Do smart pupils study classical languages (CLs), or does studying CLs make pupils smart? Anglo-Saxon research suggests that Latin pupils surpass their Latin-less peers in L1 ability c.q. English, and that studying Latin possibly has a positive effect on mathematical skills as well. However, it has not been settled whether these results are due to preselectivity (pupil influx) or cognitive transfer (one task affecting another). While a scientific answer is still lacking, the educational value of the CLs continues to be debated across Europe − not least in Flanders, where tens of thousands of youngsters study them in secondary school. By means of a longitudinal observational study I will assess transfer in the areas of Dutch L1 ability, L2 and mathematics achievement, and general cognition. No earlier Flemish study investigated cognitive effects of CL learning following pupils across years. Preselectivity will be ruled out by matching the comparison group on IQ. I will collect and compare data from schools that use distinct teaching methods, which likely influences the degree of transfer. In addition to the main study, the characteristics of CL learning suspected to cause transfer will be simulated in a short-term artificial grammar learning experiment. Apart from its applied importance in Flemish education, this research addresses the lack of didactic research on the CLs in the Low Countries and makes a contribution to the academic discussion on cognitive transfer.