Writing research articles in English is a common requirement in doctoral studies in the natural sciences; however, learning to write the research article genre is challenging, particularly in a foreign language (L2). A potential resource for learning the RA genre is giving and receiving peer review. L2 writers at the undergraduate level have been found to benefit from PR, but less  is known about the learning of L2 writers at the graduate level who are writing for specialized discourse communities. The aim of this dissertation is to describe how a group of L2 doctoral students in the natural sciences used online peer review in a research writing course. Inductive analysis was used to categorize the kinds of review comments that 11 course participants gave and received. In another study, three students’ revised texts were analyzed in detail to see how they used peer comments. To explore  students’ perceptions of using the PR activity, pre and post course interviews were inductively analyzed.  Findings were interpreted using Vygotskian constructs of learning in order to see where mediation likely occurs.  The combined studies suggest that reviewers adopted roles that influenced what they noticed, analyzed and languaged. As a group, they focused on the lexical and syntactic precision of peers’ texts, as well as the organization, cohesion, voice, stance and research knowledge. Writers used the intent of the review comments approximately 40 percent of the time, but this usage reflected only a small portion of the writers’ revision activities that occurred in response to review.  Other activities included composing, re-writing, investigating, interviewing outsiders, and re-ordering the texts.  Writers found precision and organizational comments most useful. Findings from these combined studies indicate that peer review can be a potentially powerful tool for doctoral students to familiarize themselves with discipline-specific research articles.

This thesis discusses rhetorical features of Indonesian research articles (RAs) in three disciplinary areas: Economics, Education and Psychology. These were written by Indonesian speakers and published mainly in university-based scientific journals. The main focus of this thesis is on the examination of the patterns of communicative purposes or ‘Moves’ and their subsequent elements or ‘Steps’ of the introduction sections of these articles. The analyses include the examination of communicative purposes and persuasive values of the texts, linguistic resources used to materialise the communicative purposes and persuasions, and the cultural factors (ie. norms, beliefs and values) and scientific practices and academic writing conventions underlying the specific rhetorical features. ¶ This study found that the macro rhetorical structure of the Indonesian RAs (ie. the Introduction, Methods, Results and Discussion or IMRD pattern) is relatively similar to that of the English RAs except that, unlike in English RAs, the conclusion and suggestion section in the Indonesian RAs have a separate section. However, the communicative purposes and persuasions in the introduction sections in the two groups of the RAs (English and Indonesian) are relatively different. Differences are also found in the way that rhetorical works use the linguistic resources to materialise the communicative purposes and persuasions in the introduction sections of the two groups of RAs. Some of the rhetorical differences are because of the differences in the research practices and scientific writing conventions in Indonesian and in English speaking countries, while others are because of cultural differences reflected in the two languages. ¶ The pedagogical implication of this study is that the Indonesian RA genre needs to be explicitly taught to Indonesian students, particularly university students in order to give them more access to the content of Indonesian research, and to develop skills needed by Indonesian researchers and research writers. For this purpose, an appropriate approach needs to be developed; that is to teach the generic features of Indonesian RAs such as those in social sciences written in Bahasa Indonesia or Indonesian.

This study investigated the use of lexical bundles in research articles written in English by Turkish scholars. For the purpose of the study, a corpus of published research articles produced by Turkish scholars in six different academic disciplines was collected. The four-word lexical bundles that appeared at least twenty times in this one million word corpus were identified and further analyzed both structurally and functionally based on the previous taxonomies developed by Biber, Johansson, Leech, Conrad and Finegan (1999) and Biber, Conrad and Cortes (2004). The results of this study revealed that the lexical bundles found have structural correlates as well as strong functional features that help to construct discourse in academic writing. The conclusions drawn from this study could be applied to the teaching of academic genres to researchers in English as a Foreign Language context and are expected to provide insights for further corpus-based studies in academic writing.

This dissertation reports on the results of a study into the characteristics of epistemic topoi and the methods of their identification in a corpus of biomedical publications. The study was conceived in response to the need for a systematized description of the organization of argumentative text and discourse. This need is well recognized in knowledge-intensive fields: information processing, storage, and retrieval; corpus analysis and natural language processing; data mining, knowledge management and translation; professional training and education. The study followed the design of a situated study combined with a methodological inquiry. I used inductive methods to describe the features and functions of recurrent patterns of argumentative and linguistic organization. This part of the study consisted in close reading of a corpus of fifty-five NTG papers and rhetorical and linguistic annotation of seventeen clinical studies (45,599 words) selected from the corpus. The data was generated by means of rhetorical and linguistic analysis. Visual annotation played an essential role in the identification and description of the argumentative patterns, complementing the traditional methods of corpus analysis. Forty-eight basic and nine composite epistemic topoi forming the superstructure of the papers were identified in the corpus. The topoi were found to be loosely associated with the IMRD structure and signalled with configurations of lexicogrammatical, semantic, deictic, and coreferential features. The topoi were classified according to the modes of reasoning and textual and discursive functions. The obtained results confirmed earlier insights into the links of linguistic patterning with text and discourse semantics. A significant outcome of the linguistic analysis is a catalogue of linguistic features that were found to have regular links with the topoi in the corpus. The role of linguistic configurations as identifiers of argumentative meanings makes them a valuable medium of text and discourse analysis. By linking the argumentative meanings to the surface features of text and discourse, the analysis of linguistic configurations presents informatics practitioners with an alternative to the current methods of natural language processing and knowledge management. The catalogue of linguistic features and a detailed description of the study design make the presented findings amenable to secondary analysis, extrapolation, and generalization. The auxiliary objectives of this study were a survey of argumentative practices represented in the corpus and a review of the state of epistemic research. The results of the survey and review suggest that agonistic reasoning practices and over-reliance on reductionist models have negative implications for research writing and communication. Specifically, they hamper analysis of argumentative organization of natural text and discourse. As an alternative to agonistic argumentation, I propose an argumentation model based on Aristotle’s and Kneale’s conceptions of situated knowledge and learning. The model of textual and discursive organization that accommodates situated knowledge and learning is political stasis. This model can be used as a heuristic and analytic tool. In this dissertation I use it as an explanatory conception and as a system of reference points for identifying significant research trends both in argumentation studies and in clinical NTG research.

This dissertation reports on the results of a study into the characteristics of epistemic topoi and the methods of their identification in a corpus of biomedical publications. The study was conceived in response to the need for a systematized description of the organization of argumentative text and discourse. This need is well recognized in knowledge-intensive fields: information processing, storage, and retrieval; corpus analysis and natural language processing; data mining, knowledge management and translation; professional training and education. The study followed the design of a situated study combined with a methodological inquiry. I used inductive methods to describe the features and functions of recurrent patterns of argumentative and linguistic organization. This part of the study consisted in close reading of a corpus of fifty-five NTG papers and rhetorical and linguistic annotation of seventeen clinical studies (45,599 words) selected from the corpus. The data was generated by means of rhetorical and linguistic analysis. Visual annotation played an essential role in the identification and description of the argumentative patterns, complementing the traditional methods of corpus analysis. Forty-eight basic and nine composite epistemic topoi forming the superstructure of the papers were identified in the corpus. The topoi were found to be loosely associated with the IMRD structure and signalled with configurations of lexicogrammatical, semantic, deictic, and coreferential features. The topoi were classified according to the modes of reasoning and textual and discursive functions. The obtained results confirmed earlier insights into the links of linguistic patterning with text and discourse semantics. A significant outcome of the linguistic analysis is a catalogue of linguistic features that were found to have regular links with the topoi in the corpus. The role of linguistic configurations as identifiers of argumentative meanings makes them a valuable medium of text and discourse analysis. By linking the argumentative meanings to the surface features of text and discourse, the analysis of linguistic configurations presents informatics practitioners with an alternative to the current methods of natural language processing and knowledge management. The catalogue of linguistic features and a detailed description of the study design make the presented findings amenable to secondary analysis, extrapolation, and generalization. The auxiliary objectives of this study were a survey of argumentative practices represented in the corpus and a review of the state of epistemic research. The results of the survey and review suggest that agonistic reasoning practices and over-reliance on reductionist models have negative implications for research writing and communication. Specifically, they hamper analysis of argumentative organization of natural text and discourse. As an alternative to agonistic argumentation, I propose an argumentation model based on Aristotle’s and Kneale’s conceptions of situated knowledge and learning. The model of textual and discursive organization that accommodates situated knowledge and learning is political stasis. This model can be used as a heuristic and analytic tool. In this dissertation I use it as an explanatory conception and as a system of reference points for identifying significant research trends both in argumentation studies and in clinical NTG research.

Applied linguistics has lately been seen in studies of formulaicity of language operating through recurrent word combinations. The present study deals with one type of word combinations, namely lexical bundles (LBs), which are defined as a sequence of three or more words that frequently co-occur in a particular register (Biber et al., 1999). The present study is a corpus-based analysis of four-word lexical bundles extracted from Music research articles (RAs). The Corpus of Music Research Articles (CMRA) of one million words was created in order to perform structural classification of the retrieved lexical bundles and an analysis of their functions. The CMRA includes 110 articles collected from international music journals from various music subdisciplines. In order to find which lexical bundles were characteristic of music research specifically, the findings were compared to previous research based on other academic disciplines. The list of 218 lexical bundles was compared to the one of three different subject areas (Jalilifar et al., 2016) with the purpose of identification of discipline-specific LBs (n=102) which included 20 topic-specific bundles; and general lexical bundles (n=116) which included 56 core bundles shared among Music and three subject areas (Art and Humanities, Sciences and Social sciences). Structurally and functionally, the analysis of the extracted lexical bundles demonstrated that native English expert writers predominantly used preposition-based phrases (50%), with respect to structure; and research-oriented bundles (74%), with respect to function. The findings have pedagogical applications and could be used in courses in English for Specific Purposes.

Scholarly impact and visibility have traditionally been assessed by estimating the number and quality of publications and citations in the scholarly literature. But now scholars are increasingly visible on the Web. Twitter, Mendeley and other social media platforms are quite well-known and interesting tools for sharing research articles among scholars. Based on the capabilities of altmetrics, this thesis analyses the altmetric coverage and impact of LIS articles published by Swedish universities during 2013 to 2017. It also tries to paint a picture of demographic of people engaged with these articles. The most common altmetric sources are considered using a sample of 170 LIS journal articles. The findings were interpreted using two different sociological theories: the normative theory of Merton´s norms and the theory of social constructivism. The result of the study showed that Mendeley has the highest coverage of journal articles (65 percent) followed by Twitter (33 percent) while very few of the publications are mentioned in blogs or on Facebook. Researchers were the main Mendeley readers of articles with 53 percent, followed by the general public at 26 percent. This study, on the other hand, found out that public users were the main group that shared articles on Twitter. The list of articles with a high number of tweets showed that most topics in the field of LIS are associated with bibliometrics and citation impacts. The results demonstrate that the adoption of altmetric methods within the LIS fields is inevitable, but several issues must be considered to identify their potential.

Researchers have made significant progress in information extraction from short documents in the last few years, including social media interaction, news articles, and email excerpts. This research aims to extract technical entities like hardware resources, computing platforms, compute time, programming language, and libraries from scholarly research articles. Research articles are generally long documents having both salient as well as non-salient entities. Analyzing the cross-sectional relation, filtering the relevant information, measuring the saliency of mentioned entities, and extracting novel entities are some of the technical challenges involved in this research. This work presents a detailed study about the performance, effectiveness, and scalability of rule-based weakly supervised algorithms. We also develop an automated end-to-end Research Entity and Relationship Extractor (E2R Extractor). Additionally, we perform a comprehensive study about the effectiveness of existing deep learning-based information extraction tools like Dygie, Dygie++, SciREX. The research also contributes a dataset containing novel entities annotated in BILUO format and represents the baseline results using the E2R extractor on the proposed dataset. The results indicate that the E2R extractor successfully extracts salient entities from research articles. / Master of Science / Information extraction is a process of automatically extracting meaningful information from unstructured text such as articles, news feeds and presenting it in a structured format. Researchers have made significant progress in this domain over the past few years. However, their work primarily focuses on short documents such as social media interactions, news articles, email excerpts, and not on long documents such as scholarly articles and research papers. Long documents contain a lot of redundant data, so filtering and extracting meaningful information is quite challenging. This work focuses on extracting entities such as hardware resources, compute platforms, and programming languages used in scholarly articles. We present a deep learning-based model to extract such entities from research articles and research papers. We evaluate the performance of our deep learning model against simple rule-based algorithms and other state-of-the-art models for extracting the desired entities. Our work also contributes a labeled dataset containing the entities mentioned above and results obtained on this dataset using our deep learning model.

Made available in DSpace on 2016-04-28T18:22:57Z (GMT). No. of bitstreams: 1 Etelvo Ramos Filho.pdf: 6034350 bytes, checksum: aa0c4cb5722bdf425e95516446096188 (MD5) Previous issue date: 2014-12-10 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / This dissertation describes how a corpus of 900 research articles written in English by researchers of 10 fields of study and nine different origins is mapped onto the dimensions of variation proposed by Biber (1988). Its theoretical and methodological underpinnings are provided by Corpus Linguistics and Multidimensional Analysis. The former is an area in Applied Linguistics in which language is seen as a probabilistic system and for whose studies computational tools and corpora are used. The latter is a corpus based approach for the study of dimensions of variations, which uses statistical procedures to identify relationships between linguistic features and registers in large amounts of texts. The methodology included the compilation of a corpus (Corpus of English Research Articles CERA), which is composed of articles collected using the Internet, its processing and analyses based both on origin and field. The result of the analyses for both the origin and the field shows that the corpus is composed of articles whose features characterize them in the 5 dimensions of variation proposed by Biber (1988) as: 1) being informational, 2) being non-narrative, 3) having explicit reference, 4) having non-explicit persuasion, and 5) having abstract information / Fundamentada teórico e medotologicamente na Linguística de Corpus e na Análise Multidimensional, esta tese analisa um corpus de 900 artigos de pesquisa em língua inglesa, escritos por pesquisadores de dez áreas, provenientes de nove diferentes origens. Para tanto, a pesquisa se apoia em uma área da Linguística Aplicada que vê a língua como um sistema probabilístico e para cujos estudos são utilizados ferramentas computacionais e corpora. A Análise Multidimensional é uma abordagem baseada em corpus para o estudo de dimensões de variação que usa procedimentos estatísticos para identificar relações entre traços linguísticos e registros em grandes quantidades de textos. A metodologia incluiu a compilação de um corpus de estudo (Corpus of English Research Articles CERA), composto de artigos coletados por meio da internet, que envolveu processamento do texto e análises por origem e área de estudo. A partir do mapeamento nas cinco dimensões de variação de Biber (1988), o resultado das análises por origem e área mostra que o corpus é composto por artigos cujos traços os situam nos seguintes polos: 1) produção informacional, 2) preocupação não-narrativa, 3) referências explícitas, 4) persuasão não-explícita e 5) informação abstrata

This presentation covers the origin, history and scope of Partnership: the Canadian Journal of Library and Information Practice and Research; the editorial process and timeframes; research and scholarship for librarians; the peer review process; the components of a good research article; practical tips on what editors look for in a manuscript submission with examples from Partnership journal. The presentation is intended for anyone writing a research article but should be particularly helpful to first-time authors.