2025 Theses Doctoral

Dual number systems within and between languages: Cognitive processing in simple and complex numerical tasks using behavioral and electrophysiological methods

Oh, Elizabeth Dayoung

This research explores how language influences arithmetic processing, focusing specifically on how people's experiences with different number systems affect numerical cognition. Two studies compared Korean adults, who speak one language but use two distinct number systems—Sino-Korean (SK) and Native Korean (NK)—with Mandarin–English bilingual adults, who use two separate languages for math. Each group had extensive arithmetic experience primarily with one number system. Korean adults used SK numbers for formal math and rarely used NK numbers in arithmetic contexts.

Similarly, bilingual adults mostly learned and practiced arithmetic in Mandarin, starting from preschool through high school, giving them more experience with Mandarin numbers compared to English. Study 1 explored how numerical experience influences basic numerical processing using a parity-judgment task combined with brain responses (ERPs). Participants included 15 Mandarin–English bilingual adults and 15 Korean adults. Their task was to quickly decide whether numbers were odd or even while also occasionally switching between number systems—English and Mandarin for bilinguals, and Sino-Korean (SK) and Native Korean (NK) for Koreans.

Both groups performed similarly overall, but interestingly, participants made more errors when switching from their less-preferred number system (English or NK) to their preferred one (Mandarin or SK). The ERP analysis focused on two processes: attention (measured by the P2 response) and inhibitory control (measured by the N2 response). Both groups showed similar increases in the P2 response when switching number systems or detecting changes in numerical parity, suggesting they had comparable attentional processing. However, bilingual participants had shorter P2 latencies than Korean speakers, indicating they processed attention-related aspects of the task more quickly or efficiently. Additionally, bilingual adults showed stronger N2 responses during switching, indicating higher demands on inhibitory control. Korean adults, by contrast, showed smaller N2 responses, suggesting less inhibitory effort, likely because they were switching number systems within one language rather than across two languages.

Study 2 moved beyond simple parity judgments to examine more complex arithmetic tasks—addition, subtraction, and multiplication. Korean adults performed these tasks using two number systems (SK and NK), while bilingual adults used two different languages (Mandarin and English). For Korean adults, addition and subtraction showed no differences in reaction times between SK and NK. Interestingly, multiplication was different: Korean adults solved multiplication problems more slowly (by about 14%) when using their more familiar SK system. In contrast, bilingual adults showed clear reaction-time advantages when solving addition and subtraction problems in Mandarin, their preferred math language. This suggests that arithmetic retrieval depends strongly on the language in which arithmetic was initially learned, rather than just on general numerical experience or familiarity. However, multiplication tasks showed no reaction-time differences between Mandarin and English, suggesting that more complex arithmetic might rely on procedural strategies that reduce language-specific advantages.

Overall, these findings suggest that using two languages requires greater inhibitory control effort, even when behavioral performance appears similar between bilinguals (two languages) and Korean speakers (two number systems within one language). For addition and subtraction tasks—which mainly involve procedural—familiarity alone did not create differences in performance. Korean adults performed equally well using both their familiar (SK) and less familiar (NK) number systems, while bilingual adults solved problems significantly faster in their preferred, first-learned language (L1). For multiplication, however, tasks depend more heavily on memorized facts and verbal memory. Here, the advantages of language familiarity or initial language learning faded, and both groups performed similarly.