Polysemy or monosemy: Interpretation of the imperative and the dative-infinitive construction in Russian Egbert L.J. Fortuin Abstract: In the literature there has been much debate concerning the question of whether forms are essentially _monosemous_, that is, associated with _one_ abstract meaning, or whether such abstractions are principally underspecified; according to the latter approach, meanings are essentially _polysemous_, that is, forms are associated with different interrelated meanings. Many studies that deal with this problem are highly theoretical, and do not support their empirical claims with extensive analyses of specific empirical data. The focus on the theoretical aspect of the phenomenon of meaning leads, in some cases, to particular shortcomings. Monosemous approaches frequently leave the process of interpretation of abstract meanings unexplained, and in many cases definitions of meanings are so abstract that they also describe oppositional forms. In polysemous analyses, however, the criteria for distinguishing different uses are not always clear, and intermediate uses are often not accounted for. Moreover, polysemous analyses often fail to point at the shared features of different interrelated uses, which may stand in opposition to other forms. In this dissertation I provide further insight into the phenomenon of polysemy versus monosemy by giving a detailed analysis of the interaction between meaning and context against the background of the semantic system in which the forms occur. The expressions that I analyze are the imperative and the dative-infinitive (DI) construction in modern Russian. The main aim of these analyses is to account for the different uses of these forms/constructions. In Chapter I, I give a short introduction to the dissertation. Before analyzing the forms under discussion, in Chapter II I explain the structure of meaning by discussing how meanings are learned, and how they function in the linguistic structure. Following Bartsch (1999), I argue that in order for the linguistic structure to be stable, it is necessary that forms are associated with different concepts; however, this is only possible if different perspectives enable the language user to differentiate between them. It is also important that concepts are not overextended under a perspective, and that the use of an expression is delineated by oppositional forms. I further argue that although forms are associated with different uses, it is possible in many cases to abstract from these different uses on different levels, and to define a general meaning. The general meaning can best be seen as a _frame_ within which the different uses of an expression may occur. Such a frame cannot be seen as a definition in the strict sense, as it does not predict the possible uses of a word, but rather describes the common features of a word, which may stand in opposition to other uses. The notion of 'frame' points to two things: (i) the frame can be seen as a _restriction_ on the use of a particular form, or put differently, a restriction on the extensions of a particular form, securing stability of the linguistic system; and (ii) it is within the possibilities provided by the frame that _different uses can be distinguished_. I also argue that it is not possible to give an adequate answer the question of _when_ uses of a form can be seen as different since there are no clear and discrete criteria for distinguishing different uses. The main part of the dissertation consists of a detialed analysis of the Russian imperative and the Russian dative-infinitive construction. In Chapter III, I discuss the meaning and use of the Russian imperative. I define a basic meaning of the imperative that can be seen as an abstraction from so-called 'directive' uses and 'hortative' uses. These uses have basic uses themselves, and extensions from these basic uses by the process of selection, and possibly cancelling of features under perspectives provided by contexts. The process of extension by feature selection (backgrounding, highlighting, cancelling) occurs in different degrees (corresponding to the number of selected features), such that some instances of the imperative can be seen as borderline cases between different uses. The different uses should therefore be seen as _usage types_. These usage types correspond to _context types_. Context types are constituted by collections of formal features that correspond to clear examples of different semantic types. Although it is not possible to give a necessary and sufficient definition for all the uses of the imperative it is possible to abstract from the uses on different levels, and point at shared features of the different imperative uses, that do not occur with oppositional forms. The approach to the study of the imperative that I advocate is an intermediate position between monosemous approaches and polysemous approaches. It shares with monosemous approaches the idea that some collection of features (viz. directivity) can be seen as a necessary and sufficient condition for the correct understanding (rather than correct use) of the imperative, and it shares with polysemous approaches the idea that different uses have a more or less independent status, and that can be analyzed in terms of extensions of other uses. In Chapter IV, I discuss the dative-infinitive construction. I show how the different modal uses of the construction can be derived from its component parts, and how the distribution of the construction can be motivated by its meaning. I argue that the assignment of the dative to the infinitive predicate is always connected to an RQWLF modal meaning, that is, the realization of the infinitive situation by the dative participant is presented as something which is accordance with the normal or inevitable way things go, rather than as the result of the intention of the dative participant. More specifically, I argue that the idea of recipienthood of a situation presupposes an initial information state where the dative participant is _not_ associated with the realization of the infinitive situation (or in the case of negation, where the dative participant _is_ associated with the infinitive situation), which is then (implicitly) contradicted. I argue that the verbal or predicative element of the construction cannot be seen as a _meaning_, but must rather be seen as the _interpretation_ that is the result of the association between the non-expressed infinitive agent and the dative subject. It is therefore incorrect to posit modal logic operators, or non-expressed modal elements for the construction. In my analysis I point at the shared features between the dative-infinitive construction proper (with dative nouns or pronouns), and the occurrence of the second dative. I argue that if the second dative is analyzed as a special instance of the DI-construction, it is possible to motivate its distribution. Such an analysis provides a deeper level of understanding than syntactic analysis that do not take meaning into account in a systemantic way (more specifically analyses working within the framework of Generative Grammar, or Lexical Functional Grammar). I will argue that these analyses are characterized by the following shortcomings (i) non-motivated rules are postulated to explain the linguistic phenomenon in question; (ii) model-theoretic notions that are postulated as explanatory devices have in fact no real explanatory value, because they are partly defined in terms of the phenomena they aim to describe and explain, (iii) linguistic phenomena that are formally unified (different occurrences of the dative case) are treated as non-related phenomena, such that arbitrary distinctions between linguistic data are made, and (iv) the models do not adequately explain the occurrence of the second dative, and make the wrong predictions.