Showing papers by "Tony McEnery published in 2001"

A Corpus-Based Study

Abstract: Starting in the early 1990s, work began at Lancaster University on the annotation of corpus data for the purpose of examining coreference in corpora of English (Fligelstone 1992). This work continued throughout the 1990s (McEnery, Tanaka, and Botley 1997; Garside, Fligelstone, and Botley 1997; Botley 1999), and this article represents part of its output. In this article, we will focus on one aspect of our work—the annotation and subsequent exploration of demonstrative pronouns (this, that, these, and those)—in three 100,000-word corpora of English.

Demonstratives in English: a corpus based study.

Abstract: L'A. presente une etude descriptive des pronoms demonstratifs (this, that, these et those) de l'anglais a partir d'un large corpus de donnees. Cette analyse est guidee par differentes questions recurrentes dans la problematique de recherche des anaphores demonstratives : 1. le fonctionnement anaphorique des demonstratifs, 2. le rapport entre type de discours et l'emploi des demonstratifs, 3. l'opposition entre demonstratifs anaphoriques et demonstratifs deictiques, 4. la delimitation entre anaphore, deixis, exophore.

Proximal and distal demonstratives: a corpus based study.

Abstract: In this article, we examine claims made about the conditions under which proximal (this, these) and distal (that, those) demonstratives are chosen in English. In doing so, we will be exploiting corpus data that have been annotated to show anaphoric relationships in texts (Fligelstone 1992; Botley and McEnery 2001). 1 In choosing to focus on proximal/distal demonstratives, we were aware that a number of claims had been made about the behavior of these words in English. Our goal was to take some of these claims and reevaluate them in light of a large amount of corpus evidence. The claims we investigate in this article arise from the work of Ariel (1988, 1990). Ariel proposed an influential model of the relationship between the choice of anaphoric expressions and the cognitive statuses of different discourse entities referred to by such expressions. This model is known as the accessibility scale and is described in the next section.

Multi-word unit alignment in English-Chinese parallel corpora

Abstract: Multi-word unit (MWU) alignment in bilingual/multilingual parallel corpora is an important goal for natural language engineering. An efficient algorithm for aligning MWUs in different languages could be of use in several practical applications, including machine translation, lexicon construction and cross-language information retrieval. A number of algorithms have been proposed and tested for this purpose, including collocation and association-strength testing statistics (Dagan et al., 1994; Smadja et al., 1996), n-gram, approximate string matching techniques (ASMT), finite state automata, (McEnery et al., 1997), bilingual parsing matching (Wu, 1997), and a hybrid connectionist framework (Wermter et al., 1997). Despite the work undertaken to date, however, reliable and robust MWU alignment remains an elusive goal.

On the search for metadiscourse units

Abstract: 1. Introduction To what extent do writers anchor their discourse in the current discourse situation and make the presence of the writer and/or reader overt? What types of situations are at hand when writers refer to their text as text, to themselves or to their readers? To what extent are texts monologic or dialogic? These types of questions have recently attracted a lot of attention within research on metadiscourse. Metadiscourse refers to discourse about ongoing discourse and is interesting to study from the perspective of how the writer's or reader's presence in the text are made explicit. It has been studied by various branches of linguistics, for example text linguistics and is rapidly becoming a dynamic field of research. One particularly interesting perspective is the study of cultural differences in the use of metadiscourse. Several researchers have shown that metadiscourse typically differs across cultures (e g Cultural differences in writing have been studied within the field of contrastive rhetoric, on the basis of the primary hypothesis that there are culture-specific patterns of writing, and that these cause interference in L2 writing (see Connor 1996:90). In this paper, I will report ongoing research into the use of metadiscourse in written argumentative texts by native and non-native speakers of English. The study is corpus-based and comparative, contrasting the use of metadiscourse by Swedish advanced learners' writing in English to the writing of native speakers of British and American English. 1 All writers are university students. The argumentative essays are full-length, and are available within the framework of the International Corpus of Learner English (Granger 1993). In case there may be cultural differences in the use of metadiscourse, the British English and American English parts of the control corpus are kept separate. One of the aims of my thesis is to investigate metadiscoursal patterns with explicit reference to the writer or reader. Some examples of metadiscourse specify discourse acts that the writer intends to perform are, for instance: