This thesis describes the development of Dapper (`Determiner And PrePosition Error Recogniser'), a system designed to automatically acquire models of occurrence for English prepositions and determiners to allow for the detection and correction of errors in their usage, especially in the writing of non-native speakers of the language. Prepositions and determiners are focused on because they are parts of speech whose usage is particularly challenging to acquire, both for students of the language and for natural language processing tools. The work presented in this thesis proposes to address this problem by developing a system which can acquire models of correct preposition and determiner occurrence, and can use this knowledge to identify divergences from these models as errors. The contexts of these parts of speech are represented by a sophisticated feature set, incorporating a variety of semantic and syntactic elements. DAPPER is found to perform well on preposition and determiner selection tasks in correct native English text. Results on each preposition and determiner are discussed in detail to understand the possible reasons for variations in performance, and whether these are due to problems with the structure of DAPPER or to deeper linguistic reasons. An in-depth analysis of all features used is also offered, quantifying the contribution of each feature individually. This can help establish if the decision to include complex semantic and syntactic features is justified in the context of this task. Finally, the performance of DAPPER on non-native English text is assessed. The system is found to be robust when applied to text which does not contain any preposition or determiner errors. On an error correction task, results are mixed: DAPPER shows promising results on preposition selection and determiner confusion (definite vs. indefinite) errors, but is less successful in detecting errors involving missing or extraneous determiners. Several characteristics of learner writing are described, to gain a clearer understanding of what problems arise when natural language processing tools are used with this kind of text. It is concluded that the construction of contextual models is a viable approach to the task of preposition and determiner selection, despite outstanding issues pertaining to the domain of non-native writing.

La tesi si colloca nell'ambito di studi sulla fraseologia nell'inglese prodotto da apprendenti. Presenta uno studio empirico delle sequenze di parole più ricorrenti in un corpus di inglese prodotto da apprendenti di livello avanzato durante chat asincrone in contesto universitario italiano. Secondo la letteratura d'area, sia nella lingua scritta che in quella parlata, le sequenze di parole degli apprendenti rivelano una scarsa attenzione alla variazione del registro a seconda del mezzo di comunicazione usato. Al fine di verificare la presenza di questa caratteristica in un tipo di comunicazione che si trova in posizione intermedia tra i due poli del continuum esistente tra parlato e scritto, la presente ricerca ha analizzato quantitativamente e qualitativamente le sequenze di parole più frequenti nel corpus di comunicazioni mediate dal computer (CMC) raccolto nell'ambito della presente ricerca. Successivamente, le sequenze più frequenti sono state confrontate con quelle estratte da due corpora di interlingua inglese prodotta da apprendenti italofoni, uno di testi scritti (ICLE, Granger et al. 2002) e uno interviste orali (LINDSEI, Gilquin et al. 2010 ). Il confronto ha rivelato che le sequenze più ripetute dagli apprendenti hanno caratteristiche distintive nei vari media e supporta solo in parte i precedenti studi in materia. Ciò è probabilmente dovuto sia alle caratteristiche di informalità e immediatezza della comunicazione mediata dal computer, che ai vantaggi motivazionali e al diverso tipo di elaborazione linguistica connaturato alla CMC. Per l'apprendente la CMC non presenta la stessa pressione comunicativa del parlato e, allo stesso tempo, egli ha la possibilità di monitorare la propria produzione in quanto distanziata da sé dal mezzo elettronico. / The present dissertation contributes to studies of phraseology in learner English. It is an analysis of recurrent sequences of words in a corpus of learner Computer-mediated Communication. English, collected by means of asynchronous chats in an Italian university context. Previous research has argued that the use of recurrent word sequences plays a major role in learner English fluency both in writing and in speech, and is one of the factors behind learner English register failures. Using a corpus-driven approach, the study analyses the most frequent word sequences extracted from the specially compiled Learner Chat Corpus (LCC). To determine the level of adaptation of learner English to different registers, data regarding 3-word sequences from LCC is compared with the Italian subcomponents of a well-known corpus of learner writing (ICLE, Granger et al. 2002) and a corpus of learner speech (LINDSEI, Gilquin et al. 2010 ). The cross-corpus comparisons provide evidence that learners employ combinations which make their English suitable to the mode they are using for communication. Quantitative and qualitative findings from the present research support only in part previous studies of learner English in terms of recurrent sequences. This is probably due both to the informality and spoken-like quality of CMC, and to its motivational advantages and processing differences connected to the fact that learners can monitor their output while communicating because learner language production is distanced by the electronic means.

Automatic syntactic analysis is essential for extracting useful information from large-scale learner data for linguistic research and natural language processing (NLP). Currently, researchers use standard POS taggers and parsers developed on native language to analyze learner language. Investigation of how such systems perform on learner data is needed to develop strategies for minimizing the cross-domain effects. Furthermore, POS taggers and parsers are developed for generic NLP purposes and may not be useful for identifying specific syntactic constructs such as subcategorization frames (SCFs). SCFs have attracted much research attention as they provide unique insight into the interplay between lexical and structural information. An automatic SCF identification system adapted for learner language is needed to facilitate research on L2 SCFs. In this thesis, we first provide a comprehensive evaluation of standard POS taggers and parsers on learner and native English. We show that the common practice of constructing a gold standard by manually correcting the output of a system can introduce bias to the evaluation, and we suggest a method to control for the bias. We also quantitatively evaluate the impact of fine-grained learner errors on POS tagging and parsing, identifying the most influential learner errors. Furthermore, we show that the performance of probabilistic POS taggers and parsers on native English can predict their performance on learner English. Secondly, we develop an SCF identification system for learner English. We train a machine learning model on both native and learner English data. The system can label individual verb occurrences in learner data for a set of 49 distinct SCFs. Our evaluation shows that the system reaches an accuracy of 84\% F1 score. We then demonstrate that the level of accuracy is adequate for linguistic research. We design the first multidimensional SCF diversity metrics and investigate how SCF diversity changes with L2 proficiency on a large learner corpus. Our results show that as L2 proficiency develops, learners tend to use more diverse SCF types with greater taxonomic distance; more advanced learners also use different SCF types more evenly and locate the verb tokens of the same SCF type further away from each other. Furthermore, we demonstrate that the proposed SCF diversity metrics contribute a unique perspective to the prediction of L2 proficiency beyond existing syntactic complexity metrics.

This study aims to investigate possible factors contributing to subject-verb agreement errors in Swedish junior and senior high school students' English written production. The sample data is collected from the Swedish Learner English Corpus (SLEC), which comprises student texts produced in a classroom setting. The texts are randomly chosen but evenly distributed in terms of binary gender, school year, and type of high school program. In this study, the texts included in the scope are written by students attending a Swedish-speaking school with Swedish as their first language. Errors are classified as overgeneralization or transfer and further classified in relation to the subject type, the verb type, and the distance between the subject and the verb. The classification of all correct instances of subject-verb agreement is also performed to further investigate possible error explanations. A total of 41 agreement errors were found in 24 texts written by students in the 9th and 11th grades. The results show that overgeneralization is more frequent than transfer errors. Overgeneralization suggests that the students are aware of the third-singular form but do not always apply it correctly, while transfer errors show a potential lack of awareness or attention to the form. In both cases, the errors indicate that these students have not automatized the principle yet. Errors are often related to subject types “a pronoun” or “a noun/noun phrase" and the verb be, which is the most frequently used verb. Most of the errors occur when the subject and the verb are in immediate contact, and more than half of them involve a relative pronoun as subject, which indicates that the learners have misinterpreted the grammatical principle or have not fully acquired it. Overuse of the third-person singular form can also be an effect of teaching and explicit learning, which makes learners apply the form whenever it seems possible and relevant.

Proper spelling is important for efficient communication between people with different first languages in the 21st century. While Swedish functions as an intranational language within Sweden, it sees little to no use outside of Scandinavia. English fills the role as a second language that all Swedish students must learn, yet more focus appears to be given to grammar rather than spelling. Spelling is important and knowing the kinds of spelling errors Swedish learners of English tend to make can help educators improve the spelling proficiency of their students. The aim of this study is to investigate the spelling errors made by senior high school students in Sweden by analyzing a collection of essays written by students and gathered in the Uppsala Learner English Corpus (ULEC). The results of this study show that spelling proficiency nearly doubled for students in their third year in senior high school compared to their first year, yet the distribution of spelling errors remained the same. Additionally, some particular sounds that appear to be especially problematic for Swedish spellers were identified, such as /ə/, /l/, /s/ and /k/. / Korrekt stavning är viktig för effektiv kommunikation mellan människor med olika modersmål i tjugohundratalet. Medans svenska fungerar som ett språk mellan människor inom Sverige, ser det lite till ingen nytta utanför Skandinavien. Engelska fyller rollen som andraspråk som alla svenska elever måste lära sig, ändå sätts mer fokus på grammatik över stavning. Stavning är viktig och att veta vilka typer av stavfel som svenska elever brukar göra på engelska kan hjälpa lärare förbättra elevernas stavningskunskaper. Syftet med den här studien är att undersöka svenska gymnasielevers felstavningar i Engelska genom att analysera en samlig essäer skrivna av studenter och samlade i Uppsala Learner English Corpus (ULEC). Resultaten från den här studien visar att stavningskunskaperna hos eleverna hade nästan fördubblats när de gick det tredje år på gymnasiet jämfört med när de gick det första, men att fördelningen av stavfel förblev densamma. Dessutom har vissa ljud identifierats som verkar vara särskilt problematiska för svenska elever att stava, såsom /ə/, /l/, /s/ and /k/.

This is a study that employs error analysis to investigate written production in English, by Swedish upper secondary learners of English, in order to determine which linguistic errors most commonly occur amongst this group, and to compare the results between first-year students and third-year students for a possible indication of which error types continue to occur throughout upper secondary school. The error categories included in this study are grammatical errors and errors as a result of transfer. The variable of gender will also be taken into account, due to the statistics and previous research that show female students tend to achieve higher results in academics. The purpose of the study is to gain a better understanding of how Swedish upper secondary learners acquire English, and to uncover which areas are most challenging for them, in the hopes of highlighting areas within ELT that may need revision. The participants of the study are students attending Swedish upper secondary schools, year 1 and 3. The analyzed data was collected from the Uppsala Learner English Corpus (ULEC), which consists of texts produced by Swedish learners of English attending middle school and upper secondary school.     The results show that certain error categories and types are consistently challenging for both first-year students and third-year students, which provides an indication of which areas in ELT might be lacking. Within the grammatical error category, all groups demonstrated a significant lack of knowledge pertaining to subject-verb agreement, as well as prepositions, which are both to a certain degree attributed to the first language; meaning, they may be the result of transfer. The male students were shown to outperform the female students; however, the female third-year students produced fewer errors than their male counterparts, which suggests a faster progression. The male third-year students were shown to have the same error rate as the male first-year students, which suggests a slower progression. While the third-year students produced fewer errors overall, the error types they struggled the most with are the same error types most commonly occurring in the first-year group, suggesting pedagogical remediation is needed.

This study investigates spelling errors among Swedish students of English. The purpose of this research is to develop a better understanding of the causes of spelling errors and to investigate whether there are any differences between program orientation and gender. The participants are English A students from Swedish high schools in year 1. The data analysed in this study consists of essays written by the participants which have been gathered into the Uppsala Learner English Corpus (ULEC). The results showed that the academic program has a lower error rate when compared to the vocational program and that female students display a lower error frequency than male students. These results reflect the same pattern found by previous studies on achievement between the programs and across gender. However, when comparing all variables, both gender and program, the results demonstrate that academic male students had the lowest error rate of all groups, which could partially be explained by the fact that the academic male students have a higher level of engagement with video- and online games when compared to the other groups. The language processes that the participants found the most challenging were letter omission, letter insertion and letter substitution. These results follow the same pattern highlighted in previous studies on spelling among both native speakers and learners. Finally, the exchange between the vowels <a>, <e> and <i> was found to be a frequent error among the students. Experiencing difficulties with the correct usage of vowels can be explained as a transfer error.

The aim of the study is to find out to what extent Swedish learners of English, in the first year of upper secondary school, make prepositional errors in their written production, and to what extent these errors can be attributed to negative transfer, overgeneralization and simplification by conducting an Error Analysis. A comparison between gender and type of program, academic and vocational, is made to find out in which type of program most errors appear and if there is any difference in terms of gender.  The data is annotated from the Swedish Learner English Corpus (SLEC), which consists of argumentative essays written by Swedish learners of English, and it consists of 24 randomly selected texts based on the variables binary gender, type of program, Swedish as their L1, school year, and English course. All the texts selected are written by students in the first year of upper secondary school studying the course English 5. The results of the study reveal that Swedish learners of English struggle with prepositional usage. In total, 649 prepositions were identified in the 24 texts. Out of these, 72 (11.09%) were used incorrectly. The most frequently used prepositions involved in these errors are of, for, in, to, and with. Most errors appear when prepositional phrases function as post-modifiers in noun phrases. Substitution is, by far, the most common type of error found, meaning that the students replace the correct preposition with an incorrect one. The results thus show that the students seem to be aware that a preposition should be used although they fail to choose the correct one. Female students make more prepositional errors than male students; similarly, students attending vocational programs make more prepositional errors than students attending academic programs. Most errors are cases of overgeneralizations, followed by negative transfer from Swedish, and simplification. However, many of the errors can still be attributed to negative transfer which suggests that, even though Swedish and English are similar languages which could lead to positive transfer, this does not seem to fully apply to prepositions.