‘Floating signifier’: all of a sudden I started to see this term everywhere. As if this famous denomination, coined by Claude Lévi-Strauss in his Introduction to the Work of Marcel Mauss, had literally invaded the whole field of critical theory – philosophy, political theory, gender studies, race studies, all domains that have paradoxically challenged structuralism over the last fifty years. Whatever the critiques of structuralism, the concept of ‘floating signifier’ has remained continuously central in poststructuralist discourses, in all sorts of contradictory and unperceived ways.

In his Introduction, Lévi-Strauss develops his conception of language through his discussion with Mauss. Language, he says, has to be approached from two inseparable perspectives: an evolutionary and genetic one, and a linguistic and synchronic one. The first perspective is vertical. It deals with the origin and development of language through time. The second is systematic and horizontal. It studies the relations between the elements of language: signs (signifieds and signifiers), phonemes and morphemes. Levi-Strauss's theory of the floating signifier, the meaning of which I will recall in a moment, follows both lines: it has an evolutionary and a structural dimension.

I want to argue here that, instead of maintaining these two dimensions, diachronic and synchronic together, poststructuralist thinkers have only retained the second one. They have taken for granted the existence of an empty square, a void or blank space, a signifier zero in the symbolic chain of language without ever interrogating its empirical and biological origin. The floating signifier has become a ready-made, indefinite element, able to play all possible parts, a critical skeleton key whose origin, once again, has remained unquestioned.

What is this origin? What does it say about language – and poststructuralism itself? Such are the issues I intend to deal with here. First, after resituating Lévi-Strauss's argument and recalling his definition of the floating signifier, I give examples of the proliferation of the concept of ‘floating signifier’ in various theoretical contexts. Second, I return to the aforementioned double take on language – evolutionary and structural – with the help of current trends of research in neurolinguistics. I am referring in particular to the ‘microgenetic’ theory of language brought to light by the neurologist Jason Brown and analysed by the neuro-anthropologist Terence Deacon.