My thesis is an attempt to give a minimalist account of word formation in Swedish. (See Chomsky (1993) and (1994).) In the main part of my thesis I demonstrate how the proposed system accounts for the central aspects of Swedish word formation. Simple words, compounds, and derivations alike are described as formed by Generalized Transformations, primarily by Merge, but also by Move. From a lexicon containing all morphemes in a language, inflection and a lexical stem may be enumerated. Merge unites the two, forming a minimal projection, of which inflection is the head and the lexical stem the complement. The lexical stem is argued to be devoid of word class features, these being solely a function of the inflectional part. Since Merge is free and operating at no cost, the operation may apply recursively. Hence compounding is analysed in terms of subsequent adjunction to a word or a word-like category. If the lefthand segment is itself a compound the presence of a linking element is called for. A linking element is analysed as inflection devoid of features needing checking. The role of the linking ele-ment is to create asymmetry inside words. It is suggested that compounding in a language lacking linking elements is restricted to a certain extent. Derivational suffixes are analysed as elements binding a theta-role related to the host (or rather imposed on the host). Suffixation also includes control of theta-roles, which may be assigned by the derived word to complements, in some cases with the aid of "substitute predicates", such as the preposition av and / or the possessive marker. Derivational prefixes are assumed to be either scope-taking elements or elements with a binding frame, just like suffixes. The major difference between suffixes and prefixes is that prefixes are unable to assign theta-roles. The different major word classes are assumed to correspond to different major ontological categories. The result of the merging of a stem with nominal inflection is interpreted as a Thing, and the result of the merging of a lexical stem with verbal inflection is interpreted as an Event. The thesis also discusses the universality of word formation rules, i.e. in what respect languages may differ with respect to how words are formed.