Universals of language have been studied extensively for the last four decades, allowing fundamental insight into the principles and general properties of human language. Only incidentally have researchers looked at the other end of the scale. And even when they did, they mostly just noted peculiar facts as ''quirks'' or ''unusual behavior'', without making too much of an effort at explaining them beyond calling them ''exceptions'' to various rules or generalizations. Rarissima and rara, features and properties found only in one or very few languages, tell us as much about the capacities and limits of human language(s) as do universals. Explaining the existence of such rare phenomena on the one hand, and the fact of their rareness or uniqueness on the other, is a reasonable and interesting challenge to any theory of how human language works. The present volume for the first time compiles selected papers on the study of rare linguistic features from various fields of linguistics and from a wide range of languages.

The papers in this book describe and analyze rara in individual languages, covering an extraordinarily broad geographic distribution, including papers about languages from all over the globe. The range of theoretical subjects discussed shows an enormous breadth, ranging from phonology through word formation, lexical semantics to syntax and even some sociolinguistics.

This book investigates paradigms of person - both independent pronouns as well as bound person marking. Based on empirical and theoretical grounds, the author argues that the notion 'number' has to be redefined to deal with the cross-linguistic variation of person marking. Equipped with a new definition, a typology of the paradigmatic structure of person marking is presented, incorporating data from around 400 languages. Nothing appears to be impossible for the paradigmatic structure, although some patterns are clearly more probable than others are. Starting from the more commonly occurring patterns, the diachronic dynamics of paradigmatic structure are investigated by comparing close relatives that differ slightly in the structure of their person paradigms.