While the defining cases of diglossia offered in Charles Ferguson's 1959 article have long been useful as vehicles for introducing this important form of societal multilingualism, they are also problematic in that they differ from each other in a number of significant ways. This article proposes a modified and more precise framework in which a distinction is made between three sub-categories of diglossia. The first, "traditional diglossia" consists of diglossia in pre-modern societies where a sacred language serves as H for an entire civilisation, and it is proposed that Classical Chinese, Sanskrit, Latin and Classical Arabic be viewed as prototypes of this category. This category is closer to Ferguson's original description, and should be viewed as the most typical manifestation of diglossia. A second category, "revived diglossia", includes cases in which ancient languages were revived by communities in colonial contexts for the purpose of buttressing incipient nationalism. A third and much smaller category, "modern diglossia" is made up of cases of speech communities in modern societies where a modern standard language functions as the H variety. Distinguishing between these sub-categories will allow scholars to make stronger and more precise generalisations about diglossia. (Contains 9 notes and 1 table.)