This paper investigates the accentual properties of post-accented second-declension nouns in Russian, such as kišhká ‘intestine’, zvezdá ‘star’, dušá ‘soul’, dólja ‘portion’, gubá ‘lip’, and ruká ‘hand, arm’, as well as problems arising from their treatment in A. A. Zaliznjak’s Grammatičeskij slovar΄ russkogo jazyka (1980). These nouns contain post-accented stems to which either accented or unaccented inflectional endings are attached. Morphologically, the stress in the inflected forms of these nouns should be predictable based on the accentual property of the stem, i.e. stress should fall on the inflectional ending in all forms, with the exception of the genitive plural zero ending, in which case stress should fall on the stem.
However, significant variety exists in the stress patterns of these nouns in Russian, attested by Zaliznjak’s sorting of these patterns into four categories, labeled B, D, E, and F, which consist of six stress charts, or sxemy udarenija. Across these categories, one may find both stem and ending stress in every grammatical case, both singular and plural. The morphological determination of stress patterns in these nouns, mainly in the plural, has largely been overridden by paradigmatic leveling, similar to that of neuter nouns, such as oknó / ókna, where columnar stem stress in the plural is opposed to columnar end stress in the singular.
One problematic aspect of Zaliznjak’s categories is the classification of singularia tantum nouns in either category B or D; category B nouns, such as kišká, have columnar stress on the ending in both singular and plural, while category D nouns, such as zvezdá, have columnar stress on the ending in the singular and on the stem in the plural. Out of the approximately 680 second-declension post-accented nouns in Russian, 260, or about 38 percent, are singularia tantum. How does one determine, for instance, that erundá ‘nonsense’ is a category B, while edá ‘food’ is a category D, given the fact that neither has plural forms, the only forms that differentiate categories B and D? Similarly, Zaliznjak’s classification of singularia tantum nouns as having a e/ë alternation is puzzling when such an alternation occurs only in the plural.
Also to be considered in this paper is the productive connection between Zaliznjak’s categories. In his 1977 article on the accentuation of monosyllabic masculine nouns, Zaliznjak discusses how stress mobility has to do with the familiarity of the noun. A number of the nouns in question here have alternative stress patterns, with both stem and ending stress permissible in various inflected forms, reflecting, perhaps, an increase or decrease in familiarity among speakers of Russian. Some such familiarity may be professionally motivated, such as the case of ískra ‘spark, flash’, where the form iskrá is given as the professional term. Both forms, however, have stem stress in the plural: ískry. In these respects, Zaliznjak’s separation of such nouns into separate categories fails to reflect their accentually complex nature.