No description available.

在大學入學考的英文科目中，英文作文的考題以敘述文與論說文兩種類型最為常見。特別注意的是，於每年二月份所舉行的大學學測的英文考科中，2004年至2014年間，敘述文的出現次數高達八次。因此，如何教導學生進行英文的敘述文寫作是每位高中英文老師所要面對的課題。此外，記敘文與敘述文均為描寫形式的文章，學生容易對這兩種文體產生混淆，可能影響學生的敘述文體的英文寫作而造成負面的影響。 為了幫助學生能夠了解敘述文體的特性，本研究經由以英文母語者的文章整理出七項關於敘述文的特性。此外，本研究也利用敘述文的七項特性來檢驗五篇高中英文教科書的文章與五十篇學生的英文文章，除了審視這五十五篇文章是否達到七項特性之外，也利用學生敘述文的分數發現會影響分數的語言特徵。最後，藉由研究中的發現給予英文教科書編寫與英文敘述文教學及學習上的建議。 / The most common styles of English writing in the college entrance exam are narratives and expository. Furthermore, narrative writing occurs more frequently than expository in the General Scholastic Ability Test (GSAT) held in early February every year. During the years 2004 to 2014, writing a narrative with sequential pictures was used eight times to evaluate Taiwanese senior high school students' English writing abilities. Therefore, the teaching and learning of English narratives is of a great importance for teachers and students in Taiwanese senior high schools. In order to help the teaching and learning of narrative writing, this study aims to find out the linguistic features of the narratives based on the analysis of English narratives written by native speakers. Then, according to the seven linguistic features of the narrative found in this study, fifty-five narratives from Taiwanese English textbooks and senior high school students were investigated to know whether the seven linguistic features were presented. Furthermore, to determine the significant grades-related linguistic features of students' narratives, the grades of students' narratives and the linguistic features are examined in this study. Lastly, by using the findings of the study, some pedagogical implications are offered for English textbooks, the teaching and learning of the narrative style.

This master thesis work presents a novel method of automatic essay scoring for Swedish national tests written by upper secondary high school students by deploying neural network architectures and linguistic feature extraction in the framework of Swegram. There are four sorts of linguistic aspects involved in our feature extraction: count-based,lexical morphological and syntactic. One of the three variants of recurrent network, vanilla RNN, GRU and LSTM, together with the specific model parameter setting, is implemented in the Automatic Essay Scoring (AES) modelling with extracted features measuring the linguistic complexity as text representation. The AES model is evaluated through interrater agreement with human assigned grade as target label in terms of quadratic weighted kappa (QWK) and exact percent agreement. Our best observed averaged QWK and averaged exact percent agreement is 0.50 and 52% over 10 folds among our all experimented models.

Written skills are an essential evaluation criterion for a student’s creativity, knowledge, and intellect. Consequently, academic writing is a common part of university and college admissions applications, standardized tests, and classroom assessments. However, the task for teachers is quite daunting when it comes to essay scoring. Then Automated Essay Scoring may be a helpful tool in the decision-making by the teacher.  There have been many successful models with supervised or unsupervised machine learning algorithms in the eld of Automated Essay Scoring. This thesis work makes a comparative study among various neural network models with supervised machine learning algorithms and different linguistic feature combinations. It also proves that the same linguistic features are applicable to more than one language.  The models studied in this experiment include TextCNN, TextRNN_LSTM, Tex- tRNN_GRU, and TextRCNN trained with the essays from the Automated Student Assessment Prize (ASAP) from Kaggle competitions. Each essay is represented with linguistic features measuring linguistic complexity. Those features are divided into four groups: count-based, morphological, syntactic, and lexical features, and the four groups of features can form a total of 14 combinations.  The models are evaluated via three measurements: Accuracy, F1 score, and Quadratic Weighted Kappa. The experimental results show that models trained only with count-based features outperform the models trained using other feature combinations. In addition, TextRNN_LSTM performs best, with an accuracy of 54.79%, an F1 score of 0.55, and a Quadratic Weighted Kappa of 0.59, which beats the statistically-based baseline models.

Spelling development is a linguistic process which involves the interaction of phonological, orthographic, and morphological knowledge (Bahr, Silliman, & Berninger, in press). It is also clear these linguistic factors are influenced by a person's dialect. Previous research has indicated that use of African American English (AAE) does influence spelling performance (Kohler, Bahr, Silliman, Bryant, Apel, & Wilkinson, 2007); however, few studies have considered how dialect use influences spelling as a function of spelling task (i.e., real vs. non-word tasks), error category (phonological, orthographic, or morphological) or grade. A secondary goal was to note if dialectal or developmental errors predominated in the noted misspellings. The Phonological, Orthographic, and Morphological Assessment of Spelling (POMAS, Silliman, Bahr, & Peters, 2006) was used to provide a fine-grained analysis of the spelling errors of 80 typically developing African American children in grades 1 (n = 39) and 3 (n = 41). These children were screened for language ability and they were determined to be AAE speakers by observing their use of phonological and/or morphosyntactic dialect features when retelling a story. Age-appropriate real word and non-word spelling tasks were developed which incorporated common features of AAE. A three-way ANOVA revealed that differences in error frequency were dependent upon word type, error type and grade. On the real word spelling task, children in both grades made more orthographic errors than phonological or morphological errors. On the non-word spelling task, students in both grades made fewer orthographic errors and students in grade 3 made significantly more phonological errors, while the number of phonological errors noted remained fairly constant across tasks for the children in grade 1. Common misspelling patterns revealed developmental errors, as well as errors attributed to AAE. A closer look at the occurrence of AAE features revealed that first graders were more likely to reflect dialectal patterns in their spelling than the third graders. This is possibly due to differences in exposure to the academic register and experience in code-switching. Finally, the real words elicited more AAE features than non-words suggesting that phonetic and linguistic contexts might influence the occurrence and use of AAE.

Readability assessment (RA) is to assign a score or a grade to a given document, which measures the degree of difficulty to read the document. RA originated in language education studies and was used to classify reading materials for language learners. Later, RA was applied to many other applications, such as aiding automatic text simplification.  This thesis is aimed at improving the way of using Transformer for RA. The motivation is the “pipeline” effect (Tenney et al., 2019) of pretrained Transformers: lexical, syntactic, and semantic features are best encoded with different layers of a Transformer model.  After a preliminary test of a basic RA model that resembles the previous works, we proposed several methods to enhance the performance: by using a Transformer layer that is not the last, by concatenating or mixing the outputs of all layers, and by using syntax-augmented Transformer layers. We examined these enhanced methods on three datasets: WeeBit, OneStopEnglish, and CommonLit.  We observed that the improvements showed a clear correlation with the dataset characteristics. On the OneStopEnglish and the CommonLit datasets, we achieved absolute improvements of 1.2% in F1 score and 0.6% in Pearson’s correlation coefficients, respectively. We also show that an 𝑛-gram frequency- based baseline, which is simple but was not reported in previous works, has superior performance on the classification datasets (WeeBit and OneStopEnglish), prompting further research on vocabulary-based lexical features for RA.

Concepts such as “women’s language” and “men’s language” suggest differences between how men and women speak, often concerning stereotypes. However, some research within the field of linguistics presents evidence showing little or no difference. This study aims to investigate linguistic differences between male and female characters, respectively, in The Office and analyze whether these findings correspond with, or challenge stereotypes associated with “men’s” and “women’s language”. Specifically, the analysis focuses on the lines assigned to the male and the female characters, respectively. The data was retrieved by closely watching eight episodes from two seasons and transcribing the lines spoken by male and female characters. The research methods employed are qualitative conversational analysis (CA) and quantitative content analysis. The results reveal several differences between how the male and the female characters speak in The Office. The female characters’ lines exhibit linguistic features associated with “women’s language” and lines borne out by the male characters are characterized by linguistic features typical of “men’s language”. Furthermore, these differences seem to correspond with stereotypes of gendered language features. In conclusion, the study suggests that the TV show adheres to stereotypes, potentially reinforcing stereotypical characterizations of how men and women speak. Additionally, this study suggests further research in the field of gender and language within TV shows to explore differences and the effects of these.

Ever since we entered the digital communication era, the ease of information sharing through the internet has encouraged online literature searching. With this comes the potential risk of a rise in academic misconduct and intellectual property theft. As concerns over plagiarism grow, more attention has been directed towards automatic plagiarism detection. This is a computational approach which assists humans in judging whether pieces of texts are plagiarised. However, most existing plagiarism detection approaches are limited to super cial, brute-force stringmatching techniques. If the text has undergone substantial semantic and syntactic changes, string-matching approaches do not perform well. In order to identify such changes, linguistic techniques which are able to perform a deeper analysis of the text are needed. To date, very limited research has been conducted on the topic of utilising linguistic techniques in plagiarism detection. This thesis provides novel perspectives on plagiarism detection and plagiarism direction identi cation tasks. The hypothesis is that original texts and rewritten texts exhibit signi cant but measurable di erences, and that these di erences can be captured through statistical and linguistic indicators. To investigate this hypothesis, four main research objectives are de ned. First, a novel framework for plagiarism detection is proposed. It involves the use of Natural Language Processing techniques, rather than only relying on the vii traditional string-matching approaches. The objective is to investigate and evaluate the in uence of text pre-processing, and statistical, shallow and deep linguistic techniques using a corpus-based approach. This is achieved by evaluating the techniques in two main experimental settings. Second, the role of machine learning in this novel framework is investigated. The objective is to determine whether the application of machine learning in the plagiarism detection task is helpful. This is achieved by comparing a thresholdsetting approach against a supervised machine learning classi er. Third, the prospect of applying the proposed framework in a large-scale scenario is explored. The objective is to investigate the scalability of the proposed framework and algorithms. This is achieved by experimenting with a large-scale corpus in three stages. The rst two stages are based on longer text lengths and the nal stage is based on segments of texts. Finally, the plagiarism direction identi cation problem is explored as supervised machine learning classi cation and ranking tasks. Statistical and linguistic features are investigated individually or in various combinations. The objective is to introduce a new perspective on the traditional brute-force pair-wise comparison of texts. Instead of comparing original texts against rewritten texts, features are drawn based on traits of texts to build a pattern for original and rewritten texts. Thus, the classi cation or ranking task is to t a piece of text into a pattern. The framework is tested by empirical experiments, and the results from initial experiments show that deep linguistic analysis contributes to solving the problems we address in this thesis. Further experiments show that combining shallow and viii deep techniques helps improve the classi cation of plagiarised texts by reducing the number of false negatives. In addition, the experiment on plagiarism direction detection shows that rewritten texts can be identi ed by statistical and linguistic traits. The conclusions of this study o er ideas for further research directions and potential applications to tackle the challenges that lie ahead in detecting text reuse.

<p>The literature on J. K. Rowling’s Harry Potter is prolific. People have written on various topics dealing with issues such as translation, etymology and diverse areas concerning the language. In this essay, I examine whether linguistic features such as reporting verbs, adverbs of manner and adjectives contribute to the depiction of heroic and villainous characters. Before conducting this research, my assumptions were that there would be a great difference in the value of the words depending on the character they were associated with. I wanted to see if the heroic characters used verbs and adverbs with positive connotations, and the villainous characters verbs and adverbs with negative connotations. I also wanted to know if the adjectives describing the characters could, in themselves, clearly indicate whether a character was a hero or a villain.</p><p>The results of my research suggested that the choice of particular verbs and adverbs contributed only indirectly to the depiction of the characters. Without context, it was not possible to know if the character was a hero or a villain simply by identifying the verbs and adverbs used to describe their speech. By contrast, the choice of particular adjectives did appear to indicate more clearly whether a character was hero or villain. Finally, the results of my research indicated that context, rather than the use of particular linguistic features was often the most important factor in contributing to the portrayal of characters in the novel.</p>

The literature on J. K. Rowling’s Harry Potter is prolific. People have written on various topics dealing with issues such as translation, etymology and diverse areas concerning the language. In this essay, I examine whether linguistic features such as reporting verbs, adverbs of manner and adjectives contribute to the depiction of heroic and villainous characters. Before conducting this research, my assumptions were that there would be a great difference in the value of the words depending on the character they were associated with. I wanted to see if the heroic characters used verbs and adverbs with positive connotations, and the villainous characters verbs and adverbs with negative connotations. I also wanted to know if the adjectives describing the characters could, in themselves, clearly indicate whether a character was a hero or a villain. The results of my research suggested that the choice of particular verbs and adverbs contributed only indirectly to the depiction of the characters. Without context, it was not possible to know if the character was a hero or a villain simply by identifying the verbs and adverbs used to describe their speech. By contrast, the choice of particular adjectives did appear to indicate more clearly whether a character was hero or villain. Finally, the results of my research indicated that context, rather than the use of particular linguistic features was often the most important factor in contributing to the portrayal of characters in the novel.