This work presents a hypothesis concerning the motives of the difficulties faced by contemporary phonological theories – including those that adopt a dynamic theoretical primitive – in setting theirselves free from the atomic, static and discrete conception of sign. For such, a retrospection of this conception’s antecedents and its relations with the development of the ‘sounds of speech’ sciences was attempted. At the end of the work, the conceivable consequences of adopting a dynamic theoretical primitive in writing systems studies were raised.