Tagmemic theory and practice had its roots in the phonemic view of Sapir (1925), in the clause analysis of Bloomfield (1933), in the Fries (1952) view of the English language, and in the practical application of linguistics to the analysis of preliterate languages by Cameron Townsend (founder of the Summer Institute of Linguistics). These all had their impact on the growth of tagmemics in the context of issues in American descriptive linguistics of the mid-20th century. An initial development of tagmemic theory came from the attempt to apply to the study of grammatical analysis the principles of phonemic analytical procedures involving contrast (for native speakers), variants (often unnoticed by speakers), and the structural positions (distribution) in which the units appropriately occurred. For example, the subject of a clause contrasts with an object in terms of the meaning of subject-as-actor versus object-as-undergoer; it can have variants, in that ‘a dog’ can be replaced in the same subject slot by a phrase such as ‘the cat’; and the English subject position normally precedes the predicate.