Exploring Vocabulary Acquisition in a Language Learning Application: An Error Analysis of the L2 English and Swedish Lexical Items
Hyytiäinen, Kamilla (2023)
Hyytiäinen, Kamilla
2023
Englannin kielen ja kirjallisuuden maisteriohjelma - Master's Programme in English Language and Literature
Informaatioteknologian ja viestinnän tiedekunta - Faculty of Information Technology and Communication Sciences
This publication is copyrighted. Only for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2023-05-12
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-202304254380
https://urn.fi/URN:NBN:fi:tuni-202304254380
Tiivistelmä
This thesis examines the lexical errors that the Finnish learners of a language learning application called WordDive made in the English and Swedish preparation courses. The aim was to first investigate which words were the easiest and the most difficult ones and then categorize the lexical errors produced by the users to examine commonalities and differences between the target languages and the cohorts. Another goal for the study was to find out whether the learners’ L1 (Finnish) interferes with L2 English and Swedish and causes errors. The material included data about the easiness of the studied items and the errors made, and the data was analyzed both quantitatively and qualitatively. The hypothesis was that the learners make relatively similar errors in both languages, but the English influence would be more prominent in the Swedish study items.
There were slight differences found in both the levels of difficulty and easiness between the languages as well as the error types. The overall view was that Finnish students seemed to master both difficult and easy English words better than Swedish words. The easiest study items included short words and often the Finnish translation resembled the L2 form, whereas the most difficult words were longer, included multi-unit words and lacked similar forms between the L1 and L2.
The lexical errors were classified into four main categories of which word form, word use and word meaning, derived from Nation’s (2001) aspect of word knowledge and non-words from the data. Furthermore, the subcategories were formed partially along Meriläinen’s (2010) transfer study as well as in data-driven fashion, to serve the purpose of the error analysis. The error distribution revealed that word meaning errors comprised the largest category in English, while word form errors dominated in Easy Swedish and non-word errors in Difficult Swedish. Some interesting findings were made about the L1 interference: it was English and Swedish rather than L1 Finnish that were interfering the L2 production. Therefore, the results deviated slightly from the hypothesis in that also Swedish had influence on the spelling of English words in some instances. The English and Swedish influence in vocabulary acquisition may be explained with their cross-linguistic similarity, contrary to Finnish.
All in all, the results indicate that the learners of WordDive’s preparation courses used different techniques to apply their prior linguistic knowledge to the second language learning, but regardless of that, the largest error subcategories included words without semantic resemblance, synonym errors, words with semantic association, affix errors and phonetic errors that had little influence from the L1. Those areas of vocabulary acquisition could therefore be taken into consideration when developing vocabulary exercises for L1 Finnish students. Some improvements are suggested regarding the WordDive method and application to prevent the users from producing the frequently occurring errors, such as providing more hints about the context of the word beforehand as well as classifying the answers semantically close to the target word as almost correct.
There were slight differences found in both the levels of difficulty and easiness between the languages as well as the error types. The overall view was that Finnish students seemed to master both difficult and easy English words better than Swedish words. The easiest study items included short words and often the Finnish translation resembled the L2 form, whereas the most difficult words were longer, included multi-unit words and lacked similar forms between the L1 and L2.
The lexical errors were classified into four main categories of which word form, word use and word meaning, derived from Nation’s (2001) aspect of word knowledge and non-words from the data. Furthermore, the subcategories were formed partially along Meriläinen’s (2010) transfer study as well as in data-driven fashion, to serve the purpose of the error analysis. The error distribution revealed that word meaning errors comprised the largest category in English, while word form errors dominated in Easy Swedish and non-word errors in Difficult Swedish. Some interesting findings were made about the L1 interference: it was English and Swedish rather than L1 Finnish that were interfering the L2 production. Therefore, the results deviated slightly from the hypothesis in that also Swedish had influence on the spelling of English words in some instances. The English and Swedish influence in vocabulary acquisition may be explained with their cross-linguistic similarity, contrary to Finnish.
All in all, the results indicate that the learners of WordDive’s preparation courses used different techniques to apply their prior linguistic knowledge to the second language learning, but regardless of that, the largest error subcategories included words without semantic resemblance, synonym errors, words with semantic association, affix errors and phonetic errors that had little influence from the L1. Those areas of vocabulary acquisition could therefore be taken into consideration when developing vocabulary exercises for L1 Finnish students. Some improvements are suggested regarding the WordDive method and application to prevent the users from producing the frequently occurring errors, such as providing more hints about the context of the word beforehand as well as classifying the answers semantically close to the target word as almost correct.