Experiments have been conducted in which the subjects incrementally constructed dependency trees of Russian sentences. The subject was successively presented with growing initial segments of a sentence, and had to draw syntactic links between the last word of the segment and the previous words. The subject was also shown a limited right context-a fixed number of words following the last word of the segment. The results of the experiments show that the right context of 1 or 2 words is sufficient for confident incremental parsing of Russian narrative sentences.

Abstract. The paper presents the module of interactive word sense disambiguation and syntactic ambiguity resolution used within a machine translation system, ETAP-3. The method applied consists in asking the user to identify a word sense, or a syntactic interpretation, whenever the system lacks reliable data to make the choice automatically. In lexical disambiguation, part of man-machine dialogue refers to the analysis phase, while the other part is activated during transfer. For this purpose, entries of the working dictionaries of the system are supplemented with clear diagnostic comments and illustrations that enable the user to choose the most appropriate option and in this way channel the course of system operation. Introductory Remarks. ETAP-3 OverviewETAP-3 is a full-scale rule-based machine translation system that serves RussianEnglish and English-Russian pairs and has a number of small prototype modules for Russian-German, French-Russian, Russian-Korean, Russian-Spanish and ArabicEnglish translation. The MT system is developed as part of a multipurpose linguistic processor at the Laboratory of computational linguistics, Institute for Information Transmission Problems in Moscow [1][2][3][4]. Other modules of the processor include a parsing tool for deep syntactic tagging of text corpora, a UNL enconverter and deconverter tool, and several smaller-scale components (a module of synonymous paraphrasing of sentences, syntax checker, and a computer-assisted language learning tool).

No abstract

No abstract

Experiments have been carried out in which human subjects incrementally constructed dependency trees of English sentences. The subjects were successively presented with growing initial segments of a sentence, and had to draw syntactic links between the last word of the segment and the previous words. They were also shown a fixed number of lookahead words following the last word of the segment. The results of the experiments show that lookahead of 1 or 2 words is sufficient for confident incremental parsing of English declarative sentences.