PART 1: DIACHRONIC EVIDENCE

 

Constructions with 'take' in Latgalian: The limits of diachrony

 

 

Nicole Nau

Department of Scandinavian Studies, Adam Mickiewicz University in Poznan, Poland. E-mail: naunicol@amu.edu.pl

 

 

---

ABSTRACT

A collection of Latgalian oral folktales published in 1895 shows a great frequency of multi-verb constructions with a modifying verb 'take', including Serial Verb Constructions. These constructions are not found in Old Latgalian written texts, while in modern writing, only one type is attested: pseudo-coordination. Although the documentation of Latgalian spans almost three centuries, it is not possible to show grammaticalization paths of multi-verb constructions, as these are register-specific.

Keywords: Latgalian; serial verb constructions; pseudo-coordination; take verbs; grammaticalization; register variation

---

 

 

1. Introduction

Constructions found in a language at one point in time are the result of historical linguistic processes, and synchronic variation may give hints about the nature of these processes as well as show the potential for future developments. This has been one of the cornerstones of research on grammaticalization since the 1980s (cf. Lehmann 1985). In this paper, I draw attention to the limits of projecting historical processes based on synchronic variation, especially for lesser-documented languages. I argue that variation across registers is an important factor in language change and too often is neglected in such analyses. Constructions typically arise in certain registers, where they fulfill specific functions. Some multi-verb constructions are strongly connected to oral fictional narratives (folktales), a register that is doomed to disappear or change significantly with the spread of literacy practices in the modern world. The language investigated is Latgalian, an East-Baltic language closely related to Latvian, spoken in the eastern part of Latvia in Central Europe (see Nau 2011 for general information on the language and its structure). Its documentation reaches back to the 18th century. However, before the 20th century, writing was practiced by only a few authors, most of them clergy with Polish as their main language. Modern written Latgalian, in turn, is heavily influenced by Standard Latvian, as speakers of Latgalian usually acquire and practice literacy in Latvian.

Multi-verb constructions where a verb with the basic meaning 'take' has lost its literal meaning and acts as a modifier of another verb are found all over the world. The construction most typical for European languages is pseudo-coordination (PC), where 'take' and the main verb are linked by an additive element but grammatically treated as one predicate. When this linker is omitted, the construction starts to resemble a serial verb construction (SVC). For overviews see Coseriu (1966) and Ross (2017); for comparisons of pseudo-coordination and serial verb constructions see Andrason (2019), Andrason, Gebka-Wolak, and Moroz (2022, this volume), Ross (2016; forthcoming). Nau et al. (2019) investigated pseudo-coordination with 'take' in contemporary written Baltic, Slavic, and Fennic languages, drawing data from Internet-based corpora of Latvian, Lithuanian, Polish, Russian, Estonian, and Finnish. They found that contemporary Baltic written languages have a stable, PC-type construction which is more frequent and more grammaticalized than in neighboring languages and that there is little evidence for a beginning SVC.¹ As will be shown below, this is also true for modern written Latgalian as documented in the only available corpus, which contains texts published between 1988 and 2011 (MuLa).

However, a collection of Latgalian folktales gathered in the 1890s from oral sources (Ulanowska 1895) shows quite a different picture. These texts are rich in multi-verb constructions, including various types of constructions with 'take'. In Section 2 these types will be presented in detail. In Section 3 I will shortly present the constructions represented in the corpus of contemporary texts and those found in written sources of the 18th and 19th century; the latter analysis is necessarily fragmentary, as there are no historical corpora available. My conclusion, presented in Section 4, is that the data from different time periods are not compatible and that no paths of development can be drawn from them.

 

2. Constructions with 'take' in oral folktales

The folktales gathered and published by Polish ethnographer Stefanija Ulanowska (1895) are a unique document of a register that is no longer found in modern Latgalian: orally transmitted and orally performed traditional narratives. The doculect is the central Latgalian dialect of Vilâni with some traces of other dialects. To facilitate searching within the collection, the texts of the original edition were transliterated into a modern writing system, and some inconsistencies in morphological forms were leveled out. The resulting corpus contains 50,655 words.²

The oral nature of the texts is evident in syntax and paragraph structure. A graphic sentence contains sequences of short clauses that are separated by commas or dashes and may be linked by a connector. Extract (1) is an example of such a graphic sentence. The first two lines contain multi-verb constructions with 'take'.

The presence or absence of a comma (cf. 1a vs 1b) probably does not reflect differences in prosody but seems to be random; most instances where 'take' is followed by another verb contain a comma. I assume that lines (1a-d) each correspond to one intonation unit and contain one clause, and treat (1a) and (1b) as tokens of the same construction. Here, pajimt 'take' and the following verb form a complex predicate with a single meaning (there is no act of taking something) and shared arguments. This construction is classified as an SV (serial verb) construction. In contrast, instances where 'take' and the main verb are linked by the additive connective i, as in (2), represent pseudocoordination (PC), cf. Andrason, Gebka-Wolak, and Moroz (2022, this volume).

It is however not straightforward whether SVC and PC in the Latgalian data are really two different constructions or variants of one construction. There seems to be no functional difference between the construction with and without the linker i. Furthermore, this word is not always a connective (as it seems to be here), but is also used as a focus particle, especially before verbs, for example Jei i atzagulës 'she did lie down, she indeed lay down' (after having been told to do so; B26); jis i izgaisa! 'he disappeared!' (unexpectedly; B24).

When other predicates follow, it is often not so clear whether 'take' and an adjacent lexical verb form a complex predicate. In (3), 'take' and 'count them' express one event, followed by another event, 'say'. But the verb 'take' may also be more loosely connected to the following verb, which then rather forms a predicate on its own (4). This leaves 'take' as a kind of lexically empty predicate without its own arguments except for the subject, which is shared with the following verb(s). This situation is typical when a chain of actions carried out by the same actor is named, for which 'take' signals the beginning, cf. (4).

We may thus distinguish three constructions (or variants) where 'take' functions as a modifier of a following verb or verb phrase, cf. Table 1.

However, the distinction between SVC and Auxiliary 'take' is not straightforward, and there are many examples which may represent either type. There were probably prosodic differences which are lost in the transcription.

All three constructions with 'take' emphasize the following action(s). They allow the speaker to gain momentum, so to say, and invite the listener to focus on the action. A high density of actions is characteristic for folktales, and consequently a high frequency of verbs is a register feature of these texts. More precisely, the texts show an outstanding frequency of third-person forms of verbs in past and present tense. The "verbiness" of the texts is further increased by constructions where verb forms do not strictly contribute to the content but have other functions. Further examples of such constructions found in the corpus are iteration ('they walked, walked' = 'they walked for some time') and backstitching or bridging ('They ate. Having eaten, they went on').

The constructions discussed in this paper only appear with 'taking' verbs. There are some examples with the verb giut 'catch, seize, get hold of, but the most conventionalized modifying verb is the general verb for 'take', jimt. It often appears with a lexically empty prefix pa-, especially in past tense. Table 2 shows the frequency with which third-person present- and past-tense forms of (pa)jimt appear in three formally distinguished constructions.

The folktale corpus provides adequate material for making informed guesses about the development of these constructions out of constructions where 'take' has its literal meaning. Taking an object (into one's hand, etc.) is often the first part of a chain of related actions, where the object taken is either an instrument (cf. 5) or a patient or theme (cf. 6). Constructions where (pa)jimt 'take' has (more or less) its literal meaning and the act of taking is followed by at least one other action are about as frequent as the constructions discussed above (see Table 3 below). There are several formal variants, some of which come close to the multi-verb constructions using 'take'. In (5) and (6), pajimt 'take' has its literal meaning and an overtly expressed object following the verb.

The constructions shown in (5) and (6) allow the shortening of the clause that expresses the main action by naming one of its participants in a previous clause ('took an axe and broke out windows' vs 'broke out windows with an axe'). Direct objects may be omitted if their referent is given in the context. This may lead to constructions which are formally similar to PC or SVC, but express two different actions, cf. (7).

On the other hand, the literal meaning of 'take' may fade while it still takes a direct object: in (8), the actor hardly 'takes' the door. Withpajëme durs 'took the door', the object of the main action ('close') is introduced, but by thus splitting the action over two predicates it also makes this action more prominent. It is thus functionally equal to the PC construction in (9).

Thus, the corpus shows the gradual fading of the lexical meaning of (pa)jimt 'take' as well as various tendencies of object placement and object deletion that lead to a greater cohesion of 'take' and the following verb. Both are processes of grammaticalization that may result in multi-verb constructions (PC and SVC). For both, the point of departure is constructions where 'take' has its literal meaning and explicit arguments; these are especially frequent in folktales. Table 3 shows the distribution of all occurrences of (pa)jimt 'take' in the corpus. Idiomatic uses such as 'take a wife' are counted separately; "'take' alone" means that the act of taking is not immediately followed by another action by the same actor.

 

3. Constructions with modifying 'take' in modern and old Latgalian writing

The synchronic variation which is so well attested in the folktales has almost no links to constructions documented in earlier and later written texts. I will discuss contemporary Latgalian and then present data from historical texts.

In MuLa, the corpus of modern written Latgalian (ca. one million words), only 12 instances of a modifying 'take' were detected: 10 in third-person forms, one in the imperative, and one in the infinitive. All these instances represented the PC construction with the linker i 'and'. There were also relatively fewer instances of constructions where an act of (literal) taking was followed by one or more other actions (Type B), and more instances with 'take' as the only verb in a clause (Type A). Idiomatic uses are more numerous and varied (for example, 'take part', 'take into consideration'). Compare the figures in Table 4 with those in Table 3. Taking into account that MuLa is about 20 times bigger than the folktale corpus, the difference in the frequency of 'take' in all uses is dramatic.

A formal difference not shown in Table 4 is that in the modern language, jimt is used more often for 'take' than the prefixed pajimt: of the twelve instances of PC in MuLa, nine had jimt and three pajimt, while in the folktale corpus pajimt is found more often (cf. Table 2). Modern Latgalian is here more similar to modern standard Latvian and Lithuanian (cf. Nau et al. 2019). This is also true for another feature: while in the folktales all verbs combined with a modifying 'take' express a voluntary action by a human actor, in MuLa five of the 12 instances of a PC contain a verb expressing an involuntary (non-controlled) change of state ('be born', 'die', 'appear', 'disappear', 'fall'), and one has an inanimate subject (10).

The situation in old written Latgalian is more difficult to estimate, as no corpus of old Latgalian texts is available. Therefore, no comparable quantitative data can be offered here. Searching through individual texts for constructions with a modifying 'take' yielded a few examples of two asyndetic constructions: one with the past active participle of 'take' and one in the imperative. In the oldest preserved Latgalian book, the Evangelia toto anno 1753, one instance of a construction with a participle of 'take' is attested, cf. (11)

The use of the participle in this sentence may have been inspired by the Greek original (και επιλαβόμενος Ίάσατο αυτόν, literally 'and having seized cured him'), but similar constructions are also attested in Old Lithuanian and are probably genuinely Baltic. The German and Polish translations use a finite form of 'take', as does the English King James Bible: and He took him, and healed him. In (11), the object shared by 'take' and 'cure', the pronoun 'him', stands between the two verbs, and the whole is presented as one clause without separating punctuation.

One hundred years later, we find a paraphrase of this Bible verse in a book by Kossowski (1852), also using the past participle but putting the shared object after the second verb; cf. (12)

This participle construction is not attested in MuLa. In the folktale corpus, there are two instances of a participle of lexical 'take' followed by another action (Type B in my typology). It is possible that the participle construction is an antecedent of the SV construction in the folktales, but it may also have been restricted to bookish registers.

Kossowski (1852) also has an example of the imperative variant (13):

Like (12), (13) is a retelling of a Bible verse (Mark 10:21). Neither the Greek original nor the early translations in other European languages have the verb 'take' here-instead, we find constructions with 'go': King James: go thy way, sell whatsoever thou hast; literally the same in German (Luther: Gehe hin, verkaufe alles, was du hast) and Polish (Biblia Wujka: idz / cokolwiek maszprzeday). Replacing 'go' with 'take', the author (born 1798 in Kaunas District, Lithuania, to a Polish family; from 1818 studying and working in Latgalia) draws on the resources of his repertoire of living languages.

The SVC with an imperative plays an important role among constructions with a modifying 'take' in modern Polish (cf. Gebka-Wolak 2012; Andrason, Gebka-Wolak, and Moroz 2022, this volume), but has only a marginal status in contemporary Baltic languages (cf. Nau et al. 2019:254). The folktales contain only one example, which is not a "good" instance of SVC, as the two verbs are divided by a subject pronoun (pajem tu, nükaunpati lyluokü viersi 'take (you), kill the biggest bull', B53).

In old Latgalian texts, I have so far found no examples of pseudo-coordination, nor of an SVC in past or present tense. Investigations of a larger body of texts may show that they existed; however, they could not have been frequent.

 

4. Conclusions

The folktales analyzed in Section 2 showed a wealth of synchronic variation, where constructions containing the verb (pa)jimt 'take' could be arranged on grammaticalization continua with respect to the lexical vs modifying meaning of 'take', its relative independence, and the degree of cohesion of the construction. However, when comparing this situation with data from modern and old written Latgalian, it is not possible to show the development of multi-verb constructions over the centuries. The reason lies in differences of register and medium. In the folktales, multi-verb constructions are part of the linguistic strategies used in oral narratives to organize a text that consists mainly of actions, for vivid telling and easy understanding. Authors and readers of written texts have different needs. In modern written narratives, pseudo-coordination is the only attested multi-verb construction; it is associated with emphasis and the speaker's stance. For Old Latgalian texts, data is too scarce for judgements about the functions of multi-verb constructions. Table 5 summarizes the main constructions with a modifying 'take' found in the investigated documents.

The aims of this paper were to show that synchronic variation is not always compatible with the results of diachronic investigations and to draw attention to the importance of register differences. While my findings may be disappointing to scholars interested in the development of multi-verb constructions in Baltic, it may serve as a point of departure for more detailed diachronic studies based on more varied written and spoken data-especially from the 20th century.

 

Abbreviations

acc - accusative, add - additive particle or connective, dem - demonstrative, dim - diminutive, f - feminine, gen - genitive, imp - imperative, inf - infinitive, m - masculine, nom - nominative, pa - active participle, PC - pseudo-coordination, pl - plural, prs - present, pst - past, pvb -preverb (verbal prefix), sg - singular, SVC - serial verb construction

 

Sources

Evangelia toto anno 1753. Pirmâ latgaliesu grâmata [The first Latgalian book]. Reprinted 2004 with an index and an afterword by Anna Stafecka. Riga: LU Latviesu valodas instituts.

Kossowski, T. 1851. Conciones. Mocibas Uz iadnas Nedelas wissa goda jaunin jaunparakstiejtas nu Lejksnas Baznickunga Tomasza Kossowska. 2 volumes.

MuLa = Müsdienu latgaliesu tekstu korpuss (MuLa) [Corpus of contemporary Latgalian texts]. 1 million words. Access at http://www.korpuss.lv/id/MuLa. [Accessed in September 2021]

Ulanowska, S. 1895. Lotysze Inflant Polskich, a w szczególnosci z gminy Wielonskiej powiatu Rzezyckiego. Obraz etnograficzny przez Stefaniç Ulanowska,. Czçsc III. Zbiór Wiadomosci do Antropologii Krajowej, Tom XVIII, Dz. II, 232-492. Kraków 1895.

 

References

Andrason, A. 2019. A pseudo-coordinated Serial Verb Construction "wziAc i v2" in Polish. Slovo a Slovesnost 80: 163-191.         [ Links ]

Andrason, A., M. Gebka-Wolak and A. Moroz. 2022. The rise of the WZIAC (TAKE) Serial Verb Construction in Polish. This volume.

Coseriu, E. 1966. «Tomo y me voy». Ein Problem vergleichender europàischer Syntax. Vox Romanica 25: 13-55.         [ Links ]

Gebka-Wolak, M. 2012. Wez sie zastanów! Problem status gramatycznego przyimperatywnego wei. Poradnik Jçzykowy 4: 49-63.         [ Links ]

Ivulâne, B. 2015. Palïgnozïmë lietotu darbïbas várdu sistëma latviesu valodá [The system of verbs with auxiliary meaning in Latvian]. PhD thesis, LU Riga.         [ Links ]

Lehmann, C. 1985. Grammaticalization: Synchronic variation and diachronic change. Lingua e Stile 20: 303-318.         [ Links ]

Nau, N. 2011. A Short Grammar of Latgalian. (Languages of the World / Materials 482) München: Lincom Europa.         [ Links ]

Nau, N., K. Kozhanov, L. Lindström, A. Laugalienè and P. Brudzynski. 2019. Pseudocoordination with 'take' in Baltic and its neighbours. Baltic Linguistics 10: 237-306. https://doi.org/10.32798/bl.365        [ Links ]

Ross, D. 2016. Between coordination and subordination: typological, structural and diachronic perspectives on pseudocoordination. In F. Pratas, S. Pereira and C. Pinto (eds.) Coordination and Subordination: Form and Meaning - Selected Papers from CSI Lisbon 2014. Newcastle upon Tyne: Cambridge Scholars Publishing. pp. 209-243.         [ Links ]

Ross, D. 2017. Pseudocoordinación del tipo tomar y en Eurasia: 50 anos después [Pseudocoordination with take and in Eurasia: 50 years later]. Presented at Lingüistica Coseriana VI, Lima, Peru. Presentation available at http://publish.illinois.edu/djross3/files/2013/11/Ross-tomar-y.pdf. [Accessed 12.12.2021]

Ross, Daniel. Forthcoming. Pseudocoordination, Serial Verb Constructions and Multi-verb Predicates: The Relationship between Form and Structure. Book based on PhD dissertation, University of Illinois at Urbana-Champaign, 2020.         [ Links ]

 

 

1 Ivulâne (2015), who sees an incipient SVC in the Latvian construction with 'take', has a slightly different (broader) understanding of this term.
2 The texts in original writing, a preliminary transliteration, the original Polish translation as well as a translation into English are available online at http://inne-jiezyki.amu.edu.pl/Frontend/Text/1. To facilitate finding a cited extract, reference to the number of the tale in the collection is given in this paper, for example B02 = Basnie 2.