In this interview, Dr. Anna Schmitt from Magrid speaks with Dr. Mila Marinova, a cognitive psychologist at the University of Luxembourg, about her research on the intersection of language, mathematics, and cognitive development. Dr. Marinova’s work explores how children and adults acquire numerical skills in multilingual contexts, and how factors such as math anxiety, socio-cultural background, and language of instruction influence mathematical learning.
Focusing on both experimental studies and practical implications, she sheds light on the challenges and opportunities faced by multilingual learners. This conversation offers valuable insights for educators, researchers, and policymakers seeking to support equitable and effective math education in diverse classrooms.
Could you briefly describe your research interests and work at the University of Luxembourg?
a cognitive psychologist at the University of Luxembourg, I work within the Cognitive Neuroscience Unit. My research sits at the intersection of cognitive-developmental psychology and neuroscience, with a particular focus on numerical cognition—how humans learn and process numbers, and how mathematical skills develop across the lifespan.
Currently, my work is primarily experimental, exploring how numerical and mathematical competencies emerge and develop in both children and adults. I am especially interested in how these competencies are shaped by various factors, including affective influences (such as anxiety), gender, multilingualism, and the socio-cultural environment. By socio-cultural environment, I refer to influences such as parent-child interactions involving numerical content, socioeconomic status, and immigration status. More broadly, I am also interested in the neurocognitive mechanisms of learning in both typical and specific populations.
In addition to research, I wear many other hats. I am actively involved in teaching, student supervision, and the training of the next generation of scientists. I particularly enjoy teaching courses on the scientific method, experimental research, and learning to bachelor’s and master’s students. I am also deeply passionate about skill development, offering workshops on academic writing, thesis preparation, and critical thinking. My teaching extends beyond the university setting, as I regularly facilitate critical thinking workshops for doctoral candidates, medical professionals, adolescents, and the general public, in both formal and informal contexts.
Furthermore, I am deeply committed to public outreach, striving to make scientific knowledge accessible to wider audiences. As an active science communicator, I engage with the public through my Instagram account (@dr.mila.marinova) and my website (milamarinova.com), where I focus on topics such as critical thinking, decision-making, disinformation, and scientific literacy. Through my communication work, I aim to integrate insights from psychology and neuroscience to make research findings understandable and actionable, ultimately supporting informed decision-making and public engagement with science.
Could you provide a summary of your latest publications related to bilingualism and mathematics?
First, I would like to briefly address the broader relationship between language and mathematics learning and performance, as this context is essential for understanding my research (for an overview, see Schiltz et al., 2024).
Although it is a common myth that language and mathematics are unrelated, current research—including my own—demonstrates a strong and intricate connection between these domains. Numerical systems are, in fact, linguistic systems, and as such, our language skills are deeply involved when we learn and use mathematical concepts. This connection becomes especially evident in multilingual individuals whose experiences with mathematics are shaped also by the linguistic context in which they are learning.
Multilingualism affects various numerical learning processes, from learning to count to understanding complex mathematical problems. This influence is particularly pronounced for individuals whose language of instruction at school differs from their mother tongue (e.g., Greisen et al., 2021). Our research at the University of Luxembourg, consistent with international findings, shows that such individuals often face additional challenges compared to peers who learn in their first language.
However, multilingualism should not be viewed as a disadvantage in itself—in fact, being multilingual carries cognitive and practical benefits. The key lies in understanding how to navigate these dynamics effectively in educational contexts so that multilingualism can be a resource rather than a barrier.
The country of Luxembourg and its multilingual educational system is a very interesting case and the “perfect laboratory” for someone like me who studies the interaction between language and mathematics. Just to remind the readers that in Luxembourg’s educational system, the language of formal instruction (LI1) is German in primary school, and then these learners switch to French in high school (LI2). This multilingual educational system possesses many practical advantages and certain challenges for learners. In my own research on multilingualism and mathematics, I focus on three main areas:
First, I study how multilingualism impacts the early stages of numerical concept learning in young children:
Our ongoing studies suggest that second-language learners may face disadvantages when acquiring basic numerical skills, such as counting. Since counting forms the foundation for later mathematical understanding, difficulties in this area can cascade into more advanced learning challenges. We are currently investigating how these early struggles affect the development of more advanced numerical concepts such as succession (that for each natural number n, there is a successor n+1), the understanding of infinity, and how formal and informal learning experiences can influence their early stages of numerical concept development (Marinova & Schiltz, 2024; see also Cheung et al., 2025; Schneider et al., 2020).
A second line of research focuses on the role of affective factors, such as anxiety and attitudes, in math learning among mono- and multilingual individuals. From studies in mono-linguals (including studies where the language status has not been specifically assessed), we know that high levels of anxiety and specifically math anxiety can negatively influence math performance (Vos et al., 2023; Marinova & Vos, 2024).
In multilingual individuals, this topic is not very well understood, but our preliminary data suggest that second-language learners, especially at the start of secondary school, tend to experience higher levels of math anxiety and math avoidance. Interestingly, this anxiety seems to decrease over time, likely as students gain confidence and familiarity with mathematical concepts in their second language. In adults, we observe clear language preferences for different types of tasks. For example, in one study, university students expressed a preference for taking a general knowledge exam in German, but would opt for French when taking a math exam—highlighting that the language of mathematics holds special significance for multilingual individuals.
Finally, I also study how basic numerical concepts are represented in the multilingual brain:
Although this line of research is still ongoing, preliminary findings, combined with previous work in our lab (e.g., Van Rinsveld et al., 2016;2017), suggest that numbers and mathematical operations may be represented and processed differently depending on the language used. This points to deep, brain-level interactions between language and mathematics, implying that whether a person solves a math problem in their first or second language could influence not only performance but also the underlying cognitive processes.
It is important to note that many of these findings are still preliminary and in the process of peer review, so they should be interpreted cautiously. Nonetheless, they open important avenues for understanding how multilingualism interacts with mathematical thinking and how educational systems might better support multilingual learners.
Regarding the publications related to Magrid*: Do you notice any similarities or any differences?
I see many meaningful parallels between my works and Magrid’s scientific work and its mission as an educational technology. In both cases, there is a clear recognition that language—particularly for non-native speakers—can present significant challenges to the acquisition of early number skills. This understanding is crucial, as language barriers can often hinder children’s early numeracy development.
At the same time, visual-spatial skills play a critical role in the development of early numeracy, not only in typically developing children but also in those from specific or vulnerable populations. I believe Magrid’s approach—focusing on strengthening early numerical understanding through visual-spatial skills—is a particularly powerful and effective strategy. By reducing reliance on language, this method can help mitigate educational inequalities and provide more equitable access to foundational math skills.
Moreover, once these numeracy skills are established through visual-spatial training, they can be complemented and further enhanced through targeted language-based interventions, creating a comprehensive and inclusive framework for supporting children’s mathematical development.
How can your studies have a concrete impact on mathematics learning among young children?
My work examines the critical role language plays in mathematical learning and performance and the fact that, for multilingual individuals, this relationship is shaped by a complex interplay of cognitive, emotional, and contextual factors. Understanding these dynamics is crucial for creating inclusive and effective educational practices that recognize multilingual learners’ challenges and strengths. However, studying how multilingualism impacts early cognitive development is still a relatively new field because cultural and language diversity is relatively new in human history. In addition, studying multilingual populations is a big challenge because the profiles are very unique and very diverse. Yet, we already do know some things, which I believe can make a difference in our educational systems.
One of the immediate and impactful steps I believe we can take is raising teachers’ awareness about the critical role language plays in mathematics learning. Teachers carry immense responsibilities and manage numerous tasks daily, but it is still a widespread misconception—in Luxembourg and globally—that mathematics is a “language-free” subject. In reality, language is the foundation for learning mathematics. Understanding and expressing mathematical ideas, reasoning through problems, and interpreting tasks are all mediated by language. Promoting dialogue and collaboration with teachers to make this connection explicit would be a crucial step forward.
Furthermore, recognising and embracing linguistic and cultural diversity within the classroom can foster a more inclusive and supportive learning environment. One important change could be approaching the language of instruction as a second (or even third) language for many children rather than assuming full proficiency from the outset. Teaching it formally as a foreign language could significantly support multilingual learners.
At the higher policy level, evidence-based educational practices and policies are essential for meaningful and sustainable change. I believe that research—like my own—can help inform these practices. Ultimately, I am optimistic that, as research continues to inform practice and policy, we will see progressive improvements in how mathematics is taught and learned, particularly for children navigating the challenges of multilingual education.
Would you like to add any insights related to mathematics and learning disorders in a bilingual context?
Yes, I believe this is a vital and often overlooked area when considering mathematics learning, especially in multilingual contexts. Developmental dyscalculia, or specific learning disorder in mathematics, is a neurodevelopmental disorder characterised by persistent difficulties in acquiring basic numerical and mathematical skills. These difficulties can have profound consequences, affecting academic achievement and an individual’s broader personal and professional life if left unaddressed.
However, when it comes to bilingual and multilingual learners, there is still a significant lack of research specifically focused on how learning disorders like dyscalculia present and manifest in these populations (see Schiltz et al., 2024, p.2424). This presents a real challenge, as multilingualism introduces unique cognitive and linguistic dynamics that can both mask and mimic mathematical learning difficulties
For instance, studies have shown that higher bilingual proficiency may support mathematics performance in children with mathematical difficulties, potentially through higher working memory (Swanson et al., 2018). Yet, mathematical vocabulary can be a specific point of struggle for second-language learners with math difficulties, especially in arithmetic tasks (e.g., Powell et al., 2020). This indicates that bilingualism has a particular, but not very well understood, implication for children with math difficulties.
One major concern is that language proficiency plays a critical role in how children perform on diagnostic assessments, many of which are language-heavy and designed for native speakers. As a result, there is a real risk of both over- and under-diagnosis of dyscalculia in multilingual learners, depending on whether their language challenges are mistaken for math difficulties or vice versa. For example, the data from Luxembourg’s national school monitoring program show that if we apply the same diagnostic cut-offs for all children, a disproportionately high number of second-language speakers would be classified as having math difficulties, while some first-language speakers who struggle might be missed (Martini et al., 2021). This highlights the urgent need for diagnostic tools that take into account children’s language backgrounds, with adjusted norms that reflect the linguistic diversity of the classroom.
Additionally, dyscalculic children often show weaker language skills in general, compared to typically developing peers, meaning that multilingual children with dyscalculia may be twice disadvantaged—both in terms of language and mathematics. There is a pressing need to develop and implement assessments that are linguistically sensitive and equitable, and promising steps are being made in this direction by my colleagues at the Luxembourg Center of Educational Testing (LUCET) and the Cognitive Neuroscience Unit, who have recently developed a diagnostic test for mathematical learning impairments in multilingual contexts (Hilger et al., 2024) .
References:
Cheung, P., Bourque, T., Ang, D., & Merkley, R. (2025). Early number acquisition in bilingual children. Cognitive Development, 73, 101541.https://doi.org/10.1016/j.cogdev.2024.101541
Greisen, M., Georges, C., Hornung, C., Sonnleitner, P., & Schiltz, C. (2021). Learning mathematics with shackles: How lower reading comprehension in the language of mathematics instruction accounts for lower mathematics achievement in speakers of different home languages. Acta Psychologica, 221, 103456. https://doi.org/10.1016/j.actpsy.2021.103456Get rights and content
Hilger, V., Ugen, S., Romanovska, L., & Schiltz, C. (2024). Diagnose spezifischer Lernstörung im Bereich Mathematik in einem multilingualen Bildungskontext. Nationaler Bildungsbericht Luxemburg .
Marinova, M. & Vos, H. (2024). Minding the gap in primary school: examining the interplay between working memory, math anxiety, spatial anxiety, and mathematical ability. (Poster Presentation) available at https://orbilu.uni.lu/handle/10993/60808
Marinova, M. & Schiltz, C. (2024). Learning to count in a multilingual environment. (Poster Presentation) available at https://orbilu.uni.lu/handle/10993/64509
Martini, S., Schiltz, C., Fischbach, A., & Ugen, S. (2021). Identifying math and reading difficulties of multilingual children: Effects of different cut-offs and reference groups. In A. Fritz, E. Gürsoy, & M. Herzog (Eds.), Diversity Dimensions in Mathematics and Language Learning (pp. 200–228). https://doi. org/10.1515/9783110661941-011
Powell, S. R., Berry, K. A., & Tran, L. M. (2020). Performance differences on a measure of mathematics vocabulary for English learners and non-english learners with and without mathematics difficulty. Reading & Writing Quarterly, 36(2), 124–141. https:// doi.org/10.1080/10573569.2019.1677538
Schiltz, C., Lachelin, R., Hilger, V., & Marinova, M. (2024). Thinking about numbers in different tongues: An overview of the influences of multilingualism on numerical and mathematical competencies. Psychological Research, 1-16. https://doi.org/10.1007/s00426-024-01997-y
Schneider, R. M., Sullivan, J., Marušič, F., Biswas, P., Mišmaš, P., Plesničar, V., & Barner, D. (2020). Do children use language structure to discover the recursive rules of counting?. Cognitive psychology, 117, 101263.https://doi.org/10.1016/j.cogpsych.2019.101263
Swanson, H. L., Kong, J., & Petcu, S. (2018). Math difficulties and working memory growth in English language learner children: does bilingual proficiency play a significant role? Language, Speech & Hearing Services in Schools, 49(3), 379–394. https:// doi.org/10.1044/2018_LSHSS-17-0098
Van Rinsveld, A., Schiltz, C., Brunner, M., Landerl, K., & Ugen, S. (2016). Solving arithmetic problems in first and second language: Does the language context matter? Learning and Instruction, 42, 72–82. https://doi.org/10.1016/j.learninstruc.2016.01.003
Van Rinsveld, A., Dricot, L., Guillaume, M., Rossion, B., & Schiltz, C. (2017). Mental arithmetic in the bilingual brain: Language matters. Neuropsychologia, 101, 17–29. https://doi.org/10.1016/j. neuropsychologia.2017.05.009
Vos, H., Marinova, M., De Léon, S. C., Sasanguie, D., & Reynvoet, B. (2023). Gender differences in young adults’ mathematical performance: Examining the contribution of working memory, math anxiety and gender-related stereotypes. Learning and Individual Differences, 102, 102255. https://doi.org/10.1016/j.lindif.2022.102255
About Magrid:
*Cornu, V., Schiltz, C., Pazouki, T., & Martin, R. (2017b). Training early visuo-spatial abilities : A controlled classroom-based intervention study. Applied Developmental Science, 23(1), 1‑21. https://doi.org/10.1080/10888691.2016.1276835
Jung, S., Meinhardt, A., Braeuning, D., Roesch, S., Cornu, V., Pazouki, T., Schiltz, C., Lonnemann, J., & Moeller, K. (2020c). Hierarchical Development of Early Visual-Spatial Abilities – A Taxonomy Based Assessment Using the MaGrid App. Frontiers In Psychology, 11. https://doi.org/10.3389/fpsyg.2020.00871
Pazouki, T., Cornu, V., Sonnleitner, P., Schiltz, C., Fischbach, A., & Martin, R. (2018d). MaGrid : A Language-Neutral Early Mathematical Training and Learning Application. International Journal Of Emerging Technologies In Learning (iJET), 13(08), 4. https://doi.org/10.
