linguistlist.org>
Editor for this issue: Lydia Grebenyova <lydialinguistlist.org>
lscp.ehess.fr>Workshop on Early Phonological Acquisition October 6-8, 2001 Carry-le-Rouet (Marseilles), France Invited speakers: Michael Brent, Washington University Emmanuel Dupoux, Laboratoire de Sciences Cognitives et Psycholinguistique Mark Hale, Concordia University Bruce Hayes, UCLA Peter Jusczyk, Johns Hopkins University John Kingston, University of Massachusetts Jim Morgan, Brown University Marina Nespor, University of Ferrara Joe Pater, University of Massachusetts Janet Pierrehumbert, Northwestern University Douglas Pulleyblank, University of British Columbia Jim Scobbie, Queen Margaret University College Dan Swingley, Max Planck Institute for Psycholinguistics Bruce Tesar, Rutgers University Call for Posters Phonological acquisition has been a research topic in both theoretical linguistics and experimental psychology. Much progress has been made on either side, and researchers from these disciplines can undoubtedly benefit from each other's findings. On the one hand, phonologists typically gather early production data to describe the various stages in the acquisition process. They have provided much evidence that children's productions evolve from universally unmarked structures to the marked structures present in their ambient language. However, experimental research with infants shows that phonological acquisition begins right after birth and develops considerably during the first year of life, i.e. before the first words are uttered. Although the development in production resembles the one in perception in certain respects, there are also many divergences. A valid theory of phonological acquisition should contain a description of the initial state and hence take into consideration the earliest perception data. Phonologists have also proposed models of how phonological systems might be acquired by young children. These models contain concrete algorithms by which children could derive parts of the phonological grammar of their language, but they are based upon the assumption that children have access to individual word forms, often coupled with their meaning. This latter assumption is highly unrealistic for the same reason that much of the native phonology appears to be acquired before word segmentation is in place and before lexical acquisition has started altogether. Experimental psychologists, on the other hand, have gathered data concerning the perceptual capacities of infants from birth to 18 months of life. They have thus provided evidence that during this period, infants build a phonological representation of their ambient language, and, consequently, come to perceive speech sounds in a language-specific fashion, much the same way as adults do. However, these data pertain to a relatively small number of languages, and hence to only a small amount of variation that is attested in the phonological systems of human language. Moreover, although experimental psychologists have started to propose models concerning the acquisition of basic phonological parameters such as the segmental inventory and surface syllable structure, they typically are not concerned with the acquisition of phonological rules. Besides phonology and experimental psychology, phonetics and diachrony are equally relevant to the study of early phonological acquisition. As to phonetics, it is often assumed that infants build a phonological representation of their native language by passing through a stage of a universal phonetic representation. It is thus important to carry out detailed phonetic analyses of the incoming speech signal. As to diachrony, under the common assumption that language change is induced by children acquiring their native language, diachronic data may provide insight into the acquisition process. Abstracts are being solicited for POSTER presentations. Please send 2 copies of an anonymous 2-page abstract plus one copy with your name, affiliation, and e-mail address to: Sharon Peperkamp Workshop on Early Phonological Acquisition Laboratoire de Sciences Cognitives et Psycholinguistique EHESS - CNRS 54, Bd. Raspail 75270 Paris cedex 7 France or to: sharonMail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue
epfl.ch>Call for Papers for a Special Issue of the Journal Natural Language Engineering on Robust Methods in Analysis of Natural Language Data Special Issue guest editors: Afzal Ballim Vincenzo Pallotta Department of Computer Science Swiss Federal Institute of Technology - Lausanne. The automated analysis of natural language data has become a central issue in the design of Intelligent Information Systems. The term "natural language" is intended to cover all the possible modalities of human communication and it is not restricted to written or spoken language. Processing unrestricted natural language is still considered as an AI-hard task. However various analysis techniques have been proposed in order to address specific aspects of natural language. In particular, recent interest has been on providing approximate analysis techniques, assuming that perfect analysis is not possible, but that partial results are still very useful. There are many ways in which the topic of robustness may be tackled: as a competency problem, as a problem of achieving interesting partial results, as a shallow analysis method, etc. What they have in common is that no simple combination of "complete" analysis modules for different linguistic levels in a chain can give a robust system, because they cannot adequately account for real-world data. Rather, robustness must be considered as a system-wide concern. We consider of central interest improving and integrating various processing methods with respect to the following issues: * Extending coverage * Improving efficiency * Disambiguation ability * Approximate processing * Enhancement of underlying theories Robustness may be seen as an engineering "add-on" - something that we add to a system to take account of the inability of our theories to cope with real-world data - or as a basic element of our theories - our theories are developed to admit that understanding of the domain can be incomplete. Both approaches may be valid under certain circumstances. The main goal of this Special Issue of the Natural Language Engineering journal is devoted to advances in fields like artificial intelligence, computational linguistics, human-computer interaction, cognitive sciences who are faced with the problem of feasible and reliable NLP systems implementation. Theoretical aspects of robustness in NLP are welcome as well as engineering and industrial experiences. We invite papers on all topics related to Robustness in Natural Language Processing and Understanding, including, but not limited to: Text Analysis Knowledge and Information Extraction Spoken Dialogue Systems Multimodal Human-Computer interfaces Natural Language Architectures Distributed NLP NLP and Soft Computing Semantics Underspecification Multimedia Document Analysis Robust Parsing Incremental Parsing Discourse analysis Summarization Complexity of linguistic analysis Hybrid methods in computational linguistics Text Mining Corpus linguistics Indexing and Information Retrieval SUBMISSION PROCEDURE: We are expecting full papers to describe original, previously unpublished research, be written in English, and not be simultaneously submitted for publication elsewhere (previous publication of partial results at workshops with informal proceedings is allowed). Papers should be formatted according to the NLE journal instructions and should be between 15 and 25 pages long. The preferred formatting system is LaTex, which can be used for direct typesetting, and a style file is available through anonymous ftp from the following address: ftp.cup.cam.ac.uk/pub/texarchive/journals/latex/nle-sty/. In case of difficulty there is a helpline available on e-mail: texlineMail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue