Russian (Indo-European, Slavonic)

1. 1st person singular non-past

The 1st person singular non-past of some second conjugation verbs, which are defective according to the normative standard; the forms below are taken from Zaliznjak's (1977) morphological dictionary. The gap corresponds to a set of stem-final consonant alternations found in this conjugation class. Examples of the relevant alternations are given below, comparing defective and non-defective verbs; note that these alternations are obligatory for verbs of the second conjugation.

s ~ š
z ~ ž
t ~ č (or šč)
d ~ ž
st ~ šč
lisit′
'be cunning'		 prosit′
'ask'		 buzit′
'make a fuss' 		porazit′
'strike'' 		očutit-sja1
'find oneself' 		tratit′
'waste'		 pobedit′
'defeat'		 budit′
'wake'		 šelestet′
'rustle'		 pustit′
'release'
1SG -------- prošu -------- poražu -------- istraču -------- bužu -------- pušču
2SG lisiš′ prosiš′ buziš′ poraziš′ očutiš-sja istratiš′ pobediš′ budiš′ šelestiš′ pustiš′
3SG lisit prosit buzit porazit očutit-sja istratit pobedit budit šelestit pustit
1PL lisim prosim buzim porazim očutim-sja istratim pobedim budim šelestim pustim
2PL lisite prosite buzite porazite očutite-s′ istratite pobedite budite šelestite pustite
3PL lisjat prosjat buzjat porazjat očutjat-sja istratjat pobedjat budjat šelestjat pustjat

There is both synchronic and diachronic evidence that the problems with this form are connected with some aberration in the stem alternation. Synchronically, the behavior of the perfective passive participle provides a clue. By default, the perfective passive participle undergoes the same stem alternation as the 1st singular. But the perfective passive participles of defective verbs (note that many do not have them, being either imperfective or intransitive, or both) in Zaliznjak' s dictionary all display consonant alternations that are not drawn from the set shown in the table above. These are of two types:

Thus synchronically there is evidence that something may be amiss with the 1st person singular. Indeed, such aberrant alternations are also found non-standardly in the 1st person singular of defective verbs. Diachronically, these aberrant patterns were prescribed for the normative standard in earier periods for at least some verbs, though they never seem to have caught on. In part, these alternation patterns came along with verbs borrowed from Church Slavonic, the ecclesiastical and literary language current into the 18th century, which was similar to yet morphologically distinct from Russian in a number of respects (for more detail on this history, see Baerman 2008).

2. Words with phonologically automatic stress retraction

This pattern of defectiveness is found in different word classes but under the same morphological conditions, namely when:

The problem environment is thus one where there is an automatic shift of stress onto the stem brought about by a null anding, and where this is the only shift of stress within the paradigm; words which already contain a grammatically conditioned alternation of stress onto the stem in some parts of the paradigm are exempt. This affects three word classes, described by Es′kova (1989):

2.1 Nouns

A null ending is found in the genitive plural of 2nd declension (feminine) and 4th declension (neuter) nouns, and defectiveness is found in these classes. In theory, the nominative/accusative singular of 1st declension (masculine) nouns with end stress could also be affected, but they are not:

'dream'
(2nd declension)
'letter.DIM'
(4th declension)
'table'
(1st declension)
singular plural singular plural singular plural
NOM mečt-á mečt-ý pis′mec-ó pis′mec-á stól stol-ý
ACC mečt-ú mečt-ý pis′mec-ó pis′mec-á stól stol-ý
GEN mečt-ý *méčt pis′mec-á *pis′méc stol-á stol-óv
DAT mečt-é mečt-ám pis′mec-ú pis′mec-ám stol-ú stol-ám
LOC mečt-é mečt-áx pis′mec-é pis′mec-áx stol-é stol-áx
INS mečt-ój mečt-ámi pis′mec-óm pis′mec-ámi stol-óm stol-ámi

The problematic nature of this genitive plural -- at least for the noun 'dream' -- appears to be confirmed by data from the Russian National Corpus. (And most of the few examples are clearly meant to be facetious.)

token counts of 'dream'
(queried 10/08)
singular plural
NOM 2771 1980
ACC 929
GEN 996 13
DAT 362 221
LOC 459
INS 643 348

Pertsova (2006), looking only at the 2nd declension type, takes this as evidence that the defectiveness of the genitive plural is due a ban on a stress shift that would yield a form not otherwise licensed by the lexical entry. In her view, the nominative singular of any noun is stored in the lexical entry, as are any grammatically conditioned stress shifts, which restricts the problem to null endings outside of the nominative singular, and to words which have no grammatically conditioned stress shifts. In Zaliznjak's dictionary, approximately half of the roughly 100 words that fulfil these conditions are said to be defective, the rest are said to be normal, so this by no means represents an absolute constraint.

To the extent that this constitutes an explanation, its domain of application is the stem, not the lexeme as such. Thus, there a few neuter nouns which display a three-way stem alternation, with two stems in the singular (nominative/accusative vs. the other cases) and one in the plural.3 Out of ten such nouns, three are said by Zaliznjak to be defective in the genitive plural, e.g.:

'temple (of the head)'
singular plural
NOM/ACC témja teme[n]-á
GEN téme[n′]-i --------
DAT téme[n′]-i teme[n]-ám
LOC téme[n′]-i teme[n]-áx
INS téme[n′]-i teme[n]-ámi

Stress alternates between the singular and plural, thus the lexeme as such has a grammatically conditioned stress alternation. Nevertheless, within any single stem the stress is fixed.

2.2.1 Historical aside

This pattern with with 2nd declension nouns constitutes the earliest type of defectiveness that was overtly discussed: Barsov (c. 1785) lists some 21 nouns he claims have no genitive plural, all of the fixed end stress type. While there is little overlap with Zaliznjak's list (many having since adopted a different, alternating stress pattern), it does suggest that the factors involved have been in force for some time.

2.2 Adjectives

Defectiveness in adjectives is restricted to the so-called short form adjectives, for which a bit a of background is in order. Russian adjectives inflect according to two patterns, the long form and short form. The long form is, in a sense, the all-purpose adjective, occurring both as attribute and as predicate complement, and inflecting for all cases. The short form is exclusively a predicate complement, and occurs only in the nominative. Nearly all short form adjectives have a long form counterpart, though the reverse is not the case. The choice of short or long form as a predicate complement is largely semantically determined; in brief, the short form is used for properties which are in some way contingent (e.g. temporary). What is relevant here is that the masculine singular has a null ending, and so resembles the genitive plural as described above.

Short form adjectives overwhelmingly display stem stress or grammatically conditioned stress shifts. Es′kova (1989) lists only nine short form adjectives with end stress, of which five are said not to use the masculine singular:

end stress long
form (typical)
end stress short forms
non-defective
 defective
'laughable' 'laughable' 'sick' 'by birth;
own'
M.SG smešn-ój smešón bólen ------
F.SG smešn-ája smešn-á bol′n-á rodn-á
N.SG smešn-óe smešn-ó bol′n-ó rodn-ó
PL smešn-ýe smešn-ý bol′n-ý rodn-ý

Note that the zero ending induces a morphophonological alternation in these examples, namely the insertion of a so-called 'fleeting vowel' (orthographically o or e) between the stem-final consonants.4 The non-defective items stress either the fleeting vowel (e.g. smešón) or the initial syllable (e.g. bólen).

2.3 Imperatives

The imperative ending has two allomorphs normally predictable from the stress of the non-past paradigm (represented here by the 1st person singular): -í if end stressed and null if stem stressed:

end stress
'speak'		 stem stress
'permit'
INF govo[r′]-ít′4 pozvó[l′]-it′
1SG govo[r′]-ú pozvó[l′]-u
IMPV govo[r′]-í pozvól′

An exception to this are stems in -Vj, which take a null ending regardless of stress. For the end-stressed type, this again yields an automatic retraction of stress (note that where the reflexive marker -sja appears, it is ignored by stress assignment rules). This small group is particularly erratic:

Type 1
Type 2
Type 3
boját′sja
smeját′sja
stoját′
normal
'chew'		 normal
'sing'		 defective
'rot'		 normal
'be afraid' 		defective
'be startled' 		normal
'start laughing' 		defective
'make fun of' 		normal
'stand' 		defective
'insist'
INF žev-át′ p-ét′ gn-ít′ boj-át′-sja uboj-át′-sja zasmej-át′-sja zasmej-át′ postoj-át′ nastoj-át′
1SG žuj-ú poj-ú gnij-ú boj-ú-s′ uboj-ú-s′ zasmej-ú-s′ zasmej-ú postoj-ú nastoj-ú
IMPV žúj pój *gníj bój-sja ?ubój-sja zasméj-sja ?zasméj postój ?nastój

3. Genitive plurals not covered by §2.1

With asyllabic stems, the null genitive plural ending that is found with some classes of nouns poses obvious difficulties. Though insertion of a 'fleeting vowel' (see §2.2 above) is frequent in words with a syllable-final consonant clusters (e.g. nominative singular kukl-a 'doll' ~ genitive plural kukol), where the consonant cluster itself constitutes the whole stem, the genitive plural appears simply to be avoided. Most of the roughly half-dozen examples cited by Zaliznjak are words whose application in the plural must in any case be fairly limited, e.g. t′m-a 'darkness', mgl-a 'mist'. The converse of this is found with zl-o 'evil', which in the plural has only the gentive zol.

'mist'
(2nd declension)
'evil'
(4th declension)
singular plural singular plural
NOM t′m-a t′m-y zl-o *zl-a
ACC t′m-u t′m-y zl-o *zl-a
GEN t′m-y *tem zl-a zol
DAT t′m-e t′m-am zl-u *zl-am
LOC t′m-e t′m-ax zl-e *zl-ax
INS t′m-oj t′m-ámi zl-om *zl-ámi

Earlier descriptions (e.g. Chernyshev 1915) characterize zlo as singularia tantum, with the genitive plural exclusively in the idiom 'the lesser of two evils'. But in fact, the genitive plural may be used in contexts without even the remotest connection to this idiom, as examples from the Russian National Corpus indicate.

Finally, the neuter nouns ígo 'yoke' and èxo 'echo', are said to be defective in the genitive plural, even though: (i) they are stem-stressed, so a null genitive plural would not induce a stress shift, and (ii) their roots are not asyllabic. In the case of èxo, at least, this may be due to homophony avoidance, since the expected form coincides with the common interjection ex, expressing derision or disgust.

4. Oblique cases of nékto 'somebody, a certain' and néčto 'something'

For this material , see here.

Acknowledgements

Thanks to Dmitry Idiatov for some important corrections.

Notes

1 This verb has the reflexive suffix -sja/-s′.

2 Again, this is according to normative sources. Some speakers find such forms acceptable.

3 The square brackets indicate a phonemic transcription, as opposed to the pure transliteration applied elsewhere (phonemic transcription in cases where the transliteration would be uninformative or misleading). The alternation between the stem-final palatalized [n′] and non-palatalized [n] is lexically stipulated.

4 In some contexts the fleeting vowel is phonologically predictable (e.g. syllable final consonant clusters of rising sonority). Elsewhere, it is lexically specified.

References

Baerman, Matthew. 2008. Historical observations on defectiveness: the first singular non-past. Russian Linguistics 37/1. 81-97.

Barsov (c. 1785): published as Tobolova, M. P., & Uspenskij, B. A. (eds), 1981. Rossijskaja grammatika Antona Alekseeviča Barsova. Moscow: Izdatel′stvo moskovskogo unversiteta.

Černyšev, Vasilij I. 1915. Pravil′nost′ i čistota russkoj reči. Vol. 2: Časti reči [The correctness and purity of Russian. Vol. 2: Parts of speech] (3rd edition). Saint Peterburg: Tipografija M. Merkuševa.

Es′kova, N. A. 1989. Mexaničeskij akcentnyj sdvig kak pričina defeknosti paradigm v russkom literaturnom jazyke [Automatic shift of stres as a cause of defectiveness in literary Russian]. Problemy strukturnoj lingvistike 1985-87. 96-103.

Pertsova, Katya. 2005. How lexical conservatism can lead to paradigm gaps. In: J. Heinz, et al. (eds), Working Papers in Phonology 6 (UCLA Working Papers in Linguistics, 11). Los Angeles: UCLA. 13–38.

Zaliznjak, Andrej A. 1977. Morfologičeskij slovar′ russkogo jazyka [Morphological dictionary of Russian]. Moscow: Nauka