In general, it's far from useless to investigate the nature of X before having a clear definition of X. Perhaps the definition will come as a result of the investigation. Perhaps there will never be a clear definition, but paradigm cases are easy to come by (that's arguably the case with free will).