This work reconsiders the well known problem of the uniqueness condition attached to definite descriptions, which has been subject to many a controversy since Russell (1905). A first part examines the different arguments against the claim of a strong uniqueness, trying to make explicit the underlying theoretical hypothesis, mainly those about the representation of meaning and its relationship with reference. A second part will show that, moreover, the above mentioned arguments are valid only because they do not take into account some phenomena that show the existence of effective linguistic mechanisms and advocate for a strong uniqueness. A different approach is considered in the third part, and applied to five types of phenomena.