The chapter resumes the topic of the ubiquity of metonymy in lexicogrammar expounded by Langacker (this volume). It is argued for a metonymic motivation of "prototypical" and "non-prototypical" meanings of a number of constructions, among them some morphological constructions (derivation, conversion, compounding), the quantifier a lot, instances of polysemous extension, and a clausal construction (the epistemic conditional). The paper also discusses several cases of metonymy-motivated non-prototypical lexical meaning that often involve a change in grammatical behavior (e.g. the emergence of the intransitive "slimming" sense of the verb reduce). It is demonstrated that metonymy can also motivate constructional form (a number of instances are discussed in the chapter). If the set of forms of a construction is regarded as a small cognitive category where canonical forms have prototype status, then it should be subject to (some of) the same cognitive operations (among them metaphor and metonymy) as other categories. Finally, the author argues that metonymy is fundamentally inferential and that its motivational and referential roles follow from its inferential nature. (Antonio Barcelona)