Complexity and Accuracy in Multilingual Finnish University Students’ Speech: : A Quantitative Comparison of Grammatical and Lexical Measures in L1 Finnish, L2 English, and L3 Swedish
Elo, Henna (2023-11-21)
Complexity and Accuracy in Multilingual Finnish University Students’ Speech: : A Quantitative Comparison of Grammatical and Lexical Measures in L1 Finnish, L2 English, and L3 Swedish
Elo, Henna
(21.11.2023)
Julkaisu on tekijänoikeussäännösten alainen. Teosta voi lukea ja tulostaa henkilökohtaista käyttöä varten. Käyttö kaupallisiin tarkoituksiin on kielletty.
suljettu
Julkaisun pysyvä osoite on:
https://urn.fi/URN:NBN:fi-fe20231212153490
https://urn.fi/URN:NBN:fi-fe20231212153490
Tiivistelmä
This study investigates how grammatical and lexical notions of complexity and accuracy differ in multilingual speakers’ oral production of L1 Finnish, L2 English, and L3 Swedish, and how the components of complexity and accuracy are related across the examined languages.
The data consist of oral narratives in L1 Finnish, L2 English, and L3 Swedish produced by 21 Finnish university students collected for the project MultiFluency, funded by the Swedish Cultural Foundation in Finland (2020–2023). The focus of this study is on comparing complexity and accuracy in the L1, L2, and L3 from the perspective of grammar and lexicon. The following measures of complexity are used in the analysis: mean length of AS-unit, number of different verb forms, type-token ratio, and Guiraud index for nouns. Accuracy is measured with the percentage of error-free AS-units, errors per 100 words, grammatical errors per total number of words, and lexical errors per total number of words.
Both measures of grammatical complexity revealed that L2 English was significantly more complex than L1 Finnish and L3 Swedish, and L1 Finnish was significantly more complex than L3 Swedish. The findings regarding lexical complexity were contradictory, as the highest type-token ratio was observed in L3 Swedish and the lowest in L2 English, while the Guiraud index for nouns demonstrated the greatest lexical diversity in L2 English. Regarding the accuracy measures, L1 Finnish was significantly more accurate than L2 English and L3 Swedish, and L2 English was more accurate than L3 Swedish. Correlational analyses were used to examine the extent to which complexity and accuracy were correlated within and across the complexity and accuracy measures. In total, two statistically significant correlations were found between the complexity measures and fifteen between the accuracy measures in the L1, L2, and L3. Furthermore, three statistically significant correlations were observed between complexity and accuracy in L2 English and L3 Swedish. All statistically significant correlations between complexity and accuracy showed support for the Cognition Hypothesis by Robinson (2001).
The present study delivers a pioneering contribution to the field by being the first to examine both complexity and accuracy across the same speakers in three different languages. The findings showed that this approach provides a more comprehensive representation of the relationship between complexity and accuracy across multilingual speakers. However, this multilingual approach should be further explored in future studies to consolidate the understanding of universal and language-specific trends from the cross-linguistic viewpoint.
The data consist of oral narratives in L1 Finnish, L2 English, and L3 Swedish produced by 21 Finnish university students collected for the project MultiFluency, funded by the Swedish Cultural Foundation in Finland (2020–2023). The focus of this study is on comparing complexity and accuracy in the L1, L2, and L3 from the perspective of grammar and lexicon. The following measures of complexity are used in the analysis: mean length of AS-unit, number of different verb forms, type-token ratio, and Guiraud index for nouns. Accuracy is measured with the percentage of error-free AS-units, errors per 100 words, grammatical errors per total number of words, and lexical errors per total number of words.
Both measures of grammatical complexity revealed that L2 English was significantly more complex than L1 Finnish and L3 Swedish, and L1 Finnish was significantly more complex than L3 Swedish. The findings regarding lexical complexity were contradictory, as the highest type-token ratio was observed in L3 Swedish and the lowest in L2 English, while the Guiraud index for nouns demonstrated the greatest lexical diversity in L2 English. Regarding the accuracy measures, L1 Finnish was significantly more accurate than L2 English and L3 Swedish, and L2 English was more accurate than L3 Swedish. Correlational analyses were used to examine the extent to which complexity and accuracy were correlated within and across the complexity and accuracy measures. In total, two statistically significant correlations were found between the complexity measures and fifteen between the accuracy measures in the L1, L2, and L3. Furthermore, three statistically significant correlations were observed between complexity and accuracy in L2 English and L3 Swedish. All statistically significant correlations between complexity and accuracy showed support for the Cognition Hypothesis by Robinson (2001).
The present study delivers a pioneering contribution to the field by being the first to examine both complexity and accuracy across the same speakers in three different languages. The findings showed that this approach provides a more comprehensive representation of the relationship between complexity and accuracy across multilingual speakers. However, this multilingual approach should be further explored in future studies to consolidate the understanding of universal and language-specific trends from the cross-linguistic viewpoint.