Natural Negation Modeling and Processing
Outline
Objective
The main objective of the NEXING project is to contribute for improving
the automated mapping between (orthographic) form and (linguistic) meaning,
on the one hand, and between (linguistic) meaning and knowledge (representation),
on the other hand, in what concerns natural language negation.
Motivation
The processing of natural negation offers some of its most notable
puzzles at the interfaces of Semantics with Syntax and Pragmatics. Note
that negation may be expressed only once although several negative expressions
are involved, as in the so-called negative concord constructions such as
Port.: "Nunca ninguém lê as propostas com atenção" / lit.: "Never
nobody reads the proposals with attention" / "Nobody ever reads
proposals attentively". On the other hand, negation can be expressed
even though no overt negative expression occurs, as in the so-called counterfactual
conditionals such as "If Clara had arrived, Rui would be happier"
- a construction without overt negative expressions that nevertheless informs,
among other things, that Clara did not come. Also the interface between
language and cognition offers significant challenges in this respect since
negation is at the heart both of several quantification-based linguistic
phenomena and of the operations of formal thought (attained by 11-14 year
old children), although the implications of these deep commonalties between
language and cognition have been waiting to be fully clarified.
Approach
NEXING is a multi-disciplinary project fostering the convergence of
methods, results and expertise from Informatics, Applied Logic, Cognitive
Psychology and Formal Linguistics in the areas of the lexicon, syntax,
semantics, pragmatics and reasoning. The workplan is divided into several
tasks covering Negative Concord and Polarity, Counterfactuals, Deductive
Inference, Computational Modeling, and Language Engineering Resources Development
(Corpus and Linguistic Database).
Participants
Research centers
Given its intrinsic multidisciplinary program, the NEXING project is
a common enterprise of three academic research centers affiliated with
three Universities: University
of Algarve (UALG), University
of Coimbra (UC) and University
of Lisbon (UL).
The participating centers have been conducting research on artificial
intelligence, cognitive science and natural language science and technology:
The IPCDVS-Institute
for Cognitive Psychology is a research institute of the Faculty
of Psychology and Education Sciences of Coimbra, UC; the LabMAC-Laboratory
of Computational Models and Architectures is hosted by the Department
of Informatics of the Faculty
of Sciences of Lisbon, UL; and the UCH-Faculty of Human Sciences is a division of UALG.
Team
António Horta
Branco (coord.), UL
Luís Miguel Gomes, UC
José Leitão, UC
Pedro Santos, UALG
João Silva, UL
Ana Paula Silveira, UC
Funding
The project is funded by the FCT-Foundation
for Science and Technology of the MCT-Portuguese
Ministery of Science and Technology under the contract FCT/SAPIENS99/34076/99.
The project life is planned to span over 24 months. Check here for a press
release of FCT about the project.
Results (June 2003)
Publications
-
Branco, António Horta and João Silva, Forth., "A Metric for the Efficency of Accurate Tagging Procedures", In Proceedings of RANLP 2003, International Conference on Recent Advances in Natural Language Processing, Bulgarian Academy of Sciences, Borovets, Bulgaria, 10-12 September, 2003.
-
Branco, António Horta and João Silva, Forth., "Morpho-syntactic Tagging without Training Corpus or Lexicon: How Far is it Possible to Get?".
In Proceedings of XVIII Encontro Anual da APL.
-
Santos, Pedro, Forth., "O que é dito: o caso condicional". In Proceedings of XVIII Encontro Anual da APL.
-
Branco, António Horta and Tiago Henriques, Forth., "Aspects of Verbal Conjugation and Lematization: Generalizations and Algorithms". In Proceedings of XVIII Encontro Anual da APL,
-
Santos, Pedro, Forth., "Duas Espécies de Condicionais?".
In Proceedings of Primeiro Encontro Nacional de Filosofia Analítica, May 2002, Coimbra.
-
Branco, António Horta, Forth., "Anaphora Dualities and the Semantics of Nominals".
In Proceedings of Primeiro Encontro Nacional de Filosofia Analítica, May 2002, Coimbra.
-
Santos, Pedro, 2003, "On two alleged (semantic) classes of non-counterfactual conditionals", Psychologica 32, pp. 171-183.
-
Branco, António Horta, 2003, "Nominals are doubly dual", In Workshop Notes of the First International Workshop on Current Research in the Semantics-Pragmatics Interface, Michigan State University, pp. 18-22.
-
Branco, António Horta and João Silva, 2003, "Contractions: breaking the tokenization-tagging circularity", In Mamede et al. (eds.) Computational Processing of the Portuguese Language, Springer, LNAI2721, pp.167-170.
-
Leitão, José, 2003, "Modal Logic in a production system model of syllogisitic reasoning", Psychologica 32, pp. 341-364.
-
Branco, António Horta, 2003, "Anaphor Resolution: Is the search optimization flawed?", Psychologica 32, pp. 77-89.
-
Branco, António Horta and João Silva, 2003, "Tokenization of Portuguese: resolving the hard cases", Technical Report TR-2003-4, Departament of Informatics, University of Lisbon.
-
Branco, António Horta and João Silva, 2002, " EtiFac: A Facilitating Tool for Manual Tagging".
In Proceedings of XVII Encontro Anual da APL, pp.81-90.
-
Branco, António Horta and Tiago Henriques, 2002, " Probabilistic PP Attachment: Hindle and Rooth's Procedure Applied to Portuguese". In Proceedings of XVII Encontro Anual da APL, pp.69-80.
-
Branco, António Horta, 2002, "Anaphoric Binding and Phase Quantification".
In Proceedings of Encontro Comemorativo do 25° Aniversário do Centro
de Linguística do Porto, 22-24 November 2001, CLUP, Porto, pp.59-68.
-
Branco, António Horta, José Leitão and João Silva, 2002, "Nexing Corpus: A Corpus of Verbal Protocols on Syllogistic Reasoning", in Proceedings of LREC2002, pp.397-403.
-
Branco, António Horta, 2002, "Binding Machines". Computational Linguistics, 28-1, MIT Press.
-
Branco, António Horta and Crysmann, Berthold, 2001, "Negative
Concord and the Distribution of Quantifiers", in Yves d'Hulst, Jan
Schroten and Johan Rooryck (eds.), Romance Languages and Linguistic Theory. John Benjamins.
-
Branco, António Horta, 2001, "Natural Negation and Quantification: semantic explorations into negative concord and other syntactic puzzles". In Questão, Universidade do Algarve. pp. 29-48
-
Leitão, José, 2001, "An ACT-R Model of
Syllogistic Reasoning". In Proceedings of the 2001 fourth International
Conference on Cognitive Modeling. July 26-28, George Mason University,
Fairfax, USA. Mahwah,NJ: Lawrence Erlbaum Associates.
-
Branco, António Horta, 2001, "Duality and
Anaphora". In Robert van Roy and Martin Stokhof (eds.), Procedings of the 13th Amsterdam Colloquium, 17-19
December 2001, ILLC, Amsterdam, pp.49-54.
Journal (special issue)
-
Leitão, José (ed.), 2003, Psychologica, 32 Special Issue on Language and Reasoning.
Presentations
-
Branco, António Horta, 2003, "Nominals are doubly dual", First International Workshop on Current Research in the Semantics-Pragmatics Interface, Michigan State University.
-
Branco, António Horta and João Silva, 2003, "Contractions: breaking the tokenization-tagging circularity", PROPOR2003, 6th Workshop on Computational Processing of the Portuguese Language - Written and Spoken, Universidade do Algarve, Faro, Portugal, June 26-27, 2003.
-
Branco, António Horta, 2003, "Anotação: Em direcção à eficiência máxima", Conversas d'Horal, Centro de Linguística da Universidade de Lisboa, 10 April, 2003.
-
Santos, Pedro, 2002, "O que é dito: o caso condicional". XVIII Encontro Anual da APL, Porto, 2-4 October 2002.
-
Branco, António Horta and Tiago Henriques, 2002, "Ambiguity Resolution in the Lemmatization of Complex Inflectional Systems", CLIN 2002, Thirteenth Meeting of Computational Linguistics in the Netherlands, November 29, 2002, University of Groningen, The Netherlands.
-
Branco, António Horta and João Silva, 2002, "Morpho-syntactic Tagging without Training Corpus or Lexicon: How Far is it Possible to Get?".
XVIII Encontro Anual da APL, Porto, October 2002.
-
Branco, António Horta and Tiago Henriques, 2002, "Aspects of Verbal Conjugation and Lematization: Generalizations and Algorithms", XVIII Encontro Anual da APL, Porto, October, 2002.
-
Branco, António Horta , 2002, "Nexing Corpus", talk at the Annual Workshop of the Typological
Database Project, OTS, Utrecht, September, 2002.
-
Santos, Pedro, 2002, "Conditionals and Contexts". Fourth European Congress for Analytic Philosophy, Lund, 14-18 June 2002.
-
Branco, António Horta, 2002, " Anaphora Dualities and the Semantics of Nominals". Primeiro Encontro Nacional de Filosofia Analítica, Coimbra, May 2002.
-
Branco, António Horta , 2002, "As Línguas Naturais na Nova Ordem Tecnológica", invited talk at the project Enciclopédia e Hipertexto, Faculty of Sciences, University of Lisbon, Lisbon, April 3, 2002.
-
Branco, António Horta and Silva, João, 2001, "Etifac: um facilitador
de etiquetagem morfo-sintáctica". Encontro Nacional da Associação
Portuguesa de Linguística, 2-4 October, 2001, Faculdade de Letras
da Universidade de Lisboa.
-
Santos, Pedro, 2001, "Conditionals: against the Apartheid view". First
Latin Meeting in Analytic Philosophy. 1 July 2001, Faculdade de Letras
da Universidade de Lisboa.
-
Branco, António Horta , 2001, "Nexing and more: some of the projects at
the University of Lisbon", invited talk at the Workshop of the Typological
Database Project, OTS, Utrecht, June 29-30, 2001.
-
Branco, António Horta , 2001, "Reference Processing and its Universal
Constraints", invited talk at Quid Novi 2001?, CLUP-Centro de Linguística
do Porto, Porto, June 2001.
-
See also the presentations in the Nexing meetings.
Software
-
LX-Suite
-
Full coverage, disambiguating POS tagger (LX-tagger)
-
Sentence chunker (LX-chunker)
-
Computational grammar fragment
-
Manual tagging assistant (Etifac/Crivo)
-
Tokenizer (LX-tokenizer)
-
Corpus developping assistant (NCT- Nexing Corpus Tool)
-
A computational model of human syllogistic
reasoning, instantiating a theory of syllogistic reasoning performance using ACT-R (SYRE).
Language Engineering Resources
-
Portuguese word list of closed categorial classes
-
Corpus (see a sample here; go to the Nexing corpus page)
-
Guidelines for transcription of spoken data
-
Collection of audio recorded protocols with syllogistic reasoning
Cooperation
Cooperation with related projects
-
Members of the project are participating in the international EC funded
project LTRC- Language Typology Resource Center, coordinator OTS, Utrecht.
Visiting scholars
-
Rui Pedro Chaves (CLUL, Lisbon)
-
Bruno Emmond (University of Montreal)
-
Cristina Quelhas (Instituto Superior de Psicologia Aplicada, Lisbon)
-
Amália Mendes (CLUL, Lisbon)
-
Peter Yule (University of London)
-
João Balsa (University of Lisbon)
-
João Branquinho (University of Lisbon)
-
Kai von Fintel (MIT)
-
Pieter Seuren (Max Planck Institute for Psycholinguistics)
-
David Matos (Instituto Superior Técnico, Lisbon)
Availability
-
For the current state of development of deliverables and their availability
contact the project coordinator.
Meetings
International seminars
-
2002: 18 May, Coimbra. See here for details.
-
2002: 2 Apr, Faro. See here for details.
-
2002: 20 Feb, Coimbra. See here for details.
Open meetings
-
Other working meetings involving every participant in the project, and open to
external participants:
2001: 8 Jan, Lisbon; 13 Feb, Lisbon; 16 Apr, Coimbra; 9 Jul, Lisbon; 25 Oct, Lisbon; 11 Dec, Lisbon; 2002: 18 Mar, Lisbon; 18 May, Coimbra; 30 October, Lisbon; 2003: 30 June.