The verbs of motion form an anomalous subclass within the Russian aspectual system. Whereas the typical Russian verb is usually described in terms of perfective/imperfective pairings, the verbs of motion revolve around a pairing of two imperfective forms, here termed the determinate and indeterminate. Furthermore, when prefixed, determinate verb stems pattern as perfective verbs, while indeterminate stems can be both perfective and imperfective. The exact nature of the relationship between the verbs of motion and the aspectual system seen in other verbs is still an open issue, though there is very little in the system of motion verbs not also seen in non-motion verbs.
The typical non-motion verb in Russian centers around an imperfective prototype, the most basic form from which other forms are ultimately derived (e.g., читатьI 'read', писатьI 'write', прыгатьI 'jump')1. These can derive three broad types of perfective forms, usually be prefixation or suffixation: the natural pair that has nearly identical meaning (прочитатьP 'read', написатьP 'write', прыгнутьP 'jump'), the Aktionsart pairs that express some sort of inherently perfective lexical aspect (почитатьP 'read for a bit', пописатьP 'write for a bit', допрыгатьсяP 'get into a mess'), and the regular prefixed perfectives that have a new meaning and are capable of further derivation (дочитатьP 'read up to', подписатьP 'sign', спрыгнутьP 'jump off') (Timberlake 2004; he calls the last two categories “quantizing” and “qualitative” respectively). This last group can in turn form secondary imperfectives by suffixation (дочитыватьI 'read up to', подписыватьI 'sign', спрыгиватьI 'jump off'). The schematic below shows the structure of derivations from a single prototype.
The motion verb in Russian has two prototypes, one indeterminate expressing manner of motion, iteration, or non-linear paths and one determinate expressing single linear motions (e.g., ходитьII 'walk, go', идтиID 'go'). These share a single natural pair (пойтиP 'go, set off'). Through prefixation, the determinate form derives perfectives marking direction of motion, which then in turn form secondary imperfectives using a form identical or related to the indeterminate stem (выйтиP 'exit', выходитьI 'exit') (Timberlake 2004). Separately, the indeterminate prototype may also derive Aktionsart pairs and regular prefixed perfectives (сходитьP 'go and come back', выходитьP 'nurse, care for'), the latter of which can then form regular secondary imperfectives (выхаживатьI 'nurse'). These relationships are diagrammed below.
These derivations do not appear to fit nicely into the standard model seen in non-motion verbs. In particular, there are three primary pairings in need of additional explanation: determinate/indeterminate, imperfective prefixed indeterminate/perfective prefixed indeterminate, and prefixed determinate/prefixed indeterminate.
1. The Relationship between Determinate and Indeterminate Forms
The Russian verbs of motion uniquely consist of two prototypes, the determinate imperfective and indeterminate imperfective, that share a very strong affinity with each other that is not simply semantic, but pervades these verbs' derivational systems. The determinate and indeterminate stems are related for all motion verbs except идтиID/ходитьII 'go on foot' and in part ехатьID/ездитьII 'go by vehicle', but the derivation of one stem from the other is quite heterogenous, ranging from Indo-European o/e ablaut, to Proto-Balto-Slavic statives, to Proto-Slavic iteratives. Since the actual creation of these stems spans thousands of years, for all intents and purposes the derivation is irregular and non-productive from a synchronic perspective (Schuyt 1988, Dicky 2010).
One question that arises is whether the determinate and indeterminate forms for a given verb of motion are actually paired morphologically, or whether they are simply perceived as pairs due to their semantics (Горбань 2002).
Historically, well into the early Slavic written record, it certainly appears that the determinate and indeterminate forms were independent lexemes, and that the affinity for each other seen in the modern language was either very weak or absent. The predecessors of the modern indeterminate verbs were likely manner of motion verbs (*xoditi 'walk', *jězditi 'ride', cf. *jьti 'go', *jě(xa)ti 'travel') or iteratives (*nositi 'carry repeatedly', cf. *nesti 'carry once') rather than true verbs of motion; these could be coerced into motion verbs through prefixes (*vъ-xoditi 'walk into', cf. *vъ-jьti 'go into'). This analysis is supported by the etymology of the indeterminate forms, their apparent semantics in early Slavic texts and in South Slavic (where the indeterminate/determinate contrast never developed), and the strong correlation in their adverbial usage when compared with verbs of manner in other Indo-European languages2 (Dicky 2010).
However, in modern Russian, determinate and indeterminate stems have seem to have much more in common than just semantics. This is most obvious in prefixed forms, where the determinate/indeterminate stems are repurposed as perfective/imperfective stems highly productively3. These prefixed pairings of verbs (входитьI/войтиP 'enter', относитьI/отнестиP 'carry away', etc.) clearly form standard Russian aspectual pairs, expressing the very same contrasts that non-motion aspectual pairs do4.
However, the grammaticalization of the indeterminate/determinate pairings is not complete, in that certain forms exist outside of this relationship. Notably, indeterminate verbal stems can behave completely independently of their determinate counterparts, deriving perfective forms such as выходитьP 'nurse', заездитьP 'wear out, exhaust [by driving/riding]', наплаватьP 'sail/swim [a certain distance]' that have no relation whatsoever with the corresponding determinate verbs идтиID/ехатьID/плытьID. In contrast, the determinate stems always coexist with indeterminate stems and are completely incapable of deriving non-suppletive secondary imperfectives with bases like *-идывать, *-ехивать, *-плывывать. Verbs of motion outside of the system of verbs of motion are the subject of the following section.
2. The Relationship between Imperfective and Perfective Prefixed Indeterminate Forms
Indeterminate verbal stems are capable of deriving both perfective and imperfective verbs by prefixation. Imperfective derivatives almost always express motion, combining the action described by the verb root with the direction given by the prefix; these verbs are always paired with perfectives derived from the corresponding determinate stem. Perfective prefixed indeterminates, however, can form secondary imperfectives according to the same processes used by non-motion verbs, in particular with the imperfectizing suffix -ыва-: выходитьP 'nurse' → выхаживатьI, доноситьP 'wear' → донашиватьI. In fact, if the imperfective prefixed indeterminate verbs are put aside, the remaining derivatives of indeterminate prototypes are indistinguishable from well-behaved non-motion verbs; triplets such as носитьII → доноситьP → донашиватьI perfectly parallel писатьI 'write' → дописатьP 'write up' → дописыватьI 'write up'. Perfective prefixed indeterminate verbs also share other properties in common with other perfective verbs, such as stressing the prefix вы- in perfective forms but not in imperfective ones: вы́ходитьP 'nurse', выходи́тьI 'exit' (Schuyt 1988, Titelbaum 1990, Janda 2010).
The diachronics of this situation are reasonably clear: the perfective prefixed indeterminate forms reflect the fact that the determinate and indeterminate verbs were originally wholly separate lexemes with no unusual affinity, so at one point the verbs now called indeterminate were freely able to take perfectizing derivational prefixes. The imperfective prefixed indeterminate verbs, on the other hand, are the result of generalizing the model of prefixed motion verbs, where the indeterminate stem always marks imperfective forms. The latter group has come to dominate, with the perfective forms typically described as exceptional in most pedagogical materials, despite the fact that they conform to the typical patterns of Russian derivation; they are the “exception to the exception” within the realm of motion verbs (Janda 2010).
While at first glance it seems as though the existence of both perfective and imperfective derivatives of indeterminate stems is yet another complicating factor that makes verbs of motion break the traditional pattern where prefixation results in perfectivization, this arrangement is not at all unique to verbs of motion. For instance, the verb па́датьI 'fall' forms derivatives of both aspects, as in напа́датьP 'fall in large amounts' and напада́тьI 'attack' (Schuyt 1988). We will return to напа́да́ть and how it might help to explain the derivational processes at work in verbs of motion in the following section.
3. The Relationship between Prefixed Determinate and Indeterminate Forms
One question that remains unresolved is the underlying relationship between perfective prefixed determinate verbs of motion (like уйтиP 'leave') and imperfective prefixed indeterminate verbs of motion (like уходитьI 'leave'). As discussed earlier, these two groups clearly pattern as aspectual pairs of one another. However, there are two takes on the underlying origin of the imperfective forms: Is a verb such as уходитьI a secondary perfective derived from уйтиP that for some reason is reliably suppletive, always based off the indeterminate stem, or does уходитьI derive from ходитьII in parallel to how уйтиP derives from идтиID, as the morphology suggests?
The former view (the secondary derivation hypothesis), argued by Isačenko (1960) among others, derives its strength primarily from the fact that it brings prefixed verbs of motion in line with the rest of the Russian aspectual system, only leaving behind the unpaired (in)determinate verbs as aspectual oddities. That is, the imperfective verb идтиID is prefixed to form the perfective уйтиP, which then derives a secondary imperfective уходитьI; this is conceptually the same as derivation among non-motion verbs, with the one exception that Russian secondary imperfectives never use suppletive stems. It also seems to run afoul of normal Russian rules of derivation, since it denies the fact that the verb ходитьII is the source of уходитьI, instead placing the suppletive уйтиP into the role of middleman; when taken at face value this seems patently illogical, since why would one posit two rounds of suppletion in order to achieve the net effect of simply adding a prefix (Janda 2010)?
A newer view (the coderivation hypothesis) is posited by Janda (2010), who argues that the secondary derivation hypothesis is overly complex and opts for morphologically more transparent approach of deriving уходитьI directly from ходитьII. She correctly states that this approach resolves the morphological issue by resorting to typical Russian derivational procedures, and argues in addition that the reason why some derivatives of indeterminate verbs are imperfective and others perfective ultimately comes down to semantics. However, she does not discuss the ways in which this significantly complicates the analysis as well, since it requires positing an entirely new mechanism for creating aspectual pairs in Russian: that the perfective and imperfective can be independently derived and then be brought together to form a new pair. It also runs into problems when the imperfective prefixed verbs of motion have some sort of mutation or additional morphology present. For example, the imperfective counterpart to приехатьP 'arrive by vehicle' is приежзатьI, not *приездить as the indeterminate form ездитьII would suggest. Historically the form приезжатьI is related to ездитьII, simply with the formant *-aj- added (early Proto-Slavic *jēsd-ei-tei → *prī-jēsd-i-ai-tei). Why would the addition of a prefix to ездить force the addition of an imperfectizing formant, when other verbs need no such formant?5
In this instance, I am inclined to side with the secondary derivation hypothesis due to evidence from non-motion verbs. The case of напа́дать/напада́ть is quite explicative, since it actually closely parallels the verbs of motion without having the extra complicating factor of the determinacy contrast overlaid on top of it. This verb is one of the less frequent cases of the verbal prototype/morphologically-simplest form being perfective, namely пастьP 6. Like most other verbs with a perfective prototype, it had a derived imperfective partner formed with the suffix *-aj-, падатьI. Since prefixed perfectives are based on whatever the verbal prototype is, we get the form напастьP 'attack' derived from пастьP rather than from падатьI, since пастьP is morphologically unmarked. When напастьP derives a secondary imperfective, it is based on the stem -падать rather than a regularized *-падывать because the natural pair пастьP/падатьI already provides a perfect aspectual template with no prefixes; the stress shifts from па́датьI to напада́тьI because of a general Russian stress rule where the suffix *-aj- is always stressed in secondary imperfectives7 (Schuyt 1988).
However, perfective prototype verbs in Russian have long since lost their productivity, with all new verbs having imperfective prototypes. There was thus likely pressure to derive perfectives from the imperfective stem падатьI as well, thereby resulting in forms like напа́датьP 'fall in large amounts' (note the stress, since this is not a secondary imperfective). In a way, the verb 'fall' appears to have two prototypes, both пастьP and падатьI.
This situation is almost identical to the verbs of motion, which clearly have two prototypes. If we momentarily take a logical leap and call ходитьII/идтиID a natural aspectual pair, with “imperfective” ходитьI? And “perfective” идтиP?, we have a perfect parallel with the case of пастьP/падатьI: we call идтиP? the primary “perfective” prototype, which derives an “imperfective” natural partner ходитьI? by no-longer-productive means; we then can make prefixed forms like выйтиP from the “perfective” prototype by prefixation, which then derives a secondary imperfective выходи́тьI by mirroring the prototypical natural pair. At the same time, there is pressure to treat ходитьI? as a prototype since it is “imperfective”, and thus it can derive a small number of its own perfectives, such as вы́ходитьP 'care for, nurse'. Here again we see aspect-driven stress rules apply, distinguishing the two senses of выходить.
If we continue with the assumption that ходитьI?/идтиP? form a natural pair, we can also explain the mutations in prefixed imperfectives. Amongst speakers of the East Slavic languages, there was a pressure to distinguish between morphologically-identical verbs that had distinct aspects, hence the differing stress patterns in напада́тьI/напа́датьP or выходи́тьI/вы́ходитьP above. One approach, seen in Ukrainian and Belarusian, is to generalize different stress patterns to as many verbs as possible, so we see contrasts like нано́ситиI 'bring'/наноси́тиP 'bring in a certain amount' (Russian наноси́тьI/P for both) or наїзди́тиI 'bump into'/наї́здитиP 'ride a certain distance' (Russian наезжа́тьI/нае́здитьP) (Schuyt 1988). Russian, opting against the stress approach in most instances, instead took to altering the imperfective stem by doubly-marking the imperfective aspect when stress shifts were not an option.
The above-outlined approach essentially eliminates any barriers between integrating verbs of motion with non-motion verbs, with one significant caveat: the fact that it relies on treating the determinacy contrast as a one of perfectiveness, when this is obviously not the case in actual usage. However, we can say that the indeterminate/determinate pairs pattern lexically and morphologically identically to imperfective/perfective pairs, both in prefixed and unprefixed forms, even if they do not actually express the imperfective and perfective aspects.
There are also semantic reasons to warrant treating (in)determinate verbs as “(im)perfective-like”. Indeterminate verbs express certain qualities like imperfective verbs, such as their proclivity towards iteration and the fact that indeterminate verbs are use far more frequently than determinate verbs to express atelic actions (e.g., ребенек ходит “the baby is walking”, мы ходили по парку “we walked around the park”). It is likely this fact that caused these formerly independent lexemes to become intertwined aspectually in late Common Slavic and old Russian as well: one of the reasons Russian lost the aorist and imperfect tenses was that their functions were by and large duplicated by perfective and imperfective past tense verbs respectively. Since verbs of motion like *iti 'go' tend to be telic and verbs of manner like *xoditi 'walk' tend to be atelic, the verbs of motion were more likely to appear in the aorist, while the verbs of manner would appear in the imperfect (Dicky 2010). When these two tenses were lost, these words were brought together as a perfective/imperfective pair (most clearly evident in prefixed verbs) by virtue of appearing in many of the same contexts of perfective and imperfective verbs.
While the Russian verbs of motion have historically been analyzed in a variety of ways that puts them at odds with much of the rest of the Russian verbal system, just about every quirky aspect of the class can also be observed amongst the non-motion verbs. Pairs such as падатьI/(у)пастьP have a network of derivations that is structurally identical to the verbs of motion. The only truly exceptional feature is the determinacy contrast in unprefixed verbs of motion, which however seems to fulfill the same role as the imperfective/perfective contrast derivationally, though not syntactically and only very marginally semantically.
Dicky, S. M. “Common Slavic 'indeterminate' verbs of motion were really manner-of-motion verbs”. In New Approaches to Slavic Verbs of Motion, ed. V. Hasko & R. Perelmutter. Philadelphia: John Benjamins Publishing Company, 2010, p. 47-110.
Janda, L. A. “Perfectives from indeterminate motion verbs in Russian”. In New Approaches to Slavic Verbs of Motion, ed. V. Hasko & R. Perelmutter. Philadelphia: John Benjamins Publishing Company, 2010, p. 125-140.
Schuyt, R. N. “The Indeterminate Verbs of Motion and the Morphology of Aspect”. Studies in Slavic and General Linguistics. Vol. 11. Dutch Contributions to the Tenth International Congress of Slavists. 1988, p. 481-494.
Timberlake, A. A Reference Grammar of Russian. Cambridge: Cambridge University Press, 2004, p. 115-116, 398-415.
Titelbaum, O. A. “Prefixed Russian Verbs of Transposition”. Russian Linguistics. Vol. 14, No. 1. 1990, p. 37-46.
Горбань, О. А. Древнерусские глаголы движения в системе языка и в тексте. Volgograd: Volgograd State University Press, 2002, p. 26-54.
Исаченко, А. В. Грамматический строй русского языка в сопоставлении с словацким. Bratislava: Press of the Slovak Academy of Sciences, 1960, p. 337-344.
1) Superscript letters will be used as shorthand to mark various verbal categories: I (imperfective), P (perfective), II (imperfective indeterminate), ID (imperfective determinate).
2) Modern Russian is a strongly satellite-framed language, where verb stems encode manner of motion while prefixes and prepositions encode direction. Common Slavic appears to have been a hybrid verb-framed/satellite-framed language like most other Indo-European languages. Verb-framed constructions use a manner of motion verb in an adverbial role to express how movement took place (e.g., “the bottle left, floating” for “the bottle floated away”). In the oldest Slavic texts, we see similar usage of the indeterminate forms in adverbial roles that are not possible in modern Russian (Dicky 2010).
3) Mutations in the indeterminate stem in prefixed imperfectives will be discussed in section 3.
4) For instance, prefixed determinate verbs in the past tense like войтиP mean that the agent (intransitive)/patient (transitive) is still in the final destination, while prefixed indeterminate verbs like входитьI mean it is no longer there. This contrast parallels the use of non-motion imperfectives to mark “annulled action”; in this case, with входитьI, the action of entering is annulled by virtue of the agent no longer being in the place where he/she entered (Timberlake 2004).
5) It is not feasible to claim that this is simply a historical oddity, since it clearly was productive quite recently as well. In the case of тащитьID/таскатьII 'haul', the most recent addition to the Russian verbs of motion, the stem used in derived imperfectives is -таскивать, which uses the imperfectizing suffix -и/ыва- that is very much still productive in the modern language.
6) In modern Russian the verb пастьP is archaic, having largely been replaced by the prefixed упастьP, but this does not impede the core argument. ПастьP was simply displaced by упастьP after the system described was organized.
7) We see this rule also at work in derived forms of the verb of motion бегатьII 'run', which also contains the suffix *-aj-. In the non-prefixed form бе́гатьII, the stem is stressed; in prefixed forms such as прибега́тьI 'come running', the suffix is stressed. A similar argument can be made for пла́ватьII/-плыва́ть 'sail, swim' and other verbs. There is no useful information in the case of лета́тьII/-лета́ть 'fly', since the ending is always stressed here. (Dicky 2010)