[LFG] Collaborative book on MWE Representation and Parsing - Call for Chapters

Yannick Parmentier yannick.parmentier at univ-orleans.fr
Fri Jan 22 16:19:04 UTC 2016


***********************************************************
Collaborative book on MWE Representation and Parsing
Call for Chapter proposals
***********************************************************

Provisional title
*****************
"Mutliword Expressions: Representation and Parsing"

Coordinators/editors
********************
Yannick Parmentier (Orléans) & Jakub Waszczuk (Blois)
parseme-wg2-book at listes.univ-orleans.fr

Content
*******
While Multiword expressions (MWEs) have been attracting attention for
a long time because of their idiosyncratic properties going beyond
word boundaries, they remain a challenge for both linguistic theories
and natural language applications.

A contrastive study on the representation of MWEs within various
languages and frameworks has been carried out recently in the context
of the Working Group 1 of the PARSEME Action. This study will appear
as a special issue of the "Empirically Oriented Theoretical Morphology
and Syntax" series at Language Science Press
(http://langsci-press.org/).

The planned volume will appear within the same series and aims at
providing the NLP research community with a sharing of experience from
MWE experts who have been experimenting with the design of linguistic
resources allowing for MWE parsing.

Possible topics for contributions include:

A) Implemented grammars:

    How to encode the syntactic and/or semantic properties of MWEs
    within formal grammars ?
    How to account for various degrees of flexibility of MWEs ?
    How to represent MWEs which exhibit both syntactic variability and
    semantic non-compositionality ? Is there a connection between
    semantic and syntactic flexibility? 
    How do linguistic theories and / or grammatical frameworks
    influence MWE representations within implemented grammars ?
    How to evaluate the coverage of implemented grammars with respect
    to MWEs ? 
    Can we define abstract models (e.g. metagrammars) of MWEs'
    properties that would capture linguistic richness of MWEs
    independently of particular grammatical frameworks ? If so, how
    can such models help to reduce the cost of MWE-aware resource
    development ? Can they adapt to different languages ?
    How to represent MWEs within strongly lexicalized formalisms ?

B) Challenging MWE examples:

    What types of "transformations" (such as passivization, internal
    modification, pronominalization, participation to long distance
    dependency phenomena, control and binding phenomena) can be
    applied to MWEs? Can those transformations be easily captured
    within (lexicalized or not) formal grammar ?
    How to account for morphological flexibility at the lexical
    level ? at the grammatical level (e.g. agreement) ? 
    How to tackle language-specific challenges in encoding MWEs in
    formal grammars ? 

C) Parsing MWE:

    What statistical and / or symbolic frameworks support MWE
    parsing ?
    How to enrich parsers with information about MWEs within
    treebanks ?
    How to reconcile linguistic precision and computational efficiency
    when parsing MWEs ? 
    How do syntactic parsing and semantic representation relate in the
    context of MWEs ?

D) Cross-linguistic comparison of MWE representations and analyses

    How to reuse material from  grammars for other languages or
    frameworks (cross-language and cross-framework grammar design) ?
    Can MWE parsing benefit from representations and annotations of
    MWEs in other languages ? Other frameworks ? 

Submission
**********
We invite the submission of abstracts of chapters (2 pages) by March
31, 2016.

The outline should clearly express the topic, indicate whether
electronic resources are available and give the status of the work
(completed or in progress).

Selected abstracts will be presented during the automn meeting of the
Working Group “Parsing Techniques for MWE” of the COST Action IC 1207
PARSEME (Parsing Multiword Expressions) in Dubrovnik, Croatia (26-27
September, 2016).

Reimbursement for participation at this meeting might be available for
authors of selected abstracts according to COST regulations. Please
contact the editors for details.

Submissions should mention “MWE Volume” in the subject line and be
sent to parseme-wg2-book at listes.univ-orleans.fr

Preliminary schedule
********************

January 2016: Call for abstracts sent out

March 31, 2016: Deadline for abstract submissions

Mid June 2016: Notification of pre-selection

September 26-27, 2016 (Parseme meeting in Dubrovnik): Presentation of
the contribution proposals.

October 31, 2016: deadline for first versions of the chapters.

December 14, 2016: comments, notifications of acceptance sent out.

January 30, 2017: deadline for final versions.



More information about the LFG mailing list