All different - all equal? Towards cross-language benchmarking using samples of oral production in French, German and Italian