Bilingualism and knowledge representation

Henrik Saalbach, Roland H. Grabner

The improvement of foreign language skills is a great challenge for educational policy in all modern societies in general, and in multilingual societies such as Switzerland in particular. Therefore, multi- and bilingual education programs are currently receiving a lot of public attention and promotion.
One innovative way of fostering foreign language skills in school is immersion teaching, i.e., the teaching of non-linguistic subjects (e.g., mathematics or science) in a foreign language. Immersion programs have generally been considered a very efficient way of teaching because foreign language skills are improved as a side-effect, without additional instructional effort. However, little research has been done on the underlying cognitive mechanisms of learning in bi- or multilingual pedagogical settings. In spite of the obvious benefits and the popularity of immersion teaching, it is far from clear whether or not there are any risks for students participating in such programs.
The studies we are currently conducting examine the role of the first language and the instruction language as well as possible “cognitive costs” of bilingual learning environments.

The representation of arithmetic knowledge in bilinguals: An fMRI study

The aim of this project is to explore the neurocognitive basis of bilingual learning and to ask whether and to which extent these programs tend to generate “cognitive costs”. Previous research on the cognitive effects of specific languages in monolingual and bilingual speakers within various domains has accumulated a large body of findings which suggest that language is not only a means of communication but also a means of representing knowledge about the world. The language in which knowledge is acquired has turned out to have a considerable effect on the way information is stored, on which cognitive operations can be performed with it, and on how easily it can be accessed. Therefore, we will investigate whether the representation of learned information in the brain depends on the language in which learning took place, what difficulties arise when the language of learning and the language of application are not the same, and whether these potential difficulties depend on the kind of information.
Given our group’s expertise in research on mathematical cognition and the unequivocal importance of mathematics in school, we will investigate these issues of bilingual education, using behavioral and brain imaging methods, in the domain of arithmetic problem solving.
This research will have an impact on several issues which are widely discussed in the general public as well as in the scientific community, such as the underlying cognitive mechanisms of bilingual learning and the cognitive and neurophysiological differences between arithmetic operations. Finally, our research project represents an innovative and reasonable way of applying methods from the cognitive neurosciences to answer questions pertaining to the educational field, since our hypothesis concerning bilingual learning cannot be tested by solely relying on behavioral methods.

The relation between mathematical competence and bilingualism in Turkish-German primary school students

Mathematical competence is partly based on language-dependent modeling. Well developed language skills are therefore an important prerequisite for successful mathematics performance at school. This is supported by international comparison studies on school achievement, such as PISA or IGLU, which revealed that language deficits are a major barrier for bilingual students from immigration families with respect to their achievement in various subjects including mathematics. While these problems are commonly traced back to the lack of sufficient skills in the language of instruction, the role of first language competencies remains unclear. It has been suggested that mastery of first language skills is of great importance not only for learning a second language but also for conceptual development in general. As a consequence, deficits in the first language would result in the impairment of, both, second language learning and conceptual learning. On the other hand, it has been argued that high competencies in at least one language result in better performance by bilinguals in tasks which draw on attention control processes.

Against this theoretical background we are conducting a research project on mathematical skills of Turkish-German bilingual and German monolingual third-graders. Research questions are: first, whether and to which extent proficiency in the first language affects performance in complex mathematical problem solving processes in the second language (language of instruction); second, whether and to which extent students with low and high competencies in both languages show a cognitive disadvantage or advantage in their math performance; and, third, whether the representation of knowledge gained during instruction depends on the specific language of instruction, thus validating potential findings from the study described above on students within an actual school setting.

Collaborators

- Ilonca Hardy, Universität Frankfurt
- Sebastian Kempert, Universität Frankfurt