1. INTRODUCTION This paper examines the status of the modal particle1 bi/by in three Slavic languages: Bulgarian, Macedonian and Russian, with a focus on the Russian data. The evolution of this modal particle in Slavic constitutes in many respects a canonical example of grammaticalization. A lexical item, in this case the verb to be in the aorist tense, was used with the l-participle to form an irrealis mood.2 Over time, in some of the Slavic languages, this auxiliary has lost inflectional properties and become an invariant particle. This shift from lexical item to grammatical form is accompanied also by increased restriction in syntactic position. Such changes are typical of the grammaticalization process. The comparison of Bulgarian, Macedonian and Russian data is valuable for two reasons. First, the Russian and South Slavic data illustrate some of the possible range in both rate and type of change involved in grammaticalization. Second, with this varied data before us we can address two frequently posed questions in work on grammaticalization: 1) What motivates grammaticalization?, and 2) What accounts for differences among languages in rate and type of change? This study suggests that answers to both of these questions are to be found in an appreciation of the structural differences among these three closely related languages. 2. WHAT IS GRAMMATICALIZATION? The term grammaticalization has primarily been understood to refer to a diachronic process, i.e., the process by which lexical or content items assume functions more grammatical in nature (or grammatical forms become even more so). However, much recent scholarship views grammaticalization from a synchronic perspective as well.3 While appreciating the historical nature of grammaticalization-the process-the focus in such works is on the results of this process at a given point in a language's development. Primarily synchronic studies, informed by the understanding that inherent features of the grammaticalization process such as different rates of change and less than complete applicability of a change, help us appreciate that what is evidenced synchronically at any given point may be some ambiguous or recalcitrant data. Such work underscores the essentially noncomplete nature of grammars and offers an explanation for those areas in a language's grammar which are in flux. The diachronic and synchronic approaches suggest different but interrelated models for analyzing data. The diachronic perspective is concerned primarily with how the historical process of change is conceptualized. The model usually employed is the cline of grammaticality, for example: content item > grammatical word > clitic > inflectional affix (Hopper and Traugott 1993: 7) The synchronic approach seeks to establish parameters for assessing the degree to which a given form in the contemporary language is grammaticalized. Measures to assess the degree of grammaticalization include such factors as decategorialization: lack of inflections associated with a particular category; recategorialization: reinterpretation as member of a different grammatical class; and lack of syntactic autonomy. Clearly these two analytical frameworks are linked. The cline of grammaticality posits that changes associated with the process of grammaticalization follow a path of predictable morphological developments. Synchronic measures of grammaticalization such as decategorialization and lack of syntactic autonomy can be shown to correlate with points along this historical continuum. For example, the position of a clitic is subject to greater syntactic restrictions than that of a grammatical word. This paper considers grammaticalization first and foremost from a synchronic perspective. I am not concerned so much with the evolution of the bi/by forms, but rather with their current status. There are two stages to this analysis. First I assess the degree to which the modal form bi/by is grammaticalized in contemporary Bulgarian, Macedonian, and Russian. …
Read full abstract