Methodological issues in exploring task-based speech production processes