1 Introduction

Dealing with actionality and pluractionality (a.k.a. ‘verbal plurality’),¹ this article takes up a rather well-known topic from Lithuanian morphology, which has also come up several times as the subject of studies on dialect morphology at least since Fraenkel (1936); see §1.2. However, we are unaware of corpus-based studies targetting the role of token frequency and the light they can shed on the distribution of verbal suffixes and their functions in unconstrained discourse. We may therefore consider our study the first token-based investigation, at least of verbal morphology, in Lithuanian dialectology. Before presenting it, we want to clarify the main conceptual premises and give a brief survey of extant research.

Pluractionality “is a phenomenon that marks the plurality or multiplicity of the situations (i.e. states and events) encoded by the verb through any morphological mean that modifies the form of the verb itself” (Mattiola 2019, 4). It should be distinguished from event plurality, which refers to “any linguistic means of expressing a multiplicity of events”, thus comprising also adverbials and other means outside the verb (Cabredo Hofherr & Laca 2012, 1). Pluractionality encompasses functions which are related either to the repetition of situations as a whole or to repetitions of sub-events within one global event (or occasion; for such distinctions cf. Cusic 1981, 61). The distinction hinges, of course, on the overarching time-frame (or interval) in relation to which distinct events can be “counted”. We do not intend to go into any more detailed theoretical discussion (or ways of formalisation), but on a more intuitive account the distinction can be illustrated as linguistically relevant with examples such as the following ones from Russian:

(1)

(2)

In these examples the verb forms prygaet ‘jumps’ and kašljal ‘coughed’ denote situations that are internally fragmented into sub-intervals; this fragmentability is part of the meaning of these verbs, which are therefore often called ‘multiplicatives’ (Xrakovskij 1987; 1997), and they have many features in common with progressive aspect. Concomitantly, in (1 – 2) the respective situation as a whole is marked by an adverbial either indicating habituality (obyčno ‘usually’) or a rate at which the repetition occurs (každye dvadcat’ minut ‘every 20 minutes’). These temporal adverbials scope over the internally fragmented situations denoted by the verbs; in (1 – 2) the narrower (= internal) and wider (= external) scope is indicated by subscripted indexes. This scope relation corresponds to the semantic “insertion” of internal sub-intervals into larger external intervals and the different finegrainedness, which cannot be reversed. ² In this sense, the external interval sets a frame for the count of repetitions of sub-intervals inherent to the semantics of multiplicative verbs: in each single interval “counted” by the external time adverbial a certain “amount” of subevents of the process denoted by the multiplicative verbs takes place. This also shows that external and internal event plurality operate independently from one another and, after all, that they represent two different types of repetition.

In order to not mix them up, an indication of internal subinterval properties will be called ‘repetitive’ (following Cusic 1981, 80ff.), whereas the count of the external interval will be named ‘frequentative’; ‘iterative’ will be employed as an unspecific umbrella term. These rather simple terminological decisions are meant to make maximally transparent the cut between event-internal and event-external pluractional meanings and to remain as close as possible to those terms which seem to have become most commonplace in the typological literature.³ Moreover, on this account, habituals and unrestricted (regular or irregular) repetition of global events belong to event-external plurality. Habituals can acquire meanings of dispositional or circumstantial modality (compare John repairs any car . John is able to repair any car), although this need not be the case (e.g., Usually / Every Monday, John repairs cars does not trigger a modal reading). On the other hand, multiplicative verbs, or verb forms, often yield progressive meanings (which refer to processes consisting of multiple phases without a natural segmentation), or the differences between repetitive and progressive interpretation are blurred; this happens if no identical subevents can be figured out or this is communicatively irrelevant (compare The girls are giggling or They were pouring water into the pool). We will indicate such problems in our analysis (§2.1.3).

Now, while verbs like Russ. prygat’, Lith. šokti ‘jump’ and Russ. kašljat’, Lith. kosėti ‘cough’ can readily be used as repetitive without any special morphology (just as their English equivalents), there are morphemes able to mark a verb as repetitive, i.e. as indicating event-internal plurality. One such morpheme is the Lithuanian suffix -inė-, which is productive both with unprefixed and prefixed verbs (e.g., šokti => šok-inė-ti ‘jump’, prašyti => praš-inė-ti ‘ask, request’; at-sakyti => atsak-inė-ti ‘answer’). Others, like Lith. -dav-, serve to mark event-external plurality; this meaning is syncretic with past tense (e.g., ger-ti ‘drink’ => ger-dav-o ‘s/he / they used to drink’).

In the remainder of this section we will survey the state of the art on these two suffixes, both for the standard language and in the dialects (§1.1–1.2), and we will give a brief account of the corpus (§1.3) which serves as the basis for our analysis presented in §2. A summary will be given in §3.

1.1 The frequentative suffix {dav}

This suffix marks the past habitual. Even though traditionally verbal forms with -dav- are treated as a separate tense (Lith. būtasis dažninis laikas ‘past iterative tense’), the opposition between simple past and -dav- forms has also been viewed as an aspectual one (Holvoet/ Pajėdienė 2004, 124). For comprehensive functional accounts cf. Arkadiev (2012, 78–84) and Sakurai (2015), from where these Standard Lithuanian examples are cited:

(3)

(4)

The habitual-past suffix is a rather recent innovation restricted to the Aukštaitian area and some adjacent Samogitian dialects (Stang 1942, 172f., Toporov 1961, 55, Zinkevičius 1966, 356). Historically it comes from an iterative suffix -dau- (Pakerys 2017). Consequently, dav-forms are attested only for some Lithuanian dialects, namely Aukštaitian and nearby Samogitian. It has been noted that South Aukštaitian dialects rarely use these forms or do not use them at all (Zinkevičius 1966, 356, Vidugiris 2004, 233).

1.2 The repetitive suffix {inė}

The suffix {dinė}⁴ differs from {dav} in many respects. First, it is used in the derivation of stems with values usually associated with imperfective aspect, but it is only marginally employed to denote habitual actions. It rather marks repetitive meanings; compare examples from the Corpus of Contemporary Lithuanian (CCL: http://tekstynas.vdu.lt/):

(5)

(6)

(7)

The stem įkalbinėti is derived from įkalbėti ‘persuade, talk into’ and primarily denotes goal-directed speech activity by which some addressee is to be persuaded to do (or to think) something. The present (ex. 5) and past tense (ex. 6) forms of įkalbinėti are compatible with repeated situations “counted” as a whole. However, the external interval for iterated actions has to be marked by some suitable adverbial, like vis ‘always, all the time’ in (5 – 6). Otherwise įkalbinėti just denotes intense goal-directed action within one larger interval. Thus, for instance, įkalbinėjo ją parašyti apie savo keliones (see ex. 6) would just mean ‘s/he / people were trying to persuade her to write about her journeys’, and this is compatible with repetition over a couple of larger intervals, but does not denote it by itself. In (7)įkalbinėti is in the subjunctive and coordinated with the atelic activity verb ginčytis ‘quarrel, argue’.

Therefore, while {dav} belongs to the domain of event-external plurality, {inė} should be considered a marker of event-internal plurality, or even of actionality inasmuch as the borderline with the progressive meaning is fuzzy. As such, {inė} can non-redundantly combine with {dav}, but only in the order {inė}+{dav}; cf.:

(8)

Second, {inė} is not restricted to the past domain, but freely combines with any tense and mood (see ex. 5 – 7). For these two reasons, {inė} has been compared to the Slavic suffix {iva} used to derive imperfective stems from prefixed perfective stems (so-called secondary imperfectivization). However, suffixation with {inė} is quite common also for simplex stems (Kardelis/Wiemer 2003, 65f.; see also §2.1.3), a phenomenon which in the neighboring Slavic languages occurs only rarely.

Third, the origin of {inė} and its spread within Aukštaitian markedly differ from the provenance and areal spread of {dav}. Suffixation with {inė} was apparently much more frequent in Lithuanian varieties surrounded by East Slavic (as in extinct or moribund insular dialects) or in closer vicinity to East Slavic. Thus, since Fraenkel (1936), attention has been drawn to an increased productivity of {inė} in the borderland with Belarus and in (now mostly extinct) insular Lithuanian dialects in Belarus (Vidugiris 1961; 1998; 2014, 224, Wiemer 2009, 361f.). Fraenkel (1936, 104) spoke of “exact imitations of Slavic differences of aspect and modes of action“⁵ in the border region with Belarus, while Rozwadowski (1995, 130) noted that in the dialect of Zietela (Belarus) verbs in -(d)inė- were used without any particular actional values.⁶ Anyway, it is not too difficult to show that even under intense contact conditions the relevant Lithuanian varieties do not use these resources in the derivation of verb stems with tightly-knit functions of grammatical aspect to the same extent as do Slavic languages (Kardelis/Wiemer 2002; 2003, 64, Pakerys/Wiemer 2007). In this respect, dialects do not very much differ from the standard language (for which cf. Wiemer 2001, Arkadiev 2012, 58 – 60, Arkadiev et al. 2015, 31 – 35). Nonetheless, there remains the (much more interesting) question of whether contact with (East) Slavic enhances the propensity for using verbal suffixes to derive stems with changed actionality or functions of pluractionality – both on type and on token level. Only the type level has been studied to some extent (cf. Kardelis/Wiemer 2003, 59 –66; Wiemer 2009, 361 – 363, 367 – 385), while the token level has not been studied at all. Our case study provides the first attempt at doing so.

1.3 The TriMCo-corpus

Our research is based on the Lithuanian part of the TriMCo-corpus, which was created in connection with the TriMCo-project (www.trimco.uni-mainz.de).⁷ The corpus consists of dialectal texts transcribed in ElAN (https://tla.mpi.nl/tools/tla-tools/elan/). The largest part of the texts belongs to mixed Belarusian varieties from the Slavic-Baltic contact zone in Belarus, Lithuania and Latvia (Latgalia). Smaller portions come from Russian dialects in the Pskov region and from Latgalian. The Lithuanian part, although considerably smaller than the Belarusian one, is currently much better prepared for token- based research, as it has been annotated morphologically at a sufficiently reliable level by one of the authors of this article. It is planned to make the corpus publicly accessible; it forms part of the network SpoSla .Corpus-based Research into Sociolinguistic and Dialectal Variation in Slavic Languages); see http://parasolcorpus.org/Spoken-Slavic/. For more details see Table 1 and Wiemer et al. (2019). ⁸

As can be inferred from Table 1, the three Lithuanian subcorpora belong to different regions of the Aukštaitian area and are of rather unequal size. The East Aukštaitian and South Aukštaitian parts are of comparable size only if we unite the two smaller parts, which, though belonging to the South Aukštaitian dialect area, are separated by the state border between Lithuania and Belarus.

Map 1 shows the Baltic dialect continuum, Map 2 zooms into Map 1 and illustrates how the spots where the texts of the corpus were recorded relate to this continuum.

Map 1
Lithuanian dialects9

Map 2
Places of recordings used in the Lithuanian part of the TriMCo Corpus (East Aukštaitian vs. South Aukštaitian)10

2 Token-based investigation of the suffixes

Here we offer our analysis of the two suffixes that mark pluractionality in Lithuanian dialects — -(d)inė- and -dav-. There are some other suffixes that can add a pluractional meaning to a verb, namely -dy-, as in atei-ti ‘come’ : atei-dy-ti ‘come repeatedly’, and -o-, as in neš-ti ‘carry’ : neši-o-ti ‘carry repeatedly; wear’. But since these are less productive, we did not take them into account.

2.1 -inė- vs -dinė-

There has been some inconsistency in how to treat the segments -inė- and -dinė- — as two distinct suffixes or as allomorphs of one and the same suffix. In Standard Lithuanian, the solution is rather straightforward, as the distribution is practically complementary: -inė- appears if the stem ends in a consonant, whereas -dinė- occurs with stems ending in a vowel; compare, for instance, šok|ti ‘jump’ : šok-inė-ti ‘jump repeatedly’ and jo|ti ‘ride’ : jo-dinė-ti ‘ride repeatedly’. They thus should be regarded as allomorphs of the same suffix. The data from Lithuanian dialects are less straightforward: for instance, in the South Aukštaitian part of the TriMCo corpus, both -inė- and -dinė- can be attached to the same root; compare mušinėti (1 token) : mušdinėti (3 tokens) ‘beat repeatedly’, pirkinėti (1 token) : pirkdinėti (4 tokens) ‘buy repeatedly’ etc.¹¹ We therefore start by discussing the differences and similarities between the two variants in the TriMCo corpus.

2.1.1 Areal distribution

Let us first check if these two variants are evenly distributed across the dialect areas represented in the TriMCo corpus (East Aukštaitian vs. South Aukštaitian divided into two parts, i.e. on either side of the Lithuanian–Belarusian border). Table 2 shows that there is a statistically significant dependency between the variant of the suffix and its areal distribution. The variant -dinė- is very rare in East Aukštaitian, whereas in either part of South Aukštaitian its frequency is considerably higher. Although still -inė- clearly outnumbers -dinė-, the distance is much closer.

Importantly, the most frequent lexeme with the suffix -inė- is važinėti ‘drive’. It makes up 49% of all repetitives with the suffix -inė-; see Table 3.

We may thus assume that the higher percentage of the verb važinėti among all repetitives with the suffix -inė- in South Aukštaitian is the result of an extensive spread of the variant -dinė-, i.e. verbs ending with a consonant that would take -inė- in Standard Lithuanian or East Aukštaitian, combine with -dinė- in South Aukštaitian, but važinėti remains as a lexicalized variant.¹² However, we see that in East Aukštaitian the variant -inė- shows a less skewed distribution across lexemes than in the other zones. Let us exclude the biasing lexeme važinėti from the statistics and look at the resulting figures:

After the biasing lexeme is excluded, we can ask whether -dinė- is more productive in the Southern zone. Support for this is provided by the number of different lexemes attested with each variant in the different dialect zones:

The lexical diversity of -inė- is significantly larger in the East than in the other regions, while the lexical diversity of -dinė- shows an opposite distribution. All these observations indicate that the two variants -inė- and -dinė- are possibly not allomorphs; instead they clearly tend toward complementary areal distribution.

2.1.2 Derivational properties

As already stated, the dialectal distribution of the variants -inė- and -dinė- is different from the standard language. However, it is only the variant -dinė- that broadens its derivational properties: -inė- always follows a stem ending with a consonant, while -dinė- can follow both consonants and vowels; cf. Table 5.

Another difference between the two variants suggested in the literature is that -inė- derives stems based on the past tense, while -dinė- attaches to the infinitival stem (Vidugiris 1961, 220). According to the TriMCo data, this claim holds true for -dinė-, whereas -inė- can derive verbs from various stems (not only from the past tense stem); compare the present tense stem išein-inė-dam-i ‘leaving’ (ei-ti.inf ~ ein-a.prs-3 ~ ėj-o.pst-3) (East Aukštaitian, Kalv), the widespread verb važ-inė-ti ‘drive’ derived from važiuoti (važiuoj-a.prs-3 ~ važiav-o.pst-3) (here, however, it is not clear from which stem), or dav-inė-ti ‘give’ which is clearly derived from the past stem (duo-ti.inf ~ duod-a.prs-3 ~ dav-ė.pst-3).

It has also been claimed that in South-Eastern Lithuanian dialects verbs with the suffix -(d)inė- are usually prefixed (Vidugiris 1998, 185). Furthermore, at least in the Zietela dialect, the variant -dinė- is more often attached to prefixed verbs than -inė- (Vidugiris 1961, 220).

Our corpus data show that both variants combine with prefixed verbs quite often, but the difference between them is not so straightforward. At first sight, the data from the TriMCo corpus covering both East and South Aukštaitian dialects show that -dinė- is attached to prefixed verbs more often than -inė-; see Table 6.

However, if we exclude the non-prefixed stem važinėti, the difference between the two suffixes is not significant anymore.

The preference for prefixed verbs that both variants demonstrate might have to do with the actional classes of the stems they attach to. We followed the classification of Lithuanian actional verb classes offered in (Arkadiev 2009; 2011; 2012), which itself is based on Tatevosov (2002; 2005); cf. also Tatevosov (2016). Thus, we coded the verbal stems to which -inė- and -dinė- attach as strong telic (as in (9), where the verb is derived from parduoti ‘sell’), weak telic (as in (10), where the verb is derived from šokti ‘jump’), and processual¹⁴ (as in (11), where the verb is derived from važiuoti ‘drive’). The difference between the telic classes is that weak telic in the simple past forms can have either a process or an event interpretation, while strong telic verbs yield only an event interpretation.

(9)
¹⁵

(10)

(11)

Again, at first glance, our data show that in most cases the suffix -inė-/-dinė- is attached to telic (strong or weak; i.e. those that can or must have an event interpretation in the simple past) verbs (72%, i.e. 128 out of 179 tokens). If the lexeme vazinėti (derived from the processual verb važiuoti) is excluded, the percentage of derivations from telic verbs grows even bigger (98%, i.e. 109 out of 111 tokens).

Now, if we compare the two variants with regard to the telicity of the verbs they are attached to, the results are similar to the test on compatibility with (non)prefixed verbs: the difference is significant only if the biasing lexeme važinėti is included in the count; see Tables 7 and 7a.

That is, the suffix -(d)inė- is rarely attached to processual verbs (if we exclude the lexeme važinėti, there are only two examples). We may also infer that the main function of the suffix is to make telic stems available for marking ongoing processes.

It is worth noting that certain prefixed verbs with the suffix -(d)inė- are not derived with the suffix from prefixed verbs but are rather prefixed derivatives of repetitive verbs. That is, the prefix is the last member in a derivational chain which telicizes the already suffixed stem. This is often the case especially for the prefixed derivatives of važinėti; compare važi-uo-ti ‘drive’ (processual) > važ-inė-ti ‘drive repeatedly; in different directions’ (processual) > pra-važinėti ‘spend by driving’ (telic):

(12)

This order of derivational steps can be assumed since the prefix does not introduce a border (goal, source, or whatever) of some directed movement, but only puts a temporal limit to the atelic action or, as in (12), marks the achievement of some non-spatial boundary which follows from this action.

2.1.3 Semantics

In accordance with the meanings distinguished in Geniušienė (1989), but adapted to the terminology introduced in §1, we coded the semantics of the verb forms suffixed with -inė-/-dinė- as follows:

(i) frequentative meaning: Prototypical multiple events presuppose the same participants of the situation; see (13). However, sometimes different multiple subjects or multiple objects can be involved; see (14). Finally, the repetition of events often yields the meaning of ability; see (15).

(13)

(14)

(15)

(ii) repetitive meaning; see (16).

(16)

(iii) processual meaning; see (17).

(17)

Not all examples are straightforward. See (16), where one could also argue for event- external pluractional meaning (‘several asking events occurring in one larger interval’).¹⁶ Or compare (18), in which even the broader context does not allow for a specific interpretation, i.e. both repetitive (multiple actions in a specific situation) and habitual (‘when someone marries an alcoholic, they have to…’) are possible:

(18)

Such cases were coded separately (as ‘NA’) and excluded from the further statistics.

As regards the semantics of the verb forms, the comparison of the two variants did not yield any significant difference; see Table 8.

2.1.4 Interim conclusions

Even though the variants -inė- and -dinė- differ in their areal distribution (the variant -dinė- is much more frequent in South Aukštaitian) and in some derivational properties (-dinė- is always attached to the infinitival stem, whereas -inė- shows some variation), there is no ground to view them as two separate suffixes. The apparent difference between the two variants in their compatibility with various types of verb stems (with respect to their actionality class or derivational structure: prefixed vs. non-prefixed) turns out to be illusory once the biasing lexeme važinėti ‘drive’ is excluded. Therefore, we can regard -inė- and -dinė- as phonetically distributed allomorphs in Standard Lithuanian and East Aukštaitian and as free variants of the same suffix in South Aukštaitian.

2.2 Areal distribution of -(d)inė- in comparison with -dav-

The suffixes -dav- and -(d)inė- are both employed to mark pluractionality. However, their distribution within Lithuanian dialects is not even.

Forms with the suffix -dav- are attested only for some Lithuanian dialects, namely Aukštaitian and neighboring West Samogitian dialects. It has been noted that South Aukštaitian dialects rarely use these forms or do not use them at all (see §1.1). The data from the TriMCo corpus support this claim and clearly show that dav-forms appear in East Aukštaitian much more often than in South Aukštaitian. In the South Aukštaitian data from Belarus such forms turned out to be almost absent.

It has also been noted that the suffix -(d)inė- is, in turn, more frequent in the South Aukštaitian dialects, especially in the dialects located in Belarus (e.g., Vidugiris 1998: 184). The data from the TriMCo corpus support this claim as well.

Table 10 shows that -(d)inė- forms occur in the South Aukštaitian dialects of Belarus more than twice as often than in the East Aukštaitian dialects (the odds ratio is 2.27).

As has been noted for the standard language, -dav- and -(d)inė- can be combined; compare, for instance, pardav-inė-dav-o » ‘used to be occupied with selling’, atkalb- inė-dav-o ‘used to advise against’, perim-inė-dav-o ‘used to take over’ (see §1.2). In our corpus, only seven such cases were observed, but all of them are from the East Aukštaitian part.

(19)

At first sight, the number of distinct lexemes with which this suffix combines (= productivity) does not vary much over the regions (we would expect a higher number in the South Aukštaitian dialects):

Comment (also for Table 11a): Figures in brackets indicate the amount minus those stems with more than one derivative; for instance:

(20)

However, if we put the South Aukštaitian dialects in Lithuanian and Belarus together (which yields subcorpora of a more comparable size), the results are different: in South Aukštaitian this suffix combines with 35 distinct lexemes (or 29 if we account for identical stems).

It has also been claimed that dav-forms in South Aukštaitian were driven out by inė- forms (Zinkevičius 1966, 356, Vidugiris 1998, 184, Roszko & Roszko 2006, 165). If this is true, we expect to have a larger number of simple past forms in the south than in the east. Thus, if we look at the percentage of simple past forms of verbs with the (d)inė-suffix in the TriMCo corpus, we see that, on the one hand, the South Aukštaitian subcorpora both show a higher percentage of simple past tokens for -(d)inė- than does the East Aukštaitian subcorpus; on the other hand, we also observe a somewhat higher percentage of simple past forms with the suffix -(d)inė- in South Aukštaitian in Lithuania, but not in Belarus, which is not what we would expect.

Moreover, Table 13 shows that, even though the percentage of simple past (as opposed to other verb forms) is slightly higher in South Aukštaitian, the difference is not statistically significant. Thus, corpus data do not really confirm the assumption that -(d)inė- has been expanding at the expense of -dav-.¹⁹

The aforementioned hypothesis also entails that in the past tense verbs with the suffix -(d) nė- in South Aukštaitian express habitual meanings more often than in East Aukštaitian. Again, the difference between dialect areas turns out to be illusory; see Table 14. On the one hand, if we look only at the simple past forms, we see that the dependency between the two factors (event type and region) is statistically significant (which supports the hypothesis).

On the other hand, if we compare present tense forms with the suffix -(d)inė- in different dialect areas, the data also shows a dependency between the meaning of the verb form and the dialect region; see Table 15.

Thus, even though the habitual meanings of (d)inė-forms are more common in South Aukštaitian, it is not the result of their spread (only) into the past tense. If anything, higher frequency of -(d)inė- at the expense of -dav- is a by-product of a global spread of -(d)inė- irrespective of tense (or other inflectional characteristics). This might hint at a higher similarity between South Aukštaitian (d)inė-forms and Slavic secondary imperfectives. In Standard Lithuanian, telic verbs can be used in the present tense to describe habitual events (Dumašiūtė 1961, Sawicki 2000, Arkadiev 2012, 56). We may assume that in South Aukštaitian in such contexts the -(d)inė-forms are used. However, at this stage this is just a tentative suggestion that should be tested on larger corpora.

3 Conclusions

Before drawing more general conclusions, let us summarize our findings on -(.)inė- and -dav- arrived at on the basis of the TriMCo corpus.

1. 1. As for the studied dialect areas, the variants -dinė- and -inė- can be considered allomorphs only in East Aukštaitian, contrary to South Aukštaitian, where they do not occur in complementary distribution regarding structural properties of the stems they attach to. That is, the two variants do not really show complementary areal distribution, but only a slight tendency.
2. 2. The variant -dinė- is more frequent in South Aukštaitian than in East Aukštaitian. This regards not only token frequency, but also lexical productivity, i.e. the degree to which the given morpheme is attested with different stems. However, since the absolute numbers of tokens found in the corpus are not very high, we have to be particularly careful with conclusions concerning productivity.
3. 3. Verb forms with -(d)inė- are more frequent in South Aukštaitian than in East Aukštaitian, however the lexical productivity of the iterative suffix is roughly the same across the dialect zones.
4. 4. The percentage of the past habitual (-dav-) among the past tense forms is different across the dialects. Past habitual is almost absent in the South Aukštaitian dialects of Belarus.
5. 5. Even though verb forms with the suffix -(d)inė- are more frequent in South Aukštaitian than in East Aukštaitian, this does not necessarily imply that these forms encroached on the domain of the past habitual forms. It seems that habitual meanings are primarily expressed by the simple past in South Aukštaitian. A slightly higher token frequency of -(d)inė- in South Aukštaitian seems to apply irrespective of any particular tense.

Our more general conclusions are as follows. As we stressed in our overview (§§1.1-1.2), dialectological literature has at best been restricted to type-based analyses; there have been no token-based analyses at all. We argue that a quantitative, token-based approach, for which annotated corpora are indispensible, yields much more revealing insights into the variation and spread of morphosyntactic phenomena, such as those investigated in this article.

1. 1. Only token-based analyses can substantiate claims concerning the spread of a construction. They allow us to disclose biases within some larger area (or between two areas), which remain unaccounted for if only type-based data are considered. The latter ones lie at the basis of atlases and traditional isoglosses, both single ones and bundles.
2. 2. By using inferential statistics, we can make claims about the significance of biases (i.e. skewed distributions). This is important to emphasize, since it allows us to separate impressionistic observations from more systematic differences claimed to exist between varieties and/or between forms and their functions.
3. 3. Corpus data are also indispensible, although not by themselves sufficient, for an assessment of productivity (i.e. flexibility of lexical input); cf. Baayen (2009) for an overview. Good gauges and methods of productivity are hard to come by; a central problem is the non-comparability of corpora of different sizes and of the variances of types (H. Baayen, p.c.). In any case, however, the road toward suitable indicators of productivity leads via type/token ratios for slots in specific constructions (or morphemes of word forms), and to establish these, corpora are indispensable.
4. 4. Token-based analyses allow for regression methods (sometimes even if data is sparse), i.e. methods that weight factors against each other and allow us to filter out the most relevant ones (which, again, cannot be discerned by the naked eye).

Acknowledgments

We owe words of gratitude to Peter Arkadiev, who participated in the analysis of corpus data at an early stage. We also appreciate the valuable input of an anonymous reviewer and some minor remarks by another reviewer. We would like to thank Wayles Browne for proofreading our English. Regardless of this, the sole responsibility for the analysis and conclusions is ours.

References

Arkadiev, Peter. [Arkad’ev, Petr M.] 2009. Teorija akcional’nosti i litovskij glagol [A theory of actionality and the Lithuanian verb]. Balto-slavjanskie issledovanija 18 [Baltic-Slavic investigations 18], 72–94.

Arkadiev, Peter. 2011. Aspect and actionality in Lithuanian on a typological background. In langues baltiques, langues slaves. Daniel Petit, Claire Le Feuvre & Henri Menantaud, éd. Paris: CNRS Editions. 61–92.

Arkadiev, Peter. [Arkad’ev, Petr M.] 2012. Aspektual’naja sistema litovskogo jazyka (s privlečeniem areal’nyx dannyx) [The aspectual system of Lithuanian (with an account of areal data)]. In Issledovanija po teorii grammatiki. Vyp. 6: Tipologija aspektual’nyx sistem i kategorij [Investigations in the theory of grammar. Issue 6: The typology of aspect systems and categories]. (= Acta Linguistica Petropolitana, t. VIII, č. 2). Vladimir A. Plungjan, ed. Sankt-Peterburg: Nauka. 45–121.

Arkadiev, Peter, Axel Holvoet & Björn Wiemer. 2015. Introduction: Baltic linguistics – State of the art. In Contemporary Approaches to Baltic Linguistics. Peter Arkadiev, Axel Holvoet & Björn Wiemer, eds. Berlin, Boston: De Gruyter Mouton. 1–109.

Bertinetto, Pier Marco & Alessandro Lenci. 2012. Habituality, Pluractionality, and Imperfectivity. In The Oxford Handbook of Tense and Aspect. Robert Binnick, ed. New York: Oxford University Press. 852–880. https://doi.org/10.1093/ oxfordhb/9780195381979.013.0030.

Cabredo Hofherr, Patricia & Brenda Laca. 2012. Introduction – event plurality, verbal plurality and distributivity. In Verbal Plurality and Distributivity. Patricia Cabredo Hofherr & Brenda Laca, eds. Berlin, Boston: De Gruyter Mouton. 1–24. https:// doi.org/10.1515/9783110293500.

Corbett, Greville G. 2000. Number. Cambridge etc.: Cambridge U.P.

Cusic, David Dowell. 1981. Verbal Plurality and Aspect. Stanford, CA (unpubl. PhD thesis).

Dumašiūtė, Zelma. 1961. Dabartinės literatūrinės lietuvių kalbos veiksmažodžio vienti- sinio esamojo laiko reikšmė ir vartojimas [The meaning and use of the simple present tense in contemporary Standard Lithuanian]. Vilnius (unpubl. doctoral dissertation).

Fraenkel, Ernst. 1936. Der Stand der Erforschung des im Wilnagebiete gesprochenen Litauischen. Balticoslavica 2, 14–107.

Geniušienė, Emma Š. 1989. Mul’tiplikativ i iterativ v litovskom jazyke [The multiplicative and the iterative in Lithuanian]. In Tipologija iterativnyx konstrukcij [Typology of iterative contructions]. Viktor S. Xrakovskij, ed. Leningrad: Nauka. 122–132.

Holvoet, Axel & Jūratė Pajėdienė. 2004. Laiko kategorija ir laiko formos [The category of tense and tense forms]. In Gramatinių kategorijų tyrimai [Investigations on grammatical categories]. Axel Holvoet & Loreta Semėnienė, eds. Vilnius: LKI. 121–140.

Zinkevičius, Zigmas. 1966. Lietuvių dialektologija (Lyginamoji tarmių fonetika ir morfologija) [Lithuanian dialectology (A comparative phonetics and morphology of dialects)]. Vilnius: Mintis.

Kardelis, Vytautas. 2009. Šiaurės rytų aukštaičiai vilniškiai: ribos ir diferenciacija [Northeast Aukštaitian dialects: borders and differentiation]. Vilnius: Vilniaus universiteto leidykla.

Kardelis, Vytautas & Björn Wiemer. 2002. Ausbildung von Aspektpaarigkeit in litauischen Grenz- und Inseldialekten (am Beispiel von Sprechverben). Linguistica Baltica 10, 51–80.

Kardelis, Vytautas & Björn Wiemer. 2003. Kritische Bemerkungen zur Praxis der Erstellung litauischer Wörterbücher, insbesondere von Mundarten – am Beispiel des slavischen Lehnguts und des ‘veikslas’. In Prace bałtystyczne. Język, literatura, kultura. Norbert Ostrowski & Ona Vaičiulytė-Romančuk, eds. Warszawa: Wydział Polonistyki UW. 45–72.

Mattiola, Simone. 2019. Typology of Pluractional Constructions in the Languages of the World. Amsterdam, Philadelphia: Benjamins.

Pakerys, Jurgis. 2017. On the development of past habitual from iterative in Lithuanian Baltistica LII-2, 295–323. https://doi.org/10.15388/baltistica.52.2.2324.

Pakerys, Jurgis & Björn Wiemer. 2007. Building a partial aspect system in East Aukštaitian Vilnius dialects of Lithuanian (Correlations between telic and activity verbs). Acta Linguistica Lithuanica 57, 45–97.

Roszko & Roszko 2006. Lithuanian frequentativum. Études cognitives 7, 163–172.

Rozwadowski, Jan. 1995. Litewska gwara okolic Zdzięcioła na Nowogrodczyźnie (Dzieło pośmiertne). Opracowanie: Adam Gregorski. Konsultacja: Jan Safarewicz) [The Lithuanian dialect of Zdzięcioł (Lith. Zietela) in the Nowogródek region (A posthumous work). Preparation: Adam Gregorski. Consultation: Jan Safarewicz]. Cracow: PAN.

Sakurai, Eiko. 2015. Past habitual tense in Lithuanian. In Contemporary Approaches to Baltic Linguistics. Peter Arkadiev, Axel Holvoet & Björn Wiemer, eds. Berlin, Boston: De Gruyter Mouton. 383–436.

Sawicki, Lea. 2000. Remarks on the category of aspect in Lithuanian. Linguistica Baltica 8, 133–142.

Stang, Christian S. 1942. Das slavische und baltische Verbum. Oslo.

Šluinskij, Andrej B. 2006. K tipologii predikatnoj množestvennosti: organizacija semantičeskoj zony [On the typology of verbal plurality: the organisation of a semantic domain]. Voprosy jazykoznanija 2006-1, 46–75.

Tatevosov, Sergej G. 2002. The parameter of actionality. Linguistic typology, 6(3), 317– 401.

Tatevosov, Sergej G. 2005. Akcional’nost’: tipologija i teorija [Actionality: typology and theory]. Voprosy jazykoznanija 1, 108–141.

Tatevosov, Sergej G. 2016. Glagol’nye klassy i tipologija akcional’nosti [Verb classes and the typology of actionality]. Moscow: JaSK.

Baayen, Harald. 2009. Corpus linguistics in morphology: Morphological productivity. In Corpus Linguistics. An International Handbook, vol. 2. Anke Lüdeling & Merja Kytö, eds. Berlin, New York: Mouton de Gruyter. 899–919. https://doi.org/10.15 15/9783110213881.2.899.

Toporov, Vladimir N. 1961. K voprosu ob ėvoljucii slavjanskogo i baltijskogo glagola [On the issue of the evolution of the Slavic and the Baltic verb]. In Voprosy slavjanskogo jazykoznanija 5. Vladimir N. Toporov, ed. Moskva: Izd-vo AN SSSR. 35–70.

Vidugiris, Aloyzas. 1961. Veiksmažodžiai su priesagomis -inėti ir -dinėti Zietelos tarmėje [Verbs with the suffixes -inėti and -dinėti in the Zietela dialect]. lietuvos TSR Mokslų Akademijos darbai, serija A, 2(11), 219–231.

Vidugiris, Aloyzas. 1998. Dėl daugkartinio veiksmo raiškos pietrytinėse lietuvių kalbos tarmėse [On the iterative meaning in the Southeastern Lithuanian dialects]. Lietuvių kalbotyros klausimai 39, 183–191.

Vidugiris, Aloyzas. 2004. Zietelos lietuvių šnekta [The Lithuanian dialect of Zietela]. Vilnius: Presvika.

Vidugiris, Aloyzas. 2014. Lazūnų šnekta [The dialect of Lazūnai]. Vilnius: LKI.

Wiemer, Björn. 2001. Aspektual’nye paradigmy i leksičeskoe značenie russkix i litovskix glagolov (Opyt sopostavlenija s točki zrenija leksikalizacii i grammatikalizacii) [Aspectual paradigms and the lexical meaning of Russian and Lithuanian verbs (A comparative attempt from the point of view of lexicalization and grammaticalization)]. Voprosy jazykoznanija 2001-2, 26–58.

Wiemer, Björn. 2009. Zu entlehnten Präfixen und anderen morphosyntaktischen Slavismen in litauischen Insel- und Grenzmundarten. In Von Zuständen, Dynamik und Veränderung bei Pygmäen und Giganten (Festschrift für Walter Breu zu seinem 60. Geburtstag). Lenka Scholze & Björn Wiemer, eds. Bochum: Brockmeyer. 347–390.

Wiemer, Björn, Aksana Ėrker & Kirill Kožanov. 2019. Korpus slavjanskix i baltijskix govorov TriMCo: struktura, celi i primery primenenija [The TriMCo corpus of Slavic and Baltic dialects: structure, goals, and examples of application]. Balto- slavjanskie issledovanija 20 [Baltic-Slavic investigations 20], 122–143. DOI: 10.31168/2658-5766.2019.20.6.

Wood, Esther Jane. 2007. The Semantic Typology of Pluractionality. Berkeley: University of California (unpubl. PhD thesis).

Xrakovskij, Viktor S. 1987. Kratnost’ [Event plurality]. In Teorija funkcional’noj grammatiki, t. 1: Vvedenie. Aspektual’nost’. Vremennaja lokalizovannost’. Taksis [A theory of functional grammar, vol. 1: Introduction. Aspectuality. Temporal localization. Taxis]. Aleksandr V. Bondarko, ed. Leningrad: Nauka. 124–152.

Xrakovskij, Viktor S. 1997. Semantic types of the plurality of situations and their natural classification. In Typology of Iterative Constructions. Viktor S. Xrakovskij, ed. Munich, Newcastle: Lincom Europa. 3–64.

Abbreviations

1: 1st person

2: 2nd person

3: 3d person

acc: accusative

dim: diminutive

f: feminine

fut: future

gen: genitive

hab: habitual

idef: indefinite marker

ill: illative

inf: infinitive

ins: instrumental

ipfv: imperfective

loc: locative

m: masculine

neg: negation

nom: nominative

pa: active participle

pl: plural

prs: present

pst: past

pvb: preverb

rep: repetitive

rfl: reflexive

sg: singular

subj: subjunctive

Notes

Información adicional

Lithuanian dialectal transcription: All Lithuanian examples cited from the TriMCo corpus use additional IPA diacritics: ː for long vowels, ˑ for half-long vowels, ʲ for palatalization. Stressed vowels are marked by capital letters.