Unanswered questions of Slavic aspect: a reply to ‘Thoughts on the ‘Typology of Slavic aspect’ ’

Propose theories which can be criticized. Think about possible decisive falsifying experiments—crucial experiments. But do not give up your theories too easily—not, at any rate, before you have critically examined your criticism.   Karl Raimund Popper, The problem of demarcation (1974).¹

1 Introduction

In Russian Linguistics 39(2), 2015, we published an article in which we review the theories of Slavic aspect as put forward by Adriaan Barentsen and Stephen Dickey, coined ‘East-west theory of Slavic aspect’ (EWT)² by us. The current issue of RL contains a response to our article by Stephen Dickey. Dickey writes:

“I cannot escape the impression that the article is ultimately a critical review of ‘Parameters’. Though ‘Parameters’, as any study, has its flaws and shortcomings, I think the overall analysis stands, especially because in the sixteen years since its publication, during which I have continued my analysis of the usage and form of aspect in Slavic languages, I have failed to find any data that contradict the overall analysis in a significant way.”

Dickey’s conclusion about the overall analysis of the EWT is in line with our general conclusion which is unequivocally positive: “The EWT of Slavic aspect is currently the only theory that gives a typology of Slavic aspect. […] We have argued that the theory is able to explain the entirety of the data in an elegant and convincing way.” (Fortuin and Kamphuis 2015, p. 203) However, regardless of our positive evaluation, Dickey concludes:

“I find Fortuin’s and Kamphuis’ review of the ‘East-west theory of Slavic aspect’ hard to understand, primarily because it omits a basic description of what the ‘East-west theory of Slavic aspect’ is, and thus inevitably downplays the role of my work in it, while simultaneously singling out ‘Parameters’ for all manner of criticism […], occasionally ignoring its discussions of data that they claim are problematic. […] Their interest does not seem to lie in big issues, but of playing Gotcha! with single examples that are not really the counterexamples they claim and/or require more analysis than they bother to provide.”

Perhaps we did not state clear enough in our article that ‘Parameters’ is the first real comprehensive overview and comparative analysis of Slavic aspect which, as such, was a major step forward in the study of Slavic aspect. However, we do not recognize ourselves in this criticism, and, as we will show, in his reply Dickey does not address the ‘big issues’ we raised in our article, such as whether the EWT approach to explain the data in terms of different invariant meanings is sufficient to explain all the data, and if so, what those meanings would look like. In this article, we will address these big issues again, and offer some new directions for future research. But, before going into this, it is perhaps interesting to provide some historical context to the birth of our article.

Our paper arose out of discussions we had as members of the Amsterdam-Leiden Slavic Aspect Circle, led by Adriaan Barentsen. The starting point of all these discussions was, and still is, the main theoretical tenet of the approach to aspect as advocated by Adriaan Barentsen which we state here again: The differences in use of aspect between the Slavic languages can be explained with reference to different meanings of aspect in these Slavic languages. While being enthusiastic adherents of this approach to Slavic aspect, we also still had several questions. We decided to pose our questions in a more systematic manner in the form of an article, the main goal of which was to further develop and strengthen the theory. It was our conviction that such a contribution was much needed, especially since there are relatively few discussions about the theoretical aspects of this approach. Given the fact that Adriaan Barentsen’s approach to Slavic aspect is very similar to that of Stephen Dickey, which is confirmed by Dickey (2000, p. 138; Dickey and Kresin 2009, p. 125), we decided to devote our article to both authors and theories. Our goal was not to play Gotcha!, and our goal was certainly not to review Dickey’s (2000) ‘Parameters’ fifteen years after it appeared, but to critically examine the ‘loose ends’ of the theory, not with the aim to falsify it, but to try to account for loose ends within the theoretical framework. Again, for the record: the majority of data is very well explained within the EWT, as Dickey (2000) already shows.

We state in our article that “[i]t should be noted that there is no such thing as a single fully explicit EWT or paradigm, as for example laid down in a single book or article. In fact, there are two authors who (largely independently from each other [sic!]) have developed very similar theories, sharing their central hypotheses.” (Fortuin and Kamphuis 2015, p. 164). In light of this remark, Dickey’s complaint that we suggest that his work was influenced by Barentsen (e.g. Dickey, p. 3, 4), as such downplaying his role, is unwarranted.³ Furthermore, we did not mean to say that Dickey’s work (or Barentsen’s for that matter) cannot be seen as a fully explicit theory, but we only wanted to point out that our way of presenting a single theory which combines ideas of both Barentsen and Dickey (coined EWT by us), is not something that these authors themselves do. In addition to that, in our discussion we did not limit ourselves to the discussion of ‘Parameters’ only (i.e. one single book), but we took no less than ten academic contributions by Dickey, including work which had not yet been published when we wrote our review, in addition to various publications by Barentsen. As such we limited ourselves to a discussion of Dickey’s and Barentsen’s work within the EWT, and did not include various other authors who have discussed comparative aspect in various other frameworks such as Richardson (2007) or Benacchio (2010) (but see Fortuin 2009, for a discussion of Richardson (2007), and Fortuin and Pluimgraaf (2015), who apply important insights from Benacchio to the EWT).

But let us now focus on some empirical data we critically analysed in our article and Dickey’s response. In our review, we chose a number of important contexts in which the eastern and western languages in Slavic differ (viz. habitual contexts, narrative contexts and retrospective contexts) and presented some data connected to them that are less easily accounted for within EWT. We did not choose these contexts randomly, but we chose them exactly because they are discussed extensively in the literature, and provide a very good insight into the way the various Slavic languages differ. In his response Dickey discusses some of the issues raised by us, framing them as ‘problematic empirical claims’. In the present reaction, we will try to keep the number of examples limited and refer to our prior article (Fortuin and Kamphuis 2015) for a more detailed discussion, but for the sake of clarity will repeat some examples from the prior article, and add some new material as well. The order in which we will discuss the various contexts, is the order in which they were presented in our previous article, which is also the order in which Dickey discusses them in his response. However, the contexts only provide a window on the use of aspect; we also would like to connect the discussion of the contexts to what we think are still unanswered questions within the EWT. These questions (one might call them ‘big issues’) are:

• What kind of situation or event can function as the reference point needed for the use of the eastern pf?

• How is it possible that ipf verbs in the eastern group, expressing isolation from a context, can sometimes still be used in narrative sequences of events?

• How can the definition of the western ipf in EWT as ‘assignability to more than one point in time’ account for uses of the western ipf for events without a clear duration or process phase?

There are some more overarching questions as well:

• How can we explain differences in aspect usage between languages that are considered to have the same aspectual meanings?

• What does it mean that even though the meanings of the aspects differ, there are still many contexts in which there appears to be no clear difference in meaning?

• How do we deal with the so-called ‘transitional languages’ within the theory?

• Is ‘sequential connection’ or ‘temporal definiteness’ a feature of the context, or of the eastern pf aspect, or of both?

• Is it possible to distinguish the concept of ‘sequentiality’ from the concept of ‘definiteness’?

2 Habitual repetition

What kind of situation or event can function as the reference point needed for the use of the eastern pf?

Dickey discusses our analysis of habitual constructions containing pf present forms. In general, the use of these forms is far more restricted in this construction in the east than it is in the west. Habitual contexts in the east only allow for pf present forms if the repeated events are linked to each other in a chain, or in the case of irregular or occasional repetition if one of the repeated events is presented as an example of the whole chain of repeated events (i.e. vivid-exemplary meaning, or singularization). Within the EWT this restriction is explained with reference to the feature of sequential connection (or temporal definiteness) of the eastern pf: only if the event can be linked to a preceding or subsequent event or reference point can the pf be used.⁴ In contrast to what Dickey says, we do not consider chains consisting exclusively of pf predicates to be the norm in contexts of repetition of sequences of events, but we only pose the question how a pf event can be sequentially linked to a subsequent ipf event. An example is provided in the following sentence, in which we find a pf present in the subordinate clause, which is sequentially connected to the ipf present tense in the main-clause:

• (1) Často, kogda on ljažet^pf spat’, emu delaetsja^ipf vdrug strašno (…).⁵

  ‘Often, when he lies down to sleep, he suddenly starts to feel terrible.’

    (Ru; Fortuin and Kamphuis 2015, p. 175; RNC: F. K. Sologub.

    Teni i svet. 1910)

This usage seems to be at odds with the definitions of the ipf in the east as proposed by both Barentsen and Dickey, which indicates that the feature of sequential connection is either absent or negated in the case of the ipf (Fortuin and Kamphuis 2015, p. 175).⁶ Let us take another look at the definitions provided by Dickey (2000, pp. 22, 24, 54, 108–109, 125):⁷

• eastern pf: ‘assignment of a situation to a single unique point in time’; or ‘the situation being locatable to a unique juncture in time relative to other situations’

• eastern ipf: (opposite of the pf): ‘the non-assignment of a situation to a single unique point in time’; or ‘the situation not being locatable to a unique juncture in time relative to other situations’.

In his reply, Dickey writes: “In my approach, the qualitative temporal indefiniteness of the Russian ipf means that there is no assertion that the situation is uniquely located in sequence with other situations, which does not mean that that situation cannot be a situation that constitutes one of the qualitatively different situations for a pf predicate.” Note that the definition Dickey gives here, differs from the one in Dickey (2000), to which we responded in our 2015 article. In fact, Dickey (2000, p. 22) calls the ipf aspect antithetical to sequentiality, which is very different from ‘not asserting’ sequentiality. It also seems to be different from the explanation in Dickey and Kresin (2009, p. 126) in terms of “the inability of a situation to be assigned to a single, unique point in time relative to other states of affairs”. Note that Dickey does not make explicit whether he regards his 2017 explanation of the definition as a departure from the original definition, but it seems to be the case that he did in fact change his definition in order to accommodate our criticism. As such, the current definition seems to be in accordance with our own stance on the meaning of the Russian imperfective, as we will explain below.

The question that we asked ourselves in 2015 was: how can a pf situation derive its unique location from and be in a sequence with a situation that is coded with an ipf predicate, without that situation being in the same sequence and uniquely located with respect to the event that is expressed by the pf verb? In other words: can A be in a sequence with B, without B being in a sequence with A? And how is such a contrasting relation possible with an event expressed by an aspectual form that signals isolation from a context, which, according to Dickey, is another way of phrasing the meaning of the ipf aspect in Russian (Dickey 2000, p. 109)? Note that Dickey himself appears to be aware of this possible weak point of the theory, since he writes in section 3.1. of his response: “One could assume that it is at odds with my formulation of the meaning of the Russian ipf as formulated in ‘Parameters’. Dickey indeed provides explanations as to why we find an ipf present tense in such chains, for example by arguing that the ipf is used to indicate overlap with a subsequent event, or because it indicates that the chain of events stops there, leaving the reading in media res (cf. Kamphuis 2016, pp. 188–190 for Old Church Slavonic). But this still does not answer our question of how the previous pf situation derives its unique position in time by linking to the subsequent ipf situation which itself, as Dickey argues, is not assignable to a single unique point in time.

One possible way out is to reinterpret the negative definition of the eastern ipf as provided by Dickey (or Barentsen; see Fortuin and Kamphuis 2015, p. 167), which may be the direction to which Dickey (2017) points in his adapted definition of the ipf aspect in Russian. The EWT definitions of the eastern aspect essentially state that whereas the pf always expresses a total and sequentially connected event, the ipf can express a total event, but never a sequential connection.⁸ We think this could be rephrased to say that whereas the pf signals to the addressee that the event should be construed as total and as sequentially connected (i.e. assignable to a unique point in time), the ipf does not signal this, but is, at the same time, not necessarily incompatible with terminative events that are fully complete, as in the case of the general factual use of the past tense as in (2), or in some cases complete terminative events that are part of a sequence of events, as in (3) that we presented in our article:

• (2) Moi sotrudniki našli magazin, gde pokupali^ipf ėti gvozdi.

  ‘My employees found a store where they bought these nails.’

    (Ru; Č. Abdullaev. Angel boli. Putešestvie po Apenninam)

• (3) Ix sxvatili^pf, doprašivali^ipf, potom rasstreljali^pf.

  ‘They arrested them, interrogated them, and then shot them.’

    (Ru; Fortuin and Kamphuis 2015, p. 183; RNC: Načal’nik razvedki.

    Soldat udači)

In our view, it is incorrect to say that the ipf expresses totality in (2) in the same way as the pf can express totality, and it is equally incorrect to say that in sentences like (3) the ipf expresses both totality and sequential connection (i.e. assignment to a unique point in time). As such we should make a distinction between totality, or sequentiality, as part of the meaning of a form (as in the case of the pf) and totality and sequentiality as part of the use or interpretation of a form in a particular context (which is possible in the case of the ipf; cf. Dickey 2000, p. 264). This means that in (2) the context allows for the ipf verb to be used in order to defocus the idea of totality, and similarly in (3) the use of the ipf verb indicates that one does not focus on the idea of totality and sequentiality, but instead on the (processual phase of) the event itself, for example to create a rupture in the narration.⁹ As such, we can stick to a negative definition of the ipf (‘the ipf does not portray the event as total and sequentially connected’), as long as we make a distinction between meaning and contextually and constructionally determined interpretation or use. This way of seeing things also makes it possible to understand why an ipf verb can serve as a reference point for the sequential connection of a pf verb. The ipf aspect does not resist a total interpretation, or even sequential connection, it simply does not express totality or sequentiality.

Dickey also refers to our treatment of other habitual contexts that allows for pf verbs, namely cases of singularization in which the event has a potential character as in sentences like Ona rešit^pf ljubuju zadaču (‘She solves / can solve any problem.’), in which the event could occur at any moment in time Bondarko (1971, p. 23). Dickey remarks: “[…] on p. 176, regarding their example 15, Fortuin and Kamphuis discuss the relevance of contingency for the potential function of the pf present, as if such usage were problematic for the theory offered in ‘Parameters’; they seem unaware that I discuss this kind of usage and its connection to temporal definiteness on pp. 85–86 of ‘Parameters’.” With respect to our specific potential example, we do not claim that this usage is problematic for the theory (in fact, in our article we discuss how it is treated within the theory by Barentsen), but we would like to understand to what kind of situations links can be made and why in certain contexts it appears to be possible to let the pf verb ‘create’ its own reference point (cf. Zel’dovič 2002, p. 31), while in others that appears to be impossible. In other words: why does a pf verb not automatically create the idea of contingency in all habitual contexts: ‘X (with a pf present verb) will happen whenever the right circumstances occur’? Or put differently, what are the contextual requirements for a situation to link sequentially to a reference point? Similar questions are posed in our section 3.3. (Fortuin and Kamphuis 2015, pp. 177–181) on habitual expressions in the past, in which we discuss exceptional cases in Russian in which a pf past occurs in the case of habitual contexts, which is not mentioned by Dickey in his review. Having said that, as Dickey rightly points out, our example 15 (p. 176) could be seen as an instance of the potential use of the pf, which Dickey (2000, p. 85–86) discusses as well. As we stated in our article, this potential-modal function of the Russian pf aspect can adequately be accounted for in the EWT by pointing at the contingency effect created by the sequential connection (whenever X arises, Y will occur; cf. Forsyth 1970, p. 174; Barentsen 1995, p. 21). However, the fact that the Czech pf present can have a similar function (see our fn. 39, p. 181), is harder to account for, since in the EWT, sequential connection is not a part of the meaning of the western pf.¹⁰ Note, however, that this observation is not in accordance with Dickey (2000). Dickey (2000, p. 86) states that “the so-called potential meaning of the pv is not a recognizable category in the western languages”, and argues that in Czech the interpretation (or rather inference) of ability can only occur in statements of facts that are characterizations, such as the following:

• (4) Otec vykouří^pf denně 30 cigaret.

  ‘Father smokes 30 cigarettes a day.’  (Cz; Dickey 2000, p. 85)

Dickey argues that in such contexts Russian requires an ipf, since the pf only occurs in sentences in which the ability is a prediction (cf. Dickey’s 2000 example 44b versus 46). It is not immediately clear to us, however, how Dickey would analyse Czech sentences like the ones below. They have a modal character, and cannot be seen as general statements on a par with the Czech example (4) given above, and instead have a prediction character like the corresponding Russian examples:

• (5) Peter Falk v roli geniálního detektiva Los Angeleské policie, který vyřeší^pf každý případ.¹¹

  ‘Peter Falk in the role of the genius detective of the Los Angeles police, that can solve any case.’   (Cz)

• (6) Jsme přímo v centru Liberce, kde najdete^pf vše pod jednou střechou!¹²

  ‘We are right in the center of Liberec, where you can find everything under one roof!’

    (Cz)

Also note that Stunová (1986, p. 484) remarks that the modal-potential use of the pf present can be found in Czech as well as in Russian, even though she links the Russian use, just as Dickey does, to the feature of sequential connection. As such, this topic relates to one of our more overarching questions: What does it mean that even though the meanings of the aspects differ, there are still many contexts in which there appears to be no difference in meaning?¹³

In our conclusion on habitual expressions (section 3.4) we start out by saying that the “EWT is able to account for the observed variation across Slavic”, but we also state that we would like to gain a better understanding of the concept of ‘reference point’. In sum: our questions regarding the nature of the reference point or contrasting situation still remain unanswered; our proposal in the current article for the rephrasing of the definition of the ipf aspect in the eastern languages only provides a partial solution to the problem.

3 Narrative sequences of events

How is it possible that ipf verbs in the eastern group, expressing isolation from a context, can sometimes still be used in narrative sequences of events?

In his discussion of narrative sequence of events Dickey starts out by quoting our article:

“Clearly, past tense sequences of events are not by nature incompatible with the ipf aspect in the Eastern languages, as is also shown by a number of counterexamples from Russian that Dickey presents.” (Fortuin and Kamphuis 2015, p. 186)

And then remarks:

“I think that this statement is false; moreover, it contradicts their own observation on p. 182, that ipf verbs in Russian “are generally not allowed in sequences of events in the past tense.” If the ipf aspect were not “by nature incompatible” (p. 186) with sequences of events in the eastern languages, ipf verbs would occur much more frequently than they do, to an extent on a par with the western languages.”

We find the reasoning Dickey presents here hard to follow. In our article, we remarked that, in Russian, ipf verbs “are generally not allowed in sequences of events in the past tense” (p. 182). In other words, in Russian, one usually finds pf verbs in narrative sequences of events. So far so good. Our quote later on, however, follows a Bulgarian example with ipf imperfects in a sequence of events (Fortuin and Kamphuis 2015, p. 186; example 38). Bulgarian, a member of the eastern group, allows for this usage in which the events are presented in a more detached way. As such, the Bulgarian data show that “past tense sequences of events are not by nature incompatible with the ipf aspect in the Eastern languages”, even though Russian does not allow for this particular use of ipf verbs.

This raises two questions. First, there is the difference between Russian and Bulgarian. This touches on our question: How can we explain differences in aspect usage between languages that have the same aspectual meanings? Should we argue that the meaning of the ipf is the same in Russian and Bulgarian (qualitative temporal indefiniteness), and attribute the different aspectual behaviour to the different tense systems? Note, however, that even if we do so, our Bulgarian example still shows that ipf verbs in the eastern group, of which the meaning is defined as ‘the non-assignment of a situation to a single unique point in time’, are not incompatible with past tense sequences of events.

Second, within the EWT, narrative sequences of events with ipf verbs in the eastern languages are a problem (as we already observed in the preceding section, in which we tried to remedy this by arguing that even though the ipf does not signal sequentiality, it is not incompatible with sequential contexts either): the ipf negates the idea of sequential connection, which means that it could be expected to be incompatible with narrative sequences of events. Or to put it in terms of Dickey’s terminology, the ipf cannot be anchored to one single and unique point in time, which is typically what happens in a narrative sequence of events. As such, the data we find both in Bulgarian and Russian (presented by Dickey himself in his review), pose a challenge to the theory.

With respect to our discussion of Bulgarian (Fortuin and Kamphuis 2015, p. 185; example 37), Dickey only focuses on one example (our example 37) with an ipf aorist in a sequence of events, and rightly points out that we should have picked another example from Lindstedt (1985; 27b instead of 26b; adding the phrase edin čas ‘for one hour’). However, that does not change our argument, since this alternative example is still an example of an eastern language allowing an ipf verb in a sequence of events. Examples like this with verbs like eat and drink are rather easy to find:

• (7) Tăkmo beše doprjal^{pf, plupf} čašata do ustata si, toj se sprja^{pf, aor} i malko smuteno kaza^{pf, aor}:

  — Vse edno! Kogato složil sveti Blan pri zajcite, mislex, če ne šte vidja pari ot nego.

  Ot dve godini ne sa go iskali. No vidite li, na svetiite ne im minava modata.

  Toj pi^{ipf, aor} i prodălži^{pf, aor}:

  Xajde, da pijnem ošte po edna.

  ‘He had just touched his mouth with the cup, when stopped, and said, a little embarrassed:

  “It doesn’t matter! When Saint Blanc served with the rabbits, I thought that I would not see any money from him. After two years they did not want him anymore. But you see, saints do not go out of fashion.”

  He drank and continued:

  “Come on, let me have one more glass.”

    (Bg; Guy de Maupassant, Un Normand)¹⁴

In this case the drinking event is subsequent to the speaker talking, and precedes his continued talking. Also compare the following Bulgarian example with ipf aorists, in which pf verbs are used in the Russian version:

• (8) I pogledna^{pf, aor}, i eto pri glavata mu pita pečena na žaravata i stomna s voda. I jade^{ipf, aor} i pi^{ipf, aor}, i pak legna^{pf, aor}.   (Bg)

  I vzgljanul^pf Ilija, i vot, u izgolov’ja ego pečenaja lepeška i kuvšin vody. On poel^pf i napilsja^pf i opjat’ zasnul^pf.  (Ru)

  ‘And he looked, and, behold, there was a cake baken on the coals, and a cruse of water at his head. And he did eat and drink, and laid him down again’

    (1 Kings 19:6)¹⁵

Dickey does not discuss our other Bulgarian example 38 with ipf imperfects, which constitutes yet another construction in which a language from the eastern group allows an ipf verb in a sequence of events, showing that ipf verbs are not by nature incompatible with that context. It appears that, for some reason, Bulgarian is more prone to accepting ipf verbs in past sequences of events than Russian. So the questions remains: why is Ru ipf more incompatible with this context than Bulgarian, if they share the same aspectual meaning?

Moreover, ipf verbs do not only occur in past sequences of events in Bulgarian, but also in Russian, even though to a lesser extent. Dickey (2017, examples 3, 4; Table 1) discusses marginal Russian data which support our analysis that ipf aspect is not by its very nature incompatible with this context. The data Dickey presents suggest that the use of ipf verbs in narratives in Russian is declining since they were more frequent in the nineteenth century than in the twentieth century. Furthermore, the data also show that the lexical meaning of the verbs also plays an important role.¹⁶

All this shows that incompatibility in this context is more a question of gradation than of absolute incompatibility.

3.1 Present tense narratives in Czech

In his discussion of narrative contexts, Dickey focuses solely on past narrative sequences and on the Bulgarian and Russian data. He does not comment on our treatment of the present tense narration (section 4.3, pp. 186–187) which contains data that require more explanation in the EWT, i.e. Czech constructions in which the ipf present tense is used to refer to single non-durative events, even though the event cannot clearly be assigned to more than one point in time, which is claimed to be the meaning of the western ipf in the EWT, as in the following example from our article (p. 187; example 43):

• (9) Půl páté. Vstávám^ipf, zvedám^ipf tašku, přehazuju^ipf si její dlouhé ucho přes rameno a vycházím^ipf na ulici.

  ‘Four thirty. I get up, take the bag, sling the long strap over my shoulder and go out onto the street.’   (Cz; Procházková; cited after Esvan 2015, p. 214)

This use of ipf verbs is a narrative style which Esvan (2015, p. 214) calls ‘tabular present’. It is typically used to present events in an isolating manner. This example poses two problems. First, the events described in (9) do not appear to refer to events with a certain duration and are therefore difficult to account for in the EWT. We will discuss this topic in detail in Sect. 4. Second, it appears that the ipf in (9) is used to deny, or defocus, the sequential connection between the events, resulting in an isolating presentation. However, (absence of) sequential connection is not a part of the meaning of the pf or the ipf aspect in the western languages in the EWT. This, again, touches upon the question: What does it mean that even though the meanings of the aspects of different Slavic languages differ, there are still many contexts in which there appears to be no difference in meaning?

4 Imperfective statements of fact

How can the definition of the western ipf in Dickey’s version of EWT as ‘assignability to more than one point in time’ account for uses of the western ipf for events without a clear duration or process phase?

Our discussion of retrospective contexts contains a critical discussion of ipf statements of fact. This discussion is important because it challenges the idea that in the west, the meaning of the ipf is ‘assignability to more than one point on in time’, as already stated in Sect. 3. It does so by discussing examples in which the idea of duration is not clearly present. In his response, Dickey does not refer to our earlier examples establishing the same theoretical problem, such as Czech constructions in which the ipf present tense is used to refer to single non-durative events (cf. (9) above), nor does he discuss our, at least to our mind, most convincing examples 61 and 62, 64, 92. We will repeat our 62 and 61 below as (10) and (12):

• (10) Kdo dával^ipf ten gól?  (Cz)

  ‘Who scored (lit. gave) that goal?’

The event in this example is hard to interpret as expressing ‘more than one point in time’; the ipf aspect is chosen, because the emphasis is not on the result of the action, but on the circumstance of the action, in this case the question who gave the goal. Cf. uses such as the following:

• (11) Na Baníku zápas skončil 3:3 a já dával^ipf na 2:2.¹⁷   (Cz)

  ‘At the Banik stadion the match ended in 3:3, and I was the one who scored to make it (2:2).’

This is a so-called actional general factual, a well-known context for the ipf aspect in Russian as well. However, in Czech there are more restrictions on the use of the ipf aspect to defocus the result than there are in Russian, as show the following example of an existential general factual:

• (12) Už jste někdy dal^pf / *dával^ipf gól?   (Cz)

  ‘Have you ever scored (lit. given) a goal?’

In this case it is not possible to use an ipf verb, even though the emphasis of the utterance does not seem to be on the result of the event, but on the mere fact of whether the event has ever occurred. In Russian an ipf verb would have been used in this construction. However, whatever the difference between (10) and (12) may be, it is not the duration of the event, be it real time duration, or presented duration.

As such, it is all the more surprising that Dickey states that “[it]t bears repeating, however, that Fortuin and Kamphuis have not given any really clear examples of such cases.” Opinions can differ as to how convincing examples are, however, our point is that stretching the definition of ‘assignability to more than one point in time’ to fit examples like the ones we give, reduces the explanatory power of the definition. In other words: when an event that at first sight has no clear process phase, is said to be presented in a durative manner because they are expressed by an ipf verb, what does durativity still mean? Dickey goes on to stress that the general factual use of the ipf is much more constrained in Czech than in Russian, illustrating this with an example with Russian naxodit’ (versus Czech nacházet ‘find’), but we came to the same or at least similar conclusion, providing an example with Russian terjat’ versus Czech ztrácet ‘lose’ (see Fortuin and Kamphuis 2015, p. 197).¹⁸

5 Concluding remarks by Dickey

In our article, we conclude, based on the various data we discuss, that the definition given for the western ipf aspect in terms of ‘assignability to more than one point in time’ cannot be maintained, and it would be better to provide another definition, namely ‘non-totality’, which Dickey finds a reasonable one, even though he remarks that “they have not argued for non-totality as the meaning of the western ipf earlier in the article (in fact they seem to have rejected it on p. 171).” This must be a misunderstanding, perhaps due to a lack of clarity on our side, since the section he refers to is a summary of the EWT theory (including his theory as presented in ‘Parameters’), which indeed seems at odds with the idea of non-totality for the west, because, as we have shown, in the west we find instances of use of the ipf with complete (i.e. total) telic events. It was exactly these kinds of examples that made Dickey argue that the western ipf expresses ‘assignability to more than one point in time’, which does not exclude the possibility of expressing totality.

Dickey further argues that we have not defined non-totality, but we have only pointed out where we think it is present,¹⁹ further remarking that “[m]y hypothesized meaning of the western ipf, assignability to more than one [conceptual] point in time, whatever its flaws, is at least positively defined.” Dickey does not, however, provide any arguments as to why a negative definition must be seen as something undesirable. In fact, note that Dickey himself negatively defines the ipf aspect for the eastern language as ‘the non-assignment of a situation to a single unique point in time’ (Dickey 2000, pp. 108–109). We think it is not problematic to provide a negative definition for the western ipf as well, as long as ‘non-totality’ is interpreted in such a way that it is compatible with total events, when the totality is for some reason defocused (cf. Sect. 7 for a more detailed discussion).

Finally, Dickey remarks that our proposal for the definition of the meaning of the western ipf relies on an extension of a prototype, in a manner similar to that assumed for the meaning of the western ipf in ‘Parameters’ (Dickey 2000, p. 278), where he suggests that:

“[…] it might ultimately prove useful to assume that the western impv, in addition to the central node of quantitative temporal indefiniteness [assignability to more than one point in time] has an associated, peripheral node of qualitative temporal indefiniteness [no assignment to a single point in time relative to qualitatively different states of affairs].”

This statement basically says that the prototypical meaning of the ipf in the west is ‘assignability to more than one point in time’, but that in some (but not yet identified) contexts this meaning will perhaps not work, and one has to propose the meaning of the eastern ipf ‘non-assignment to a unique point in time’ (i.e. the opposition or negation of temporal definiteness (sequential connection)) for the western ipf. This statement is rather counterintuitive since it would imply that in the west the ipf in some contexts negates the meaning of sequential connection, even if this meaning is not part of the (prototypical) meaning of the western pf.

In any case, by postulating ‘peripheral nodes’ the demarcation between the eastern and western pf and ipf aspect becomes less clear, and it raises the question whether two different meanings should be maintained, or whether it is better to provide an analysis in terms of the same meaning with gradual differences. Besides that, and perhaps more importantly, with a peripheral meaning (‘node’) it would be useful to have an explanation for the distribution of the peripheral uses of the ipf aspect, which is, as far as we can see, missing in Dickey’s analysis.²⁰

6 Two remaining issues

In our view, there are two important issues within EWT which are still unanswered and which need further clarification. One issue we raised in both articles, is how one should deal with instances in which the use of the eastern and the western pf or ipf appears to be similar, even though their meaning is supposed to be different (see for example our discussion about the modal use of the pf present in Russian and Czech in the present article). This can also be illustrated with the pf future tense, in which both the eastern and the western pf appear to function in the same way:

• (13) Přijdu^{pf, fut}!   (Cz)

  Pridu^{pf, fut}!   (Ru)

  ‘I will come!’

According to the EWT we would have to argue that in Czech we find the meaning of ‘totality’, whereas in Russian, we find the meaning of ‘totality’ plus a sequential link to the moment of speech. This analysis is, at least at first sight, counterintuitive since the function of the construction seems to be the same.²¹

Another issue which we briefly mentioned in our article (Fortuin and Kamphuis 2015, p. 205), is the existence of transitional languages, as well as differences within the same aspectual group. See for example Fortuin and Pluimgraaf (2015) for the observation that there are differences between Czech and Slovene in the use of aspect in the imperative; and the data on performatives including those provided by Wiemer (2014) and Dickey (2000) himself, which show that perfective performatives are probably more restricted in Czech than in Slovene, or at least occur with different lexical verbs, making Dickey’s (2000, p. 175) claim that the east-west division in Slavic is evident in the case of coincidence (i.e. performatives) perhaps too strong or at least too general. Dickey also mentions the existence of transitional languages as a remaining issue. Within the present version of EWT, these languages seem to employ different meanings of the pf and ipf depending on the type of context. This could raise doubt about whether one can actually say that the eastern pf (and ipf) on the one hand, and the western pf (and ipf) on the other, can be seen as two separate and discrete meanings.

7 Possible directions for future research

In general, we still believe that the EWT is able to explain aspectual differences between modern Slavic languages in an adequate way and we believe that the insights of Barentsen and Dickey are of paramount importance for the future development of aspect theory in Slavic. In our first article, as well as in the present one, we have focussed on data that are more challenging to explain within the EWT of aspect. Discussing these issues is important to further develop and strengthen the theory.

Our proposal for the adaptation of the theory as put forward in our prior article, concerned the definition of the ipf in the western languages. The definition ‘assignability to more than one point in time’ is difficult to maintain for those examples in which the event has no clear process phase, or in which a potentially durative event is not presented as having any duration. At the same time, Dickey convincingly shows that instances in which the ipf occurs with complete telic events are much more common in Russian than, for example, in Czech.

Our proposal is to define the ipf in the western language in a similar way as the ipf in the eastern languages, namely as a negation of the meaning of the pf aspect, hence as non-totality (cf. Stunová 1986, 1991). This may seem at odds with examples in which the ipf aspect refers to a fully complete event. When we say ‘non-totality’, however, we do not mean to say that an ipf verb cannot refer to such a fully complete event, but that the ipf verb does not signal the attainment of the inherent boundary, i.e. a change in situation. This does not mean that it negates a change in situation, though. One could phrase this as the defocussing of totality.²² The shift in focus can occur for various reasons; in the present article we have shown cases in which the meaning of totality in Czech is defocussed in favour of a focus on, for example, an isolating manner of presentation (cf. (9)), or a focus on the circumstances of the event (cf. (10)). Such usage of the ipf aspect in Czech, is rather similar to how the ipf aspect in the eastern languages is used. However, in the eastern languages this usage is much more frequent, mainly because of the stronger restrictions on the use of the pf aspect, which always needs a contrasting situation, which is apparently harder to come by in some situations than in others.

We do not have an immediate and straightforward solution for the remaining issues discussed in Sect. 6. A possible direction for further research could benefit from the following insight: even though the meaning of sequential connection does not appear to be gradable (something is either sequentially connected, or it is not), the data suggest that the requirements needed for the meaning of totality to be expressed in the various Slavic languages can be described in terms of a gliding scale, in which the East-West division only points to the most extreme positions within the Slavic aspectual continuum. This continuum does, however, suggest that the feature which sets the Russian pf apart from a language like Polish or Czech, cannot be sequential connection or temporal definiteness as part of the invariant meaning of the pf aspect by itself, and that an extra dimension of explanation is needed.

Since the idea of sequential connection is closely connected to the context requirements within which the pf can occur, and since these requirements differ per languages, it makes sense to first look at the context for this extra dimension (cf. Barentsen 2008, p. 33). As Barentsen and Dickey show, the idea of totality can be more, or less easily emphasized depending on the context. For example, if a fully realized telic event occurs in a chain of events, or if a fully realized telic event is projected into the future, it is easy to emphasize the totality of the event. There are more factors that appear to play a role in the choice of aspect (cf. Wiemer 2008; Dickey 2015, pp. 184, 193–194). Some obvious factors are: the verb form (e.g. differences between past and present, or between finite and infinite forms), the lexical meaning of the verb (certain classes of verbs allow specific uses; cf. Dickey 2000, pp. 179–181) and the grammatical construction (e.g. the difference between main and dependent clause). Hence, the extra dimension of explanation can be subdivided into many smaller factors. This raises the following two questions, with connected hypotheses:

• Question 1: Would it be possible to argue that sequential connection is not part of the Russian or the eastern pf alone, but that it plays a part in every Slavic language to varying degrees (cf. Dickey 2000, p. 278), depending on context, verb form, lexical meaning of the verb, and grammatical construction? This could explain why the pf present in the western languages can also express potentiality, and why the ipf aspect in the western languages can occur in contexts like the ones in (9), (10) and (11), as well as the fact that sequentiality does not appear to play a clear role in some contexts in the eastern languages, like in the ones that are set in the future. Hypothesis 1: sequential connection (temporal definiteness) appears as part of the meaning of the pf aspect in all Slavic languages, but the degree to which it appears depends on various factors which should be determined per language. As a follow-up one could pose the following question: is it possible to identify a common characteristic of the factors that trigger the use of either the pf or ipf aspect in an individual Slavic language? And if so, does the same common characteristic play a role in the choice of aspect in the other Slavic languages as well?

• Question 2: Would it be possible to come to an overarching Slavic hierarchy of contexts with regard to the choice of aspect? To give an example, we could argue that past tense narrative contexts, due to their combined meaning of ‘past’ and ‘narrative’ meaning, more easily trigger sequential connection than, for example, performative contexts or habitual contexts. It seems that in performative contexts the totality of the event (which would call for the use of the pf) competes with the coincidence with the moment of speech (which would call for the use of the ipf). In case of habitual contexts there is competition as well, when the individual events are total, while the overall habitual context is presented as open-ended (cf. Mønnesland 1984). The competition results in different aspect choices in the various Slavic languages. Hypothesis 2: there is a hierarchy of contexts that trigger the use of the pf (and consequently ipf) aspect. This hierarchy may be hard to establish, but a comparison of the factors between Slavic languages, as well as more insight into the development of the functions of the aspect could provide the necessary information. The ideal picture would be that the occurrence of the pf or ipf in a certain context would predict the occurrence in another context and that the contexts that most easily allow for either pf or ipf would diachronically be the oldest contexts, while those that are less typical throughout Slavic would be of relatively recent origin (cf. Dickey 2012, pp. 43–44 for the ipf general factual in Russian).

Systematic corpus-based research combined with extensive research with native speakers is necessary to see what factors trigger the pf or ipf aspect in the various Slavic languages and how the various factors that play a role in the context are responsible for the choice of aspect. It remains to be seen whether invariant meanings of aspect can explain all differences and similarities between modern Slavic languages, or that a more fine-grained analysis based on the combination of various factors is necessary. As said, we think that both more insight into the development of the aspectual systems within Slavic, as well as into the so-called transitional languages is necessary for this research. We hope that a fruitful discussion of the theory in the coming years will help clarify these issues.