MARKUS RODUNER

Institute of the Lithuanian Language

Research interests: Lithuanian syntax, Baltic syntax, historical syntax, areal linguistics, translation studies.

AMBIPERSONAL – IS THIS CATEGORY ALSO APPLICABLE TO THE BALTIC LANGUAGES?

Ambipersonal: A Suitable Category for the Baltic Languages too?

ABSTRACT

This article examines how the category of ambipersonal is realized in Lithuanian and—briefly—in Latvian by constructions with passive participles. Since ambipersonal can be expressed by various means even within one and the same language, the focus here is on constructions with passive participles. It is precisely in ambipersonal expressed by constructions with passive participles that a certain degree of grammaticalization can be discerned. The relationship between the passive in these languages and the aforementioned category is also discussed.

ANNOTATION

This paper discusses the realisation of the ambipersonal using participle forms, traditionally called passive participles, in Lithuanian and – to a much lesser extent – Latvian. As different means can be used to express the category of the ambibersonal, we concentrate in this paper on constructions with passive participles as a mean of expressing the ambipersonal. This decision was made, as these constructions with neuter passive participles show a high degree of grammaticalisation. We further discuss the connection between the passive and the ambipersonal.

INTRODUCTION

The term ambipersonal was coined by the Finnish linguist Hannu Tommola (Tommola 1993), who introduced it while examining issues of diathesis (grammatical voice) in Finnish and Estonian, cf. (1), (2).

KEYWORDS: ambipersonal, passive, grammatical voice, Baltic languages, areal linguistics.

KEYWORDS: ambipersonal, passive, voice, Baltic languages, areal linguistics.

106

Acta Linguistica Lithuanica LXIV–LXV

Ambipersonal as – ar ši kategorija tinka ir baltų kalboms?

(1) Finnish. Tanskassa puhutaan tanskaa. Denmark. LOC.¹ speak. AMB. PRAES. Danish. Danish is spoken in Denmark (Karlsson 2000: 190).

(2) Estonian. Tehti ära suur töö. work. AMB. PRAET. much large work. NOM. A great deal of work was done (Tommola 1997: 3).

These examples clearly illustrate the problem of interpreting verb voice in Baltic Finnic languages. In older grammars of Baltic Finnic languages, the category called passive is more reminiscent of the impersonal passive in German (or other languages) (German unpersönliches Passiv² and can be formed from almost all types of verbs, even stative and unaccusative ones³ (cf. Lithuanian augama, krentama, užmiegama). Similar constructions are found in many other languages too, but Baltic Finnic languages have no grammatical means of expressing the agent⁴ and, depending on the verb’s valency, only a two-participant or one-participant “passive” is possible. In those languages, ambipersonal is grammaticalized.

Wolfgang Schlachter, one of the first to examine in detail the problem of Finnish passive, begins his work with the telling title Hat das Finnische ein Passiv (“Does Finnish have a passive?”, Schlachter: 1985) by citing a claim by Oswald Szemerényi (1970) that applies to Baltic Finnic languages, but not only to them: it is particularly important to distinguish a two-participant passive construction from a three-participant one, because the three-participant construction (with an expressed agent) is merely a transformation of the corresponding active sentence (thus only a syntactic variant), whereas the two-participant construction (without an expressed agent) is necessary because the absence of an agent may carry informational value (an indefinite agent, deliberately un-

¹ Abbreviations used in the article: ACC. – accusative, LOC. – locative, NOM. – nominative, GEN. – genitive, AMB. – ambipersonal, PRAES. – present tense, PRAET. – past (simple) tense, ART. – article, PAG. – auxiliary verb, PPP. – past passive participle, PPA. – past active participle, languages: Irish – Irish, Est. – Estonian, Lat. – Latvian, Lith. – Lithuanian, Fr. – French, Finn. – Finnish, Ger. – German, Welsh – Welsh.

² Plg. Tommola (1997): „в структурном отношении данное явление рассматривалось традиционно как особый случай пассива: „имперсональный“ пассив“.

³ In many languages there is no passive of such verbs, because the subject of the active sentence is not an agent but most often an experiencer.

⁴ It is true that this option exists in Estonian, and the third participant can be expressed by a postpositional construction pool(t) ‘from, on the part of [...]’ (cf. Tauli 1983: 32). In their article, Manninen and Hiietam (2005) also give examples from both Estonian (with the aforementioned postposition poolt) and Finnish (with the preposition toimesta ‘at someone’s direction’) in which the agent can be expressed even with synthetic (“true”) ambipersonal forms, or where the sentence semantics compel this interpretation. This phenomenon should be further investigated using a larger corpus in order to determine how widespread it already is in these languages and whether it is continuing to spread and giving rise to a full-fledged passive.

Straipsniai / Articles

107

M A R K U S R O D U N E R

vardijamas veikėjas ir kt.). Fredo Karlssono suomių kalbos gramatikoje (Karlsson 2000: 187) rašoma, kad šios kalbos pasyvą geriau būtų vadinti „indefinitu“. Kaip matome, nė vienas šių kalbininkų, aprašydamas Baltijos finų kalbų neveikiamąją rūšį, nebuvo linkęs ją sutapatinti su žinomesnių Vakarų Europos⁵ kalbų pasyvu.

In linguistics, ambipersonal sentences such as (1) have traditionally been called impersonal passives, because the agent / experiencer is not overtly expressed. However, this is problematic because they can be formed from all types of verbs. Describing them as impersonals (cf. Holvoet 2001, Blevins 2003, and others) is problematic from a semantic point of view, because the agent / experiencer is unexpressed only on the surface of the sentence, but is understood from context as ‘animate’ or even ‘human’—in example (1), these are the inhabitants of Denmark. Therefore, the term ambipersonal (with an ambivalent agent / experiencer referent) would better fit this phenomenon. Moreover—as discussed in the next paragraph—this allows the expression of ambipersonal to be compared across the world’s languages using a range of means, not only verb forms.⁶

Pats Tommola (1997) diskusijoje dėl ambipersonalo kategorijos savo pirmtaku įvardija Igorį Melčuką, pasiūliusį „diatezes, kuriomis supresuojami aktantai, pavadinti „supresyvu“ (Mel’čuk: 1991: 65). Ir Tommola vartoja sąvoką supresyvas (plg. jo 1993 m. straipsnio pavadinimą), tačiau taip vadina supresyvizacijos, sintaksinės operacijos, kuria supresuojamas aktantas, rezultatą, o ambipersonalu – semantinė-pragmatinė kategorija, kuri gali būti reiškiama įvairiomis konstrukcijomis, pvz., sugramatinama atskiromis veiksmažodžio formomis kaip suomių ir estų kalbose. Ambipersonalo apibūdinimas kaip semantinės-pragmatinės kategorijos – kaip teiginių, kuriuose veikėjas neįvardijamas (supresuojamas), tačiau referentiškai identifikuojamas arba konkrečiai, arba kaip generinis – leidžia Tommolai (1997) be sugramatinto ambipersonalo Baltijos finų kalbose aptarti įvairių Europos kalbų strategijas ambipersonalui reikšti, tarp jų antrojo vienaskaitos veiksmažodžio formą (pvz., rusų kalboje), atskirus ambipersonalinius įvardžius (pvz., vok. man, pranc. on), sangrąžines veiksmažodžio formas (pvz., italų, švedų kalbose) ir kt. Taigi ambipersonalo raiškos priemonių esama gana įvairių, veiksmažodžių rūšies formų atžvilgiu jie taip pat nevienareikšmiai – vartojamos ir pasyvinės, ir sangrąžinės, ir aktyvinės veiksmažodžio formos.

This article will focus on the use of passive forms to express ambipersonal in the Baltic languages, primarily Lithuanian.

⁵ “More familiar” because, at the other edge of Europe, one of the “passive” forms in Celtic languages also has the function described here; cf. Welsh siaredir (speak AMBP. PRAES.) Cymraeg (‘Welsh’) yma (‘here’)—‘Welsh is spoken here’—or Irish Labhraítear (‘speak’ AMBP. PRAES.) Gaeilge anseo—‘Irish is spoken here’.

⁶ A similar case, in terms of diversity of expression, is the category of evidentiality.

108

A c t a L i n g u i s t i c a L i t h u a n i c a L X I V – L X V

Ambipersonalas – ar ši kategorija tinka ir baltų kalboms?

THE AMBIPERSONAL IN LITHUANIAN

Why does this paper begin its argument with the Finnic languages neighboring the Baltic languages? Because the interpretation of the passive voice in the Baltic languages as a full-fledged passive is almost as problematic as it is in the Baltic Finnic languages.

Following the authors mentioned above, this article uses the term full-fledged passive only for the three-place passive, that is, one in which the direct object of an active sentence is promoted to subject position, while the agent (of the active sentence) is expressed by a prepositional construction or an oblique case—the kind found in many (modern) Western European languages:

(3a) German Der Mann schlägt ART. MASC. NOM. man beat. PRAES. 3. SG. den Hund ART. MASC. ACC. dog. ACC. SG. The man beats the dog.

(3b) German Der Hund wird vom ART. MASC. NOM. dog. NOM. SG. PAG. PRAES. 3. SG. by Mann geschlagen man beat. PPP. The dog is beaten by the man.

Although the statements the dog is beaten by the man and the dog is the man’s beaten are both grammatical in Lithuanian and comply with all the rules of grammar, they are not very natural⁷. For example, of the 202 examples in DLKT of the typical accusative-taking verb mušti ‘beat’, often cited to illustrate the passive, only 12 examples of the present passive participle⁸ mušamas in DLKT occur with a genitive agent. Only one of these could be described as a complete passive sentence (example 4):

(4) Behold, Maleckis, although a professor, is himself being beaten by the authorities!

In the other examples, the passive participle is used attributively (3), or the agent is not specified (4).

⁷ Plg. Ambrazas 2006: 492: „Subjekto (agentinis) kilmininkas tose konstrukcijose [pasyvo su neveikiamųjų dalyvių vardininko derinamomis formomis – M. R.] vartojamas retai.“

⁸ In this article, most of the argument will be based on examples with present passive participles, since constructions with past passive participles are often also used to express other categories—for example, resultative meaning (cf. Geniušienė 1988) or evidentiality (cf. Holvoet, Semėnienė 2004).

Straipsniai / Articles

109

M A R K U S R O D U N E R

(5) Even today, in almost every daily newspaper one can read about young brides beaten by their husbands, mistreated by their mothers-in-law, or driven to suicide by unbearable lives—and sometimes murdered.

(6) The person was beaten and strangled, while his home was ransacked in search of money.

Sentences of the latter type—without an agent expressed in the genitive case—are fairly common in contemporary standard language. Although some features, such as the participle agreeing with an element expressed in the nominative (the experiencer/patient), indicate that the direct object is promoted to subject position, the rare expression of the agent and the infrequent use of the personal passive point to its connection with the ambipersonal.

Unlike a fully personal passive, statement (6) has no true active counterpart. Its “active” counterpart they beat and strangled the person and ransacked his home, searching for money also leaves the agent unexpressed (they). The agent can be identified only from the context, just as can the agent in the “passive” sentence (5). However, this by no means implies that it is impersonal, since, as in example (1), the participant here is a living being identifiable from the context—in this case, criminals acting very actively indeed. Thus, both statements are good examples of the ambipersonal category and show that it can be expressed by various linguistic means.

Tommola (1997) argumentuoja, kad ambipersonalo funkcija yra kitokia nei triasmenio pasyvo. Pastarajame tik keičiama aktualioji skaida, t. y. aktyvinio sakinio tiesioginis papildinys yra iškeliamas į veiksnio poziciją. Dėl to pasyvas būdingas kalboms, kuriose jis yra priemonė aktyvinio sakinio tiesioginiam papildiniui tematizuoti. O baltų ir Baltijos finų kalbose tam užtenka argumentus sukeisti vietomis (Gediminas įkūrė Vilnių – Vilnių įkūrė Gediminas). Björnas Wiemeris (2001: 9) teigia, kad tokių sakinių kaip Jokios klaidos nepastebėta arba Čia vakar buvo valgoma priskyrimas pasyvo kategorijai yra problemiškas, kadangi trūksta privilegijuoto sintaksinio argumento (angl. privileged syntactic argument). Lyginant teiginį Jokios klaidos nepastebėta su teiginiu Dvi klaidos pastebėtos aiškiai matyti, kad ir pastarojo sakinio funkcija yra ambipersonalinė. Abejuose teiginiuose veikėjas numanomas, tik supresuojamas.

One characteristic of the personal passive is that, because the direct object is promoted, it can be formed only from transitive verbs. Sentences with intransitive verbs cannot be passivized. Lithuanian constructions with passive-voice participles of intransitive verbs seemingly contradict this claim; compare (7):

(7) The expedition members spent four days in the Lukina Jama cave. They slept in base camps set up in the cave at depths of 750 m and 950 m (DLKT).

110

Acta Linguistica Lithuanica LXIV–LXV

Ambipersonal as – ar ši kategorija tinka ir baltų kalboms?

Given its valency, it would be impossible to form a personal passive sentence with the verb miegoti (‘to sleep’). Nor is there any possibility of making the participle in the predicate agree with anything, since there is nothing to promote to subject position.

However, verbs such as miegoti (‘to sleep’) or kosėti (‘to cough’), although intransitive, can be “passivized” as ambipersonals (impersonal passives) in some languages, for example German; compare (8):

(8) es wurde auf dem Boden geschlafen

PAG. PRAET. 3. SG. on ART. DAT. M. floor sleep. PPP.

was slept on the floor / we slept on the floor

In German, a construction with the ambipersonal pronoun man can also be used synonymously, for example man schlief (sleep PRAET. 3. SG.) auf dem Boden. In both statements, the sleepers are simply not specified, but are understood to be the expedition members—or we, if the account is given by an expedition member.

The subject of such verbs is active (an agent), and the verb is called unergative. However, in the same language, German, sentences with unaccusative verbs such as fallen (‘to fall’) or ersticken (‘to suffocate’) cannot be “passivized.” For example, *es wurde in den Abgrund gefallen (“there was falling into the abyss”) is ungrammatical, whereas it is entirely grammatical in Lithuanian, although perhaps not very frequent (the LKET query krentama returns only 6 examples). Ambipersonal forms of intransitive verbs are clearly identifiable: they are the neuter forms of passive-voice participles.

As we have seen, the personal passive exists in standard Lithuanian, but is rarely used. In many dialects, sentences⁹ with indeclinable participles are widespread and resemble the Baltic Finnic ambipersonal even more, since the participle does not agree with a nominative element:

(9) fields is being plowed (Ambrazas 2001: 407)

(10) well, the monument is built (Krokšlys, Varėna district, fieldwork recording)

In such sentences, the direct object is marked with the nominative case but is not raised to subject position, and the predicate participle does not agree with it. Nominative marking does not necessarily entail raising to subject position, because in these dialects the nominative is also used to mark direct objects in sentences without an overt subject (cf. Ambrazas 2001 and the literature cited there). In this case, the absence of agreement is not problematic either, since the nominative marks not a grammatical

⁹ They also occur in texts, but, like nominative objects, more often in prewar texts. Among more than 15,000 examples from DLKT with neuter passive participles, collected for the dissertation, only one sentence with the structure mentioned here was found.

Straipsniai / Articles

111

MARKUS RODUNER

subject. This is also confirmed by the few dialectal examples in which the experiencer is marked not by the nominative but by the accusative, cf. su geiniu iš dreivių medų kopama (Ambrazas 1990: 205)¹⁰.

No examples were found in constructions with neuter passive participles involving first- or second-person pronouns (in any or all three numbers), or sentences with direct objects in the accusative—apparently, in that case only an ambipersonal construction with an agreeing participle is possible, cf. *mane buvo mušama – aš buvau mušamas¹¹. This is interesting because in Finnish these very pronouns are expressed in the accusative in ambipersonal sentences, while nouns are in the nominative.

Besides the means already mentioned, ambipersonal constructions in Lithuanian—as, for example, in Russian—can also be expressed using the second-person singular (sometimes with the additional word žmogus).

(11) Russian Когда влюблен, не замечаешь When be_in_love PPP. NOM. SG. M. not notice PRAES. IND. 2. SG. недостатков любимого человека. Shortcoming GEN. PL. beloved GEN. SG. M. person GEN. SG. When you are in love, you do not notice the shortcomings of the person you love.

The generic use of the second person is also confirmed by examples such as German man (pron. AMB.) ist (‘be’ PRAES. 3. SG.), was (‘what’ ACC.) man isst (‘eat’ PRAES. 3. SG.), which, like its Lithuanian equivalent esi, ką valgai, can also be expressed in the second person: du bist, was du isst.

AMBIPERSONAL IN LATVIAN

In Latvian, passive forms are formed with the auxiliary verbs tikt (less often also tapt, kļūt, in eventive passives) or būt ‘be’ (in stative passives), and –ts– participles. As Nicole Nau (1997: 37; see also Holvoet 2001: 370) points out, almost all verbs can be passivized¹².

¹⁰ Such constructions with past passive neuter participles and an accusative object are also quite common in some Slavic languages (cf. Wiemer, Giger 2005)—they are resultatives.

¹¹ In the Irish ambipersonal construction, too, the object of an active sentence is not raised to subject position, cf. Maraíodh (‘kill’ AMB. PRAET.) (léi) (‘with her’) é (‘he’ ACC!): literally “was killed (with her) him”—“he was killed by her, she killed him.”

¹² Axel Holvoet’s (ibid.) restriction—that, unlike in Finnish, the copula būti cannot be passivized/ambipersonalized in Latvian—also applies to Lithuanian: the sentence *būta gudraus / gudriam is ungrammatical; it becomes grammatical only when an agent is also specified with the copula: jo esama

112

Acta Linguistica Lithuanica LXIV–LXV

Ambipersonalas – ar ši kategorija tinka ir baltų kalboms?

(12) There were read (PAG. PRAET. 3P.) in front (‘in front’) all the burlak novels then in existence, and witches were talked about (‘talk’ PPP. NOM. SG. M.). All the burlak novels in existence at the time were read, and witches were talked about (Nau 1997: 38).

(13) So tada ‘then’ was divided (PAG. PRAET. ‘divide’ PPP. NOM. SG. M.). Then it was divided.

Nau maintains that the ambipersonal (passive) is not frequently used in spoken language, especially with first- or second-person agents (1997: 38), while Sturla Berg-Olsen argues that the use of the passive in Latvian is more restricted than in languages with fixed word order (2001: 220). However, it should be noted that the function of the ambipersonal here, too, is to mark the agent as indefinite—and animate, often I or we; cf. (13). This could explain a fact mentioned by Nau in the same place: that a sentence with a “passive” can be combined with an impersonal sentence containing a third-person plural verb; cf. Vecākais brālis tika ievainots un viņu paņēma vācu gūstā—“The eldest brother was wounded and they took him prisoner.” According to the category set out here, both sentences are ambipersonal and can therefore be combined.

In terms of the fact that a verb’s transitivity is almost irrelevant to its “passivization,” Latvian is very similar not only to Lithuanian but also to the neighboring Finnic languages.

A feature Latvian shares with Finnic languages, but not with Lithuanian, is that in standard Latvian, as in many dialects, it is not possible to express the agent in an “action passive” formed with the auxiliary verb tikt (cf. Berg-Olsen 2001: 219, Nau 1997: 38, Holvoet 2001: 371). However, Berg-Olsen (ibid.), citing Endzelīns (1951: 984), gives examples from older Latvian texts in which the agent is expressed—once in the genitive case (as in Lithuanian), and once with the preposition no (‘from’):

(14) Balti man vaidzini ne ūdeņa balināti [...] asarām balināti no nelieša My white cheeks are not bleached by water [...] bleached by tears, by the enemy

Tačiau šie pavyzdžiai abu yra ne veiksmo pasyvai. Alfredas Gātersas (1993: 295–97) pažymi, kad latvių dainose asmeninis veikėjas beveik niekada neįvardijamas, ir pabrėžia konstrukcijų su –ts– dalyviais artumą konstrukcijoms su predikatyvais.

cunning. Another parallel is Irish, where almost all verbs can also be ambipersonalized, except in sentences with the copula is. “Third, with the exception of sentences formed with the copular verb is, all Irish sentences have an impersonal counterpart: this includes intransitives, copular sentences formed with the copular verb bí, and even passives, together with the expected transitive sentences; only transitive clauses have passive counterparts” (Noonan 1994: 5).

Straipsniai / Articles

113

MARKUS RODUNER

Holvoetas (2001: 332–335) atkreipia dėmesį į faktą, kad latvių kalboje šalia 3 daugiskaitos asmens ambipersonalui reikšti vartojamas ir vienaskaitos 3 asmuo¹³. Tame pačiame straipsnyje įžvelgiamas ir skirtumas tarp abiejų 3 asmens formų vartojimo – daugiskaita vartojama nereferentiniam, generiniam veikėjui išreikšti, vienaskaita referentiniam¹⁴. Tame pačiame straipsnyje nurodoma dar kitų į pasyvą panašių konstrukcijų, kurios yra bendros latvių kalbai ir kaimyninėms finų kalboms.

The stative passive, like Lithuanian constructions with predicative ta-participles, will not be discussed here because of their links with resultative and predicative constructions¹⁵.

THE EXPRESSION OF AMBIPERSONAL IN SOME OTHER EUROPEAN LANGUAGES

As mentioned in footnote 3, at the other end of Europe, in the west, the Celtic languages also have grammaticalized ambipersonal forms. Modern Irish even distinguishes two “passives”—the ambipersonal and the “true” passive (cf. Noonan 1994):

(12) Rinneadh a faire agus a hadhlacadh. do AMB. PRAET. his wake and his burial his wake and burial were organized / someone organized his wake and burial¹⁶ (Nolan 2005: 3)

(13) Bhí sé á bhualadh aici¹⁷ was he at+his beating-VN by+her He was being beaten by her (Noonan 1994: 6)

It is clear that in Irish—as in the other Celtic languages—ambipersonal is grammaticalized, and there are separate means of expression for the passive. The latter—for expressing the passive

¹³ Since finite third-person verb forms in the Baltic languages do not differ in number, he relied on verb forms in which such a distinction exists, for example participles.

¹⁴ In this context, it would be interesting to investigate whether the use of the second-person singular and third-person plural ambipersonal in Lithuanian differs semantically in a similar way.

¹⁵ There is considerable literature on this topic, as well as on referring to the agent with the genitive or other cases/constructions; for the agent genitive, for example, see Ambrazas 1990, Holvoet 2000, and the literature cited there.

¹⁶ “Someone organised his burial and wake. [do’(x) CAUSE [make’(x, a faire agus a hadhlacadh)]], where x is an unspecified individual of animate and human type, and interpreted as ‘(some)one’, that is, a specific (because they exist) indefinite person.“

¹⁷ This statement, made without naming the agent, is ambiguous: He was being beaten or He beat him.

114

Acta Linguistica Lithuanica LXIV–LXV

Ambipersonalac – ar ši kategorija tinka ir baltų kalboms?

to express – in the related Welsh language, as in, for example, Latvian or Estonian, there are several different constructions (cf. Skrzypiec 2000). And although the construction is ambipersonal, there are nevertheless means of expressing the agent suppressed in it (and this is reminiscent not only of the Baltic but also of the Germanic languages).

Ambipersonal as a semantic-pragmatic category is expressed in one way or another in many European languages, for example German or French, by a separate pronoun – German man, French on¹⁸. However, it should be emphasized that the domain of use of the ambipersonal category is not identical in these languages. This can give rise to difficulties when translating from one language into another. The author of this article has heard advanced German-language students produce sentences such as man hat in Mannheim studiert, a literal translation of the Lithuanian sentence studijuota Manheime: the ambipersonal form studijuota in Lithuanian is translated by the corresponding ambipersonal pronoun man. However, unlike in Lithuanian, the pronoun man cannot be used in German in this case, since a German speaker will not understand that it can also mean I¹⁹.

This very feature of Lithuanian once again has a counterpart in Finnish, where the pronoun me (‘we’) can already be used colloquially with an ambipersonal form, replacing the “true” first-person plural form in the paradigm (Christen 1998: 52). However, in none of the Baltic languages is such a combination—the first-person plural pronoun + an ambipersonal verb form (*mes einama)—used. The generalized use of the first-person plural pronoun is readily understandable. Like the third-person plural, this pronoun can refer to an indefinite group of people or to everyone. The use of the ambipersonal instead of the first-person singular is surprising at first glance. However, Masayoshi Shibatani (cf. scheme 1998: 97) notes that the use of ambipersonal voice with passive morphology is restricted in many languages according to the animacy hierarchy. For example, in Nepali it is permitted only when the unmentioned agent, understood from the context, is a first person. This is justified by that person’s maximal control over the action²⁰.

¹⁸ Incidentally, in French too, unmentioned agents are often we, cf. on y va? ‘shall we go?’, or on a bien mangé ‘we ate well, one ate well’.

¹⁹ There are sentences in which man in German can also be interpreted as a substitute for ich (‘I’): Man (impersonal pronoun) hat (‘have’ 3 SG. PRAES. AKT.) halt so seine Beziehungen – “we all have our connections,” lit. “one has one’s connections.” However, this is a special case in which I equals all of us. Cf. from DLKT: Moreover, there is now evidence that winters used to be very severe in those latitudes.

²⁰ Regarding Lithuanian, Masayoshi Shibatani (ibid.: 98) states that it allows only human beings (English human) as agents in the ambipersonal. However, examples such as lojama (DLKT – Fig. 2) or miaukiama (DLKT – Fig. 1) can also be found in Lithuanian, although it must be acknowledged that these are isolated examples.

Straipsniai / Articles

115

MARKUS RODUNER

CONCLUSIONS

Unlike in neighboring Finnic or Celtic languages, Baltic languages do not have a single grammaticalized verb form to express the ambipersonal. Only constructions with a passive participle in the neuter gender, and in Latvian with an uninflected masculine participle, could be called “true” ambipersonals. Nevertheless, the similarity between Baltic and Baltic Finnic languages is evident: in both language groups the ambipersonal (“impersonal passive”), as one would expect, can be formed from both transitive and intransitive verbs—there is almost no restriction by verb type (the ambipersonal is even possible with genuinely unaccusative verbs or with the verb “to be”). In Latvian, specifying the agent of transitive verbs / the experiencer of intransitive verbs in the ambipersonal (only in the progressive passive) is impossible; in Lithuanian it is rare. This suggests that the ambipersonal (elsewhere called impersonal) is basic, while the personal passive is a secondary phenomenon²¹. Moreover, in Baltic languages, as in Finnish or Estonian, the ambipersonal is often used where many Western European languages would use sentences with a subject expressed by the pronouns we or I.

SOURCE

DLKT – Corpus of the Contemporary Lithuanian Language. Available online: donelaitis.vdu.lt.

LITERATŪRA

Ambrazas Vytautas 1990: Сравнительный синтаксис причастий балтийских языков. Vilnius: Mokslas.

Ambrazas Vytautas 1994: On the interpretation of Lithuanian constructions with neuter passive participles. – Linguistica Baltica 3, 7–11.

Ambrazas Vytautas, sud., 1997: Lithuanian Grammar. Vilnius: Lietuvių kalbos institutas.

Ambrazas Vytautas 2001: On the development of the nominative object in East Baltic. – CBL, 391–412.

Ambrazas Vytautas 2006: Lietuvių kalbos istorinė sintaksė. Vilnius: Lietuvių kalbos institutas.

Berg-Olsen Sturla 2001: Subjects and valency-changing mechanisms in Latvian. – Sprachtypologie und Universalienforschung 54(3), 209–225.

²¹ Tokį triasmenio pasyvo atsiradimą iš ambipersonalo įvairiose kalbose nurodo ir Shibatani – pagal principą „maksimaliai padidink gramatinės reikšmės skirtumą“ (Shibatani 1998: 95).

116

Acta Linguistica Lithuanica LXIV–LXV

Ambipersonalās – ar ši kategorija tinka ir baltų kalboms?

Blevins James P. 2003: Passives and impersonals. – Journal of Linguistics 39, 473–520.

Christen Simon 1998: Unpersönliche Konstruktionen und sekundäre Personifizierung. – Linguistica Baltica 7, 51–61.

Dahl Östen, Koptjevskaja-Tamm Maria, sud., 2001: The Circum-Baltic Languages. Typology and Contact. Amsterdam/Philadelphia: John Benjamins.

Endzelīns Jānis 1951: Latviešu valodas gramatika. Rīga: Latvijas valsts izdevniecība.

Gāters Alfrēds 1993: Lettische Syntax. Die Dainas. Frankfurt am Main: Peter Lang.

Geniu šienė Emma, Nedjalkov Vladimir P. 1988: Resultative, Passive and Perfect in Lithuanian. – Typology of Resultative constructions, ed. by Vladimir P. Nedjalkov. Amsterdam: John Benjamins, 369–386.

Holvoet Axel 2000: Once more on the Baltic genitive of agent. – Baltistica 35, 45–58.

Holvoet Axel 2001: Impersonals and passives in Baltic and Finnic. – CBL, 363–391.

Holvoet Axel 2004: Evidencialumo kategorija. – Gramatinių kategorijų tyrimai (Lietuvių gramatikos darbai 2), red. Axel Holvoet, Loreta Semėnienė. Vilnius: Lietuvių kalbos institutas.

Karlsson Fred 2000: Finnische Grammatik. Hamburg: Helmut Buske.

Larsson Lars-Gunnar 2001: Baltic influence on Finnic languages. – The Circum-Baltic Languages. Typology and Contact. Amsterdam/Philadelphia: John Benjamins, 237–253.

Lees Aet 2006: The Undergoer in Passive Clauses in Estonian and Finnish. – Selected papers from the 2005 Conference of the Australian Linguistic Society, ed. by Allan Keith. Prieiga internete: http://www.als.asn.au.

LKG – Lietuvių kalbos gramatika 3. Sintaksė, red. Kazys Ulvydas. Vilnius: Mokslas, 1976.

Manninen Satu, Hiietam Katrin 2005: Some thoughts on Balto-Finnic passives and impersonals. – The Department of English in Lund: Working Papers in Linguistics 5, 65–90.

Mathiassen Terje 1996: Tense, Mood and Aspect in Lithuanian and Latvian (Meddelelser 75). Oslo: Universitet i Oslo, Slavisk-baltisk institutt, 1996.

Mathiassen Terje 1996 (1997): A Short Grammar of Latvian. Columbus, OH: Slavica.

Mel’čuk – Мельчук Игорь 1991: Ещё раз к вопросу об эргативной конструкции. – Вопросы языкознания 4, 46–88.

Nau Nicole 1998: Latvian. München / Newcastle: Lincom Europa.

Nolan Brian 2005: The role of definiteness in the impersonal passives of Modern Irish. – Proceedings of the RRG 2005 conference at the Academia Sinica. Taipei, Taiwan. Prieiga internete: http://linguistics.buffalo.edu/people/faculty/vanvalin/rrg/Brian Nolan.

Straipsniai / Articles

117

MARKUS RODUNER

Noonan Michael 1994: A tale of two passives in Irish. – Voice: Form and Function, ed. by Barbara Fox, Paul Hopper. Amsterdam: John Benjamins. Prieiga internete: https://pantherfile.uwm.edu/noonan/www/2Irish.passives.pdf.

Perlmutter David M. 1978: Impersonal passives and the Unaccusative Hypothesis. – Proceedings of the 4th Annual Meeting of the Berkeley Linguistics Society. UC Berkeley, 157–189.

Schlachter Wolfgang 1985: Hat das Finnische ein Passiv. – Finnisch-Ugrische Forschungen: Zeitschrift für finnisch-ugrische Sprach- und Volkskunde 47(1), 1–144.

Shibatani Masajoshi 1998: Voice parameters. – Kobe papers in Linguistics 1, 93–111.

Skrzypiec Andrzej M. 2000: A preliminary study of Welsh passives and their Polish equivalents. – Poznan Studies in Contemporary Linguistics 36, 151–157.

Szem erény Oswald: Einführung in die vergleichende Sprachwissenschaft. Darmstadt: Wissenschaftliche Buchgesellschaft.

Tauli Valter 1983: Standard Estonian Grammar. Part II. Syntax (Acta Universitatis Upsaliensis. Studia Uralica Upsaliensia 14). Uppsala: Uppsala Universitet.

Tommola Hannu 1993: Ambipersoonainen Suppressiivi: diateesista suomessa ja virossa. – Studia comparativa linguarum orbis Maris Baltici 1. Tutkimuksia syntaksin ja pragmatiikan alalta, sud. Valma Yli-Vakkuri. Turku: Turun yliopiston suomalaisen ja yleisen kielitieteen laitoksen julkaisuja, 41–96.

Tommola Hannu 1997: О супрессиве и об амбиперсонале. – Труды по русской и славянской филологии. Лингвистика, Новая серия, вып. 1, 173–187. Tartu: Tartu Ülikooli Kirjastus. Cituojama iš rankraščio.

Wiemer Björn 2001: Partizipien zwischen Syntax, Semantik und Pragmatik: Ein Überblick zu aspektuellen, diathesebezogenen und diskursrelevanten Eigenschaften im modernen Litauischen. – Sprachbeschreibung & Typologie (Publikation zum Workshop vom 16. Dezember 2000 in Bern). Arbeitspapiere 38, sud. B. Wälchli, F. Zúñiga. Bern: Institut für Sprachwissenschaft, 65–81.

Wiemer Björn, Giger Markus 2005: Resultativa in den nordslavischen und baltischen Sprachen. München: Lincom.

Vihman Virve-Anneli. Nothing Personal. Some detransitivizing devices in Estonian. Prieiga internete: www.ling.ed.ac.uk/~pgc/archive/2001/virvegc01.pdf.

Vihman Virve-Anneli 2004: Valency Reduction in Estonian (daktaro disertacija). Prieiga internete: http://linguistics.buffalo.edu/people/faculty/vanvalin/rrg/Vihman_PhD_dissertation.pdf.

118

Acta Linguistica Lithuanica LXIV–LXV

Ambipersonalas – ar ši kategorija tinka ir baltų kalboms?

Ambipersonal: A Suitable Category for the Baltic Languages too?

S U M M A R Y

In this paper, the author discusses, if the category „ambipersonal“ as postulated as a semantic-pragmatic category with grammaticalised realisation in Finnish and Estonian by Hannu Tommola (1993, 1998) has its counterpart in the neighbouring Baltic languages Lithuanian and – to a much lesser extent discussed – Latvian. The Lithuanian and Latvian – and the Baltic Finnic – passive share the feature that virtually any verb can be passivised. In both Baltic languages constructions with neuter participles (Lithuanian) or non-agreeing masculine participles can be identified as ambipersonal, the easiest to identify being constructions with intransitive verbs. In Latvian as in the neighbouring Finnic languages, the actor cannot be overtly named in constructions involving the auxiliary tikt, denoting actions not states. In Lithuanian, this is possible, also, as shown on the base of the electronic corpus of contemporary Lithuanian, quite rare with constructions involving participles in -ma. We therefore can make the conclusion that the ambipersonal meaning is basic in the Baltic languages and that the full three-argument passive is nearly not or weakly developed.

Įteikta 2010 m. lapkričio 30 d.

M A R K U S R O D U N E R

Lietuvių kalbos institutas

P. Vileišio g. 5, LT-10308 Vilnius, Lietuva

[email protected]

Straipsniai / Articles

119