Please use this identifier to cite or link to this item: doi:10.22028/D291-25146
Title: Prosody, focus, and focal structure : some remarks on methodology
Author(s): Batliner, Anton
Language: English
Year of Publication: 1994
OPUS Source: Saarbrücken, 1994
SWD key words: Künstliche Intelligenz
DDC notations: 004 Computer science, internet
Publikation type: Report
Abstract: Prosody falls between several established fields as e.g. phonetics, phonology, syntax, and dialogue structure. It is therefore prone to misconceptions: often, its relevancy is overestimated, and often, it is underestimated. The traditional method in linguistics in general and in phonology in particular is the construction and evaluation of sometimes rather complex examples based on the intuition of the linguist. This intuition is replaced by more or less naive and thus non-expert subjects and inferential statistics in experimental phonetics but the examples, i.e. the experimental material, are often rather complex as well. It is a truism that in both cases, conclusions are made on an "as if'; basis: as if a final proof had been found that the phenomenon A really exists regularily in the language B. In fact, it only can be proven that the phenomenon A sometimes can be detected in the production of some speakers of a variety of language B. This dilemma matters if prosody has to be put into practice, e.g. in automatic speech and language processing. In this field, large speech databases are already available for English and will be available for other languages as e.g. German in the near future. At least in the beginning, the problems that can - hopefully - be solved with the help of such databases might look trivial and thus not interesting - a step backwards and not forwards. "As if'; statements (concerning, e.g., narrow vs. broad focus) and problems that are trivial at face value (concerning, e.g., the relationship between phrasing units and accentuation and the ontology of sentence accent) will be illustrated with own material. I will argue that such trivial problems have to be dealt with in the beginning, and that they can constitute the very basis for the proper treatment of more far reaching and complex problems.
Link to this record: urn:nbn:de:bsz:291-scidok-41024
Series name: Vm-Report / Verbmobil, Verbundvorhaben, [Deutsches Forschungszentrum für Künstliche Intelligenz]
Series volume: 58
Date of registration: 22-Aug-2011
Faculty: SE - Sonstige Einrichtungen
Department: SE - DFKI Deutsches Forschungszentrum für Künstliche Intelligenz
Collections:SciDok - Der Wissenschaftsserver der Universität des Saarlandes

Files for this record:
File Description SizeFormat 
report_58_95.pdf239,53 kBAdobe PDFView/Open

Items in SciDok are protected by copyright, with all rights reserved, unless otherwise indicated.