The expression of change-of-state in the Finnic languages

4.1.1 “To come” and “to go”

Both “to come” and “to go” are old motion verbs, which are traced back to the Proto-Finno-Ugric. The equivalents for “to come” are the following: Est. tulla, Liv. tūlda, Fin. tulla, Ingr. tulla, Vot. tulla, Kar. tulla, Lud. tulda, and Veps tulda. The equivalents for “to go” are Est. lähe-, Liv. lǟdõ, Fin. lähteä, Ingr. lähtiä, Vot. lätä, Kar. lähtie, Lud. lähtä, and Veps l’ähtta. Although both verbs occur as change-of-state predicates in the Finnic languages, “to go” appeared to be marginal in all the languages in the data set (typically only 1–2 instances could be attested among 100 examples; see Table 1 and Appendix 1).

As claimed by Huumo and Sivonen (2010), the Finnish mennä “to go” can be associated with negative (including abnormal, unexpected) changes, whereas the Finnish tulla “to come” tends to occur with positive (including normal) changes; cf. mennä koomaan (go:inf coma:ill) “go into coma” vs tulla tajuihinsa (come:inf concious:pl:ill:3px) “regain consciousness”. (This makes the situation similar to the English go and come described in Section 2.2.) However, the conceptual metaphor good is here, which is at play in the case of tulla “to come”, or more precisely, the quality “good” is shown to also capture the meanings “available”, “conscious”, “accessible”, “canonical”, “brought about intentionally by an agent”, etc. (Huumo and Sivonen 2010). As a result, agentivity is said to explain the choice of tulla in example (13a) and the lack of agentivity in the choice of mennä in (13b).

Whereas the Finnish tulla appears to be a common change-of-state predicate, the Estonian counterpart is used less frequently (Pajusalu and Tragel 2007). Still the Estonian tulla is also associated with notions such as “conscious” (e.g. tulla teadvusele come:inf conscious:all “regain consciousness), “accessible”, and “available”; it is also used with professions, e.g. (14). In both languages, there are GM (13a) and SM cases (14).

The Estonian minna “to go”, in turn, in the case of animate referents, is argued to express “involuntary change in mental or physical condition, particularly towards a state of increased agitation” (15); whereas with inanimate referents (e.g. weather conditions), it is claimed to express increase or change in them (16a) (decrease is typically covered by jääda “to remain/stay”; cf. 16b; Pajusalu and Tragel 2007).

Previous results on Livonian show that neither tūlda “to come” nor lǟdõ “to go” commonly function as change-of-state predicates (Norvik 2014a). The domains listed above are frequently expressed by īedõ “to remain/stay”, which can be argued to be the most general change-of-state predicate in Livonian (see more in Section 4.1.2).

As regards the Finnic languages represented in the data set, only Ingrian and Votic showed considerably more instances of “to come” (see Table 1 and Appendix 1). In both languages, the expression of change was mainly conveyed using the GM construction as in (17). There unmarked constructions conveying change-of-state were mostly instances of an adjective complement used in a comparative form. In fact, majority of the examples contained an adjective in the complement position; there was only one clear example of a substantive used as the subject complement. The data sets of Ingrian and Votic also contained examples of experiential and existential clauses, where tulla was used to express change (e.g. 18). There were no examples of the Ingrian tulla used in SM construction or possessive constructions, but this could be attributed to the scarcity of data. As tulla was in both languages spread across various constructions to express change, it could be regarded as a common change-of-state verb in these languages. To compare, the data sets of the other Finnic languages contained only a few examples of “come” in one to two different constructions:

Instances with “to go” turned out to be infrequent in the case of each language – among 647 instances there were only 12 examples (see Table 1 and Appendix 1). Thus, there is no one Finnic languages in which “to go” would be the most widely spread change-of-state predicate. In Votic that contained relatively more instances of “to go” (altogether nine instances), further research is needed to determine its more precise usage. The nine examples in the data set expressed negative change (see 19) or change in mental or physical condition towards increased agitation (see 20).

The data set also contained instances like (21) and (22) that do not express change in the prototypical sense but rather reveal something about the origins of the development. For instance, example (21) presents the source but does not say that something turns into something else; example (22) can be seen as an extension of prototypical motion (eht “evening” is not an animate agent), but again, it is not possible to apply the schema X located in state 1 → X located in state 2 (cf. Section 1). Thus, both usages seem to be close to the source meaning of the construction.

4.1.2 “To remain/stay”

The predicate associated with the meaning “to remain/stay” is present in all the Finnic languages: Est. jääda, Liv. īedõ, Fin. jäädä, Vot. jäävvä, Ingr. jäävvä, Kar. jeähä, Lud. d’iädä, and Veps jädä. They are usually traced back to the Proto-Finnic root *jǟ-, which is considered to be a loan from Proto-Indo-Iranian (ETY).

Previous research on Estonian, Livonian, and Finnish demonstrates that while the Finnish jäädä is the least productive in the sense of expressing change (Huumo 2007: 89, Tommola 1995: 126), the Livonian īedõ shows the widest range of usages (Norvik and Prillop 2017). Estonian jääda, in turn, appears to display certain restrictions: it is often associated with the expression of negative change, such as falling ill (23a); to convey a positive change (e.g. getting healthy), another verb has to be used – saada ““get; become” (23b; Pajusalu and Tragel 2007). Norvik and Prillop (2017) show that instances of negative change may also be subsumed under the terms “difference from the majority”, “unexpectedness,” or the like:

Different from Estonian, the Livonian īedõ cannot be explained in terms of negative versus positive change as īedõ is used for both; cf. examples (24a) and (24b). In addition, īedõ can also convey emergence of something, change in physical or mental conditions (25), where Estonian or Votic would use the verb meaning “to go” (see examples (15) and (20), respectively, in Section 4.1.1). Although according to the Livonian–Estonian–Latvian dictionary (Viitso and Ernštreits 2012), the verb lǟ’dõ “to go” could be used in Livonian as well (e.g. ma’rrõ lǟ’dõ “become agitated”), īedõ in any case proves to convey change-of-state in the widest sense in a variety of constructions (further examples and discussion available in Norvik and Prillop 2017).

The data sets of other Finnic languages/language varieties (except Lutsi and Votic) tended to contain only some instances of “to stay/remain”-verbs per 100 examples (Table 1 and Appendix 1). The few instances that occurred in the data set could primarily be associated with negative change, such as becoming poor (ex. 27). In addition, the corresponding instances often expressed remaining in a previous state (ex. 26), or remaining in a state after a particular entity had emerged (ex. 28). (Norvik and Prillop 2017 discuss similar examples in Estonian and Livonian.)

To compare, in Estonian but also in Finnish, remaining in a state after an emergence of a particular entity may also be expressed using the verb “to come” (Erelt 2005: 23, VISK § 904). The data set contained one such example in Votic (29). Dahl (2000: 353) has called the respective use the “turn out” use, the motivation of which becomes clear also from the translations of examples (28) and (29) into English:

As regards the examples in the data set, only Lutsi jääda and Votic jäävvä had considerably more occurrences. Although they mainly conveyed change-of-state, neither of them revealed such a general usage as the Livonian īedõ. Primarily they could be associated with negative change (e.g. getting old, blind, ill, drunk, etc.) or with a situation becoming more passive (e.g. becoming silent; both usages are listed by Pajusalu and Tragel 2007: 297 for the Standard Estonian jääda). However, it is of interest that the Lutsi data set contained a few instances of positive change. In the case of example (30), it becomes clear from the context that the predicate with jääda is used to denote getting rich. Thus, here, the usage of Lutsi jääda shows parallels with the usage of the Livonian īedõ rather than the Estonian jääda. In Estonian, to denote getting rich, the verb saada “to get; to become” should be used instead; using jääda could be possible only if one would like to express continuation of the previous state, i.e. being rich and remaining rich. As the Latvian palikt “to remain/stay” also shows a wide range of uses in the spoken language and can be held responsible for the broad usage of the corresponding verb in Livonian (Norvik and Prillop 2017), such broader usage of Lutsi jääda confirms the claims about the possible influence of Latvian:

Further proof about the extension of the verb meaning “to remain/stay” in a language contact situation comes from Harju-Madise, a subdialect of Western Estonian, which has been under the Swedish influence. It is possible that the broader usages of the Estonian jääda can be attributed to the contact with Swedish (Evi Juhkam, personal communication, March 22, 2019). For instance, the Harju-Madise example (31) contains jääda to express change, whereas in literary Estonian another verb (e.g. muutuma “to change”) would be used in this case. In Swedish, the corresponding verb is bli, which in addition to change also conveys future (Dahl 2000). Such development cannot be detected in the Finnic languages, even in Livonian, where īedõ “to remain/stay” is the most general change-of-state verb. Most probably, the reason is that in Livonian there is already another future copula līdõ (see Section 4.1.4).

The instances conveying change-of-state in the data set tended to be GM. Both in Lutsi and in Ludic, there was one unmarked example (for Ludic, see Ludic e.g. 26). In Lutsi, the corresponding example was an instance of a complement used in a diminutive form (bednakene “poor”). For drawing further conclusions, however, one needs additional examples.

4.1.3 “To get; to become”

A verb with the meaning “to get” can be found in all the Finnic languages: Est. saada, Liv. sǭdõ, Fin. saada, Ingr. sāvva, Vot. sāta, Kar. soaha, Lud. suada, and Veps sada. They are traced back to the Proto-Uralic root *saγe-; the original meaning is usually claimed to be that of “to come” (ETY, Häkkinen 2005). This means that in the Finnic languages there are two change-of-state predicates, which originally mean “to come” (see also Section 4.1.1).

As regards the Finnic languages, the verb “to get; to become” can be found to express change-of-state primarily in the southern Finnic languages. The Estonian saada is usually described as a commonly used change-of-state predicate; it is primarily associated with desired states, positive developments, brought about by intentional action (32; Pajusalu and Tragel 2007, Tragel and Habicht 2012). Similar notions are also associated with the Finnish tulla “to come” (see Section 4.1.1). The fact that Estonian saada is also found to express non-evaluative meaning or neither desired outcome nor intentional action (ex. 33; ibid.) could point to the generalization of the predicate.

The Livonian sǭdõ is also shown to function as a change-of-state predicate, however, its usage seems to be associated with literary texts rather than with clear semantic restrictions (a similar case was introduced for Czech in Section 2.2). Whereas in Bible texts, the Livonian sǭdõ appears to be the main means used to convey change-of-state, in colloquial texts the verb īedõ “to stay/remain” is common (see Norvik 2014a). This can be argued to mirror the Latvian model; as in literary Latvian, the verb kļūt is typically used to mean “to become” (roughly corresponds to Livonian sǭdõ), whereas in colloquial Latvian, the verb palikt “to stay/remain” tends to be used instead (roughly corresponds to the respective verb in Livonian – īedõ) (Norvik and Prillop 2017). Still, one can notice differences in the underlying constructions. Latvian kļūt takes the nominative case with adjectival complements, whereas the PP with par is used with substantives denoting professions, e.g. kļūt par ārstu “to become a doctor” (Kalnača 2013: 638). In Livonian, in turn, the GM construction is used in both cases; see (34) and (35). However, the Livonian GM constructions containing a substantive reveal a clear case of Latvian influence: they contain the Latvian prefix pa although the translative marker -ks already expresses change (see Norvik and Prillop 2017; for Latvian par, see also Kalnača 2013: 638).

The data sets compiled for the purposes of the present study revealed a wide range of uses for Lutsi saada – about three fourth of the instances collected from Lutsi texts were examples of saada in various constructions (see Appendix 1). Moreover, in addition to the expression of change-of-state (36), there were instances of saada functioning as a future copula, i.e. cases where the reading “to become” seemed to remain in the background (37). It is possible that here we are dealing with an internal development as the Estonian saada does not occur alone to express future (it also requires the “be”-verb in the infinitive form, i.e. saab olema “will be”) and the respective verb in Latvian (kļūt) does not show further development into a future copula. It is, however, possible that the Latvian future copula būt “will be” has played a role in this development.

Whereas in the case of Estonian and Livonian one mainly encounters GM constructions (33) or SM constructions (Pajusalu and Tragel 2007, Norvik 2014a); in Lutsi, the unmarked cases also turned out to be common (see 36 and 37). It may be argued that the unmarked constructions also license the future meaning more easily.

As regards other Finnic languages, the Ingrian saavva and Votic saata also occurred as change-of-state predicates (see 38). As in Ingrian, tulla “to come” turned out to be quite frequent among 100 examples in Ingrian (saavva had 17 occurrences and tulla had 9); further research is needed to determine the differences between them. Currently, they are both indicated as common change-of-state devices in the Votic- and Ingrian-speaking areas (see Map 1). In any case, from the part of “to come”, Ingrian and Votic show more similarities with Finnish, whereas from the part of “to get; to become” – with Estonian. The Finnish saada-verb even appears to be absent in become-constructions, the verb tulla or other expressions of change-of-state are used instead (see Pajusalu 1994: 91; Jokela and Nummila 2015); instances like (39) are regarded as archaic/poetic (see KS). Such examples of saada are closer to the original meaning “to come” (see etymological note above). The connection with motion is also visible in the Finnish verb saapu/a “arrive” (and its Estonian counterpart saabu/da) derived from saa/da (ETY, Häkkinen 2005: 1097). For a similar use of tulla “to come”, see example (22) in Section 4.1.1.

Among 100 examples there were no instances of Ludic suada, Veps sada, and Karelian soaha used as change-of-state predicates.

4.1.4 “Will be”

In all the Finnic languages, there are at least traces of the futurate/modal verb, which goes back to Proto-Finno-Ugric. The cognate forms in the Finnic languages are the following: Est. leeda, Liv. līdõ, Fin. lienee, Ingr. leenöö, Vot. leevvä, Kar. lie(nöy), Lud. liettä, and Veps lindä (ETY, SSA). Whereas only the Old Literary Estonian and Insular dialects reveal examples of Estonian leeda, and the Finnish lienee is mainly found in the modal meaning; in the other Finnic languages the cognate verb can even appear as a future copula; see Veps example in (40; for more information, see Norvik 2015).

It is highly probable that originally we are dealing with a verb which can be subsumed under the concept become rather than will be. There are clear indications of the change-of-state usage even in few instances of Livonian līdõ used in a participle construction clearly expressing becoming (Norvik 2015: 52–3). To compare, the Ludic and Karelian cognates occurring in the data set were commonly used to change-of-state; differently from Livonian, they also have the past tense form, which can be first and foremost associated with change, as in (41).

The data set contained 213 instances of will be, which made it the most frequently occurring verb in the data set (the next most frequent verb was be born/give birth with 139 instances). However, 90 of these instances were noted down from Valdai Karelian. These were followed by 48 examples in Ludic. In Votic, Veps, and Ingrian, there were fewer examples, which all occurred in the present tense.

Especially in Valdai Karelian and Northern Ludic, the unmarked constructions turned out to be the commonest although there were some instances of SM or GM constructions. As example (41) shows, the unmarked construction was possible even in cases that clearly marked change-of-state. Thus, although one would want to associate goal-marking with change, it is not necessarily the case, especially in the easternmost Finnic languages.

Valdai Karelian and Northern Ludic dialect as well as the northernmost areas of Central Ludic stood out as next to clear instances of becoming (41), there were cases of being (42–43), i.e. Ludic liettä and Valdai Karelian lie(nöy) also functioned as future copulas (for Ludic liettä, see also Kehayov et al. 2013). The distinction between the two readings could mostly be made relying on the broader context; however, in the case of future time reference it was not always clear whether the primary meaning is that of “to become” or “to be (in the future)”. The fact that these two meanings intertwine has been already stressed before (e.g. Dahl 2000: 351). The Ingrian, Votic, and Veps examples, which only occurred in the present tense form, confirmed this claim: it was often hard to tell the difference between become and be as often both seemed to be relevant (44). However, if to compare the Ingrian and Votic tulla “to come” used with reference to future, the focus of tulla tends to be on becoming (see examples 17 and 18 in Section 4.1.1); whereas with leevvä/leenöö, it is on being (44). In addition, the suffix -ne- may be responsible for the accompanying modal meanings, as the suffix is a potential suffix conveying epistemic modal meaning (Laanest 1975: 155). However, as has been shown, the presence of -ne- does not necessarily evoke a modal meaning (see discussion in Norvik 2013).

4.2.1 “To make/do”

In Veps, the verb t'ehtakse, which is a reflexive form of the verb t'ehta “to make/do”, appeared to be a general change-of-state predicate (45–46). It is described as a change-of-state verb also by Grünthal (2015: 146–147). Among the 56 instances of t'ehtakse, the unmarked construction (45) turned out to be the commonest (20 examples); there were only three instances of GM and two instances of SM constructions (46 and 47, respectively). The examples of GM constructions in the data set were all instances of substantives in the complement position (e.g. ex. 46).

The Veps t'ehtakse could be regarded as a general change-of-state predicate as in addition to the examples mentioned above, it also appeared in the existential, experiential, and possessive constructions (see Appendix 1). These occurrences could be associated with change in a physical/physiological state (ex. 48, which is an instance of an experiential construction) or in the general state of affairs (ex. 49, which is an example of a possessive construction).

The Veps t'ehtakse most probably reveals a case of pattern-borrowing (PAT-borrowing) as in Russian the corresponding reflexive verb (s)delat'sja “to make/do” can also be found as a change-of-state verb in various constructions; see the GM construction in (50) (cf. 45) and the experiential construction in (51) (cf. 48). According to Sakel (2007: 15), PAT-borrowing has to do with borrowing the “organization, distribution and mapping of grammatical or semantic meaning, while the form itself is not borrowed”. The Veps case is an instance of borrowing the semantic meaning. To compare, whereas the Veps t'ehtakse tends to occur in unmarked constructions, the various examples of the Russian (s)delat'sja in Russian–Estonian and Estonian–Russian dictionaries (VES, EVS) did not contain any unmarked constructions; the typical construction appeared to be the GM construction as in example (50).

Although the Veps t'ehtakse was commonly used to express change-of-state, there were no instances that would reveal its further development into the future copula. In fact, in the data set t'ehtakse mainly appeared in the past tense (there were only two instances containing a present tense among 56 examples). It is most probably because Veps already has a future copula – lindä “will be”, which may also show instances of intertwining of being and becoming. This makes the Veps situation similar to Livonian, where īedõ “to stay/remain” only occurs as the general change-of-state predicate but no further development into future can be detected as there is the future copula līdõ “will be”, which is a cognate to Veps lindä (see more in Section 4.1.4).

Among the examples collected from the other Finnic languages, there were hardly any examples of a “to make/do”-verb used to express change-of-state. Still one such example seemed to occur in Lutsi (52). Differently from Veps, the Lutsi tetäˀ is not a reflexive verb. But as it appears, the uses of non-reflexive can also be attested in other languages; e.g. in Texas English, make may be used to convey “to become”: She made a teacher = “She became a teacher” (Clancy 2010: 35).

4.2.2 “To be born; to give birth”

Predicates meaning “to give birth”/“to be born” were attested to express change-of-state in Livvi-Karelian (53) and Ludic (54). Differently from Veps t'ehtakse, which is a native verb employed to express change-of-state as in Russian (i.e. a case of PAT-borrowing), the Ludic rotie(kseh) and Livvi-Karelian rodida/rozdizetta are instances of matter-borrowing (MAT-borrowing) as the “material” is taken from Russian: < rodit'sja “to be born”, rodit' “to give birth” (for MAT-borrowing, see Sakel 2007). Both the non-reflexive and the reflexive forms were used (cf. the reflexive uses of the Veps t'ehtakse in Section 4.2.1).

The Ludic rodida/rodizetta as well as Livvi-Karelian rotie(kseh) occurred in various constructions to convey change-of-state. Most typically, change was expressed by unmarked forms; there were only a few instances of the GM construction (54) (this was true also for the Veps t'ehtakse; see Section 4.2.1). There were also examples conveying change in physical/physiological state (55), and the emergence of something (56). The Ludic data set did not contain any examples only of SM construction, which does not necessarily mean that the SM construction is impossible. However, as regards Ludic, rodida/rodizetta appeared to be as a general change-of-state predicate only in Central Ludic; in Northern Ludic, there were only few examples of rodida/rodizetta as liettä “will be” is used instead (see Appendix 1).

In Central Ludic as well as in Livvi-Karelian, one can observe further development into the future copula (for Ludic, see also Kehayov et al. 2013). Namely, there are instances such as (57) and (58) that primarily convey being with reference to the future rather than becoming. Such further development is probably licenced by the fact that liettä/lie(nöy) “will be” do not function as primary future copulas in these areas. The Central Ludic liettä had only 14 occurrences of which 7 examples were instances of an existential clause, which could be rather associated with the emergence of something rather than prototypical change-of-state usage. Thus, the usage of the Central Ludic rodida/rodizetta and Livvi-Karelian rotie(kseh) “be born/give birth” can be argued to reveal a case of renewal/replacement as they seem to be used in the same functions as the native verbs liettä and lie(nöy) “will be” (for renewal/replacement, see also Section 2.1).

In the North Russian dialects, the respective Russian verbs rodit'/rodit'sja are attested to mean “begin”, “become”, “appear”, “take place”, etc. (SRNG). The fact that we are dealing with an areal-specific phenomenon found further support from other Karelian dialects. For instance, the Valdai Karelian data set contained only a few instances of rotie(kseh) that all occurred in an existential construction to express change in weather conditions (59). As the verbs studied in this article were all used in such function (see also example 17 for “to come” and example 39 for “to get”), it is possible that such usage represents an initial stage on the path towards “real” change-of-state predicate: