=Paper=
{{Paper
|id=Vol-1341/preface-1
|storemode=property
|title=None
|pdfUrl=https://ceur-ws.org/Vol-1341/preface.pdf
|volume=Vol-1341
|dblpUrl=https://dblp.org/rec/conf/argnlp/Green14
}}
==None==
https://ceur-ws.org/Vol-1341/preface.pdf
Preface
Elena Cabrio1 , Serena Villata1 , and Adam Wyner2
1
INRIA Sophia Antipolis - Mediterranee
2
Department of Computing Science, University of Aberdeen
Large amounts of text are added to the Web On the one hand, text analysis is a promising
daily from social media, web-based commerce, approach to identify and extract arguments from
scientific papers, eGovernment consultations, and text, receiving attention from the natural language
so on. Such texts are used to make decisions in the processing community. For example, there are ap-
sense that people read the texts, carry out some in- proaches on argumentation mining of legal doc-
formal analysis, and then (in the best case) make a uments, on-line debates, product reviews, news-
decision: for example, a consumer might read the paper articles, court cases, scientific articles, and
comments on an Amazon website about a camera, other areas. On the other hand, computational
then decide which camera to buy; a voter might models of argumentation have made substantial
read various political platforms, then vote. An an- progress in providing abstract, formal models to
alyst or consumer of such corpora of text is con- represent and reason over complex argumentation
fronted by several problems. The information in graphs. The literature covers alternative models,
the corpora is distributed across texts and unstruc- a range of semantics, complexity, and formal dia-
tured, i.e. is not formally represented or machine logues.
readable. In addition, the argument structure - jus- Yet, there needs to be progress not only within
tifications for a claim and criticisms - might be im- each domain, but in bridging between textual and
plicit or explicit within some document, but harder abstract representations of argument so as to en-
to discern across documents. As well, the sheer able reasoning from source text. To make progress
volume of information overwhelms users. Given and realize automated argumentation, a range of
all these problems, extracting and reasoning with interdisciplinary approaches, skills, and collab-
arguments from textual corpora on the web is cur- orations are required, covering natural language
rently infeasible. processing technology, linguistic theories of syn-
tax, semantics, pragmatics and discourse, domain
To address these problems, we need to develop
knowledge such as law and science, computer sci-
tools to aggregate, synthesize, structure, summa-
ence techniques in artificial intelligence, argumen-
rize, and reason about arguments in texts. Such
tation theory, and computational models of argu-
tools would enable users to search for particular
mentation.
topics and their justifications, trace through the ar-
To begin to address these issues, we orga-
gument (justifications for justifications and so on),
nized the seminar Frontiers and Connections be-
as well as to systematically and formally reason
tween Argumentation Theory and Natural Lan-
about the graph of arguments. By doing so, a
guage Processing, which was held July 21-25,
user would have a better, more systematic basis
2014 at the University Residential Center, Berti-
for making a decision. Clearly, deep, manual anal-
noro, Italy. It was attended by 29 participants.
ysis of texts is time-consuming, knowledge inten-
The papers in this CEUR Workshop Proceedings
sive, and thus unscalable. Thus, to acquire, gener-
are the outcome of the workshop, ranging over a
ate, and transmit the arguments, we need scalable
host of topics, empirical approaches, and theoreti-
machine-based or machine-supported approaches
cal frameworks.
to extract and reason with arguments. The ap-
plication of tools to mine and process arguments
would be very broad and deep given the variety of
contexts where arguments appear and the purposes
they are put to.