1. Introduction
In the earliest attested stages of Swedish, the verb varda ‘become’ was used both as a copula, taking adjectival and nominal complements, and as a passive auxiliary, combining with a passive participle. See the runic inscriptions in (1).Footnote 1
Around 1300 the Middle Low German (MLG) verb blîven ‘remain’ was borrowed into Swedish as bliva. In the earliest examples it retains the original meaning ‘remain’, as shown in (2).
But around two hundred years later, bliva had acquired the meaning of varda, ‘become’, as shown in (3a), and had become the more frequent of the two. It was also used as a passive auxiliary just like varda, as shown in (3b).
The change in meaning from the static ‘remain’ to the dynamic ‘become’ is unexpected. The aim of this article is to pinpoint the crucial stages in the development, in particular the increased use with adjectives denoting a change of state. We identify the subsequent extensions in the use of bliva and the contexts critical for establishing the new meaning. By looking at the various constructions in which varda and bliva were used and tracing their relative frequencies, we can establish a timeline of the change, its stages, and the critical contexts.
Based on earlier research (Markey Reference Markey1969; Rosenthal Reference Rosenthal1984; Skrzypek Reference Skrzypek2020, Reference Skrzypek2024a,b,c; Woźnicka Reference Woźnicka2024), we have identified 1450–1600 as the critical period for the change and have conducted a corpus study of seven Swedish texts spanning this period. In addition to the original meaning of the verb bliva ‘remain’, it is increasingly used with the new meaning ‘become’, which continues to be expressed by the verb varda as well. When bliva is used meaning ‘become’, it is more and more likely to appear in new contexts, including the passive construction. Gradually, the original meaning ‘remain’ is no longer expressed with bliva; in order to express this meaning, either adverbs such as kvar ‘remaining’ and åter ‘again’ are added, or the newly formed förbli ‘remain’ (also of MLG origin) is used. At the final stages of the development, the use of varda is very limited (Skrzypek Reference Skrzypek2020, Reference Skrzypek2024a).Footnote 2
We also consider the role played by the continued contact with MLG, which subsided first in the early 1500s. The contact with German continued well into the 1600s, but gradually the influence involved High German rather than Low German. Earlier research has shown that the two varieties differ with respect to the use of the verb blîben; in particular, the only meaning of the verb in High German is ‘remain’ (Woźnicka Reference Woźnicka2024).
The paper is organized as follows. In Section 2 we review previous research and give the background to the current investigation. In Section 3 we present the texts used, discuss their limitations, and give a quantitative overview of the use of varda and bliva in these texts. In Section 4 we present the various constructions in which varda is used in the chosen texts and in Section 5 we do the same for bliva. This is followed by a qualitative analysis of the uses of bliva as meaning either ‘remain’ or ‘become’ in Section 6. Section 7 is devoted to the rise of bliva as a passive auxiliary, the main factor behind the increased use of bliva. We discuss the factors influencing the development in Section 8 and end with some concluding remarks in Section 9.
2. Background
In order to express the meaning ‘remain in a location’, the common strategy in Old Swedish was to use posture verbs such as sitja ‘sit’ and liggia ‘lie’, as well as staþna ‘stay’ (stanna in present-day Swedish) and, to some extent, even vara ‘be’.Footnote 3 The earliest examples of bliva in Swedish texts also have the meaning ‘remain in a location’, and are most often used with locative adverbials, as in (2). But soon the verb bliva was also used as a copula with adjectival complements, with a more general meaning ‘remain’, as shown in (4).
The earliest example of bliva with the ‘become’ meaning is found in Codex Bureanus, the Swedish translation of a compilation of Christian legends (Legenda Aurea), from around 1330; see (5) and compare the similar phrase in (1a).
Although the source expression, the MLG verb blîven, was mainly used with the meaning ‘remain’, there are also contexts in which it could acquire the meaning ‘become’. One such context is precisely with the adjective tod ‘dead’, as suggested by Markey (Reference Markey1969:77); consider the MLG example in (6).
According to Markey (Reference Markey1969:77), the MLG expression dot blîven tended to be used with a locative adverbial, and expressions like he blef dot originally meant ‘he remained there dead’.Footnote 4 In a similar vein, Tamm (Reference Tamm1890–1905) suggests that the meaning of the Swedish han blev död was han blev död kvar på platsen ‘he stayed dead in the location’. This ambiguity of meaning: ‘remain’ or ‘become and remain’, together with the high frequency of the expression, may have led to a change of meaning, to just ‘become dead’. As such, it could easily be seen as a parallel to the Swedish varda död, as in (1a), and was probably the first context in which the similarities between varda and bliva were exploited.
In Swedish texts from the 1300s and 1400s, the expression bleff dödhir is particularly frequent, gradually replacing vard dödher. It is likely that it became a lexicalized phrase with the meaning ‘died’. In the religious text Själens tröst (ST) from around 1420, there are 47 examples of bliva död ‘become dead (= die)’, as shown in (7), and not a single one of varda död.Footnote 5
Example (8) is taken from a passage describing how a blind man regained his sight after having prayed at the temple. Here the meaning of bliva can perhaps be better rendered as ‘become and remain’ or ‘turn into’, as is the case for bliva död. The verb is usually not used for a temporary change, but about a lasting one.
Apart from the lexicalized bliva död, the earliest example of the ‘become’ meaning, shown in (9a), comes from Sju vise mästare (SVM), a text from the early 1400s. In this text bliva is otherwise used with the meaning ‘remain’, as shown in (9b).
In the 1300s and early 1400s, bliva could be used with its original meaning ‘remain’, in contrast to varda, which always had the meaning ‘become’. Consider the two examples from Själens tröst in (10) about a son who thought that his father led an unjust life.
In (10a) the son becomes a hermit. After some time he considered returning to his former life, but came to his senses and remained a hermit, as shown in (10b).
Apart from the occasional uses of bliva as ‘become’ before 1450, the dominant use of the verb is in its original meaning ‘remain’. The ‘become’ meaning becomes common first during the 1500s. Skrzypek (Reference Skrzypek2020:322) quotes early examples of the passive construction with bliva from Historia Trojana from 1529, but it is not until after 1550 that the construction is well established (Skrzypek Reference Skrzypek2024a:Table 8). These observations lead us to focus on the period 1450–1600 as the relevant time to investigate the change in meaning of the verb bliva.
3. Corpus investigation
The study is based on a collection of seven Swedish texts spanning 100 years, from Lucidarius (1487) to Brahe’s chronicle (1585). We include two gospels from the translation of the New Testament (1526), Matthew and Luke, as they were most likely translated by two different translators (Lindblad Reference Lindblad1971, Wollin Reference Wollin2024), with divergent use of the verb bliva (Skrzypek Reference Skrzypek2024c). All texts are available in a digitized format, which allows a complete search.Footnote 6 We have excerpted all instances of the two verbs, varda and bliva, from these texts. The sources are presented in Table 1 together with the total number of occurrences.
Excerpted texts and number of occurrences of varda and bliva
Despite efforts to collate a balanced corpus, there are some differences between the sources: Luc, Linc, NT Luk, NT Matt, and Troj are translations, whereas Petri, Swart, and Brahe are all historical chronicles originally written in Swedish. There are also considerable discrepancies in length. As with all diachronic research, the choice of source texts is limited. However, these texts have all been used extensively in previous linguistic research and we believe that they are suitable for the aims of the present paper.
The material for the present study consists of 3274 examples of either varda or bliva. The frequencies of each verb (per 1000 words) are given in Figure 1.
The frequencies of varda and bliva per 1000 words.
As can be seen, varda is more frequent than bliva in the oldest texts in the corpus (Luc and Linc) and in the NT. The proportions change in Troj; bliva gains in frequency while varda becomes more and more marginalized. In the youngest text, Brahe from 1585, there are only five instances of varda, despite the fact that this text is comparable in length to both NT Luk and NT Matt.
In order to trace how bliva gradually took over several functions from varda, we first identify the functions of varda and their relative frequency in the following section and then, in Section 5, turn to the functions of bliva.
4. The use of varda
In this section we show in which constructions varda is used in the corpus.Footnote 7 As already shown in (1), varda was used with adjectival complements. It was also used with nominal complements. Examples from our corpus are shown in (11a–c).
Another common construction involved present participles (glossed as aprt, active participle), as in (12).
When combined with a passive participle, varda functioned as a passive auxiliary, as in (1b) and (13).
In addition, varda could be used without complements, in what we refer to here as an absolute use, where the meaning of the verb is ‘to come into existence’; see (14).
Very rarely, we also find examples of varda with a PP complement or adverbial, as in (15).
However, we do not find varda with locative adverbials unless there are other complements present; see the discussion in Section 7.1.
The constructions vary greatly as to how frequently they are used, as can be seen in Figure 2.Footnote 8
Constructions with varda per 1000 words.
Adjectival and participial complements are found in all the texts studied; the passive participles dominate, except for Linc and Troj. The gospels stand out with a high number of present participles with varda.Footnote 9 This construction was used to express future reference, as shown in (16a). It was likely modelled on the MLG future construction with werden and can be found in other texts from the 1500s, as shown in (16b).Footnote 10
However, bliva does not replace varda in this construction, and the whole construction disappears from Swedish around 1700 (Skrzypek Reference Skrzypek2024a). We do not consider it further here.
5. The use of bliva
As already mentioned, the earliest instances of bliva in Swedish are used with a locative adverbial, underlining its original meaning ‘remain’, as in (2). These are the examples found in religious prose from the 1300s and early 1400s (Skrzypek Reference Skrzypek2020, Reference Skrzypek2024a). With locatives, the original meaning of bliva, ‘remain’, is also found in our later texts, as shown in (17).
Just like varda, bliva was used with adjectival complements, as shown in (18), as well as with nominal complements, as shown in (19).
Examples with present participles can be found, as shown in (20), as can absolute uses of bliva without complements (meaning ‘to come about’), as shown in (21). Both uses are rare.
As bliva gains in frequency, the use with passive participles becomes more prominent, starting in NT 1526 and Troj 1529, as shown in (3b), and (22).
An overview of the constructions with bliva is shown in Figure 3. Comparing the results with those given in Figure 2, we can see that the rise in frequency of bliva is almost entirely due to the rise in frequency of the construction with a passive participle, i.e. the passive construction (see Section 7). We note that adjective and locative complements are possible all through the period. NP complements are hardly used before NT Matt. Note the low number of examples with present participles, a difference from varda. This is presumably due to the fact that the future meaning of varda + present participle was not taken over by bliva (cf. (16)).
Constructions with bliva per 1000 words.
6. The meaning of bliva
In this section we trace the gradual increase of examples with the new meaning ‘become’. In order to determine which meaning was intended, we judged each example in its context. Most of the time it was straightforward to determine the intended meaning, but in some cases both the ‘remain’ and the ‘become’ meaning were plausible; these examples are shown in grey in Figure 4, which illustrates the distribution (% of all bliva) of the two meanings throughout the corpus. Since the number of occurrences varies between the texts (see Table 1), we give the absolute numbers in parenthesis following the name of the text.
The meaning of bliva in the corpus (% of all bliva in a given text; absolute numbers shown after the text abbreviation).
We see that bliva is mainly used as ‘remain’ in the oldest texts in the corpus, Luc and Linc. The innovative meaning ‘become’ is dominant in all other texts. NT and Troj pattern with the younger texts in the sample, with Matthew being more progressive than Luke.
We now look more closely at the factors which influence the meaning of bliva. If there was a locative argument present, the meaning was always ‘remain’, as shown in (23).
The ‘remain’ meaning was increasingly reinforced by a particle, as in (24a, b).
In examples with adjectival and nominal complements, it was often the case that we needed to look closely at the context in order to determine the meaning. Consider the following passage from NT Luk 1:20–22, where the angel Gabriel tells Zachariah that his wife Elizabeth would become pregnant. Zachariah reacts with incredulity and the angel tells him that he will become speechless, using a future reference thet skall skee ‘it shall be’; see (25a).
When Zachariah comes out of the temple, he still cannot speak, kunde han intit tala ‘he could not speak’, which makes it plausible that bleeff måållöösz in (25b) should be interpreted as ‘remained speechless’.
In the material, we see a growing tendency for the ‘become’ uses of bliva with adjectival complements, shown in Figure 5.
The meaning of bliva with adjectival complements.
The height of each bar corresponds to the number of occurrences of [sbj bliva adj] in each text. The differences are partly a result of the differences in text length and do not illustrate the frequencies of the construction (these are shown in Figure 3). We note that the ‘remain’ meaning gradually disappears; in Troj (1529) the verb is only used with the ‘become’ meaning with adjectives. Furthermore, with adjectives that are etymologically adjectivized participles (Hellqvist Reference Hellqvist1922), e.g. dödher ‘dead’ and rädder ‘scared’, bliva is always interpreted as ‘become’ in the entire corpus; see (26).
There are only a few examples in the early texts where bliva is used with an NP complement, [sbj bliva np]; see Figure 6. When the construction starts to be used more frequently, from NT 1526, the ‘become’ meaning dominates. As with adjectival complements, the meaning of the verb depends on the context, as shown in (27a, b).
The meaning of bliva with nominal complements.
In (27a) the combination wara och bliffua strongly suggests the interpretation ‘be and remain’, whereas the temporal adverb sedan ‘then’ in (27b) invites the inference that he became chancellor.
The results for bliva with present participles, [sbj bliva aprt], are the least informative, which is in part due to the fact that the construction is infrequent. There are nine occurrences in Petri (1530), all with the verb meaning ‘remain’, but just one in Troj (1529), where the verb is used as ‘become’; see Figure 7.
The meaning of bliva with present participles.
Generally, the [sbj bliva aprt] construction allows for both ‘remain’ and ‘become’ uses of the verb. In this construction, we find quite a lot of variation and a number of ambiguous cases, as late as in Swart and Brahe; consider example (28).
In (28), it is conceivable that the soldiers went and stopped (blev ståendes lit. ‘became standing’), but it is equally conceivable that that they were already on a plain and stayed there (blev ståendes lit. ‘remained standing’). The exact reading is not of great consequence, but the example shows that in certain contexts the meaning of bliva was ambiguous. In general, the ‘remain’ meaning is more likely with participles derived from posture verbs;Footnote 11 the nine examples in Petri are all of sittia ‘sit’, hengia ‘hang’, standa ‘stand’, or liggia ‘lie’.Footnote 12 Participles derived from other verbs favour the ‘become’ reading of bliva in this construction. However, as (28) shows, at times it is impossible to decide the meaning of bliva in its context.Footnote 13
7. The passive auxiliary
Perhaps the most striking change is the replacement of varda by bliva as a passive auxiliary; see examples (1b), (3b), and (22), and the overview in Figure 8. The numbers given by each text, e.g. Luc (46), refer to the total number of occurrences of the passive construction with varda or bliva.
Choice of auxiliary in the periphrastic passive (% of all occurrences with passive participles; absolute numbers shown after the text abbreviation).
In the two oldest texts, Luc and Linc, there are no clear occurrences of bliva as a passive auxiliary, whereas in the youngest text, Brahe, 161/163 auxiliaries are bliva. Footnote 14 Gradually, the frequency of bliva rises and the frequency of varda diminishes.
The first clear examples of passive with bliva are found in the first translation of the New Testament from 1526, as shown in (29a), and in Historia Trojana from 1529, as shown in (29b). All occurrences are interpreted as ‘become’.
As can be seen in Figure 8, NT Luk and NT Matt are the last texts in which the passive with varda is more frequent than the passive with bliva. In the entire NT varda is used in 90% of all passive constructions (Skrzypek Reference Skrzypek2024c). There are, however, discrepancies between the different NT books, which most likely reflect the fact that they were translated by different people (Lindblad Reference Lindblad1971, Wollin Reference Wollin2024). In Luke we find bliva in 10% of the periphrastic passive constructions, but the percentage is as high as 34% in Matthew (Skrzypek Reference Skrzypek2024c). In the later texts in our corpus, varda can still be used as a passive auxiliary, although gradually bliva takes over.
7.1 A possible bridging context
Although there are no clear examples of bliva used as a passive auxiliary in the oldest text Luc from 1487 (see Figure 8), there are some examples with passive participles that might provide a potential bridging context for the introduction of bliva as a passive auxiliary; consider (31).
At first glance, (31) appears to be an example of a passive construction with auxiliary bliva and passive participle ynnelykter. We would like to argue, however, that another reading is also possible and potentially more likely. Elsewhere in Luc, bliva is used in its original meaning ‘remain’ in over 70% of the examples; see Figure 4. In (31) there is also a locative adverbial j modrens lifwe ‘in the mother’s womb’, and a temporal adverbial signalling durativity jx manadha ‘nine months’. Both these featues are more compatible with ‘remain’ than with ‘become’. The participle ynnelykter would under this interpretation not be a complement of bleef but rather an attribute, predicated of the subject, as shown by the agreement. We suggest that given the dominant meaning of bliva in Luc, example (31) should be read as ‘remained confined in her womb’ rather than the passive ‘was locked (by somebody) in her womb’. Such contexts can thus constitute bridging contexts for the rise of the new meaning and, in particular, the use of the verb as a passive auxiliary. Similar uses of attributive participles are found with posture verbs such as sitta ‘sit’ both in Old Swedish, in (32a), and in present-day Swedish, in (32b).Footnote 15
7.2 Animacy in periphrastic passives
In present-day Swedish, use of the periphrastic bliva-passive is strongly correlated with animate subjects and the morphological s-passive is preferred with inanimate subjects (Engdahl Reference Engdahl, Lyngfelt and Solstad2006, Laanemets Reference Laanemets2012). It appears that this difference was present also in Old Swedish. In an investigation of the use of passive in early Swedish laws from around 1225 to 1350, Holm (Reference Holm1952:261) found that around 80% (169/213) of passive examples with varda had subjects that denoted animate entities. This is a noticeably higher proportion than for the more common morphological s-passives, which are the main object of Holm’s study. He estimates that the percentage of animate subjects with s-passives varied between 10% and 30% in the different laws (Holm Reference Holm1952:265). We have not investigated s-passives in our corpus, but Skrzypek (Reference Skrzypek2024b) shows that bliva-passives tend to be used with animate subjects in texts from the late 1500s and 1600s. In this section we investigate whether there are any differences with respect to animacy in the choice of subject in the competing periphrastic passives with varda and bliva in our corpus from 1450–1600. Figure 9 shows the percentage of animate and inanimate subjects in passive constructions with varda and bliva.
Subject animacy in the varda-passive and bliva-passive (% of all occurrences with passive participles; absolute numbers shown after the text abbreviation).
In two texts, Linc and Brahe, the varda-passive is used exclusively with inanimate subjects. As there are only three instances of the construction in Linc (and none of the bliva-passive), this is not very informative. Brahe, on the other hand, seems to have switched to bliva as a passive auxiliary, as there are also only two instances of the varda-passive in the text, compared to 161 passives with bliva.
Starting from NT 1526 there is a clearer preference for animate subjects with bliva than with varda, in particular in NT Matt (1526) and Troj (1529). This is partly due to the choice of the verbs used; in NT, many of the passive constructions with bliva are constructed with participles derived from verbs such as bota ‘heal’, mätta ‘satisfy, feed’, and hälsa ‘greet’, which take animate objects (subjects in the passive clause). Similarly in Troj there are several occurrences of the bliva-passive with såra ‘hurt’, gripa ‘capture’, and slå ‘kill’, which typically take animate objects. However, these verbs are also used with varda in passive constructions. It is possible that the tendency towards animate subjects is a consequence of the low frequency of the construction rather than a genuine difference between varda and bliva. At the same time we may note that in texts with low frequencies of the bliva-passive, such as NT Luk, NT Matt, and Troj, when the construction is used, there is a preference for animate subjects.
It is also interesting to look more closely at the chronicle by Olaus Petri, which has an equal number of varda- and bliva-passives (see Figures 8 and 9). It turns out that although we can discern a preference for inanimate subjects in varda-passives and animate subjects in bliva-passives, there is a lot of variation. The same verbs can be used with either inanimate or animate subjects, as shown in (33a–d).
There is nothing in our data that would unequivocally indicate why varda or bliva is chosen in a given context. However, we note that for authors who use both bliva and varda frequently and in equal proportions, i.e. Petri (1530) and Swart (1560), there is a slight preference for animate subjects with bliva (see Figure 9). This preference becomes even clearer in the late 1600s and 1700s, with almost 70% of animate subjects in the bliva-passive (Skrzypek Reference Skrzypek2024b).
8. Discussion
The change in meaning of the verb bliva from ‘remain’ to ‘become’ was a lengthy process, closely connected with the syntactic status of bliva, which can be divided into two major changes: first from a lexical verb to a copula and then from a copula to a passive auxiliary.
The first stage of the development took place soon after the verb was borrowed into Old Swedish, in the mid 1300s. Skrzypek (Reference Skrzypek2020) quotes several examples from the late 1300s in which the verb is already used as a copula with adjectives and noun phrases, in particular adjective phrases. It is in this context that we find most variation and also a clear tendency for the ‘become’ meaning to gain in frequency with time (see Figure 5). Notably, with adjectives which etymologically are adjectivized participles, such as dödher ‘dead’ or rädder ‘afraid’, the verb’s meaning was always ‘become’; see (26). The crucial context was most likely the phrase bliva död ‘die’, especially if we accept Tamm’s and Markey’s proposals that the ‘remain’ meaning was supported by locative adverbials, as in ‘he remained there, dead’; see (6) and (23). Note that in this analysis, the adjective remains outside the argument structure of the verb bliva and serves rather as a predicative attribute, clarifying in what way the subject remained at some location. However, as död ‘dead’ implies a change of state (from ‘alive’), it is also conceivable that the change of meaning took place through this context, as argued by Markey (Reference Markey1969:77); MLG blîven could in fact be interpreted as ‘become and remain’ rather than just ‘remain’ and from there reinterpreted as ‘become’.
Once the potential ‘become’ (or ‘become and remain’) meaning was established, bliva with adjective and noun phrases was ambiguous and its interpretation highly dependent on the context, as shown e.g. in (25). It is possible that the contrast between varda and bliva was gradually obscured, from ‘become’ versus ‘remain’ to ‘become’ versus ‘become and remain’, resulting in a partial overlap between them. Nevertheless, remnants of the ‘remain’ meaning persist in the material, so that we find bliva used with locative adverbials also in the later texts, as shown in (23). We can also see this in an example from NT in (34).
In (34) there is a contrast between wardher uptagen, implying a new state, and bliffuer quar låten where the ‘remain’ aspect of bliva resonates well with quar ‘remaining’.Footnote 16 In the mid 1500s, the prefix för (from MLG vor) could be added to bliva in order to emphasize the ‘remain’ meaning; see an early attestation in (35).
However, the development proceeds against the backdrop of an already existing periphrastic passive construction with the verb varda, which always was interpreted dynamically, as ‘become’. During the period investigated, both verbs were used as copula verbs and could be used interchangeably, as in (36).
There are also examples which show that the ‘remain’ meaning of bliva could still be contrasted with the ‘become’ meaning of varda in the early 1500s. Skrzypek (Reference Skrzypek2020:319) quotes the example in (37) from a religious text showing that the two verbs could still be perceived as contrasting.
The second major stage in the development was the further grammaticalization of bliva into a passive auxiliary. This was only possible after the new ‘become’ meaning was well established, but the presence of the varda-passive and other constructions with varda may have acted as a catalyst for the development when the meaning of bliva became polysemous. Figure 3 shows that it was precisely the use in passive constructions that resulted in bliva’s rapid rise in frequency, eventually replacing varda. Although the changes in Swedish took place during a period when there was considerable language contact between Swedish and MLG (Wessén Reference Wessén1954, Höder Reference Höder2010, Rambø Reference Rambø2010), the development in Swedish is very different from that in MLG. As we saw in Section 2, although the meaning of blîven in MLG was ‘remain’, the verb could also be used with the adjective tod ‘dead’, meaning ‘become and remain’. However, the ‘become’ meaning of blîven was seldom utilized in MLG (Rosenthal Reference Rosenthal1984, Woźnicka Reference Woźnicka2024), and there is hardly any evidence that blîven could be used as a passive auxiliary, similarly to what happened in Swedish.Footnote 17 In Middle High German, which gradually replaced MLG as the relevant contact variety in Sweden (Braunmüller Reference Braunmüller2000), blîven only had the meaning ‘remain’ and could not be used as a passive auxiliary; see also Kotin (Reference Kotin2000).
Returning briefly to the contrast in meaning between ‘remain’ and ‘become’, they are actually closer than it might seem. Löbner (Reference Löbner1990) and Steinitz (Reference Steinitz1999) argue that the two verbs are dual operators (Ger. duales Paar), related to each other by inner and outer (= dual) negation (i.e. become = NEG remain; remain = NEG become).Footnote 18 Hansen & Heltoft (Reference Hansen and Heltoft2011:193) describe the related Danish verb forblive as a verb that belies that the change implied by blive took place.Footnote 19 However, it is first when the ‘become’ meaning is well established that bliva develops into a passive auxiliary, a process that is well documented cross-linguistically (Haspelmath Reference Haspelmath1990:38 and references therein).
9. Concluding remarks
In this article we have concentrated on describing the change of meaning of the Swedish verb bliva from ‘remain’ to ‘become’, detailing its timeline and most important stages. We have also discussed possible factors influencing the change. The extension of the meaning of bliva from ‘remain’ to ‘become’ takes place in the late 1400s and early 1500s. During this period the verb continues to be used with both meanings. However, the ‘become’ reading gradually gains in frequency and the ‘remain’ meaning disappears and is retained only in expressions with locative adverbials, usually with particles like kvar ‘remaining’ to clarify the intended reading; see (24).
Throughout the period studied, the interpretation of the verb is to a large extent dependent on its complements. With locative adverbials, bliva is a lexical verb with ‘remain’ as the only reading; with passive participles it is an auxiliary with ‘become’ as the only alternative. In the other constructions, where the verb is used as a copula, both readings are possible and the context is necessary to identify the intended meaning of the verb. In particular, this concerns the adjectival complements; see examples (25a, b) and Figure 5.
The change of meaning takes place gradually, but 1450–1600 is the critical period. At the onset of the period we find occasional examples of bliva meaning ‘become’ but none in its auxiliary function. This is first attested in the New Testament (1526) and Historia Trojana (1529). The latter text, traditionally included in the Old Swedish period (1225–1526), in this respect seems to pattern with Early Modern Swedish (1526–1732).
Nya testamentet 1526 is the first complete translation of the New Testament into Swedish and a document constitutive for the periodization of the Swedish language, marking an important stage in the history of the language. As we noted in Section 7, there are discrepancies between different parts of the text, with Luke representing the more conservative and Matthew the more progressive group. However, we find examples of bliva used with the meaning ‘become’ in all the books in the text and, more importantly, used as a passive auxiliary, even in the more conservative parts (Skrzypek Reference Skrzypek2024c). This suggests that at the time of the NT translation the development was sufficiently advanced for the bliva auxiliary to be part of the grammar of all the translators.
We have identified a potential bridging context in which bliva is used with a passive participle which is predicated of the subject without being a complement of the auxiliary; see example (31). It is easy to see how the whole construction could be reinterpreted, so that the participle was seen as part of the verb’s valency, which gradually led to the rise of the bliva-passive. However, this was only possible once the ‘become’ meaning was established.
When used as passive auxiliaries, we find an increasing difference in the choice of subject type whereby varda tends to be used with inanimate subjects and bliva with animate subjects (see Section 7.2). This suggests that the correlation between the bliva-passive and animate subjects which we find in present-day Swedish is discernible at the early stages of the rise of the construction. A preference for different types of subjects could explain why, at least in the 1500s, we find two seemingly identical periphrastic constructions, but more research is necessary.
The traces of the ‘remain’ meaning become more and more rare and are almost absent in the texts from the late 1500s. By that time, bliva is used in the same contexts as the original varda. As a result, the frequencies of varda dwindle and eventually the verb disappears from the standard language. Similar developments have taken place in the closely related languages Danish and Norwegian, which were also influenced by Middle Low German and which had close contact with Swedish.Footnote 20 We find early examples of the blive-passive in Danish in the late 1400s, some forty years earlier than in Swedish.Footnote 21 More research is necessary to chart the developments in Danish and Norwegian, and to compare them with those we have documented for Swedish, so that a more complete picture can be painted of how bliva replaced varda.
Acknowledgements
This research was supported by the Polish National Science Center (grant number 2021/41/B/ HS2/00011) to the first author. We would like to thank the editors of the journal and three anonymous reviewers for their valuable and constructive comments.
Open access
