1 Introduction

In this paper I will describe the referential and pragmatic-discourse functions of three Lithuanian reference impersonal constructions, which I will label 2sg-imp (1), 3.imp (2) and ma/ta-imp (3).

(1)

(2)

(3)

All three constructions are used to encode events where the agent – obligatorily human – is de-focused: semantically, it is very low in individuation (non-referential); morpho- syntactically, it is realized as zero as in (1) and (2), where the subject is encoded through verbal agreement only, or it is completely demoted, as in (3). 2sg-imp, 3-imp and ma/ta-imp overlap to an extent, as they are all agent-demoting constructions: however, each one has specific semantic properties and discourse functions. In the following sections, I will describe the referential properties as well as the discourse-pragmatic functions of 2sg- imp, 3-imp and ma/ta-imp in contemporary Lithuanian. The paper is organized as follows: In Section 1. I offer a typological overview of impersonal constructions and in Section 2. I introduce 2sg-imp, 3-imp and ma/ta-imp in Lithuanian. In Section 3, I present my data and the results of my analysis of the corpus. Finally, in Section I draw some conclusions.

1.1 Impersonal constructions from a typological perspective

In typological literature, different kinds of constructions have been subsumed under the label ‘impersonal’ (cf. Siewierska 2008): constructions lacking a syntactic and referential subject, such as meteorological verbs; constructions with a dummy, non-referential subject such as es ‘it’ in German; constructions lacking a grammatical but not a semantic subject, such as dative experiencer constructions in many Indo-European languages; and, finally, constructions featuring non-referential grammatical subjects, such as the man-construction in German and the ‘vague they’ construction in English. In this paper, I follow the definition of impersonal constructions adopted in Siewierska (2008) and Malchukov & Ogawa (2011, 20) as “constructions lacking a referential subject”. This definition is broad enough to subsume both constructions that qualify as impersonal from a syntactic point of view (lack of agreeing subject altogether; cf. example 4a) and from a functional point of view (lack of referential subject; cf. examples 4b and 4c, where they and it are non-referential):

(4)

The sentences exemplified in (4) above all qualify as impersonal, albeit because of different criteria. In (4a) the main participant (the experiencer) is indeed highly identified and specific, but it is not a grammatical subject: therefore, the sentence is syntactically impersonal. The example sentences (4b) and (4c) are not syntactically impersonal, as they feature grammatical subjects triggering agreement on the verb (they are / it rain-.): in (4b), however, the grammatical subject is non-referential and non-specific and in (4c) the grammatical subject is a dummy, and there is no real referential argument. Therefore, (4b) and (4c) are functionally impersonal.

Malchukov & Ogawa (2011) classify impersonal constructions into three types: A-impersonals, T-impersonals and R-impersonals. A-impersonals are sensitive to reductions in the animacy/agentivity of the subject. In 4a. above, for instance, the experiencer is coded as an oblique (dative) instead of being coded as a nominative subject, because it is not agentive. T-impersonals are sensitive to a reduction in topicality: in the French sentence in (5a) the participant moqueurs ‘scoffers’ is the grammatical subject and the topic; in (5b), the role of grammatical subject is taken over by the dummy il ‘it’, while the participant moqueurs ‘scoffers’ is in focus and stripped off agreement control (cf. Malchukov & Ogawa 2011, 30).

(5)

Finally, R-impersonals are sensitive to reduction in referentiality: in (6a.) Mr. Darcy, a character from Jane Austen’s Pride and Prejudice, declares his love to Elizabeth Bennet, the protagonist of the novel, referred to in the text by the deictic you. In (6b.), instead, you has a non-referential, generic interpretation (‘anyone who would utter such a remark in Britain would not get away with it’):

(6)

In European languages, R-impersonals are expressed by a variety of constructions (Siewierska 2011, 58): pronominalized forms of the numeral ‘one’, as in English oneor Italian uno; impersonal pronouns such as German man and French on; personal pronouns used non-referentially, such as the English they in (4b) above and you in (6b) and zero- subject constructions like the obligatory ellipsis of the third person plural pronoun in Italian (7b); indefinite pronouns such as English someone or, finally, morphological impersonals such as the Romance reflexive impersonal or the German impersonal passive (8).

(7)

(8)

1.2 Impersonal constructions in discourse

What all impersonal events have in common is the demotion/ de-focusing of the agent and the consequent irrelevance of the agent’s perspective (Sansò 2003, 251). In impersonal constructions (A-impersonals, T-impersonals and R-impersonals alike), the agent (or the main participant, which can also be an experiencer) is always de-focused (Myhill 1997): morpho-syntactically, being encoded as an oblique or as zero; semantically, being encoded by a non-referential pronoun; pragmatically, being non-topical. According to Giacalone Ramat & Sansò (2007), demoted agents can have three main interpretations: species generic, where the agent represents virtually all of humanity (9a); non-referential indefinite, where the agent represents a given sub-group of humanity (‘anyone’; 9b); referential indefinite (or ‘vague’), where the agent represents a specific group of people, which is left unspecified (‘someone’; 9c).

(9)

A further development of impersonal constructions is their ability to encode specific reference: the agent is specific and defined, and even co-referential with the speaker (10; Giacalone Ramat & Sansò 2007; Siewierska 2011).

(10)

Furthermore, impersonal constructions have also developed pragmatic functions.

Impersonals can be used to express humility and politeness. In some Polynesian languages, for instance, the Proto-Austronesian pronoun *kita ‘1pl.incl’ has been reanalyzed as kita ‘1sg.incl’ and it is used to convey generic reference (‘one’; ‘anyone’) as well as to encode first person singular expressing emotional involvement of the speaker (Moyse-Faurier 2011, 600–601). In English, 2sg-imp are typically used to convey high subjectivity and the speaker’s identification with the referent (Ushie 1994, 144; see also Kitagawa & Lehrer 1990).

1.3 2sg-imp, 3-imp and participial-imp from a European cross-linguistic perspective

As seen above (Section 1.1), European languages make use of different constructions to encode R-impersonals: pronominalized forms of the numeral ‘one’, impersonal pronouns. personal pronouns used non-referentially, zero-subject constructions, indefinite pronouns and morphological impersonals. In this paper, however, I will only focus on three constructions, featuring a non-referential 2sg¹ personal pronoun, a 3(sg/pl) personal pronoun or a participial form of the verb: I label them 2sg-imp, 3-imp and participial-imp.

2sg-imp are attested in many European languages; their semantic and pragmatic uses have mostly been studied in Germanic, Slavic and Romance (Deringer et al. 2015). 3-imp are also widely attested in Europe (Siewierska 2011 and Siewierska & Papastathi 2011 for an overview). In some languages, such as Russian and Italian, 3-imp require a zero subject: overt subjects are interpreted as referential by default (cf. example 7 above). As for the label 3-imp, a note is required. Siewierska (2011) labels this construction 3pl-imp, because in all the languages she analyzes the used pronoun/bound forms are plural (English they, Italian bound verbal agreement). However, Lithuanian lacks a morphological distinction between 3sg and 3pl: gyventi‘live.inf’ – gyvena ‘live.prs.3sg/pl’. A disambiguation is only possible by means of the personal pronouns (jis.3sg.m; ji.3sg.f; jie.3pl.m; jos.3pl.f), but, crucially, impersonal uses of the third person are always pronoun-less in this language.² Therefore, for Lithuanian, 3pl-imp should be redefined as 3-imp; in this paper, I extend this label also to languages other than Lithuanian.

Participial-imp are not as common in Europe as the two other strategies. They are found in Germanic, Baltic, Slavic and Finnic (Siewierska 2008, 21): notable examples are the German passive impersonal (11a, 11b), the Polish no/to construction (13) and the Lithuanian ma/ta constructions (see Sections 2 and 3). In European languages, participial impersonals feature a form identical with or derived from a participle, usually past passive: therefore, they share morphology with passives, cf. the following German examples (11a and 11b).

(11)

The structure of (11a) and (11b) is almost identical; however, in (11b) the main verb wurden ‘became’ agrees with the subject, zwei Frauen ‘two women’, while the sentence (11a) is subjectless. In German the impersonal construction cannot have an object: if an object is present the main predicate agrees with it, and the whole construction is formally an agentless passive (12).

(12)

In the remainder of this paper, I will not take into account agentless passives. These fall into the definition of impersonal constructions by virtue of their agent-demoting function: as Malchukov & Ogawa (2011, 36ff.) notice, agentless passives have a very broad functional range, cross-cutting the domains of A-, T- and R-impersonals. However, syntactically they are personal, as the semantic patient is the syntactic subject and triggers verbal agreement (when the language has it, cf. German example 12). In participial-imp such as the Lithuanian ma/ta-imp, on the other hand, the patient does not trigger agreement and it can even be encoded as a syntactic object in the accusative case: the two constructions are therefore to be kept apart. Similarly, in Polish, the no/to impersonal construction has always an active alignment, with the object in the accusative case (13)³: thereby, in Polish impersonal and agentless passive are both morphologically and syntactically distinct.

(13)

2sg-imp, 3-imp and participial-imp differ from one another not only formally, but also semantically and functionally. First of all, they differ in their referential range, that is, the pool of possible implied agents. As far as the inclusion of speech act participants is concerned, 2sg-imp usually include the speaker, while 3-imp usually exclude them. Moreover, 2sg-imp cannot be used in a referential indefinite meaning (*you broke into my house yesterday, meaning ‘someone’). Participial-imp show more cross-linguistic variety: in Finnic, they can include the speaker, too, whereas in Polish the referential range of the no/to construction coincides with that of 3-imp (Siewierska 2008, 7; 21).

The three constructions have been shown to be different also on the stylistic and pragmatic level. In some languages, such as English or Danish, 2sg-imp and 3-imp belong to the colloquial register, while in others, such as Modern Greek, they are also used in the literary language (Sansò 2006). Pragmatically, 2sg-imp have been shown to be preferred to express empathy (Deringer et al. 2015).

2 2sg-imp, 3-imp and ma/ta-imp in Lithuanian

In Lithuanian, 2sg-imp and 3-imp are formally identical with the respective deictic forms. However, in 3-imp the pronouns jie/jos ‘they.m/f’ must be dropped (14): overt 3pl pronouns are always interpreted referentially. On the other hand, the pronoun tu ‘you.sg’, while being usually dropped, can also be retained: in (15), both the pronoun-less form prašai ‘ask for.2sg’ and the pronoun tu in tu esi ‘you are’ are found.

(14)

(15)

The participial impersonal in Lithuanian is formed by means of the non-agreeing (old neuter form) of the present (marking simultaneity) passive participle (in -ma; 16) or the past (marking anteriority) passive participle in -ta (17). They can be formed from both agentive as well as non-agentive (unaccusative) verbs:

(16)

(17)

Ma/ta participles are used to form canonical personal passives, which can be realized as agented or agentless (18a, 18b). In canonical passives, the participle in ma/ta agrees in gender and number with the subject.

(18)

Just as in the Polish no/to construction mentioned above, Lithuanian ma/ta-imp may display an active alignment, where the object of the impersonal construction is in the accusative case, the case of direct objects (19a). However, while in Polish the accusative encoding of the object of no/to impersonals is obligatory, in Lithuanian it is not. In fact, accusative objects of ma/ta-imp are very infrequent: objects appear usually in the nominative case (19b; see. the discussion in Spraunienė et al. 2015, 339–340).

(19)

The ma/ta construction has also developed a further function, namely evidential/ inferential (see Wiemer 2006; Usonienė & Šinkūnienė 2017). In their evidential function, ma/ta participles require an overt agent, which may not be human, and cannot have an overt auxiliary (20, 21): these requirements differentiate the evidential from the impersonal construction, where the agent must be covert and semantically human. It must be remarked, however, that the evidential function of ma/ta-imp is very limited in contemporary Lithuanian: “the evidential -ma/-ta construction is actually obsolete in modern Lithuanian discourse; it is occasionally used in specific contexts and it is nearly restricted to one verb, namely būti ‘be’” (Usonienė & Šinkūnienė 2017, 35).

(20)

(21)

So far, scientific scholarship about ma/ta-imp has mostly focused on their syntactic properties and on their development into an evidential construction (see inter alia Wiemer 2006; Spraunienė et al. 2015; Usonienė & Šinkūnienė 2017 and references therein). The functional (referential and pragmatic) properties of ma/ta-imp, as well as 3-imp and 2sg-imp, instead, have been discussed in lesser detail. In the remainder of this paper, I will consider only the impersonal ma/ta construction, and I will disregard the evidential/inferential one.

Table 1 shows the referential range of these constructions (Žeimantienė 2005, 2006; Geniušienė 2016; blank cells represent situations not explicitly mentioned in literature). 2sg-imp behave in Lithuanian quite similarly as in other European languages: they are used to express a state of affairs that refers to the speaker (or to the addressee) and to anyone, who is in the same situation as them. 3-imp are the rarest form of impersonal constructions; ma/ta-imp have the widest referential range.

According to Geniušienė (2016), who compares the use of 3-imp and agentless passives/ impersonal constructions (in her paper she does not distinguish between the two), 3-imp are used when the emphasis lies on the indefiniteness of the agent; the agentless passive or the impersonal construction are instead used when the emphasis lies on the event itself.

My investigation builds on the studies by Žeimantienė (2006, 2005) and Geniušienė (2016). Differently from Žeimantienė, however, I do not consider all possible Lithuanian impersonal constructions (for instance, indefinite pronouns such as kas ‘someone’, modal predicates such as galima ‘it’s possible’, the indefinite noun žmogus ‘person’), focusing solely on 2sg-imp, 3-imp and ma/ta-imp.

3 Referential, semantic and pragmatic properties of Lithuanian 2sg-imp, 3-imp and ma/ta-imp

3.1 The corpus

In order to analyze the actual use of Lithuanian 2sg-imp, 3-imp and ma/ta-imp, I have created a convenience sample of occurrences. The Lithuanian National Corpus is not morpho- syntactically annotated and, thus, it is very difficult to extract impersonal constructions. My corpus of occurrences is based on the following sources: the Lithuanian translation of Antoine de Saint-Exupéry’s novella Le Petit Prince (henceforth LPP); the Lithuanian translation of J. K. Rowling’s novel Harry Potter and the Philosopher’s stone (henceforth HP); the first 11 pages of the thread “Lietuvių kalba” on the Internet forum supermama. lt (https://www.supermama.lt/forumas/topic/692682-lietuviu-kalba) and the short story Aš mirštu, tu miršti, jis (ji) miršta by Jurga Ivanauskaitė (1989).

These sources are extremely different from each other: LPP and HP are translations, where the influence of the original may have led to linguistic choices that would have been different in an original text. I have chosen to examine them because they offer an easy way to find impersonal constructions in their Lithuanian translations. I have checked translation equivalents of French on-imp, English 3-imp (impersonal they) and 2sg-imp (impersonal you). French on has a very wide referential range (species generic, non-referential indefinite, referential indefinite, specific): therefore, different Lithuanian translation equivalents may be determined by different referential semantics, blended out in the original⁴; conveniently, on has no direct equivalent in Lithuanian, and therefore there is no risk of the original form being directly calqued into the translation. As for the English 3-imp, I have not considered instances where they anaphorically refers to an antecedent such as anybody or somebody, or to a noun referring to a person of indefinite gender (such as teacher).

The two original Lithuanian texts are also extremely different. The forum thread “Lietuvių kalba” is about the use of Russian and Polish languages in Lithuania: the main questions of the thread, “Should Russian and Polish speakers learn Lithuanian? Should Lithuanian speakers also speak Russian and Polish, if talked to in these languages?”, generated a very heated and highly emotional debate. The forum texts are written in an informal style, very similar to spoken language. They often do not follow standard orthographic and punctuation rules; sometimes they present ungrammatical constructions (perhaps due to haste in writing), not representative of standard Lithuanian. I have not edited the examples, except I added missing diacritics for ease of reading.

The short story by Ivanauskaitė (1989), on the other hand, is a dramatic piece of fiction, describing the experience of a young Lithuanian girl confronting drug addiction and depression. The stylistic register is quite formal. In this text, I only found three occurrences of referential impersonal constructions. Tables 2. and 3. offer an overview of my occurrences sample.

The tables show that 2sg-imp is the most frequent impersonal construction in both translated and original texts. 3-imp, on the other hand, is by far over-represented in the translated texts (namely, in HP), and it is very marginal in the original Lithuanian texts. This is probably the result of the influence of the original English text; it must be noted, however, that in nine cases the translator of Harry Potter chose another strategy to translate 3-imp, and in one case she chose to use a ma/ta-imp.

3.2 Referential properties of 2sg-imp, 3-imp and ma/ta-imp

As seen above, both cross-linguistic analyses as well as analyses of Lithuanian data point out to a difference in the referential range of 2sg-imp, 3-imp and ma/ta-imp. My analysis confirms these results, as shown in Tables 4a. to 4d.

In my corpus, non-referential indefinite, generic agents (‘anyone’) are always coded either by 2sg-imp or ma/ta-imp (22, 23), while referential indefinite agents (‘someone’) are coded only by 3-imp (24). In (22) and (23) the referent is generic: anyone would/ should behave in the same way in that situation (learning the country’s language, speak in Lithuanian). In (24), on the other hand, the speaker does not know who exactly is calling, but the pool of possible referents is limited to the employees of the clinic.

Non-referential indefinite agent

(22)

(23)

Referential indefinite agent

(24)

Specific reference in the 3rd person can be coded by 2sg-imp, 3-imp and ma/ta-imp alike (25, 26, 27). In (25), the scenario is fictional: however, in this scenario, the speaker knows who is the shopkeeper, who represents thus a specific, referential referent. Similarly, in (26) the referent is a specific work colleague of the speaker.

Specific reference 3sg/pl

(25)

(26)

In (27), the author is speaking about the Little Prince, who is referred to by the personal pronoun il ‘he’ in the second sentence. The impersonal construction – on in French, 2sg in Lithuanian – in the first sentence has therefore a clear third person reference.

(27)

Specific reference in the 1st person may be encoded by 2sg-imp: in (28) the impersonal on clearly refers to the author himself (also reference by the possessive mon ‘my’) and in the English translation it is translated as ‘I’.

Specific reference 1sg

(28)

Crucially, all 2sg-imp, even when they have a contextually specific reference, have the pragmatic inference that the expressed experience is common to everyone: in (27) and above, the specific referent is contextually clear, but the experience is related as universal, applicable to anyone who would be in the same situation. However, the use of 2sg-imp to express a clear specific reference may not be typical for Lithuanian, and it may be due to the influence of the source text. In a later translation of The Little Prince, (27) is translated with a first person plural (Kartais būname išsiblaškę ‘sometimes we are absent-minded’) and (28) with a first person singular (nesu piešęs ‘I have not drawn’; LPP2, 21, 106).

Sometimes, 2sg may be ambiguous between specific 2sg reference and universal, impersonal meaning. Example (29a) is taken from Harry Potter, where a character, Hagrid, addresses directly another one, Harry. Hagrid is explaining to Harry what is the Gringott, the Wizards’ Bank, and the 2sg form may be understood as being at least in part deictic and referential. However, Harry has never expressed any intention to rob the bank: Hagrid’s utterance, thus, must be understood as a generic warning and not one aimed solely at Harry. In fact, later in the dialogue, Harry repeats the question to Hagrid (29b), again in the second person, without switching to the first person, as it would be expected if (29a) would have been a purely deictic form. Therefore, I have counted the instance in (29a) in my corpus as having a generic reference.

(29)
⁵

My analysis confirms the results of previous studies: 3-imp have their semantic referential nucleus in the expression of referential indefinite agent, 2sg-imp are confined to the expression of generic and specific agents and ma/ta-imp are able to code all types of reference.

3.3 Discourse functions and pragmatic properties of 2sg-imp, 3-imp and ma/ta-imp

As shown in the foregoing sections, the three constructions under investigation have a different referential range: 2sg-imp denote generic agents, and are always inclusive of the speaker; 3-imp mostly denote referential indefinite agents and ma/ta-imp are extremely flexible, being able to express all types of reference. Their use, however, is not only determined by reference: discourse functions such as the establishment of empathy or the emphasis on the indefiniteness of the agent are relevant as well.

3.3.1 2 sg-imp

The pragmatically marked character of 2sg-imp has already been pointed out in literature. Deringer et al. (2015) claim that 2sg-imp have the fundamental function of expressing “generalized empathy that is, “empathy with the members of a class of referents over which a generalization is made, and solidarity between the speech act participants” (Deringer et al. 2015, 313). Kitagawa & Lehrer (1990, 750) state, that 2sg-imp are found in the following narrative contexts: (a) ‘situational insertion’ (‘I hit a guy who swung at me. You react instinctively at a time like that’), (b) ‘moral or truism formulation’ type (‘you kill yourself to raise your kids properly, and guess what happens’), (c) ‘life drama’ (‘you are in Egypt admiring the pyramids and feeling that you have really left your own world and time behind when suddenly you meet your next-door neighbor from home.’). In (a) you can be replaced by anyone/one and in (b) by one, with no semantic or discourse-effect changes. In (c) contexts, instead, 2sg-imp are the only strategy available in order to preserve the intended dramatic effect (Kitagawa & Lehrer 1990, 751). ‘Life drama’ situations are those, where the referent is specific, and it is identical with the speaker: either in real life, as if would be the case if the speaker had really been to Egypt and was relating on something happened to them, or in a potential, irrealis situation, in which case the speaker describes an experience to which they can relate, and expect the hearer to relate as well (‘if I/anyone were in this situation, then I/anyone would do so- and-so’; ‘simulation’ in terms of Deringer et al. 2015). In the forum texts, 2sg-imp are mostly used in opinion statements (30) and ‘life drama’ situations (31).

(30)

(31)

In (31) we have no way of knowing whether the reference to the speaker is real (they have really been in Russia and got such an answer) or simulated (the speaker imagines themselves in a possible situation they relate to emotionally). In any case, the use of 2sg-imp adds to the dramatic/comic effect of the narrative in a way that would be lost with any other strategy; the writer in the forum is using this ‘real life situation’ to make a point: in Russia, people knew about Lithuania much less than the Soviet propaganda would have liked Lithuanians to believe. The switch between the impersonal galima deklaruoti ‘it is possible to declare; one can declare’ and the 2sg-imp of the following predicates clearly marks not a switch in reference itself, but in perspective and inclusiveness: ‘one can declare’ excludes, at least for the argument’s sake, the speaker (who does not agree with what was declared), while ‘you were in some Vologda and […]’ includes them, expressing their true opinion. In (32) the speaker openly declares that the 2sg-imp statement is their opinion:

(32)

3.3.2 3-imp and ma/ta-imp

In terms of discourse functions, already Geniušienė (2016, 254) claims that in Lithuanian texts 3-imp are preferably used when the emphasis is on the indefiniteness of the agent and ma/ta-imp when the emphasis is on the action itself. Geniušienė’s claim agrees with the fundamental difference in reference between 3-imp and ma/ta-imp: the first refer to a subgroup of humanity (Sansò 2006, 268), whereas ma/ta-imp can refer to the whole of humanity, a subgroup thereof or even a specific individual, including the speaker. Therefore, 3-imp are eminently suited to encode those situations, where the speaker wants to focus on the fact that the action has been fulfilled by an unspecified ‘someone’. Ma/ta-imp are instead preferred to encode situations where the agent is completely irrelevant: either because it can be anyone or because it is specific but irrelevant to the development of the story. In (33)⁶, the linguistic competence and behavior of the person, answering in Lithuanian, is the subject of the whole text: the agent, therefore, though indefinite, is focused on. In (34), instead, the emphasis is put on the fact that, no matter what, Americans will get an answer in German; the German-speaking shopkeepers are not relevant to the subsequent narrative development. Similarly, in (35), the co-worker asking the question is irrelevant to the development of the story: what was important to the speaker was the content of the question.

(33)

(34)

(35)

3.4 Deictic meaning and discourse function

The original deictic meaning of the constructions highly correlates with their functions (Sansò 2006; Siewierska 2011; Kitagawa & Lehrer 1990). Ma/ta-imp with no overt agent and a formally impersonal form are the best candidates for encoding events, where the agent is either completely irrelevant or, though being relevant, should not overshadow the event itself. Similarly, 2sg-imp, having in their non-generic, deictic use a reference to a speech act participant, are the best candidates for encoding specific inclusive reference (1sg). 2sg-imp are therefore also quite apt to express agentless generic events, where the agent is generic (and implied to be ‘all of humanity’). Its empathy-generating pragmatics can be seen as a result of the implication ‘everyone is the same as me and you’: the speaker’s experience is presented as universal, thereby establishing empathy with the adressee/reader. The empathy effect is strenghtened by the fact that 2sg-imp directly address the adressee/reader, making the statement ‘personal’. Finally, 3-imp have the narrowest functional range, because they cannot express any other person than their original personal function (third): as they are intrinsically exclusive, 3-imp are bad candidates to express reference to ‘anyone’, as this necessarily also includes the speaker (Sansò 2006).

4 Conclusions

Despite the relatively small number of analyzed occurrences, the results of the analysis presented in the foregoing sections depict a clear picture of the semantics and functions of Lithuanian 2sg-imp, 3-imp and ma/ta-imp, and show that the referential properties and pragmatic-discourse functions of these constructions are in line with those already described for other European languages.

As for their referential properties, ma/ta-imp are quite flexible, being able to encode all types of reference (generic, referential indefinite and specific); 3-imp and 2sg-imp have a more restricted referential range, the first being restricted to the encoding of referential indefinite agents and specific 3rd person agents, and the second to the encoding of generic and specific agents. Having a wide referential range, ma/ta-imp may replace both 2sg-imp and 3-imp. However, these constructions differ greatly on the pragmatic- discourse level, and are therefore not always freely interchangeable (cf. also Geniušienė 2016). 2sg-imp are pragmatically marked, expressing empathy. They are used to express a strong emotional identification: 2sg-imp are extremely frequent in the forum texts and in The Little Prince, which are both text characterized by a high level of emotionality.

In discourse, 3-imp are preferred to ma/ta-imp whenever the indefiniteness of the agent is focused on; ma/ta-imp are instead preferred whenever the focus is on the event itself.

As for their distribution, ma/ta-imp and 2sg-imp are by far more frequent than 3-imp. Siewierska (2011) puts Lithuanian in the group of languages that “virtually lack 3(pl)- imp”: while it is true that this strategy is the least frequent, it is nevertheless present even in my small sample, in all four sources (in the translation of Harry Potter, 3-imp are more frequent than in the original Lithuanian texts, probably due to the influence of the original text). 2sg are more frequent in stylistically informal texts – it is by far the predominant strategy in the forum texts, while it is completely absent in the more formal language of Ivanauskaitė’s short story. Ma/ta-imp are stylistically neutral, and they are found in both original Lithuanian sources.

Acknowledgments

I would like to acknowledge the precious advice I got from Anastasia Bauer, Daniel Bunčić, Maria Katarzyna Prenner, Jogilė Teresa Ramonaitė and Björn Wiemer on different topics discussed in this paper. A thank you goes to Thomas Stolz and colleagues at the University of Bremen for letting me use their corpus of The Little Prince translations. Also, I would like to thank Aurelija Usonienė and two anonymous reviewers for their helpful suggestions and corrections. All remaining errors and shortcomings are my responsability only.

References

Bahr, Robert. 2013. Die Dublonen des Samiel. Published on-line: http://www.lulu.com/shop/robert-bahr/die-dublonen-des-samiel/paperback/product-21377282.html.

Creissels, Daniel. 2009. Impersonal pronouns and coreference: The case of French on. Ms, University of Lyon.

Deringer, Lisa, Volker Gast & Florian Haas. 2015. Impersonal uses of the second person singular and generalized empathy: An exploratory corpus study of English, German and Russian. In The Pragmatics of Personal Pronouns. Sandrine Sorlin & Laure Gardelle, eds. Amsterdam: Benjamins. 311–334. https://doi.org/10.1075/ slcs.171.15der.

DW: Projekt Deutscher Wortschatz. http://wortschatz.uni-leipzig.de.

Geniušienė, Emma. 2016 [1973]. The relation between the indefinite personal and the passive in Lithuanian. In Passive constructions in Lithuanian. Selected works of Emma Geniušienė. Anna Kibort & Nijolė Maskaliūnienė, eds. Amsterdam: John Benjamins. 247–268.

Giacalone Ramat, Anna & Andrea Sansò. 2007. The spread and decline of indefinite man-constructions in European languages. In Europe and the Mediterranean as Linguistic Areas. Paolo Ramat & Elisa Roma, eds. Amsterdam/Philadelphia: John Benjamins. 95–131. https://doi.org/10.1075/slcs.88.07gia.

Haspelmath, Martin. 2001. Non-canonical marking of core arguments in European languages. In Non-Canonical Marking of Subjects and Objects. Alexandra Y. Aikhenvald, R. M. W. Dixon & Masayuki Onishi, eds. Amsterdam/Philadelphia: John Benjamins. 53–83.

Holvoet, Axel. 2001. Impersonals and passives in Baltic and Finnic. In Circum-Baltic languages. Volume 2: Grammar and Typology. Östen Dahl & Maria Koptjevskaja- Tamm, eds. Amsterdam/Philadelphia: John Benjamins. 363–389. https://doi.org/10.1075/slcs.55.04hol

Kibort, Anna. 2008. Impersonals in Polish: An LFG perspective. Transactions of the Philological Society 106 (2), 246–289.

Kitagawa, Chisato & Adrienne Lehrer. 1990. Impersonal uses of personal pronouns. Journal of Pragmatics 14, 739–759. https://doi.org/10.1016/0378-2166(90)90004-w

LBA: Bovet, Félix & Louis Bonnet. 2012 [1900]. La Bible (Traduction de la Bible annotée). ThéoTeX. Published electronically: www.theotex.org.

LSG: Segond, Louis. 1910. La Bible. Traduction de Louis Segond. Geneva: Alliance Biblique Universelle.

Malamud, Sophia. 2012. Impersonal indexicals: one, you, man, and du. The Journal of Comparative Germanic Linguistics 15 (1), 1–48. https://doi.org/10.1007/s10828-012-9047-6.

Malchukov Andrej L. & Akio Ogawa. 2011. Towards a typology of impersonal constructions: A semantic map approach. In Impersonal constructions. A cross- linguistic perspective. Andrej Malchukov & Anna Siewierska, eds. Amsterdam: John Benjamins. 17–54. https://doi.org/10.1075/slcs.124.02mal.

Moyse-Faurier, Claire. 2011. Impersonal constructions in some Oceanic languages. In Impersonal constructions. A cross-linguistic perspective. Andrej Malchukov & Anna Siewierska, eds. Amsterdam: John Benjamins. 581–606. https://doi.org/10.1075/slcs.124.20moy.

Myhill, John. 1997. Towards a functional typology of agent defocusing. Linguistics 35, 799–844.

Sansò, Andrea. 2003. The network of demotion. Towards a unified account of passive constructions. In Meanings Through Language Contrast, Vol.1. Ken Turner & Katarzyna Jaszczolt, eds. Amsterdam: John Benjamins. 245–260. https://doi.org/10.1075/pbns.99.20san.

Sansò, Andrea. 2006. ‘Agent defocusing’ revisited. Passive and impersonal constructions in some European languages. In Passivization and typology. Form and Function. Werner Abraham & Larisa Leisiö, eds. Amsterdam: John Benjamins. 232–273. https://doi.org/10.1075/tsl.68.15san.

Siewierska, Anna. 2008. Ways of impersonalizing. In Current Trends in Contrastive Linguistics. Maria delos Angeles Gomez Gonzales, Lachlan Mackenzie & Vladimir Panov. Juk and gi, and “particles” in contemporary Lithuanian Elsa Gonzalez Alvarez, eds. Amsterdam: John Benjamins. 3–26. https://doi.org/10.1177/1461445609342501

Siewierska, Anna. 2011. Overlap and complementarity in reference impersonals. In Impersonal constructions. A cross-linguistic perspective. Andrej Malchukov & Anna Siewierska, eds. Amsterdam: John Benjamins. 57–89. https://doi.org/10.1075/slcs.124.03sie

Siewierska, Anna & Maria Papastathi. 2011. Towards a typology of third person plural impersonals. Linguistics 49(3), 575–610. https://doi.org/10.1515/ling.2011.018.

Spraunienė, Birutė, Auksė Razanovaitė & Erika Jasionytė. 2015. Solving the puzzle of the Lithuanian passive. In Voice and Argument Structure in Baltic. Axel Holvoet & Nicole Nau, eds. Berlin: John Benjamins. 323–365. https://doi.org/10.1075/vargreb.2.07spr.

Ushie, Yukiko. 1994. Who Are You? And What Are You Doing? Discourse and Pragmatic Functions of the Personal Pronoun You in Conversational Narratives.https://www.researchgate.net/publication/281119326_Who_Are_You_And_What_Are_ You_Doing--Discourse_and_Pragmatic_Functions_of_the_Personal_Pronoun_ You_in_Conversational_Narratives

Usonienė, Aurelija & Jolanta Šinkūnienė. 2017. Potential vs use: revisiting an evidential participial construction in Lithuanian. In Evidentiality Revisited: Cognitive Grammar, Functional and Discourse-Pragmatic Perspectives. Juana I. Marín- Arrese, Gerda Haβler & Marta Carretero, eds. Amsterdam/Philadelphia: John Benjamins. 297–340. https://doi.org/10.1075/pbns.271.08uso.

Wiemer, Björn. 2006. Relations between Actor-demoting devices in Lithuanian. In Passivization and Typology: Form and function. Werner Abraham & Larisa Leisiö, eds. Berlin/Amsterdam: Benjamins. 274–309. https://doi.org/10.1075/tsl.68.16wie

Zifonun, Gisela. 2001. ‚Man lebt nur einmal‘. Morphosyntax und Semantik des Pronomens man. Deutsche Sprache 3, 232–253.

Žeimantienė, Vaiva. 2005. Einzelsprachliche Realisierungen des Subjekt-Impersonals: Das Beispiel deutscher man-Sätze und ihrer Entsprechungen im Litauischen. Kalbotyra 55 (3), 81–90.

Žeimantienė, Vaiva. 2006. Deutsche subjektlose Passivkonstruktionen und „man-Sätze“ im Vergleich zu ihren Entsprechungen im Litauischen. In Menschen – Sprachen – Kulturen. Lucyna Wille & Jaromin Homa, eds. Marburg: Tectum Verlag, 379–387.

Primary sources Lithuanian original texts

Ivanauskaitė, Jurga. 1989. Aš mirštu, tu miršti, jis (ji) miršta. In Kaip užsiauginti baimę (novelių ir apsakymų rinkinys). Vilnius: Vaga. Forum: https://www.supermama.lt/forumas/topic/692682-lietuviu-kalba [accessed on 12.11.2019]

Glosario

Abbreviations

1, 2, 3: first, second, third

acc: accusative

COND: conditional

DAT: dative

DEF: definite

DIM: diminutive

F: feminine

FUT: future

GEN: genitive

HAB: habitual

IMP: impersonal

IMPER: imperative

INDEF: indefinite

INF: infinitive

INS: instrumental

LOC: locative

M: masculine

N: neuter

NEG: negation

NOM: nominative

PAP: past active participle

PFV: perfective

PL: plural

PP: past participle

PPP: past passive participle

PRAP: present active participle

PRPP: present passive participle

PRS: present

PST: past

PSGER: past gerundive

SG: singular

SUBJ: subjunctive

SUPER: superlative

Translations

HP English:: Rowling, Joanne K. 1997. Harry Potter and the Philosopher’s Stone.London: Bloomsbury

HP Lithuanian:: Rowling, Joanne K. 2000. Haris Poteris ir isminties akmuo. Vilnius: Alma Litera.

LPP English:: Saint-Exupéry, Antoine de. 1996 The Little Prince. London: Heinemann

LPP French:: Saint-Exupéry, Antoine de. 1943. Le Petit Prince. Paris: Gallimard

LPP Lithuanian:: Saint-Exupéry, Antoine de. 1995. Mažasis Princas. Vilnius: Džiugas.

LPP2 Lithuanian:: Saint-Exupéry, Antoine de. 2011. Mažasis Princas. Vilnius: Alma littera.

Notes