure transformations that are
not at the center of interest of the computer science community.
    As a research area that is crucially located between different disciplines, text technolo-
gy is a strongly interdisciplinary research attempt. It combines research methodologies
from the already mentioned disciplines like linguistics, computational linguistics, ar-
tificial intelligence, and computer science. Whereas the domain of interest remains in-
the realm of the humanities, namely texts of all sorts and types, from a methodological
Editorial
Editorial
Uwe Mönnich, Kai-Uwe Kühnberger
iv
point of view text technology is primarily concerned with formal sciences. With the
emergence of new types of text (mainly based on the success of the world-wide-web),
that are no longer linearly structured, but contain hyperlinks and multi-media data, new 
interaction paradigms and usability aspects are provoked and generate new needs for
finding, retrieving, and editing information. Furthermore new techniques for archiving
multi-modal linguistic knowledge need to be developed and implemented that have firm-
ly established the importance of the interdisciplinary research endeavor text technology.
    In comparison to text technology, the term ontology as it is used in technical disciplines 
has a rather different history. Picking up a term that has a history of more than two
thousand years in philosophy, researchers in artificial intelligence introduced ontolo-
gies into their discipline as a means to represent conceptual background knowledge in
expert systems (Brachman and Schmolze, 1985). During the further development it
turned out, that many applications, most recently web applications, would strongly
benefit from a sound basis on which semantic information can be coded (Daconta
et al., 2003). It is a rather natural idea to integrate ontological knowledge into current
text technological applications. The result is an enrichment of structural information:
for example, taking annotation graphs as structural representation formalisms into
account, adding ontological knowledge to annotation graphs enlarges the structural
representation of text data by semantic knowledge.
    The present volume is the first part of a double volume about “Ontologies in Text
Technology” covering the theoretical basis of the topic. It contains a representative
sample of cutting-edge work in the foundations of combining text technology and
ontologies for state-of-the-art techniques of processing texts in language technology.
Volume II entitled “Applications of Ontologies in Text Technology” will be published in
January 2008 and will contain more applied work in the area of anaphora resolution,
discourse parsing, and extracting synonymy relations and lexico-semantic classes from
text.
    The origins of this double volume go back to the workshop “Adaptive Ontologies on
Syntactic Structures” held in conjunction with the 28th Annual Meeting of the German
Association of Linguistics (DGfS) at the University of Bielefeld in February 2006. As
a follow-up workshop, the editors organized an international workshop in Osnabrück
in September 2006 entitled “Ontologies in Text Technology – Approaches to Extract
Semantic Knowledge from Syntactic Information”. The proceedings of this workshop
contain six page papers of the participants and were published in the PICS series
(Publications of the Institute of Cognitive Science). Due to the fact that with Guus
Schreiber and Klaus Schulz two distinguished keynote speakers, gave inspiring talks
in this workshop, the workshop attracted many internationally well-known researchers
working in text and language technology. Because of the great success of this workshop 
the idea was conceived to provide a possibility to present the results of this workshop
to a broader audience. It was decided that the participants of the workshop should be
invited to submit full and extended versions of their papers for a journal publication.
After a thorough further reviewing process and a revision of the accepted full articles, 
the result is the present double volume of the GLDV-Journal for Computational Linguistics 
and Language Technology.
v
2 The Research Unit 437 Text Technological Information Modeling
During the last six years the development of text technology in Germany was strongly
influenced by the research unit 437 “Text Technological Information Modeling” funded
by the German Research Foundation (DFG). This research unit is an interdisciplinary
research endeavor carried by the Universities of Bielefeld, Gießen, Dortmund, Tübingen,
and Osnabrück. Starting in the year 2001, this group constitutes the largest collaborative
research project devoted to text technological issues and has provided the basis for text
technological research in Germany. Currently this research unit is in its final funding
year. In order to get a better impression of the overall project, a concise overview of
the involved sub-projects of the second phase of this research unit is given:
    • Secondary Structuring of Information and Comparative Analysis of Discourse.
       Principal Investigator: Dieter Metzing.
 
    • Induction of Document Grammars for the Representation of Logical Hypertextual
       Document Structures.
       Principal Investigator: Alexander Mehler.
    • Text-Grammatical Foundations for the (Semi-)Automated Text-to-Hypertext Con
       version.
       Principal Investigator: Angelika Storrer.
    • Generic Document Structures in Linearly Organized Texts: Text Parsing Using
       Domain Ontologies and Text Structure Ontologies.
       Principal Investigator: Henning Lobin.
    • Adaptive Ontologies on Extreme Markup Structures.
       Principal Investigators: Uwe Mönnich, Kai-Uwe Kühnberger.
    Although the research unit tries to cover all aspects of current text technological
activities, it is easily possible to identify certain core aspects that play a central
role in all sub-projects. Examples for such vertical topics of the whole research unit
are ontologies, annotations, markup standards, and processing aspects of texts. All
these topics play an important role in all participating projects. Some aspects of
these vertical topics of the research unit are also represented in this double volume
of the GLDV-Journal. The present volume focuses on the foundations of theories
for developing, characterizing, coding, learning, and adapting ontological background
knowledge as a crucial challenge for the semantic annotation of text documents. Some
of the sub-projects of the collaborative research unit mentioned above are represented
in this volume. Others will document aspects of their work in Volume II “Applications
of Ontologies in Text Technology”. We think that we can provide by this not only a 
representative documentation of text technology in general, but also a representative
collection illustrating the research unit 437 in particular.
vi
3 The Structure of Volume I
This first volume “Foundations of Ontologies and Text Technology” contains articles
concerned with the methodological basis of using ontologies in text technology. Two
aspects need to be distinguished in this context: the syntactic aspect attempts to focus
on the underlying languages and data structures used for coding technologically rele-
vant information, as for example, markup standards like XML and annotation graphs 
as a means to code linguistic information. Complementary to the syntactic level, the
semantic aspect deals with properties of ontological knowledge for text technological
applications. Both topics include aspects of learning and adaptation: learning and
adaptation of ontologies is a research field that is of great importance for the future,
because hand-coded ontologies are tedious, time-consuming, and expensive to create
(Perez and Mancho, 2003). But also on the syntactic side, there is the need for the
development of learning mechanisms: learning text types based on structural informa-
tion only, without any information about their content, turns out to be possible in many
cases. In the following, we will summarize major aspects of the articles included in this
volume.
    Lexical-semantic networks like the well-known WordNet (Fellbaum, 1998) together
with its versions in other languages like RussNet or GermaNet are not only a de
facto standard for several applications in text technology, but can also be seen as
prototypical examples where ontological knowledge can successfully be applied in text
technology. In their article “Domain Ontologies and Wordnets in OWL: Modelling
Options”, Harald Lüngen and Angelika Storrer question the common conversion stan-
dard to interpret synsets and lexical units of WordNet as OWL individuals. The article
provides arguments for a different conceptual view, namely that synsets and lexical
units need to be interpreted as concepts instead. Technically this results in a different
modeling of codingWordNet ontologies in the OWL format. The authors base their claim 
on an evaluation of OWL representation models for WordNet variants like GermaNet
combined with TermNet.
    The second article of this volume “Automatic Ontology Extension: Resolving Incon-
sistencies” by Ekaterina Ovchinnikova and Kai-Uwe Kühnberger continues the discussion
of ontologies by focusing on learning and adaptation aspects of ontologies against the
background of new input. The work follows the tradition to represent ontological
knowledge using description logic, therefore it considers ontology design from a logi-
cal perspective (Baader et al., 2003). Due to the fact that automatically generated and
automatically updated ontologies face the problem of becoming inconsistent, the paper
provides an automatic procedure for resolving occurring inconsistencies in ontology
design. Potential inconsistencies in ontology design are restricted to logical ones, in
particular, the overgeneralization of concepts and polysemy problems are discussed in 
detail. The authors propose an algorithmic solution for an automatic resolution based
on the minimal non-conflicting substitute.
    Related to the question of how to consistently extend ontologies by dynamic updates
is the question of how the population of ontologies with existing data sources can be
achieved. The article “Integration Languages for Data-Driven Approaches to Ontology
vii
Population and Maintenance” by Eduardo Torres Schumann, Uwe Mönnich, and Klaus
Schulz proposes a new integration language that is capable of generating new entries
in large-scale ontologies based on structured data. By an intelligent user interface
an efficient way to supervise the population of the ontology can be provided. The
authors embed their work into a system that is able to encode large amounts of data
like encyclopedic and common purpose knowledge: this large-scale knowledge base is
called EFGT net (Schulz and Weigel, 2003), a precisely defined framework designed for
various NLP applications.
    Text technology is concerned with different types of text. Not only that different
types of text have different content, often they differ significantly on a structural level
as well. Concerning webpages one can distinguish, for example, homepages of scien-
tists, blogs, or online stores from each other by structural features (Lindemann and Littig,
2006). The article “Structural Classifiers of Text Types: Towards a Novel Model of Text
Representation” by Alexander Mehler, Peter Geibel, and Olga Pustylnikov discusses
possibilities to learn text types solely on the basis of structural information without
having any content information. The authors show that the document object model
(DOM) can be used, in order to code structural information of texts. The authors
propose different learning mechanisms for achieving this task like quantitative structure 
analysis (QSA) and several variants of tree kernels. The article adds also an evaluation
of these learning algorithms based on a large newspaper corpus.
    Graph structures play an important role in coding linguistic and textual information.
Prominent examples are annotation graphs, which are used to represent multi-layered
information about language, like phonological, grammatical, semantical, and pragmati-
cal information, as well as non-linguistic information (gestures or cultural background).
On the other hand, tree structures can be used in order to analyze text types. The
article “Towards a Logical Description of Trees in Annotation Graphs” by Jens Michaelis
and Uwe Mönnich focuses on logical descriptions of annotation graphs, one of the
major data resources for text technological applications. The authors present results for
characterizing a large class of annotation trees, namely, single time line, multiple tiers
(STMT) models, which constitute a subclass of annotation graphs in the sense of Bird
and Liberman (2001), and from which multi-rooted trees can be constructed. Besides
other technical results, the article provides also a spelled-out algorithm for tree-like
graph transduction from a given STMT model into a multi-rooted tree. The result
is a uniform and mathematically rigorous format for the syntactic representation of
annotation graphs. Taking into account that multi-rooted trees and Bird-Liberman
annotation graphs play a prominent role in archiving and coding texts, this work can
be considered as a theoretical basis for annotation tasks in general.
4 Acknowledgments
This volume would not have been possible without the help of many people. Most
importantly the guest editors want to thank the editors-in-chief of the GLDV-Journal for
Computational Linguistics and Language Technology, Alexander Mehler and Christian
Wolff. Their encouragement and support in all phases of the emergence of this double
viii
volume has been irreplaceable in completing it. Furthermore we want to thank the
German Research Foundation for financial support of the research unit 437 “Text
Technological Information Modeling” and particularly the speaker of this research unit,
Dieter Metzing.
    Last but not least, the editors want to thank the program committee for their
careful evaluations of the submitted papers. The quality of this volume is also a direct
consequence of the work these reviewers invested. The program committee consisted
of the following researchers (in alphabetical order): Irene Cramer, Thierry Declerck,
Stefan Evert, Pascal Hitzler, Wolfgang Höppner, Helmar Gust, Marcus Kracht, Edda
Leopold, Alessandro Moschitti, Larry Moss, Rainer Osswald, Olga Pustylnikov, Georg
Rehm, Hans-Christian Schmitz, Bernhard Schröder, Uta Seewald-Heeg, Manfred Stede,
Markus Stuptner, Frank Teuteberg, Yannick Versley, Johanna Völker, Armin Wegner,
and Christian Wolff.
References
Baader, F., Calvanese, D., McGuinness, D. L., Nardi, D., and Patel-Schneider, P. F., editors
    (2003). The Description Logic Handbook: Theory, Implementation, and Applications.
    Cambridge University Press, New York.
Bird, S. and Liberman, M. (2001). A Formal Framework for Linguistic Annotation. Speech
    Communication, (33):23–60.
Brachman, R. and Schmolze, J. (1985). An Overview of the KL-ONE Knowledge Representation
    System. Cognitive Science, 9:171–216.
Daconta, M., Obrst, L., and Smith, K. (2003). The Semantic Web: A Guide to the Future of
    XML, Web Services, and Knowledge Management. John Wiley and Sons.
Fellbaum, C., editor (1998). WordNet. An Electronic Lexical Database. MIT Press.
Lindemann, C. and Littig, L. (2006). Coarse-Grained Classification of Web Sites by their
    Structural Properties. In Proc. of WIDM’06, pages 35–42.
Perez, G. A. and Mancho, M. D. (2003). A Survey of Ontology Learning Methods and
    Techniques. OntoWeb Delieverable 1.5.
Schulz, K. and Weigel, F. (2003). Systematics and Architectures for a Resource Representing
    Knowledge about Named Entities. In Proceedings Workshop on Principles and Practice     
    of Semantic Web Reasoning, pages 189–207.



Editorial

Geeb, Franziskus

Chatbots in der praktischen Fachlexikographie / Terminologie

Lu, Xiaofei

A Hybrid Model for Chinese Word Segmentation

Torres Schumann, Eduardo

Schulz, Klaus U.

Integration Languages for Data-Driven Approaches to Ontology Population and Maintenance

Not a member yet

Experiments on Lexical Chaining for German Corpora: Annotation, Extraction, and Application

Manually vs. Automatically Labelled Data in Discourse Relation Classification: Effects of Example and Feature Selection

Analysis of E-Discussions Using Classifier Induced Semantic Spaces

Editorial

Chatbots in der praktischen Fachlexikographie / Terminologie

A Hybrid Model for Chinese Word Segmentation

Integration Languages for Data-Driven Approaches to Ontology Population and Maintenance

Domain ontologies and wordnets in OWL: Modelling options

Editorial

Chatbots: Are they Really Useful?

231

253