This presentation is based on the framework of cognitive linguistics, which assumes that grammar results from metaphorical and metonymical extensions of construals of perceptual experiences (for more on the basic tenets of cognitive linguistics, see Laura Janda's position piece at http://www.indiana.edu/~slavconf/SLING2K). In this framework, the form-meaning relationship is very real. This means that a given form (such as a preposition or a morphological category such as case) always bears its meaning (which is coherent though it is often complex), and that if form is different, meaning is also different (i.e., true synonymy is rare, though meanings may be similar). On the basis of nearly fifteen years of research on case semantics, the author has established that each case represents a coherent semantic network, and that all uses of a given case access at least some portion of this semantic network. The uses of cases and prepositions are neither arbitrary nor meaningless. The present paper is an attempt to take this research a step further by comparing systematic semantic relationships among the cases, thereby targeting semantically significant junctures in the grammatical landscape of case, namely those that are potentially ambiguous. A semantic relationship exists when there is some degree of synonymy (semantic contiguity or overlap) and/or multiple constructions are available to express various construals of a perceived experience. The ultimate goal is to explore the various ways in which a given experience can be cognitively and linguistically manipulated.
The six grammatical cases of Russian are not discrete semantic islands. The cases interact with each other, variously dividing up and sharing meanings and syntactic roles, and thus weaving the whole of the case system into an interconnected, interdependent entity. I will attempt a brief typology of inter-case relationships in Russian, based on three parameters: 1) the number of cases engaged in a semantic relationship, 2) the type of semantic relationship, and 3) the factors contributing to the semantic relationship.
1) The number of cases engaged in a semantic relationship: 1 x 1, 1 x 1 x 1, 2 x 2
By far the most common relationship is that of a pair of cases, as the instrumental and genitive in this pair of phrases: stradat' bessonnicej vs. stradat' ot bessonnicy, two overlapping yet non-identical construals of an experience expressed in English as 'suffer from insomnia'.
Less common, but by no means rare, are threesomes of cases engaged in a similar semantic task, as we see with the nominative, instrumental, and accusative cases, all of which can express a possessed object, as in the following three construals of a given perceived reality: U nego ogromnye sredstva vs. On raspolagaet ogromnymi sredstvami vs. On imeet ogromnye sredstva 'He has enormous means/great wealth'.
There are additionally a few instances where a given concept can be rendered with two different arrays of cases, such as verbs meaning 'teach' which can either mark the learner with the accusative and the subject matter with the dative (uchit'), or the learner with the dative and the subject matter with the accusative (prepodavat').
2) The type of semantic relationship: contiguous, overlapping, or virtually synonymous.
Contiguous meanings amount to a division of labor--these meanings may touch, but do not actually overlap. An example of contiguous meanings are the destinational uses of the prepositions v and na with non-humans in the accusative case and the parallel destinational use of the preposition k with humans in the dative case. The choice is forced by the nature of the destination, but a very strong parallelism exists. Like complementary distribution in phonology, this differentiation is symptomatic of a greater unity.
Overlapping meanings provide choices for speakers, since they share at least some portion of meaning and syntactic function, although usually one option will be preferred. Certain verbs denoting waiting and wanting can appear with either the accusative or the genitive cases, and the choice is largely determined by context: Boris zhdet avtobus/avtobusa 'Boris is waiting for the/a bus'.
Occasionally the overlap between case meanings is so extensive that the resulting expressions are virtually synonymous; for example the use of the genitive or accusative with zhal'/zhalko 'pity, sorry', or the use of the accusative with izuchat' 'study' as opposed to the dative with uchit'sja 'study'.
3) The factors contributing to the semantic relationship: metonymic reduction, construal, abstract similar meaning, same/similar lexical trigger.
In comparison with a complete path (Bill walked over the hill), the endpoint of a path (Sally lives over the hill) is a metonymic reduction. Metonymic reduction relates the destinational (path) meanings marked with the accusative of the prepositions v, na, o, za, pod with their corresponding (and contiguous) locational (endpoint) meanings marked with the locative case for v, na, o, and with the instrumental case for za, pod.
The speaker can choose to foreground, background, imply blame, etc. By imposing various case roles on the noun phrases in an utterance. Although Ja xochu spat' 'I want to sleep' and Mne xocùetsja spat' 'I feel like sleeping' can describe the same objective reality, the speaker is assigning a volitional role to the nominative subject in the first sentence, but claiming that the corresponding dative in the second sentence is but a victim of circumstance. Similarly, Ivan razbil mashinu otca makes a relatively neutral statement, whereas Ivan razbil otcu mashinu focuses on the father's distress, though both mean 'Ivan smashed up father's car'.
Often a number of means (including various cases) are used to express a set of similar concepts with overlapping or contiguous meanings. For example, Russian has many ways to express directionality in the domain of intention (many of which are roughly equivalent to English 'for'), among them the prepositions v, na, pro, and za with the accusative case, do, dlja, radi with the genitive case, as well as the bare dative case (as in Vot pis'mo emu 'Here's a letter for him').
Sometimes different case uses are triggered by the same or similar lexical items, such as verit' 'believe', which can combine with both the bare dative case and the preposition v with the accusative case. The verb napolnit' 'fill' uses the instrumental case to mark the substance doing the filling, but the related adjective polon 'full' can use the genitive case for the filling substance. The same/similar triggers indicate that there are some semantic points of contact between these case uses, again usually involving overlapping or contiguous meanings.
This paper will examine these and other factors in an exploration of the interconnections among the Russian cases (from a database of inter-case relationships currently numbering several dozen).