Multidisciplinary Perspectives on Signalling Text Organisation: Introduction to the Special Issue.

Le lecteur trouvera une version francaise de cette introduction ci-dessous. Texts are organised wholes. Understanding a text entails constructing a representation of its organisation. Several research domains, with different assumptions and objectives, have taken an interest in the devices which seem to help readers in this process. As a consequence, research concerned with the signalling of text organisation is far from constituting a u...


Multidisciplinary Perspectives on Signalling Text Organisation: Introduction to the Special Issue
Perspectives multidisciplinaires sur la signalisation de l'organisation textuelle : introduction au numéro thématique Texts are organised wholes.Understanding a text entails constructing a representation of its organisation.Several research domains, with diff erent assumptions and objectives, have taken an interest in the devices which seem to help readers in this process.As a consequence, research concerned with the signalling of text organisation is far  om constituting a unifi ed fi eld.The notion of signal itself may be associated with diff erent key concepts according to discipline and theoretical underpinning: document structure, discourse organisation, layout structure, text architecture, etc.As far as function is concerned, signals may be seen as discourse construction devices, as metadiscourse, as reading or processing instructions, as traces of the writer's cognitive processes, as cues revealing the author's intentions… This special issue, which follows on  om the MAD 2010 workshop in Moissac (http://w3.workshop-mad2010.univ-tlse2.), is a fi rst step towards the development of multidisciplinary approaches: by bringing together perspectives  om diff erent disciplines interested in the signalling of text organisation -linguists, computational linguists, psycholinguists, educational/cognitive psychologists, document designers-we aim to contribute to mutual understanding Lydia-Mai Ho-Dac, Julie Lemarié, Marie-Paule Péry-Woodley et Marianne Vergez-Couret and cross-fertilization.We urge readers to resist the temptation to stay smugly within their own fi eld, and invite them to fi nd out what research in other disciplines brings up in relation to text organisation signals.

2
Though this issue is modestly multidisciplinary in the sense that it mostly brings together research  om diff erent fi elds, it is also more ambitiously so in the fi rst two papers, which present approaches associating several disciplines: cognitive psychology and linguistics for the fi rst, information design (and architecture…) and linguistics for the second.

3
In their paper on "Understanding How Headings Infl uence Text Processing", Julie Lemarié , Robert F. Lorch and Marie-Paule Péry-Woodley show that the lack of integration of linguistic and psychological research concerning signals in general and headings in particular leads to blind spots; they present a general theoretical  amework, SARA, that provides a linguistically-based approach to analyse signalling devices and generates psychological predictions on their eff ects on text processing by the reader.They illustrate how this integrative  amework may be used to investigate how headings infl uence text processing.

4
In their contribution, "Towards a Pattern Language Approach to Document Description", Robert Waller , Judy Delin and Martin Thomas draw connections between information design, discourse research and corpus linguistics.They introduce the concept of pattern languages and pattern libraries, originating in architecture and widely used in so ware engineering and interaction design, as the basis for a functional description of the graphic structure of discourse.In information design, patterns are described as clusters of graphical and linguistic features that recur as regular solutions to particular design problems in particular contexts.The authors argue for a corpus approach with systematic multiple tagging of patterns in diff erent genres and describe their own work in this direction.Among the features to be tagged, they stress the linguistic component of patterns, which they see as interacting closely with the graphical components.

5
This view of signals in terms of recurrent clusters of graphical and linguistic features is also defended by Lydia-Mai Ho-Dac , Cécile Fabre , Marie-Paule Péry-Woodley , Josette Rebeyrolle and Ludovic Tanguy in "An Empirical Approach to the Signalling of Enumerative Structures".They report on a large-scale corpus-based study of enumerative structures made possible by the recent availability of a corpus hand-annotated with discourse structures and cues.Taking the annotated structures as their starting point, and through the use of statistical methods, they propose an analysis of signalling as complex confi gurations of cues.On this empirical basis, they question the notion of textual metadiscourse as separate  om propositional content, stressing instead the interplay of the three Hallidayan metafunctions: ideational, textual and interpersonal.

6
The next paper, "Genre Realised in Theme: The Case of News Reports and Commentaries", by Julia Lavid , Jorge Arús and Lara Moratón , is also rooted in the systemic functional perspective.The authors propose a comparative analysis of the thematic features appearing in "thematic heads" i.e. elements occurring in the preverbal zone at the beginning of a sentence.Their observations are based on a systematic human annotation of the semantic properties of these thematic heads in news reports vs. news commentaries.Their results show that each genre selects diff erent "thematic participants" (e.g. the Sayer in reports) and uses diff erent "thematic realisations" (e.g.concrete nouns in reports vs. abstract nouns in commentaries).Such semantic features are presented by the authors as features which, in a broad sense, "signal" these genres.
The next two papers are contributions to a bottom-up approach to text organisation using the  amework developed within Segmented Discourse Representation Theory (SDRT).Bottom-up approaches to discourse structure attempt to give a representation of the structure through discourse relations between elementary segments.Relations may be signalled in text by means of so called discourse markers.
The article by Charlotte Roze , Laurence Danlos and Philippe Muller , "LEXCONN: A French Lexicon of Discourse Connectives", contributes to the theme of the special issue by presenting a lexicon of 328 French discourse connectives or discourse markers along with their syntactic categories and the discourse relations they may express within the SDRT set of relations.The methodology consists fi rstly in applying various criteria, including syntactic, semantic and discourse criteria, to the identifi cation of discourse connectives, and secondly in observing their context to associate each of them with one or several discourse relations it may express.Finally, the authors reveal some problem cases when no discourse relation in SDRT can describe the discourse contribution of the connective.

9
In their article, "Signalling Elaboration : Combining French Gerund Clauses with Lexical Cohesion Cues", Marianne Vergez-Couret and Clémentine Adam tackle the question of the signalling of text organisation in cases where there are no obvious discourse markers, as it is usually the case for the Elaboration relation.The authors focus on this relation and on its automated detection.In the absence of specialised discourse markers, they investigate the presence of lexical cohesion links involved in the interpretation of Elaboration , and how to bring them to light using a lexical resource based on distributional similarity in combination with an ambiguous marker of Elaboration , the French gerund construction.

10
The last paper describes a study which associates a linguistic perspective on signalling with a psycholinguistic experimental approach: "Children's Use of Comparative Text Signals: The Relationship between Age and Comprehension Ability" by Bonnie J.F. Meyer, Melissa N. Ray and Wendy Middlemiss .The authors off er a very original developmental perspective on the topic: they investigate how children and young adults acquire and use comparative signalling devices.Their results show diff erent patterns of acquisition as a function of young readers' comprehension skills.