Seminar on L2 Vocabulary Development

Rosa Salmela - 10:15 - 12:00, Room- 9-1017

Vocabulary Assessment in Finnish and Swedish

Vocabulary knowledge refers to both the number of words a person knows and the depth of their knowledge of those words. It is also closely related to general language proficiency. Short, validated tools for measuring vocabulary knowledge are therefore useful in both research and applied settings. Tests based on the visual lexical decision task (VLD) already exist in several languages, including English and French (Brysbaert, 2020; Lemhöfer & Broersma, 2012). In our study, we developed a novel online test, Lexize, based on VLD, to measure the vocabulary knowledge of adult L2 learners of Finnish. The Lexize test proved effective in distinguishing between L1 and L2 speakers. Among L2 learners, scores correlated strongly with self- assessed proficiency and language exposure. In native speakers, test scores correlated with self-ratings, Finnish school grades, and age, and higher Lexize scores were also associated with higher education levels. The results suggest that Lexize is a useful tool to assess Finnish vocabulary knowledge for non-native speakers and to some extent, for native speakers. Lexize has also been adapted into a developmental version, demonstrating that in children, vocabulary knowledge varies by grade, gender, and home language (Bertram et al., 2025). Currently, a Swedish-language version is under development. In Lexize‐SWE, we focus specifically on the role of exposure, which may vary considerably across individuals in Finland, where Swedish is a minority language. Both the Finnish and Swedish versions of the Lexize test are freely available athttps://lexize.org/.

Lars Bokander & Anna Ingves - 13:15 - 15:00, Room- 7-0043

Can Vocabulary Be Tested with Non-Existent Words? Constructing and Using Pseudowords in Vocabulary Measurement

Imaginary words that resemble real words in the target language have long been used in vocabulary testing. However, this practice still faces multiple methodological and practical issues. Over the years, vocabulary researchers have experimented with constructing pseudowords in different languages, studied the effects of using pseudowords in vocabulary tests aimed at first- and second-language speakers, and debated the validity and reliability of such tests. In recent decades, algorithm-generated pseudowords have become more widely available (e.g. Keulers & Brysbaert, 2015). Nevertheless, creating pseudowords for smaller languages continues to rely on the researcher’s knowledge of the target language. In our research into vocabulary testing in Swedish (Bokander, 2021; Ingves, 2024), we encountered various issues regarding the construction and usage of pseudowords. Swedish, being a smaller language, is not included in previously available tools for pseudoword construction. Language-independent pseudoword tools have only recently been published (König et al., 2020; New, 2024), but their practical usefulness in vocabulary testing for individual languages has yet to be investigated. According to König et al., it is now possible to construct new Swedish pseudowords based on just 100 Swedish words and evaluate existing Swedish pseudowords. In the upcoming talk, we will provide an overview of previous research into pseudoword construction and usage in vocabulary testing and the tools available for pseudoword construction. We will also discuss the role of pseudowords in our own work with Swedish vocabulary tests.