Viewing the Verb Classifier Hypothesis from a noun classifier language: a contrastive analysis of verbal aspects in Russian and Japanese

Japanese belongs to the numeral classifier type of languages (Aikhenvald, 2019), where default bare nouns cannot directly combine with number words, similar to mass nouns in languages with grammatical number. To enumerate noun referents, they must be individualized and thus made countable by adding a numeral classifier (Bisang, 1993; Downing, 1996). From the perspective of speakers of grammatical number languages, “all nouns are like our mass nouns” (Lucy, 1992, p. 83).

 The study of numeral classifiers has contributed to exploring verbal classification systems and verb classifiers (McGregor 2002). The Verb Classifier Hypothesis in Slavic languages applies the concept of the nominal classifier system to the verbal aspectual category (Janda et al., 2013; Dicky & Janda, 2015). It posits that perfectivizing aspectual prefixes in Slavic function as verb classifiers, added onto bare imperfective verbs in a manner analogous to numeral classifiers attaching to bare nouns. It is grounded in the noun-verb analogy that perfective is a discrete solid object and imperfective is a fluid substance (Janda, 2004). Given this, our question is: what can be said about the verbal aspectual category in Japanese, a classifier language, from the perspective of the Hypothesis, which is built on a grammatical number language such as Russian? The question concerns the conceptual integrity of the linguistic phenomena observed in different parts of one language (Ikegami, 1993). In this regard, we can presume that the motivational concept underlying the representation of noun referents should recur in the representation of verb referents in each language type. At the same time, it is crucial to take into account the fundamental difference in the aspectual characterization of bare verb forms: in Russian, they are imperfective, whereas in Japanese, they are considered perfective (Kudo, 1995; Suda, 2010).  

Our study aims to examine the similarities and differences between Russian bare verbs (i.e., imperfectives) and Japanese bare verbs in search of answers to the above-mentioned question. At this stage, we focus on Russian imperfectives that pair with the so-called “purely aspectual perfectives,” or “Natural Perfectives” (Janda’s terminology), and their correspondences in Japanese, alongside Japanese bare verbs and their correspondences in Russian. We examine text data obtained from the Japanese-Russian / Russian-Japanese parallel corpora of the Russian National Corpus and from the author’s own parallel corpus of literary works. 

The corpus-based empirical study reveals a variety of instances where Russian imperfectives correspond to Japanese bare verbs (despite their aspectual characterization), as well as cases where they do not. In the latter cases, Japanese bare verbs are used to express the initial and final phases of actions conveyed by Russian perfectives. Given these mismatches between Russian and Japanese bare verbs, how can we identify traits of noun-verb analogy in these target languages? This leads us to the following hypothesis. Perfectivizing an imperfective verb in Russian entails counting all the internal temporal phases of the action, thereby imposing a heterogeneous reading on the preceding process denoted by the corresponding imperfective verb. The grammatical system of imperfective-perfective pairing enables them to function in a “built-in mode” (cf. the property of objects – Quine, 1960). In contrast, Japanese bare verbs express the generic concept of action without specifically singling out its internal parts (cf. the part-whole structure of substances – Bunt, 1985), thereby allowing for multiple aspectual readings depending on lexical and contextual factors. The weak notion of decomposing a process into distinct phases prevents it from being perceived as a heterogeneous contour. As a result, Japanese bare verbs convey a broader range of aspectual meanings than Russian imperfectives, extending into the semantic domain of Russian perfectives.