1. Introduction: nominal aspect

For decades, verbal aspect, nominal quantity and the interplay between the two have been popular research topics in linguistics (a thorough survey of the literature is given in Sasse Reference Sasse2002). In particular, formal semantics (for in-depth studies see e.g. Verkuyl Reference Verkuyl1972, Reference Verkuyl1993; Krifka Reference Krifka, Bartsch, van Benthem and van Emde Boas1989) has studied the interdependence of aspectual and quantificational factors, demonstrating that aspect must be considered a compositional phenomenon. Many analogies between the aspectual (verbal) and quantificational (nominal) domains have been pointed out, and it has been demonstrated in detail how clause-level aspect depends on the contribution of many clausal elements, not just the verb. Especially Verkuyl (Reference Verkuyl1972, Reference Verkuyl1993) has emphasized the significance of the arguments of the verb for aspect. The situation is perhaps clearest in transitive clauses in which the object is an incremental theme (see Dowty Reference Dowty1991) and thus participates in the event gradually. For instance, in She was painting the fence the activity proceeds along the fence gradually and reaches its endpoint when the whole fence has been painted. The extent of the fence (its quantity) thus sets boundaries to the duration of the painting, since the painting can only last as long as there remains fence to be painted. Such dependence of clausal aspect on nominal quantity will be referred to in the present paper as nominal aspect.

Among phenomena related to nominal aspect, the concept of boundedness is of central importance. As is also well known, the opposition between bounded and unbounded is common to both verbal aspect and nominal quantity (see Talmy Reference Talmy2000: 42–88 for a detailed comparison). In the aspectual domain, boundedness depends first and foremost on the verb and its inherent aspectual nature, also known as Aktionsart. At the clausal level, however, other factors such as aspectual constructions (e.g. the progressive), durative modifiers (for two hours), and locatives indicating a source or a goal contribute to the aspectual interpretation. In recent literature on aspect, it has become customary to distinguish two different levels of aspect. Sasse (Reference Sasse2002: 203–204) calls these aspect₁ (viewpoint aspect, including the perfective/imperfective dichotomy and its associates, often expressed by grammatical means) and aspect₂ (the Aktionsart type of aspect, more directly dependent on the lexical meaning of the verb). Smith (Reference Smith1997: 3) draws a similar distinction between situation types (Sasse's aspect₂) and viewpoint types (Sasse's aspect₁). She distinguishes five main types of situations: states (static, durative, e.g. know), activities (dynamic, durative, atelic, e.g. dance), accomplishments (dynamic, durative, telic, consisting of a process and its outcome, e.g. build), semelfactives (dynamic, atelic, instantaneous, e.g. sneeze) and achievements (dynamic, telic, instantaneous, e.g. find). States and activities are atelic and thus inherently unbounded, i.e. have no natural end point. Accomplishments, semelfactives and achievements are bounded situation types: they have an endpoint which they reach either after a gradual development (accomplishments) or instantaneously (achievements and semelfactives). A crucial difference between them is telicity: the completive types accomplishments and achievements bring about an end result, whereas semelfactives are atelic. This means that in the completive event types the state that follows the event differs from the state that precedes it, whereas in semelfactives the two states are (conceptualized as) similar. Smith also distinguishes three viewpoint aspect types that the speaker may select when viewing a situation: these are the perfective viewpoint (viewing a situation in its entirety, including both initial and final endpoints if such exist), the imperfective viewpoint (viewing only part of a situation, including neither the initial nor the final endpoint) and the neutral viewpoint (flexible, including the initial endpoint of a situation and at least one internal stage).

A rapidly growing literature has addressed the phenomenon called aspectual coercion, (informally) the alteration of the aspectual meaning of a linguistic element (typically a verb) to semantically match its context. Coercion takes place, for instance, when a telic verb is used in a progressive construction (for details, see e.g. Pustejovsky & Bouillon Reference Pustejovsky and Bouillon1995, de Swart Reference Swart1998 and Michaelis Reference Michaelis2004). De Swart (Reference Swart1998: 349) defines aspectual coercion as ‘an implicit, contextually governed process of reinterpretation which comes into play whenever there is a conflict between the aspectual nature of the eventuality description and the input condition of some aspectual operator’. Topics studied under the rubric of coercion within the formal semantics tradition cover phenomena ranging from verbal Aktionsart to nominal quantity.

In the nominal domain, boundedness depends on whether the NP refers to a closed (bounded) or an open (unbounded) quantity. For clarity, I will follow Larjavaara (Reference Larjavaara1991) in using the terms open vs. closed to refer to this nominal quantificational opposition, and reserve the terms bounded and unbounded for verbal aspectual use. In incremental situations, open quantity gives rise to unbounded aspect and closed quantity to bounded aspect. Another pair of terms used for this purpose were introduced by Krifka (Reference Krifka, Bartsch, van Benthem and van Emde Boas1989), who distinguishes quantized (closed/bounded) and cumulative (open/unbounded) expressions in both the aspectual and the quantificational domains. In the nominal domain, only mass nouns and plural forms are typically able to indicate both open and closed quantities (of substances and multiplicities), whereas singular count nouns refer to indivisible entities, which inherently constitute closed quantities. However, indivisible entities may also participate in events incrementally, as in The potter was making a flower pot.

With mass nouns and plurals, the opposition between open and closed quantities also correlates with definiteness: definite mass nouns and plural forms generally indicate closed quantities, while indefinite ones indicate open quantities. For instance, in The flour is in the cup the definite subject NP refers to a closed quantity of flour, whereas in There is flour in the cup the quantity of the flour is open (‘some flour’). Count nouns, on the other hand, always indicate a closed quantity, since their referent is conceived as an indivisible whole. This closedness does not correlate with definiteness, cf. The strawberry [closed quantity, definite] is in the cup vs. There is a strawberry [closed quantity, indefinite] in the cup (for such relationships between definiteness and quantification, with special reference to English and Finnish, see Chesterman Reference Chesterman1991). However, as is well known, the boundary between mass and count nouns is fuzzy, and the speaker has the ultimate option of representing even prototypical count nouns as mass-like, as in the following example from Langacker (Reference Langacker1991: 73): After I ran over the cat with our car, there was cat all over the driveway. According to de Swart (Reference Swart1998: 352; see also Bach Reference Bach1986), such examples also involve a coercion, and the literature uses the notions of a universal grinder and a universal packager to reflect (respectively) the nature of the operations at work when we use mass nouns in a count-noun-like fashion, and when we use count nouns as mass-like.

In her cognitive linguistic study of Russian aspect, Janda (Reference Janda2004) makes a detailed comparison between nominal referents (in her terms, the idealized cognitive model (ICM) of matter) and verbal aspect. Her general objective is to explain aspectual phenomena as a metaphor of the ICM of matter. She draws a basic distinction between discrete solid objects and fluid substances (basically the count vs. mass distinction), and argues that a similar distinction can be made in aspect, with perfective aspect corresponding to a solid object and imperfective aspect to a fluid substance. Crucial features of a discrete solid object include the fact that it has inherent shape and edges, cannot be easily penetrated, cannot occupy the same place as another discrete object, is stable, and can be grasped (Janda Reference Janda2004: 475–476). Another feature that distinguishes solid objects from fluid substances is what Janda calls ‘streamability’: fluid substances can ‘stream’ and participate in events gradually (incrementally), whereas solid objects cannot. What makes Janda's work especially fruitful for studies dealing with the interplay of nominal quantity and aspect is that she not only discusses these domains separately but also introduces many conceptual operations that are available to the speaker for manipulating these notions. Such operations may result, for instance, in the conceptualization of a solid object as fluid-substance-like (‘pulverization’, cf. Langacker's example above), or the conceptualization of a fluid substance as filling a container with firm boundaries, or the placement of a solid object within a fluid substance, and so on.

In spite of these similarities between verbal aspect and nominal quantity, there are also important differences regarding the nature of boundedness (closedness) in the two domains. Perhaps the most important difference is that verbal boundedness is always boundedness in time and correlates with the temporal features of the event: whether the event is ongoing or terminated (imperfective vs. perfective), and whether it has an inherent endpoint or not (atelic. vs telic). Nominal quantity, on the other hand, is not necessarily distributed over time, though it may be, if the referent participates incrementally. In the verbal domain, boundedness is both a lexical semantic feature of particular verbs and a feature of concrete usages of verbs (or of whole clauses). Though these levels correlate, they need not always coincide, as shown by the extensive studies on aspectual coercion (e.g. Michaelis Reference Michaelis2004 and the references mentioned there). For instance, the verb run indicates a basically unbounded activity, but sentences like He ran a mile or He ran from school to the station refer to a bounded situation that ceases after the indicated distance has been traversed. In the nominal domain, only count nouns indicate a closed quantity as part of their lexical meaning. In contrast, mass nouns are neutral with respect to quantity, since mass nouns receive their quantificational interpretation only in concrete usage events. For instance, the noun flour designates a substance, but without the context of an actual usage event it is impossible to tell whether the quantity of the substance is to be conceived as closed (as in The flour is in the cup) or open (as in There is flour in the cup).

Typically, the referent of a count noun participates in the situation throughout and does not set a boundary to its duration. For instance, in The girl patted the dog, the quantity of the participants ‘girl’ and ‘dog’ does not set boundaries to the duration of the ‘patting.’ Both the girl and the dog participate in the event throughout as wholes (not incrementally), both already exist before the event and continue their existence after the event. Neither of the participants is ‘consumed’ during the event. This conceptualization is especially common and natural if the referent of the NP is a solid object, whereas fluid substances more typically participate gradually.

Indeed, the situation is different when a nominal referent (typically a substance) participates in the event incrementally. In this case the extent of the referent also determines the duration of the event. An incremental participant is ‘consumed’ (either concretely or figuratively) during the process, as in The boy was eating a sandwich or The professor was reading the thesis, where the process reaches its endpoint when the whole referent of the object has participated in it. Here the participation of the nominal referent is incremental in an objective sense, i.e. the eating and reading are actually affecting (or directed at) different parts of the sandwich or the thesis at subsequent points of time. Since a sandwich and a thesis are bounded entities, the activities that involve them as incremental participants are understood as (aspectually) telic: they reach their endpoint when the whole entity has participated. However, an incremental participant can also be quantitatively open, resulting in aspectual unboundedness of the situation; here the situation is continuously ‘fed’ by the incremental participant with new substance, as in Water was pouring out of the drain. Indeed, as pointed out by Verkuyl (Reference Verkuyl1993) and Krifka (Reference Krifka, Sag and Szabolcsi1992: 50), it may not be the activity of ‘eating’ or ‘reading’ as such that constitutes the boundedness in examples like He ate the sandwich or She read the book; rather, it is the nature of the object that creates the boundedness (i.e. we typically eat entities or portions of some kind, and read books, papers and other things, whose extension is bounded). It is also possible that the incremental participation of a nominal referent in a situation is based on subjective factors, such as an incremental scanning of a nominal referent which is not affected itself or does not undergo a change in the course of the event (see Talmy Reference Talmy2000: 71 for details).

In the following sections, I will study nominal and verbal aspect and their effects on the case marking of the object in Finnish, which reflects both aspectual and quantificational factors. I begin by introducing the classical view of Finnish object marking in section 2, where it is shown how the case marking is determined by a complex interplay of aspect and quantity, and often results in massive ambiguity. In section 3, I discuss certain problematic usages that seem to contradict the general principles of object marking. These usages often (though not always) involve incremental objects. In section 4 I discuss examples where it is the subject (not the object) whose referent participates incrementally, and show how this affects the object marking. In section 5 I return to one of the problematic object types discussed in section 3, the so-called quasi-resultative type, and show how this static type can be associated with dynamic types where either the subject or the object indicates an incremental participant. Section 6 sums up the results of the study.

2. The Finnish object marking system: an overview

One of the most problematic issues in Finnish grammar is without doubt the case marking of the object (for literature in English, see e.g. Denison Reference Denison1957; Heinämäki Reference Heinämäki, Groot and Tommola1984, Reference Heinämäki, Bache, Basböll and Lindberg1994; Helasvuo Reference Helasvuo2001; Huumo Reference Huumo2005). The case marking of the object is part of a more extensive system of oppositions among the grammatical cases of the language. The partitive case marks the object under three conditions: a) in negated sentences, b) in aspectually unbounded sentences (more precisely: atelic, progressive, cessative, and irresultative (semelfactive) sentences), and c) in sentences where the object NP refers to an open, indefinite quantity. Its counterpart is a morphologically heterogeneous category which will be referred to here (following the recent descriptive grammar ISK Reference ISK=Auli, Vilkuna, Korhonen, Koivisto, Heinonen and Alho2004) as the total object, designating what might be called completeness of quantity or aspect or both. The case opposition between the partitive and other grammatical cases works not only in the marking of the syntactic object (Heinämäki Reference Heinämäki, Groot and Tommola1984, Reference Heinämäki, Bache, Basböll and Lindberg1994; Chesterman Reference Chesterman1991; Huumo Reference Huumo2005), but also of the existential subject (see Helasvuo Reference Helasvuo2001, Huumo Reference Huumo2003) and the predicate nominal (Huumo Reference Huumo2009), though under different conditions. Furthermore, the ‘total’ cases that participate in the system vary from one subtype to another. In existential subjects and predicate nominals, the total-case counterpart of the partitive is always the nominative. In object marking, the situation is more complicated: the total object can be marked with either the nominative, the genitive, or the irregular accusative form of personal pronouns (ending -t). Historically the object-marking genitive was an accusative case with the ending *-m, but it coalesced with the genitive -n as the result of a sound change whereby word-final -m became -n. This sound change caused a reanalysis of many infinitival constructions in the language, such that many accusative-marked objects of finite verbs were reinterpreted as genitive-marked subjects of infinitives (see e.g. Anttila Reference Anttila1989: 103–104).

The choice of the case is based on an interplay of multiple factors, including quantity, definiteness, the count vs. mass distinction, negation, subject-verb agreement, and aspect. Among these factors, quantity and the count vs. mass distinction are involved in all three main syntactic contexts of the case opposition (existential subject, predicate nominal and object). Negation requires the partitive marking of objects and existential subjects but not of predicate nominals. Pure aspect (i.e. without quantity) affects the case marking of the object only. In the following discussion, I will not deal with predicate nominals (for these, see Huumo Reference Huumo2009) but will focus on object marking; subjects will be treated in section 4.

As mentioned above, there are three main factors that affect the case marking of the object: quantity, aspect, and negation. The received view of Finnish grammar (e.g. Chesterman Reference Chesterman1991: 93; Vilkuna Reference Vilkuna2000: 119; ISK Reference ISK=Auli, Vilkuna, Korhonen, Koivisto, Heinonen and Alho2004: 887) maintains that negation is the strongest of these conditions, followed by aspect and quantity. That negation is the strongest condition is shown by the fact that (practically) all objects of negated clauses take the partitive case, whereas the case marking in affirmative clauses is determined by a complex interplay between quantity and aspect. The effects of quantity can be seen most clearly in expressions of a punctual achievement, since these indicate a situation that is indisputably bounded and brings about an end result; thus aspect will not be a reason for using the partitive (certain exceptions are discussed in Huumo, Reference Huumoforthcoming). Consider (1)–(3), whose punctuality is also reflected in the fact that they reject direct durative modifiers (which in Finnish are marked like the total object and correspond semantically to the English for an hour type)Footnote ².

1. (1) Löys-i-n kirja-n (* minuuti-n). find-pst-1sg book-tot minute-tot ‘I found a/the book (*for a minute).’

1. (2) Löys-i-n voi-ta (* minuuti-n). find-pst-1sg butter-par minute-tot ‘I found some butter (*for a minute).’

1. (3) Löys-i-n voi-n (* minuuti-n). find-pst-1sg butter-tot minute-tot ‘I found the butter (*for a minute).’

In (1) the object is a count noun and refers to a bounded entity (thus to a closed quantity); therefore the total object must be used. The punctual nature of (1) also excludes an aspectual partitive. In (2) and (3) the object is a mass noun, which can be differentially quantified by means of the case opposition. The partitive in (2) indicates that the quantity of the butter is open (‘some butter’), whereas the total object in (3) indicates a closed quantity of a definite referent (‘the butter’).

At the other end of the aspectual scale we have atelic verbs that indicate an unbounded situation, i.e. an activity or a state. In such examples the unbounded aspect requires the partitive even if the object is a count noun or a mass noun indicating a closed quantity. Such examples do allow direct durative modifiers (consider examples (4) and (5); however, see also section 3.2 for exceptions).

1. (4) Pitel-i-n käde-ssä-ni kirja-a ~ voi-ta (tunni-n). hold-pst-1sg hand-ine-1sgpx book-par butter-par hour-tot ‘I was holding ~ held [a/the] book ~ [the/some] butter in my hand (for an hour).’

1. (5) Rakasta-n sinu-a ~ olut-ta. love-prs.1sg you-par beer-par ‘I love you.’ ~ ‘I love beer.’

Example (4) indicates an activity and example (5) a state. As shown by the English translation of (4), the object marking does not reveal whether the quantity of the butter is open or closed. This is because the unbounded aspect demands the partitive in any case. The aspectual function of the partitive thus ‘conceals’ the quantity of the referent. In (5) the situation is abstract and therefore the question of quantity does not arise: the object does not refer to an actual instance of beer that could be quantified but is understood as generic. Note, too, that the unboundedness of the situation in (4) can be understood in two different ways: as progressivity (‘was holding’) or as cessativity ‘held’, where the event has ceased without reaching an end point (Sasse Reference Sasse2002: 206 uses the term delimitative for this type of aspect). In some languages there are aspectual grams that specifically indicate delimitative meanings; for instance Russian has verb prefixes whose meaning is roughly ‘to do X for a while’ (see Dickey Reference Dickey, Divjak and Kochanska2007 and Janda Reference Janda2007, who calls such events complex acts). As example (4) shows, there is no specific way of indicating cessativity in the Finnish system; this is just one possible interpretation of the polysemous partitive object. However, the fact that it is the partitive and not the total object that is used to indicate cessativity clearly shows how the total object requires more than just a perfective viewpoint (i.e. the fulfilment of an accomplishment) to be available.

Examples like (4) also demonstrate that a partitive object is able to indicate pure quantity only if it does not indicate unboundedness, that is, only in aspectually bounded examples like (2). Similar observations have been made in the literature on aspectual coercion. For instance, de Swart (Reference Swart1998: 349) argues that a progressive predication such as Anna was running a mile does not licence the inference that Anna actually ran a mile; when focusing on the ongoing activity, the progressive conceals the projected end result of the event. Thus, according to de Swart, the semantics of the progressive involves ‘stripping an event of its culmination point’ (p. 355). Similarly, the Finnish partitive in unbounded sentences often conceals the projected overall quantity of the object NP. Unlike the object in examples (2) and (3), the object in example (4) is unable to indicate the quantity of its referent by case alteration, because the unboundedness of the aspect excludes the total object altogether. Among the conditions motivating object case marking, accordingly, aspect appears to be stronger than quantity. This is why many grammars of Finnish (e.g. Chesterman Reference Chesterman1991, Vilkuna Reference Vilkuna2000) represent the three conditions for the partitive object in the form of the hierarchy negation>aspect>quantity, where negation is the strongest condition (requiring the partitive irrespective of quantity and aspect) and aspect is the second-strongest (requiring the partitive irrespective of quantity). However, as the following discussion will show, this hierarchy is a simplification, and matters are in fact more complicated.

Leaving aside punctual and atelic examples for the moment, I will now discuss the case marking in examples that indicate accomplishments, i.e. non-punctual telic events that develop gradually towards their endpoint. In such examples the total object indicates the combination of bounded aspect and closed quantity, whereas the partitive usually has many possible interpretations. In this type, the partitive but not the total object allows a direct durative modifier in the same clause. Consider examples (6) and (7), where the star in parentheses indicates the ungrammaticality of the total object version in the presence of the durative modifier.

1. (6) Lu-i-n kirja-a ~ (*) kirja-n (tunni-n). read-pst-1sg book-par book-tot hour-tot ‘I was reading the book ~ read the book (for an hour).’ [par] ‘I read the [whole] book.’ [tot]

1. (7) Sö-i-n puuro-a ~ (*)puuro-n (tunni-n). eat-pst-1sg porridge-par porridge-tot hour-tot ‘I was eating (the) porridge’ ~ ‘I ate (some of the) porridge (for an hour).’ [par] ‘I ate up the porridge.’ [tot]

As the English translations indicate, the partitive indeed has a considerable variety of possible interpretations. First, it can have the aspectual function of indicating either a progressive or a cessative reading where the event has been interrupted (terminated before reaching its possible endpoint). These readings are both possible in the partitive versions of (6) and (7), as the English translations show (‘I was reading the book’ vs. ‘I read the book (without finishing it)’ in (6); ‘I was eating the porridge’ vs. ‘I ate (some of the) porridge’ in (7)). Another reading of the partitive is the quantificational one, which applies if the event itself has been completed but the quantity of the object NP is open. When considering examples (6) and (7), it should be kept in mind that the hierarchy negation>aspect>quantity represents the situation from the viewpoint of the partitive object, not of the total object, which can only be used if the sentence does not meet any criterion of the partitive (however, in section 3.2, this principle too turns out to be problematic). Thus aspectual factors (boundedness) alone cannot trigger the total object unless the quantity indicated by the object NP is closed.

If we consider example (7) in more detail, we see that the alternative interpretations of the partitive set up another kind of a hierarchy. With the progressive reading, the partitive object conceals the (projected) quantity of the object referent: the example does not tell whether the projected endpoint of the event would involve a closed or an open quantity of porridge (the opposition between ‘I was eating porridge’ and ‘I was eating the porridge’). The same ambiguity arises in the cessative interpretation where the activity has been interrupted: again, we do not know whether the event had been proceeding towards an endpoint involving a closed quantity (‘I ate some of the porridge’) or not, in which case the partitive object indicates open quantity of the porridge eaten so far (‘I ate some porridge’). In sum, examples (6) and (7) show that the total object is used only if the situation has both reached its endpoint and has affected a closed quantity (‘I ate up the porridge’). They also show that the multifunctional nature of the partitive often gives rise to ambiguity, and when indicating a more dominant function (e.g. progressivity) the partitive conceals less dominant features that it can indicate in other contexts (quantity).

Finally, it needs to be pointed out that, in addition to progressivity and cessativity, there is yet another aspectual function for the partitive, which in the Finnish tradition is referred to as irresultativity (discussed in Huumo Reference Huumoforthcoming). An irresultative situation is (typically) punctual (at least bounded) but fails to bring about a substantial end result that would motivate the use of the total object and is thus atelic. Its opposite is a resultative situation which is conceptualized as involving a fundamental, often irreversible change of state in the object referent. On the basis of the resultativity vs. irresultativity opposition, Finnish punctual verbs can be divided into two groups: resultative (accomplishment/achievement) verbs like tappaa ‘kill’, ostaa ‘buy’, huomata ‘notice’ and löytää ‘find’ that take the aspectual total object, and irresultative (semelfactive) verbs like tönäistä ‘nudge’, mulkaista ‘glance’ and lyödä ‘hit’ that take the aspectual partitive object. The resultative total object thus indicates a transition: the event brings about a change, after which it does not return to its original state but enters another one. The partitive, on the other hand, indicates that no such transition takes place and that immediately after achieving its target state the situation returns to the original state (or to a state that is conceptualized as similar to the original state). Croft (Reference Croftforthcoming) calls these different situation types directed achievements and cyclic achievements, respectively, and points out that directed achievements can be further divided into two classes: reversible ones, such as open, and irreversible ones, such as break or die. A different terminology is used by Smith (Reference Smith1997), who draws a distinction between two punctual situation types, achievements (which result in a change) and semelfactives (which do not bring about a change). At a crude level, such distinctions also reflect the resultativity opposition between the Finnish object markers. However, there are certain uses of the partitive that do not properly fit such a classification. For instance, Leino (Reference Leino1991) points out that the verb vaihtaa ‘change’ often takes the partitive object in cases where one realization of the object replaces another and where the two realizations both belong to the same category – even though such a change would count semantically as an accomplishment (8).

1. (8) Vaihdo-i-n paita-a.change-pst-1sg shirt-par ‘I changed my shirt’.

In example (8) the person takes off one shirt and puts on another, which would be expected to count as an accomplishment, but nevertheless the partitive object is used. What is also noteworthy in (8) is the number of the object: it is in the singular though actually there are two referents.

Other exceptional instances include punctual verbs that allow both a resultative and an irresultative interpretation of the situation they indicate (and thus an alteration between the partitive and the total object). The best-known example is the verb ampua ‘shoot’ in (9), discussed by Heinämäki (Reference Heinämäki, Groot and Tommola1984) .

1. (9) Ammu-i-n lintu-a ~ linnu-n.shoot-pst-1sg bird-par bird-tot ‘I was shooting [aiming at] the bird.’~ ‘I shot [at] the bird [but failed to kill it].’ [par] ‘I shot the bird [dead].’ [tot]

With the total object, the verb ‘shoot’ indicates resultativity, i.e. the death of the bird as a result of the shooting. With the partitive object it indicates the irresultativity of the shooting. In addition, the partitive also has a prospectiveFootnote ³ reading where it indicates a phase that precedes the actual event. Thus (9) also shows that achievement verbs must be divided into two classes: proper achievements that do not allow the partitive object to indicate a prospective viewpoint aspect (e.g. löytää ‘find’ in examples (1)–(3)) and extended achievements that allow the prospective viewpoint aspect.

The resultative vs. irresultative opposition is also relevant with certain verbs that indicate the causation of a gradual change-of-state in the patient, e.g. lyhentää ‘shorten’ and lämmittää ‘warm’. Smith (Reference Smith1997: 24) calls such events degree predicates and Croft (Reference Croftforthcoming) directed activities; and both argue that such verbs indicate a directed change that has no natural endpoint. However, an endpoint can be conceptualized into the event, and this opposition is reflected in the Finnish object marking: with the total object these verbs indicate an abrupt, absolute change (‘make short/warm’), but with the partitive object they indicate a gradual, scalar change with no particular endpoint (‘make [somewhat] shorter/warmer’). The situation is further complicated by the fact that these verbs are not punctual but durative, and therefore the partitive can also receive the progressive reading. Consider (10).

1. (10) Lyhens-i-n hame-tta ~ hamee-n.shorten-pst-1sg skirt-par skirt-tot ‘I was shortening the skirt.’ ~ ‘I shortened the skirt [somewhat].’ [par] ‘I shortened the skirt.’ [=made it short] [tot]

It should be pointed out that the aspectual opposition here called resultativity vs. irresultativity is quite different from the opposition observed in accomplishments: here it is related to the nature of the end result of the event, and the relevant factor is not whether the event has been brought to its possible endpoint or not (as in ‘reading a book’) but the nature of the endpoint itself. In many grammatical descriptions of Finnish, however, the term irresultative has also been used to refer to durative events that have not been brought to their endpoint, and it needs to be emphasized that my usage of the term is somewhat narrower. It is also worth noting that although (10) would most naturally be classified as an accomplishment, its object marking seems to correspond to that of accomplishments only partially (i.e. when the partitive indicates progressivity). The fact that the irresultative reading is also possible for the partitive shows that such situations can also be conceptualized as semelfactives.

Next consider negation. Keeping in mind that negation requires all objects to be marked with the partitive, we can expect that negated examples will have even more possible readings than affirmative examples like (10). Indeed, as pointed out by Itkonen (Reference Itkonen1976), examples like (11) may have as many as six possible readings, taking into account both aspect and quantity, if we consider all alternatives from the viewpoint of the corresponding affirmative sentence that has been negated.

1. (11) En voidel-lut ~ suks-i-a.neg.1sg wax-pst.conneg ski-pl-par ‘I was not waxing ~ did not wax [the] skis.’

Note first that the affirmative counterpart of (11) could have either the total object (‘I waxed the skis’) or the partitive object (‘I was waxing [the] skis’; ‘I waxed some [of the] skis’). The total object would indicate bounded aspect together with a closed quantity, which in turn could be understood in two ways: it could consist either of one pair of skis or of a larger but closed quantity of (pairs of) skis. The affirmative partitive in turn could have two aspectual readings (the same as in example (10)). First, it could have the progressive reading, which would again conceal whether the event is conceptualized as an accomplishment or an activity, i.e. whether it is progressing towards a closed quantity of skis or not (‘I was waxing the skis’ vs. ‘I was waxing skis’). Second, the partitive could have the cessative reading, maintaining the same ambiguity of the overall quantity (‘I waxed some [of the] skis’). The negative partitive conceals all these possibilities and focuses only on the fact that the event did not take place at all.

To sum up, it is easy to see that the general function of the partitive is to indicate the incompleteness of the situation in one way or another (a generalization suggested by Leino Reference Leino1991): either the situation does not take place at all (negation), or it does not reach an endpoint (atelic/cessative/progressive), or it reaches its endpoint but does not affect a closed quantity of entities. It is also noteworthy that the partitive often implies either an actual or a potential continuation of the event. In the progressive reading, the event is actually ongoing, and in the cessative reading it could potentially be carried on further. In the reading involving an open quantity, the event could (in many cases) be continued so as to affect a larger quantity of entities. It is only the irresultative partitive that lacks such an implication – this is because the event type indicated by the verb is itself one lacking a relevant end result. The total object, on the other hand, indicates that the event has been brought to its endpoint and that the activity has affected a closed quantity in full. Thus a potential continuation of the event is excluded, though it may be possible to repeat the event (e.g. to read a book again).

The above discussion also shows that the lexical aspectual situation type of the examples is in many ways crucial to the selection of the object marker. In contrast, viewpoint aspect can be explicitly expressed by the object case only in certain situation types. The unbounded situation types, i.e. states and activities, require the partitive object irrespective of quantity and viewpoint aspect (examples (4) and (5); exceptions will be discussed in section 3). In unbounded situation types the partitive thus cannot unequivocally indicate viewpoint aspect, though it can often be interpreted with this kind of a meaning. Among the bounded situation types, accomplishments trigger the total object if the situation has been brought to its endpoint and if the quantity of the object NP is closed and its entire referent has participated in the event. In accomplishments, the partitive can indicate a cessative or progressive viewpoint aspect, or (if not interpreted as indicating one of these viewpoint aspects) an open quantity of the object referent. The two punctual situation types, achievements and semelfactives, do not allow the progressive or cessative readings for the partitive, since the situations they indicate are not durative. Proper achievements (‘notice’, ‘find’) take the aspectual total object (or the quantificational partitive), but extended achievements (in Finnish, verbs meaning e.g. ‘shoot’, ‘buy’, ‘kill’) allow the aspectual partitive to indicate a prospective viewpoint aspect – that the event is about to take place or that the referent of the subject is attempting the achievement. Semelfactives always take the partitive object, since in spite of their temporal boundedness they are atelic and lack an end result. However, as example (8) shows, the division of punctual verbs into achievements and semelfactives is far from clear-cut, and as example (9) shows, some punctual verbs (‘shoot’) can take both kinds of object marking depending on the conceived nature of the end result. An even more complicated case in point involves so-called degree predicates like ‘warm’, ‘lengthen’ or ‘shorten’ (example (10)). These resemble accomplishments in that they are durative and thus allow the progressive partitive (10). However, the object marker also participates in the resultative vs. irresultative opposition with degree predicates: it can reflect the conceived nature of the end result (gradual vs. absolute). The overall situation is summarized in table 1, which indicates the readings available for the partitive object in each situation type.

Table 1

The functions of the partitive object

Key: Prog.=Progressive; Cess.=Cessative; Prosp.=Prospective; Irres.=Irregulative

In each row of the table, the possible readings of the partitive are marked with a plus (+) mark. As can be seen, there is a plus mark immediately following the name of certain situation types. These are the atelic situation types that exclude the total object altogether: activities, states and semelfactives. The other situation types, which do not have a plus mark after their name, allow viewpoint aspect and quantity to determine the case of the object. As can also be seen, some of the plus marks in the table are in parentheses. A bare plus mark indicates a function that can be the sole factor motivating the partitive object. Parentheses around the plus mark indicate that the partitive allows the interpretation with the function in question, in addition to another function. This happens in situation types that require the partitive in any case but can have different aspectual readings for it. For instance, the partitive object in an activity expression can have either a progressive or a cessative reading, but since the atelic nature of the situation type requires the partitive anyway, these functions remain only possible alternative interpretations. In contrast, the partitive object of an accomplishment verb can specifically indicate progressive or cessative aspect, or open quantity. The same principles work in quantification: in atelic situation types that require the aspectual partitive anyway, the partitive cannot unequivocally indicate open quantity (it remains indifferent to quantity). In contrast, the telic situation types allow the partitive to have a quantificational reading – but only if its viewpoint aspectual readings do not apply. If interpreted as indicating viewpoint aspect, the partitive again becomes indifferent to quantity.

Above I have discussed object case marking from the point of view of the partitive object, trying to pin down the factors motivating it within each aspectual type. Alternatively, the system can be approached from the point of view of the total object, the usage conditions of which turn out to be simpler: the total object is used in (affirmative) accomplishment and achievement expressions with a perfective viewpoint (thus excluding prospective, progressive and cessative viewpoints), on condition that the object indicates a closed quantity that has participated in the event in its entirety.

At a crude level, the opposition between the partitive and the total object thus seems to follow the principle suggested by Leino (Reference Leino1991): the partitive indicates incompleteness of the situation in one way or another. This explanation strives towards a central objective of all scientific investigation: generalization. However, an adequate explanation must also be able to deal with problematic instances, and one should not strive towards generalization at the expense of the facts. Though a widespread consensus prevails among scholars of Finnish grammar that the three factors of negation, aspect and quantity explain the case marking of the object, and that more general terms such as boundedness or incompleteness represent useful and valid generalizations, there remain a few problematic instances that are not easily explained by such general principles. These include particular uses of both the partitive and the total object. First, the partitive is used in certain examples that are aspectually bounded and cannot proceed further in time; thus the total object would be expected (Huumo Reference Huumoforthcoming). Second, the total object is used in certain examples indicating situations that are aspectually unbounded; thus the partitive would be expected. Some of these uses have been discussed in some detail in the literature, while others have escaped the attention of scholars.

In the following discussion it will be argued that the generalizations concerning Finnish object marking have given too much weight to what has been taken to be the prototype situation for these oppositions, that is, an aspectual opposition which involves as one pole a telic, non-punctual and thus an accomplishment-type situation that gradually proceeds towards its endpoint, finally reaches it and thereby attains completion. In English, such examples include I took ~ was taking the letter to the post office (non-incremental object) or I read ~ was reading the book (incremental object); to understand the Finnish object marking properly we need to pay attention to incremental and non-incremental instances in each opposition type.

The problematic instances to be discussed in the following sections involve expressions that indicate unbounded situations but nevertheless select the aspectual total object. In section 3 I will argue that the motivation for using the total object in such instances is indeed boundedness of aspect or quantity (or both), but a type of boundedness that differs fundamentally from the canonical kind that gets realized only after cumulating gradually over time. The type of boundedness that will be relevant here, to be called continuous boundedness, is one that obtains throughout the entire situation, at any given point of time during its duration. In section 3.1, I will discuss two types of usage of the total object that have received attention in earlier work, i.e. habitual expressions and so-called quasi-resultative sentences (Huumo Reference Huumo2005). Habitual expressions are aspectually dual in the sense that they indicate the iteration of an event that may itself be bounded. Since the quantity of the iterations is unbounded, the overall situation is also unbounded, even though its component situations may be bounded. It will be shown that in habitual examples, Finnish aspectual object marking reflects the nature of the component situations, not the overall habitual situation, although the duration of the latter can however be indicated by durative adverbs. The quasi-resultative type consists of stative situations which, according to the analysis of Huumo (Reference Huumo2005), are conceptualized as ‘frozen’ fictive accomplishments that extend over time. In section 3.2 the focus is on two usage types that have not been discussed previously: continuous completives (accomplishments/achievements) and continuously bounded quantities. It is in these latter two types that we need to draw a distinction between overallaspect (classical aspect, proceeding and cumulating in time) and another type of aspect, henceforth called the transient aspect, which is orthogonal to the time axis and thus gives a characterization of the ‘cross-section’ of the situation at any given point of time, i.e. whether the indicated relationship is conceptualized as bounded or unbounded at any given point of time during the event. It is argued that transient aspect plays a central role in Finnish object marking and needs to be taken into account in order to give a coherent picture of the grammatical systems involved in aspectual object marking.

3. The total object as indicator of continuous boundedness

The fact that the total object can be used in unbounded examples shows that the general rules of Finnish object marking fail to explain certain usages that are of crucial importance to the theory of aspect. In this section it is argued that even in such problematic instances the function of the total object is to indicate boundedness (of aspect or quantity), but of a type that differs from the canonical type where boundedness equates to the reaching of an endpoint where the event itself culminates. In the instances to be discussed below, boundedness prevails continuously: the situation does not develop towards an endpoint but consists of a continuous force-dynamic interaction (in the sense of Talmy Reference Talmy2000: 409–470) that serves to keep the situation complete at each point of time, as in the English example This brush cleans the conveyor belt before it enters the machinery (‘continuously keeps the conveyor belt in the state of total cleanliness’).

In the same way as canonical boundedness, continuous boundedness can be based either on aspect alone or on a combination of aspect and quantity. When the continuous boundedness of a situation is purely aspectual, the situation is conceptualized as a continuous completive (a higher-order term for accomplishments and achievements) where a force-dynamic interaction between an actor and an undergoer keeps the undergoer in a state conceived as completed, or resultative, as the classical Finnish terminology puts it (see also Huumo Reference Huumoforthcoming). This means that the situation cannot develop towards an even greater perfection; on the other hand, without the continuous force-dynamic interaction the situation would cease to obtain and the undergoer would leave its resultative state. Correspondingly, when the continuous boundedness of the situation is motivated by a combination of bounded aspect and closed quantity, the object is incremental and changes its reference through time but at the same time indicates closedness of quantity at each individual point of time. More concretely, such a sentence means that continuously all substance (or multiplicity of individuals) that enters the situation is brought into a new, resultative state by the force-dynamic interaction indicated by the verb.

In sections 3.2 and 3.3 I will discuss such examples in detail; I will refer to the two types as continuously closed quantities and continuous completives. First, however, I will discuss in section 3.1 two other types where the total object is used in unbounded sentences. These types have been dealt with in the literature to some extent, and they are closely related to the new types introduced here. The first type consists of habitual or generic examples, and the second type of so-called quasi-resultative examples (for these, see Huumo Reference Huumo2003). After this, I will proceed to continuous boundedness.

4. Effects of an incremental subject

In the previous sections I have discussed in detail the Finnish object marking from the viewpoints of aspectual boundedness of the situation and the quantity of the object referent. However, other NPs of the sentence may also contribute to such interpretations. In this section I will take a closer look at the effect of the subject on object marking. Like the object, the subject can be incremental (e.g. Water was dripping from the tank), and this may result in unbounded aspect (see also Verkuyl Reference Verkuyl1972). In Finnish transitive sentences, the subject must take the nominative case, and therefore features such as open quantity or unbounded aspect cannot be marked on subjects (e.g. by using the partitive). The situation is different, however, with intransitive/existential subjects, which do allow the use of the partitive to indicate open quantity and incremental nominal aspect (though not pure aspect; for details, see Huumo Reference Huumo2003 and the references mentioned there).

Since the focus of this paper is on object marking, I will only discuss intransitive or existential subjects briefly. It suffices to say that these subjects have similar quantificational options as object NPs: they can indicate open or closed quantities, and both kinds can participate in the event incrementally or non-incrementally. Consider the intransitive example (38) with an achievement verb and a nominative subject that indicates a closed quantity. This causes a bounded reading for the sentence, irrespective of whether the guests arrive all at once (collective reading) or one by one (distributive reading). Because the overall quantity of the guests is closed, the situation reaches its endpoint when all the guests have arrived. In contrast, example (39) has an existential partitive subject, which indicates an open quantity (and does not trigger number agreement in the verb which is in the singular), and therefore the distributive reading results in unboundedness of the overall situation (‘guests kept arriving’).

1. (38) Vieraa-t saapu-i-vat. guest-pl.nom arrive-pst-3pl ‘The guests arrived.’

1. (39) Viera-i-ta saapu-i. guest-pl-par arrive-pst.3sg ‘[Some] guests arrived./Guests kept arriving.’

In the same way as the total object in some of our earlier examples (e.g. (25)), the nominative subject can also indicate a closed transient quantity (‘continuously all’), as in (40).

1. (40) Iltapäivä-n aja-n museovieraa-t tul-i-vat sisä-än sivuove-sta. afternoon-gen time-tot museum.visitor-pl.nom come-pst-3pl in-ill side.door-ela ‘During the afternoon, the museum visitors came in through the side door.’

The example refers to a situation where continuously all visitors had to use the side door to enter the museum. In this respect it is similar to transitive examples like (25) above, where the total object indicates a closed transient quantity (though in (40) the closed transient quantity is not foregrounded by the case marking since the transitive subject is in the nominative anyway). Also its opposition with the partitive subject is similar to that in object marking. In the same way as the partitive object (if used in (25)), the use of a partitive subject in (40) would indicate an open transient quantity (‘for the whole afternoon there arrived guests through the side door’). In that case the focus would be local and restricted to the side door only (see Huumo Reference Huumo2003), and the sentence would not exclude the possibility that other guests arrived through other doors.

Now consider transitive examples that indicate a situation where the referent of the subject participates incrementally. In the following examples the subject refers to an elongated entity (‘freight train’) that gradually passes the referent of the object (‘semaphore’). In (41) the total object indicates that the situation is completed, whereas in (42) the partitive indicates progressivity.

1. (41) Tavarajuna sivuutt-i opastime-n.freight. train pass-pst.3sg semaphore-tot ‘The freight train passed the semaphore.’

1. (42) Tavarajuna sivuutt-i opastin-ta.freight. train pass-pst.3sg semaphore-par ‘The freight train was passing the semaphore.’

The referent of the subject in (41) and (42) is a bounded entity and thus its quantity is closed. Its incremental participation determines the duration of the event, and the progressive (42) indicates that the whole train has not yet passed the semaphore. As usual, the progressivity of the situation is marked on the object by using the partitive case. This may seem somewhat paradoxical, given the fact that in (42) it is the referent of the subject whose incremental participation gives rise to the incompletedness of the situation. However, the grammatical system of Finnish does not allow the marking of progressivity on the subject NP.

Now consider what happens if the subject is incremental and refers to an open quantity, each of whose (conceivable) components completes the aspectually bounded event designated by the verb. With a distributive reading, such incremental plural subjects give rise to an iterative interpretation, and in spite of the overall unboundedness, the total object can be used to indicate the boundedness of the individual component situations (43).

1. (43) Juna-t ohitta-vat nykyään asema-n pysähty-mä-ttä. train-pl.nom pass-prs.3sg nowadays station-tot stop-inf-abe ‘Nowadays trains pass the station without stopping.’

In (43) the total object indicates that each train accomplishes the event of passing the station. Example (43) resembles the generic and habitual examples discussed in section 3.1. In fact, (43) is understood as generic; the only difference from those earlier examples is that now it is the referent of the subject that changes in each component situation, whereas in our earlier examples it was the referent of the object. Example (43) shows, once again, how Finnish object marking pays attention to the internal aspectual nature of the individual component events, not to the overall unbounded nature of the holistic situation (which is unbounded). If we again compare Finnish with Spanish, we see that Spanish uses an imperfective verb form in similar examples, e.g. los trenes pasaban la estación sin parada ‘the trains passed[imperfective] the station without stopping.’ Thus, in contrast with Finnish, the Spanish system of aspectuality conceptualizes such situations at the holistic level, where they are unbounded. Finally, the unboundedness of example (43) is also shown by the fact that in spite of the total object it is possible to add a direct durative modifier that indicates the duration of the overall situation (44).

1. (44) Jo vuode-n aja-n juna-t o-vat ohitta-neet asema-n pysähty-mä-ttä. already year-gen time-tot train-pl.nom be-prs.3pl pass-ptcp station-tot stop-inf-abe ‘For a year already, the trains have passed the station without stopping.’

More problematic are cases where the incremental subject NP is headed by a singular mass noun and refers to a homogeneous substance where no individual referents can be distinguished. From this it follows that the situation cannot include iterative components: there are no distinguishable component referents to perform such component events. Recall again our earlier examples (25)–(27) (the ‘buying electricity’ type), where the object NP referred to a substance whose transient quantity was closed but whose overall quantity was open. We saw that in such instances Finnish uses the total object to indicate the closedness of the transient quantity, in spite of the fact that no individual accomplishments can be distinguished. Remarkably, we find the total object also in examples where the subject refers to such an open overall quantity which ‘streams’ through the situation and keeps the event completed in a continuous and seamless sense. Consider (45) and (46).

1. (45) Venäjä-ltä tule-va sähkö ylittä-ä raja-n Imatra-lla. Russia-abl come-ptcp electricity cross-prs.3sg border-tot Imatra-ade ‘The electricity coming from Russia crosses the border at Imatra.’

1. (46) Nykyään vesi alitta-a keskusta-n tunneli-n kautta. nowadays water pass.under-prs.3sg city.centre-tot tunnel-gen through‘Nowadays the water passes under the city centre through a tunnel.’

In (45) and (46) we have a subject that refers to a substance (electricity, water) that participates in the event incrementally: the event is being ‘fed’ by new substance all the time. The overall quantity of the substance is thus open, and this results in unbounded aspect – there is no end point to the event. In (45), for instance, new electricity keeps coming across the border all the time. Unlike (44), however, (45) and (46) do not designate iterative events that consist of distinguishable component situations. The events they designate continue seamlessly and homogeneously in time, but the total object is used nevertheless. If the partitive object were to be used, it would indicate a progressive meaning, which would make these examples awkward, since it would require the subject NP to indicate a closed overall quantity. In (45) this would result in a reading like ‘The [particular] electricity coming from Russia is crossing the border at Imatra’; cf. example (42) with the freight train. Remarkably, thus, the partitive object would require the overall situation to be bounded and then select a progressive viewpoint on it, whereas the total object allows the unbounded conceptualization of the overall situation.

In the following examples (47)–(48), on the other hand, the subject (‘traffic’) allows the iterative conceptualization of the situation from the viewpoint of individual participants in the traffic, i.e. cars driving past the accident site. Therefore, in spite of being formally a mass noun (‘traffic’), the subject in these examples can be coerced into an interpretation where it refers to a replicate mass (a series of countable objects). The consequences of such a shift in conceptualization are demonstrated by the object case marking in (47) and (48); the difference in interpretation of the durative modifier is particularly relevant.

1. (47) Liikenne kiers-i onnettomuus-paika-n kiertotie-tä pitkin yli puoli tunti-a. traffic bypass-pst.3sg accident-site-tot detour-par along over half hour-par ‘The traffic bypassed the accident site via a detour for more than half an hour.’ [the traffic arrangement existed for half an hour]

1. (48) Liikenne kiers-i onnettomuus-paikka-a kiertotie-tä pitkin yli puoli tunti-a. traffic bypass-pst.3sg accident-site-par detour-par along over half hour-par ‘The traffic bypassed the accident site via a detour for more than half an hour.’ [it took the cars half an hour to bypass the site]

In (47), which has the total object, ‘half an hour’ indicates the overall duration of the situation during which the detour was in effect. In contrast, in (48) ‘half an hour’ is the time it takes any individual car driving in the traffic to pass the accident site by the detour. This shows clearly that the partitive in (48) does not indicate the unbounded aspect of the overall situation but instead takes on a specific progressive reading that is internal to the conceivable component situations. This is also why the durative modifiers in the two examples set a temporal boundary to the event at different levels. Because the component situations are conceptualized as progressive in (48), they allow the durative modifier to indicate their temporal extension. In contrast, the total object in (47) makes the component situations themselves bounded, and therefore the durative element must relate to the overall situation by indicating its duration.

Given that the durative modifier has such different functions in the two examples, it is even possible to create an (admittedly marginal) example containing two different durative modifiers, one indicating the duration of the overall situation and the other the duration of any individual component situation. Consider (49), which means that there was a traffic arrangement in effect for the whole morning such that it took each car half an hour to pass the accident site.

1. (49) Koko aamu-n liikenne kiers-i onnettomuus-paikka-a kiertotie-tä pitkin yli puoli tunti-a. whole morning-tot traffic bypass-pst.3sg accident-site-par detour-par along over half hour-par ‘For the whole morning, the traffic bypassed the accident site via a detour [that took] half an hour [to drive].’

To sum up our observations concerning transitive examples that combine telic verbs and incremental subjects, we have seen that the partitive object can only be used to indicate progressivity of a telic event but not to indicate the general unboundedness of the overall situation that results from an open overall quantity of the subject NP. For the event to be telic, the incremental subject must be understood as indicating a closed quantity that participates in the event incrementally: the event reaches its endpoint when the whole referent of the subject has participated (as in example (41) with the freight train). Another possible interpretation is the one where the progressivity concerns the actions of individual components of the subject referent, such as the cars moving in the traffic in (48), where the focus is on the individual cars that make up the stream of traffic.

What the examples discussed in this section show, then, is that Finnish aspectual object marking not only reflects aspectual and quantificational features of the expression from the viewpoint of the object NP itself, but sometimes it reflects the aspectual effects of the subject NP as well. Against this background, it may be easier to see why the total object is used in examples like (47). At first sight the total object appears paradoxical here, since the event is clearly unbounded at the holistic level. However, considering the opposition between (47) and (48) we see that the opposition between the partitive and the total object reflects the internal aspectual nature (progressive vs. completed) of the component events that constitute the unbounded holistic situation. The focus of attention in example (48) is on the (conceivable) individual components of the subject referent, each of which accomplishes one instantiation of the situation (in its turn). This can be so even in cases where no discrete components can be distinguished, i.e. where the referent of the subject is a homogeneous substance such as electricity or water (examples (45) and (46)). Such examples in fact resemble the continuity type discussed in 3.2, only now the homogeneous substance participating in the event is referred to by the subject NP and not by the object NP. The object marking of examples like (45) and (46) can be explained by assuming that it reflects the aspectual nature of a cross-section of the homogeneous event – e.g. in (46) the overall aspectual situation is unbounded since the situation is continuously ‘fed’ with new water, but at the same time it is bounded from the point of view of any conceivable subquantity of the water that passes under the city via the tunnel.

5. A different perspective on quasi-resultative examples

We can now return to the riddle of the total object in quasi-resultative (QR) examples, discussed in section 3.1. Recall that the paradox was that the total object, which canonically indicates aspectual boundedness, is used in sentences that refer to completely static situations. The explanation I offered in 3.1 was that such cases involve a meaning component of fictive dynamicity (a fictive accomplishment) that motivates the object marking. But considered objectively, QR examples would appear to fulfil all the main criteria for the partitive object: they indicate a static situation that does not develop towards a culmination, is unbounded, and represents a minimal degree of transitivity.

The discussion in sections 3 and 4, however, suggests that QR sentences can in fact be seen as an extreme instance of the continuous accomplishment type discussed in 3.3, where the referent of the subject continuously exerts a force-dynamic effect on the referent of the object and keeps it in a state that it would otherwise leave. As we have also seen, both the subject and the object can be incremental in such examples, i.e. they can change their reference over time so that new entities (or new substance) enter the situation continuously, so as to complete and re-complete the event over and over. For instance, in the examples with an incremental subject, each (conceivable) component of the subject referent accomplishes the event in turn, making the object referent enter a condition or a state or just maintaining it there. Depending on the nature of the subject referent, such a situation can be iterative, as in The visitors push the door open when they enter the museum, or continuous, as in The water pushes the hatch open when it streams out of the container, or something between the two, as in Dark clouds conceal the moon from sight. The Finnish equivalents of these examples all take the total object (assuming that the event is resultative, i.e. the door and the hatch are kept fully open, and the moon is kept completely hidden behind the clouds); consider (50).

1. (50) Pilve-t peitt-i-vät kuu-n.cloud-pl.nom conceal-pst-3pl moon-tot ‘Clouds concealed the moon.’

Example (50) allows three interpretations that constitute a continuum from iterative to quasi-resultative. We can conceptualize the clouds, first of all, as a group of individuals, in which case the sentence indicates an iterative accomplishment: each cloud conceals the moon in turn. Second, the clouds could be seen as a seamless, substance-like mass where it is impossible to distinguish boundaries, or even as a homogeneous substance that conceals the moon in a continuous sense. Third, the example allows a static interpretation, assuming that the clouds concealing the moon are not moving; and this static interpretation brings us to a classical quasi-resultative example of the locative type (cf. examples (16) and (17)). The quasi-resultative type, accordingly, turns out to be one extreme (static) on a continuum extending from iterative via homogeneous to static.

A similar continuum can be distinguished in examples where it is the object that is incremental (and where the subject is a discrete individual maintaining its identity throughout the event). As with subjects discussed above, the continuum of objects likewise extends from discrete individuals that participate sequentially (This machine cleans the recycled bottles), via a continuous substance that ‘streams’ through the situation (e.g. The catalytic converter cleans the exhaust gas of the car), and finally to examples where the referent of the object remains the same (The brush cleans the conveyor belt). Against this background, the total object of the QR type certainly appears less paradoxical than it does if we consider it as an isolated, exceptional phenomenon.