In the lead article to the volume, Ronald Langacker makes another important contribution to the debate on the nature of grammar. He challenges the prevalent view in structuralist and formal (generative) linguistics that linguistic elements are discrete and combined in well-defined ways to yield more complex structures. This presumption of discreteness and determinacy of grammar, Langacker argues, is undermined by research in Cognitive Linguistics. For example, there exist no clear boundaries between linguistic meaning and encyclopedic knowledge (two domains that are usually kept strictly apart in formalist models). Also, the ubiquity of active zones (as in My cat bit your dog, where strictly speaking it is not the cat but the cat's teeth that can bite (part of) the dog), metonymy, metaphor, and conceptual blending speak against the idea that grammar is discrete and determinate. Note that Langacker uses the term metonymy in a narrow sense and a broad sense here. When talking about the nature of grammar, what he has in mind is the wide sense of metonymy as a property characterizing grammar in general. In his contribution, he argues that grammatical relationships between two elements are not precisely determined. Langacker concludes that a grammatical relationship between two elements can be established on the basis of the mental access they provide to the entities needed for conceptual integration. A case in point is the relationship between possessor and possessed in an expression such as Mary's book, which without further context, remains indeterminate as to what kind of relationship obtains between Mary and the book. In other words, the exact relationship between Mary and book has to be metonymically inferred. (Ronald W. Langacker)