Classification of Author and/or Genre? The Impact of Word Length

190 Russian texts – letters and poems by three different authors – are analyzed as to their word length. The basic question concerns the quantitative classification of these texts as to authorship or as to text sort. By way of multivariate analyses it is shown that word length is a characteristic of genre, rather than of authorship. 1 Word Length and the Quantitative Description of Text(s) and Author(s) This study focuses on word length. Word length is a central characteristic in the framework of quantitatively oriented linguistics. In fact, the study of word length can be traced back to a hundred year long tradition (as to a historical and methodological survey of these studies, cf. Grzybek 2004). Knowing this historical background, it is evident that word length, as it is studied today, is no isolated characteristic. The basic question of the present study is to what degree word length may contribute to the discrimination of authors and genres. An answer to this question will not only shed light on specific factors influencing word length; it will also provide an argument if word length is an appropriate variable to describe an author’s individual style, or the stylistic traits of specific genres. The discussion of these questions has a history of its own: as opposed to the field of quantitative typology of texts (cf. Alekseev 1988, Pieper 1979), approaches in the realm of stylometry (cf. Martynenko 1988) assume that the individual style of texts and/or authors can be quantitatively described. Part of this research has concentrated on the question of authorship attribution, particularly applying quantitative methods to decide doubtful cases of authorship (cf. Marusenko 1990). In a way, these approaches have paved the 1 This study has been conducted in context of research project # 15485 («Word Length Frequencies in Slavic Texts»), financially supported by the Austrian Research Fund (FWF); cf.: http://www-gewi.uni-graz.at/quanta. 2 Within a synergetic approach, word length is closely interrelated with other linguistic levels and units, and it is well known that word length interacts, e.g., with the number of phonemes (in a given inventory), with lexicon size (cf. Köhler 1986), with polysemy (cf. Altmann et al. 1982), or word length and word frequency (Strauss et al. 2004, with a survey of the Zipfian tradition).