Cross-linguistic influence in third language acquisition: Sentence processing and comprehension by child heritage bilinguals

Projektinformationen

TomRis

ID Finanzhilfevereinbarung: 101107162

DOI

10.3030/101107162

Projekt abgeschlossen

EK-Unterschriftsdatum 24 April 2023

Startdatum 1 September 2023

Enddatum 31 Oktober 2025

Finanziert unter

• Marie Skłodowska-Curie Actions (MSCA)

Gesamtkosten

Keine Daten

EU-Beitrag

€ 226 751,04

Koordiniert durch

UNIVERSITETET I TROMSOE - NORGES ARKTISKE UNIVERSITET
Norway

Periodic Reporting for period 1 - TomRis (Cross-linguistic influence in third language acquisition: Sentence processing and comprehension by child heritage bilinguals)

Berichtszeitraum: 2023-09-01 bis 2025-10-31

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

Understanding how children learn multiple languages is essential for creating fair and effective education systems in Europe. European societies are increasingly multilingual, with a growing number of children acquiring more than one language from an early age. Many children use their native language at home, receive education in a different societal language, and acquire additional languages in school. Despite this diversity, educational systems largely rely on monolingual norms when assessing language development and academic performance. These conditions can lead to misunderstandings about multilingual learners’ abilities and result in unequal educational opportunities.
The TomRis project addresses this challenge by investigating how multilingual children process and use an additional language, focusing on children exposed to a home language. TomRis examined how different linguistic systems influence each other during comprehension and production. The project aimed to generate evidence that reflects the real experiences of multilingual learners.
Situated at the intersection of linguistics, psychology, and education, TomRis contributes to the social sciences and humanities by offering insights into online language processing, learning, and variation in multilingual contexts. These insights are directly relevant to European and national priorities, which emphasise inclusion, equal access to education, and recognition of linguistic diversity. Empirically grounded knowledge from TomRis supports equitable assessment practices and educational approaches that reflect the realities of multilingual classrooms. The project contributes to long-term goals in EU and Norwegian language and education policy, including social cohesion, educational equity, and informed decision-making.

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

TomRis investigated moment-to-moment sentence comprehension in child heritage speakers—children exposed to a minority language (Turkish), acquiring the societal language (Norwegian), and learning English as a third language. By focusing on online language processing in these bilingual and multilingual children, the project goes beyond previous research that relied mainly on adult data or offline measures.
The study combined eye-tracking, grammaticality judgement (GJT), and story-telling production tasks to examine how home and societal languages influence third language acquisition. Eye-tracking measured reading times and gaze patterns while children processed morpho-syntactic structures such as definite articles, modifier–noun agreement, topicalisation, and adverb placement. Statistical models were used to analyse eye-tracking data, and GJT performance was examined using both overall accuracy and signal detection measures, allowing reliable comparisons across participants and sentence types. The story-telling task captured naturalistic language use. Language proficiency in English, Norwegian, and Turkish was included as a covariate to ensure that observed effects reflected cross-linguistic influence rather than general language ability.
Preliminary findings show that heritage children display distinct processing patterns compared to monolingual peers, reflecting flexible use of cues from both their home and societal languages. Results from eye-tracking and grammaticality judgements revealed property-specific cross-linguistic influence. Eye-movement measures showed that children were sensitive to grammatical violations during reading, while judgement data indicated that competing cues from their first and second languages influenced how confidently and consistently they evaluated sentence correctness.

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

By examining online language processing in child heritage speakers, TomRis advances beyond previous research that has largely relied on adult participants or offline measures. Heritage Turkish participants showed intermediate processing across grammatical properties. For definite articles and quantifier–noun agreement, eye-movement measures such as gaze and regression-path durations were slower than those of Turkish monolinguals but closely aligned with Norwegian peers, indicating early sensitivity to morpho-syntactic violations during reading.
For adverb placement and topicalisation, intermediate reading times reflected competition between structural cues in the first and second languages. Grammaticality judgement and signal detection analyses confirmed this pattern offline: reduced sensitivity and a more liberal response tendency were observed for adverb placement and topicalisation, whereas performance on definite articles and quantifier–noun agreement was more stable. Together, these findings show that heritage speakers can detect grammatical violations during online processing, while offline judgements are more strongly shaped by competing cross-linguistic cues.
Scientifically, these results contribute to models of multilingual processing and heritage grammar by demonstrating how cues from the first and second languages are weighted during real-time comprehension. Educationally, they inform teacher training and curriculum development, supporting effective multilingual education strategies. By publishing results in open-access journals and depositing anonymised datasets in FAIR-compliant repositories, the project ensures transparency, reproducibility, and broad scientific reuse, facilitating further research and long-term impact.

Figure 1 illustrate the predicted probabilities of correct responses for both grammatical and ungram