Dissertations & Theses
http://hdl.handle.net/10106/1173
2024-03-28T10:15:35Z
-
Semantico-pragmatic aspects of the metalinguistic uses of comparatives in Mandarin Chinese and English
http://hdl.handle.net/10106/31748
Semantico-pragmatic aspects of the metalinguistic uses of comparatives in Mandarin Chinese and English
This dissertation examines a group of Metalinguistic Comparatives (MCs) in Mandarin Chinese, exploring what they suggest about the universality of comparatives. I show that Chinese MCs encode a more fine-grained scalarity of the subjective attitude conveyed, which varies in orientation, i.e., positive or negative, and strength, i.e., subtle or strong. By analyzing their uses in conversational settings, I claim that MCs are pragmatical devices for speakers to reject or rectify an utterance. In examining the shared Chinese ‘than’ marker in both Negative MCs (NegMCs) and Rhetorical Comparatives (RCs), I show that both subtypes are comparatives with a contrastive and negative sense, rather than a description of a degree-differential ordering relation. Furthermore, I propose a Logic Convertibility analysis for comparatives, showing a comparative is logically equivalent to the negation of the flipped inequality relation, i.e., (d1 d2) NOT (d2 d1), which builds the foundation for an implied negative inference to be retrieved. Finally, I suggest English comparatives are pragmatically ambiguous in that they descriptively encode a degree-differential inequality relation but can simultaneously produce a metalinguistic reading, i.e., to convey an evaluative attitude.
2023-08-23T00:00:00Z
-
Artificial Operators: Function and Use in English
http://hdl.handle.net/10106/30986
Artificial Operators: Function and Use in English
This dissertation provides a descriptive account of the use of ‘artificial operators’ on Reddit. I adopt the term artificial operators to refer to written symbols (e.g., #, ™, ©, ®) with overlapping linguistic and metadiscursive properties that are leveraged for a pragmatic effect (e.g., I love her even though she’s problematic™). I employ a mixed-methods approach, using a combination of corpus, experimental, and machine learning methods. Using a 1.2 billion word diachronic corpus of comments from the popular forum website, Reddit, I demonstrate that these operators are used in a small percentage of Reddit communities. Operator usage is therefore often community-specific and provides a useful indicator of the shared repertoire that exists within online Communities of Practice (Lave and Wenger, 1991; Wenger 1998). Operators often function as stance markers, contributing to all three components of stancetaking (evaluation, positioning, alignment). Operators also interact with adjectives by upscaling, or in some cases, downscaling the adjective meaning. Input from social media users indicate that there are perceivable differences between genuine and figurative uses of the operators but do not show perceivable differences between the meaning contributed by the four distinct operators under controlled conditions. Feature importance scores from machine learning models suggest that author-related features are more important than subreddit-related features in modeling operator use. Ultimately, this dissertation shows that artificial operators are a pragmatic resource that authors use to perform a variety of functions, including stance marking, upscaling, and indicating community membership. This work contributes to broader research around pragmatics in computer-mediated communication (CMC) which has shown that authors use CMC cues and other online textual resources not as a replacement for paralinguistic cues, but as new ways to create nuanced and sophisticated meaning.
2022-08-12T00:00:00Z
-
EMPLOYING A PARALLEL CORPUS-BASED APPROACH IN TEACHING SEMANTIC PROSODY AND COLLOCATIONAL BEHAVIOR TO ARABIC EFL LEARNERS
http://hdl.handle.net/10106/29980
EMPLOYING A PARALLEL CORPUS-BASED APPROACH IN TEACHING SEMANTIC PROSODY AND COLLOCATIONAL BEHAVIOR TO ARABIC EFL LEARNERS
This dissertation is intended to investigate if, and to what extent, a web-interface parallel corpus known as Reverso Context can assist Arabic EFL learners in addressing two aspects of word knowledge: semantic prosody and collocational behavior. A convergent mixed method design is adopted in this study in which one group of undergraduate L1 Arabic students are asked to do a pretest that is followed by a pedagogical intervention over the course of three 3-hour sessions and then a posttest is administered again with the same group of students. The posttest is followed by a one-on-one interview with the students and the course instructor to obtain a well-rounded view of their experiences with the pedagogical material and the new resource that they have been introduced to. The results reveal that the students are capable of using the parallel corpus (Reverso Context) effectively in semantic prosody identification and non-congruent collocation translation. The interview demonstrates that in spite of some difficulties the students encounter with Reverso Context, the students’ perceive the new resource positively and that it might be used to increase autonomy and discovery learning. The study also illustrates how Reverso Context can be implemented effectively to obtain the maximum benefit of this resource in a classroom setting with some pedagogical implications for EFL teachers. In addition, the study concludes with some tips for future researchers on how to better evaluate the efficacy of parallel corpora in foreign language pedagogy.
2021-08-10T00:00:00Z
-
LEARNING EFFICIENCIES FOR DIFFERENT ORTHOGRAPHIES: A COMPARATIVE STUDY OF HAN CHARACTERS AND VIETNAMESE ROMANIZATION
http://hdl.handle.net/10106/29693
LEARNING EFFICIENCIES FOR DIFFERENT ORTHOGRAPHIES: A COMPARATIVE STUDY OF HAN CHARACTERS AND VIETNAMESE ROMANIZATION
Chiung, Wi-Vun Taiffalo
In order to address the question of whether or not to abandon Han characters (Hanji), it is important to evaluate empirically the efficiency of Han writing. The purpose of this study is to compare the efficiency of learning to read and write in Hanji versus learning to read and write in phonemic writing systems, such as Vietnamese Chu Quoc Ngu (CQN) or Mandarin Bopomo. Three experiments were conducted in this study. The first experiment focused on a study of reading comprehension; the second one focused on a study of accuracy of writing dictation; and the last was a study of oral reading. A total of 453 subjects from Taiwan and 350 subjects from Vietnam were involved in the experiments. Subjects consisted of elementary school and college students. The reading comprehension tests were divided into groups Hanji, Bopomo, and CQN, in which subjects were examined with reading texts in Hanji, Bopomo, and CQN, respectively. The results of the reading comprehension tests reveal no statistically significant difference between Hanji and CQN groups. However, students from the second to fifth grades in the Bopomo group had significantly lower scores than students in the other groups. In dictation tests, subjects were divided into groups Taiwanese and Vietnamese. Tests in each group were given in soft and hard articles. The statistical results of tests on soft article reveal that students in both Taiwanese and Vietnamese groups significantly increased their score each year until the fourth grade, by which time they had the same statistical score as college students. As for tests on hard article, Taiwanese students spent more years in the acquisition of Hanji, and even the sixth graders' scores do not statistically reach the same level as college students. However, Vietnamese students had reached a college level at the fifth grade. Errors in the dictation tests were also analyzed, and twelve error types were found in the Taiwanese group. The major errors were made due to similarity in sound between correct and incorrect Han characters. The phonetic similarity errors account for 85.70% in the dictation test two. In addition to dictation tests, CQN also showed superiority in oral reading tests. The results indicate that CQN beginners are able to produce about 90% accuracy in oral reading after three or four months of learning, and reach nearly 100% accuracy a year later. In short, these results lead to the conclusion that Vietnamese CQN is more efficient than Chinese characters in learning to read and write.
2003-05-01T00:00:00Z