Russian verbal aspect, along with the associated derivational morphology, is widely acknowledged as a particularly problematic issue both in linguistics and in language pedagogy. Aspect involves the way in which human beings conceptualize time, and because it provides evidence that different languages use different cognitive strategies, this issue has important implications for cross-linguistic typology and human cognition.

All Russian verbs clearly express aspect, in all forms and all contexts, as either Perfective or Imperfective. This bipartite division in the lexicon has long been the motive behind the traditional “pair” model of Russian verbs according to which a given verb is “paired” for aspect, and verbs are assumed to have Perfective and Imperfective “partners”. Janda 2007 demonstrated that the “pair” model is inadequate and that it is necessary to recognize four types of Perfective verbs, based on both semantics and derivational patterns of behavior. This can be demonstrated with the Imperfective verbs pisat’ ‘write’ and čixat’ ‘sneeze’ and the Perfective verbs derived from them via prefixes and suffixes. When pisat’ describes a completable activity, like pisat’ dissertaciju ‘write a dissertation’, it is possible to form a Natural Perfective with the same meaning: napisat’ dissertaciju ‘write a dissertation’. The verb čixat’ ‘sneeze’ can’t describe a completable activity, so it has no Natural Perfective. Both verbs can be modified to describe other completable activities, which are Specialized Perfectives, as in vypisat’ ‘issue, prescribe’ (as in: ‘issue a fine, prescribe medicine’) and vyčixat’ ‘expel by sneezing’. Specialized Perfectives are the only Perfectives that regularly form derived Imperfectives by means of suffixation, as in vypisyvat’ ‘issue, prescribe’. When a verb describes a non-completable activity, such as pisat’ naučnuju fantastiku ‘write science fiction’ or čixat’ ‘sneeze’, it is possible to form a Complex Act Perfective describing an activity that is bounded in time but does not reach a result, such as popisat’ ‘write for a while’ and počixat’ ‘sneeze for a while’. If a verb forms a Complex Act Perfective and describes a cyclical repeated activity like čixat’ ‘sneeze’, it is possible to form a Single Act Perfective that describes just one cycle, as in čixnut’ ‘sneeze once’. There are twelve cluster structures observed among Russian verbs with various combinations of Perfectives.

The Cluster Model of Russian Verbs was originally proposed in Janda 2007 based on a database representing all morphological classes of verbs in Russian. Janda & Korba forthcoming confirmed the model using a database of high-frequency verbs (all verbs found in a first-year and a second-year textbook of Russian) and suggested pedagogical applications. Janda forthcoming b examined the semantic motivations that align with the morphological patterns in Russian aspectual clusters. The remaining works to date are devoted to problematic subsystems of Russian aspect that have traditionally been considered “deviant” but are easily accounted for in the Cluster Model: the bi-aspectual verbs (Janda forthcoming a and c) and the motion verbs (Janda forthcoming d).

Janda forthcoming c tested an implication of the Cluster Model against language data provided by a search engine. The implication was that bi-aspectual verbs would be much less likely to form Complex Act Perfectives than other verbs. This hypothesis was confirmed with a statistically significant correlation (p < 0.0001). The Cluster Model presents several more implications that can be tested using quantitative language data. Below are some examples of possible subprojects; any one of these subprojects (or any subset of them) would be an excellent starting point for the dissertation to be written by the PhD student. All of these subprojects also address a larger issue of type vs. token frequency which is central in the debate over frequency and cognitive entrenchment (cf. recent work by Bybee and Schmid):

a) Russian uses eleven “empty” prefixes to form Natural Perfectives, and seven of the cluster structures attested in Russian include a Natural Perfective. Is there a correlation between cluster structure and the prefix selected for a Natural Perfective? We believe there is and suggest testing this hypothesis by comparing the distribution of cluster structures and the distribution of Natural Perfectives. The results could tell us something about the meaning of the prefixes, since the cluster structure of a verb is also dependent on its meaning. Type vs. token frequency can be examined in connection with this project by comparing various parameters, such as the most frequent “empty” prefixes, the frequencies of the verbs that use them and the frequencies of the associated cluster structures.

b) Janda forthcoming a shows that approximately 30% of Russian aspectual clusters present a situation in which there is no one-to-one correspondence between form and meaning (cluster elements), and in most such situations a single form plays two roles in a cluster. Thus, for example, pročitat’ can serve as both a Natural Perfective meaning ‘read’ (a whole book, for example) or as a Complex Act Perfective meaning ‘read through a given time period’ (the whole night). This is particularly pronounced among motion verbs, where a form such as zaxodit’ can be both a Complex Act Perfective meaning ‘begin to walk’ and a secondary Imperfective meaning ‘stop by’. What pattern of frequency do we find for these verbs in actual use and what does it tell us? This question can be answered by analyzing corpus data, which can tell us about the relative “strengths” of the various components in Russian aspectual clusters. Type vs. token frequency is likely a factor, since it appears that high-frequency verbs from small classes, such as motion verbs, often participate in such patterns.

c) One prefix, s-,  can form all four types of Perfectives, and it seems that nearly all prefixes can form Specialized Perfectives, but little more is known about the distribution of prefixes relative to the types of Perfective verbs in Russian. A quantitative study could both establish all the possibilities and weigh their relative “strengths” in terms of frequency. This would contribute to an understanding of the meanings of the prefixes, which remains a difficult problem in Russian linguistics.

d) Some cluster structures are very common, some are less common and a few are quite rare. How does the frequency of cluster structure relate to other factors, such as the frequency of the verb itself, the meaning of the verb, and the prefixes associated with it?

e) All of the above subprojects could be extended diachronically to study the historical development of clusters among Russian verbs.

The subprojects listed above all share a common theme, that of using quantitative data and subjecting them to statistical analysis within the framework of Cognitive Linguistics. This methodology is at the cutting edge of Cognitive Linguistic research, and has been facilitated by recent advances in technology (universalization of Unicode; coupling of online corpora to advanced search engines) that has made huge advances in both the quantity and quality of access to natural language data. Quantitative methods were arguably one of the hottest topics at the most recent International Cognitive Linguistics Conference (held in Krakow in July; cf. presentations there by Schmid, Gilquin, Newman, Geeraerts, Glynn, Veal and others – program available at  http://www.iclc2007.pl/). The “Exploring Emptiness” project is also promoting quantitative data analysis, and statistical expertise is among the qualifications for both the PhD and Post-Doctoral positions associated with that project.