Grammaticalization is a well-attested process of linguistic change in which a lexical item becomes a function word, which may be further reduced to a clitic or affix. Proponents of the universality of grammaticalization have usually argued that it is unidirectional and have thus found it auseful tool in linguistic reconstruction. In this book Prof Norde shows that change is reversible on all levels: semantic, morphological, syntactic, and phonological. As a consequence, the alleged unidirectionality of grammaticalization is not a reliable reconstructional tool, even ifdegrammaticalization is a rare phenomenon. Degrammaticalization, she argues, is essentially different from grammaticalization: it usually comprises a single change, examples being shifts from affix to clitic, or from function word to lexical item. And where grammaticalization can be seen as a process, degrammaticalization is often theby-product of other changes. Nevertheless, she shows that it can be described, like grammaticalization, in a principled way, in order to establish whether a change in a word has been from more to less grammatical or vice versa, and the stages by which it has become so. Using data from differentlanguages she constructs a typology of degrammaticalization changes. She explains why degrammaticalization is so rare and why some linguists have such strongly negative feelings about the possibility of its existence. She adds to the understanding of grammaticalization and makes a significantcontribution to methods of linguistic reconstruction and the study of language change. She writes clearly, aiming to be understood by advanced undergraduate students as well as appealing to scholars and graduate researchers in historical linguistics.