The linguistic structure of number words can influence performance in basic numerical tasks such as mental calculation, magnitude comparison, and transcoding. Especially the presence of ten-unit inversion in number words seems to affect number processing. Thus, at the beginning of formal math education, young children speaking inverted languages tend to make relatively more errors in transcoding. However, it remains unknown whether and how inversion affects transcoding in older children and adults. Here we addressed this question by assessing two-digit number transcoding in adults and fourth graders speaking French and German, that is, using non-inverted and inverted number words, respectively. We developed a novel transcoding paradigm during which participants listened to two-digit numbers and identified the heard number among four Arabic numbers. Critically, the order of appearance of units and tens in Arabic numbers was manipulated mimicking the “units-first” and “tens-first” order of German and French. In a third “simultaneous” condition, tens and units appeared at the same time in an ecological manner. Although language did not affect overall transcoding speed in adults, we observed that German-speaking fourth graders were globally slower than their French-speaking peers, including in the “simultaneous” condition. Moreover, French-speaking children were faster in transcoding when the order of digit appearance was congruent with their number-word system (i.e., “tens-first” condition) while German-speaking children appeared to be similarly fast in the “units-first” and “tens-first” conditions. These findings indicate that inverted languages still impose a cognitive cost on number transcoding in fourth graders, which seems to disappear by adulthood. They underline the importance of language in numerical cognition and suggest that language should be taken into account during mathematics education.
The Contribution of Numerical Magnitude Comparison and Phonological Processing to Individual Differences in Fourth Graders’ Multiplication Fact Ability
