Netiesioginės nuosakos (modus relativus) paplitimas ir kilmės problema

LIETUVIŲ AREALINĖS LINGVISTIKOS KLAUSIMAI LIETUVOS TSR MOKSLŲ AKADEMIJA V. AMBRAZAS THE DISTRIBUTION AND ORIGIN OF THE INDIRECT MOOD (MODUS RELATIVUS) 1. CONSTRUCTIONS WITH ACTIVE PARTICIPLES 1.1. The forms of the nominative case of Lithuanian active participles that occupy the verbum finitum position in a sentence and contrast with the direct mood on the basis of the meaning of a reported or otherwise indirectly experienced action are usually called the indirect mood (modus relativus, hereafter II); cf.: he is said to have lived in the village (ind. mood) and he lives in the village (direct mood). The term “indirect mood,” first used in 1937 in Gimtoji kalba1, has become established in Lithuanian linguistics in recent decades2. Although it is sometimes debated3, it is nevertheless more convenient and precise than the previously used term “participial speech,” which denoted many different cases of participle use. Like the other Lithuanian moods, II has a specific modal semantics and a defined paradigm of forms. 1.1.1. In contemporary literary Lithuanian, the indirect mood paradigm consists of the following forms: 1. Simple forms: present stem, singular: metęs, -anti, plural metę, -ančios; past frequentative stem, singular metęs, -usi, plural metę, -usios; past simple stem, singular mesdavęs, -usi, plural mesdavę, -usios; future stem, singular mesiąs, -ianti, plural mesią, -iančios. 2. Compound forms. Inchoative active forms: past frequentative stem, singular buvęs bemetęs, buvusi bemetanti, plural buvę bemetę, buvusios bemetančios; future stem, singular būsias bemetęs, būsianti bemetanti, plural būsią bemetę, būsiančios bemetančios; active perfect forms: present stem, singular esąs metęs, esanti metusi, plural esą metę, esančios metusios; past frequentative stem, singular buvęs metęs, buvusi metusi, plural buvę metę, buvusios metusios; 1 Žr. J. Lazausko str. „Penkt oji nuosaka“ GK 1937, Nr. 2, p. 24–6 su pritariamu redakcijos prierašu. 2 Cf. J. Balkevičius, Sint. 94 ff.; P. Gailiūnas, LKG II 58–9; N. Sl žienė, LKG II 167 ff.; A. Paulauskienė, Veiksm. 158 et al.; on the term modus relativus in Lithuanian, cf. E. Fraenkel, IF XLVI 44; P. Trost, LKK III 220; W. K. Matthews, SEER XXXIII 370, E. Hofmann, Indogermanica 29; I. Marvan, SIPr IV 254, B. Сиртаутас, Глагольное сказуемое в совр. литовском лит. языке, Авт. канд. дисс., Вильнюс, 1968, p. 5; Ia. J. Endzelins, DI.I 144; LaGr. 902; M. Rudzīte, Dial. 138, 241, 374; A. Bergmane, MLVG I 624 et al. Paradigms of forms with corresponding meanings in other languages are likewise generally regarded as separate moods and called modus obliquus (A. Sauvageot, BSL L. 2 208; W. K. Matthews, SEER XXXIII 370), commentativus (W. Fiedler, WZL XV 3 565), imperceptivus (V. Koseska, SFPS XIII 166), narrativus (B. Panzer, OSS III 70 and in Bulgarian grammars), auditivus (J. Endzelins, LaGr. 902; В. З. Панфилов, Взаимоотношения языка и мышления, Москва, 1971); according to the universal table of modality and moods proposed in B. Panzer, OSS III 70—modus narrativus (auditivus). Thus, the term modus relativus used for Lithuanian and Latvian is traditional. On the theoretical foundations and rationale for distinguishing II in Lithuanian and Latvian, see I. Marvan, SIPr IV 253–4. 3 A. Girdenis, V. Žulys, Baltistica IX (2) 210. 7 past anterior tense, singular: būdavęs metęs, būdavusi metusi, plural: būdavę metę, būdavusios metusios; future tense: singular: būsiąs metęs, būsianti metusi, plural: būsią metę, bū siančios metusios; passive voice with present passive participles: present tense, singular: esąs metamas, esanti metama, plural: esą metami, esančios metamos; past frequentative tense, singular: buvęs metamas, buvusi metama, plural: buvę metami, buvusios metamos; past anterior tense, singular: bū davęs metamas, būdavusi metama, plural: būdavę metami, būdavusios metamos; future tense, singular: būs iąs metamas, būsianti metama, plural: būsią metami, bū siančios metamos; passive voice with past passive participles: present tense, singular: esąs mestas, esanti mesta, plural: esą mesti, esančios mestos; past frequentative tense, singular: buvęs mestas, buvusi mesta, plural: buvę mesti, buvusios mestos; past anterior tense, singular: bū davęs mestas, būdavusi mesta, plural: būdavę mesti, būdavusios mestos; future tense, singular: būs iąs mestas, būsianti mesta, plural: būsią mesti, bū siančios mestos. The paradigm of NN forms is symmetrical with the paradigm of the indicative mood. Both the simple and compound tense forms of NN contrast with the corresponding indicative forms, cf.: yra, buvo, būdavo, bus metęs: esąs, buvęs, būdavęs, būsiąs metęs. The same relationship holds between the indicative and the passive forms of NN yra, buvo, būdavo, bus metamas: esąs, buvęs, būdavęs, būsiąs metamas; yra, buvo, būdavo, bus mestas: esąs, buvęs, būdavęs, būsiąs mestas. 1.1.2. NN differs from the other Lithuanian verb moods in that its forms are not inflected for person. In this respect, it can only conditionally be regarded as part of the system of finite verb forms. Its place among the other moods is also distinctive. Literary language uses NN forms which, on the basis of the meaning of an indirectly reported action, are correlated with the so-called optative mood (or the third person of the imperative mood), e.g.: tenėša: tenėšas, -anti; tenėšą, -ančios; tegu(l) neša: tegu(l) nešąs, -anti, tegu(l) nešą, -ančios. In this case, the overlap between the indirect and optative (imperative) moods is explained by the fact that the forms of the so-called optative mood tenėša, tegu(l) neša, like tenėš, tegu(l) neš, in fact have indicative endings and may be regarded as indicative forms with the particles te-, tegul (cf.: benėša, benėš4). In the Eastern Aukštaitian dialects (around Subačius and Karšakiškis), NN also has optative forms correlated with the conditional mood, cf.: eitų: eitąs, -anti; būčiau ėjęs: būčias ėj ęs5. This is because NN expresses the modality of an action from a somewhat different perspective than the conditional mood, namely, according to the nature of the information available to the speaker and the speaker’s view of the action’s actuality. However, these dialectal forms are narrowly distributed and do not constitute a formal obstacle 4 This also applies to the corresponding Latvian forms, e.g. lai ved; lai vedot, cf. I. Marvan, SlPr IV 255; for other formal and semantic motivations, see P. Пазухин, Так называемое „повелительное наклонение“, „Studia Rossica Posnaniensia“, VI, Poznań, 1975, p. 94. 5 E.g.: kad norėtu.s, tai jimtu.s ir padari.tu.s Sb; jėgu būt.čiu.s žinoj.si.s, nebūt.čiu.s nė iš vieta.s ju.dimi.si.s Krs, see A. Kučinskaitė, K. Morkūnas, MAD 1 (16) 215 ff.; A. Vidugiris, LKK VII 190–1. 6 R. Jakobson regards the description of an action on the basis of testimony, i.e. on the basis of a report about that action, as a separate evidential category; see Р. О. Якобсон, Шифтеры, глагольные категории и русский глагол, — Принципы типологического анализа языков различного строя, Москва, 1972, p. 101. On this basis, paradigms of forms with meanings similar to those of Lithuanian and Latvian NN are distinguished in many languages belonging to different families, e.g. Bulgarian, Macedonian, Albanian (see 3.6.1–2), Estonian, Livonian (see 3.5), Turkish, Turkmen (see 3.6.3), Arabic (Б. М. Гранде, Введение в сравнительное изучение семитских языков, Москва, 1972, p. 397), Georgian, Svan (J. Lohmann, KZ LXIII 45 ff.; У. Д. Салтиховшвили, Sintagmatika 146 ff.), Native American languages (Р. О. Якобсон, ib. with lit.), Nganasan (Н. М. Терещенко, ЯН СССР III 431), etc. 8 to consider nominative participles denoting a reported action as one of the moods of the Lithuanian verb. 1.1.3. The most important structural features of NN forms that distinguish them from other participial nominatives are the position of the finite verb in the sentence and the absence of a finite form of the auxiliary verb. In terms of verb position, the participles forming NN differ from the corresponding attributive and semi-predicative participles (cf. the aching hand recovered and his hand aching). Furthermore, when forming NN, the masculine singular and plural nominatives of present active participles usually have a short form (historically, a consonant-stem form), e.g.: gyvenas, gyveną, whereas attributive participles more typically have the longer, newer forms (gyvenantis, gyvenantys).7 NN differs from predicative participles that form part of the system of compound (periphrastic) tenses in lacking a finite indicative form of the auxiliary verb; cf. yesterday he had written a letter and he has (had, will have) written a letter. Past frequentative and future active participles are not used with an auxiliary verb and are therefore the clearest representatives of NN in this respect. In present-day literary language, present active participles are also almost never used with the present-tense form of the auxiliary verb būti, and they combine with its past simple, past frequentative, and future forms only when they have the affix be- and form the so-called inchoative tenses. Thus, when occupying the finite-verb position without a copula, present active participles are always understood as NN forms. Past simple active participles belong to the subsystem of compound perfect tenses and, when used without a finite form of būti, can be regarded as present perfect forms with a zero copula, e.g., he is tired. Such nominatives of past simple active participles (especially prefixed forms with perfective meaning) are homonymous: whether they mean the present perfect indicative or the past simple NN depends on the context. The same applies to the compound passive past simple forms buvęs metamas and buvęs mestas, where a zero copula is likewise possible. Only in the three cases just mentioned is the opposition between the indicative and the indirect mood formally unexpressed (neutralized). In cases of neutralization, the meaning of NN can be made explicit by adding another participle of the auxiliary verb būti to the past simple participle, that is, by using the compound form jis esąs (buvęs) pavargęs instead of the simple form jis pavargęs. Used in place of simple forms, compound perfect NN forms emphasize the modal nuance of the action8. A similar function is served by the neuter form esą of the participle, which developed from NN and became an parenthetical word (less often būsia LKŽ I 1209 or even būsin ib.), expressing the speaker’s uncertainty or doubt about the reality of the action9. 7 Only in some dialects (mainly along the Nemunas) are the long singular nominative forms of present active participles used alongside the short ones, e.g.: sakė, jis žinantis / žinąs visą miestą Ig, Grl, Kdn, or even predominate (Zp, Plūt, Jsv, Krkn, Ss). Such usage is not characteristic of the literary language. Except for J. Balčikonis’s writings from his final period, in which he generally sought to avoid short forms, NN in the literary language is almost always formed by the short nominative forms of present active participles. A tendency to formally differentiate attributive and predicative participles is also evident in other languages, especially Slavic ones (see 2.2.5.2 for these). Latin participles with a meaning close to that of verbs also generally retain the consonant stem, whereas those that have become adjectival tend to shift to the i-stem; cf. J. Safarewicz, The development of participles in Latin, – Linguistic studies, Hague – Paris, 1974, p. 245. 8 Cf. the corresponding (only more emphatically expressed) function of compound (or emphatic) indirect-mood forms in Latvian (see 1.3.1) and Bulgarian (see 3.6.1). 9 Cf. modal particles used for the same function, or words that become such particles, in other languages: Russian мол (from молвити B. Fraenkel, Baltosl. 40), Sanskrit iti (W. D. Whitney, SkrGr. 89, 405), and so on. 9 1.2.1. The semantics of NN forms consist of specific modal meanings. These mood forms denote: 1) a relayed or otherwise indirectly experienced action, or another person’s speech or thoughts, e.g.: I heard that he had drunk well10; others say that he too had got his share Tt; here, they say, there used to be forests Krš; When our old woman used to tell about the corvée, how they would go, how they would work Mšk; he only asks whether I will have fifty litas for the medicine Šll; he’ll do that, he’ll do that as it should be, nothing wor Plvn; mistaken in thinking there would be no storms BalčHP 25; the shepherd knew that the old man had money Ms; 2) an action that is doubted, not known for certain, or regarded as unreal, e.g.: we thought the little bees were buzzing over JV 912; perhaps he is said to have separated from his wife SmIn; the buried money is burning Vrn; that they had supposedly gone to the races Mžš; the parents had been worrying whether Peter, having shown himself among people, might bring some misfortune upon himself Myk-PutS 514; it seems the people had come out of their graves VrnV; 3) an action experienced suddenly and unexpectedly, giving rise to astonishment11, e.g.: I go out into the field, and lo, the clover is warm Krp (LKŽ VII 589); whenever there was a village dance, look—there he already is, coming Jnš (LKT 214); I go to see what’s wrong with that dog, and look—I’ve found a hedgehog Sk (LKT 153); I stretch out and shine a light—and the cat has crept into the granary and got trapped Rt (LKT 89); 4) an action inferred from its results, e.g.: Ah, look, you too already seem to understand Lithuanian… SimonR IV 74; those cats seem not to have licked their own bowl clean Vvr; the father inspected the horses and saw that they had been with a smoky horse Grg; apparently those children, little rascals, had been in the beans: the biggest tracks had been left Tt (LKT 108); the bull had apparently broken loose, knocked the potato ridges over, and rooted around Krtn (LKT 76). Thus, the semantics of NN are not homogeneous: they include meanings that in other languages may be marked by separate moods, e.g. relaying—by the narrative, incongruity with reality—by the irrealis, doubt—by the dubitative, suddenness and astonishment—by the admirative, and so on.12 In Lithuanian, all these meanings are linked by a single NN paradigm. However, they are not equally widespread, nor are they equally characteristic of all participial forms. The most widespread is the meaning of a relayed, indirectly experienced action, closely associated with a shade of doubt (things relayed may by their very nature seem doubtful to the speaker). These meanings are found in participles of all tenses. The meaning of a suddenly experienced action is more characteristic only of present active participles (especially with the prefix be-) and, less often, past frequentative active participles. These forms may also convey the meaning of an action inferred from its results, which is the rarest of all. The latter two meanings are also closely related. Thus, the semantics of NN consists of two interrelated groups of meanings that share the same morphological expression and are henceforth referred to collectively by the term ‘indirectly experienced action’. It should be noted that an indirectly experienced action may also be expressed by indicative forms, especially when relaying or doubt is marked in the sentence by lexical means. The constructions he said, he will come and he said, he will apparently come are interchangeable in contemporary literary language. This is because NN forms a privative opposition with the indicative. NN is the marked member of this opposition, and the distinguishing semantic feature is the meaning of an indirectly experienced action. 10 Predicate examples in the article are phonetically transposed into the literary language while preserving their morphological and syntactic features, and are identified by LKŽ abbreviations. Examples from dialect texts published in the nineteenth and twentieth centuries are also presented in current orthography. 11 Cf. “We often make use of participles for unexpected, sometimes seemingly belated observations, experiences, and remarks in general” J. Jablonskis, RR I 527. 12 Cf. B. Panzer, OSS III 63 ff., 70. 10 1.2.2. Forms of NN are used: 1) in subordinate clauses of complement sentences, when the main clauses contain comment-taking (i.e. phrasal-control)13 verbs, most often denoting sensations, perception, thought, and the conveyance of information, e.g.: he said that there was no rye Slm; I heard that he was not at home Skd; 2) in the second clauses of asyndetic sentences, usually following verbs with the same meaning, e.g.: Ona had dreamed that she was going everywhere, talking everywhere Jrb; he said that the thunder had burned down the cottage Ktk; 3) in independent sentences, e.g.: the aitvarai would carry things to whoever had them BsMt I 129; the potter, weeping, would be pouring the horse with a stick over the pots, until they had all scattered into shards BM 52. In the first (preposed) clauses of asyndetic sentences, NN participles occur relatively rarely, e.g.: so he had beaten the wolf, they say KzR; she had sung all kinds of songs, they say Vvr. In subordinate and synonymous asyndetic sentences, the modal character of the participle’s meaning is brought out by the verb in the other clause14; in independent sentences, it becomes clear from the context and situation. 1.3.1. In literary Latvian, the indirect mood (atstāstamā izteiksme) is generally formed by fossilized present- and future-tense active participles with the suffix -ot <*ant (celot, celšot, etc.). The forms of participles that do not inflect for case or gender esot, tiekot (tapot); būsot, tikšot (tapšot) serve as auxiliaries in compound tenses. The fossilized future-tense active participles now have only the meaning of NN, and from the perspective of the present-day language are no longer classified as participles at all15. Past-tense active participles are now regarded as paradigmatic forms of NN only when they occur with a participial copula, e.g., esot runājis, that is, only so-called compound present-tense forms are used to express past action. Thus the modal function of the participial copula is more pronounced in Latvian than in Lithuanian. In this way, homonymy between NN and the compound present indicative is avoided in the literary language paradigms. In fact, however, in Latvian dialects and, to some extent, in the language of belles-lettres, past-tense active participles are still used with the meaning of NN without the copula esot, e.g., a certain father had had three sons16, that is, just as in Lithuanian. In the High Latvian (aukšzemnieku) and some other dialects, the NN active participle may still agree in gender and number with a noun or a pronoun in the nominative, e.g.: they had run up, she cannot catch her breath17. Forms of NN with agreement are also attested in old Latvian texts; cf. in Adolphi’s grammar nom. sing. masc. sarguots, fem. sarguoti; nom. pl. masc. sarguoši, fem. sarguošas, and so on.18 NN was also formed with the neuter form of the present-tense participles esu ~ Lithuanian esą, which is still known in dialects, e.g., es esu bijis (R. Grisle ib. 252). Thus the Latvian indirect mood consists historically of the same participial forms as in Lithuanian, only more altered at present. In addition, following the pattern of the indicative paradigm19, indirect 13 For the terminology, see the author, LKK XV 210 with references. 14 Sometimes it is inserted into a sentence or clause containing an NN participle and becomes an parenthetical word, e.g.: the hospital, they say, had been Sd. 15 J. Endzelīns, LaGr. 1009; A. Bergmane, MLLVG I 624, 661 ff.; J. Kabelka, La, 155, 166. 16 J. Endzelīns, ib. 977; see also A. Bergmane, ib. 625. 17 J. Endzelīns, ib. 977; M. Rudzīte, Dial. 374. 18 P. Šmits, RKr XIV 29, 38; J. Endzelīns, ib. 977–8; R. Grisle, LVLIR VII 251–2. 19 J. Endzelīns, DII 144. Since the Latvian debitive also intersects with the conditional mood (cf. būtu jāmet, būtu bijis jāmet), I. Marvan, SIPr IV 255, proposes separating it from the other moods and treating it as a separate verbal category. 11 necessitative mood (debitive) forms, e.g.: esot jāmet, būšot jāmet; esot bijis jāmet, būšot bijis jāmet. As a result of these later changes, Latvian NN became morphologically more clearly defined20. Semantically, Latvian NN corresponds to the principal, i.e. the first and second, group of meanings of Lithuanian NN. It is usually used to denote an action that is reported, not known for certain, or doubted. 1.3.2. No participles with a clear meaning of indirect mood are attested in Old Prussian texts. However, Old Prussian participles can occur in the position of a verbum finitum and, by syntactic function, correspond to NN forms. The syntactic precursors of NN include present active participles used after the relative pronoun kawids in subordinate clauses of complex sentences, e.g.: tans… prei isspreštun dāst | en kawidsei debīkun nautīn bhe wargan stai gurijnai malnijkiku embaddusisi 71₂₋₅ — Er… zuerstehen gibt | in was grosser Not vnd Gefahr die armen Kindlein stecken; kai ious… sen seilin pomīrit quoietē en kawidsu debīkan powargsennien | bhe nautien | schis malnijkiks… embaddusisi 71₂₉₋₃₂ — das ihr… mit fleiß bedenken wolt | in was grossem Jammer vnd Noth diß Kindlein… stecket21. The autochthonous character of such usage is confirmed by the fact that a finite verb form in the original is translated by a participle (cf. the corresponding Lithuanian participial constructions with conjunctions in 1.9.3). 1.4. The indirect mood in both Lithuanian and Latvian, as a particular verb category, is relatively recent and developed from forms of nominal origin. Although the participles that constitute it (formerly verbal adjectives), with the suffixes nt and us, were inherited from the period of Indo-European unity, such usage is not attested in most related languages. Nor is it reflected in the oldest texts of the Indo-European languages (e.g., Hittite texts, the Vedas, or Homer’s epic). This suggests that NN developed during the independent evolution of the Baltic languages, and that its basis should be sought in the use of participles in these languages. Various views have been put forward concerning the origin of NN in the Baltic languages. 1.4.1. B. Delbrück regarded the accusative with participle construction (accusativus cum participio) as the basis of the indirect mood (VS II 491) and argued that, from expressions such as said his father would come (would be coming), alongside which synonymous expressions with the indicative were used, said that his father would come, the construction said his father would come could have arisen, followed by said that his father would come. J. Endzelīns agreed with this view (LaGr. 976–977, 979), adding that contamination between the accusative-with-participle construction and synonymous expressions with indicative verbs may also have facilitated the nominative-with-participle (nominativus cum participio, hereafter VD) construction, e.g., he is said to know a lot. According to E. Hofmann, Baltic NN developed from indirect speech expressed by constructions of the accusativus cum gerundio type22. G. Bense likewise derives Latvian NN from an accusative construction as its original source (“letzten Endes”) (LKK VI 192). 1.4.2. A. Potebnia (Зап. I–II 230–1) linked the indirect mood with VD (nominativus cum participio) and related constructions. He argued that the indirect 20 Cf. I. Marvan ib. 253. 21 See J. Zubaty, IF VI 302; E. Fraenkel, Baltosl. 46, contra J. Endzelin, AprGr 22 E. Hofmann, On the Use of Participles in the Skireins, Indogermanica, Festschrift für W. Krause, Heidelberg, 1960, p. 29. 12 The nuance of hearsay or doubt may initially have arisen from the lexical meaning of the verb used in these constructions (e.g., dûmoti, dėtis), and may later have been reinforced by conjunctions such as buk, nei, e.g., bijojos, dûmodami, buk dwase kokę regį; dėtis turėjo, ney esą teisus (the construction itself was initially indifferent with respect to modality). Such constructions correspond fully to Slavic ones: mъnjaacho (aky) ducho videste. The Baltic languages, however, went further: they began to use participles even when the subject of the participial action did not coincide with the subject of the verb’s action (e.g., I do not know whether he will have come 225, vyns lika atsacit, ka tēvs majās ne esot 227). Regarding as more recent only the constructions common in Lithuanian and Latvian with a different subject of the participial action, A. Potebnia also pointed out circumstances that contradicted this view: heterosubject constructions are closer to nominal constructions and are less unified, whereas language development proceeds precisely in the direction of sentence integration.23 1.4.3. E. Tanglis, Acc. 50–51, derives the indirect mood directly from VD (nom. c. participio) constructions with non-reflexive verbs, which, in his view, arose later and had not yet become widespread in the 16th–17th centuries. E. Tanglis regarded the indirect mood as a new phenomenon that apparently took hold only during the period of written tradition. I. Marvan similarly explained the internal causes behind the emergence of NN24, adding that VD with non-reflexive verbs were variants of subordinate clauses and could, by analogy, have acquired conjunctions and a separate subject of the action. Since these authors derive VD from an accusative construction, they postulate approximately the following developmental chain for NN: says herself to have come → says himself to have come → says (that) he has come → says (that) father has come → father has come. 1.4.4. J. Kazlauskas explained the formation of NN quite differently (LKK IV 86–87, Ist. gr. 402–403). He thought it might be a remnant of a periphrastic optative (*bi negelbas, *bi nekelės), whose existence in Lithuanian was demonstrated by Ch. S. Stang (NTS XVIII 348–356, VGr. 429), on the basis of conditional constructions with participles, e.g., jeib negelbas, jeib nekelės, and the like. In other Indo-European languages, indirect speech is often expressed by the optative, so this may also have been the case in Lithuanian. Once the burden of optative meaning had shifted from the auxiliary verb bi, which had become a particle, to the participle, that particle could have disappeared. Thus, the origin of NN is explained in rather contradictory ways. To clarify it somewhat, it is useful to examine the use of NN at the beginning of the written tradition, its spread, and its relationship to other constructions of similar meaning. 1.5. What do we know about the use of NN at the beginning of the Lithuanian written tradition? E. Tanglis, Acc. 50–51, found no examples of NN in old texts. Independent NN constructions, formally distinct from VD, are indeed not characteristic of the language of the writings by J. Bretkūnas, M. Daukša, and others that he examined. Individual cases in which participles approximate the indirect mood (e.g., DP 51236, BrB I Moz 48, 18; Joz 10,1) can also be interpreted as compound tense forms with zero copulas25. However, an examination of other 16th–17th-century sources leaves no doubt that NN was fully formed at the beginning of the written tradition. In this respect, the writings of S. Vaišnoras are particularly significant. In his translations “Žemčiūgos” and “Apie popiežiškąją mišį,” we find 35 fully independent (with a separate 23 More recently, V. Sirtautas has expressed a similar view, Kalbotyra XXV (I) 95. 24 H. Марban, Kalbotyra IV 36–37; Baltistica V(1) 18. 25 For example: Ka wel paskui to daug kitų szwentių / miniku / ir mergu padariusias DP 51236 (Co tež potym wiele inszych świętych… vczynilo W 876). 13 examples of NN with the verb as subject26. Since they have not been discussed anywhere to date, we will describe them in somewhat greater detail. 1.5.1. With the exception of one case with optative meaning (see 3.3.2), all NN forms in S. Vaišnoras’s translations occur with the modal particle būk, which emphasizes a nuance of uncertainty or doubt27. NN with this particle is still widely used in Samogitia (e.g., around Telšiai, Šalantai, Len kimai, Paplė, Tytuvėnai, Užventis, Kalnėnai, Šiauliai, Raseiniai, and Laukuva), as well as in northeastern Lithuania (around Joniškis, Pasvalys, Linkuva, Daugailiai, Dusetos, and Miežiškiai), and sometimes even in southern dialects, where NN participles are very rare overall. S. Vaišnoras usually uses these participles in subordinate clauses of complex sentences, with the conjunctions jog, kaip and the relative pronouns kuris, kursai. When NN occurs in a clause of a complex sentence that has no conjunction, the same particle būk serves in its place. Subordinate clauses with NN forms are often very far removed from the main clauses, separated from them by whole sequences of other clauses, and are consequently fairly independent. In reporting indirect speech, the main clause or preceding context contains the verbs sakyti, perkalbėti, taryti; paspirinti “assert,” rašyti, for example: Teipajeg saka ghie… jog spawiednie būk priwalingas essąs ischskaitimas wissų griekų 14822; skaisczęi nor perkalbeti | kaip būk macis tokia io walgimo | ir kitiems tos regetoiems ant gera atenti PM 3223; netikumus staczei uschstoie taridami… jog ipaczia wiera ing Christu | būk nesanti dalis Pakutos 14815; raschiti | ir neprotingai pastiprinti dressa… Buk Chrikschtas… ne essas apmasgoimu 132a21. When conveying thoughts and beliefs that the author regards as unreal and untrue, NN participles follow the verbs dūmoti, laikytis, mokinti, pramanyti, pramanymūs ginti: (the Stoics) dumoie | būk Diewas ir ischkadingu paiudin ojimu (177) žmones biauribump werčęs 177–177a. Popiežischkei… klajojme laikosi | jog būk dūna ingi kuna Christaus butischkai persimainanti PM 30a14; mokin teipajeg. Iog atgimimas ir ischganimas mussu būk ne nūg Chriks ch ta kabąs 133a12. Nessa žmogischkasis protas pramana | Diewa santi sudcze swietischku | kursai sawo dingscze būk danguie ischkadingus pastatimus raschęs 16421; pramanimus ginna: Pirmiausei iog Diewas… nekur ius žmones … ant amžinoses prapulties būk paskires 178a23. 1.5.2. In S. Vaišnoras’s writings, indirect mood is most often formed by present active participles (25 examples), and less often by past active participles (5 examples). The participle of the verb būti (12 examples), functioning as a copula, is linked with nouns (ne essas apmasgoimu 132a24, uschmokejmu… essas 264a6, kitu nosrais essa 3223), adjectives (priwalingas essąs 14823, nei naudingas nei ischkadingas… essas 132a7), and their 26 In two of the examples, the spelling makes it unclear whether we have participles or simply finite forms of the indicative: ischrinktieghi būk palaika (-ą or -o?) malone 17913, Apaschtolai būk… laike (-ę or -ė?) PM 376. The other cases raise no doubts. Six constructions occur in the book “Apie popiežiškąją mišią” (PM), and 29 in “Žemčiūga” itself. 27 The same particle is also used several times with the indicative in the same sense (11a16, 13320, 1785–8, 178–178a, PM 2321, 27a11–13), and in one place with a participial genitive: ghie ir regima affiera… (25) be iokia ischstowincza grunta | kaip buk priwalomo sancze | be reikos stata PM 25–25a; for corresponding cases in present-day language, see LKŽ I 1136; cf. also E. Fraenkel KZ XIV (Erg. H.), 36, 98 ff. Dll. būk sen. in older writings is also used as part of the conjunction būk… alba, būk… “whether… or,” for example: būk tatai Namai alba Miestas būtų BrB III Moz 25, 34 (cf. also 17, 10; 22, 18; 22, 28; IV Moz 15, 2; V Moz 12, 3; 13, 6, etc.). 14 with neuter forms (priwalu essa PM 33a₁₈) and a prepositional construction (nūg czesa Apaschtolu santi PM 39₆). With present passive participles, the participle of the verb būti forms compound NN passive forms: Adoma butibe… muodemaĩ buk santi attolinama 69a₁₇; tie | kad isch schos giwatas ischeiti… nūg biaurjubu grieku buk essa nuczistijemi 263a–264; pramana ipaczus pagawimus | kureis buk žmones ir be to vreda Baszniczes esq priwerczemi 113a₁₂; neatmestiemus | kurie… buk niekudais ne essa atgimdomi 133a₁₇. In S. Vaišnoras’s translations, NN also includes combinations of the converbal form esant with the neuter form of an adjective priwalu and the predicative adverb gana, as well as impersonal converbs reikent, tereikiant, which have the meaning of the corresponding neuter participial forms (cf. reikią, tereikią): norėtu tu perkальbeti | ghiems nebereikent walgjti | bet buk essant ganna PM 32a₆; Paskiausei mokin | iog buk apie nuleмima ir isch-rinkima Diewa reikent suditi | tiektaĩ isch tiuau nūdarbiu 179₁₉; Enthusiastai… saka… buk reikent laukti pagawimu gwoltinguiu 69a₂₃ (opposites) nūg Euangelias pradesti | ir tq be žokana tiektaĩ per Euangelia buk teireikent sakiti ir mokinti 75a₂₃. Such expressions following stated or implied (75a₂₃) verbs of explanation are very close to the accusative with a converb; cf. mokiną reikjant sūdyti, sako reikjant laukti. It was precisely in impersonal accusative-with-participle constructions that the converb first began to spread, displacing the neuter participial form; see Baltistica IV (2) 197 ff. Apparently, there was a tendency to generalize the converb form in other cases of impersonal usage as well; this is also indicated by expressions cited by F. Kurschat, such as aš misliju, (kad) šiandien lysiant KGr. 424, 426, alongside lysia K I 284. 1.5.3. As in the present-day language, in S. Vaišnoras’s writings NN can occur alongside the indicative with exactly the same meaning, for example: anis toktai… raschti | ir neprotingai pastiprinti dressa: 1. Pirmiausei | Buk Chrikschtas wlosnai kalbant | ne essas apmasgoimu atgimima … bet Figurischkai (132ᵃ)… buk tiektaĩ issimano widutini atgimima bei atnaughinima Dwasses schwentoses | kursai ir be Chrikschta | ischrinktkusamp buk galis nusidūti ant amžina ischganima 132a–133. 1.5.4. The examples discussed show that the NN constructions found in S. Vaišnoras’s writings correspond essentially to the present-day language, both in their structure and in their meanings. NN functions here as an established grammatical means of expressing a reported, doubtful, or unreal action. The fact that NN is most clearly attested in the sixteenth- and seventeenth-century sources in S. Vaišnoras’s writings is attributable primarily to the author’s Samogitian origins, since indirect mood is most widespread there even today, and to the polemical purpose of the books themselves. In setting out views alien to him, the translator sought to emphasize their unreality by every linguistic means and to distance himself from them (this is also shown by the above-mentioned use of the particle būk with indirect-mood forms and, in some places, with indicative forms). On the other hand, the avoidance of NN in most other writings of that period can also be explained in part by their nature and content: the religious truths set out in them and the events described were regarded by their authors as indisputably true and certain. Since NN has also been used in Latvian since the seventeenth-century written sources, there is no doubt that it arose in the Eastern Baltic dialects during their prehistoric development. 1.6.1. The questionnaire of the Atlas of the Lithuanian Language helps establish the distribution of NN across the area of present-day Lithuanian and its relationship with other constructions of similar meaning, particularly the responses to items 120, 145ᵇ, 151, 175, and 176 of the atlas’s data-collection program 15 questions28. Combining those answers with data from various dialect texts29, it is possible to delineate at least approximately the distributional boundaries of NN and some related participial constructions. The area in which NN is used is marked in Fig. 1. The distributional boundaries of syntactic phenomena fluctuate much more than those of phonetics or morphology, and are therefore considerably harder to establish. Individual syntactic constructions, being more or less motivated in nature, often travel far beyond their principal area of use through folklore texts, idioms, and other means. Thus, the isoglosses for individual participial constructions marked in the diagrams are not absolute but relative: they indicate areas in which the likelihood of choosing a particular construction is greatest and in which tendencies toward its use are most pronounced, although corresponding utterances may also occur elsewhere. The nature of the dialectological material used also makes it impossible to establish strict boundaries for the use of participial constructions. Not all the questions mentioned were formulated with sufficient precision in the atlas data-collection questionnaire (e.g., question 176 presents an alternative: whether the accusative with a converb or indirect speech is used, although in fact both cases are possible in many places). Moreover, some answers written by less experienced assistants simply repeat the sentences in the questionnaire, so their evidential value is doubtful. Such answers were given the least consideration; the analysis relied above all on authentic sentences recorded at the site under investigation. Where possible, answers to several intersecting questions in the questionnaire were compared with one another (e.g., 151 and 176; 120 and 145b), and boundaries were drawn on the basis of the comparison (disregarding contradictory, inconsistent, and suspicious answers). Agreement with dialect texts of the same dialect published by another transcriber or available in manuscript was considered an argument confirming the reliability of an answer. 1.6.2. NN is most widespread in the Samogitian dialect. In the east, the isogloss marking intensive use of NN roughly coincides with the current boundary of the Samogitian dialect, while in the south it also encompasses part of the Western Aukštaitian dialect (including all the Aukštaitians of the Klaipėda region, to whom, as older recorded texts show, at least some of the former East Prussian Aukštaitians should also be assigned in this respect). In this area, present, past simple, and future active participle forms are used intensively; they may occur either entirely independently (e.g., in recounting a longer text) or after verbs sakyti, girdėti, manyti, rodytis ‘say, hear, think, seem’, and the like, with or without conjunctions. In Užnemunė, NN is less frequent; in many places only constructions with past simple active participles have been observed, which can also be regarded as present perfect forms with a zero copula. A second prominent center of NN use lies in the northeastern corner of the Lithuanian-speaking area. Its western edge in the north also roughly corresponds to the boundary between the Panevėžys and Kupiškis dialects of Eastern Aukštaitian, but does not reach Anykščiai or include the southern part of the Eastern Aukštaitian dialect. Present and future active participles are used intensively with the meaning of NN in the northern part of this area (around Rokiškis, Kupiškis, Dusetos, and Zarasai), whereas south of Utena only past simple forms are more commonly found. The two centers of intensive NN use mentioned above are connected in the north by a very narrow strip running along the Latvian border. To its south, across a wide gap (almost from Šiauliai to Anykščiai), NN is not used at all or occurs only sporadically. It is also entirely uncharacteristic of the Southern Aukštaitians, whose dialect area after verbs of saying uses- 28 See Lietuvių kalbos atlaso medžiagos rinkimo programa, LTSR MA Lietuvių kalbos ir literatūros institutas, Vilnius, ²1956. 29 Of these, the sources used most extensively were texts published in the reader Lietuvių kalbos tarmės, ed. E. Grinaveckienė and K. Morkūnas, Vilnius, 1970, in the “Lietuvių kalbotyros klausimų” and “Kalbotyra” series, and recorded in diploma theses by VVU Lithuanian philology students—descriptions of individual dialects. 16 1. NN with present participles 2. VD with reflexive verbs (užsakėsi vaizduotis) 2. Order No. 1026 The past simple forms of the participles do not differ at all from the compound perfect tenses. 1.7.1. A comparison of the distribution areas of NN and the nominative with a participle (VD) after verbs of saying (see Fig. 1) shows that they largely coincide. This overlap is particularly pronounced in the Samogitian dialect area. The available data suggest that NN is used somewhat more intensively than VD in the lower Dubysa region and along the Nemunas30. In northeastern Lithuania, where VD is formed with non-reflexive verbs of saying, the two constructions cannot be distinguished at all: the utterance sako šiandien dirbęs ‘he says he worked today’ may be regarded both as VD and as NN, with the subject of the participial action coinciding with that of the verb. The fact that no pause is made here between the verb and the participle is not a specific feature, since very often there is no pause after the verb even when the participle has its own distinct agent, e.g.: (people) sako devintą dešimtį einanti Vvr; jau vakar vakare sakė mirusi Grg; girdėjau ãnas didelius mokslus aĩns Ut and so on.31 The similarity in meaning and the coincidence of their areas of use indicate close ties between NN and VD. The latter construction, unlike NN, is characteristic of many related languages. It is based on verb phrases with a nominal constituent (the so-called predicative attribute, or adjunct), apparently inherited from the period of Indo-European unity. All this seems to support the above-mentioned hypothesis of E. Tangl and I. Marvan, according to which NN developed from VD. However, other circumstances do not permit its acceptance. 1.7.2. In sentences that are not intonationally divided, such as sako jis dabar negeriąs Rs ‘he says he is not drinking now’32, we clearly distinguish two syntactic units: sako ‘he says’ and jis dabar negeriąs. ‘he is not drinking now.’ In this respect, utterances with participles fully correspond to utterances with finite verb forms, which often follow without a pause after verbs such as sakyti, girdėti, manyti, žinoti ‘say, hear, think, know’ and the like, e.g.: sako nebus vasaros ‘they say there will be no summer’ Krk; girdėjau seniai nebegeri ‘I heard you haven’t been drinking for a long time’ Vk; see also V. Sirtautas, ibid., 95. The verbs sakyti, girdėti ‘say, hear’ and the like tend here to lose their semantic and syntactic independence and approach modal words (cf. modal words that developed from verbs: matyt, mat, girdi ‘apparently, indeed, you see’). Yet there is no doubt that such sentences arose from two separate utterances, which are still often separated by intonation, cf.: sako, dabar jau namus statos ‘they say, they are already building a house now’ Prn. In exactly the same way, corresponding utterances with participles are separated by intonation, e.g.: sako, vyresnysai jau daktaru esąs ‘they say, the elder is already a doctor’ Kp; žmonės kalba, tu pačią primušęs ‘people say you beat her yourself’ Svn. The analogy with verbal utterances leads us to suppose that such sentences have retained an older semantic and intonational division than the sentences mentioned above that lack pauses. Exactly the same analogy with verbal utterances is also evident in sentences in which the subject of the actions denoted by the participle and the verb is the same: a) sentences not divided intonationally, e.g.: sakė neturinčios laiko ‘she said she had no time’ Krtv; pirmą sykį girdžiu turįs ponus ant savo žemėsŽemR₁ II 24; Janė sakė Petrą sutikus ‘Janė said she had met Petras’ Kp; žmonės kalbėjo matę Kraujelį ‘people said they had seen Kraujelis’ Vžns; sako porą tik stikliukų buvęs gėręs; sakė nešančios vežimų kraut ‘they say he had drunk only a couple of small glasses; she said she was carrying loads to put on the carts’ Vl (cf.: tėvas sakė eis valgyt ‘father said he would go eat’ Krk); 30On the other hand, VD constructions with passive participles (of the duodasi mušamas ‘he lets himself be beaten’ type) have also been noted by K. Sirtautas (KK 9.31) in places where NN is not characteristic, for example, around Šeduva, Radviliškis, and in Alksnėnai. 31V. Sirtautas, Kalbotyra XXV (1) 94–95, provides a collection of such examples and notes their rarity. In literary written language, where texts are punctuated with commas according to syntactic structure and meaning, sentences without a pause between verbum dicendi and a participle or another verb are in fact quite rare. In folk speech, however, they are even more frequent and widespread than corresponding sentences with pauses. This is shown by texts recorded directly from dialect speakers and by observation of folk speech. 32Similarly in SInt, Vvr, Rt, Grg, Šl, Šlu, Krš; Kp, Ut, Mlt, etc.; also with past active participles: Kv, Dr, Šv, Vn, Klp, Skd; Čk, Jrb, Trg, Pnv; Trgn, Jž, Kp, Skp, etc. 18 b) sentences divided intonationally: he says he is carrying hay as a hired hand Plng; the neighbor woman told [me] that she had ridden a cart from Vilnius Mlt; mother said that she had almost cried when they led [her] out like a cow Rk; [he] said [he] had fixed it well Šl; the child said he had met a wolf Km; [she] said she would come next time Šlu (cf. Ona said she had found boletus mushrooms Sv). After reflexive verbs such as sakytis, pasisakyti, pasakotis, girtis ‘to say, declare, tell, boast’ and the like, both intonationally divided and undivided expressions are also used, e.g.: he said he had no money Grz; he says he ate only because he was starving Mžk; cf. he said (that) he would go Ker, Užg, Kš, Vlkn, etc. If we assumed that NN developed directly from VD, the oldest forms would have to be considered monosubject constructions not divided intonationally, e.g., he said he had returned, and from these we would derive two-member sentences he said, having returned and he said, his father had returned. Such an assumption, however, would be contradicted by the general direction of development of the syntactic system, which is clearly evident in corresponding sentences with verbs. That the evolution of the sentence in Baltic, as in other Indo-European languages, proceeded through the synthesis of separate utterances rather than through analytical division is shown not only by the entire history of the complex sentence, but also by processes that took place within the sentence, such as the conversion of individual nominal sentences into sentence constituents and their adverbialization (cf. nominative-derived adverbs gana ‘enough’, gėda ‘shame’, gaila ‘pity’, etc.; see K. Ulvydas, LKK XI 7 ff.), as well as the transformation of utterances with independent meaning matyti, girdėti, žinia, aišku, žinoma, suprantama ‘it is seen, heard, known, clear, of course, understandable’ and others into parenthetical words, etc. It is very difficult to believe that the development of participial constructions proceeded in the opposite direction and that initially monolithic expressions of the type he said he had returned later split into separate, relatively independent units, even acquiring different subjects of the action. The formation of NN can also be explained in a more natural way, one more consistent with the evolution of syntactic structure, by recognizing the archaic character of the asyndetic complex sentences of the type he said, having returned and he said, his father having returned, widely attested in dialects. For this, attention should be paid to constructions in which participial nominatives have the same form and function as when they form the indirect mood. 1.8.1. In its principal structural features—the position of the finite verb and the absence of an auxiliary—the indirect mood corresponds first of all to past active participles used without a copula, which from the perspective of the modern language form present perfect forms, cf. father returned (: had returned: used to return: will have returned). The past simple in NN differs from these only in modal meaning. However, modal meaning often emerges only from context and situation, and at times it may be entirely indeterminate. Even after verbs of perception and information that indicate reported speech, a past active participle may be understood in either way, especially at the edge of the NN area, where the corresponding present active participles are used little or not at all, e.g.: they say he is already dead now Škn; that one, I heard, hanged himself Mlt; they say that the candles of those whose candles have gone out have altogether forgotten about the other world Vl; it seems the people have come from the graves Vrnv, etc. On the other hand, past active participles with a present-tense copula are sometimes also used in the Samogitian dialect to denote an indirectly experienced action, and thus correspond fully in meaning to NN forms, e.g.: there is talk that people have hanged themselves by that stream Rsn; they say he has left Kin; they say the cart has fallen in Šlu; that woman comes to say that the child has pulled up a carrot Kin; look, the hayloft has been there Akm; how much she has written! Skd; similarly in the writings of S. Vaišnoras, cf. saka ghie… iog Sunus Diewa buk beiktinu ir paslapczu pastatimu Diewa / ne wissa Swieta (178)… yra kentejes ir numires MT 178–178ᵃ. Therefore, the NN area 2* 19 can now be defined more clearly only by the limits of use of present and future active participles. 1.8.2.1. Present active participles are no longer part of the system of compound tenses in the literary language; therefore, when they occur without an auxiliary verb in the position of a finite verb, they are understood as NN forms. This state of affairs, however, is not ancient. First, present active participles with be- in the Samogitian dialect can be used to denote an ongoing action without having a clear modal meaning, e.g.: the cow is soft (milking), the milk is flowing Pln; I came home and she was walking Vkš; in that garden the grass is trampled like this—so trampled, and there such a little shepherd is grazing the sheep SlančP 13; cf. also N. Slizienė, LKK IV 70. That this usage is archaic is also borne out by examples from folk songs, whose rhythmic structure makes it possible to preserve syntactic archaisms more effectively, cf.: The little dawn is dawning, the little sun is rising, the little head is aching Sk (LKK III 84), what is that young man thinking, what is the young one pondering, forging a golden ring, and hammering my little name KlvD 163. Present active participles without the affix be- are also used, e.g.: a worker is needed at home, mother is ill Pj (LKT 114); she has two brothers, and two sisters are there Plng. Thus the affix be- has not been associated with the admirative meaning of participles since ancient times. Present active participles often have this meaning even without this affix, e.g.: He forgot to cross himself and, turning the other way, sees—the devil is standing there Slm; I look and look, I watch, and [there it is]—arrived Tln; meanwhile, participles with be- often also denote an action reported or learned from others, e.g.: I heard he had been preparing to mow Krtn; they said he was studying to be a tractor driver Vg. On the other hand, in the western part of the Samogitian and Western Aukštaitian dialects, present active participles are used with the auxiliary verb būti ‘to be’ and form a fairly complete compound-tense paradigm: present tense I am, you are, he/she is… going (masc.), going (fem.); past simple I was, you were, he/she was… going (masc.), going (fem.); future I will be, you will be, he/she will be… going (masc.), going (fem.); conditional mood I would be, you would be, he/she would be… going (masc.), going (fem.) (no forms of the past frequentative or imperative mood have been observed, but these too are entirely possible), e.g.: look how I am talking (=as I speak) Mž; he is ill Žd, Prk, Klp, Eig; that one is so knowing Skd, Kl; someone is in a hurry Šl; he was a little ill Sl, Klm; it will soon be raining Klp; when there is fog, Varniai will smell; he will be living with a worm when they send him to the front Šts; and if they themselves had been like that, nothing would have been lacking except swan’s milk Brs (LKŽ VII 77). Corresponding participial forms with the affix be- are also used, but only with an emphasized durative nuance, e.g.: my cow is very good for milking: she has filled up, and the milk will be flowing Pln (LKT 79); good thing you didn’t go: you would have been ill Eig (LKT 73); I would not be living, but they fixed things now Žr. More complex compound-tense forms are also occasionally formed with a present active participle, e.g.: if I had been capable, I would have gone Skd; cf. (heretics) klaydawo / kayp butu anas wienu noru ne sutwertu turis essus (bladžili / iakoby on mial mieć iednę wolą nie stworzoną) SP I 193₁₂. Past-tense forms are sometimes also used to describe an action that has begun but been interrupted; that is, in meaning they correspond to the past inceptive tense formed with the affix be-, e.g.: I heard he was leaving the house to go to him, but returned Trgn. 1.8.2.2. The archaic character of this usage is confirmed by linguistic monuments of the sixteenth and seventeenth centuries. In the writings of J. Bretkūnas and especially S. Vaišnoras, who represent precisely the dialects mentioned, present active participles with an auxiliary verb are used with the meaning of compound tenses, e.g.: Kaip (Ponas) ira milins Szmones, Wiss i io Schwenteiei ira tawa Rankoie BrB V Moz 33,3 (Wie hat er die Leute so lieb; dilexit populos); netarent ios miegi | bet 20 they abide and rule over the living MT 260a₂₁; O Kursai Sunui netikjs jra, ne reges giwatas MT 262a₁₇–₁₈; Wina schaknis (pabr., virš.: medis)… turreia tris schakkas ir buwa szellioens (iš -enti), auga ir szideia BrB I Moz 40,10 (und er grúnete, wuchs und blúete); Diewas jstate tq sawedima… kad dabar (žmonės)… prigimtaie Teisibe buwa tebturi; MT 221₁₃–₁₄; There will be a day (the body) does not decay MT 245₄; Many things are indeed said thus | as if one were praying to God | that they… might be saved MT (PM) 29a₁₁–₁₂ (see also N. Sližienė LKK IX 80). Such constructions occur somewhat less often in other linguistic monuments; cf. And now the Son of Man will be sitting at the right hand of God’s power DP 163₁₁ (bedzie siedzial W 13; erit sedens); Ah, what will be said then SG 71₇. In the writings of S. Vaišnoras and some other authors (M. Daukša, K. Sirvydas), the participle esąs, -anti is especially favored in constructions with an auxiliary verb, e.g.: whoever truly is MT 7₈; The world from the beginning is a liar MT 42₂₀; I would teach him… likewise is MT (BĮ) XVIII₁₁–₁₃; described… the being of the ancients | what was there MT 150₁₅; which would be a true Religion MT (LĮ) XXIV₁₂; That he would not be an evildoer | we would not have handed him over to you DP 163₃₃ (By to nie byl zloczyńcá W 279₁₉; si non esset hic malefactor); If this were of this world | my servants would nevertheless fight for me DP 164₉ (Być z tego świata bylo krolestwo moje W 280₅; si ex hoc mundo esset regnum meum)³³. This combination of a participle and an auxiliary verb may in turn be linked with other participles or adjectives to form a compound copula, e.g.: that the Church of Christ would be established | or held by some other Apostle… (37ª) therefore it should also be mentioned well in that regard MT (PM) 37ª–38; all those parts will be consecrated in one place | not meaning that they themselves would be holy, corrupted, and useless MT (PM) 36₁₄, 34. Present active participles with the prefix be-, when combined with the auxiliary verb būti, often form the same compound past-tense forms in J. Bretkūnas’s writings as do other present active participles. Here the prefix be- emphasizes and highlights the duration of the action, e.g.: And when he returns there, behold, he was still offering his burnt offering IV Moz 23, 17 (sihe | da stund er bey seinem Brandopffer LB); A man… brought back… the Midianite woman… while all the assembly of the children of Israel were weeping at the entrance IV Moz 25,6 (die da weineten für der thür LB; And Saul stayed at Gilgal (was still there) I Sam 13, 7 (Saul aber war noch zu Gilgal LB); Boaz was just coming from Bethlehem Rut 2,4 (Boas kam eben von Bethlehem LB); And they were just burying the firstborn IV Moz 33, 3 (und begruben eben LB). The same compound forms are sometimes also formed with the auxiliary verb tapti; cf. I went then and washed, and became seeing ChB Jon 9, 11. Apparently, the emphasis on duration by the prefix be- paved the way for the inchoative past-tense meaning, which becomes apparent when another event interrupts the ongoing action denoted by the participle 33 See also MT (BĮ) XVIII₂₀, (LĮ) XXIII—XXIV, 150₁₅, 150ª, 161ª, 181₂₀, 183–183ª, 269ª₂, (PM)14ª₁₄, 14ª₂₀, 15ª. Cf. N. Sližienė, LKK IX 80; J. Palionis, LLK 198. 34 Similarly MT 77₁₉, 151₁₅, 268ª₁₂, (PM)25ª₁₂–₁₃, 34ª₁₉–₂₁. Altogether, in S. Vaišnoras’s writings we find 22 constructions with esas, -anti and a finite form of būti, including 3 with present-tense forms and 19 with conditional forms. 21 action, cf. Tawa tarnas buwa beganans Awis sawa Tiewo, ir ateija Łewas BrB I Sam 17, 34 (hütet die schafe LB). Occasionally, in old writings, present active participles with the prefix be- are also associated with a present-tense form of the auxiliary verb, that is, they appear to form a kind of present inchoative tense (but without the meaning of an interrupted action), which is already missing from the paradigm of inchoative tenses in the modern language, e.g.: Tu awinélio newirk, kolei ira beszindans (prieg Pieno Augives) BrB II Moz 34,26 (wenns noch an seiner muetter milch ist LB; in lacte matris suae V); Sakie iems su dziauksmu didziu / Iog est begulis edziosu BrGD XLVII (p. 83), likewise Mž 193. 1.8.2.3. Thus, predicative present active participles, which in the contemporary literary language are used only as NN forms, may once have had no definite modal meaning. In the position of verbum finitum, they occurred alone or with the auxiliary verb būti, with which they formed constructions like those of past simple active participles. Present active participle forms with and without the prefix be- had identical syntactic properties and differed only in a stronger or weaker meaning of duration. Thus, constructions with present active participles were also close to NN, just as the so-called perfect-tense forms with past simple active participles remain close to it today. Evidence for reconstructing such a state can also be found in the system of the modern language. This includes, first of all, the incompleteness of the so-called inchoative-tense paradigm (present-tense forms are lacking), the absence of a general meaning (the meaning of the temporal relation between actions is usually borne by present participles only with past-tense forms of the auxiliary verb, and then only in a particular context; cf. N. Slizienė, LKK IV 67; LKG II 146), the use of present active participles with the prefix be- in the admirative sense, and their correlation with past simple active participles that lack this prefix, cf.: After three days the parents look out—and the whole flock of geese is crawling into their yard LTt III 296 and Look, what a big hazel tree has grown! Kp. 1.8.3. The relationship between compound tenses and NN is also indicated by the similarity of their areas of use. Compound tenses, like NN, are most widely used in the area of Samogitian and the western part of the Western Aukštaitian dialect region (which also includes the Lithuanian dialects of former East Prussia). In the northeastern corner of Lithuania, where NN is also alive, only perfect forms of the past simple and future tenses occur more frequently. In general, the area of distribution of NN roughly coincides with the area of most intensive participle use. The facts cited here suggest that active participle nominatives, when used in the position of verbum finitum, once did not differ from NN forms in their morphological and syntactic properties. What is now called the indirect mood was merely a special case of the predicative use of the participle nominative, distinguished from other cases by a certain nuance of modal meaning. 1.9. The predicative use of participles is a legacy of the distant past. Being forms of nominal origin, participles may already in the earliest stages of the development of Indo-European languages have formed nominal, or verbless, sentences35, of the type father tired alongside father good or father master. Initially, such sentences apparently differed in their use from corresponding sentences with verbs from the roots *es, *bhū, 35 Cf.: K. Brugmann, Synt. 69—70; J. Zubatý, Studie a články, II, Praha, 1954, 126 tt.; A. Meillet, MSL XIV. 1 tt.; Introduction à l’étude comparative des langues indo-européennes, Paris, 1934, p. 356—7; J. Bloch, MSL XIV. 1 56 tt.; E. Benveniste, BSL XLVI 27 tt.; Ch. Guiraud, La phrase nominale en grec, Paris, 1962, p. 152 tt.; W. P. Lehmann, PIE 115—6. 22 which could have a specific lexical meaning36. Over time, the old nominal sentences were incorporated into the system of verbal sentences as its members with zero copulas37. Nominal sentences with participles could not be isolated from this process, and brolis einąs | grįžęs-type utterances likewise had to attach themselves to the brolis yra | buvo | būdavo | bus grįžęs paradigm. However, the system of compound tenses that began to form on its basis never fully crystallized in Lithuanian, and many predicative participles denoting an action (especially present-tense forms) remained outside it, retaining to this day features of their earlier predicative use. 1.9.1. Even those participles that, from the perspective of the modern language, are regarded as equivalents of the present perfect often differ in meaning from compound forms with a copula. N. Šližienė, LKK XI 20, noted that active past simple participles without the auxiliary verb būti in the present tense are used “to create an artistic picture of the present,” e.g.: Jonas išsitiesęs kniūpščias pas krosnį ant moliuko; trumpa jam vietos, todėl kojas ligi kelių pastatęs aukštyn pagal sieną; galva antrame gale nusikišusi ilgiau suoliuko ŽemR I 86. However, creating an artistic picture only partly characterizes the specific use of such participles. More accurately, participles without an auxiliary verb are generally used to describe a static situation and the qualitative state of individual things. Present active participles are used in this way as well (usually with be-), e.g., O ans jau nuo ryto atsigulęs ir bedribsąs Rs. By contrast, participles with an auxiliary verb are more characteristic of narrative. This difference between the use of participles with and without an auxiliary verb is significant because it partly reflects the difference between old verbal and nominal sentences. In fact, sentences with a nominal predicate and no copula in Old Indian, Greek, and Celtic are used to express constant, “timeless” states, whereas forms with a copula generally denote changing processes related to time38. 1.9.2. Relics of the predicative use of participles are found in fossilized constructions in which an auxiliary verb form is entirely impossible. Many of them have a pronounced modal and expressive nuance. Instead of a finite verb, negative participial forms (especially neuter forms) are used to express negation emphatically, with a certain emphasis, e.g.: It will never be that I should have anything more to do with him Jnšk; It will never be that I go after him Ds (LKŽ I₄ 1215); I shall no longer live (= I shall no longer live) here—if only things would go well this time! Sdk. It’s raining. — As if it weren’t raining! Ds (LKŽ VII 589). Well, brother, without having swung it, without having taken it off… VaižgRR I 421. The dog would buy meat, only he has no money Baran (APh I 98). So you didn’t know; you should have answered him something Alk (LKŽ VII 833). The neuter form nematę is used in wishes that have become idiomatic expressions, e.g.: Kotas jo nematę! Ds (LKŽ VII 914); devynios jo nematę — jau išsibaigė (died) Lkm (LKŽ I 459); Bala nematę, nebeprikrsu to cukraus Slm (LKŽ VII 914); bala jo nematę Kp; Versmė ją nematę tos apkiuštelės Kp. The archaic character of such constructions is also indicated by the fact that the neuter form of the participle is used with the entire gender- 36 Cf. būti remnants of its concrete meaning in dialects: mes taip esam (gyvename) atklūdziai Žibiniškai (LKT 76); mes dabar būnam (= gyvenam) prie ežero Mrc; pradėjo kur (= visur) būt (= gyventi), dirbt Všn (LKT 300); žiniai pradės aukštyn būt (= augti) Ventė (LKT 144), see also LKKŽ I 1215. 37 Вяч. Вс. Иванов, Общ. 268. 38 See E. Benveniste, BSL XLVI. 1. 31 ff.; J. Humbert, Syntaxe grecque, Paris, 1960, p. 66; Ernout–Thomas, Syntaxe latine, Paris, ²1953, p. 146–8; P. Chantraine, TCLC IX 25; Ch. Guiraud, ib. 64 ff. 23 with nominal forms. In meaning, the participle here approaches the optative, which often expresses a wish (cf.: Grausmas jo nematai! Ds), whereas corresponding constructions with the indicative (galas tave nematė Dkš, velniai jo nemato Grž LKŽ VII 914) are clearly later. The optative meaning is occasionally also found in forms of other participles, e.g., Kad tu skradžemėn nuėjęs, kad tu velij prapuołus Gry (LKK XIII 122). In this respect, the frozen neuter form of the participle is especially characteristic in a formula repeated in a folk tale: Atadārą (tarm.: atadɔ.ru.) vaikėliai durytėlės BsP II 327, plg. atadɔ.ru. durelās Verksnionys LKT 299 ir Atidarai darai (duris), jūs mani vaikeliai Lt III 86⁳⁹; here the participle already has a clear imperative meaning. Participles that have retained a predicative function are also found in utterances in the form of a rhetorical question with kur or various cases of kas, e.g.: kur tu matęs, kad aš jai užsileisčiau! Vb (LKŽ VII 915), nu kur tu matęs: oru pavažiuot žmogų, nukris Joniškėčiai LKT 245; Na, kur nesimeldęs, anoks maldininkas KatkR 259; Ką tu belaukęs senatvės? ib. 190; Kur jis geriau ras, ko jis neėjęs? ib. 219; Žinau, kur nežinojęs ib 186; kame buvusi grėbėja! Snarglį pirmiau nusivalyk! Rdn (LKT 75). Some of these utterances have become wholly desemanticized, turning into idioms, e.g.: kur buvę, apsirgo pono dukte Daunoriai (LKK III 201); kur buvus nebuvus, ateis lapė Lt III 147; Kas tai bėsaš (“means nothing”) mums vainiką nupinti Šts (LKŽ I₁ 1217); Kas nebuvęs, plauksiu į jūrą Plng (LKŽ I₁ 1216). Similarly, participles (most often with the particle ar) are also used in constructions expressing alternatives, e.g.: O tai neklaũžada vaikas, ar jam sakęs, ar ne Slm; Ar sakęs, ar šuniu lojęs – neklauso, ir tiek Lnkv (LKŽ VII 661); tiek anam sakęs, tiek nesakęs: pro vieną ausį įein(a), pro kitą išein(a) Jnš; Kirtęs nekirtęs – kumelės stovi sau, galvas nuleidusios Žem (LKŽ V 855); compare also cases with a corresponding meaning in which the participle occurs alongside a verb: sakyк jam nesakęs – nuleidžia pro ausis Jnš; teipos ponas buvęs, ponas bus Ppl; bus kas esąs (nebuvęs) Als “a very big deal” (LKŽ I 1216); and emphatic constructions with a participle and a verb from the same root: augęs paaugęs žals pušynėlis pas tėtušio dvarelį Šl (LKŽ I 383). 1.9.3. Nominative participles can also occur in the position of verbum finitum, denoting in the sentence not the main action, but a secondary, subordinate one. Postpositive (with respect to verbum finitum) participles in constructions with the conjunctions kiek, kur, kada, iki, ligi and the cases of the pronoun kas correspond in meaning to verbs used in subordinate clauses of complex sentences, e.g.: vaikas… turėjo duonos kiek noris ValPŽK 4; žvejas… nuėjo per mišką, pats nesumanydamas, kur ir ko einas TMPs 27; jau neatmenu, kada beragavęs žuvies Lt III 377; liepė karalius pilną žąką primeluot, ik sakysias užriškit Jrk 46; jiems turėję vergauti lig atduodantys ar lig atdirbantys DB 169; nėr kas jam darbą Dgls; ji… ėmė pasakoti, ko ieškanti Lt III 290; gera būtų mergiotė, tik džiauna ką tik nutvėrusi Sml; see the author’s LKK IX 43–62. When following verbs of communication (especially verba percipiendi et dicendi), such constructions are close to NN, e.g., pagrįžę namo, (the brothers), nieku neprasitarė Eglei, ką padarę Lt III 299, and when they have a subject different from that of the verb, they coincide with it entirely, e.g., ir karčūninks (= karčemininkas) klausė, kur jis einąs Glbr (Mit VI 46). After verbs of other meanings, however, participles denote a secondary action without any NN overtones; cf. eina, kur budelis įsakęs Dr. Iš 39The formula’s archaic character is also confirmed by the ancient reduplication; cf. sut rinko trinko tėvelio grindos NS 341, išleido leido plonus tinklelius LTt 68; išveža veža mano bernužėlį LTt₁ II 58. 24 In constructions with conjunctions, the relationship between the meanings of participles denoting reported action and, more generally, subordinate action is clearly evident. The meaning of reported action characteristic of NN is secondary here, dependent on the lexical expression of the verbs, and forms that carry this meaning constitute only a certain subgroup of predicative participles. 1.9.4. Prepositive participles joined to a finite verb form by the conjunction ir likewise denote a secondary action and function in the sentence as equivalents of verbum finitum, e.g., vilkas iš baimės galą pabėgėjęs ir pastipo BsP IV 23540. Such constructions, characteristic of folklore and early written texts, often also have a modal meaning: as with NN, they often emphasize the unexpectedness of an action or surprise at it. 1.9.5. The use of past active participles in the position of a second verb is also associated with modal meanings (command, prohibition, negation, etc.) and expressive overtones: a) to emphasize the action of a main verb from the same root, e.g.: gerk gėręs, brolty, dūmok ir namo JD 72; guliu pasigulęs, kad ir neužmingu Plng41, similarly with negative verbs, e.g.: kad šoks nešokęs zuikis ir kudl LTt 343, sometimes also with a conjunction, e.g. eik kad ėjės, kad nė kuo apsiaut Gs (LKŽ II 771); b) to express an alternative (also with a verb from the same root), e.g.: sakyk jam nesakęs — nuleidžia pro ausis, c) to express a negative or undesirable action (with a verb from a different root), e.g.: tylėk neburbėjęs Vkš (LKŽ I 959)42. The cases of predicative participle use listed here have all the characteristics of syntactic archaisms. Within the system of Lithuanian syntactic devices, they occupy a peripheral, isolated position and do not fit any conventional type of participle use. They are generally found in phraseologically frozen or fossilizing expressions. In contemporary language they are unproductive and are disappearing; they are encountered only occasionally in colloquial speech or belles lettres, where they are stylistically marked. They are linked to NN not only by formal identity but also by modal and expressive overtones, which indicate a tendency to assign modal meaning to forms that are disappearing and no longer fit the changed system. All this suggests that the foundation of the indirect mood was constituted by an archaic use of predicative participles. This assumption is confirmed by constructions with passive participles. 2. CONSTRUCTIONS WITH PASSIVE PARTICIPLES 2.1.1. In Lithuanian, neuter forms of passive participles, with or without an agent in the genitive, often have a modal meaning: present tense (brolio) gyvenama, sėjama, past tense (brolio) gyventa, sėta. Compound past-tense forms are also used occasionally: (brolio) būta gyventa. Their use was aptly described by J. Jablonskis, who observed that with such expressions the speaker “seems to want to say that he has only just seen it, has only just noticed this trick. (…) as if he wants to report that the news he is now telling us was conveyed to him only by some incidental trifle that he himself has just noticed, or by another person” (RR I 465). At the same time, J. Jablonskis noted that such expressions are close to “narrative speech” 40 For a fuller discussion, see the author, Baltistica I (1) 53—66. 41 Cf. corresponding Greek constructions with participles from the same stem: φυγῇ φεύγειν, ἀκούσας… ἤκουσα and so forth; see R. Kühner, GrGr. 656. 42 See also J. Balčikonis, Baltistica III (2) 204; author, LKK VII 72—3, LKG II 353—4. 25 and are not used in all dialects: “The Samogitian expressions ‘There was a house here’, ‘People slept in the forest here...’ are replaced by their own expressions such as ‘There used to be a house here; it stood here’, ‘The forests slept here’... (RR I 466; cf. also ib. 530). Later grammars either did not examine the modal meaning of the neuter forms of passive participles at all, or were content with a summary of J. Jablonskis’s ideas. More attention has been paid to the voice meaning of these forms. It has repeatedly been noted that neuter forms of passive participles are often formed from intransitive verbs and lack the meaning characteristic of other passive forms1 or correspond to active-voice forms2. The relation of neuter forms of passive participles to other passive forms was recently discussed by V. Sirtautas, Kalbotyra XXI 71 ff., and E. Geniušienė, ib. XXIV (1) 27 ff.; they concluded that, in terms of syntactic features, constructions with genitive neuter forms fully correspond to the passive. The passive origin of the neuter forms of passive participles is well argued by the authors cited; it is entirely clear that they should be examined as a particular group of passive forms. Yet their specific meaning and synonymy with active-voice forms, especially with active participles of modal meaning, are equally clear. It is precisely because of this modal meaning that constructions with neuter forms of passive participles have recently sometimes even been classified as NN3, albeit inconsistently and with some vacillation4. What in fact prevents these forms from being included in NN is the incompleteness and indeterminacy of their paradigm, their close relation to the passive, and their similarity to compound passive tenses. 2.1.2. On the basis of their modal meaning, the present- and past-tense neuter forms of passive participles can contrast with the indicative present and simple past forms, e.g.: brolio gyvenama: brolis gyvena (cf. brolis gyvenąs); brolio gyventa: brolis gyveno (cf. brolis gyvenęs). However, in formal terms they entirely correspond to compound present-tense passive forms with a zero copula5. They may be understood both as equivalents of brolio yra gyvenama / gyventa and as expressions with a certain modal meaning, depending on context and situation (cf. žiūrėk, kaip čia tavo brolio gyvenama!). Corresponding neuter forms of transitive verbs are equally ambiguous (e.g., čia jo miežiai sėjama / sėta). Only the construction brolio būta gyventa—contrasting with the indicative pluperfect form brolis buvo gyvenęs (cf. brolis buvęs gyvenęs)—has no equivalent in the system of indicative compound tenses. However, it is rather isolated, since constructions with present-tense participles (e.g.: *brolio esama gyvenama / gyventa, *brolio būta gyvenama), which could contrast with other indicative compound tenses, have not been attested in the dialects and could at most be used accidentally and sporadically6. Moreover, in many places the meaning of the construction būta gyventa is conveyed by the regular compound passive past-tense form buvo gyventa, which correlates with (yra) / buvo / būdavo / bus gyventa, cf.: 1 P. Trost, LKK III 220 ff.; N. Sližienė, LKK IX 74, LKG II 152. 2 W. K. Matthews, SEER XXXIII 81, 352 ff.; A. Paulauskienė, Veiksm. 157. 3 J. Balkevičius, Sint. 95—6; N. Sližienė, LKG II 168; LKK XV 78; P. Gailiūnas, LKG II 59; E. Geniušienė, Kalbotyra XXIV (1) 31. 4For example, P. Gailiūnas, Kalbotyra IV 76, does not classify passive participles as NN, while in LKG II 59 they are mentioned only briefly, in a single sentence at the end of the chapter. N. Sližienė also omits them from the conjugation paradigm of compound verb forms in LKG II 173—8. 5 On these, see N. Sližienė, LKK IX 74, LKG II 53 ff. 6 J. Balkevičius included the constructions esama mokoma, esama mokyta, būta mokoma in the table of predicate expression, Sint. 95, but gave no examples of them. 26 we had already gone to bed when he came Kp; I had fallen asleep, you see Mlt; I did not buy it for nothing; it had already been brought Slm; his berries had already fallen in Brž; he had already left Alz, Aln, Mlt, VIkv et al.; Aneliutė had come; I did not find him; he had already left Aln; then I understood what had been left here Kp; someone had sown rye Auk; writing was forbidden Stlm (LKT 310); I had served for twenty years; here they had gone (the cows) Klp; everything had been done at my place: ploughed, cooked, washed Rgv; earlier, at that time, the hay had already been cut Trg; you will go, and it will be finished Mžš; we will already have gone Ds; Do not go; he will already have left Kp; perhaps it really will already have been brought Mlt; that horse will have been stolen and sold by him Kp. Even with a present-tense copula, the neuter singular form of the passive participle is often used with the same meaning as in constructions without a copula, e.g.: apparently, people have lived here at some time Sk; I have fallen down Sr; the chest is congested; medicine has been drunk Nmš; look, the hares have scattered to the apple trees Vkš. On the other hand, constructions without a copula may also lack more distinct modal nuances of meaning. For example, such expressions widespread in the Aukštaitian dialects as the neighbor has threshed, they have left, said, done… mean the same as the neighbor has threshed, they have left, said, done, and the participles may be regarded as neuter compound present-tense forms with a zero copula, cf. the grain is threshed / has been threshed. The agreeing form of the passive participle is occasionally used alongside the copula būta, e.g.: the money had been steeped in sour milk in the church cellar Neršionys (LKT 310). In Samogitian dialects, where feminine forms are used more widely in place of the neuter, they are sometimes found in the constructions under discussion as well, e.g.: their institutions are full (of teachers) stuffed; how many of those roots have been peeled Pj⁷. 2.1.3. Thus, the neuter forms of passive participles (unlike the nominative forms of active participles discussed above) do not constitute a separate paradigm that could correlate with the paradigm of the indicative. In terms of structure and grammatical tense meaning, they are affiliated with compound passive tenses⁸. There is no grammatical basis for treating them as forms of the indirect mood, since separate morphological categories can be formed only by specific systems of forms that contrast with others in terms of certain generalized meanings. The neuter forms of passive participles also do not form a clear correlation with the NN (they cannot be regarded as NN passive forms because of their meaning) and do not fit into it. The modal meanings these forms most often (though not always) have in a sentence are not sufficient grounds for distinguishing them from compound tenses as a separate mood, since in many cases these meanings are determined by context; moreover, they can also be expressed by other compound-tense forms, used with or without an auxiliary verb. For example, the meaning of an action experienced unexpectedly and prompting astonishment (i.e., the admirative) can also be expressed by the masculine and feminine nominative forms of passive participles: he goes in and looks—his brother is nailed to a block Grš; the girl comes to bring oats to the geese; she sees that all the geese have been slaughtered Norvaišiai LKT 244 (cf.: look, the hens have already gone into the house Škt LKT 159, see 1.2.1). Compound future forms with active and passive participles very often express a doubtful or conjectured action, but they are not therefore considered a separate mood (cf. a monster 7 Cf. also clear compound passive-tense forms: from of old, a plot of land had been left from the manor estate Žr; everything has already been cut, hauled in, and threshed Als; one must live out however much God has ordained Trš. 8 E. Geniušienė also discusses them as separate instances of the passive-voice system, Диате́зы и залоги в современном литовском языке, – Типология пассивных конструкций, Диате́зы и залоги, Ленинград, 1974, p. 209 ff. 27 he will have stolen it Tlž; the people of the household will have gathered and come up with that name SmIn; he will surely already have ridden away Brž; perhaps it will already have been brought Kp)⁹. All this shows that Lithuanian has not special neuter forms of the passive voice in the indirect mood, but rather a modal use of passive participles attached to compound tenses. 2.2.1. The constructions with neuter passive-participle forms closest to compound tenses are those such as rye is being sown | has been sown, e.g.: in the evening, the rope is being twisted KzR; potatoes are already being dug Ds; the little holes are full of honey that has been poured in Ad; the vimba are fastened to the beams Šk; the doors are open and left that way KzR, and also in Rk, Švnč, Gdr, Jnš, Tršk, Trk, Vv, Grl, Zp, Plv, etc.¹⁰; similarly in old texts, e.g.: Tulszis… su Vksūsu inpilt buwa SG 21₅; latru kaulai lauzsit buwa BrG 101, see A. Bezzenberger BGLLS 159–160. These constructions differ from instances of the rye is being sown | sown type in that the participle here does not agree with the nominative of a noun (pronoun). That nominative may also be unexpressed; in that case, the construction has a clearly impersonal character, e.g., (now) it is being sown | has been sown11. (Impersonality is also evident in constructions with a nominative when the neuter form is placed first, cf. being sown | sown is the rye.) In their function and syntactic relations, the neuter forms of passive participles here correspond to neuter adjectives, cf. beer is tasty and tasty, used interchangeably with agreeing masculine and feminine forms beer is tasty. Admittedly, this parallelism in the present-day language is only partial: not all neuter forms with a stative meaning can be replaced by agreeing adjective forms, and vice versa¹²; their differences in meaning and use must have become especially pronounced when neuter forms developed into predicative words having mainly a stative meaning. In the literary language, non-agreeing two-member constructions with neuter forms are disappearing, but they are still well preserved in the eastern part of the Aukštaitian dialect area, and their Baltic character is confirmed by clear Old Prussian equivalents, e.g.: erains bou sei poklusman (=Jēderman sey vnterthan) 57₉, stai perōni… ast poklusman (die Gemeine… ist vnterthan) 65₁₆¹³; also with past-tense passive participles, e.g.: stai grikai ast… etwierpton (die sünde seien… vergeben) 43₂₁; stai gannai bousei pomeston (Die Weiber sein vnterthan) 65₁₂¹⁴. In Latvian impersonal constructions we already find masculine passive-participle forms, used similarly to the feminine forms in Lithuanian Samogitian, cf. Latvian nav gulēts ‘there has been no sleeping’, bij ilgi domāts un smagi iets15 and Samogitian Lithuanian as it is commanded, so work Als, everything already cut, hauled in, threshed Als, from of old a plot had been left Žr, one must live out however much God has ordained Trš. The construction type is preserved here, only one form having been replaced by another, because the neuter gender has become more greatly reduced. The regular distribution of non-agreeing two-member and one-member constructions throughout the Baltic languages shows that they do not— 9 See also N. Sližienė, LKK XI 34; LKG II 158. Such future-tense forms are widespread even in dialects where forms of other compound tenses are hardly found. Simple future-tense 1st-person forms are also used with potential, optative, and imperative meanings; see A. Jakulis, Lietuvių kalbos būsimojo laiko istorija, cand. diss. manuscript, Vilnius, 1967, 121 ff. 10 Cf. also K. Brugmann, LB 319; J. Schmidt, Pluralb. 162 ff.; J. Endzelīns, ИОРЯС XXII. 2 190 ff.; J. Jablonskis, RR I 464 ff.; Ch. S. Stang, VGr. 265; Z. Zinkevičius, Dial. 268. 11 From the perspective of present-day Lithuanian, such constructions are considered impersonal compound forms of the passive voice; see N. Sližienė, LKK IX 74; V. Sirtautas, Kalbotyra XXI 72; E. Geniušienė, Kalbotyra XXIV (1) 27 ff., with literature. 12 J. Jablonskis, RR I 465; A. Valeckienė, DB 555. 13 See also R. Trautmann, AprS. 244; J. Endzelīns, AprGr. 100. 14 For more examples, see R. Trautmann, ibid. 259; J. Endzelin, ibid. 199. 15 A. Blinkena, MLLVG II 533, W. K. Matthews, SEER XXXIII 360. 28 may have arisen through the erosion of adjective and participle inflection or under the influence of other languages16, while their explanation by the substantival use of neuter adjectives17 is precluded by the much lower productivity of such usage. Thus, there is no basis for treating the indeclinable constructions of the Baltic languages as newer, having developed from agreeing ones. 2.2.2. Counterparts to the Baltic indeclinable constructions18 are found in numerous Indo-European languages; cf. Old Church Slavonic skorpija sěklivo; ugodno… krotostb; Old Russian епвхъ сладко а человѣкъ падко, Gothic þannu gatauran ist marzeins galgins Gal 5, 11 “ergo evacuantum est scandalum crucis,” Greek οὐχ ἀγαθὸν πολυχρονίζειν, Latin mors omnium rerum extremum and so on19. Neuter forms of participles in the indeclinable constructions of the Slavic languages are used in exactly the same way as in Lithuanian, from the earliest monuments, that is, before the disappearance of the inflection of non-pronominal participles, e.g.: Old Church Slavonic glasъ tryby uslyšano budeť, Old Russian головы поимано (1314), Old Czech města bylo dobyto20. The absence of agreement here likewise cannot be explained by later syntactic processes21. These constructions are relics of earlier epochs in the development of the sentence. Historical and typological studies of Indo-European sentence structure suggest that during the period of Indo-European unity, predicative nominal forms expressing a state were invariable, indeclinable, and probably corresponded to bare stems22. It is precisely such bare-stem forms that Baltic and Slavic neuter adjectives and participles in *-o have in the constructions under discussion23. Thus, Lithuanian neuter forms of non-pronominal participles functioning in the verb position were probably associated with nominals of various genders from the time when the present-day adjective inflection had not yet developed, and “bare nominal stems used in the predicative function were not agreed with the subject” (J. Kazlauskas, Ist. gr. 304)24. Bearing in mind predicative adject- 16 The influence of German, which Ch. S. Stang sees in VGr. 179 even in constructions such as Old Prussian sta ioumas ni ast labban 57,7 (=Lith. tai jums nėra gera), in our view, could only have stimulated the use of neuter forms with masculine and feminine nouns and pronouns in Old Prussian. 17 K. Brugmann, Synt. 174–5. 18 Cf. E. Fraenkel, Baltosl. 38. 19 F. Miklosich, VGr IV 29; Ф. И. Буслаев, Ист. Гр. 446; А. А. Потебня, Зап. I–II 114 tt.; Ш 358 tt.; B. Deibrück, VS. Ш 247–248; И. Эндзелин, ИОРЯС XXII 192 tt.; O. Lagerkrantz, Präd. 19–20. 20 F. Miklosich, ib. 365; А. А. Потебня, Зап. Ш 359 tt.; В. И. Борковский — П. С. Кузнецов, Ист. гр. 423 tt.; Ф. П. Филин, Происх. 495 tt.; J. Bauer, OSS II 334. 21 The view that Old Russian indeclinable constructions are connected with the loss of participial inflection (П. С. Кузнецов, МИРД, III 71 tt.; С. М. Обнорский, Очерки 159; З. М. Петрова, УЗ ЛПИ ССП,36 tt.) or with their later origin from verbal utterances (А. А. Шахматов, Синт. I 119, Т. П. Ломтев, Очерки 209 tt.) is contradicted by the presence of indeclinable constructions in the earliest monuments (cf. Ф. П. Филин, ib. 499), their spread across various Slavic languages, and clear parallels in other Indo-European languages. On the archaic nature of these constructions, see also О. С. Мельничук, Вступ. 383; Розв. 142–3. 22 Cf. W. Lehmann, PIE, 69 ff., 199 ff., 208 with literature. 23 On the antiquity of the neuter nom.-acc. form in *-o, see S. Agrell, Neutr. 18; T. Milewski, Indohitt. 22–23; V. Mažiulis, Balt. 84 ff. with literature; J. Kazlauskas, Ist. gr. 124. The traditional view that the final -a of Baltic neuter forms (gera, eita etc.) reflects < *-od (R. Trautmann, AprS. 244, J. Endzelīns, ИОРЯС XXII 192, AprGr. 99 ff., etc.) is less plausible, if only because it isolates o-stem nominals from other nominals whose neuter form has long been regarded as a bare stem (cf. K. Brugmann, Grdr. II₂ 144). The Baltic and Slavic data do not show an inflectional -d/t, nor do the neuter nominals once borrowed from Baltic into the Finno-Ugric languages; see J. Kazlauskas, Ist. gr. 123; V. Mažiulis, ib. 85. 24 The antiquity of the use of predicative, indeclinable neuter forms (e.g.: Lithuanian šalta, salta, gražu, Old Church Slavonic sladko, Gothic filu) was emphasized by O. Lagerkrantz, Präd. 12, 19 ff., who proposed that originally every Indo-European predicative adjective had a neuter singular form, with gender and number distinctions emerging later. 29 the archaic character of the forms of pronouns and indeclinable participles, and the indisputable parallels in usage in related languages, make it possible to regard the model (alus) gardu, as well as its variant (rugiai) sėjama / sėta, as Indo-European²⁵. 2.2.3. It is not difficult to see that this model represents nothing other than a nominal sentence with an uninflected main constituent (A. Valeckienė, DB 557). Thus, when examining the use of indeclinable participles in the neuter gender, we once again arrive at the nominal sentence on which constructions with active participles are based. As already mentioned, in these constructions the active participle is most often agreed with the nominative expressing the agent. However, agreeing constructions are not the only ones and are evidently not the oldest. That in the past the main constituent of a nominal sentence could have been an unagreed active participle lacking special inflection becomes clear from the use of its neuter forms with nouns and pronouns of both masculine and feminine gender, e.g., broliai | seserys sėja, sėję, sėdavę, sėsią. Such usage, preserved across a large part of the Lithuanian-speaking area²⁶, cannot be recent, since the gender opposition is well maintained in the inflectional paradigms of both adjectives and participles; it can be explained only as a syntactic archaism, assuming that “initially the neuter singular form could have been associated with a subject of any number and any gender” (J. Endzelīns, ИОРЯС XXII 198)²⁷. Alongside two-member expressions of the sisters went type with neuter forms, we also find one-member ones, e.g., (today) sninga, (formerly) snigė, snigdavę, (tomorrow) snigsią (cf.: sako par jum visą vasarą lyją Rk; kur tau neliję Ds; jau buvo praušę, kai sukilo Grž LKŽ I, 515; sniegutėlio pasnigę Lz LKT 395). These forms are used with impersonal meaning also in various constructions with conjunctions, e.g.: nėra kas daro, kur eina and so on; see also 1.9.3. They are also found in sentences expressing the presence of certain things, e.g.: čia vabalų esą Mž; vilkų esą Rk; dabar da teip tebesa Rk; rugių nesą Slm, Vkš; labai mėlynių buvę Kp, as well as with neuter adjectives, e.g.: aje gi kaip gražu būdavę Ut, cf. Jų darbai bei buriškos budavonės tikt bjauru esą bei gėda žiūrėti DonR 55. Such sentences, from which esą apparently developed as a modal word (cf. esą, jūs turit išeit Šk), are equivalent in meaning to sentences with neuter forms of indeclinable participles, cf.: kumelei nieko nešama Ds; kitiems pasaulyje esama dar blogiau LKŽ III 740; kiek čia esama nelaimių VaižgRR II 115; šiąnakt vilkų būta Plš; seniau ten būta grybų Mlt. The syntactic position of neuter forms of active and passive participles is entirely identical here, and the two can freely alternate. Thus, the constructions of the simplest and apparently oldest type, with neuter forms of active and passive participles, essentially coincide in structure. Both are based on nominal sentences inherited from the period of Indo-European unity; it is only in constructions 25 The antiquity of the model (alus) gardu is also confirmed by the use of neuter adjectives in Hittite, e.g. aššu ‘good, well’, mekki ‘much’; see E. Benveniste, Hitt. 64. 26 See Z. Zinkevičius, Dial. 378–9, 381–2, and the card index of the Atlas of the Lithuanian Language. Neuter forms of active participles in the past definite tense are particularly widespread; they did not become standard in the literary language, apparently only because of paradigm levelling and the influence of the written language of Lithuanians in East Prussia, especially F. Kuršaitis’s grammar; see V. Rimsas, Kalbotyra XXIV (1) 80. 27 Likewise, the neuter form of active participles in the past tense is now used in the Lazūnai, Gudeliai, Kabeliai, and some other dialects on the margins of the Lithuanian-speaking area, e.g.: tį boba būvo až pečių pruklaupį, vilkas išgąsti, rasizōji guli Lz (E. Grinaveckienė, LKK III 200); viškia pabuvai: ir pavalgi*, ir alkanā Gdlš (LKT 369). However, the generalization of the neuter form here may also be later, arising from the fossilization of participles (cf. similarly fossilized semi-participle forms in Gudžiūnai, LKT 368, Zietela, LKG II 378, etc.). 30 with passive participles, their older structural model (i.e. with unagreed neuter forms) is better preserved and more generalized. 2.2.4 The subsequent development of constructions with neuter forms is connected with the incorporation of passive participles into the verbum finitum system and the formation of the passive. In the Eastern Baltic languages, the passive has not everywhere become fully established. In some Lithuanian dialects (e.g. Southern Aukštaitian and some Eastern Aukštaitian dialects), it consists only of impersonal constructions with neuter participles. However, passive forms with masculine and feminine participles are widespread in Žemaitian and Western Aukštaitian, as well as in some Eastern Aukštaitian dialects. They have been used in many sixteenth- and seventeenth-century texts from the very beginning of written records. As the passive took shape, the adnominal genitive, long used in possessive constructions, acquired the meaning of an agentive genitive (genetivus auctoris)28. The use of the genitive with *to-suffix participles that have or acquire a passive meaning (especially those close to nouns) is also attested in Old Indo-Iranian29, Greek30, and perhaps Old Persian31, but nowhere was it developed as fully as in Lithuanian. The gen. auctoris was generalized here with passive participles of both transitive and intransitive verbs, and alongside namas tėvo statoma constructions of the type tėvo sėdima spread. In languages where the agent of a passive participle came to be expressed by other cases and prepositional constructions (also mostly of possessive origin), these were sometimes likewise associated with neuter forms of passive participles; cf. Latin dictum est a te, Old High German iz ist gescriban fona dir (P. Trost, LKK III 221). Such usage became particularly widespread in Russian dialects, where the expression of the agent varies greatly: alongside the predominant construction Y + genitive, e.g.: мпу овцы у волка зарезано (В. А. Матвеенко, МИРД II 125), the agent is also found in the instrumental (деньги тестям подарено ib. 128), with от+genitive (от Петьки писано ib. 130), and so forth. In such cases, neuter participle forms are also made from intransitive verbs, e.g. у меня в хозяйство вхожено32. Although the expression of the agent differs, the typological parallelism of these constructions with Lithuanian ones is evident. Comparing them, W. K. Matthews, SEER XXXIII 365, concluded that in both language groups they were based on impersonal sentences with neuter forms of passive participles. 28 On the relation of the gen. auctoris to the possessive genitive, see K. Brugmann, Grdr. II₂ 601; B. Delbrück, VS. I 348; E. Fraenkel, Kas. 95 with references; Baltosl. 37 with references; H. Pedersen, KZ XL 152. On the gen. auctoris in Latvian, see J. Endzelīns, LaGr. 561. The possessive origin of the Lithuanian gen. auctoris is also shown by the possessive pronoun forms used in this function: mano, tavo atnešta (ne manęs, tavęs), cf. E. Fraenkel, Kas. 96; V. Žulys, Baltistica V (2) 170. 29 K. Brugmann, ib.; B. Delbrück, AiS. 153; VS I 348; E. Fraenkel, Baltosl. 37. 30 K. Brugmann, ib. 601; LB 321. 31 See H. Hübschmann, Cas. 270, B. Delbrück, VS. I 348; E. Benveniste, BSL XLVIII 52 ff. with references. The constructions of the type manā kṛtam found in Old Persian correspond exactly in structure to Lithuanian māno kūrta (gen.-dat. + neuter form of a *to verbal adjective), but the passive character of the *to forms used in them has recently been debated: E. Benveniste, ib. 52–62, regards it as a verbal noun and the construction itself as possessive; in contrast, G. Cordona, Lg XLVI 1–12. The difference of opinion apparently arises from the insufficient differentiation of this construction from possessive ones, that is, from a very early stage in the development of the passive; D. I. Edelman, ВЯ 1974. 1 23–24, regards them as passive-possessive, unlike M. M. Guchman, Эпр. 71–72, who sharply distinguishes the possessive construction from the passive one. On the origins of the gen. auctoris in other Indo-European languages, see G. Cordona, ib. 2–3. 32 Н. Б. Кузьмина – Е. В. Немченко, Синт. 109; pp. 17–116 there discuss in detail passive constructions with neuter forms in Russian dialects, with references. 31 (e.g., mak būtum ne bŭtmo, cf. Lithuanian taip nebūta, čia visai nelyta lmb), which were later extended with different cases. This conclusion should be supplemented: in both Lithuanian and Russian, not only impersonal but also two-member personal constructions with an uninflected participial form should be regarded as primary, of the type rugiai pjauta or bambka ubūto33. By their structural type, the constructions examined here resemble ergative constructions, in which the agent of an active action is expressed by an oblique case and the object by the direct case (nominative)34. Since many scholars also reconstruct an ergative or ergative-like sentence structure for the Proto-Indo-European period35, it is sometimes suggested that constructions with neuter forms of uninflected participles may be relics of that old, pre-nominative sentence structure36. However, the Lithuanian genitive, like the Russian instrumental, in these constructions is conditioned by the passive form of the uninflected participle and is inseparable from other instances of the use of the genitive (respectively, instrumental) auctoris, while prepositional phrases with oblique cases in Russian dialects (e.g., у него ушло) are probably of even more recent origin37. Only nominal constructions that once laid the foundation for the passive construction with an oblique case38 could be considered a relic of remote antiquity. The convergence of Lithuanian and Russian passive constructions with ergative ones, meanwhile, is apparently just as much a consequence of later development as the transformation into ergative constructions, in Old Persian39 and Armenian40, of inherently nominative constructions with verbal nouns bearing the suffixes to and to. Nor do we have any evidence to suggest that the nominative case with a neuter form of an uninflected participle could once have functioned as an agentive nominative with a verb, i.e., that constructions of the type rankos kelta could have meant “the hands raise”41. It would be wholly incomprehensible and inexplicable for such a nominative to have become an agentive genitive. 2.2.5. Differences in the expression of the action’s agent and variation in oblique cases show that in the Eastern Baltic and Slavic languages, constructions with neuter forms of uninflected participles developed independently. However, the parallel development of these constructions in both language groups proceeded in a very similar direction, although under the influence of different morphological processes. 33 They are sometimes extended by an instrumental agent already in Old Russian monuments, e.g.: медъ дано бысть богомъ (in a text of 1076; see Ф. П. Филин, Происх. 495). 34 Because of this typological similarity, W. K. Matthews simply calls constructions such as jų laikoma arklys and the like a Lithuanian “ergative” (ibid. 352, 364); V. S. Khrakovsky makes a similar observation about Russian, ВЯ 1972, 5 35 ff. 35 Plg. C. C. Uhlenbeck, IF XII 170–1; KZ XXXIX 600 tt.; H. Pedersen, Hitt. 83 tt., A. Vaillant, BSL XXXVII 93 tt., E. Laroche, BSL LVII 23 tt.; B. B. Иванов, Общ. 51 tt., A. Н. Савченко, Эрг. 74–90; Ср. гр. 364 tt.; V. Mažiulis, Balt. 79–80. 36 Plg. G. Janke, Pass.; M. M. Гухман, Эрг. 68 tt. 37 П. Я. Черных, Ист. гр. 281–2. Since the construction у + genitive spread in Karelia and the northern Russian dialects, W. Veenker, Substr. 137 ff., conjectures that it was stimulated by Finnish constructions with the adessive. 38 Cf. M. M. Гухман, Эрг. 72; for Lithuanian, A. Valeckienė, DB 556–7. The use of the genitive with verbal nouns (nomina actionis) to denote the agent is also considered archaic; see Э. Бенвенист, ОЛ 163–4; В. В. Шеворошкин, ВЯ 1957, 6 89; conversely, C. Watkins, “Remarks on the genetiv,” – To honor R. Jacobson, II, The Hague, 1967. 39 Д. И. Эдельман, ВЯ 1974. 1 23–24; А. А. Пирейко, Эрг. 135–141; М. М. Гухман, ib. 71–2; žr. dar 31 išn. 40 A. Meillet, MSI, XI 375 tt., 385; Altarm. 74, 114; H. Pedersen, KZ XL 152 tt. 41 J. Kazlauskas, Ist. gr. 332–3; cf. A. Girdenis and V. Žulys’s criticism of this view, Baltistica VIII (2) 201. 32 2.2.5.1. In Lithuanian, the disappearance of neuter-gender nouns must have played a major role here42. With the disappearance of the neuter gender in nouns and most pronouns, the use of neuter-gender adjectives and participles was largely restricted to the predicative function (the nominal use of neuter-gender forms is comparatively rare and unproductive). P. Trost convincingly demonstrated that under these circumstances the neuter-gender forms of adjectives and passive participles, having lost their morphologically expressed connection with nouns (pronouns), must have become impersonal substitutes for verbum finitum — predicatives43. In the sentence they occupied a central position, while the nominative nouns (pronouns) associated with them came to occupy the periphery. In this way, two-member non-agreeing constructions came closer to one-member impersonal constructions. In present-day Lithuanian, this process of regrouping syntactic relations has not yet been fully completed or morphologically formalized. A noun or pronoun associated with a neuter-gender passive participle and denoting the object of a transitive verb’s action has the same nominative form as in other passive constructions, e.g.: already the rye, too, is being sown Pnd; the livestock is fed in the morning Ds; the potatoes were dug up long ago Šl; look, the money has been stolen Šmn; have the beets been hoed? Grž, and so on (cf. the rye is sown, the livestock is fed, the potatoes are dug up, etc.); the agent of the action may also be expressed here by the agentive genitive, e.g.: their horse is kept LKŽ VII 36; the potatoes were baked by those who stayed at home VaižgRR I 151; our three rows have been ploughed here Vlk; she was paid a ruble Ktk; they had grain sown here Prn; he has also bought a little horse Vs (LKT 381); I’ll go — the loops for you have been set up by me Rš. The two-member nature of the construction becomes clear when the object is a neuter-gender pronoun in the nominative, e.g.: that has been eaten by him Švnč, I have forgotten everything; I remember nothing Ssk, or when it is a synonymous masculine form, e.g. everything has been thought up by me Glv, cf. that (everything) is good. Nevertheless, constructions with neuter-gender forms of passive participles of intransitive verbs and an agentive genitive are now the most commonly used, e.g.: the soil dries and dries out Ds (LKŽ II 744); who lives here? Pl; and he, too, is going to the market Ktk; the cow has gone into the barn Vrn; he got up a long time ago Kt; I have lived everywhere Alk; he has fallen asleep / slept Jnš, Rk, Kš, Kn; the dog ran here Prn, Vlkn, Šlčn, Rk; he went into the cowshed and hanged himself Šd, you ran home Ob; there was a bird in the tree Vb; it was rained on Dj. The agent’s genitive (in this case, the genitive of the subject of a state) also occurs with neuter-gender forms of the passive participle of the verb to be, e.g.: how many crows there are! Plš; and who has been there? Dbg; it has rained Eiš; he has been there and gone Slm, Kp; it also occurs in emphatic, fossilized constructions such as there was a man — a shepherd, that’s all! Lp (LKŽ I 995), sometimes together with a second genitive corresponding to the predicative constituent in the active construction, e.g.: the wind was strong at night Ob; the machine was full Rdm. Such constructions denote certain states and are close to impersonal ones. In them, the genitive of a personal pronoun sometimes indicates not so much a person as a location; for example, expressions such as we have already finished threshing, sowing, and digging up, and the like, over a large area of the eastern Aukštaitian dialects mean approximately the same as threshing, sowing, and digging up have already been completed at our place (or around us). ⁴² Cf. B. Delbrück, VS I 407 ff. ⁴³ In this respect, P. Trost’s view differs from that of W. P. Matthews, who derives all constructions with neuter-gender forms of passive participles from impersonal constructions. While convincingly tracing the development of two-member constructions with passive participles toward impersonality, P. Trost does not address at all the inherently one-member impersonal constructions, whose archaic character, in our view, is beyond doubt. 3. Užs. Nr. 1026 33 Since constructions with participles of intransitive verbs are the most productive, syntactic relations in constructions with neuter-gender forms of passive participles of transitive verbs are also being reorganized on their model. The nominative, formerly one of the principal constituents of a two-member passive construction, is now becoming dependent on the impersonal passive-participle form that constitutes the sentence’s nucleus, and the expression a letter is being written by him, once equivalent to a letter (is) being written by him, acquires the same meaning as (he) is writing a letter, cf. guests are being awaited by him (not guests in the nominative!). In other words, what was the sentence subject becomes the object, and the two-member personal construction becomes a one-member impersonal one⁴⁴. Therefore, in these constructions the nominative case is sometimes already replaced by the accusative of the object, e.g., honey is climbed for from tree hollows Lp⁴⁵, while the genitive form of the object is already entirely regular, e.g.: she was awaited by mother Kp; pike were caught Vrn; he had found some money Rd. Syntactic relations have been regrouped similarly to those in infinitival constructions, e.g., well water is healthy to drink⁴⁶, where the nominative, initially linked to the neuter-gender form by a predicative relation, became dependent on the infinitive⁴⁷ and synonymous with the accusative, cf. well water is healthy to drink⁴⁸. The use of the object nominative with infinitives and participles in the eastern and southern Aukštaitian dialects (e.g.: where could you get a horse here, and building houses requires a lot of extra work JaBIrR I 560) helped preserve the nominative, which had acquired the corresponding meaning, also with neuter-gender forms of passive participles. 2.2.5.2. In Slavic languages, the development of constructions with neuter-gender forms of adjectives and participles was influenced by another morphological process, namely, the formation of pronominal inflection in adjectives and participles. Having acquired pronominal endings, attributive adjectives and participles formally became distinct from predicative ones. Although the neuter gender of nouns survived in the Slavic languages, the neuter-gender nom.-acc. sing. forms of predicative adjectives in -o, -e (e.g.: OCS gorbko, drevlje, Old Russian зъло, добро, бѣрзо⁴⁹), which retained the old nominal stem ending, were differentiated from inflected pronominal adjectives that agreed with neuter-gender nouns. Since the neuter-gender nom.-acc. forms had long been capable of impersonal use, with a meaning close to that of adverbs (K. Brugmann, Grdr II₂ 688 ff.), once detached from other adjectives they readily became certain adverbs of state — predicatives. Under these circumstances, old two-member non-agreeing constructions, e.g., Old Russian поручено же быть ему стража (П. Я. Черных, Ист. гр. 281) or Old Czech města bylo dobyto (J. Bauer, OSS II 334), were eventually understood as one-member impersonal constructions, and the nominative case (стража, města) as dependent on the indeclinable predicative (поручено, dobyto). The consequences were similar to those in Lithuanian— 44 B. Larin regards the nominative as a direct object in such constructions, LKK VI 100 ff.; J. Balkevičius, Sint. 189—190. 45 Cf. E. Fraenkel, Baltosl. 39; N. Sližienė, LKG II 53. 46 For the structure and use of these constructions, see J. Jablonskis, RR I 161—162, 325, 328, 454, 458, 461—462, 560—564; E. Fraenkel, Kas. 14—16; Б. А. Ларин, LKK V 100 ff.; J. Balkevičius, Sint. 85—86, 115; A. Senn, Hdb. I 467—8; A. Laigonaitė, LKG I 183; A. Piročkinas, Kalbotyra XXII (1) 61—69; J. Otkupščikovas, Dar kartą apie konstrukciją reikia vaikai guldyti, — Leksikos ir sintaksės klausimai, Šiauliai, 1974, 151 ff. with references. 47 J. Endzelīns, LjaGr. 553—554; V. Kiparsky, Baltistica V (2) 147; differently, Б. А. Ларин, LKK VI 103 ff. 48 For the relationship between constructions with the nominative and accusative, see J. Otkupščikovas, ib. 49 For these, see K. Brugmann, Grdr. II₂ 691; A. Meillet, Общесл. 378; П. С. Кузнецов, Ист. морф. 140, В. И. Борковский — П. С. Кузнецов, Ист. гр. 321. 34 in the development: the spread of impersonal constructions with neuter forms of passive participles of intransitive verbs (e.g., Russian dialect бежано со сковородкой50, Old Czech by na lov… jėno, prošeno mne51), their use with oblique agent cases and prepositions (Russian dialect у волков тут хожено, у них уехано52), and the replacement of the nominative by the accusative. For instance, the изба поставлено-type constructions most widespread in the northwestern dialects of Russian can now be regarded as impersonal constructions with an object in the nominative53, while corresponding accusative constructions are used much more widely alongside them than in Lithuanian: его убито, подошву оторвано54; in Ukrainian, constructions of the мене підбито type, attested since the 14th century, are now entirely regular55; similarly in Polish, e.g., wzięto go, podano herbatę, and so on.56 The morphologization of the object dependent on the impersonal passive-participle form has proceeded even further here than in Lithuanian. Syntactic relations have been regrouped similarly in надо трава косить-type constructions57, in which the nominative has likewise become dependent on the infinitive and functions as a synonym of the accusative58. The accusative of the object is also used with impersonal verb forms in other Indo-European languages (Oscan–Umbrian, Latin, Greek, Celtic)59. Thus, Lithuanian constructions with impersonal neuter forms of passive participles essentially correspond in structure and developmental tendencies to Slavic constructions, and have parallels in other Indo-European languages. They differ more markedly only in their modal meanings. 2.3. The impersonal neuter forms of passive participles have modal meanings similar to those of NN with active participles. They also denote: 1) a reported action, experienced indirectly, or another person's speech and thoughts, e.g.: they say, no one has lived there for a long time Kp; Vilius Karalius knows that there is still a room here SimonVK II 373; when you think how much has been endured Virškūpėnai LKT 290; they say, five (wolves) have been eaten Al; they say, they have been stealing from them Dgl; they say, I was deceived Kš; 2) an uncertain, approximately known action, e.g.: on the other side of that river there are good pastures to graze on Jnšk; there, probably, mushrooms are gathered too Kp; it was necessary to travel far to the cities to bring back 50 Н. Б. Кузьмина — Е. В. Немченко, Синт. 18. 51 J. Bauer, TLP II 234; OSS II 334. 52 А. А. Потебня, Зап. III 344; П. Я. Черных, Ист. гр. 281—2; В. И. Борковский — П. С. Кузнецов, Ист. гр. 427; И. Б. Кузьмина — Е. В. Немченко, Синт. 17 tt. 53 И. Б. Кузьмина — Е. В. Немченко, Синт. 47 tt., 51, 56 tt. 54 В. А. Матвеенко, МИРД II 103 tt.; Slavia XXIX 352 tt.; И. Б. Кузьмина — Е. В. Немченко, Синт. 37 tt. su lit. 55 А. А. Потебня, Зап. III 342; В. А. Матвеенко, Slavia XXIX 352; Ф. П. Филин, Происх. 497. 56 A. M. Lewicki, Geneza struktur typu „Podano herbatę“ PJ 1966 5, 207; В. И. Собинникова, OSS II 359—363; for Belarusian, Polish, and other languages, see also E. Fraenkel, Baltosl. 39. 57 On the origin of these constructions from possessive sentences, see А. А. Потебня, Зап. I—II 405 tt.; П. Я. Черных, Ист. гр. 288; В. И. Борковский, Гр. I 338—51; В. И. Борковский — П. С. Кузнецов, Ист. гр. 455—9; V. Kiparsky, ZfSPh XXVIII 333—42; XXXIII 263—266; Baltistica V (2) 141—8 with references. B. Larin considers the object meaning of the nominative archaic and explains it as the influence of a prehistoric substrate, LKK VI 104—6 (for a critique of this view, see G. Jacobsson, Zur Frage von Nominativ als Kasus des direkten Objekts in Slavischen, Lingua viget, Helsinki, 1964, 71—82; Ф. П. Филин, Происх. 488, V. Kiparsky, Baltistica V (2) 145). F. Filin, ib. 490 tt., considers this construction in the northern Russian dialects possibly a recent development of local origin, unrelated to the corresponding Baltic constructions. 58 Žr. И. Б. Кузьмина — Е. В. Немченко, Синт. 46 su lit. 59 H. Hirt, IdgGr. VI 15—6 su lit. 3* 35 to bring salt Btg; apparently, there had been thick pines Sd; perhaps my own had also left tracks Rdn; it (the lake) must have been large, but it will grow over Zr; perhaps a dog has bitten near the eye Ob; 3) an action experienced unexpectedly at the time of speaking, often prompting surprise, e.g.: the matchmakers came inside and looked around, [saying] that there were two girls here Ds; would you look at that—it is already being ploughed here Kp; why, look—he is still being dragged alive BM 18; I come in, and she has already gone into the forest Nmn; the dwarf… calls the girl, but she does not answer; he looks and sees that the bones have been laid out Vėžionys (LKT 378); look how many beehives have been set up—the little garden is full Vnd (LKT 222); the old woman… came out, was harvesting—the room had been filled with water Rš (LKT 337); but the horse snorted, so I came to my senses—I must have been asleep StIm (LKT 316); he finds a little mushroom, opens it, and sees that a lot of meat has been put in Spitrėnai (LKT 324); 4) an action inferred or conjectured from the results, e.g.: Cows have grazed here, as you can see from the hoofprints Skp; I see only one set of wheel tracks on the road—I must have driven here Btg; a beggar sat by the stump, a fire was lit, and he went on farther Pšl; a hare lay here, a fox bit here—I see it all—this is my home—the trees Vlk; well, he must have been a devil Onš; there has also been something here with us—has the whole little road been covered? Kč (LKT 380). Although they coincide with the main modal meanings of NN, the neuter singular forms of passive participles differ in their productivity. These forms are most often used to express an action inferred from its results or experienced unexpectedly at the time of speaking (3, 4), whereas the meanings of a reported or doubted action (1, 2) are less common, less prominent, and seem secondary. Thus, the productivity relationship between the two interrelated groups of meanings is the reverse of that found with active participles. The boundaries between individual meanings are indistinct: the same construction often expresses both an action inferred from the results at hand and one that is simultaneously doubted (e.g., apparently, he had lain down and drunk his fill Šrv), or an action reported on the basis of certain traces (e.g., having looked at their tracks, he says that a person has gone this way Ktk), and so on. The meaning of a past action inferred from its results is especially usual with neuter singular forms of past-tense passive participles and is directly related to the resultative meaning characteristic of these participles in general: when forming compound passive tenses, they likewise mean ‘a past action whose consequences are relevant now’ (N. Slėžienė LKK XV 81). Moreover, when used without a copula, neuter singular forms of passive participles much more often than nominative forms of active participles lack any tangible modal meaning. However, the coincidence of their main modal meanings allows them to be used in the same contexts as NN forms and to be regarded as their syntactic synonyms. 2.4.1. In terms of their distribution across present-day Lithuanian dialects, neuter singular forms of passive participles used with the modal meanings mentioned above differ markedly from active participles (see Fig. 2). In the Samogitian dialect, which is the principal center of NN, these forms are either not used at all (in the central area) or are rare and sporadic (on the northern and southern fringes). Passive participles are used most extensively with the meaning of NN in the eastern, southern, and western Aukštaitian dialects—and precisely in areas where NN is not characteristic. Particularly noteworthy is the productivity of neuter singular forms of passive participles in the very south of Lithuania (south of Lazdijai, Alytus, and Varėna), where active participles in the present and future tenses are not used at all with the meaning of NN. Only in the northern part of the Lithuanian language area, and especially in its northeastern corner, are active and passive participial forms used side by side with the meaning of NN. In this area, a certain differentiation of their meanings can also be observed: constructions with neuter singular forms of passive participles more often denote an action inferred from certain traces or results, or experienced unexpectedly, whereas actions reported or doubted are expressed by active participles used as indirect mood. 36 Fig. 2. 1. šo living, having lived 2. šo having been lived in 2.4.2. The coincidence of the main modal meanings of active and passive participles, despite their occurrence in different areas, is an important argument demonstrating the connection between NN and other cases of predicative use of participles. If we were to consider indirect mood as having simply developed from the nominative with a participle (VD) or the accusative with a participle (GD), it would be entirely unclear how passive participial neuter singular forms—unattested in VD constructions and occurring only rarely and sporadically in GD—could have acquired the same meanings in different areas (e.g.: having returned, I found that Antanas had arrived Jrk; I found that bread had been brought Vžns). Identical meanings could not have arisen by chance, while the data of other, especially the earliest attested, Indo-European languages do not allow us to regard them as archaic60. It seems that they arose as the nature of the predicative use of both active and passive participles changed over a certain period. The conditions for this process were prepared by more general changes in the relations and structure of nominal and verbal sentences. After the old nominal sentences were incorporated into the verbal sentence, many predicative participles remained outside it. Some became fixed in idiomatic or semi-idiomatic expressions; others acquired modal meanings and were thereby reinterpreted within the altered syntactic system. The development of modal meanings in participles was determined by a whole complex of internal and external conditions. 3. THE DEVELOPMENT OF MODAL MEANINGS 3.1. The tendency to use nominal verb forms that have lost their original function with modal meanings is quite frequently apparent in the history of Indo-European languages. Among the many such cases, the incorporation of participles, and later the supine, into the Baltic optative paradigm should be mentioned first. The periphrastic optative of the Slavic languages (e.g., Russian дал бы) is likewise formed from verbal adjectives that became participles, with the suffix -l, and an auxiliary verb1. These same participles in -l, having lost inflection, became the renarrative mood and admirative in Bulgarian (see 3.6.1 ff.). In Greek and Latin, indirect speech was expressed by infinitives (the so-called infinitivus narrationis), that is, by forms that were also nominal in nature2. In general, when a verbal form begins to disappear and its productivity declines, its use with various modal meanings typically becomes more active3. According to J. Kuryłowicz (Infl. cat. 136 ff.), even the old optative originated in a secondary, modal use of the indicative, after other forms had taken over the indicative’s original function. 3.2. The fact that an indirectly experienced action is expressed in many languages by 60The morphologization of modal meanings in Indo-European languages is not very ancient: Hittite has only the indicative and imperative; in other languages that have well preserved the archaic state, at least one further old mood is found. Thus, for the period of Indo-European unity, a system of no more than three moods is generally reconstructed; cf. A. Hahn, Subjunctive and Optative, New York, 1953, p. 56; J. Gonda, The character of Indo-European moods, Wiesbaden, p. 1956; J. Kuryłowicz, Infl. cat. 136; A. Erhart, SPFF A XX (1972) 23; P. Lehmann, PIES 105–6; differently (four moods), A. Сабченко, Ср. гр. 288 ff. Under these circumstances, there is no basis for treating the modal meaning of Lithuanian participles as primary (С. И. Груздева, ВЛУ XX 129–30). 1 Cf. K. Brugmann, Grdr. II 3² 873–4; B. Delbrück, VS. II 409; В. И. Борковский – П. С. Кузнецов, Ист. гр. 288; J. Kuryłowicz, Infl. cat. 137. 2 A. Meillet – J. Vendryes, Traité 672 ff. 3A striking example here is the modal use of the aorist in Serbo-Croatian and, in part, Bulgarian, attested from the 15th century onward. See М. Деянова, ССб 63 ff. 38 perfect. The Lithuanian perfect consists of active and passive participles with the verb to be, and its connection with NN is obvious (see 1.8.2 ff.). In some Slavic (Bulgarian, Macedonian⁴) and Romance (French, Italian, Spanish) languages⁵, new perfect forms incorporating participles and other deverbal nouns also denote an action learned about from others and recounted. That such use of the perfect is quite old is evident from Old Indian, where, already in the Vedic period, the perfect (unlike the aorist and imperfect) denoted actions of which the speaker had not been a witness and about which information had been obtained indirectly, secondhand⁶. As is well known, the Indo-European perfect (as well as its later periphrastic variants in individual language groups) denoted a state existing in the present as the result of an earlier action. On this basis, some scholars linked its modal use with the meaning of an earlier action⁷, others with the meaning of a state⁸. Examination of the history of the perfect in Greek and Old Indian⁹, where its forms have been best preserved, and of the so-called preterite-present verbs in Germanic languages¹⁰ showed that the meaning of state is primary, whereas the emergence of the meaning of past action and the incorporation of the perfect into the verbal tense system are later developments¹¹. Because of this meaning, its root o-vocalism, and other circumstances, the idea has repeatedly been put forward that the perfect developed from deverbal nouns¹². More recently, J. Kurylowicz (Apoph. 41 ff., Infl. cat. 61 ff.) and C. Watkins (IdgGr. III 1 105 ff.) have supported the hypothesis of the perfect’s nominal origin by means of the method of morphonological reconstruction. This has provided new arguments for the view that the tendency to use the perfect in the sense of NN is conditioned, at least indirectly, by its nominal nature. In other language families, too, reported facts tend to be marked by forms of nominal origin. For example, the Turkic “indirect past” forms in -gan, -myš, which initially had a perfect meaning, are derived from participles¹³. In medieval Arabic, when indirect speech was expressed, verbum finitum was replaced by nouns denoting actions¹⁴; deverbal nouns or forms developed from them are used similarly in some Caucasian languages¹⁵. The circumstance that Lithuanian ⁴ See 3.6.1. ⁵ Meyer–Lübke, Grammatik der romanischen Sprachen, III, Leipzig, 1899, § 107; O. Grünenthal, KZ LXIII 135 su lit. ⁶ See J. S. Speyer, VedS. 175; J. Lohmann, KZ LXIV 42; J. Bloch, MSL XIV 48; L. Renou, La valeur du parfait dans les hymnes védiques, Paris, 1925, p. 87. ⁷ L. Renou, ib. 88. J. Kurylowicz, BPTJ XXIX 29–30. J. Kurylowicz regards the development of modal use as a general tendency in the development of the Indo-European perfect, arising from the secondary function of the perfect (“an action prior to the moment of speech”) becoming its primary function. ⁸ O. Grünenthal, KZ LXIII 135 ff. W. Fiedler, ACB VI 369, explained the modal use of the perfect by its combining two temporal planes. ⁹ J. Wackernagel, Studies on the Greek Perfect, Göttingen, 1904; P. Chantraine, History of the Greek Perfect, Paris, 1927; L. Renou, ib. ¹⁰ Э. Бенвенист, ОЛ. 144 ff. ¹¹ Cf. J. Wackernagel, Vorles. I 166 ff.; Э. Бенвенист, ОЛ. 217, J. Kurylowicz, BPTJ XXIX 28; А. Н. Савченко Ср. гр. 326. ¹² Cf. O. Grünenthal, KZ LXIII 133 ff.; H. Hirt, IdgGr. IV 273 ff. In contrast, J. Lohmann, KZ LXIV 42 ff., 48, links the modal use of the perfect solely with its semantics, setting aside the genesis of its forms. ¹³ Н. З. Гаджиева, Очн. 179–180 with references. ¹⁴ Б. М. Гранде, Введение в сравнительное изучение семитских языков, Москва, 1972, p. 397. ¹⁵ H. Vogt, IF LII 245 ff., J. Lohmann, ib. 43 ff.; on the relation of the so-called resultative I verb forms in Georgian, denoting actions inferred from what others say or from their results, to the Latvian indirect mood, see У. Д. Салхучиашвили, Sintagmatika 146–149. 39 the fact that in the language the meaning of indirect mood was acquired by participles, most of which also denote a resultant state, is probably not accidental but grounded in more general patterns of semantic development. 3.3.1. The characteristic meaning of the Lithuanian NN—an action reported or experienced indirectly—is expressed in other Indo-European languages most often by the optative. The beginnings of such usage have been observed already in Old Indian monuments, where the optative, according to J. Gonda, makes it possible to describe an action more cautiously, with a certain reservation, and to emphasize the subjective content of a report16. In Greek, too, the optative had a nuance of caution and reservation and could denote an action whose factuality the speaker did not guarantee. In Homer’s epics the optative was only just beginning to be used to convey indirect speech17, but later, from about the time of Herodotus, it became quite common18. The expression of indirectly experienced or unreal action by the subjunctive, which incorporated the old optative, spread even further in Latin and was partly passed on to the Romance languages19. The Germanic languages also generalized the optative in this meaning20. J. Kazlauskas associated the Lithuanian indirect mood with the optative, Ist. gr. 402–3. He believed that the periphrastic optative forms reconstructed by Ch. S. Stang, NTS XVIII 348–56, once had the meaning of reported action; they consisted of an auxiliary verb and a participle, e.g.: *bi negelbas, *bi nekėlęs (= OCS ašte bi bylъ); after the burden of the optative meaning had been transferred to the participle, bi or be disappeared, and thus the periphrastic optative became the NN. The reconstruction of the periphrastic optative is undoubtedly important for the genesis of the NN: it allows us to infer that compound verb forms with participles were used in the prehistoric past of the Baltic languages and had modal meaning. However, the particle bi, which developed from the auxiliary verb, has survived in many historically attested Lithuanian optative forms containing the supine; therefore, it had not yet disappeared before the supine became established in the optative. The participle itself was displaced from periphrastic optative forms by the supine, and thus it, too, could not have remained the sole bearer of optative meaning. The periphrastic optative forms of the Slavic languages, which correspond in structure to the Lithuanian forms, do not have the meaning of the NN. Under these circumstances, deriving the NN directly from the periphrastic optative does not seem possible. In our view, the periphrastic optative could only have contributed indirectly to the development of the modal use of other participles, occurring with or without auxiliary verbs. 3.3.2. Connections between optative and translative (reported action) and dubitative (doubt) meanings are also observable in the history of Lithuanian itself. In S. Vaišnoras’s “Žemčiūga” and other writings, periphrastic forms of the conditional, composed of būti in the conditional and present-tense active participles, are sometimes used in reported- 16 J. Gonda, The character of the Indo-European moods, Wiesbaden, 1956, p. 61. Autorius duotuose pavyzdžiuose optatyvas maždaug atitinka lie. NN ir gali būti ja verčiamas, pvz., skr. gacchet kadācit, „ji ateisianti kada nors“. Žr. dar E. Fraenkel IF XLVI 44–5 su lit. 17Usually in indirect questions (J. Gonda, ib. 124), where it may be translated into Lithuanian either by the NN or by the conditional, cf. φ 368: ἐφρόντο τίς εἴη “they asked who he was” or “who he might be.” 18 B. Delbrück, VS. III 287; H. Hirt, IdgGr. VII 206; P. Chantraine, Le Grec et la structure des langues modernes, – Language, TCLG IX, 1959, p. 28; J. Gonda, ib. 125–6; Л. Ф. Лосев, О законах сложного предложения в древнегреческом языке, – Статьи и исследования по языкознанию и классической филологии, Москва, 1965, p. 42–43, 47. 19 A. Meillet–J. Vendryes, Traité 674–5; J. H. Schmalz–J. B. Hofmann, Synt. 568 with Lithuanian examples. 20 P. Diels, KZ XLI 194–8; H. Hirt, Urg. III 207. 40 describing implausible or, in the author’s opinion, disputable matters, i.e., in much the same way as the NN, e.g.: the pupil and how the Sacraments were graciously given 120₆₋₇; they say that the words of Christ… were added from eyes and smoke 160₆; await revelations from heaven / through which God may be mighty for your salvation 120₁₄; how could it have been that / so that the Spirit… had not reminded / that such an offering of unity ought to have been PM 9₁₇, 21. In subordinate clauses of complex sentences, these forms also express matters that are doubted, e.g., it seems that nowadays the sacrifice of the Mass does not belong to the priesthood MT (PM) 7₇. Conditional forms are similarly used in the present-day language, cf. he asked who he might be (alongside essą). On the other hand, in S. Vaišnoras’s writings, NN without a conjunction sometimes occurs instead of periphrastic forms of the subjunctive mood, e.g.: Nessa kad tokios treczoses nübūtas / po schei gjwatai priwalu essa (= būtų buvę) / tikek tam S. Powilas butu te pacze / … neuschnmirches PM 33₁₈, 22; schitatai wiss teip… stipru ira / kaip buk Diewas patsai Baszniczioje pribudams dargs bei ischpildas (= darytų bei išpildytų) 109₁₆.The same meaning may also be conveyed by gerundive forms, which S. Vaišnoras sometimes uses generally in place of the neuter participle forms (cf. 1.5.2), e.g.: tatai tur nusidūti ne tū budu / kaip buk dinkscze reikalingas slusbas priwalu sant ant to tiketies 202₁₂. The NN and conditional forms occur side by side in subordinate clauses dependent on the same verbs and having the same functional meaning, e.g.: (the papists) firmly maintain falsehoods, saying… that faith alone in Christ / is not part of repentance. Repentance must be sufficiently sorrowful for those who repent 148₁₃–₁₅. Thus the righteousness of Christ is greatly diminished / as if his suffering were imperfect, sufficient for sins… And his suffering and blood / were the redemption of our insufficient payment. And he has left part of our sufficiency 264–264ₐ. Such synonymous use of the NN and conditional may perhaps be explained in part by the fact that they translated verbs in what was probably the same mood (the subjunctive). Nevertheless, the points of contact between the two moods are evident and also exist in present-day language; therefore, the influence of periphrastic optative forms on the spread of the modal use of active participles is entirely possible. 3.4. The construction VD (nominativus cum participio) and other, probably even older constructions, in which predicative participles followed verbs of speech or mental perception, must have played a very important role in the “attachment” of the meaning of reported or conjectured action to nominative forms of active participles: he said (that) he had walked (see 1.7.2). In these constructions, the main informational burden usually falls on the participles, while the verbs, which formally constitute the basis of the sentence, indicate only the source of the information or some modal aspect of it. Because of this meaning, verbs of saying in sentences conveying indirect speech tend to become parenthetical expressions, modal words, or even certain particles (cf. Russian мол из молвити). Thus the meaning of reported action could easily have become associated with the predicative participles themselves and transferred to sentences in which they had long been used in the position of verbum finitum without verbs of reporting or mental perception, e.g., he walked- 21 Similarly 160₉, PM 24₂₃. 22 Evidently, without any influence from the source language, cf. Denn wann eines solchen dritten Zustands nach diesem Leben vonnöthen (J. Vietor, see W. Witte, Der Uebersetzer Simon Waischnoras d. Ae., Braunschweig, 1931, p. 104, 115; the Latin text is entirely different). Similarly MT 202₁₁–₁₃. 41 having. The GD construction inherited from the distant past may likewise have stimulated the use of participles to denote a reported action. The influence of these constructions explains well the fact that the meaning of a reported action is most characteristic of active participles in the nominative, whereas neuter forms of passive participles, which are not associated with VD and GD constructions, have this meaning much less frequently (cf. 2.3). All the circumstances discussed here were favorable to the development of the modal use of participles. Nevertheless, they do not yet allow us to answer definitively the question: why was it specifically in the Eastern Baltic languages that predicative participles acquired the meaning of indirect mood?23 Very similar conditions obtained in the Slavic languages, where participles could also have a predicative function and where GD and VD constructions were once used, but NN did not develop. Admittedly, the use of participles in the Slavic languages is generally more restricted than in the Baltic languages, and in the Germanic languages it is further reduced. Yet in Greek and Indo-Iranian, where the participial system is rich and developed and its predicative function has also been preserved, participles do not have the meaning of a reported action. Apparently, the formation of NN in the Eastern Baltic languages was facilitated not only by the internal factors discussed above, but also by certain external conditions. 3.5. As early as 1935, P. Arumaa drew attention to the similarity between the indirect, or reported, mood in Latvian and the Estonian modus obliquus (kaudne kõneviis, also referred to below as indirect mood)24. In Estonian, indeclinable forms of active participles with -vat (e.g., tulevat ‘returning, to return’, töötavat ‘working, to work’) denote an action of which the speaker has been informed by another person, and are used in direct speech to report things heard from elsewhere, e.g., meie töötavat halvasti ‘we are said to work poorly’25. As in Lithuanian and Latvian, the participle of the verb lentu ‘to be’, olevat, can function here as an equivalent of an auxiliary verb and, with participles in -nud, form periphrastic past-tense forms of the indirect mood, e.g., mina olevat kirjutanud ‘I (am said to have) written’ (Э. Пялль, ib. 156), cf. Latvian es esot rakstījis. E. Matthewss, comparing such Estonian constructions with the modal use of neuter forms of passive participles in Lithuanian (e.g., mūsų tėvelių visos tos giesmės mokėta ‘our fathers are said to have learned all those songs’), concluded that the latter is autochthonous and independent of influence from other languages (SEER XXXIII 370). V. Pisani compared the indirect mood of Estonian with the active-participle forms expressing the corresponding meaning in the Eastern Baltic languages. Because of their semantic and structural similarity, he proposed that Eastern Baltic NN be regarded as a borrowing from the Baltic-Finnic languages, or as an old common isogloss inherited from the ancestors of modern Finnish and Estonian26. I. Marvan, critically assessing V. Pisani’s hypothesis that NN was borrowed from the Finnic languages, pointed out structural differences between Estonian and Eastern Baltic NN: the participle forming the Estonian modus obliquus, with -vat, is in the partitive case and differs from participles with -van, used in constructions corresponding to Lithuanian GD and VD27. In I. Marvan’s view, the influence of the Finnic languages on the Eastern Baltic indirect mood, 23 И. Марван, Baltistica V (1) 18. 24 In an article published in Eesti Keel in 1935, Nos. 4–6, cited by J. Lohmann, KZ LXIV 46. 25 Э. Пялль, Учебник эстонского языка, Таллин, 1955, p. 155; А. Каск, ЯН СССР III 48. 26 V. Pisani, Zu einer Baltisch–Estfinnischen Partizipialkonstruktion, — Rakstu krājums J. Edzelīnam, Rīgā, 1959, p. 215 tt. 27 И. Марван, Kalbotyra IV 35–37; Baltistica V(1) 18. Ignoring this difference, A. Sauvageot, BSL L 208, derives the Estonian modus obliquus from such GD constructions as näin isän tulevan ‘I see the father coming’. 42 should be regarded as a secondary, modifying factor, while the internal, systemic conditions are primary. This view, also supported by other Baltic–Finnic parallels, is very important and helps us better understand the development of Eastern Baltic NN. 3.5.1. In the Baltic-Finnic languages, a reported action is expressed by forms of varying morphological nature and origin. Besides participles with -vat, Estonian has used deverbal forms with -ma with the meaning of NN (A. Sauvageot, BSL L 208). NN is also formed here by infinitives in -da, which were even more widespread than participles in -vat in texts up to the beginning of the twentieth century, and have only in recent decades been disappearing from the literary language28. These infinitives are likewise unrelated to GD- and VD-type constructions; A. Sauvageot, ib. 208–209, explains their modal use by the fact that the infinitive generally conveys a meaning of indeterminacy and uncertainty. In Livonian, which is characterized by many archaic features and longstanding contacts with the Baltic languages29, the forms of NN are again different: they have developed from an agent noun (nomen agentis) with -iji and form present-future and periphrastic past-tense paradigms, e.g.: ta lu’ggiji ‘he is said to be reading’, ta vol’ lu’ggiji ‘he is said to have read’30. Almost all the formatives of NN in the Baltic-Finnic languages are phonetically similar to the suffixes of Baltic participles denoting a reported action; cf. Estonian -vat and Latvian -ot, Estonian -ma and Lithuanian present-tense active passive participles in -mas, -ma, Livonian -iji and Latvian past active participles in -ujis (rakstījis), although they apparently have nothing in common in origin. Further instances of similarity between morphological formatives of different origins in the Baltic-Finnic and Baltic languages have also been noted; cf. illatives with the formative -n, which developed in distinctive ways in each of the two language families (A. Sabaliauskas, LKK VI 112 with literature). NN is not used in the other Baltic-Finnic languages (nor in the Finnic-Ugric languages generally), and there is no evidence to suggest that it once existed there and subsequently disappeared. On the contrary, the variation in forms would more likely indicate that this mood had already acquired morphological expression in the period when the separate Baltic-Finnic languages existed. 3.5.2. In both the Baltic and Baltic-Finnic languages, verbal nouns were used to express a reported action: participles, “converbs,” and infinitives. Finnic languages in general are characterized by a rich system of participles and infinitives31 and their extensive use, in many respects akin to that in the Baltic languages. Participles here are the principal means of expressing a secondary clause action; moreover, they can often have a separate agent of the action and function as equivalents of verbum finitum32. Many Finnish participial constructions (e.g., GD- and VD-type expressions33) correspond in their underlying structural pattern to Baltic ones. Following the example of the Eastern Baltic languages, Estonian even developed compound past tenses with forms of the verb olema ‘to be’ and a participle in -nud34. On the other hand, contacts with the Baltic Finns apparently helped the Eastern Baltic peoples 28 A. Kask, Die indirekte Redeweise in der Estnischen Schriftsprache, – Congressus Tertius Internationalis Fenno—Ugristarum, Tallinn, 1970, Teesid I, p. 49. 29 П. А. Аристэ, ТИЯ IV 257 тт.; Ливы и ливский язык, „Eesti NSV TA Toimetised, ühiskonnateaduste s.“, 1, 1958, p. 42 тт. 30 П. А. Аристэ, ВЭИ 15; Э. Э. Вяари, ЯН СССР III 148. 31 For example, literary Finnish has 3 participles and 5 infinitives, Livonian has 4 participles and 2 infinitives, and Karelian has 2 active and 3 passive participles, as well as 7 converbs formed from various case forms of participles; see G. Décsy, Einf. 205; Э. Э. Вяари, ЯН СССР III 148–9; Г. Н. Макаров, ib. 73. 32 Л. Хакулинен, Разв. II 248 тт. su lit.; К. Е. Майтинская, ЯН СССР II 23. 33 Cf. Л. Хакулинен, ib. 248 ff., 253. 34 П. А. Аристэ, ВЭИ 12–13; Э. Пялль, op. cit. 148–9; А. Х. Каск, ЯН СССР II 48; G. Décsy, Einf. 205; A. Sabaliauskas, LKK VI 112. 43 to this day, preserving the varied and archaic use of participles, which in other Indo-European languages was greatly reduced and fell into disuse over time. In the Slavic-language area, predicative participial constructions were also best preserved in the northwestern dialects of Russian, which form an adstratal layer in the former territory of the Finnic languages35. That the predicative use of participles, infinitives, and similar forms in the Finnic languages is very archaic is evident from the structure of their verbum finitum itself: most finite verb forms here still retain formatives shared with nouns36. 3.5.3. Thus, the meanings of directly and indirectly experienced action are morphologically differentiated only in a fairly small and bounded area of the Baltic and Baltic Finnic languages, which begins in the north with the Estonian-language area and extends south along the Baltic coast, encompassing the present-day Latvian dialects and a large part of the Lithuanian dialects. The semantic category of indirectly experienced action is realized here by morphological forms of different origins: participles (Lithuanian), indeclinable participial forms (Latvian, Estonian), infinitives (Estonian), and deverbal nouns (Livonian). However, the semantic features of this category and its relation to the speech situation are very similar throughout the area37. This category cannot be particularly archaic, since it is unknown to the closest neighbors of either the East Baltic peoples or the Estonians and Livonians. Most likely, it took shape when the unity of the Baltic Finnic languages had already broken up, while the Eastern Balts still formed a relatively cohesive linguistic community in which newly emerging phenomena could spread unhindered. The spread of the semantic category of indirectly experienced action in the East Baltic languages may be associated with their expansion into areas inhabited by the predecessors of the Livonians and Estonians in the 5th–6th centuries AD38. Long-term language contacts and extensive substrate and bilingualism zones are characteristic of almost the entire area: above all, the Livonian and Estonian substrate in northern Latvia, the Couronian–Latvian and Lithuanian languages in the western part of the area, and Selonian in southern Latvia and northeastern Lithuania. It is easy to see that NN is distributed precisely in that part of the Lithuanian-language area where Couronian and Selonian were once spoken. It was precisely through the Latgalian, Couronian, and Selonian dialects, influenced by the language of the Estonians’ predecessors39, that the category of indirectly experienced action could spread across a large part of the present-day Lithuanian-language area. (Direct Baltic Finnic influence on Lithuanian dialects does not seem to have been possible.) The intensity of the use of the indirect mood in the Baltic-language area is roughly proportional to the degree of closeness of Baltic–Baltic Finnic language contacts: it is greatest in the territory of Latvia— 35 In this case, the Baltic Finnic languages may have had a similar “conserving” influence on Baltic and Slavic participles as on nominativus cum infinitivo constructions, according to V. Kiparsky’s “refrigerator theory”; cf. V. Kiparsky, ZfslPh XXVIII 333—42; XXXIII 263—4; Baltistica V(2) 148. 36 B. Collinder, Comparative grammar of the Uralic languages, Stockholm, 1960, p. 243; G. Décsy, Einf. 160—1. Kadangi veiksmažodžio—vardažodžio kategorijos yra koreliatyvios, pastaruoju metu manoma, jog metodiškai netikslu kalbėti apie veiksmažodžių kilmę iš vardažodžių, tačiau finų-ugrų kalbų verbum finitum artimumas vardažodžių kamienams ir jų genetinis ryšys abejonių nekelia, žr. K. E. Майтинская, Оч.-УЯ 219, 300 tt., 330 tt. su lit. 37 In this respect, the semantic category of indirectly experienced action may be compared with the category of action expressed as obligatory, which is realized in Latvian and Livonian and in the Saaremaa dialect of Estonian by different forms of the debitive; cf. J. Endzelins, LaGr. 885 ff., 971 ff.; П. А. Аристэ, ТЯИ IV 288; Э. Э. Вяари, ЯН СССР II 147—8. 38 The northward migration of the Balts during this period is indicated by both archaeological and linguistic evidence (remnants of Estonian and Livonian languages in northern Latvia, Finnic toponymy, etc.); see П. А. Аристэ, ВЭИ 16 ff.; M. Gimbutas, The Balts, London, 1963, p. 141; V. Mažiulis, Baltistica X (2) 124. 39 On the indirect influence of the Baltic Finnic languages on Lithuanian through Couronian and Selonian, cf. V. Mažiulis, Baltistica X (2) 124. 44 ... in the area, smaller—in the northeastern corner and northern fringe of the Lithuanian-speaking territory and in Samogitia, while farther south it weakens and fades. In southern Lithuania, the neuter singular forms of passive participles used there to convey reported action have this meaning considerably less often than the nominative forms of active participles in the northeastern Aukštaitian and Samogitian dialects. Thus, the category of indirectly witnessed action most likely spread across the Baltic-language area gradually from north to south, weakening along the way, and perhaps never reached the territory of the Old Prussian language. 3.6. The development of the category of indirectly witnessed action in East Baltic and Baltic Finnic languages, which belong to different families, has a striking parallel at a quite different periphery of the Indo-European language area, namely, in the Balkan Sprachbund. In this regard, the data from Bulgarian are particularly interesting. 3.6.1. Bulgarian participles with the suffix l, when used in the verb position, can form the so-called renarrative forms or renarrative mood40. These express things experienced indirectly, reported, or doubtful, whose truth the speaker does not guarantee. The first- and second-person forms have the auxiliary verb съм and, being formed from the perfect stem, coincide with the indicative periphrastic perfect (e.g.: писал съм “I have written,” писал си “you have written”), whereas in the third person (which is also the person most often used41) the participle occurs without an auxiliary verb, thus differing from other compound tenses; e.g.: писал “(he) has written,” писали “(they) have written.” This distinction is not absolute, since the renarrative forms may occasionally have a third-person auxiliary verb, while the perfect may occur without one42. Forms made from the imperfect stem are used only with the meaning of a reported or doubted action, e.g.: писал съм, писал си, писал “I, you, he have/has written”43. In addition to the simple forms, the paradigm of the renarrative mood also includes the so-called emphatic forms, which are formed just like Lithuanian compound forms, that is, with a participle of an auxiliary verb meaning “to be”; e.g., бил писал “(he) had apparently written”44. They denote reported actions whose truth the speaker doubts, that is, they have a more pronounced modal nuance and in this respect likewise correspond to the Lithuanian compound NN forms45. Among the other Slavic languages, Macedonian also has renarrative forms with l participles46, and its grammatical structure essentially coincides with that of Bulgarian. Serbian— 40 The place of these forms in the grammatical system of Bulgarian is debated. In grammars they are usually regarded as one of the verb moods—a translatif, commentative, or прези-казного наклонение; see А. Андрейчин, Н. Костов, К. Мирчев, Е. Николов, Ст. Стойков, Български език, София, 1955, p. 212—234; Ю. С. Маслов, Очерк 288 ff. Since the renarrative forms also correlate with the subjunctive and imperative moods, some researchers have proposed distinguishing an entire system of secondary (renarrative) moods in Bulgarian; see Е. И. Демина, БГБЛЯ 359 ff., or treating renarrative forms as particular cases of the modal use of the indicative mood (изъявительно наклонение); see В. Станков, Врем. 155 ff., especially 162—4 with references, and more recently Е. И. Демина, Ген. 13 ff. 41 According to V. Koseska, SFPS XIII 166, third-person forms make up about 90% of all examples in the contemporary language. 42 Cf. М. Г. Рожновская, Синт. 218; Ю. С. Маслов, Очерк 250; V. Koseska, SFPS XIII 166; В. Станков, ib. 78; Е. Дьомина, БЕ XX 407 ff. 43 М. Г. Рожновская, Синт. 117. 44 Е. М. Демина, БГБЛЯ 353 tt.; В. Станков, Врем. 166. 45 For forms with a similar emphatic meaning in Albanian, see W. Fiedler, ACB VI 369. 46 H. G. Lunt, Grammar of the Macedonian literary language, Skopje, 1952, 91 tt. Tačiau makedoniečių kalboje atpasakojamosios formos vartojamos su pagalbiniu veiksmažodžiu taip pat kaip perfekto, tad formaliai nuo pastarųjų nesiskiria. 45 In Croatian, a perfect is formed with these participles, whose 3rd person, without an auxiliary verb, can be used to express an indirectly experienced action and other modal meanings47. In general, participles derived from the Common Slavic verbal adjectives with the suffix l formed a paradigm of the indirect (or, respectively, reportative) mood only in a very limited area of individual Slavic languages and, apparently, relatively late48. 3.6.2. The same l-participle forms, made from imperfect and perfect stems and denoting a reported action, are also used in Bulgarian to express an unexpectedly experienced, astonishing action. In such cases they are called the admirative. Some researchers regard the Bulgarian admirative as a variant of the reportative mood (translative or commentative)49, while others discuss it as a separate group of forms50 or as a particular case of the modal use of the indicative tenses51. However, both groups recognize a genetic relationship between the reportative mood and the admirative52. The Albanian admirative, formed with new forms of verbal nouns (the so-called short participles) and a postposed auxiliary verb meaning “have”53, likewise denotes not only an unexpectedly experienced or astonishing action, but also a reported action with a certain emotional coloring54. The two meanings are closely related (an action may be astonishing whether experienced directly or reported) and, in the view of some Balkanists, developed from the emphatic use of the perfect55; only in Bulgarian was the commentative (reportative) meaning component more prominent, while in Albanian the admirative (expression of astonishment) component was more prominent56. It is easy to see that, in the relationship between the meanings of reported and astonishing action, the Bulgarian and Albanian admirative fully corresponds to Lithuanian NN57. Bulgarian reportative-admirative forms, when used with the conjunction da, also have an optative meaning, e.g., da ne cu otшila tam! (especially after nei- 47 И. Грицкат, О перфекту без помоћног глагола и сродним синтаксичким појавама у српскохрватском језику, Београд, 1954; J. Kuryłowicz PRTJ XXIX 30–1; М. Деянова, ИИБЕ XIX 847. 48 According to K. Mirčev, Ист. Гр. 123, the modal use of the perfect began to spread in Bulgarian from the 14th century. E. Diomina, Ген. 15 and БЕ XX 410, notes that 17th-century Bulgarian literature still reflects an early period in the formation of reportative forms. 49 Ю. С. Маслов, К вопросу о системе форм пересказывательного наклонения, Сборник в чест на Александър Теодоров-Балан, София, 1956, р. 314 tt.; Озерк 250–1; W. Fiedler, ACB VI 369; WZL XV. 3 565. 50 Е. И. Демина, ВГБЛЯ 327–8 su lit. 51 В. Станков, Врем. 176 su lit. 52 At the International Congress of Balkan Studies in 1968 and the Symposium on the Grammatical Typology of the Balkan Languages in 1973, this point was also emphasized by E. Diomina, who, from the perspective of the modern language, regards the admirative and reportative forms as different categories; see Е. И. Демина, Ген. 14. 53 Dėl albanų kalbos admiratyvo žr. G. Weigand, Albanesische Grammatik, Leipzig, 1913, p. 122 tt.; W. Fiedler, ACB VI 367 tt.; WZL XV. 3 561 tt. su lit.; А. В. Десницкая, Албанский язык и его диалекты, Ленинград, 1968, р. 17tt. 54 W. Fiedler, WZL XV. 3 563–5. 55 Historically, the Albanian admirative differs from the perfect only in the position of the participle and auxiliary verb: in perfect forms, the auxiliary verb precedes the participle and is not fused with it; cf. А. В. Десницкая, ib. 17–18; W. Fiedler, WZL XV. 3 562. E. Çabej points to the gradual transition from the meaning of the perfect to that of the admirative in the written tradition of Albanian, ACB 386–7. 56 W. Fiedler, WZL XV. 3 565. 57 The similarity between the Bulgarian admirative and Lithuanian participles expressing an astonishing action was noted by A. Nepokupnas, who proposed distinguishing an admirative category in the Baltic languages as well; see А. П. Непокупный, Ареальные аспекты балто-славянских отношений, Киев, Москва, 1964, р. 26–7. 46 (for the genitive, see V. Koseska, SFPS XIII 168–9), thereby also coinciding with Lithuanian predicative participles (e.g., Nekalbėjęs! and so on; see 1.9.2). In addition, they also express a past action inferred from its results58 (cf. the fourth meaning of Lithuanian NN). Thus, the connection between the admirative and the reportative forms, the perfect, and, at the same time, the use of predicative participles is entirely beyond doubt. 3.6.3. Just as the East Baltic indirect mood is semantically close to the Baltic Finnic modus obliquus, so the reportative forms in Bulgarian, Macedonian, and Albanian are akin to the Turkish “inferential preterite,” which developed from a participle with the suffix -myš and denotes actions experienced indirectly, through another person’s account59. It is thought that this modal meaning of the tense arose on the basis of an older perfect meaning60. Inferential preterite forms derived from participles with -myš and -gan have spread areally to other Turkic languages as well (forms with -gan are attested from the sixteenth century, and those with -myš from the beginning of written records in the Oghuz languages)61. In addition, Turkmen also uses inferential preterite forms with -an62. Since the modal category of the non-evidential, that is, indirectly experienced action, is widespread in Turkic languages and has been so since ancient times, it is quite understandable that, over centuries of linguistic contact, it could also have influenced Bulgarian, Macedonian, and Albanian. Some scholars thought that the Bulgarian and Macedonian reportative forms arose under the influence of the Turkish verbal system during the period of Bulgarian–Turkish bilingualism (fourteenth–nineteenth centuries)63 or as a result of the Turkic substrate of the Proto-Bulgarians (after the seventh century)64. Closer study of the semantic structure of the categories of indirectly experienced and surprising action, as well as their distinctive morphological realization, brought to light internal preconditions for these categories, inherent in the Balkan Slavic and Albanian languages themselves65. Nevertheless, there is no doubt that this category took shape under complex conditions of intensive linguistic contact66. Its very limited area at the southern edge of the Indo-European linguistic territory and its relatively recent formation indicate that the main center of its spread must have been in the Turkic-language area67. 58 E. Diomina calls this meaning conclusive; see Е. И. Демина, Ген. 14; Е. Дво-мина, БЕ XX 411–2. 59 For this, see O. Grünenthal, KZ LXIII 134; А. И. Кононов, Грамматика современного турецкого языка, Москва – Ленинград, 1956, p. 231 ff.; Н. З. Гаджиева, Осн. 182 ff. Turkish verbal forms with -myš, like Bulgarian forms with -l, are also used as a participium conjunctum; see O. Grünenthal, ib. 134; А. Н. Кононов, ib. 251. 60 O. Grünenthal, ib. 134. 61 Н. З. Гаджиева, Осн. 179 tt., 182–3. 62 Ib. 181–2. 63 Б. Цонев, История на български език, III, София, 1937, p. 106; К. Мирчев. Ист. гр. 48, 123, 187; Л. Андрейчин, ИИБЕ II 39 tt.; Э. Бенвенист, ОЛ. 223–4. 64 K. Popov put forward this view; see Е. И. Демина, Ген. 14. 65 S. Mladenov had already considered the Bulgarian reportative forms to be autochthonous phenomena, Symb. Rozwadowski II 205 ff.; see also V. Koseska, SFPS XIII 168 and the literature cited in E. Diomina’s study ББЛЯ 314. 66 According to E. Diomina, the explanation of their origin must be sought “in the entire complex situation of long-standing linguistic contacts on the Balkan Peninsula” ((Е. И. Демина, Ген. 17). 67 Even scholars who emphasize the distinctiveness of these verbal forms in the Balkan Slavic and Albanian languages acknowledge Turkic influence on Bulgarian reportative forms (e.g., the omission of the third-person auxiliary verb) and on the use of the Albanian admirative to express indirectly experienced action; cf. W. Fiedler, WZL 565, Е. И. Демина, Ген. 15–16; А. С. Мельничук, Розв. 86–8. 47 3.7. Thus, the semantic category of indirectly experienced action is realized by very similar means at two diametrically opposed fringes of the Indo-European linguistic area—the Baltic region and the Balkans. The indirect mood of the eastern Baltic languages, the reported forms of the Balkan Slavic languages, and the Albanian admirative coincide in many features of meaning and morphological expression. The most important similarities are as follows: 1. The semantic structure of the aforementioned modal categories consists of the meanings of indirectly experienced (translative), doubtful (dubitative), or surprising action (admirative). Particularly noteworthy is the nontrivial connection between the quite distinct admirative and translative meanings, and their expression by the same forms. 2. Indirectly experienced action is expressed by participles that form part of the system of compound (periphrastic) tenses. Although the participial forms themselves have different origins, the indirect-mood paradigms formed with them are very similar in structure (cf. the compound tenses of the Lithuanian and Latvian indirect moods and the so-called emphatic Bulgarian reported forms and Albanian admirative forms). 3. In all the languages mentioned, forms expressing indirectly experienced action are genetically related to the periphrastic perfect, and in the eastern Baltic and Balkan Slavic languages they are even homonymous with it in certain cases. 4. The category of indirectly experienced action was morphologized relatively late (apparently, no earlier than the second half of the first millennium CE). These similarities cannot be accidental. They clearly indicate the typological commonality of the semantics and expression of the indirect mood and its functional equivalents. The external conditions for the formation of the indirect mood are also very similar. At both fringes of the Indo-European linguistic area, the distinction between directly and indirectly experienced action is found within a clearly defined area that, in addition to Indo-European languages, includes languages of other families (Finno-Ugric and Turkic), which together with neighboring Indo-European languages form zones of intensive mutual contact, bilingualism, or even a “linguistic union.” Although the prerequisites and beginnings of the indirect mood can also be found in other Indo-European languages, nowhere else have they been morphologically formalized. Meanwhile, in languages of other families, such as Turkic, the indirect mood or its equivalents have been used more widely and for a long time. Under such circumstances, the role of language contacts in the formation of the indirect mood is quite evident. Comparison with other languages that have corresponding modal categories helps us better understand the origin of the Lithuanian indirect mood. Its structural basis was the predicative use of participles inherited from the Proto-Indo-European period, itself based on a nominal sentence type. Its semantic prerequisites are associated with the tendencies in Indo-European languages toward modal uses of participles (especially without an auxiliary verb) and of periphrastic verb forms formed with them (especially the perfect and optative). The establishment of the reported-action meaning was stimulated by semi-predicative participial constructions: the accusative and nominative with a participle. However, contact with the nearest Baltic Finnic languages (Livonian and Estonian) played a major role in the final formalization of the indirect mood and the formation of its morphological paradigm. Literature Abbreviations 1. Monographs S. Agrell, Neutr. – Zur Geschichte des indogermanischen Neutrums, Lund, 1926. E. Benveniste, Hitt. – Hittite et indo-européen. Etudes comparatives, Paris, 1962. J. Balkevičius, Sint. – Dabartinės lietuvių kalbos sintaksė, Vilnius, 1963. 48 A. Bezzenberger, BGLS. — Contributions to the History of the Lithuanian Language Based on Lithuanian Texts of the 16th and 17th Centuries, Göttingen, 1877. K. Brugmann, Grdr. — Outline of Comparative Grammar of the Indo-European Languages, Strassburg, 1897–1916. — ” — Synt. — The Syntax of the Simple Sentence in Indo-European, Berlin — Leipzig, 1925. K. Būga, RR. — Selected Writings, I–IV, comp. Z. Zinkevičius, ed. V. Mažiulis, Vilnius, 1958–1962. G. Décsy, Einf. — Introduction to Finno-Ugric Linguistics, Wiesbaden, 1965. B. Delbrück, VS. — Comparative Syntax of the Indo-European Languages, I–III, Strassburg, 1893–1900. J. Endzelin, AprGr. — Altpreussische Grammatik, Riga, 1944. J. Endzelīns, DI. — Selected Works, Riga, 1971 ff. — ” — LaGr. — Grammar of the Latvian Language, Riga, 1951. E. Fraenkel, Baltosl. — Baltoslavica, Contributions to Balto-Slavic Grammar and Syntax, Göttingen, 1921. — ” — Kas. — Syntax of the Lithuanian Cases, Kaunas, 1928. H. Hirt, IdgGr. — Indogermanische Grammatik, I–VII, Heidelberg, 1927–1937. — ” — Urg. — Handbook of Proto-Germanic, Heidelberg, 1931–1934. H. Hübschmann, Cas. — Zur Casuslehre, München, 1875. J. Jablonskis, RR. — Rinktiniai raštai, I–II, sud. J. Palionis, Vilnius, 1957. G. Janke, Pass. — Der Ausdruck des Passivs im Altrussischen, Berlin, 1960. J. Kabelka, La. — Latvių kalba, Vilnius, 1975. J. Kazlauskas, Ist. gr. — Historical Grammar of the Lithuanian Language (Accentuation, Noun, Verb), Vilnius, 1968. R. Kühner, GrGr. — Detailed Grammar of the Greek Language, I–II, Hannover, ²1870. J. Kurylowicz, Apoph. — L’ apophonie en indo-européen, Warszawa, Kraków, 1956. — ” — Infl. cat. — The Inflectional Categories of Indo-European, Heidelberg, 1964. O. Lagerkrantz, Präd. — Indogermanisches Prädikativ, Uppsala, 1933. W. P. Lehmann, PIE. — Proto-Indo-European syntax, New York, 1974. V. Mažiulis, Balt. — The relations between Baltic and other Indo-European languages (declension), Vilnius, 1970. A. Meillet, Altarm. — Altarmenisches Elementarbuch, Heidelberg, 1913. A. Meillet – J. Vendryes, Traité — Treatise on the Comparative Grammar of the Classical Languages, Paris, 1952. F. Miklosich, VGr. — Comparative Grammar of the Slavic Languages, I–IV, Vienna, 1868–1879. T. Milewski, Indo-hitt. — L’ indo-hittite et l’ indo-européen, Cracovie, 1936. J. Palionis, LLK — Lietuvių literatūrinė kalba XVI–XVII a., Vilnius, 1967. A. Paulauskienė, Veiksm. — Dabartinės lietuvių kalbos veiksmažodis, Vilnius, 1971. M. Rudzīte, Dial. — Latviešu dialektoloģija, Rīgā, 1964. J. H. Schmalz – J. B. Hofmann, Synt. — Syntax and Stylistics, – Stolz–Schmalz, Latin Grammar, Munich, 1928, 345–850. J. Schmidt, Pluralb. — Die Pluralbildungen der indogermanischen Neutra, Weimar, 1889. A. Senn, Hdb. — Handbuch der litauischen Sprache, I, Grammatik, Heidelberg, 1966. J. S. Speyer, VedS. — Vedische und Sanskrit-Syntax, Strassburg, 1896. Ch. S. Stang, Verb. — Das slavische und baltische Verbum, Oslo, 1942. — ” — VGr. — Comparative Grammar of the Baltic Languages, Oslo–Bergen–Tromsø, 1966. R. Trautmann, AprS. — The Old Prussian linguistic monuments. Introduction, texts, grammar, dictionary. Göttingen, 1910. 4. Užs. Nr. 1026 49 W. Veenker, Substr. — The question of the Finno-Ugric substrate in the Russian language, Bloomington, 1967. J. Wackernagel, Vorles. — Lectures on syntax, with special consideration of Greek, Latin, and German, I–II, Basel, 21926–1928. C. Watkins, IdgGr. — Indo-European Grammar, edited by J. Kuryłowicz, III, Morphology, Part One, History of Indo-European Verbal Inflection, Heidelberg, 1969. W. D. Whitney, SkrG. — A Sanskrit grammar, Leipzig, 21889. Z. Zinkevičius, Dial. — Lietuvių dialektologija, Vilnius, 1966. Э. Бенвенист, ОЛ. — Общая лингвистика, М., 1974. В. И. Борковский, Гр. — Syntax of Old Russian charters (simple sentence), Львов, 1949. В. И. Борковский—П. С. Кузнецов, Ист. гр. — Historical Grammar of the Russian Language, М., 1965. Р. И. Булаев, Ист. гр. — Historical Grammar of the Russian Language, М., 61959. Е. И. Демина, Ген. — On the question of the genesis of modal categories of the Bulgarian indicative, — Symposium on the grammatical typology of modern Balkan languages, preliminary materials, М., 1973, 13–17. Н. З. Гаджиева, Осн. — The main paths of development of the syntactic structure of the Turkic languages, М., 1973. Вяч. Вс. Иванов, Общ. — The Common Indo-European, Proto-Slavic, and Anatolian linguistic systems, М., 1965. П. С. Кузнецов, Ист. морф. — Очерки исторической морфологии русского языка, М., 1959. И. Б. Кузьмина — Е. В. Немченко, Синт. — Синтаксис причастных форм в русских говорах, М., 1971. Т. П. Ломтев, Очерки — Очерки по историческому синтаксису русского языка, М., 1956. Ю. С. Маслов, Очерк — Очерк болгарской грамматики, М., 1956. А. Мейе, Общ. — Общеславянский язык, М., 1951. О. Мельничук, Розв. — Розвиток структури слов'янського речення, Київ, 1966. К. Мирчев, Ист. гр. — Историческа граматика на българския език, София, 1953. С. П. Обнорский, Очерки — Очерки по истории русского литературного языка старшего периода, М.—Л., 1946. А. А. Потебня, Зап. — Из записок по русской грамматике, I–II, М., 21958, III, М., 21968. М. Г. Рожновская, Синт. — Синтаксис прилагательного в болгарском литературном языке, М., 1970. А. Н. Савченко, Ср. гр. — Сравнительная грамматика индоевропейских языков, М., 1974. В. Станков, Врем. — Българските глаголни времена, София, 1969. Ф. П. Филин, Происх. — Происхождение русского, украинского и белорусского языков, Л., 1972. Л. Хакулинен, Разв. — Развитие и структура финского языка, I–II, М., 1953–1955. П. Я. Черных, Ист. гр. — Историческая грамматика русского языка, М., 1954. А. А. Шахматов, Синт. — Синтаксис русского языка, Л., 1925. 2. General works, periodicals, etc. ACB — Actes du Premier Congrès International des Études Balkaniques et Sud-Est-Européennes, Sofia, 1968. Baltistica — Baltistica, Studies of the Baltic languages, Vilnius, 1965 ff. 50 BPTJ — Bulletin of the Polish Linguistic Society. BSL — Bulletin of the Linguistic Society of Paris, from 1869 onward. DB — Donum Balticum to prof. Ch. S. Stang, ed. by V. Rūķe-Dra- vina, Stockholm, 1970. GK — Native Language, Kaunas, 1933–1941. IF — Indo-European Research, from 1892 onward. Indogermanica — Indogermanica, Festschrift für W. Krause, Heidelberg, 1960. Kalbotyra — Linguistics, scholarly papers of the higher education institutions of the Lithuanian SSR, Vilnius, from 1958 onward. KZ — Journal of Comparative Linguistic Research in the Field of Indo-European Languages, founded by A. Kuhn, from 1852 onward. LB — The Bible: That is, the whole Holy Scripture: Newly prepared in German by D. Marth, Luth. Wittenberg. Printed by Hans Lufft. 1549. Lg — Language, journal of the Linguistic Society of America, Baltimore, from 1925 onward. LKG — Institute of the Lithuanian Language and Literature of the Lithuanian SSR Academy of Sciences, Grammar of the Lithuanian Language, ed. K. Ulvydas, I–II, Vilnius, 1965–1971. LKK — Questions of Lithuanian Linguistics, Vilnius, from 1957 onward. LKT — Institute of the Lithuanian Language and Literature of the Lithuanian SSR Academy of Sciences, Lithuanian Dialects: A Reader, ed. E. Grinaveckienė and K. Mor- kūnas, Vilnius, 1970. LVLIR — Papers of the Institute of Latvian Language and Literature of the Latvian SSR Academy of Sciences, Riga. MAD — Papers of the Lithuanian SSR Academy of Sciences, Series A. MLLVG — Institute of Language and Literature of the Latvian SSR Academy of Sciences, Grammar of Modern Literary Latvian, I–II, Riga, 1959–1962. MSL — Memoirs of the Linguistic Society of Paris, from 1868 onward. NTS — Norwegian Journal of Linguistics, Oslo, from 1928 onward. OSS — Otázky slovanské syntaxe, I—III, Brno. PJ — Poradnik językowy, Warszawa. RKr — Rakstu krājums, Rīgas Latviešu Biedrības Zinību Komisijas izdots. SEER — Slavonic and East European Review, London. SFPS — Studia z filologii polskiej i słowiańskiej, Warszawa. Syntagmatics — Syntagmatics, paradigmatics and mutual relations at the syntactic level, 6th A. Ozols Day, materials, Riga, 1970. Slavia — Slavia, Časopis pro slovanskou filologii, Praha, 1922 tt. SlPr — Slavia Pragensia. SPFS — Proceedings of the Faculty of Philosophy of the University of Brno, Linguistics Series (A), Brno. TCLC — Travaux du Cercle Linguistique de Copenhague, Copenhague. TLP — Travaux linguistiques de Prague, Prague. V — Biblia sacra Vvlgatae editionis… Romae… MD XCIII. W — Postille… By D. Iakob Wvyka… in Kraków… 1617 (sixth edition). WZL — Wissenschaftliche Zeitschrift der K. Marx-Universität Leipzig, Ges.-Sprachwiss. Reihe. ZfslPh — Zeitschrift für slavische Philologie, Leipzig, 1925 tt. БЕ — Bulgarian Language, journal of the Institute for Bulgarian Language at the Bulgarian Academy of Sciences, Sofia. ВГБЛЯ — Вопросы грамматики болгарского литературного языка, М., 1959. Leningrad University Bulletin. Introduction to the comparative-historical study of Slavic languages, ed. O. S. Melnychuk, Kyiv, 1966. 4* 51 Questions of the Ethnic History of the Estonian People, Tallinn, 1956. Questions of Linguistics, Moscow. Proceedings of the Institute of Bulgarian Language, Sofia. Proceedings of the Department of Russian Language and Literature of the Imperial Academy of Sciences, Petrograd, 1896. Materials and Studies in Russian Dialectology. ОФ-УЯ — Основы финно-угорского языкознания, М., 1974. Slavistic Collection on the Occasion of the VI International Congress of Slavists in Prague, Sofia, 1968. Proceedings of the Institute of Linguistics of the USSR Academy of Sciences. Scholarly Notes of the A. I. Herzen Leningrad State Pedagogical Institute. Ergative Sentence Construction in Languages of Various Types (Studies and Materials), Leningrad, 1967. Languages of the Peoples of the USSR, I–V, Moscow, 1966. В. АМБРАЗАС THE SPREAD AND THE PROBLEM OF THE ORIGIN OF THE INDIRECT MOOD (MODUS RELATIVUS) Summary 1. The indirect mood (hereafter IM), or modus relativus, in the East Baltic languages is formed by active participles in *-nt- and *-us- (nominative forms in Lithuanian, and the more recent indeclinable “departiciples” in Latvian), e.g.: Lithuanian tēvas einąs, Latvian tēvs ejot ‘the father is supposedly going’. The IM paradigm is symmetrical with respect to the indicative paradigm and contrasts with it; besides simple forms, it includes compound (periphrastic) forms with active participles of the auxiliary verb būti, which contrast with the corresponding periphrastic indicative forms, e.g.: yra, buvo, būdavo, bus metęs: esąs, buvęs, būdavęs, būsiąs metęs. The principal distinguishing features of the IM are the position of verbum finitum in the sentence and the absence of a finite form of the auxiliary verb as a copula. In modern standard Lithuanian, formal opposition between the IM and the indicative is neutralized in three forms with past active participles (jis metęs, buvęs metamas, buvęs mestas), in which omission of the finite form of the auxiliary verb is possible. In Latvian, neutralization of this opposition is even more restricted. The modal semantics of the IM comprise the meanings of reportedness (translative), doubt (dubitative), unexpectedness-surprise (admirative), and inference from results (conclusive), with the closely interrelated meanings of reportedness and doubt being predominant. The distribution of the IM in Lithuanian dialects is highly uneven: it is widely used only in the Samogitian dialect, in the southern part of the Western Aukštaitian dialect adjoining it, and in the northeastern corner of the Lithuanian-speaking area (see Map 1). This same area is also characterized by the productivity of other participial constructions, including nominativus cum participio, dativus absolutus, and periphrastic verb-tense forms. By the beginning of Lithuanian and Latvian writing, the IM had become a fully established category with semantics corresponding to those of the modern language. In Old Lithuanian, the IM is most extensively attested in the works of S. Vaišnoras (35 constructions). Dialectal, folkloric, and Old Lithuanian textual evidence points to a close connection between the IM and periphrastic verb forms in the past. Active participles of the present tense- 52 einąs, bėgas, žiūris in Old Lithuanian, and partly also in modern Samogitian, still belong to the paradigm of periphrastic verb forms; they can be used in the position of verbum finitum without a reported or dubitative meaning and can combine with an auxiliary verb, while the affix be- merely emphasizes durativity (e.g., ira beszindans BrB II Moz 34, 26). On the other hand, the IM is inextricably connected with other predicative constructions of active participles without a copula in the position of verbum finitum, which occupy an isolated position in the syntactic system of Lithuanian, become idiomatic, and are already identified in synchronic study as relics of an archaic character. They are used chiefly to express “timeless” states, that is, in the ordinary function of ancient nominal sentences, e.g.: O ans jau nuo ryto atsigulęs į bedribss Rs, as well as for the emphatic expression of negation (e.g. Nebebuve, kad aš až jo eič! Ds), wishes, and commands in semi-idiomatic expressions, incantations, and the like, e.g. atadara durytėles BsP II 327. With various expressive nuances, active participles without a copula also occur as the second verb in postpositive constructions with relative pronouns and adverbs kas, kiek, kada, kur, iki, ligi and the like, e.g.: vaikas turėjo duonos kiek noris, while in preposition they may be linked by the verbal conjunction in, e.g. pabėgėjęs ir pastipo BrP IV 235. The antiquity of these constructions is confirmed by exact correspondences in Slavic, Germanic, and other related languages. Postpositive past participles are also used to emphasize an action expressed by a verb with the corresponding root, e.g., gerk gėręs, as well as to express an alternative (sakyк jam nesakęs) or an undesirable action (tylėk nebybėjęs). The genetic connection between the IM and the relic constructions mentioned above indicates that, from a historical point of view, the IM is a specialized case of the ancient predicative use of participles, distinguished from other such uses only by certain modal meanings. 2. Constructions with neuter passive participles also attest to the development of the IM from the ancient predicative function of participles, e.g.: (brolio) gyvenama, sėjama; gyventa, sėta or (brolio) būta gyven­ta. In terms of their distribution across Lithuanian dialects, these constructions differ sharply from the IM: they are most productive where the IM is not used, and are almost entirely absent from the Samogitian dialect (see Map 2). However, their modal meanings largely correspond to those of the IM; only the admirative and conclusive meanings are more productive with passive participles. The parallelism of modal meanings despite the different distributions of the IM and passive constructions points to a common development and contradicts attempts to derive the IM directly from the nominativus cum participio construction, which does not admit neuter passive participle forms at all. A comparative analysis of constructions with neuter passive participles such as brolio gyvenama / gyventa leads to the conclusion that historically they are variants of the constructions (rugiai) sėjama / sėta (cf. Old Prussian grīkai ast... etwīerpton, Old Church Slavonic glasъ... uslyšano budetъ), built according to the same structural pattern as constructions with adjectives such as (darbas) miela (cf. Old Russian зрохо сладко). The neuter form of adjectives and participles here consists of a bare stem in -o. Constructions such as (rugiai) sėjama are explained as relics of nominal sentences with an uninflected predicative constituent, inherited from the Indo-European linguistic community. Corresponding IM constructions with neuter forms of active participles have also preserved traces of ancient incongruence: they may combine with nouns of either masculine or feminine gender (cf. broliai / seserys eina) and are likewise rooted in the ancient nominal sentence. Thus, the hypothesis is advanced that the IM and passive participle constructions share a primary syntactic structure. The spread of the genetivus auctoris (which apparently originally had a possessive meaning), or of case constructions synonymous with it for marking the subject with passive participles of both transitive and intransitive verbs (e.g.: brolio sėjama, gyvenama), is considered a later transformation of the primary construction (rugiai sėjama), with typological parallels in Slavic (especially Northern Russian) and Indo-Iranian languages. In connection with the transformation of adjectives and participles of the neuter 53 …of gender into adverbialized predicates, within the passive construction the syntactic relations are reinterpreted, as a result of which the nominative in many cases is no longer understood as the subject case, but as the case of the direct object of the action expressed by the participle; cf. jo rašoma laiškas (a secondary “ergativization” of the construction). 3. The development of modal meanings in predicative constructions with active and passive participles is conditioned by a whole complex of internal and external factors. After most nominal sentences had been incorporated into the syntactic paradigms of verbal sentences as zero-copula constructions, some of the ancient predicative participles found themselves at the periphery of the syntactic system and were subsequently reinterpreted on the basis of their secondary—modal—function. The tendency to use nominal forms of the verb to express indirectly perceived action (the category of evidentiality, in R. Jakobson’s terminology) is indicated by the use of both ancient and new (periphrastic) forms of the perfect in Ancient Indian, Slavic, and Romance languages, as well as by typological parallels in Turkic, Semitic, and some Kartvelian languages. The modal reinterpretation of predicative participles may also have been facilitated by the use of participial forms to form the periphrastic optative, that is, precisely the mood which in many Indo-European languages expresses indirect speech. Data from Old Lithuanian texts attest to a connection between optative and translative meanings; cases of synonymy between the KN and the optative in the works of S. Vaišnoras also point to this. An even greater influence on the “attachment” of the meaning of reportativity to predicative active participles must have been exerted by the construction nominativus cum participio after verbs of mental perception and speech, e.g., jis sakė(si) ėjęs; this explains the marked predominance of the translative and dubitative meanings of the KN compared with the corresponding passive constructions. The external conditions that fostered the development of modal meanings in predicative participles should be regarded, above all, as the historical relations between the Baltic and Finnic languages. Comparison of Lithuanian-Latvian participial constructions with semantically corresponding Estonian (kaudne kõneviis) and Livonian constructions points to a shared semantic category of reportativity, which, in a comparatively small and closed area occupying part of the eastern coast of the Baltic Sea, is realized by various deverbal forms: participles (Lithuanian), converbs (Latvian, Estonian), infinitives (Estonian), and deverbal nouns (Livonian), although only the Lithuanian and Latvian forms are genetically identical. The spread of this semantic category may be connected with the advance of the Baltic tribes northward into regions inhabited by the predecessors of the Livonians and Estonians, which is dated to the 5th–6th centuries AD. By this time, Finnic linguistic unity no longer existed, while the East Baltic languages still formed a community within which new phenomena could spread freely. The frequency of use of the KN in the East Baltic languages roughly corresponds to the degree of their contact with Finnic languages: it is most frequent in Latvian, less frequent in northern and northeastern Lithuania and in the Žemaitian dialect, that is, in the contact zones of Lithuanian with Selonian and Curonian, which were more strongly affected by Finnic influence; in southern Lithuania the KN is already very rare. This allows us to assume that the category of reportativity spread southward in the Baltic languages from the north and probably did not even reach the territory of Old Prussian. A typological parallel to the East Baltic and Finnic KN is observed in the Balkans. The Bulgarian-Macedonian reportative forms and the Albanian admirative, whose formation was facilitated by the influence of Turkish, correspond to the Lithuanian-Latvian KN both in their semantic structure and in their morphological means of expression (deverbal nouns closely associated with periphrastic perfect forms). The grammaticalization of the semantic category of reportativity only in peripheral bands of the Indo-European area, based on the ancient predicative function of verbal nouns characteristic of many Indo-European languages, points to the considerable role of language contact in this process and provides a striking example of the spread of identical semantic categories in contact areas of genealogically unrelated languages. 54