It has been my good fortune to review two books on Slavic aspect in the space of six months, both published by Oxford University Press (see Slabakova Reference Smith2007). As stated in my previous review, every time a major work on the syntax and semantics of Slavic aspect appears, I am hopeful that it will provide our field with, or at least come close to, a definitive treatment of the subject and give an answer to the most divisive and fundamental question: whether perfective prefixes encode situation (lexical) or viewpoint (grammatical) aspect (Smith Reference Smith1997). Kylie Richardson's book, published soon after Borik (Reference Borik2006), defends the same type of answer to this question, arguing that prefixes are viewpoint aspect markers. Two crucial features of Richardson's book deserve high praise: it is wide in scope and coverage, including data from a number of Slavic languages (Belarusian, Russian, Ukrainian, Czech, Slovak, Polish, and Bosnian/Croatian/Serbian), and offers a wealth of examples that are corroborated by a respectable number of native speakers. While Richardson's book is thus definitely an improvement over Borik (Reference Borik2006), it cannot be regarded as the definitive treatment of Slavic aspect either. In the following, I will summarize Richardson's arguments chapter by chapter, and then turn to her answer to the fundamental question in the study of Slavic aspect.
After a short ‘Introduction’ in chapter 1, in which she maps out the book, Richardson spells out her assumptions about case and aspect in chapter 2, ‘Preliminaries’. She offers lucid definitions of structural case (nominative and accusative), lexical case (case that idiosyncratically depends on the verb and is part of the verb's lexical entry, e.g. the instrumental case governed by verbs like rukovodit' ‘to lead, guide’ and upravljat' ‘to manage, direct’), and semantic case (the genitive of negation and the partitive genitive, but also case linked to semantic features like definiteness and abstract vs. concrete). She forcefully makes the point that both structural case and semantic case are predictable, in the sense that a specific morphological form always brings about a specific meaning, whereas lexical case is completely idiosyncratic and hence unpredictable.
While the reader will benefit greatly from Richardson's discussion of case, the same cannot be said unequivocally for her discussion of aspect. Richardson starts out (9) by formulating the central question presented by perfective affixes as whether lexical and grammatical aspect are two distinct phenomena. That is, she outlines both her own view that there is a distinction between lexical and grammatical aspect in Slavic languages (which is in line with Smith Reference Smith1997; Filip Reference Filip1999, Reference Filip2005; Dickey Reference Dickey2000 and Borik Reference Borik2006), and her opponents' view that lexical and grammatical aspect should not be treated as separate phenomena (see Brecht Reference Brecht, Michael and Richard1985, Schoorlemmer Reference Slabakova1995, Bohnemeyer & Swift Reference Bohnemeyer and Mary2004, Borer Reference Borer2005 and van Hout Reference Hout, Melissa and Penelope2007, among others). Note that the way in which Richardson formulates the question is very different from identifying perfective prefixes as markers of either lexical aspect (the opponents' view) or grammatical aspect (Richardson's view). It might in fact be more appropriate to present the approach of the second group of authors as one that does discriminate between the two types of aspect, but proposes that only lexical aspect is marked overtly while grammatical aspect is expressed by word order or supplied by discourse context.
Another puzzling decision in this chapter concerns the definition of telicity. Richardson opts to define telicity as involving a telos, i.e. an inherent endpoint to the event, rather than non-homogeneity of the event, although she cites both approaches. Assuming the telos definition, Richardson categorizes inceptive verbs such as zarabotat' ‘to start working’ as atelic (54), due to the fact that they do not have an endpoint. However, the event of something not working and then starting to work involves a change of state and is hence not homogeneous. Consequently, when using the non-homogeneity definition, zarabotat' ‘to start working’ must be classified as telic. It is because of cases like this that Borer (Reference Borer2005) and Borik (Reference Borik2006) have cogently argued for the advantages of defining telicity in terms of non-homogeneity. As a final point in this chapter, Richardson makes the case for treating the quantization of the object as inducing telicity in some Slavic languages, using examples from Russian and Bosnian/Croatian/Serbian (22).
Richardson's arguments for the above claim invoke Dowty's (Reference Dowty1979) Imperfective Paradox (although attributed not to Dowty but to Borik's (2002) dissertation). The crucial sentences are (1) and (2).
Since the matrix clause in (1) has only an ongoing but not a present perfect inference (i.e. Ivan had not already eaten up the apple), the VP jel jabloko ‘ate an/the apple’ cannot have a complete interpretation. The matrix clause in (2), on the other hand, has both interpretations (i.e. Ivan was drinking and had already drunk some quantity of tea). Since both verbs are imperfective creation/consumption verbs, it must be the case that the quantization of the object is what makes the difference.
Chapter 3, ‘Case marking on the internal argument and lexical aspect’, starts out with definitions of the three types of perfective prefixes (the first member of the following verb pairs is imperfective, while the second member displays the different types of perfective prefix): (i) purely perfectivizing (e.g. stroit' ‘to be building’ versus po-stroit' ‘to build’), (ii) superlexical (e.g. chitat' ‘to read’ versus po-chitat' ‘to read for a while’), and (iii) lexical (e.g. bit' ‘to hit’ versus pere-bit' ‘to interfere, interrupt’). The observational adequacy of this division is beyond dispute. However, since Richardson argues that prefixes are on the whole grammatical aspect markers, she is forced to collapse the three types into two: purely perfectivizing and superlexical prefixes, on the one hand, which do not affect the telicity of the verb phrase, and lexical prefixes, on the other, which do. This amounts to the claim that some prefixes are lexical aspect markers, while others are grammatical aspect markers (although Richardson never formulates it this way). Next, Richardson makes a cogent case for prefixation being a syntactic (post-lexical) operation and, based on prefix-stacking facts, for a syntactic hierarchy of prefixes. In this chapter, the author elaborates the picture of compositional telicity as follows. In creation/consumption verbs, telicity may depend on the quantization of the object; it may also depend on the addition of a lexical prefix. For all other verbs, telicity is determined only by the addition of lexical prefixes. Lexical case-assigning verbs are always atelic. Many examples of such verbs are given in the appendix to the book. This is then the ‘case-aspect connection’, which constitutes the main line of reasoning in this volume and a valuable, innovative argument in the literature on Slavic aspect.
After disposing of potential counterexamples, Richardson provides a detailed syntactic, feature-based account of how exactly compositional aspect is derived. Instead of following Chomsky (Reference Chomsky and Michael2001) in distinguishing between interpretable and uninterpretable features, she uses Pesetsky & Torrego's (Reference Pesetsky, Torrego, Simin, Samiian and Wilkins2007) four-way distinction of features. In this system, features can be not only interpretable/uninterpretable but also valued/unvalued. Richardson proposes that in Slavic creation/consumption verbs, the functional category v has an unvalued interpretable [quantized] feature, which acts as a probe and Agrees with the [+/−quantized] feature of the object Determiner Phrase. In the case of creation/consumption verbs that are telicized by a lexical prefix, it is the prefix that supplies the value of v. In the case of non-creation/consumption verbs like push, stir, know, the lexical verb (V) is specified in the lexicon with a [−quantized] feature, and the quantization of the object does not matter. Thus, in all three cases, the accusative case is linked to the Agree relationship between an unvalued aspectual feature in v and features of elements within the vP (objects, lexical prefixes or base V). As for unaccusatives, Richardson assumes that they lack a vP-shell; hence, the unvalued quantized feature is (magically!) transferred to T(ense). This assumption allows Richardson to give a principled and neat, although somewhat stipulative, explanation of why East Slavic languages have optional accusative or nominative marking on the sole argument of unaccusative verbs, while West Slavic languages do not allow this.
Chapter 4, ‘Case and grammatical aspect in East Slavic depictives’, takes the claims of chapter 3 and applies them to the depictive construction in East Slavic languages (Russian, Ukrainian, Belarusian), which is exemplified in (3).
Richardson shows that if a depictive adjective is in the instrumental case, the state denoted by it is interpreted as bounded in time (e.g. in (3), Ivan was cured as a result of being in the hospital, i.e. a change in his health has occurred), whereas if there is case agreement between the depictive adjective and the subject, the adjective is interpreted as unbounded (e.g. in (3), Ivan may not even have been sick when he went to the hospital, i.e. no change of state occurred). She argues that these two interpretations are linked to a grammatical aspect contrast and proposes that there is an aspectual projection AspP situated above the depictive Adjective Phrase (AP). This AspP has two functions. First, it is responsible for instrumental case, that is, the instrumental case of the depictive adjective is the morphological reflex of a [+bounded] AspP. Second, the AspP serves as an escape hatch for wh-movement of the depictive adjective and is thus a phase in the sense of Chomsky (Reference Chomsky and Michael2001).
In chapter 5, ‘Extensions of the link between case and grammatical aspect’, exactly the same mechanism is extended to participial constructions, as in (3), and copular constructions.
Again, the instrumental case is argued to be the morphological reflex of a [+bounded] AspP.
Chapters 4 and 5 offer ample examples from Russian, Ukrainian and Belarusian; numerous native speakers have been consulted, and Richardson reports even their contradictory judgments. These two chapters should be required reading for any student of Slavic linguistics who needs to understand these constructions, because they nicely distill very complex data and interpretations.
Let me now take stock and summarize what aspectual projections Richardson proposes for the Slavic clause, and what morphological manifestations they have. Within the vP, there is the unvalued, but interpretable, [quantized] feature that can get valued by a lexical prefix, by an object, or by a [−quantized] V. The quantization Agree relationship is responsible for the telic meaning. The [±bounded] feature is checked in an AspP projection, which can be just below the matrix TP, or within the vP+VP shell but above adverbial and participial projections that the VP takes as arguments. The quantization calculus is morphologically related to accusative case; the boundedness calculus is related to instrumental case.
This is a very plausible picture, if we correct some claims that Richardson makes but contradicts herself later in the book. First, she defines purely perfectivizing prefixes as follows: they ‘bound the action described in the verb in time, but do not affect the lexical aspect (i.e. the telicity or atelicity) of the verb’ (53). An example is stroit' – po-stroit' ‘to be building – to build’. Later in the book, Richardson gives the pair pit' – vy-pit' ‘to be drinking – to drink up’ as an example of a verb with a lexical prefix which participates in the quantization calculus within the vP (95, examples (75), (76)). However, as even beginning learners of Russian can attest, let alone native speakers, there is no difference in the aspectual function of these verbs. Both po- and vy- are purely telicizing prefixes because they signal completion, not boundedness in time. Neither vypit' nor postroit' can denote eventualities that are interrupted without being complete, as the (infelicitous) example in (3) indicates.
Furthermore, these prefixes do not change the lexical meaning of the verb but just make the verb phrase heterogeneous (add a telos). If a prefix like vy- in vypit' is a telicity marker, in accordance with Richardson's own argument, then it is hard to see why a prefix like po- in postroit' above would not also be a telicity marker. I suggest that Richardson's proposal should be amended as follows. Both purely perfectivizing and lexical prefixes are telicity markers and reside in aspectual projections within the vP; only superlexical prefixes like po- in po-chitat' ‘to read for a while’ are markers of boundedness and thus reside in a grammatical aspect projection above the vP (contra Richardson's proposal on page 62, example (12)). Therefore, Richardson's claim that perfective prefixes are boundedness markers and not telicity markers (16) is not warranted and is in fact contradicted by her own arguments later in the book.
In summary, there is quite a lot in this book to recommend to the Slavic linguistics scholar: its wealth of data on a number of intriguing constructions in different Slavic languages, its insight as to the connection between structural case and aspect, its lucid writing style and its syntactic treatment of aspect. The latter brings Slavic perfective prefixes into the typological fold of marking temporal-aspectual meanings in dedicated functional projections, and shows that aspect marking in Slavic is indeed fully comparable, although not identical, to that in Germanic languages.
